1 Introduction

The urban mosaic is in a continuous process of change. Immigrants cluster together and mix with others; ethnic and racial groups are segregated in ghettos and slums, or they are able to escape to more liveable neighbourhoods; low-income households are forced to move to other neighbourhoods, because of urban restructuring processes; and expensive accommodations are continuously built on the most attractive sites within cities, in order to attract higher income groups. Existing neighbourhoods see their function in the urban mosaic change and experience the in- and outflow of population groups with different incomes, colour, age and household composition. They might develop from attractive neighbourhoods areas to ghettos and slums, but also the other way around.

It is nothing new that cities are divided into different kinds of areas. It is also nothing new that urban neighbourhoods continuously change in terms of function and population structure. What might be new, however, is that these changes have different causes now than in previous periods. A main argument could be that in the last three decades or so the process of globalisation has become enormously influential in explaining changes within cities. Authors like Castells (2000), Taylor (2000) and Sassen (1991, 2006) have mentioned globalisation processes as the driving force behind urban change, but also other authors, such as Ulf Stare (2004) on Stockholm and Paul Jenkins and Wilkinson (2002) on South African cities, focus on globalisation as an essential determinant of urban change. Even for Chinese cities the importance of globalisation in urban development has been mentioned (Lin, 2004).

The argument in this paper will be that attention for globalisation is useful, but that we should never exaggerate the influence of this process in a city as a whole and in parts of that city. In other words: we want to challenge the importance of globalisation when explaining urban change. Our focus will be on change in urban neighbourhoods.

In this paper we will first pay attention to the concept of the divided city. Then we will show briefly that for a long time divided cities have been explained without referring to any macro-development or megatrend at all, let alone globalisation. Still, however, many explanatory accounts are partially, or in one way or another, still valuable, as we hope to show. After this overview we will focus on the possible role of globalisation in explaining urban change. This section will be followed by a critical account of this principal role of globalisation. In this section we will focus again on the role that older theoretical notions might play and on various contingencies.

2 Divided cities

CitiesFootnote 1 consist, almost by definition, of various neighbourhoods, each with its own function, nature, architectural style, attraction, and advantages and disadvantages for various residents and visitors. In other words, the undivided city is a myth and a utopia at the same time.

In the last century, a large quantity of typologies of urban neighbourhoods have seen the light of day, some specifically for cities and most with the pretension of having more general, or even complete, validity. Almost always there is a certain dichotomy in these typologies: between prosperous and impoverished areas, between areas where only those with a high income can afford to live and areas where the poor are more or less forced to live because of the lack of alternatives elsewhere. These cities are described as divided cities (Fainstein, Gordan & Harloe, 1992), dual cities (Mollenkopf & Castells, 1991), polarised cities, fragmented cities (Burgers, 2002) and partitioned cities (Marcuse & van Kempen, 2002a). In most cases in this sort of study, a clear connection is made between a divided society and a divided city: if a society is divided, the urban space must also be divided. It is a matter of the connection between social polarisation and social inequality on the one hand and spatial segregation on the other (see, for example, Hamnett, 1994).

It is easy to come to the conclusion that a divided city exists. Within cities many neighbourhoods will differ from each other fundamentally. The concept of a dual city, however, actually seems unusable. The idea that the well-to-do live in one part of the city while the other part of the city is at the disposal of the poor ignores the fact that an important part of the city is accessible to and inhabited by households that are neither truly prosperous nor particularly poor (Marcuse, 1989). Certainly in Western cities there is no sign of a simple dichotomy between prosperous and poor: there is always evidence of a substantial middle class. A more general criticism levelled at the typologies mentioned is that they are far from being universally verified empirically. Whenever such a typology is tested empirically, it soon transpires that the reality is a good deal more complicated, or that at the particular place where the typology was tested the reality was substantially different (see, for example, Baum, Mullis, Stimson & O’Connor (2002) who tested the typology of Marcuse and van Kempen (2000b) in Australia).

Divided cities imply the existence of different areas within a city. These all have different characteristics; they may be prosperous or poor, deprived or privileged, old or new, dominated by high rise apartment blocks, or by single family homes, or contrasted on countless other factors. In this paper my focus is on these different areas within a divided city; I have chosen the term neighbourhood by which to refer to them.

