1Socio-spatial differentiation, or residential segregation, of the urban population according to social category, age, ethnic group or lifestyle is not stable over time, but subject to fluctuations in the segregation patterns that produce an “urban mosaic”. Over long time-spans, it reflects the hierarchisation and social fragmentation of society within the urban space. In this respect, the configuration of the social space is markedly time-dependent, and alters with the overall environment of social change.
2Western societies today are facing considerable economic, demographic and cultural changes, linked to the shift from an industrial to a post-industrial society, which has been underway for the last few decades. The notion of the post-industrial society, originated by the sociologists Bell (1973) and Touraine (1969), is an attempt to conceptualise the emerging organisation of society arising from technological progress (Crouch, 1999). According to Singelman (1978) the “Fordist-industrial” system based on the extraction of raw materials, their processing (production of standardised goods on a large scale) and their distribution is moving towards a system termed “post-industrial”, which is organised around services to business and enterprise, social services, and individual care services.
3In cities in particular the new organisation of society appears to produce a phenomenon of social polarisation (O'Loughlin & Friedrichs, 1996). Scientific debate suggests that a socio-spatial fragmentation is occurring inside cities. Certain sociologists are in favour of different patterns, such as that of the “dual city” (Mollenkopf & Castells, 1991, or that of the city divided into three of more parts (Häuermann & Siebel, 1987; Krätke, 1995). The social and spatial processes occurring in cities are closely linked to the position of each city within the urban hierarchy. In cities at the top of the hierarchy, i.e. the global cities (Sassen, 1991), which are characterised by their economic, political and cultural stature on world scale, (Friedmann, 1986, 1995; Sassen, 1994), social polarisation and spatial fragmentation appear more marked.
4The processes occurring in the social spaces of the different cities are numerous and diverse. In many Western cities an upgrading or “gentrification” of the central city areas is observed (Castells, 1994; Hamnett & Randolph, 1984; Häuermann, 1990; Lees, 2003; Ley, 1992, 1996). These studies show how the new attractiveness of the city centres for the more well-off is closely linked to the economic and social changes of recent decades. The de-industrialisation of the central areas is often accompanied by urban renovation projects, and is thus the starting-point or the vehicle for the social up-grading process. In addition, despite financial restrictions, urban policies have generated large prestige-related projects aiming to enhance the attractiveness of the city, in the context of new inter-city rivalries. Indeed, in a tertiary economy, prestige and the “well-being” factor are elements that have become increasingly important (Helbrecht, 2001). Finally, post-modern society has seen the emergence of a new, fairly wealthy social class adopting an “urban” lifestyle and preferring to live in the city centres (Dangschat, 1990; Ley, 1996).
5Another aspect is the fact that societies are experiencing the “second demographic transition” (Lesthaeghe, 1995). This demographic transition is characterised by an increase in the number of single-person households, divorces, and single-parent families, and also by a lengthening of life expectancy and a drop in the birth rate to below the renewal threshold (Eurostat 2004; Höpflinger, 1997). More globally, the human life-cycle in the industrial society, based on the model of the nuclear family with four main phases (marriage after leaving the parental home, birth of several children, departure of the children and widowhood) has lost its general validity. Changes have occurred, introducing numerous differentiations and a new flexibility into this classic model: many individuals remain single for some time, in a sort of “post-adolescence” phase, before setting up durably with a partner. In addition, the more fluctuating nature of partnerships has increased the number of phases in life cycles, which have become more and more individualised (for instance there may be a return to being single, or a phase of single parenthood). Finally, children are no longer the top priority in the hierarchy of values of couples (Glick, 1989; Kemper, 1985; Stapleton, 1980).
6Socio-demographic segregation in the city is often represented in a concentric ring model. According to this logic, the ageing population and people living alone are located in the city centre, and the young population and families are located in the urban periphery. Certain classic theories consider that the urban population makes residential adjustments according to stages in their life cycles (see for instance Timms, 1971). Nevertheless a marked demographic inertia can be observed. Once young households have settled, they tend to age in the same location, which produces the emergence of successive rings of ageing from the centre outwards towards the periphery (Ghékière, 1998; Wyly, 1999). Ghékière (1998, p.66) describes this process as “the wave dynamics of demographic ageing” (« la dynamique ondulatoire du vieillissement démographique ») : demographic growth occurs by way of the settlement of the successive fringes of the urban area, in the form of a wave of settlement.
