• segregation;
  • social polarization;
  • immigration;
  • Athens;
  • Southern Europe


  1. Top of page
  2. Abstract
  3. Introduction: segregation, polarization and contextual difference
  4. Segregation, social polarization and inequality in Athens
  5. Making sense of decreasing segregation within increasing social polarization and income inequality
  6. De-segregation reappraised
  7. Conclusion: the importance of context
  8. References

This article investigates trends in segregation, polarization and inequality in Athens during the 1990s, and focuses on what may seem a paradoxical coexistence of decreasing segregation with increasing social polarization and inequality. To explain this coexistence, the article examines the distinction between social polarization in specific contexts and the dominant assumptions about social polarization, which derive from the global city thesis, and the structure of the city's housing market, which prevented the substantial wave of immigration during the 1990s from intensifying segregation. Arguing for more context awareness, the article does not try to stress the specificity of the Athenian case, but, mainly, to reveal the context-dependent character of the dominant assumptions about segregation and social polarization and, therefore, to show their limitations.

Introduction: segregation, polarization and contextual difference

  1. Top of page
  2. Abstract
  3. Introduction: segregation, polarization and contextual difference
  4. Segregation, social polarization and inequality in Athens
  5. Making sense of decreasing segregation within increasing social polarization and income inequality
  6. De-segregation reappraised
  7. Conclusion: the importance of context
  8. References

This article investigates segregation, polarization and inequality trends in Athens during the 1990s and focuses on what may seem a paradoxical coexistence of decreasing segregation with increasing social polarization and inequality.

The idea behind this investigation is that the relations between these social and spatial trends may be more complex than was envisaged in the old Chicago School idea, according to which social relations can be read through their spatial attributes.1 The context dependence of this assumption2 was overlooked, owing both to the overwhelming dominance of that particular paradigm and to the strong modernist belief that if things were not happening in the same way in different parts of the world, then the disparities would eventually disappear as development stages converged.3

The same kind of contextual ‘blindness’ is also characteristic of more recent theoretical contentions. The social impact of economic restructuring in major world cities, and especially in the leading urban centres of the advanced capitalist world, is customarily summarized as social polarization, following work initiated in the early 1980s (Friedmann and Wolff, 1982) and subsequently developed by Saskia Sassen (1991; see also Mollenkopf and Castells, 1991). According to this line of thought, economic restructuring triggers changes in global city labour markets that induce social polarization, and state regulation is considered increasingly unable to act against such an outcome. An increase in segregation becomes part of this polarization dynamic as a spatial impact of the polarization processes taking place in the labour market, being linked to the increasing social segmentation between holders of different kinds of job and to the catalytic effect of increasing immigration. For Sassen, this ‘spatial polarization’ is epitomized by gentrification for/by the new job elite, the massive appropriation of prime urban space through luxurious development for financial and other leading businesses and the parallel growth of segregation for low-end job holders, minorities and immigrants (Sassen, 1991).

Although such phenomena and processes are undoubtedly present to some degree in most leading world cities, this dominant approach imposes an explanatory pattern on situations that in reality are much more varied. Social polarization has been contested both as a tendency and an outcome, even in several global cities. Hamnett (1994; 1996) has stressed the importance of the welfare state in shaping the outcome of economic restructuring in terms of polarization and provides empirical evidence to support the idea that the situation in London is characterized by professionalization rather than polarization, while Preteceille (1995; 2006) has argued the same about Paris. Both are in fact claiming that the polarization thesis is too much influenced by, and constitutes a generalization of, the realities of major US cities like New York and Los Angeles.4

Moreover, following Sassen, segregation is presented, to a far greater extent than it really is, as a corollary of economic restructuring and a spatial expression of increasing polarization. The relation between social polarization and segregation is, in fact, more complex. Polarization is not a prerequisite for segregation. The two may change in relatively independent ways since the growth of segregation is primarily dependent on the mechanisms that allocate residential space to different social groups, rather than on the degree of social polarization or even on the range of income inequality. This is especially true when the land and housing markets are regulated in a spirit of decommodification (Musterd and Ostendorf, 1998) or when marketization is not strong enough to override the influence of other factors (such as traditional family structures) in the allocation of residential space (Allen et al., 2004).5 The social polarization thesis leaves no room for substantial input by elements not immediately related to the market, and for the most part it minimizes the importance of politics and policies aiming at controlling polarization, segregation and the other socially divisive processes that are built into the operation of market mechanisms.

The shortcoming of the polarization thesis that is most relevant to my present discussion, however, relates to its contextual origins, which are often forgotten.

The growth of polarization is argued on the basis of changes in the global city context, involving the spatial concentration of financial activities and high-end producer services that give these cities their leading edge as world management centres. Social polarization, it is said, follows from the decline of industrial activities that used to provide jobs in the middle of the social hierarchy, which, in turn, gave people an average income and status. These jobs have now been replaced by more polarized jobs in the service sector (Sassen, 1991).

Regardless of the validity of this thesis, its paramount diffusion is giving the impression that polarization and segregation are growing throughout the urban world, even if in most urban areas of the Western world the socially polarizing mechanism assumed for global cities is not present. Both the social polarization thesis and the ‘underclass’ debate (Wilson, 1987; Massey and Denton, 1993; Mingione, 1996; Wacquant 1997) have contributed to creating powerful metaphors of divided cities and societies. These have subsequently become emancipated from analysis of the mechanisms and contexts that generated them. This has happened because theory is context-dependent, but not necessarily context-confined, since it is frequently projected as an explanatory overlay on different realities. The social polarization thesis has thus created a dominant way of seeing urban society extending far beyond the global city context. By producing an implicitly context-indiscriminate and seemingly unquestionable link between economic restructuring and segregation, this way of seeing reduces the need to explain segregation in specific contexts and terms, and takes for granted the existence of common generating mechanisms and processes.6

The context-indiscriminate perception of urban social reality through the polarization frame is partly related to the lack of rigour characterizing its widespread use. Thus, although the vast majority of European cities cannot pretend to be global cities, social polarization is widely used to describe the impact of de-industrialization on their social structure and especially on its lower segments. Polarization, in this case, is not understood stricto sensu as the dual process of increasing numbers at both ends of the social spectrum; a rather confused version of the polarization thesis, more linked to the metaphor of the divided city than to the analysis itself, identifies polarization with the accumulation of problems and the entrapment of people in the lower echelons of the social scale and in deprived areas. The impact of economic restructuring on the post-industrial city of Western Europe offers some recognizable elements that can be perceived as polarizing and segregating under this broad view: loss of employment and redundancy in the labour market; proliferation of professional and managerial jobs fuelled by the increasingly educated cohorts; increased gentrification within inner cities; increasing flows of immigration giving more visibility to the lower occupational categories; increasingly visible spatial concentration of ethno-racial difference. Consequently, social polarization and segregation are often used interchangeably and it seems, as Wessel (2001: 891–2) claims concerning Sassen and Castells, that even the major proponents of the polarization thesis are sometimes using this terminology in a loose way.

