Navigation – Plan du site

The “Paris Basin”, a metropolitan region

Frédéric Gilli
Cet article est une traduction de :
Le Bassin parisien 
Une région métropolitaine

Texte intégral

  • 1 This work was conducted when the author was working for the Direction Régionale de L’INSEE Picardie

1Paris has influence that extends well beyond its administrative boundaries or its urban area1. An integrated metropolitan area has formed around the French capital. Daily commuting across the Paris Basin, and the existence of industrial groups of similar type, are in favour of the existence of a functional region that integrates several administrative Regions. The industrial and business patterns in this economic region show specialisation of the different territories, alongside concentrations of activities. The organisation of the economic spaces differs depending whether it is the different urban areas in the area that are taken into account, or the regional space as a whole. The presence of the Paris urban area is however felt across all the employment zones in the basin, although this by no means justifies considering the region solely in a hierarchical perspective.

2Paris is a major world city (Sassen, 1991). On this account, its activities, population and infrastructures have often been compared to those of other cities of similar rank, whether London, New York or Tokyo. Its place within the network of contemporary capital cities is important. However a city is not solely a node in a worldwide network, it is also affiliated to different territories that possess their particular logics.

  • 2 The term “bassin Parisien” translated hereafter as “Paris Basin”, while first of all referring to a (...)

3The positioning of a global city within its local network can be a determining factor for its outreach. Obviously the vitality of the central city does not necessarily depend on the vitality of surrounding spaces of secondary rank. In Asia there are numerous instances of cities that are “isolates”. But wherever relationships do exist between a city and its territory, they partly condition the successfulness and the positioning of the central space. These relationships also of course have an impact on the vitality of the local spaces involved. This study therefore set out to explore the extent of the influence zone of Paris, and the internal organisation of the Paris Basin2. For this purpose, the focus will be on the structure of industrial spaces (localisation of activities) and their daily functioning (commuting patterns).

4We will first see the mechanisms that structure the large metropolitan regions across the world so as to understand on what scale this particular space takes on meaning. We will then consider the organisation of daily commuter flows within this functional region so as to discern the structures that emerge. Finally we will complete the description with a study of the functioning of this integrated economic space.

A positioning within international networks

The importance of the Paris hinterland

5A metropolis is a large concentration of humans and activities, where the space to which they belong is organised according to a shared logic. Thus urban sprawl generates large urban regions, the contours and organisational patterns of which are fairy blurred. Spaces that were historically mono-centric become polycentric. Polycentric spaces gradually become single entities. In either case, the logics of the spatial organisation that appears reach well beyond the bounds of the continuous built-up area, or even the urban area.

6It is obvious that the larger the central zone, the more likely are these phenomena to occur on a large scale. If we consider the “global” cities, genuine economic regions are generated by a redistribution of populations across the different spaces. Here the distances involved can be enormous. Beaverstock et al (2000) thus consider that Chicago suffers from the competition of New York, some 1000 km distant. This overshadowing is linked, in part, to effects of communication infrastructures and market area, which favour the geographical concentration of players in a globalised, oligopolistic economy (Krugman 1993). As a result it is thought that there are vast economic regions revolving around a single centre, in which the emergence of a second top-ranking centre is impossible, or at least considerably hindered. These economic regions are thus seen as organising themselves around the global city. This however does not mean that these regional spaces are orbiting exclusively around the central city. They also belong to other, parallel networks where they entertain relationships with cities on their own level (Cattan et al, 1999). Their development is nevertheless hampered by the presence of a higher-ranking city with which they have special relationships.

7The examples of the Randstad and Chicago highlight two fundamental elements for our analysis of the Paris Basin. First of all, the organisation of space as generated by global cities should be viewed by way of their place in the global network of cities (Taylor, 2001). It is indeed increasingly the relationship that a city entertains with the other cities (type and intensity) that will characterise its vitality and measure its status as a metropolis. But while the area of influence of metropolises is constantly extending, it is also useful to consider more local networks. If we restrict ourselves to the first stage of the analysis, it will only be possible to sketch out the world of global cities seen from Paris or London. These global cities are also part of their local spaces, where they dominate whole regions that extend well beyond the classically envisaged boundaries of the built-up or urban area. Thus beyond the sphere of global cities and the network that they form, we therefore need to consider the zone of influence of each. In the spaces around global cities, smaller cities draw advantage from the outreach of the main metropolis. Just as a port is the point of anchorage for a territory (its influence zone) in the network of maritime routes, and serves a hinterland of some size, a global city forms a point of anchorage of a territory within a network. The organisation of these hinterlands can vary widely. The hinterland may be indistinguishable from the central space (as is the case with the Randstad), or conversely the city may be cut off from any territory (as was the case with Hong Kong for a long period).What we will endeavour to understand here in the case of the Paris Basin is how the territories located around a global city find their place in a shared space.

The integration of local spaces

8Exploring the metropolitan nature of the space around the city of Paris is not solely a matter of asking whether Paris or its urban area have developed specific modes of organisation. It is also to ask whether or not the vast economic region around Paris forms an integrated space, coherent with shared economic logics, the functioning of which is apprehended and managed as a single entity. Let us consider that a metropolis is the combination of a human agglomeration of critical size, of recognised specialised centres of activity, and of an organisation obeying its own specific logic. Urban sprawl, the break-down of traditional residential patterns (in the form of dual residences, commuting from centre to centre rather than solely between centre and periphery) are phenomena that contribute to upsetting the traditional view of the metropolitan space. A given space may indeed only make sense on a scale that takes several “regional metropolises” into account. When we envisage the way in which Paris fits into the global network, we can also wonder if Orléans, Tours, Rouen or Amiens have specific positions in this same network. It is quite possible that the outreach of these cities should be consigned to a sub-network on the level of the Paris Basin, in which each of these spaces, including the centre, is a competence node, contributing to and drawing life from Paris the “global city”. We should also note that the particular feature of a network geography is that there is no single affiliation of a place to a set of places. Thus it is not because the Paris area forms a metropolitan region that each component exists solely within and by that network.