7Finally, migratory phenomena have been occurring on a new scale since the Second World War, and this has resulted in the cultural diversification of Western societies. During the first wave of immigration (migration of labour) in the 1950s and 1960s, and then in a second wave (migration of families joining the workers) in the1970s, groups came in from southern Europe and later from Turkey and the Maghreb, or from the former French colonies in sub-Saharan Africa. Since the end of the 1970s, a new wave of post-industrial migration can be observed characterised by motivations and countries of origin that are different from those of the first two waves (Asia in the 1970s, Eastern Europe since the fall of the Berlin wall) (Champion, 1994; de Wenden, 1999; White, 1993). The spatial patterns of location of these migrant populations take the form of numerous pockets of concentration. Because of the many filtering processes, and the frequently precarious economic situation of these populations, they often concentrate in the oldest, most run-down housing (Friedrichs, 2000; Van Kempen & Özüekren, 1998; Simon, 1998).
8In urban analysis the three socio-ecological models of the Chicago school constitute references: the concentric model developed by Burgess, Hoyt's sector model, and the multiple nuclei model developed by Harris and Ullman. Nevertheless, as a result of recent segregational processes, the configuration of social spaces is constantly changing. Consequently, the classic representations of the social space of cities require reappraisal. The work by Davies and Murdie (1991, 1994) shows that the urban model for the post-industrial society is clearly distinct from that for the industrial society. In their book “Globalizing cities”, Marcuse and Van Kempen (2000) also raise this same issue of a new spatial order.
9In France the spatial manifestation of the new social structures also holds true, and has been the subject of numerous studies. For instance, Brun and Fagnani (1994) mapped the new residential trajectories of well-off couples, and Préteceille (2006) showed the wide diversity of segregation patterns according to socio-professional category. Integrating the temporal dimension into the analysis of territorial inequalities, Martin-Houssard and Tabard (2002) have shown an increase in spatial inequalities and the emergence of the dyad unemployment/employment as a differentiation factor between 1990 and 1999.
10The present article aims to further clarify the contemporary structure of socio-spatial differentiation in the intra-urban space. Its specific approach is the comparison of the main French cities and the analysis of their socio-spatial evolution from the 1960s. This temporal horizon can be justified by the fact that the start of these social changes is generally dated from the 1960s and 1970s. Thus the analysis of these changes over a relatively long period of 30 years should enable a better understanding of the changes in their temporal dimension, and in particular enable the social transformations of the urban space to be traced. A comparison of social structures across time makes it possible to clearly evidence the social and spatial structures of the industrial city in relation to those of the post-industrial city.
11The study reported here covers the eight largest urban areas in France: Paris, Marseille, Lyon, Lille, Nice, Toulouse, Bordeaux and Nantes. Populations range from 0.7 million inhabitants (Nantes) to 11.2 million (Paris) in 1999. The large size of these cities is suited to highlighting general segregation patterns that take precedence over local spatial segregation factors. In addition, with the exception of Paris, these metropolitan spaces are comparable because they are all regional metropolises (Damette, 1994), which generally drive innovation in social change.
12The article is in three parts. First the way in which the social structure of the population in French cities evolved between 1968 and 1999 will be presented, on the scale of the urban area as a whole. The issue will be to determine the evolutions that are in common to all these cities, and which social differences characterise the eight main French cities. Secondly the different segregation processes that can be observed in these French cities from 1968 to 1999 will be presented, with a focus on analysing the localisation of the different social groups in relation to the various urban rings, which appear to be the most suitable analysis grid to capture the various processes of displacement, concentration and dispersion of the various social groups. Finally, models of the social spaces of the different cities at the end of the 20th century will be set out. These spatial models will be based on the results of the analysis of segregational processes and the exploitation of several principal component analyses, so as to determine the shape and the dimensions of the social spaces.
13Prior to the analysis of the different social redistribution processes in French cities, it will be useful to provide certain details on the methodology used.
14For this study, the different urban areas were delineated according to the definition provided by the French national statistics institute (INSEE). An urban area is defined both morphologically and functionally. It comprises two parts, firstly the urban pole, characterised by a clear urban morphology (dense built-up area), and secondly the polarised or peri-urban zone, characterised by discontinuous built-up cover, which is nevertheless a space in which the population entertains strong links with the central space (this being measurable by way of home-to-work commuting patterns).
15The delineation of these urban areas is that established for 1999. For reasons of analysis, the urban areas were divided into four rings
-
the city centre (or central commune)
-
the suburbs (in the 1982 situation)
-
the former peri-urban ring (peri-urban space in 1982)
-
the recent peri-urban ring (peri-urban space of 1999, excluding the communes of the former peri-urban ring).