Apart from the loose and empirically undocumented use of social polarization, its relation to segregation does not appear as straightforward as the ‘spatial polarization’ argumentation suggests (Sassen, 1991). Preteceille's recent work on the Paris region (2003; 2006) shows a certain polarization of the socially extreme types of residential space whose profiles became more sharply defined between 1990 and 1999. This is happening, however, in a broader Parisian context where the socially intermediate type of residential space is growing either through population increase or through the transformation of other types of (mainly working-class) space to socially intermediate space. Andersen (2004) has shown that, although policies of the Danish welfare state prevent polarization, the housing policies that have developed lately enhance segregation. Leal (2004) has shown that, although there is seemingly a decrease in segregation in Madrid due to changes in the occupational structure and mainly to the increasing numbers in the higher occupational categories, there is increasing segregation in terms of housing prices and income. In Athens, as I try to show in what follows, segregation has decreased since the early 1990s although polarization has increased during the same period. These varied patterns indicate that segregation may be produced otherwise than as an immediate corollary of social polarization and that tackling it in effective ways entails a clearer grasp of its different forms and generating mechanisms within specific contexts.

The varied forms and processes of segregation in European cities are often difficult to identify if one follows the dominant assumptions about forms of segregation and uses the measuring tools that were tailored to those assumptions. Segregation presumes, in fact, a horizontal pattern of socially diversified neighbourhood communities. This horizontal pattern and the fairly distinct socio-racial boundaries between different neighbourhood communities can be easily associated with the context in which the concept of urban segregation was generated, i.e. the fast-growing industrial metropolis of the American mid-West with its low-rise suburban sprawl, strong wave of immigration and heritage of racism from slavery and civil war.7 In that context, segregation meant spatial isolation with spatial distance more or less equating to social distance. Through its reference to a specific context — which progressively became implicit — segregation has inevitably become problematic to use in substantially different contexts.

Segregation, social polarization and inequality in Athens

  1. Top of page
  2. Abstract
  3. Introduction: segregation, polarization and contextual difference
  4. Segregation, social polarization and inequality in Athens
  5. Making sense of decreasing segregation within increasing social polarization and income inequality
  6. De-segregation reappraised
  7. Conclusion: the importance of context
  8. References

The following investigation of segregation and social polarization in Athens during the 1990s is not meant to draw attention to the peculiarity of an ‘unusual’ context, but rather to reveal the limitations and the context-related character of concepts and tools that we often consider context-independent.

Athens is a regional metropolis of four million people in South-eastern Europe. Its growth was spectacular during the first three post-war decades (Kotzamanis, 1997), when its population more than doubled, as it became the principal destination of an important internal migration. After a period of stability during the 1980s, its population started growing again due to immigration from neighbouring countries and in spite of the very low birth rate that brought Greece to the same level as Italy and Spain (Eurostat, 1996: 18).

Athens is an archetype of South European piecemeal urbanization. Its post-war growth was not driven by industrial development, which specialized in building materials and housing-related consumer goods (Economou, 1987; 1988) and followed rather than led the city's growth. Industry never became the main employer in the city's labour market and was mostly made up of traditional small-scale commodity production units, rather than of large modern industrial plants. The city's growth momentum was mainly fuelled by exogenous factors, like the crisis of the agrarian economy and post-civil-war repression, which pushed people away from the countryside, rather than by a burgeoning industrial labour market that attracted outsiders with the promise of stable and well-paid jobs and the amenities (like housing) that such jobs usually entail in other contexts (Allen et al., 2004). The incoming population was mostly employed in construction, personal and domestic services and artisan works.

The irregularity and precariousness of employment gave housing a pivotal role in the process of integrating rural migrants into urban society. Housing had to be secured first — usually through homeownership, even if this entailed self-help and, often illegal, self-construction leading to poor housing conditions — in order to make the difficult and incremental process of integration into the labour market feasible. At the same time, the abundance of the labour force (a very substantial wave of emigration to the Fordist labour markets of Western Europe developed during the peak of urbanization in Southern Europe) contributed to the residual development of local welfare states and the consolidation of family-centred welfare regimes monitored by the dominant clientelist political systems. The absence of the rationalizing effect of substantial industrial development and a developed welfare state (which impose their rationale both through the physical presence of their plants — factories, hospitals, schools, etc. — and the planning of their functionality — zoning, transport, commuting, etc.) is clearly evident in Athens' piecemeal plan where any such rationale is very difficult to locate and decipher. The piecemeal constitution of the urban fabric, a product of the unplanned aggregation of individual housing operations, especially during the first three post-war decades of intense city growth, has had important repercussions on the form and intensity of residential segregation thereafter.

Due to historical circumstances, Athens has been a rather introverted national metropolis. Traditionally difficult Greek–Turkish relations and Cold-War isolation from Greece's northern neighbours during most of the post-war period prevented Athens from assuming a more important role in its broader region. This is still reflected in the paucity of local and foreign corporate investment related to such a role and, thus, in the city's relegation to the lower ranks of metropolitan importance (Cattan et al., 1994: 89–101; Leontidou-Gerardi, 2004: 18-28; see also Taylor and Hoyler, 2000, Economou et al., 2001). As a result, the Greek capital not only does not have the profile of a global city, it also lacks the regional role that being an EU metropolis of its size would normally entail. This long introversion had a significant impact on its labour market and particularly on the profile of its higher occupational categories (managers and professionals). Consequently, the polarization of its occupational structure since the early 1990s, should not be interpreted at face value, as I will try to show in the following.

Decreasing segregation

No significant conclusions can be drawn from direct comparison of the level of racial and social segregation in different cities on both sides of the Atlantic, as attempted in Musterd (2005). It is true, however, that the level of class segregation in most European cities is much lower than the level of ethno-racial segregation in the hypersegregated American metropolises. In Chicago, for instance, only a small minority of census tracts are not inhabited by an absolute majority of a major ethno-racial group — whites, African Americans, Hispanics or Asians (Garner et al., 2007). In contrast, the broad mapping of class segregation in Athens for 2001 (see Figure 1) shows that in 70% of census tracts none of the three major occupational classes8 (see Table 1) contains more than 50% of their residents. It is clear that a very important part of the city's residential areas are mixed or relatively mixed in terms of social class.