9There is indeed no single type of economic integration, there are several. Research in economic geography has established this by distinguishing different specialisation and concentration logics operating simultaneously. When the integration of two economic spaces increases, cumulative dynamics tend, to some extent, to favour the concentration of the most productive factors in one of the regions concerned. The processes at work are numerous: they can involve the effects of market size via vertical links upstream – suppliers – or downstream – clients (Venables, 1996). There are also processes of concentration around already existing structures (Krugman, 1991) – in an economy characterised by increasing yields, any difference in size will translate into differences in productivity. The appearance of a centre-periphery geographical configuration can thus be explained merely on the basis of economic activities and dynamics (Puga, 1999). The French policies consisting in delocalising activities to provincial areas in the 1970s are thus thought to have enabled genuine transfers of activity from the Ile de France (i.e. the administrative Region covering Paris) to the surrounding administrative regions, but also to have resulted in a hierarchisation of the spaces concerned by way of concentration in certain places of activities in a particular sector, and the specialisation of certain places in a few specific activities, excluding others. In the Paris Basin there has been both a concentration of activities with high added value in Paris and the specialisation of the capital in these activities, and a concentration of production activities in the periphery, with spaces specialising in these activities. This dual pattern of concentration and specialisation is to be expected within integrated spaces. As noted by Brunet and Dollfus (1990), the division of labour that becomes established on the regional scale takes on forms that differ from the international division of labour. The different parts of the national territory are supposed to “work together harmoniously and according to their abilities towards the overall performance”. This leads these authors to observe that a main centre is rarely the only centre in a regional space, and that the spatial “translation” of this is never simple. They refer to the example of the rural areas in the Picardie or Champagne Regions, where the decisional patterns sometimes take precedence in the Parisian “centre”.

10Thus the activities of the global cities themselves depend on their positioning within these regional spaces. A refinement of the approach by way of an exploration at local scale, and a finer nomenclature for the different economic activities, give more complexity to the analysis of these relationships, and reveal a greater situational diversity. Fujita and Hamaguchi (2001) captured this phenomenon in a formalised framework: an increase in intercourse between two cities results in an evolution of their respective specialisations, and can progressively produce a hierarchised, specialised urban network, with industrial cities and tertiary cities. This dual shift can result in the disappearance of any notion of hierarchy. Although it leads to greater interdependence between zones (which also prevents the fragmentation of regional spaces) it places each zone in a leading position in its sector, enabling it to open up directly to other regional networks. Thus periphery-to-periphery relationships become established, akin to the already existing centre-to-centre relationships.

The Paris Basin, a functional inter-regional space

11Delineating boundaries is a delicate stage in the study. Indeed, it is easy to think that the space defined by existing boundaries is too restricted in scope. If Paris is really to be contrasted with the “French desert” (Gravier, 1947), it is possibly the national scale that should first of all be envisaged. The studies by Hautreux & Rochefort (1964) and Pumain (1986) are indeed in favour of this approach. This is indeed a study project in itself.

12Here we have positioned our study in relation to the administrative framework used by the Frémont mission (2002). This framework was set up in the 1970s, and has been validated several times as defining spaces that are relevant for exploring issues relating to the area of direct influence of Paris, irrespective of the conclusions reached (Thiard, 2001). It is made up of the Ile de France Region (Paris administrative Region), and the administrative Regions of Haute-Normandie, Picardie, Champagne-Ardennes, Centre, and Basse-Normandie, and also includes the départements Yonne and Sarthe.

  • 3 The French “zones d’emplois” or “employment zones” correspond to a statistical subdivision of the F (...)
  • 4 After removal of the complex case of Paris, the only urban area that is divided into several employ (...)

13This space, accounting for more than 21 million inhabitants, is mainly dominated by the capital, which accounts for half of the total population. The Centre Region (12%) ranks second, preceding Picardie with 9%. This disproportion is likewise observed in the distribution of employment as shown in Table 1. Of the ten first-ranking “employment zones”3 in the Paris Basin which comprises a hundred and one, only two are not located in the Paris Urban Area - Rouen and Caen - and none are among the first five, which are all in Paris and the inner ring of employment zones. This dissymmetry appears clearly if the different Urban Areas are the focus of the comparison, since Paris has 25 times more jobs than the second most important employment centre, Rouen. In this hierarchy, it can be noted that only the large cities in the Paris Basin are present in the top classification. However the “outpost” cities (kind of « edge cities », see Gilli, 2003) of Chartres, Evreux, Compiègne and Beauvais rank immediately next according to the approach based on Urban Areas, and are positioned on the same level as certain large cities in terms of the approach by employment zone. They are zones in which the density of employment outside the centre is high, even if the employment centre is not necessarily large. If we consider the relationship between employment in the Urban Area and employment in the employment zone4, the ten “outpost” cities emerge among the top fourteen.

Table 1 – Jobs in the Paris Basin: Paris and the main cities

Table 1 – Jobs in the Paris Basin: Paris and the main cities

14These results are in favour of the considerable integration of the peripheral spaces into a single functional region. If this is so, the exploration of commuting patterns should enable an evaluation of the scale of this integration, and the integration should also be visible in the economic space.