16Figure 1 shows this spatial configuration for the city of Lyon. As peri-urbanisation began in Europe in the 1960s, in that period the peri-urban space was virtually non-existent, and was added on as the urban agglomeration developed. To interpret the maps, this fact needs to be kept in mind.
17The data mobilised for this study is derived from the five population censuses of 1968, 1975, 1982, 1990 and 1999. A certain harmonisation was required to render the different variables comparable over time. For instance, for the 1982 census INSEE replaced the definitions of the socio-professional categories as established in 1954 by a new nomenclature, which therefore was extrapolated to the earlier censuses so as to ensure that the variables were comparable. The data is represented according to the constant geography provided above.
18The spatial grid is derived from that of the communes (the finest administrative division in France) that make up the different the different urban areas. To ensure statistical precision, the communes with a small demographic weight (extremely variable across the French territory) were grouped, with a threshold of 200 working individuals (on a one-quarter sample). Finally the city centre communes, generally the most populated, were subdivided into smaller spatial units.
Figure 1. Definition of study zones in the different urban areas: the example of Lyon.
19The urban development of recent decades saw a marked progression of urbanisation. In France, the share of the population living in one of the 354 urban areas amounted to 77% of the population in 1999 (see Julien, 2001).
- 1 All the figures given in this article were calculated according to a fixed delineation of the urban (...)
20For the eight largest French urban areas, considerable growth can be noted, ranging from 12.7% (Lille) to 62% (Toulouse) between 1968 and 1999 (Table 1)1. This growth results both from a natural population increase, and from a positive migratory balance in most of the urban areas. Nevertheless the evolution of the population is very unevenly distributed in space. Following processes of de-urbanisation (sub- and peri-urbanisation) the city centres underwent a fairly marked demographic decline up to the 1980s. Since then, the tendency towards lesser densities in the city centres has to some extent reversed in most of the cities studied. Only Paris and Marseille still had a negative migratory balance between 1990 and 1999. The largest demographic increases occurred in the suburbs, and this was supported by the massive construction of tower blocks for social housing, and of individual housing on a large scale from the 1960s. This is the spatial expression of the economic growth of the industrial society (Kesteloot, 2000). More recently, the return towards the city centres is a spatial characteristic of the post-industrial city.
Table 1: Evolution of the population in the eight largest urban areas in France (constant urban areas of 1999)
Source: INSEE census data 1968 and 1999 (calculations by the author)
21The economic changes of recent decades have had a profound impact on the socio-professional composition of the working populations in French cities. Following industrialisation, and with the increase in service-related (tertiary) activities, a marked decrease in the share of manual and unskilled workers in the working population, and an increase in the proportion of managerial and professional people, in the different urban areas between 1968 and 1999 (Figure 2). The general evolutionary trend in the socio-professional structure demonstrates marked stability. Indeed, numerous studies have shown the very great stability over time of the (functional) position of cities belonging to a given city system (Cattan, Pumain et al, 1994, Pumain 1982). If the share of managerial and upper intellectual professions is compared with that of manual and unskilled workers in the different urban areas, it can be seen that Paris confirmed its tertiary function with 21% managerial and professional jobs in the working population in 1999. In the global hierarchy, Paris has taken on a world city function, demonstrated by its financial, political, economic and cultural role at global level (Vandermotten & Vermoesen et al, 1999; Sassen, 1994).
22On account of the share of managerial and professional occupations in the workforce (17.2% in 1999), the urban area of Toulouse can be characterised as the second tertiary metropolis in France. After having been bypassed by the industrial revolution in the 19th century, by way of its economic policies in the 1950s and 1960s Toulouse has become the construction and research pole for the aeronautical industry in France. In addition to research, the economy of Toulouse is markedly specialised in the areas of computing, telecommunications, management and services to business (Jalabert, 1995; Julien, 2002).
23The urban area of Lyon is the second city in the French urban hierarchy in terms of demographic and economic weight. It was particularly hard hit by de-industrialisation (textiles, metallurgy), but finally succeeded in diversifying its economy, which now rests on tertiary activities (trade and retail, management, research, and services to business), while maintaining a certain industrial infrastructure (Bonneville, 1997; Pumain & Saint Julien, 1990). Consequently the structure of the working population in 1999 is characterised by a relatively large proportion of workers (23.3%, alongside a managerial and professional population of 15.0%).