What is a neighbourhood? There are countless definitions.Footnote 2 Megbolugbe, Hoek-Smit and Linneman (1996) give a clear overview. A first possibility is to consider a neighbourhood as a homogeneous area in terms of demographic and housing characteristics. Such a definition would exclude many areas within a city, because many urban neighbourhoods are in fact characterised by heterogeneity. A second possibility is to see a neighbourhood as an area in which the residents have a feeling of shared identity. Here again, such a definition would exclude many areas, certainly at a time in which individualism is in its heyday (see also Lofland, 1989). There is a dull, but certainly functional, third possibility: a neighbourhood as a small, statistically defined area. For this paper, I have used this neutral definition. Usually such a definition is derived from a public institution, such as a local authority. Such a description will involve all parts of a city and that is a great advantage. Let us assume that a neighbourhood consists of at least 2,000 dwellings (see also Power, 1997) and at most about 10,000. A neighbourhood is usually bounded by something physical such as a railway, roads, a park, or water.

3 Explaining urban and neighbourhood change

In explaining patterns and processes of spatial concentration and segregation, the story starts with three “traditional” spatial analysis approaches: human ecology, social area analysis, and factorial ecology. Good descriptions of these approaches have already been made elsewhere (see, e.g., Bassett & Short, 1980; Sarr, Phillips & Skellington 1989). Here we will only summarise the basic ideas of these approaches and the main criticisms (see for more elaborate accounts: Murie, Knorr-Siedow & van Kempen, 2003; van Kempen, 2002; van Kempen & Özüekren, 1998).

3.1 The human ecology approach

The enormous influence of the Chicago School with its human ecology approach is well known. The structural analysis of neighbourhood change, residential differentiation, and the concomitant processes of spatial segregation and concentration started with the human ecology tradition associated with the Chicago School (see, e.g., Burgess, 1925/1974; McKenzie, 1925/1974; Park, 1925/1974). The city developed through a competition for space to produce concentric zones (Burgess, 1925/1974), specific sectors (Hoyt, 1939) or multiple nuclei (Harris & Ullman, 1945), housing households with different resources and other characteristics. Processes of invasion and succession involved a chain reaction, with each preceding immigrant wave moving outwards and being succeeded by more recent, poorer immigrants (Park, Burgess & McKenzie, 1925/1974). The final pattern of segregation, the “mosaic of social worlds” (or a residential mosaic; see Timms, 1971) was seen as a “natural” equilibrium. It was a consequence of various processes: invasion, dominance and succession. Behind it was the idea of immigrant enclaves as transitional stages on the road to eventual acceptance and integration in the larger (American) society (Clark, 1996, p. 110).

The classical formulation of human ecology was criticised for being derived from a biological model, rather than being based on cultural and social processes (Firey, 1947; Jones, 1960; Wirth, 1944). It is argued that the social ecologists paid too little attention to how neighbourhood change actually occurs. Their explanations were insufficiently informed by empirical research referring to choice, preference and social action (e.g., Hollingshead, 1947). Moreover, their neglect of the influence of institutional and political factors, including the national and local state, rendered their account unsatisfactory, even for cities where state intervention through planning and housing provision was limited (Bassett & Short, 1989). Macro-developments or megatrends, let alone globalisation, were not mentioned. The ideas of the Chicago School were essentially American, and maybe even “Chicagoan”. They were developed in a specific time period and under a specific system: the free market economy, in which terms such as social security and housing subsidies were not common and the role of the state in general was marginal. At the same time, some crucial concepts of the human ecology approach are still important. Divided cities cannot be explained without using concepts like invasion, succession and filtering.

3.2 Social area analysis and factorial ecology

The human ecology approach was followed by positivistic-empirical approaches like deductive social area analysis (e.g., Bell, 1953; Shevky & Williams, 1949; Shevky & Bell, 1955) and inductive factorial ecology (e.g., Berry & Kasarda, 1977; Murdie, 1969; Robson, 1969). Factorial ecology uncovered the socio-spatial layout of many cities in the world, though without focusing on causality (De Decker, 1985). Census variables were selected and “run through the statistical mill of principal components analysis or factor analysis” (Bassett & Short, 1980). Many analyses revealed sectoral and zonal patterns. Differences between urban neighbourhoods could often (but not always; see Robson, 1969) be summarised by three sets of variables: socioeconomic status, family status and ethnicity (Bell, 1968).