15The activity patterns in these spaces do seem to be in line with one another, and the same national and international groups are observed, while they are less predominant in other Regions. Regarding the institutional structure of the system of production, the Paris Basin thus can be seen as an exception on the scale of the country as a whole, since it is the only area to have an economy that is markedly dominated by large groups (Hecquet and Lainé, 1997). Despite, or because of, their specialisations, the industrial areas in the Paris Basin thus belong to a large economic region that is strongly integrated. It is on this basis that the centre-periphery model, considered to characterise the Paris Basin in the 1970s and 1980s, was developed. It is from this situation that it is today evolving very fast, as its different spaces are being affected in different manners by the structural changes in activity (Thiard, 2001).

16This apparent functional unity of the Paris Basin is also seen in the main demographic trends occurring there. Whether the commuting patterns, residential migrations, the age structure of local populations or their socio-economic composition are envisaged, the different elements defining the Paris Basin can be fitted to a single functioning model (Gilli, 2002a). Even if we restrict consideration to socio-demographic criteria, which give more importance to proximity than to network relationships, the Paris “area” today extends well beyond the administrative boundaries of the Ile de France Region. We will explore this point in more detail so as to understand how this space is organised. Indeed, if we wish to capture any form of metropolitan functioning of the Paris Basin, we need first to give attention to any boundaries that can be evidenced in the study zone. If sub-spaces are shown, the way in which they are integrated into the Paris Basin will require study. It could be useful, for example, to note whether the spaces are polarised around a given city that is the only outlet, or whether exchanges with the neighbouring areas take place by simple contact along the boundaries of each of the spaces.

17Finally, the two main issues that emerge are thus as follows: firstly there is the question of the extension of the peri-urban fringes of the Paris urban area towards neighbouring administrative Regions; secondly there is the question of the nature and the localisation of specialised concentrations of employment in the Paris Basin.

The organisation of everyday spaces viewed via commuting patterns

18On the outskirts of the Ile de France (Paris) Region there are therefore spaces in which numerous jobs are located outside the clearly identified urban centres. The Ile de France is an administrative Region with its boundaries, but these appear to correspond only very partially to the functional realities of the spaces concerned. This can be clearly seen from home-to-work commuting patterns.

Different levels of territorial coherence

  • 5 It should be noted that adopting a single criterion for commuting patterns restricts the heuristic (...)

19Commuting patterns provide synthetic data on the functioning of local spaces. They are indeed the product of a geography of the residents, a geography of employment, and a geography of transport (means of travel, readiness to travel, and of course transport networks). They are used to apprehend the “urban area” of the French National Statistics Institute (INSEE), or the “urban region” used at European level. As the present aim is to understand the Paris Basin, for the rest of this study we will adopt the scale of the 101 “employment zones”, which enable easier interpretation of results than the 11 000 communes (the finest administrative division in the French system, but their size varies considerably). As the employment zones are also defined in relation to commuting patterns, their implementation will not raise major methodological problems (with the exception of Paris, but this will not be detailed here)5.

20We apply the criterion of the functional urban region to the employment zones. Namely the Paris Region (Région Parisienne) is composed of all the “employment zones” where 10% of the local active population works in Paris (Figure 1). Note that, since Paris is off-centre within the Ile de France, the administrative Region and Paris Region do not coincide exactly.

Figure 1. Share of salaried workers working in Paris

Figure 1. Share of salaried workers working in Paris

Figure 2. Working population working in the Paris Region.

Figure 2. Working population working in the Paris Region.

21The employment zones where 10% of the working population work in Paris do not cover the whole of the Ile de France Region, but they extend out into Picardie, the Region to which the Sud-Oise employment zone belongs.

22If we now consider the commuters travelling into the Paris Region, “outside” fringes can be observed, in the sense that they do not belong directly to the central space (Figure 2). These include three Ile de France employment zones that do not belong to the Paris Region (Nemours, Provins and Montereau-Fault-Yonne) as well as all the peripheral employment zones of the Paris Region (except for Sud-Ouest Champenois, where only 7% of the working population work in the Paris Region. This ring of “outside” fringes corresponds to zones that are centred on “outpost” towns and cities.

  • 6 The model that was used as the basis for this illustration is a classic gravitational model calibra (...)

23Around the centre of the built-up area, a vast revolving space is generally distinguished. This descriptive approach has the undeniable advantage of providing clear boundaries. However it does not make it possible to know if the spaces concerned actually correspond to spaces that are lived in and experienced as “close” by the resident working population. The construction of a gravitational model based on the commuting patterns between communes can evidence these preferences (Gilli, 2002a). Seen from the capital, four types of space can be distinguished from centre to periphery (Figure 36):

  • A markedly integrated space where flows are considerable and significant. Their significance is often negative (negative flows are more clearly visible in Figure 2) since travel is more difficult in dense urban zones than in rural spaces. Since the model is calibrated on the Paris Basin as a whole, intra-urban flows can evidence this barrier effect restricting travel that is formed by the city. It can be noted that this space covers the employment zones of the Paris Region.

  • There is then a ring of centres that generate convergent flow patterns. These local poles correspond to the outpost towns and cities of the Paris Region identified earlier. They exhibit a special relationship with Paris, while no proximity effect is noted with the large cities located further from Paris. This obviously does not mean that relationships with such cities are non-existent, because in that the case the flows would show negatively.