Figure 1. Census tracts in the Attica region* with more than 50% of the working population belonging to one of the three major occupational classes (2001) * The population of the Greater Athens Area covers more than 95% of the population of Attica.

Download figure to PowerPoint

Table 1.  Three major class groups in the Attica region (2001) based on the European Socio-economic Classes (ESeC)*
i Large employers, professional, administrative and managerial occupations (ESeC 1 & 2)24.7
ii Intermediate occupations, lower technician occupations, small employers and self employed occupations (ESeC 3, 4, 5, 6)36.1
iii Lower white-collar occupations, skilled and routine occupations (ESeC 7, 8, 9)39.2

Moreover, about 75% of the total employed population live in socially mixed areas. Only 10% of the higher socio-economic groups live in areas where they hold the absolute majority; for intermediate groups this percentage is very low (2%) while for lower categories it is much more substantial (31%), although this is partly due to their bigger overall size (see Table 2).

Table 2.  Distribution of major occupational classes by social type of census tract
Social type of census tractTractsClass % by type of census tract
Number%Class iClass iiClass iiiTotal
Large employers, professional, administrative and managerial occupations (i) >50% of census tract's employed population190510314
Intermediate and lower technician occupations, and small employer and self employed occupations (ii) >50% of census tract's employed population20951211
Lower white-collar and lower technical and routine occupations (iii) >50% of census tract's employed population8252011183121
No major class group >50% of census tract's employed population2,8517078776774

This rather low level of class segregation decreased during the 1990s. Indices of dissimilarity decreased between 1991 and 2001 for all occupational classes with respect to both the higher-grade professionals and the lower technical occupations in industry and construction (see Table 3).9 The greater decline in percentage points — but also in relative terms — is observed between social extremes.

Table 3.  Dissimilarity indices (IDs) for ESeC classes in Athens
ESeCIDs in respect to higher grade professionalsIDs in respect to lower technical occupations in industry and construction works
(1) Higher professionals and administrators3.84.40.648.643.4−5.2
(2) Lower professionals and administrators19.916.8−3.237.832.8−4.5
(3) Intermediate professions25.724.1−1.630.826.4−4.5
(4) Small self-employed (excluding agriculture)32.729.1−3.622.222.70.5
(5) Self-employed in agriculture65.959.6−6.358.155.5−2.6
(6) Lower supervisory and lower technician occupations34.929.6−5.330.029.8−0.2
(7) Lower services, sales & clerical occupations32.731.9−0.824.119.1−4.9
(8) Lower technical occupations47.740.8−
(9) Routine occupations46.439.6−6.813.414.10.7

It is rather remarkable that indices of dissimilarity are higher for the native-Greek population than for the total population, which includes the sizeable wave of recent immigrants. Moreover, it is the lower occupational categories, to which the bulk of recent immigrants has been relegated, that present the greatest decrease in dissimilarity from the spatial distribution of the higher occupational group (see Table 4). Immigration, and recent immigration in particular, is usually related to increasing segregation through the relegation of immigrants to the lower tier of the housing market, often leading to their spatial concentration in downgraded housing estates or other low-quality housing. It seems, therefore, paradoxical that in Athens the massive arrival of immigrants has contributed to decreasing segregation in the 1990s.10

Table 4.  Dissimilarity indices (IDs) for ESeC classes in Athens (2001). Total active population and Greek population
 IDs in respect to higher grade professionals
Within the total active population (1)Within the Greek population (2)(2)/(1)
(1) Higher professionals and administrators/managers4.44.40.0
(2) Lower professionals and administrators/managers16.817.00.2
(3) Intermediate24.124.20.1
(4) Small self-employed (excluding agriculture)
(5) Self-employed in agriculture59.664.95.3
(6) Lower supervisory and lower technician occupations29.629.60.1
(7) Lower services, sales & clerical occupations31.932.40.4
(8) Lower technical occupations40.842.21.4
(9) Routine occupations39.642.73.2

Increasing social polarization

Decreasing social segregation during the 1990s is strangely combined with increasing social polarization. Both extremes of the occupational class hierarchy have increased their shares in the active population (see Table 5).

Table 5.  Change in major occupational class groups 1991–2001
 1991 (% of total)2001 (% of total)1991/2001 (% change)
(I) Large employers, professional, administrative and managerial occupations ESeC 1 & 220.924.718.3
(II) Intermediate and lower technician occupations ESeC 3 & 622.120.09.5
(III) Small employer and self employed occupations ESeC 4 & 523.416.131.2
(IV) Lower white-collar and skilled occupations ESeC 7 & 821.620.93.4
(V) Routine occupations ESeC 912.018.352.8

The higher occupational pole, with its comparatively moderate growth, has continued the growth trend it had also presented in the 1980s. The very significant growth of the lower pole, on the other hand, has reversed the serious decline it presented in the 1980s (Emmanuel, 2002) and this is almost exclusively due to the immigrant inflow of the 1990s that raised the percentage of foreigners in the active population from 2.9% to 13.6% between 1991 and 2001 (see Table 6).

Table 6.  Change in major occupational class groups by nationality 1991–2001
GreeksForeign citizens fromGreeksForeign citizens from
Developed economy countriesOther countriesDeveloped economy countriesOther countries
(I) Large employers, professional, administrative and managerial occupations20.634.416.126.640.85.9
(II) Intermediate and lower technician occupations22.323.88.522.023.42.9
(III) Small employer and self-employed occupations24.
(IV) Lower white collar and skilled technical occupations21.116.043.819.413.132.7
(V) Routine occupations11.84.919.614.26.752.0
Total in active population97.

Growing income inequality

Income inequality also increased along with social polarization in the 1990s. According to data from household budget surveys, in the second part of the 1980s and up to the early 1990s, income inequality remained unchanged, or even slightly decreased. The highest decile in the income distribution (measured using average monthly consumption as a proxy) drew slightly nearer the average income, and the total range of income inequality remained unchanged (Emmanuel, 2004). But from the early to the late 1990s the situation changed: both the lowest and highest deciles increased their distance from the average income and the total range of inequality increased (see Table 7).

Table 7.  Income inequality: decile ratios of average monthly consumption per adult equivalent in Attica
YearDecile Ratios
1st/ All households10th/ All households1st/ 10th
  1. Source: Household budget surveys. Calculations by D. Emmanuel


Increasing inequality has affected issues directly related to segregation, like housing. In Greece, like the rest of Southern Europe, the weak welfare state has rendered home ownership an important asset across all class lines, especially due to deficient provision of social rented housing (Allen et al., 2004). The 1990s, however, introduced increasing social differentiation into access to home ownership. Between 1991 and 2001, the percentage of those living in rented dwellings increased for the lower socio-economic categories while it decreased for the higher and intermediate ones (see Figure 2; see also Emmanuel, 2004).