Critics of social area analysis and factorial ecology said it was descriptive and based on very meagre theoretical notions (Bassett & Short, 1989; Hawley & Duncan, 1957; Kesteloot, 1980; Yeates, 1989). Again, the possible role of the state was never mentioned, neither did the analysts working in this framework pay attention to macro-developments. Other critics pointed out that since most of the research had been done in the United States, the results should not be applied automatically to the European situation (see O’Loughlin, 1987). Finally, to the extent that this approach has a theoretical base, it is that the subject is a homo economicus: a fully informed individual with a perfect ability to act in an economically rational way (see Bolt & van Kempen, 1997).

3.3 Introducing the individual: preferences and choice

Since the 1970s it has been realised that decisions and the behaviour of individuals and households can have major effects on urban divisions. These ideas have been mainly developed in the housing literature, but can often without any problem be used in urban theory. In behavioural approaches developed in the housing field, choices of households are directly linked to positions and events in the family life cycle (see, e.g., Clark & Dieleman 1996; Clark, Duerloo, & Dieleman, 1997). Household characteristics are major determinants of housing (and locational) preferences (Adams & Gilder, 1976; Clark, Duerloo, & Dieleman, 1986). Phases in both the household and labour market career influence the household’s size and its preferred type of dwelling and place to live (Rossi, 1955; Speare, Goldstein & Frey, 1975; Stapleton, 1980).

The role of ethnic and ‘racial’ factors comes back in a specific behavioural approach: the ethnic-cultural approach. The general argument within this approach runs thus: housing conditions and residential patterns differ between groups, and these differences can be attributed to cultural differences between these groups. There is a clear element of “choice” in this approach, although it allows for the inclusion of constraints in the explanation (van Kempen & Özüekren, 1998). The ethnic-cultural approach opens our eyes to the fact that Asians, Turks, Moroccans and West Indians—to mention just a few of the backgrounds of immigrants in West European countries—are not single groups with unitary values (see also Özüekren & van Kempen, 2003). All kinds of subgroups might attach different meanings to many aspects of life, including the kind of housing and neighbourhood in which they would like to live (see, e.g., Ballard, 1990).

The behavioural approach has been criticised for its emphasis on demand and choice and the lack of attention for constraints (see, e.g., Hamnett & Randolph, 1988). Attention for developments and factors outside the individual, let alone megatrends like globalisation, do not get any attention.

3.4 Constraints and opportunities

Choices take place in an environment of constraints. This simple idea is the basis of Rex and Moore’s Race, Community and Conflict (1967), which is generally seen as the beginning of the neo-Weberian or institutional approach in housing research. This work is grounded in the idea that housing, and especially desirable housing, is a scarce resource and that different groups are differentially placed with regard to access to these dwellings, in which constraints play a major part. People are distinguished from one another by their strength in the housing market (Rex, 1968) and this causes better or worse possibilities to live at one place or another.

Different resources of households can be identified (see van Kempen & Özüekren, 1998 for a more elaborate overview). Financial resources refer to income, security of income, and capital assets. Cognitive resources include education, skills, and knowledge of the housing market. Political resources refer to the political power people wield, either formally or informally. And social resources refer to the contacts people have, which may help them to find suitable housing and places to live. Even the present housing situation can be seen as a resource. All these resources are highly influential in explaining the possibilities of households in the housing market and therefore form an important part of the explanation of spatial patterns within cities.

3.5 Institutional approaches

When talking about the division of households and individuals over space, the crucial role of the state and state institutions has been elaborated in the institutional approach. In some countries the welfare role of the state has always been very small: the support for the poor and provision of subsidies for the needy has always been narrowly limited. In other countries there has been an elaborate welfare state, especially since World War II. The main principles of the welfare state have always been twofold. The first encompasses an elaborate system to support those who are in a weak position on the labour market, which for instance meant financial support in situations of unemployment and illness, and support for the elderly. The second principle is the elaborate system of subsidies in all kinds of fields like housing, recreation, and social work.