  • The large cities also entertain special links with Paris. They also polarise their local spaces, but likewise have proximity relations with lower-ranking towns and cities. If we consider commuting patterns, modes of organisation of the Christaller type appear, a typical example being seen around Caen, but also around Le Mans, Reims, and to a lesser extent Orléans. The only centre-to-centre relationships observed, apart from the links with Paris, concern cities that are relatively close to one another (Orléans-Bois-Tours or Rouen-Le Havre). It can be noted that this type of organisation is closely linked to the variable used here: commuting patterns evidence proximity relationships. This therefore does not necessarily mean that the production space follows this Christaller-type layout.

  • Finally, certain spaces are only weakly linked to the rest of the Paris Basin. This involves all the fringe areas, and includes one city, Troyes.

  • Thus what is observed is a two-level radial concentric structure around the capital, made up of a strongly integrated inner ring and an outer ring linked via local hubs (the outpost cities) serving as gates to Paris Region. Beyond this, spaces organise themselves around most of the large towns and cities, with the exception of Troyes (very weakly linked to the rest of the Paris basin), and Amiens (strongly linked to Paris but polarising only a small zone beyond its own zone). This does not imply that major flows between home and work are an advantage or otherwise, from a normative viewpoint. But given present trends in the organisation of spaces (sprawl and metropolitanisation) the weakness of such flows is more the sign of a lack of metropolitan vitality than of a balanced multipolar structure enabling each person to find a job close to his or her home.

Figure 3. Attraction and barrier effects in the Paris Basin

Figure 3. Attraction and barrier effects in the Paris Basin

Poles structuring a fundamentally “Parisian” basin.

24In order to understand how this space functions, we thus need to differentiate the relationships existing in the Paris Basin and those existing inside the Paris Region. These two scales exhibit logics and levels of flows that are very different. Figures 4 and 5 thus show the flows observed between employment zones within the Paris Basin, between employment zones in the Paris Basin and the employment zones in the Paris Region, and inside the Paris Region.

25The first flows (shown in blue) confirm the previous analyses: outside the Paris Region, the employment zones incorporating large urban areas are the only ones involved in flows from one employment zone to another that register more than 2000 daily commuters. Two elements mentioned earlier are also observed, firstly the fact that local spaces are hierarchised, with the small employment zones rarely sustaining marked relationships (more than 1000 commuters) one with the other, and with the flows converging towards the “large” employment zones; secondly, the fact that these large employment zones rarely entertain relationships one with the other. Thus Rouen forms a counter-example, since its employment zone indeed sends out considerable flows towards Paris, but also towards Le Havre and Evreux. Orléans likewise exhibits a particular profile, with the employment zone fuelling the only flow greater than 1000 commuters towards a non-Parisian zone in the heart of the central urban area, Nanterre.

Figure 4. The main journey-to-work flows between employment zones in the Paris Basin

Figure 4. The main journey-to-work flows between employment zones in the Paris Basin

Figure 5. The main journey-to-work flows between employment zones inside the Paris Region

Figure 5. The main journey-to-work flows between employment zones inside the Paris Region

26Apart from the direct links with Paris, the large flows between the employment zones in the Paris Basin and those in the Paris Region are observed above all around the Paris Region. It is also the only part of the Paris Basin where large non-hierarchised flows occur, and the only part where employment zones astride two areas of influence are observed. Thus workers from Vernon commute to work in Paris while others commute to Rouen, and there are likewise large flows of working population from Pithiviers or Montargis both to Orléans and to Paris. This type of profile is however fairly specific, and exchanges with Paris are increasingly predominant.

27Thus there are numerous exchanges on either side of the boundaries of the Paris Region, which is clearly in favour of the integration of the first ring of employment zones (at least) into the Paris functional space. While 22% of the working population in the Oise département work in the Ile de France Region, there are 50000 workers from the Sud-Oise employment zone alone, or one third of the local population, who commute to work in Ile de Frace, among whom 14000 to Paris itself. If we look further than this particular instance (since Sud-Oise partly belongs to the Paris urban area) the outside fringes are clearly distinct. The employment zones of the “outpost” towns and cities are thus, along with the large cities, the only ones to exchange with Paris, and they all do so. The only towns or cities in the Picardie Region to send a large proportion of their population to Paris, apart from Amiens, Sud-Oise and Santerre-Oise, are Beauvais, Compiègne, Soissons and Château Thierry. This is also true for the Haute Normandie Region, where Vernon and Evreux are the only two towns with the exception of Rouen, and Le Havre, exchanging with Paris, and it is true too for the Centre administrative Region. In addition, local flows occur in the outside and inside fringes of the Paris Region, with convergent flows from the employment zones in the outside fringes towards zones in the inside fringes, which virtually act as airlocks. Versailles, Sud-Oise, Cergy or even Melun thus receive numerous commuters from the outside fringes, such as Chartres, Dreux or Beauvais.

28The inside fringes also play a part in the flows observed within the Paris Region. In this space the commuting patterns form a core that functions as a network. Exchanges between employment zones in Paris and the inner ring are considerable, and both peripheral and radial. In contrast, if employment zones in the second and third ring are considered, it is the radial pattern that becomes predominant. Flows between neighbouring zones exist, but do not compromise this pattern, and the main flows from Cergy, Sud-Oise, Roissy, Lagny-sur-Marne, Meaux, Melun, Evry, Orsay and even Versailles are towards the centre (Paris or zones in the inner ring).

29In conclusion (figure 6) we can observe a major central space whose direct influence extends beyond the regional administrative boundaries, and whose presence is felt much further afield again. The functioning of this space is indeed based on the existence of a “hyper-centre” whose influence is relayed as far as a network of outlying towns and cities by way of poles located on the fringes of the area under the direct influence of the central space. Despite a material break in the urban fabric (reinforced by the fact that the outposts are located beyond the green belt around the Paris Region) these towns and cities are fundamentally parts of the “lived-in” spaces that come under the strong influence of Paris. We also observe a network of metropolises of lower rank that have a significant role locally, since they genuinely structure the spaces around them.