Figure 2: The share of the different socio-professional categories in the working population (in %) in the eight main French cities in 1968 and in 1999
Source: INSEE census data 1968 and 1999 (calculations by the author)
24The urban areas of Lille and Marseille underwent a severe economic crisis as early as the 1960s. Lille, which was for a long time a major centre for the steel and textile industries, is today a large tertiary metropolis (Pumain & Rozenblat, 1996; Gamblin, 2001). The urban area of Marseille, for its part, has suffered from the loss of the economic base that for centuries revolved around the port. The unemployment rate, at 20.1%, is a clear reflection of this economic crisis. It should be noted that although this unemployment rate is the highest of all the urban areas studied here, the increase in the numbers of unemployed is a new characteristic of urban societies in general (Figure 2). In 1999 the unemployment rate was more than 11% across the urban areas considered overall, and indeed the rate for the whole of France in 1999 was 12.8%.
25The economy of the urban areas of Bordeaux and Nantes is characterised at the end of the 20th century by its diversity, without any prominent specialisation (Julien, 1995; Pulmain & Saint Julien, 1995). The structure of the working population can thus be described as fairly balanced.
26Finally in Nice the economy is strongly influenced by tourism, and also since the 1960s and 1970S by the arrival of high-tech industry (Dauphiné, 1990; Pumain & Saint Julien, 1990). Managerial and working class jobs are both under-represented, and the main characteristic of the working population is the large share of tertiary workers and people in trade and small business.
27Finally, it should be mentioned that since 1968 the number of “intermediate” professions has increased considerably, today reaching around a quarter of the working population in the different urban areas (Figure 2). A slight increase can be seen for the number of salaried workers while overall, the share of self-employed, retail traders and small business owners has decreased in the course of time.
28The geographer Clark (1987) remarked that demographic change is as important a factor for social change in cities as is economic change.
Figure 3. The share of socio-demographic groups in the city populations (in %) in 1968 and 1999
Source: INSEE census data 1968 and 1999 (calculations by the author)
29A first aspect is the considerable increase in the number of single-person households. Figure 3 shows that this type of household rose from around 20% in 1968 to more than 30% in 1999. When the different urban areas are compared directly, it can be noted that the urban area of Nice registers 36.5% single-person households, thus ranking first in 1999 (see also Ogden & Hall, 2000).
30The other change is the ageing of society, demonstrated by the increase in the number of people over 64 and the decrease in the proportion of under 18 year olds in the population. With 22% of the population over 64 in 1999, Nice is also the city with the oldest population among all the urban areas studied. If in general the increase in the proportion of people over 64 can be explained by the increase in life expectancy in the different urban areas, in the south of France this trend is boosted by the large number of retired individuals settling in that part of France. In general, the increase in the number of single-person households is however note solely explained by the increase in life expectancy, it is also related to a new phase of celibacy (post-adolescence), and to the more widespread breakdown of families. This second phenomenon has led to a doubling in the proportion of single-parent families between 1968 and 1999 in most French cities. At 16.9% in 1999, the proportion is particularly marked in Marseille.
31In 1999 the smallest proportion of people over 64 was in Paris, at only 12.5%. The population of Paris is characterised by a large 25 to 45 age group. This particular demographic structure can be explained by the “Paris filter”, attracting students and young professionals, who move out after a few years (Bonvalet & Lelièvre, 1991; Ogden & Hall, 1998).
32Regarding the percentages of people under 18 in the different urban areas, it can be noted that Lille is the “youngest” city, with a quarter of the population under 18 in 1999. The share of single-person households is the lowest.
- 2 In this article "foreign" or "immigrant" populations are defined as individuals born outside France (...)
33The presence of large ethnic groups is often an urban phenomenon. In Europe, the largest foreign (immigrant)2 populations are found in large cities such as London, Paris or Brussels (Friedmann, 1995). This concentration is attributable both to job availability and to the existence of already established ethnic networks that facilitate integration into local life for new immigrants settling in the country. The different ethnic groups are however very unequally distributed across French cities. Table 2 shows the proportions of foreign or immigrant populations in the eight largest urban areas in France in 1999. It can be noted that there is a wide variation in the share of immigrants in the populations of the cities considered. In absolute value and in percentage, the largest number of immigrants is to be found in Paris. The smallest number is in Nantes. Nevertheless, even in Nantes, the urban setting has an effect on the proportion of immigrants in the population: thus the Pays de Loire region in which Nantes is situated has 1.9% immigrants; while the urban area has 3% (see also Boëldieu & Borrel, 2000).