Figure 2. Percentage of tenants per major class group

Download figure to PowerPoint

Access to housing has been increasingly socially differentiated, not only in terms of tenure, but also in terms of housing age, space per capita, quality and comfort. In most cases, increasing inequality in housing is mainly due to the concentration of immigrants in the older, private rented, less equipped and smaller housing units (see Figures 3, 4 and 5).11


Figure 3. House tenure by nationality in Attica 2001

Download figure to PowerPoint


Figure 4. Average housing surface per capita (in m2) by nationality in Attica 2001

Download figure to PowerPoint


Figure 5. Average housing age (in years) by nationality in Attica 2001

Download figure to PowerPoint

Making sense of decreasing segregation within increasing social polarization and income inequality

  1. Top of page
  2. Abstract
  3. Introduction: segregation, polarization and contextual difference
  4. Segregation, social polarization and inequality in Athens
  5. Making sense of decreasing segregation within increasing social polarization and income inequality
  6. De-segregation reappraised
  7. Conclusion: the importance of context
  8. References

As I argued in the introduction, the pervasiveness of the socio-spatial polarization imagery of urban social reality related to the assumptions of the global city thesis, makes the desegregation trend that characterizes Athens in the 1990s a paradoxical occurrence since this period was also one of social polarization and increasing inequality. Moreover, the 1990s were a period of intense immigration; and usually the presence of immigrants and ethno-racial difference engenders more segregation, as the growth of immigrant numbers and their spatial concentration in parts of European cities is often taken as a sign of their progressive Americanization.12 It is, therefore, even more intriguing that class segregation seems to be less important within the total — compared to the Greek — population (see Table 4 above), suggesting that in Athens the presence of immigrants reduces rather than exacerbates class segregation.

How, then, can we explain the fact that during the same decade there was increasing social polarization, booming immigration and increasing inequality and, at the same time, decreasing segregation?

The profile of the growing higher occupational pole and the spatial impact of its growth

The first element in the explanation relates to the character of social polarization in Athens, which does not follow the global city model. According to the latter, the growth of the upper social pole is due to the attraction of a corporate elite; job holders with lower status potential are also attracted to take service jobs catering to the elite and low-end corporate posts, while the social middle is depleted due to the loss of stable jobs in Fordist industry. These changes induce a fairly high population turnover, high residential mobility and strong pressure on the housing market from the incoming corporate elite (through gentrification or other forms of spatial appropriation) as well as from other corporate activities in the services reclaiming centrally located urban space. All this leads to the spatial marginalization and segregation of the lower social strata, or to ‘spatial polarization’ in Sassen's (1991) terms.

In Athens, the growth of the upper occupational pole is not particularly related to the formation of a corporate elite. Table 8 shows that the growth of the categories included in the higher occupational pole is not related to higher-grade managers and administrators, but mainly to higher-grade professionals and, secondarily, to lower-grade managers.

Table 8.  Growth of the higher occupational categories 1991–2001
Managers & administrators (higher grade) (ESeC 1)
Professionals (higher grade) (ESeC 1)8.810.620.7%
Managers & administrators (lower grade) (ESeC 2)2.22.510.6%
Professionals (lower grade) (ESeC 2)

The growth in the number of higher-grade professionals is an endogenous phenomenon related to the recent development of higher education: in 1971 the proportion of higher-education graduates in the Greek population over 25 years old was 3.7% (ESYE, 1974); in 2001 it had reached 15.3% (EKKE-ESYE, 2005). This significant increase has produced a certain amount of inflation in higher educational credentials (Duru-Bellat, 2006), leading often to downgraded employment prospects for graduates — employment in small (family) businesses, self-employment and employment in the public sector rather than becoming part of the corporate elite. Among higher-grade professionals in 2001 there are a considerable number (35%) who are self-employed or small employers with less than 10 employees (EKKE-ESYE, 2005). Moreover, the growth in this, the most important and fastest-growing category within the upper social pole, is not particularly taking place in financial services, which usually fuel the growth of the corporate elite. The higher-grade managers in the broader financial and property management sector only increased from 11.2% to 12.8% of the total number of salaried higher-grade managers between 1991 and 2001, while the respective increase in salaried higher-grade professionals in the same sector during the same period was also limited (from 9.4% to 11.1%) (EKKE-ESYE, 2005).

Furthermore, a considerable percentage of the employee members of these categories — almost 30% for the higher-grade managers and almost 40% for the higher-grade professionals — works in government and social services and therefore does not form part of the corporate elite (EKKE-ESYE, 2005).

Table 9 illustrates the fact that the growth of the higher occupational pole remained, almost exclusively, a local-Greek affair — in stark contrast to what happened to the lower pole — and therefore the pressure from an incoming elite on the housing market has been minimal. At the same time, the local elite's pressure on this market has also been relatively reduced due to the traditionally low level of residential mobility even for the higher social strata (Allen et al., 2004; Maloutas, 2004).

Table 9.  Percentage composition of ESeC classes in terms of citizenship 1991–2001
GreeksForeign citizens fromGreeksForeign citizens from
Developed economy countriesOther countriesDeveloped economy countriesOther countries
(1) Large employers, higher grade professional, administrative and managerial occupations97129433
(2) Lower grade professional, administrative and managerial occupations and higher grade technician and supervisory occupations97129523
(3) Intermediate occupations98119722
(4) Small employer and self-employed occupations (excluding agriculture)98119415
(5) Self-employed occupations (agriculture)98029118
(6) Lower supervisory and lower technician occupations98029513
(7) Lower services, sales and clerical occupations96139019
(8) Lower technical occupations950571128
(9) Routine occupations960466134

Finally, the content of the ‘vanishing middle’ is also different from the global city model. In Athens, the occupational middle was not depleted by a reduction in stable and specialized working-class jobs in Fordist industries — these were never particularly well developed anyway — but by a reduction in the numbers of small employers and the self-employed, mainly in artisan works and trade. Moreover, the significant decrease in the numbers of these categories has not involved major repositioning movements in residential space, since in most cases this trend reflects the ageing of these categories and their progressive replacement in the labour market and does not involve their abrupt residential displacement as, in most cases, their members are spatially fixed for life in traditional homeownership structures (i.e. in loan-free properties).

Therefore, both the growth of the higher occupational pole and the ‘vanishing middle’ in Athens did not instigate any significant processes of spatial polarization and segregation. However, while this may explain why segregation did not increase in the 1990s, it does not explain its decreasing trend.

The spatial impact of the growing lower occupational pole

The explanation for decreasing segregation within this context of social polarization, booming immigration and increasing inequality lies in the shape of housing provision in conjunction with the place recent immigrants have occupied within it.