30Using home-to-work commuting patterns, we have thus constructed a relevant model of “lived-in” spaces on the scale of the Paris Basin. This model is based on the existence of a central metropolitan space extending over a radius of 70 km around Paris, and on regional spaces that organise themselves according to the relationship of their centres with the Ile de France centre.

Figure 6. Regional units around a vast Paris metropolitan Area

Figure 6. Regional units around a vast Paris metropolitan Area

A sectoral and functional analysis of the economy

31However this pattern of organisation is derived solely from the analysis of home-to-work commuting patterns, that is to say from an indicator that synthesises the functioning of spaces in the immediate proximity of the people. It is obvious that even if strong organic links might exist between Orléans and Caen, few individuals will commute between the two on a daily basis. These cities of some 250 000 inhabitants offer a proximity living environment, and they are not linked one to the other by fast transport facilities. Thus it is unlikely that there will be any significant commuting, while this in fact says nothing about any links that might exist in the system of production.

32Any such links are of course difficult to capture in the absence of data on telephone or trading flows. If we consider that there is a centre-periphery distribution, as expected in theory, one way of apprehending the issue could be to attempt to verify this hypothesis by using information on the structure of industry. It is possible to choose either an approach by way of socio-professional category, or an approach based on types of activity.

  • 7 The drawbacks of this choice should also be acknowledged: a single activity is allocated to a given (...)

33If the distinction between centre and periphery operates on a functional basis, with the centre directing and the periphery executing, an approach based on socio-professional categories should evidence a gradient from centre to periphery (Thiard, 2001). It will then be possible to check for the existence or otherwise of a centre-periphery structure, but this will mean that all trace of any other relationships that might exist between the spaces under study will be lost. The loss of information on sector proximities is a major disadvantage. The advantage of an approach based on activities is thus to retain the ability to analyse possible relationships between spaces by using the activities that they house. This variable also has the advantage of retaining traces of function within an activity, since manufacturing activities are distinguished from management and advisory activities in the nomenclature (Naf). When activities are integrated into enterprises, managing functions no longer appear separately. However the wave of externalisations in the 1980s and 1990s led to the development of a specific sector that is clearly distinct from executive activities broken down according to the activity. It is therefore on the structure by activity that we will conduct our study7. It is thus possible to study the spatial patterns of economic sectors in the Paris Basin.

Significant economic integration

  • 8 Starting from employment E in each zone k in region R, the Herfindahl index gives: HC=Σ(Ek/ER)2. It (...)

34To study these spatial patterns we used results obtained from the indices of distribution of activities in sectors and geographical zones based on the Herfindahl index8. Initially elaborated by Ellison and Glaeser (1997), these indices derived from the Herfindahl index have been used in the French setting by Maurel and Sedillot (1997) with a few modifications. These indices, calculated in a probabilistic perspective, enable the characterisation of spatial concentrations of activities after eliminating structural bias: the concentration of a given sector is apprehended in relation to a situation of even distribution across the different territories.

35Any element drawing away from this even distribution but not connected with the dynamics of economic agglomeration can be linked to the geographical structure of the space considered, within which industrial activity can exhibit variable degrees of concentration. The spatial concentration of a given sector of activity is likely to be greater in France, where the Paris Region carries considerable weight in national employment, than in the Netherlands where employment is more evenly distributed across regions. The Herfindahl index of a sector i, Hi, is thus modified according to the Herfindahl index calculated for all the sectors of activity, HC. The corrected index G is obtained as follows: G= (Hci)HC)/1-HC. It is positive if the sector of activity considered is more concentrated in that space than activity as a whole.

  • 9 If Xz is the number of jobs in enterprise z in sector i and Ei is the number of jobs in sector i, t (...)

36Bias can also be related to the structure of industrial activites. This occurs when the spatial concentration is wholly explained by the presence of one or two very large firms employing the majority of the labour force in the sector of activity. Thus the tyre industry in France is almost totally concentrated in Clermont Ferrand. But this is the result not so much of an agglomeration dynamic (which would be shown by a cluster of businesses) as that of the weight of the main firm, Michelin, in national employment in that sector. The corrected index therefore takes account of the effect of the distribution of jobs across the different firms. A Herfindahl index is again used here, but rather than being calculated from employment per zone in relation to overall employment, the H index is obtained from employment per firm in relation to employment overall in the sector of activity to which it belongs9. The index obtained is y=(G-H)/(1-H).

37Using the distribution of numbers of workers among the different business enterprises, these indices enables a characterisation of the spatial concentration of an activity, thus rendering it comparable with that observed in other spaces of different size and industrial structure. In this respect, these indicators are very useful for apprehending the degree of agglomeration of the economic activity in a given space. The closer the index is to 1, the more concentrated are the activities in a single spot. Thus a high index value reflects the fact that there is a single cluster for a given activity in a given space.

38Since here the same indices as those implemented by Maurel and Sedillot (1997) were used, the first step in the analysis is to compare the results. They are nevertheless derived from different databases: the activities overall considered here comprise industries (not including extraction activities), to which we added services (not including administrations and retail trade). The study by Maurel and Sedillot (1997) did not take account of these service activities. The results obtained are therefore not easy to evaluate in terms of actual values, but they are nevertheless comparable – they will systematically tend towards the same direction. The very fact of introducing services into the calculation, more concentrated, results in the mean concentration of activities in the Paris Basin being always be greater. For a given activity, the adjustment made to take account of concentration in the Paris Basin will thus always yield a value greater than it would be if the service activities had not been taken into account, and the concentration index for an industrial activity will be reduced by the integration of the localisation of services into the calculation. The Paris Basin indices for industrial sectors of activity should theoretically be lower for the same type of distribution of activities across the space. Despite this, the indices observed for the activities present in the two data bases are identical, or higher in the case of the Paris Basin. This means that, all other things being equal, activities are more concentrated in this space.