Table 2: Immigrants in the 8 largest urban areas in 1999
Source: INSEE census data 1968 and 1999 (calculations by the author)
34The different immigrant groups, classified here in four categories according to their region of origin, are very unequally represented in the different urban areas. For instance, the regional concentration of immigrants from southern Europe in Bordeaux (38.8%) and Toulouse (37.8%) reflects the geographical distance from Spain and Portugal. The marked presence of immigrants from Africa in Marseille (58.8%) shows the classic function of Marseille as the port of entry for numerous immigrant groups from Africa.
35Numerous studies on the social space of cities show that residential segregation is determined by socio-economic status, family characteristics (stage in the life-cycle) and the ethnic origin of households. To understand the evolution of the social space in its complexity and the many interactions among the different social groups, 22 variables that characterise differences in the social mosaic were considered (see column 1 in Table 4). In addition, 5 variables characterising he different living environments in the city were also taken into account.
36In a first approach to the socio-spatial evolutions, only a few indicators were considered: managerial and upper intellectual professions, workers, population aged 18 to 24, population aged 55 to 64, population aged over 64, households with a child under 6. These variables are the most significant indicators to understand the processes occurring over recent decades.
37For an analysis of changes in the geographical distribution of the different social groups in the city their localisation was represented in relation to the four concentric urban rings (Figure 1). Indeed, analyses on the city have shown that family differences often follow the concentric circle model. When socio-economic differences are envisaged, they generally comply with a sector model. However the analysis of the distribution of the different socio-economic groups showed that changes in recent decades followed the centre-periphery model, without invalidating the sector pattern which is dominated by a concentric layout across the urban area as a whole (Schwabe, 2005).
38To capture the spatial logics of localisation, a localisation quotient (LQ) was calculated, indicating the concentration of a social group in a spatial entity in relation to their total share in the space envisaged (here, the urban area, see Apparicio, 2000) This quotient enables the comparison of inter-census and inter-urban spatial structures, since it is independent from the mean. A value smaller than 1 indicates a spatial concentration under the mean; a value of 1 indicates a concentration that is equal to the mean; and a value over 1 indicates a marked concentration of the group.
39Unlike American cities, French cites have always been characterised by the presence of wealthy social classes in the city centre. Figure 4 confirms this regular feature for the managerial and professional categories. With the very small representation of workers in the centre and their marked concentration in the suburbs, it can be said that the localisation of the two groups at either end of the social hierarchy are opposite. The evolution of their localisation between 1968 and 1999 shows an increase in social disparities between city centre and suburbs from the end of the 1970s in Paris, Bordeaux, Lyon and Lille. The degree of disparity between centre and periphery remains fairly stable for Nantes and Toulouse, and decreases slightly for Marseille and Nice in this same period. This result in the longer term confirms the results reported by Lajoie (2002) and Reynard (1995) for the period 1982 to 1990.
Figure 4: Localisation of managerial /professional category and worker category in city centre and suburbs
Source: INSEE census data 1968 and 1999 (calculations by the author)
40In France the urban area of Paris can be considered as a model of the social and economic changes occurring, and their socio-spatial imprint. Paris intra muros (city centre) has long been attractive to the population, and this has led to a sharp increase in property prices over a number of decades. These increases can be explained by the fact that the migratory balance (Table 1) between the central space and the rest of the urban area is not only selective in relation to the position of the individual in the life cycle; it is also selective in relation to socio-professional category (Rhein, 1996; Noin & White, 1997). For instance, only 5.5% of the managerial category living in the city centre in 1990 left the centre for other parts of the urban area between 1990 and 1999, compared to 22.3% of the worker category (Schwabe, 2005). The same phenomenon explains the increase in socio-spatial disparities in Bordeaux and Lyon.
41In fact, the housing market plays a central part in the social stratification of the urban area. According to the classic Alonso-Muth urban economy model (Alonso, 1964), the urban area organises itself in concentric circles, where property prices decrease with the distance from the city centre. In the context of property market pressures, only the presence of a large pool of social housing can enable poor populations to continue living in the central areas of the city (Preteceille, 1997). Figure 5 shows this relationship very clearly. The more concentrated is the availability of social housing in the city centre, the greater the concentration of the worker population. This relationship is independent from the total proportion of social housing in the urban area (Table 3).