Contextual parameters may prevent theoretically expected outcomes from materializing. Hamnett (1996) and Preteceille (1995) have demonstrated this in the cases of London and Paris respectively, where local welfare regimes prevented the materialization of social polarization according to the global city thesis. Andersen (2004) has shown the same thing in Copenhagen, where despite the effective protection of the welfare state against the growth of social polarization and inequality, segregation has increased following changes in housing policy.

In Athens, the weak welfare state with its minimal housing component would in principal entail free rein for the market to shift and sort its clients, especially the new immigrants destitute of the traditional protective assets of the local population — that is destitute mainly of family self-help networks and of home ownership. And in fact it did so, which is why immigrants are in the oldest, cheapest and least comfortable part of the privately rented housing stock, as indicated earlier (see Figures 3, 4 and 5 above).

However, although this housing stock is spatially concentrated — particularly in the densely populated areas around the city centre (see Figure 6) — immigrants are not spatially isolated from higher and middle occupational groups. This is happening because they live in condominiums that are vertically segregated, with the upper floors containing more spacious apartments occupied by local intermediate and upper-intermediate groups and the lower floors, with smaller and darker apartments, occupied by immigrants and indigenous lower social strata (Leontidou, 1990; Maloutas and Karadimitriou, 2001).


Figure 6. Relative concentration of immigrants in Athens' residential space by census tract (2001)

Download figure to PowerPoint

Figure 6 shows the concentration of immigrants in the areas of vertical social differentiation around the centre that does not lead to increased segregation. In fact, the percentage of immigrants living in the central municipality of Athens is considerably higher (37.3%) than for the native population (17.8%). Immigrants are also living in some parts of the city's periphery, especially when they are engaged in agriculture and building activities. Peripheral immigrant settlement also involves almost exclusively privately rented housing, with no major spatial concentrations and no segregation that could have resulted from social or other forms of collective housing provision. Immigrants in peripheral residential areas are usually located near their place of work — commuting is much less important for immigrants than for Greek nationals (Kandylis, 2005) — in small family businesses (Sayas, 2004) or private households involving often residential proximity to their employer.

Nevertheless, each ethnic immigrant group has a more or less distinctive residential pattern (see Figures 7, 8, 9 and 10).


Figure 7. Relative concentration of Albanian immigrants in Athens' residential space by census tract (2001)* * The percentage of Albanians in the total population of Attica in 2001 was 5.5%.

Download figure to PowerPoint


Figure 8. Relative concentration of Bulgarian immigrants in Athens' residential space by census tract (2001)* * The percentage of Bulgarians in the total population of Attica in 2001 was 0.3%.

Download figure to PowerPoint


Figure 9. Relative concentration of Filipino immigrants in Athens' residential space by census tract (2001)* * The percentage of Filipinos in the total population of Attica in 2001 was 0.2%.

Download figure to PowerPoint


Figure 10. Relative concentration of Pakistani immigrants in Athens' residential space by census tract (2001)* * The percentage of Pakistanis in the total population of Attica in 2001 was 0.3%.

Download figure to PowerPoint

Albanians (the largest immigrant group in Athens comprising more than 60% of the immigrant population) follow — or rather determine — the general immigrant pattern of residential location, with a comparatively strong presence at the periphery, since they specialize in building and agriculture and less in services.

Bulgarians, the second largest immigrant group in Athens, are mainly middle-aged women working in personal services (and especially taking care of elderly Greeks); they are concentrated in densely populated areas around the city centre where upper-intermediate elderly people are abundant and housing is affordable.

Filipinos are also a group made up almost exclusively of women, and have the reputation of being the best domestic help; they are highly present in the most expensive residential areas of the city (in the north-east) where they work as live-in personnel; those who have day jobs, are mainly concentrated in a particular area of vertical segregation near the city centre, where they have established schools and other community facilities.

Pakistanis, finally, are an almost exclusively male group in the labour market, working in small industries, in transport and agriculture and are relatively absent from the centre since their residence follows the location pattern of their activities.

Therefore, despite the clearly distinctive spatial pattern of each immigrant group (and consequently the high segregation indices on separate group level), their overall segregation is not high, either because they live in vertically segregated areas or because they live in or very near their place of work; and their employment is more often in a small family business or a household rather than in a large industrial or other plant,13 both entailing a residential mix with the local population and frequently with their own employers.

De-segregation reappraised

  1. Top of page
  2. Abstract
  3. Introduction: segregation, polarization and contextual difference
  4. Segregation, social polarization and inequality in Athens
  5. Making sense of decreasing segregation within increasing social polarization and income inequality
  6. De-segregation reappraised
  7. Conclusion: the importance of context
  8. References

A last point of discussion relating to the coexistence of polarization, inequality and desegregation trends in Athens during the 1990s is that desegregation should not be taken at face value, and may not, in the end, be a cause for celebration.14 Important negative trends may be dissimulated within the aggregate de-segregation trend conveyed by decreasing dissimilarity indices, if these account for trends within particular parts of the broader area under scrutiny.

In order to overcome this shortcoming, I used a different approach to inquire into meaningful differences in segregation trends within particular social types of residential space in Athens. The first step was to create a social typology of the city's residential space and then to measure the social profile changes in each type during the 1990s.

A classification of the city's census tracts into five different social types was chosen to convey a synthetic image with rather clearly demarcated class features between the different types/clusters (see Table 10).15 According to this classification, 10% of the city's residents live in areas dominantly inhabited by higher and 22% by higher and intermediate occupational categories (clusters 1 and 2), while 17% reside in areas mainly inhabited by lower ones (cluster 5). The rest — practically the other half — live in socially mixed areas (clusters 3 and 4), which sometimes appear slightly polarized as they contain increased shares of both higher and lower occupational categories (cluster 3).

Table 10.  Social typology of Athens' residential space (2001). Class content of five residential clusters
 (I) Large employers, professional, administrative & managerial occupations(II) Intermediate and lower technician occupations(III) Small employer and self employed occupations(IV) Lower white collar and skilled occupations(V) Routine occupationsPercent in active population

The location of the two higher clusters covers most of the eastern part of the city, including the traditional strongholds of the upper socio-economic class in the city centre. The lower cluster covers the western, working-class part of the city and most of the very distant residential areas. The slightly polarized cluster is located around the centre in areas that are densely populated and most characteristic of vertical segregation, and the mixed cluster is mostly, but not exclusively, located in the best residential areas of the western (working-class) part of the city (see Figure 11), which gather the socially — but not residentially — mobile offspring of the local population (Maloutas, 2004).