39Once the biases mentioned above have been corrected, the level of concentration observed is linked to the choice of localisation of business enterprises. If a single localisation is favoured above others, the activity will be concentrated; if several alternative localisations appear the activity will be de-concentrated. In an integrated economic region, the logics of concentration and specialisation tend to favour the choice of a single site. The presence of a second site for a particularly concentrated industrial activity can be interpreted as an encroachment on a second functional region. Since greater economic integration in theory translates into an increase in the concentration and specialisation of the spaces concerned, the Paris Basin forms a more integrated economic zone than the national space as a whole.

40In the perspective of an economy that is increasingly dependent upon local dynamics (Marshall, 1898, Jacobs 1969, Baro & Sala-i-Martin 1995) indices of specialisation calculated in similar manner can evidence zones in which activity is specialised in a few sectors, without this reflecting an unbalanced industrial structure in favour of just a few firms. The advantage of this method is that it enables different levels to be studied. It is thus possible to study the specialisation of départements, of employment zones, or even communes in the Paris Basin. Although the figures obtained on the different geographical levels cannot be compared, they nevertheless enable the characterisation of differing spatial patterns.

A centre-periphery type of organisation.

41The local specialisation indices developed by Houdebine (1999), constructed in completely symmetrical manner, are interpreted in similar manner, and correct the same biases as the concentration indices. Specialisation in a zone is adjusted by way of the weight of a sector in the overall space considered, and the size of firms in that sector. A high specialisation index will be obtained either from very marked specialisation that is not offset by the successive adjustments, or from a local employment setting that is liable to be quite diverse, but within which clusters of firms are rendered prominent by the adjustments. In this second instance, the index shows the potential importance of scale economies: the local specialisation will be accounted for not by one large firm, but by a large number of firms, and by a fabric that is objectively varied.

42The specialisation indices calculated for départements show a clear centre-periphery structure: the centre presents high index values, the periphery low values, and the fringes high values again.

Figure 7. Agglomerations of specialised industries on the level of the départements

Figure 7. Agglomerations of specialised industries on the level of the départements

43The central départements in the Paris Basin (Paris, Hauts-de-Seine, Val-de-Marne, Essonne, Yvelines and Val d’Oise) owe their index values to a combination two parameters - employment being specialised in just a few activities, and numbers of jobs being distributed homogeneously among the different firms. Thus the indices are naturally quite high, and undergo only slight adjustment when the industrial structures are taken into account. The high values for these indices are also related to the activities involved, which are generally very concentrated. The fictitious job numbers on which these indices are calculated are therefore larger than real job numbers.

44Next there is a vast area covering almost the whole of the Paris Basin and in which index values are lower. This can be explained by the fact that despite an employment structure that is not very different from that of the centre, numbers are concentrated in large enterprises. The adjustment of concentration indices according to industrial structure therefore has considerable effect on these index values. The fact that the activities are relatively similar, while the industrial structures are markedly different, suggests a centre-periphery organisation where the activities are identical across the whole region, but where there is firstly a specialisation in the centre on activities based in small structures, and secondly a specialisation of the periphery in functions based on scale economies, and a dual concentration of these two functioning modes in the two regional sub-spaces, with small structures not occurring widely in the periphery and large structures being absent from the centre.

45Finally, on the fringes of the Paris Basin, high values are once again observed. These are linked to very marked local specialisations. The départements of Manche, Marne, Ardennes or Haute-Marne all exhibit industrial structures that differ little from that of the neighbouring départements, but as they are much more specialised, their weighted indices remain high. The dual dynamic of concentration and specialisation is also notable, for instance for shipbuilding (Manche) or champagne (Marne). These activities have a very large impact on the local spaces, which are virtually the only spaces involved in the activities considered.

Significant specialisation of the employment zones

46If the employment zones are considered (Figure 8) the centre and the fringes once again appear, and for the same reasons. The contour of the centre is refined: Paris, Nanterre, Boulogne, Orly, Orsay, Versailles and Roissy. It also includes fringes, since the south of the Centre Region appears. But the main characteristics of the overall central space vary little. The profile of the periphery, in contrast, changes very considerably. While the départements all have fairly similar specialisation indices, the employment zones for their part show much wider variation. This wide disparity is masked on the level of the départements: several highly specialised zones, each in a different activity, when aggregated generate a département that is not particularly specialised. This occurs in the majority of the employment zones around the western part of the Paris Region – from Beauvais to Pithiviers via the Perche area, specialisation indices in some localities are very high. In these employment zones profiles are similar to those in the fringes: specialised employment in very few activities grouped in large firms. However, the fact that the départements do not reflect this local specialisation means that each employment zone is specialised in a different activity from that of its neighbour. The apparent centre-periphery structure, relatively coherent from a functional point of view, in fact conceals very marked heterogeneity in the activities involved. Thus the peripheral areas are also specialised at a more refined local level.