Figure 5: Relationship between the concentration of social housing and localisation of workers in the city centre
Source: INSEE census data 1968 and 1999 (calculations by the author)
42To explain the evolution of socio-spatial disparities between city centre and periphery it is not sufficient to consider solely the flows between city centre and the rest of the urban area, it is also useful to analyse flows between inside and outside the urban area. Because data on this migratory balance is lacking, only an analysis of the settlement zones of new arrivals in the urban area was performed (Schwabe, 2005). It shows that the localisation of new arrivals in the urban areas is very selective with regard to socio-professional category. For instance, in Paris among all the new arrivals who settled in the urban area between 1990 and 1999 only 19.2% of the workers against 43.4% of the managerial and professional arrivals settled directly in the city centre. The increase in spatial inequalities is likewise explained in Bordeaux, Lyon and Lille by the same socially selective flows. In contrast, in Nantes, Toulouse, Marseille and Nice the settlement rates for the different categories in the city centre are more balanced. In these cities the proportion of social housing in the city centres is above that in the urban area overall, (Figure 5 and Table 3), which shows the influence of the availability of social housing on urban segregation.
Table 3: Proportion of social housing (HLM) in the large urban areas in 1999
Source: INSEE census data 1968 and 1999 (calculations by the author)
43The socio-demographic segregation in cities is frequently described as a localisation of the elderly population in the city centre and the young population and families in the peripheral areas. The different graphs in Figure 6 show that this spatial approximation was correct at one time, but that it is not at all true today or for future trends.
44If we look at the evolution of the localisation of the population over 64 in city centre and suburbs between 1968 and 1999, the first observation is that localisation is not stable over this period, and also that the maximum over-representation in the city centre can be seen for the 1980s in most of the urban areas studied. From then on, there is a marked decrease in the concentration of this age group, with the exception of Nice where the elderly population is still moderately represented in the centre. In Marseille the evolution is less marked on the graph because the weight of the central commune distorts the analysis. Regarding the presence of elderly populations in the suburbs, it can be seen that their under-representation has tended to decrease since the 1980s. This process was evidenced and analysed for the first time by Ghékière (1998) in Lille. The change in the localisation of the population aged 55 to 64 shows the same trends, with a time-lapse of around ten years (Figure 6). For this age group, there is an under-representation in the city centre and an over-representation in the suburbs in the 1999 census in the majority of the urban areas studied. Assuming they grow old where they were located in 1999, this graph suggests a clear over-representation in this zone of the population over 64 in 2009 (10 years on).
45Regarding the localisation of families with one or several children, their localisation is generally expected to be the suburbs and the peri-urban areas. If households with at least one child are considered, this pattern appears quite clearly. Nevertheless, if the young families with one or more children under 6 are compared to families overall (couples with or without children and single-parent families) new localisation trends can be seen. This shows an increase in the localisation of (young) families in the city centre and a de-concentration in the suburbs. This process is particularly marked in Lyon, where the centre-periphery migratory balance for young families is constantly decreasing, falling from 48.3% of young families leaving the centre to move to other parts of the urban area between 1968 and 1975 to 27.2% between 1990 and 1999 (Schwabe, 2005). Different studies have shown that some of these couples no longer leave the city centre at the birth of their first child. In fact, the city centre is an ideal location for the family belonging to post-industrial society: it makes it possible to reconcile contradictory requirements, such as the professional careers of women, proximity of the workplace for both partners, urban lifestyle and social relations (Rose, 1989; Ley, 1992; Brun & Fagnani, 1994; Karsten, 2003).
46Another characteristic of the urban population that has already been reported is the increase in the number of single-person households in recent decades. Their localisation is markedly in the city centres. In 1999, the proportion single-person households among households in the city centres was more than 50% in Bordeaux, Lille, Paris and Toulouse (see also Ogden & Hall, 2000). Nevertheless, with the displacement of ageing populations, the concentration of households comprising a single (elderly) person is also increasing in the suburbs. While for a long time these single individuals were above all elderly women, on account of differences in life expectancy between men and women, the present structure of these households is today radically different. Overall the new single individuals are now mainly young adults (students or individuals in their first job), adults of intermediate age who have divorced, and finally elderly single people. Figure 6 shows the localisation of young adults (aged 18 to 24). Their localisation is increasingly central in most of the urban areas. The situation is extreme in Lille where one fifth of the population (21.7%) in the city centre was between 18 and 24 years old in 1999. The absence of this phenomenon in Marseille can be explained by the large surface area of the central commune, and the location of the university campus in the outskirts (Aix en Provence). In Paris the tight property market appears to explain the balance of localisation of the 18-24 age group between centre and periphery.