Figure 11. Social clusters in Athens' residential space. Census tract level (2001)

Download figure to PowerPoint

Figure 12 shows the change in terms of class composition for the five clusters.16 In the first two clusters there was growth for the highest occupational category, stability for the three intermediate ones and decline for the lowest category. The opposite happened to the lower cluster, where there were substantial gains for routine occupations and losses for the higher occupational class. Therefore, the clusters with the higher and lower class profiles (the socially extreme ones) have respectively strengthened their profile. The polarized cluster of vertically segregated areas around the city centre has presented important losses for the higher occupational class and equally important gains in routine occupations, that led it to a more mixed social profile. Finally, the mixed cluster has remained relatively stable.


Figure 12. Percentage point change of major class groups (ESeC based) in five residential clusters controlled by the average change in each group 1991–2001

Download figure to PowerPoint

In summary, the three clusters with the clearer class profiles and comprising approximately 50% of the city's active population have become socially more homogeneous (and thus they have contributed to the increase in segregation); the large mixed cluster has remained stable (and therefore has not affected segregation) and the small mixed cluster has become more mixed and relatively polarized due to filtering-down. Thus, it has contributed to the reduction of segregation in the same way that segregation is reduced in gentrified areas during the first phases of the process.

Therefore, social segregation may be decreasing according to indices of dissimilarity for the city as a whole, but this does not prevent substantial parts of the city (those with the more marked social profile) from becoming increasingly homogenous; at the same time the overall desegregation trend is fuelled by processes such as filtering-down, whose desegregating impact may be circumstantial.

Conclusion: the importance of context

  1. Top of page
  2. Abstract
  3. Introduction: segregation, polarization and contextual difference
  4. Segregation, social polarization and inequality in Athens
  5. Making sense of decreasing segregation within increasing social polarization and income inequality
  6. De-segregation reappraised
  7. Conclusion: the importance of context
  8. References

By investigating the interrelations between segregation, social polarization and inequality in Athens, I have tried to show that they may be different from the dominant assumptions about them, and that the difference should not be simply attributed to the contextual specificity of the Athenian case, but also to the usually forgotten contextuality of the dominant assumptions and, therefore, to their limitation as explanatory devices.

The spatial dimension of social distance is certainly much more complicated than the dominant assumptions about segregation and social mix usually imply, as Chamboredon and Lemaire (1970) first revealed some time ago. Moreover, this complexity is increased by contextual difference: varied and nuanced forms of segregation may reasonably be expected in societies where class relations and not ethno-racial differences are the primordial differentiating element in urban space and where less discrimination and more egalitarian approaches underlie their regulation. Important public intervention through the welfare state and, in particular, through explicit anti-segregation housing policies in several countries around Europe are an important element expected to impede the clear ‘shifting and sorting’ operated by the housing market.17 Different forms of segregation — vertical for instance (White, 1984; Maloutas and Karadimitriou, 2001) — and less socio-spatially homogeneous city quarters may reasonably be expected in rather compact built environments with a substantially longer history, and therefore of a slow and incremental constitution under varying conditions, than in recent suburban expansions developed rapidly and in purely marketized conditions.

Different forms of segregation challenge the way we usually measure the change in segregation intensity and the way we evaluate the positive or negative character of this change. Vertical segregation in Athens, as I have tried to show, is a process that leads to decreasing segregation, but cannot be considered as an unequivocally positive outcome since it is also impregnated with a host of new socio-spatial separations and tensions (Maloutas and Karadimitriou, 2001). The same can be argued about gentrification, which is a process of social redistribution and re-appropriation of residential space in mature urban settings, where change no longer refers primarily to rapid urbanization but to internal reforming. This reforming often involves changes of scale and form for segregation through the diversification of patterns and mechanisms for the social allocation of urban space. Hamnett (2004) illustrates this diversification for East London where gentrifiers gradually replaced the shrinking working class from the 1970s and created more variegated segregation patterns at the micro-scale in the place of former broad socio-spatial divisions. A gentrified area becomes less segregated, in terms of segregation indices, at least for some time. This outcome is often accounted as positive due to the improvement of residential social mix, usually forgetting or minimizing the importance of displaced social groups (Slater, 2006).

The need to distinguish the specific processes of segregation involved in every different case stems from the specific social dynamic of each process which may be temporarily producing lower segregation indices, but may at the same time be leading towards more deprivation and inequality for certain groups. Increase or decrease in segregation indices are, therefore, not unequivocally socially negative or positive, and segregation in its simple definition and operational use is a very insufficient concept to account for the complex spatial dimension of social distance in contemporary urban settings.

The dominant assumptions about the form and the process of segregation, which usually remain implicit, are making this concept problematic when applied in contexts that partly at least contradict them. A more contextualized approach to segregation and the specific processes that structure its different forms should therefore permit us to go beyond the deforming lenses of social polarization and ghettoization, and the first step in this direction is certainly to be aware of the contextual origins and limitations of the deforming lenses themselves.