Figure 8. Specialised industrial agglomerations in employment zones

Figure 8. Specialised industrial agglomerations in employment zones

Figure 9. Specialised industrial agglomerations at the municipal scale (communes with more than 250 jobs)

Figure 9. Specialised industrial agglomerations at the municipal scale (communes with more than 250 jobs)

47If we now consider the communes, and with the exception of a few rural communes reaching the selected threshold of 250 jobs (revealing the presence of a big factory in a small village), it is above all in the urban areas (the outlines are shown in pink in figure 9) that jobs are concentrated. Inside each urban area a centre-periphery structure is once again observed. The centre houses diversified activities, and presents a structure in which large firms are very few, while the suburbs present more specialised profiles, and are more dependent upon just a small number of firms. The peri-urban spaces are similar to the rural spaces, that is to say that when there are jobs in industry, these are generally concentrated in one or two firms, overall numbers rarely exceeding 250 jobs.

48The only exception – but it is of consequence – is the Paris Region. The central space is markedly specialised, but exhibits a diversified industrial structure, while the suburbs are characterised by large firms, and the periphery is more empty. However on the fringes of the urban area there is what looks rather like a ring of communes offering more than 250 jobs. Since these communes are not clustered along a communication axis, (except for the Oise valley) or around a pole, the only element that these spaces have in common is their distance from Paris. In addition we can see here a phenomenon that we noted above: a density of employment outside the urban areas that is characteristic of the “outpost” town employment zones. This phenomenon indeed creates certain problems for the towns concerned, since it questions their ability to polarise employment and population generated by the extension of the Paris Region.

49If we now operate a change in scale, and return to the regional space rather than to the urban spaces in the Paris Basin, a centre-periphery structure no longer appears, the general layout suggesting a reticular structure. In fact, the urban areas appear as anchorage points for the activities in the peripheral regions. It can thus be noted that in the Paris Basin, the specialised employment zones never correspond to zones that include large urban areas. Among these urban areas, Caen, Le Mans or Reims are those presenting the highest index values, and they all differ from their surrounding spaces by their lower values. The commune grid provides and explanation: in the urban areas, activities are relatively diversified and the industrial structures are balanced. In addition to the centre-periphery structure that characterises relationships of the Paris Region with employment zones in the Paris Basin, it therefore appears that there is an underlying network of cities.


50Finally, the Paris Basin is an area that articulates around a central space possessing a very wide influence zone (Figure 10). The core, made up of the Paris administrative Region and extended to the zones coming under the influence of the outpost towns and cities, entertains direct relationships with the large regional cities which for their part structure their local spaces very markedly. The geography derived from these proximity relationships should however also be apprehended via an approach based on industry. A study of the spatial distribution of industrial activities thus confirms the integrated functioning of the Paris Basin overall. The general pattern of this organisation is of the centre-periphery type. The centre of the Paris Region (the Paris, Nanterre, Boulogne, Orly, Orsay, Versailles and Roissy employment zones) is specialised in the most concentrated activities, operating in small structures. The industrial structure of the periphery is generally unbalanced with just a few large firms, even if as a whole it presents the same diversity as the centre, and is positioned on the same activities (at least in the industrial sphere). Marked specialisations may appear locally, but they tend to be characteristic of the most vulnerable spaces, those depending on a small number of activities and a small number of firms. This result is fairly classic, and not necessarily negative in the perspective of an integrated economic space (Henderson, 1988). In the peripheral areas, the only spaces presenting relatively balanced profiles are the large cities. These thus provide leverage for the vitality of the centre, all the more so because they also entertain preferential relationships with it.

Figure 10. Organisation of the industrial space in the Paris Basin

Figure 10. Organisation of the industrial space in the Paris Basin
Haut de page


Barro R. , X. Sala-i-Martin, 1995, Economic Growth, MIT Press, Cambridge.

Beaverstock J., Smith R. , Taylor P., 2000, World city network: a new meta-geography?, Annals of the Association of American Geographers, 90, 123-34

Brunet R., Dollfus O., 1990, Géographie Universelle : Mondes nouveaux, Belin et Reclus.

Cattan N., Pumain D., Rozenblat C., Saint-Julien T., 1999, Le système des villes européennes, Anthropos, Paris, Economica.

Ellison G., Glaeser E., 1997, « Geographical concentration in US industries : A dartboard approach », Journal of Political Economy, 105, 889-927.

Evrard A., 2002, Projections de population à l’horizon 2030 : une croissance dans la dynamique du Bassin parisien, Insee Picardie relais, 102.

Frémont A., 2002, Le Bassin parisien, La Documentation française.

Fujita M., Hamaguchi N., 2001, Intermediate goods and the spatial structure of an economy, Regional Science and Urban Economics, 31, 79-109.

Gilli F., 2002a, « Déplacements domicile-travail et organisation du Bassin parisien », L’Espace Géographique, 2002-4, 289-305.

Gilli F., 2002b, Essai de caractérisation de l’espace industriel du bassin parisien, in Le Bassin parisien, La Documentation française.

Gilli F., 2003, Les villes avant-postes, une alternative à de nouvelles villes, Urbanisme, à paraître.

Gravier, 1947, Paris et le désert français, Flammarion, Paris.

Hautreux J., Rochefort M., 1964, La fonction régionale dans l’armature urbaine française, Ministère de la Construction, Paris

Hecquet V., Lainé F., 1999, Structures industrielles locales et formes d’organisation économiques, Economie et Statistiques, 326-327, pp 205-224.

Henderson V., 1988, Urban development; Theory, Fact and Illusion, Oxford University Press.

Houdebine M., 1999, Concentration géographique des activités et spécialisation des départements français, Economie et Statistiques, 326-327, pp 189-204.

Jacobs J., 1969, The economy of cities, Random House.