Figure 6: Localisation of the different socio-demographic groups in the city centres and in the suburbs
Source: INSEE census data 1968 and 1999 (calculations by the author)
47The socio-spatial structure of cities is often construed by way of three dimensions and the spatial patterns associated with them: social status in association with a sector model, family status with a concentric circle model, and ethnic status with a multiple nuclei model. In the previous section of this article, different processes of socio-spatial change have been shown. It can be supposed that these processes together generate a new socio-spatial structure of French cities.
- 3 For the tables concerning the other urban areas, see Schwabe 2005.
48For this reason, a principal component analysis was performed for each urban area individually. This enables, for each case, the identification of the main dimensions of the socio-spatial structure. For the analysis, 27 variables pertaining to professional status, working activity status, training, age, household and family structure, housing characteristics and ethnic group were retained (see the “variables” column in Table 4).The table giving the variable scores for the city of Lyon is given as an example3. The five main dimensions in terms of explained variance and associated spatial configuration are presented, for each urban area studied, in figures 6 and 7. These two figures providing the synthesis of spatial configurations were developed not only from maps of the spatial distribution of factorial scores associated with the different axes; they were also derived from the maps of the spatial distribution of the different variables included in the analysis. The two figures integrate the different spatial rings in each urban area, and the geographical orientation of each spatial structure represented. The map of France is represented as a hexagonal outline.
49From the analysis of Table 4 which provides the factorial scores for the 27 variables on the different axes for the city of Lyon, it can be seen that in 1999 the structure of the urban area is more complex that that predicted by the classic models. Instead of a single socio-economic dimension, two are found, instead of a just one socio-demographic dimension three are found, this being true for most of the other urban areas studied. Indeed, this greater complexity is found not only in the example of Lyon, but also in the other urban areas (Schwabe, 2005).
50The main spatial differentiation dimension is the social marginality status. It distinguishes the city quarters characterised by a large proportion of unemployed, single-parent families, immigrants from Africa and Asia, and social housing (HLM) on the one hand, and on the other spaces comprising a large proportion of individual houses. The association of single-parent families with this axis is reinforced by the priority allocation of social housing to this category of persons. This characteristic can be shown for 1999 in all the urban areas studied except Nice, where the small proportion of social housing appears to prevent the emergence of this socio-spatial dimension. In Marseille, where the structure of the social marginality space is more complex than in the other urban areas, this dimension has two components. One associates the social characteristics of this dimension with social housing in large ensembles; the other associates them with de facto (informal) social housing in the old quarters. Figure 7 shows that the social marginality is particularly prominent in some quarters in the city centre and the suburbs. In the Paris urban area, this type of area is typically suburban. Indeed, because this type of quarter is characterised by the presence of social housing, its spatial layout is strongly influenced by the distribution of social housing across the urban area (Table 3).
Table 4: Factorial structure of the social space of Lyon in 1999
Source: INSEE census data 1968 and 1999 (calculations by the author)
51The second dimension is easily recognisable as the classic social status dimension. It contrasts urban spaces characterised by a large proportion of workers with spaces in which there is a marked overrepresentation of managerial and professional categories and individuals possessing a higher education diploma. This dimension can be identified in all the urban areas. Its spatial configuration is mainly sectorial. Over time, the initial sectors have extended into the peri-urban space. In Lille the sector pattern is less clear on account of a distinct economic history, and the existence of several urban cores (Roubaix and Tourcoing) which make the layout more complex. In Paris the gentrification (embourgeoisement) of the city centre is so advanced that the sectorial layout of the city centre has been virtually replaced by a central core in which he population has high social status, from which the social sectors extend outwards to the peri-urban space.
52Regarding socio-demographic segregation, a break-down into several socio-spatial dimensions can be observed in most of the urban areas studied.
53The family status dimension observed in numerous factor analyses can only be identified in Marseille, Nice, Paris and Toulouse. In the other urban areas it is divided into two independent socio-demographic dimensions: family and age, and young families and ageing families. The dimension family and age contrasts the population under 18 and the population aged 35-44 with persons aged 65 and over and with families (couples) without children.
54The spatial distribution of family status and young family & ageing family status is given in Figure 8. The layout is mainly one of concentric rings. Nevertheless, following the demographic ageing described above, it is no longer possible to conclude to a simple centre/periphery gradient. The elderly population appears now to be increasingly located in the near suburbs, and families in the periphery. However in Nice the localisation of the elderly population in the city centre and in the suburbs is close to the mean, and very stable over time, so that the distribution of the factor scores for family status is better captured by way of a straightforward centre-periphery gradient. This can no doubt be explained by the large proportion of retirees in Nice.