  1. Top of page
  2. Abstract
  3. Introduction: segregation, polarization and contextual difference
  4. Segregation, social polarization and inequality in Athens
  5. Making sense of decreasing segregation within increasing social polarization and income inequality
  6. De-segregation reappraised
  7. Conclusion: the importance of context
  8. References
  • Abu Lugod, J. (1969) Testing the theory of social area analysis: the ecology of Cairo, Egypt. American Sociological Review 34, 189212.
  • Allen, J., J. Barlow, J. Leal, T. Maloutas and L. Padovani (2004) Housing and welfare in Southern Europe. Blackwell, Oxford.
  • Andersen, H.T. (2004) Spatial – not social polarization: social change and segregation in Copenhagen. The Greek Review of Social Research 113A, 14565.
  • Arapoglou, V.P. (2006) Immigration, segregation and urban development in Athens: the relevance of the LA debate for Southern European metropolises. The Greek Review of Social Research 121C, 1138.
  • Arbaci, S. (2007) Ethnic segregation, housing systems and welfare regimes in Europe. European Journal of Housing Policy, 7.4, 40133.
  • Arbaci, S. (forthcoming) (Re)viewing ethnic residential segregation in Southern European cities: housing and urban regimes as mechanisms of marginalisation. Housing Studies.
  • Baum, S. (1997) Sydney, Australia: a global city? Testing the social polarisation thesis. Urban Studies 34.11, 18811901.
  • Baum, S. (1999) Social transformations in the global city. Urban Studies 36.7, 10951117.
  • Burgers, J. (1996) No polarisation in Dutch cities? Inequality in a corporatist country. Urban Studies 33.1, 99110.
  • Cattan, N., D. Pumain, C. Rozenblat and T. Saint-Julien (1994) Le système des villes européennes. Anthropos, Paris.
  • Chamboredon, J.C. and M. Lemaire (1970) Proximité sociale et distance spatiale. Les grands ensembles et leur peuplement. Revue Française de Sociologie XI.1, 333.
  • Duru-Bellat, M. (2006) L'inflation scolaire. Les désillusions de la méritocratie. Seuil, Paris.
  • Economou, D. (1987) Housing policy in post war Greece. The Greek Review of Social Research 64, 56129 (in Greek).
  • Economou, D. (1988) Land and housing system. In T.Maloutas and D.Economou (eds.), The problems of welfare state development in Greece, Exandas, Athens (in Greek).
  • Economou, D., P. Getimis Z. Demathas, G. Petrakos and J. Pyrgiotis (2001) The international role of Athens. The University of Thessaly Press, Volos (in Greek).
  • EKKE – ESYE [National Centre for Social Research – National Statistical Service of Greece] (2005) Panorama of Greek census data 1991–2001. Application under experimental and restricted use at the EKKE.
  • Emmanuel, D. (2002) Segregation, polarization and inequality in the geography of Athens: the role of the housing market and urban development. Geographies 3, 4670 (in Greek).
  • Emmanuel, D. (2004) Socio-economic inequalities and housing in Athens: impacts of the monetary revolution of the 1990s. The Greek Review of Social Research 113A, 12144 (in Greek).
  • ESYE [National Statistical Service of Greece] (1974) Statistical yearbook of Greece 2003. ESYE, Athens.
  • Eurostat (1996) Social portrait of Europe. Office for Official Publications of the European Communities, Luxembourg.
  • Friedmann, J. and G. Wolff (1982) World city formation: an agenda for research and action. International Journal of Urban and Regional Research 6.3, 30943.
  • Garner, R., B.H. Hancock and K. Kim (2007) Segregation in Chicago, The Tocqueville Review, XXVIII.1, 4174.
  • Hamnett, C. (1994) Social polarisation in global cities. Theory and evidence. Urban Studies, 31, 40124.
  • Hamnett, C. (1996) Social polarisation, economic restructuring and welfare state regimes. Urban Studies 33, 140730.
  • Hamnett, C. (2004) Economic and social change and inequality in global cities: the case of London. The Greek Review of Social Research 113A, 6380.
  • Harrison, E. and D. Rose (2006) The ESeC classes summarized. Paper presented at the European Socio-economic Classification Validation Conference, Lisbon, Institute for Social and Economic Research (available at
  • Häussermann, H. (2005) The end of the European city? European Review 13.2, 23749.
  • Kandylis, G. (2005) Centrality and commuting for natives and immigrants in Athens. Unpublished communication at the Institute of Urban and Rural Sociology seminars, National Centre for Social Research, Athens (March 2005).
  • Kandylis, G., V. Arapoglou and T. Maloutas (forthcoming) Immigration and the ‘competitiveness-cohesion’ dipole in Athens. In P.Ache, H.T.Andersen, T.Maloutas, M.Raco and T. TasanKok (eds.), European cities between competitiveness and cohesion. Discourses, realities and implementation. Springer.
  • Kotzamanis, V. (1997) Athens 1848–1995. The demographic emergence of a metropolis. The Greek Review of Social Research 92/93, 330 (in Greek).
  • Leal, J. (2004) Segregation and social change in Madrid metropolitan region. The Greek Review of Social Research 113A, 81104.
  • Le Galès, P. and M. Zagrodzki (2006) Cities are back in town: the US/Europe comparison. Cahiers Européens du Pôle Ville/metropolis/cosmopolis, Centre d'Études Européennes de Sciences-Po, Paris.
  • Leontidou, L. (1990) The Mediterranean city in transition. Cambridge University Press, Cambridge.
  • Leontidou-Gerardi, K. (2004) Strategic framework for the spatial development of Athens-Attica. Research project synopsis. Department of Urban and Regional Planning, School of Architecture, National Technical University of Athens, Athens (in Greek).
  • Maloutas, T. (2004) Segregation and residential mobility. Socially entrapped social mobility and its impact on segregation in Athens. European Urban and Regional Studies 11.2, 17187.
  • Maloutas, T. and D. Karadimitriou (2001) Vertical social differentiation in Athens: alternative or complement to community segregation? International Journal of Urban and Regional Research 25.4, 699716.
  • Massey, D.S. and N.A. Denton (1993) American apartheid: segregation and the making of the underclass. Harvard University Press, Cambridge, MA.
  • Mingione, E. (1996) Urban poverty in the advanced industrial world: concepts, analysis and debates. In E.Mingione (ed.), Urban poverty and the underclass, Blackwell, Oxford.
  • Mollenkopf, J.H. and M. Castells (eds.) (1991) Dual city: restructuring New York. Russel Sage Foundation, New York.
  • Murie, A. and S. Musterd (1996) Social segregation, housing tenure and social change in Dutch cities in the late 1980s. Urban Studies 33.3, 495516.
  • Musterd, S. (2005) Social and ethnic segregation in Europe: levels, causes and effects. Journal of Urban Affairs 27.3, 33148.
  • Musterd, S. and W. Ostendorf (1998) Segregation, polarisation and social exclusion in metropolitan areas. In S.Musterd and W.Ostendorf (eds.),Urban segregation and the welfare state, Routledge, London.
  • Park, R.E. (1925 1957 ]) The urban community as a spatial pattern and a moral order. Publications of the American Sociological Society XX, 1-14. Reprinted (1957) in E.C.Hughes (ed.), Human communities. The city and human ecology. The collected papers of Robert Ezra Park, Vol II, The Free Press, Glencoe, IL.
  • Preteceille, E. (1995) Division sociale de l'espace et globalisation. Sociétés Contemporaines 22/23, 3368.
  • Preteceille, E. (2003) La division sociale de l'espace francilien. Typologie socioprofessionnelle 1999 et transformations de l'espace résidentiel 1990–99. Observatoire Sociologique du Changement, FNSP-CNRS, Paris.
  • Preteceille, E. (2006) La ségrégation sociale a-t-elle augmenté? Sociétés Contemporaines 62, 6993.
  • Rose, D and E. Harrison (2007) The European Socio-economic Classification: a new social class schema for comparative European research, European Societies 9.3, 45990.
  • Sassen, S. (1991) The global city: New York, London, Tokyo. Princeton University Press, Princeton, NJ.
  • Sayas, J. (2004) An exploration of the social and spatial division of labour in the Athenian urban space. The Greek Review of Social Research 113A, 167200.
  • Shevsky, E. and W. Bell (1955) Social area analysis: theory, illustrative application and computational procedures. Stanford University Press, Stanford.
  • Slater, T. (2006) The eviction of critical perspectives from gentrification discourse. International Journal of Urban and Regional Research 30.4, 73757.
  • Taylor, P. and M. Hoyler (2000) The spatial order of European cities under conditions of contemporary globalization. Tijdschrift voor Economische en Sociale Geografie 91.2, 17689.
  • Timms, D. (1971) The urban mosaic. Towards a theory of residential differentiation. Cambridge University Press, Cambridge.
  • Vaattovaara, M. and M. Kortteinen (2003) Beyond polarization versus professionalization? A case study of the Helsinki region, Finland. Urban Studies 40.11, 212745.
  • Wacquant, L. (1997) Three pernicious premises in the study of the American ghetto. International Journal of Urban and Regional Research 21.2, 34153.
  • Walks, R.A. (2001) The social ecology of the post-Fordist/global city? Economic restructuring and socio-spatial polarisation in the Toronto Urban Region. Urban Studies, 38.3, 40747.
  • Wessel, T. (2000) Social polarisation and socio-economic segregation in a welfare state: the case of Oslo. Urban Studies 37.11, 194767.
  • Wessel, T. (2001) Losing control? Inequality and social division in Oslo. European Planning Studies 9.7, 889906.
  • White, P. (1984) The West-European city. A social geography. Longman, London.
  • Wilson, W.J. (1987) The truly disadvantaged: the inner city, the underclass and public policy. Chicago University Press, Chicago.
  • Earlier drafts of this article were presented at the XVIth ISA congress in Durban (July 2006) and in a series of talks to several American universities (Brown, Graduate Center–CUNY, Berkeley, Buffalo, De Paul and Michigan) sponsored by the Alexander Onassis Foundation (October and November 2006). I would also like to thank Edmond Preteceille and two anonymous referees for their critical comments.