Kloosterman R.C., Lambregts B., 2001, Clustering of Economic Activities in Polycentric Urban Regions: The Case of the Randstad, Urban Studies, 38 (4), 717-732

Krugman P., 1993, First nature, second nature and metropolitan location, Journal of Regional Science, 33, 129-144.

Krugman P., 1991, Increasing returns in Geography, Journal of Political Economics, 99, 483-499.

Leroux L., 2002, Les déplacements domicile-travail en Picardie : mobilité intense et clivages entre les trois départements, Insee Picardie Relais, 103.

Marshall A.,1898, Principles of Economics, McMillan, London.

Maurel F., Sedillot B., 1997, « La concentration géographique des industries françaises », Economie et Prévision, 131, 25-45.

Puga D., 1999, The Rise and Fall of regional Inequalities, European Economic Review, 43, 303-334

Pumain D., 1986, Les migrations interrégionales de 1954 à 1982: directions préférentielles et effets de barrière. Population, 2, 378-388.

Sassen S., 1991, The Global City, Princeton, NJ: Princeton University Press

Taylor P., 2001, Visualizing a New Metageography: Explorations in World-City Space, in Dijkink G. et H. Knippenberg (eds), The territorial factor: political Geography in a globalizing world, Amsterdam: Vossiuspers UvA, 113-28.

Thiard P., 2001, Les dynamiques spatiales du bassin parisien (1975-1990). Un système spatial entre mutations du système productif et impact des politiques publiques, Thèse de doctorat, Université Paris 1.

Thisse J-F.,1997, De l’indétermination des régions et de quelques inconvénients qui en résultent, l’Espace Géographique, 2, 135-148.

Venables A., 1996, Equilibrium location and vertically linked industries, International Economic Review, 37, 341-359.

Haut de page


1 This work was conducted when the author was working for the Direction Régionale de L’INSEE Picardie

2 The term “bassin Parisien” translated hereafter as “Paris Basin”, while first of all referring to a geo-morphological structure, is widely used in French to refer to the socio-economic space under the influence of Paris, very roughly corresponding to the same boundaries.

3 The French “zones d’emplois” or “employment zones” correspond to a statistical subdivision of the French territory based on commuting patterns.

4 After removal of the complex case of Paris, the only urban area that is divided into several employment zones

5 It should be noted that adopting a single criterion for commuting patterns restricts the heuristic impact of the “lived-in spaces” thus outlined. The employment basins account for only part of the activities (working activities) and there is no objective reason to suppose that they correspond exactly, for instyance, to retail catchment areas or public services cover. It is nevertheless the information that is most easily mobilised, and that concerns regular and unavoidable travel to which numerous secondary activities are often attached.

6 The model that was used as the basis for this illustration is a classic gravitational model calibrated on commune to commune commuting patterns in the Paris Basin in the 1999 census. It explains 35% of the flows observed (partly on account of the presence of Parisian flows which have particular features) and is significant at the threshold of 1/1000: log (flow)=0.48 (working population in workplace) + 0.04 (resident working population) – 0.78 (distance) + log (1.55). This figure, and also figure 6, have already been published in the journal Espace Géographique (Gilli, 2002a)

7 The drawbacks of this choice should also be acknowledged: a single activity is allocated to a given firm, and there is no clear logic in the NAF classification enabling an operational grouping of activities that are linked one to another. Concerning this particular point, we have attempted to re-aggregate NAF600 activities so as to form coherent industrial entities.

8 Starting from employment E in each zone k in region R, the Herfindahl index gives: HC=Σ(Ek/ER)2. It varies from 1/n to 1, where n is the number of zones in the region.

9 If Xz is the number of jobs in enterprise z in sector i and Ei is the number of jobs in sector i, the Herfindahl index calculated for the enterprises gives : HΣ(Xz/Ei)2.

Haut de page

Table des illustrations

Titre Table 1 – Jobs in the Paris Basin: Paris and the main cities
Fichier image/png, 61k
Titre Figure 1. Share of salaried workers working in Paris
Fichier image/jpeg, 168k
Titre Figure 2. Working population working in the Paris Region.
Fichier image/jpeg, 164k
Titre Figure 3. Attraction and barrier effects in the Paris Basin
Fichier image/jpeg, 200k
Titre Figure 4. The main journey-to-work flows between employment zones in the Paris Basin
Fichier image/jpeg, 188k
Titre Figure 5. The main journey-to-work flows between employment zones inside the Paris Region
Fichier image/jpeg, 180k
Titre Figure 6. Regional units around a vast Paris metropolitan Area
Fichier image/jpeg, 380k
Titre Figure 7. Agglomerations of specialised industries on the level of the départements
Fichier image/jpeg, 112k
Titre Figure 8. Specialised industrial agglomerations in employment zones
Fichier image/jpeg, 148k
Titre Figure 9. Specialised industrial agglomerations at the municipal scale (communes with more than 250 jobs)
Fichier image/jpeg, 268k
Titre Figure 10. Organisation of the industrial space in the Paris Basin
Fichier image/jpeg, 216k
Haut de page

Pour citer cet article

Référence électronique

Frédéric Gilli, « The “Paris Basin”, a metropolitan region », Cybergeo : European Journal of Geography [En ligne], Espace, Société, Territoire, document 578, mis en ligne le 22 décembre 2011, consulté le 16 décembre 2017. URL : ; DOI : 10.4000/cybergeo.24913

Haut de page


Frédéric Gilli

chercheur associé au Centre d’Etudes Européennes, Sciences Po Paris,

Articles du même auteur

Haut de page

Droits d’auteur

© CNRS-UMR Géographie-cités 8504

Haut de page