55The second socio-demographic dimension is the young family & ageing family status. Here certain quarters are notably occupied by young families with at least one child under 6 and people aged 25 to 35, and others by people aged 45 to 55. This dimension can be identified in Bordeaux, Lille, Lyon and Nantes.
56The spatial distribution can be described as a prominent presence of young families not only in the urban periphery, but also in the city centre. This new pattern is the result of new behaviours by young families, often fairly well-off, who remain in the city centre beyond the birth of the first child.
57Finally, in all the urban areas except Lille and Nice, there is a young adult dimension, where there is a very marked concentration of individuals aged 18 to 24. The spatial distribution for this category shows a large concentration in the city centres, and, following the establishment of university campuses from the 1960s (Merlin, 1995), concentrations appearing in certain zones in the near suburbs. However in the Paris urban area this age group is strongly concentrated in the distant suburbs. The localisation quotient indicates an almost equal spatial distribution between suburbs and city centre. A late home-leaving age among young adults could explain this socio-spatial feature (IAURIF/INSEE, 2002).
58In Paris and Toulouse there is a particular demographic dimension. It takes the form of a contrast between the population aged 25 to 34 with that aged 45-54. The emergence of this socio-demographic dimension can be explained by a large proportion of single people between 25 and 34 (Rey, 1997; Ogden & Hall, 2000).
59Finally, a socio-ethnic dimension can be expected, and was indeed observed in this study. We have already seen the association of immigrant populations from Africa and Asia with the first dimension characterised as socio-economic. In addition, it is possible to identify a dimension that can be interpreted as ethnic and predominantly south European in most urban areas. Immigrants from southern Europe are generally concentrated in suburbs, except in Nantes and Bordeaux where they are over-represented in the city centres. Regarding immigrant populations from other countries, this group is often associated with social status.
Figure 7: The socio-economic space in the larges urban areas in France
Figure 8: The socio-demographic space in the large urban areas in France
60The analysis of the residential mosaic in the main French cities was conducted with the underlying hypothesis that the social changes of recent decades result in a new configuration of their spatial structure.
61First of all it can be noted that the changes in the different urban societies show roughly the same socio-economic, socio-demographic and socio-cultural trends.
62Secondly, the analysis of changes in the residential localisation of a few social groups makes it possible to capture the different segregation processes that can be observed in the urban space. For instance, processes of social upgrading in the city centres, the displacement of ageing populations, or the residential shift of young adults towards the city centre in most of the urban areas considered were all observed trends.
63Finally the principal component analysis provided an understanding of the new emerging spatial structures as a whole. It shows that at the end of the 20th century, the urban structures no longer correspond to the usual representation of the social space in three dimensions (economic status, family status, ethnic status). It shows the emergence of numerous other socio-spatial dimensions (social marginality status, young families, ageing families, young adults), or else classic dimensions but where the spatial structure has changed.
64These processes of residential segregation are not specific to France. Numerous studies have shown the diverse facets of change in Western cities. For instance the gentrification of the central areas has been observed in most large European cities (Vandermotten et al, 1999), the shift to new areas of ageing populations was observed by Köster (1994) or Wyly (1999), and the presence of families with one child in city centres was reported by Ley (1992) for Canadian cities or Karsten (2003) for Amsterdam. Regarding the structure of the social space overall, the factor analyses undertaken by Davies and Murdie (1991) for the 24 “metropolitan areas” in Canada show the emergence of new dimensions of social differentiation in urban space which are comparable to those identified here for the 8 main French cities (see Murdie & Teixeira, 2000). Do these socio-residential changes indicate the emergence of new laws of socio-spatial differentiation that are typical of society in the post-industrial city?
65In all events, this study also shows differences in terms of intensity, and trends in spatial evolution. For instance it has been noted that the urban area of Nice often forms and exception to the trends generally observed. This can be seen as a regional feature, possibly typical of the Mediterranean city. We can also return to the example of the Paris urban area which is clearly distinct from the other urban areas for the expression and the intensity of the socio-spatial restructuring occurring there. One feature of this urban area is that it registers the greatest increases in socio-economic disparities of all the urban areas studied. It seems that the role of Paris as a “global city” leads to this spatial pattern. This economic position is also reflected in the demographic spatial patterns. The particular distribution of young adults in this urban area can be cited as an example.