  • 1

    ‘It is because geography, occupation, and all the other factors which determine the distribution of population determine so irresistibly and fatally the place, the group, and the associates with whom each one of us is bound to live that spatial relations come to have, for the study of society and human nature, the importance which they do. It is because social relations are so frequently and so inevitably correlated with spatial relations; because physical distances so frequently are, or seem to be, the indexes of social distances, that statistics have any significance whatever for sociology. And this is true, finally, because it is only as social and psychical facts can be reduced to, or correlated with, spatial facts that they can be measured at all’ (Park, 1925[1957]: 177).

  • 2

    For several decades this ‘reading’ followed the socio-spatial patterns identified in the booming metropolis of the American mid-West in the early 20th century, codified by the Burgess model and its subsequent variations.

  • 3

    When, for instance, fewer ‘dimensions’ of residential segregation were found in the factorial ecology of Cairo (Abu Lugod, 1969, cited by Timms, 1971: 144–5) this was interpreted as an unfinished transition from Gemeinschaft to Gesellschaft. Social Area Analysis (Shevsky and Bell, 1955) had already posited social rank and family status as separate dimensions of residential separation, i.e. each one being accountable for a different spatial pattern of urban segregation. In Cairo, however, segregation patterns according to wealth (social rank) and household size (family status) were conflated since larger households were also systematically wealthier due to the possibility for rich men to marry several wives.

  • 4

    Urban Studies has hosted several other authors who have critically discussed polarization in global (or presumed to be global) cities in different parts of the world. See, for example, Baum (1997; 1999), Wessel (2000), Walks (2001) and Vaattovaara and Kortteinen (2003).

  • 5

    See Arbaci (2007) for an attempt to relate a typology of welfare regimes with (ethnic) segregation across Europe in a spirit that challenges any assumptions about direct segregation impacts from economic restructuring and presumed social polarization.

  • 6

    Burgers (1996) provides a characteristic example of this approach by explicitly claiming that, since all cities have the tendency to become ‘dual cities’, the best way to learn what is happening in cities all over the Western world is to study global cities.

  • 7

    On a broader level, Häussermann (2005) has argued that urban sociology since the Chicago School has provided rather inadequate platforms for grasping the distinctive social identity of the European city.

  • 8

    These aggregate classes are based on a recent classification proposal of Goldthorpian inspiration — the European Socio-economic Classes[ESeC] (see Harrison and Rose, 2006; Rose and Harrison, 2007).

  • 9

    These two class categories were chosen as the most consistent reference groups within the two class poles, both for their substantial numbers and the clearly different form of their spatial distribution.

  • 10

    Arapoglou (2006) has already drawn attention to the reduced segregation of immigrants in Athens following an analysis that challenges the de-contextualized conventional wisdom regarding the patterns of immigrant integration. Arbaci (forthcoming) investigates immigrant segregation in eight South European cities and argues that reduced spatial segregation is a common feature related, among other things, to the profile of local housing and broader welfare systems.

  • 11

    This corroborates Arbaci's (forthcoming) claim about the coexistence of reduced segregation with housing marginalization for immigrants in Southern Europe.

  • 12

    The convergence/divergence of US and European urban models, involving their socio-spatial structures and regulation, have often been discussed in urban sociology. See Häussermann (2005) and Le Galès and Zagrodzki (2006) for some recent insights.

  • 13

    A substantially higher percentage of immigrants is employed in small businesses (with less than 10 employees) compared to the native Greek population (Kandylis et al., forthcoming).

  • 14

    Arbaci (forthcoming) argues, for instance, that reduced segregation does not hinder housing marginalization in the South European context.

  • 15

    I constructed a very broad social typology of residential space in Athens using a combination of a binary correspondence analysis on 17 class variables to determine the main social axes of their spatial variation, and a K-means clustering to classify the city's 3,500 census tracts (average population: 1,000; average active population: 450) according to their position on the axes that resulted from the correspondence analysis. The class variables are a disaggregated version of the nine fundamental ESeC classes that (a) differentiates between managers and professionals; (b) uses separate categories for large employers and artists; and (c) distinguishes the members of intermediate and lower socio-economic classes employed in industry, construction and transport from those working in the services. Agriculture-related occupational classes have not been taken into consideration due to their limited numbers and their important spatial variance, which is not relevant to the object of this analysis. The variables used are directly available in the database application EKKE-ESYE (2005).

  • 16

    The occupational profile change was measured by the percentage point change of the major occupational groups during the 1990s in each cluster adjusted according to the percentage change of each occupational group in the city as a whole (i.e. if an occupational group's percentage increased by 3 points in cluster x, while it decreased by 1 point in the city as a whole, its performance in the particular cluster was considered +4 percentage points).

  • 17

    See Murie and Musterd (1996) for a comparison of segregation trends in UK and Dutch cities in terms of different housing policies since the 1970s.