Especially since the middle of the 1980s, the welfare activities of states have been declining or retreating. The need for cutbacks and the new liberal philosophy in which a declining role of the state was central were responsible for this. Declining incomes, especially of those who are dependent on some form of state allowance, have been the direct effects of cutbacks in government expenditure. Declining subsidies further diminished the possibilities of all kinds of households on different types of market, among them the housing market. In many Western countries, these cutbacks have intensified again since September 2001.Footnote 3

Governmental cutbacks can directly affect the incomes of those who are dependent on the state, like the unemployed, the elderly, and the handicapped. Declining incomes can directly influence the housing market opportunities of households, because they are relegated to those dwellings they can afford to pay. When these dwellings are spatially concentrated in certain areas of the cities, a new spatial division of cities may result, characterised by an increasing concentration of relatively poor households in areas with low-rent (and often low-quality) dwellings and a growing segregation of different income groups. Spatial divisions may become sharper, because of an increasing polarisation of incomes.Footnote 4

Within the institutional approach, it is worth mentioning the managerialist approach of Pahl (1975, 1977) and Lipsky (1980). Pahl suggests that social gatekeepers (like housing officers) can allocate resources according to their own implicit goals, values, assumptions and ideologies. This means that stereotypes and racism might influence their decisions (Tomlins, 1997). Local government or housing associations might decide to allocate dwellings in a certain neighbourhood exclusively to non-immigrants. Studies in the United States have shown that real estate agents are primary information brokers and major agents of change (e.g., Galster, Freiberg, & Houk, 1987; Turner & Wienk, 1993).

4 Divided cities: the influence of globalisation

It is not a very new idea that cities are part of a larger society:

“... cities are the central elements in the spatial organization of regional, national, and supranational socioeconomies by virtue of the interregional organization in a total ‘ecological field’ of the functions they perform” (Berry & Kasarda, 1977, p. 85).

Because of this interrelatedness, areas within the city are also influenced by developments and decisions on higher spatial levels. Many sociologists and geographers now agree that patterns of segregation and concentration change as a consequence of individual household decisions in response to the complex interaction of a variety of structures and developments on different spatial levels. General processes—like economic restructuring on a global level—have their impact on local situations and developments (Sassen, 1990; van Kempen & Marcuse, 1997; van Kempen & Özüekren, 1998) and on choice patterns of households (Clark et al., 1997). In order to explain (spatial) changes on the local level, we have to incorporate structures and developments on other spatial levels (see also Sarre et al., 1989; Karn, Kemeny & Williams, 1985). In the words of Phillips and Karn, we should be aware of:

“ ... a close and dynamic relationship between individual strategies, institutional behaviour and the wider social, economic and political structure” (Phillips & Karn, 1992, p. 358; see also Clark & Dieleman, 1996, p. 137).

The causes of changes within cities can for an important part be traced back to developments that take place on higher spatial levels, at least regionally but even more critically on the level of the nation or even the world. The latter, with their concomitant national and regional implications, is today generally subsumed under the concept of globalisation, a term that is often used, but not always well defined. Globalisation can comprise many processes, such as the changing spatial structure of economic activities, migration of capital, migration of people, and values and norms that spread along various parts of the world. We take it here to mean globalisation in its present configuration, that is, a combination of new technology, increased trade and mobility, increased concentration of control, and reduced welfare-oriented regulatory power of nation states.Footnote 5

Globalisation has clearly much to do with mobility of goods, of capital, of persons. The globalisation of production has for instance led to the movement of shipyards and textile production from Western European countries to countries on the Pacific Rim. Effects of this process may influence the labour market position of people in both areas. Effects of migration are also on both sides, of course. Sending countries, which are often poor countries or countries with oppression, are losing people, while other areas, mainly the more wealthy countries, receive people who have to find a home. The influx of people in a country may change the character of cities and neighbourhoods, because these areas become inhabited by new kinds of people.

Through these developments, deeper divisions are created between on the one hand areas that flourish and on the other hand areas that seem to miss out in every respect and have no function in the new network society. We find this same dichotomy on various spatial scale levels: on a world scale, a national scale, and even within cities: “Indeed, we observe the parallel unleashing of formidable productive forces of the informational revolution, and the consolidation of black holes of human misery in the global economy, be it Burkina Faso, South Bronx, Kamagasaki, Chiapas, or La Courneuve” (Castells, 2000, p. 2). Of course, we must avoid too great an oversimplification of the reality: the dichotomy will not be absolute.

Globalisation can even influence life within neighbourhoods. Basically, two lines of reasoning can be discerned. The first follows Robert Reich’s The Work of Nations (1991). He argues that local forms of social solidarity become less important, because elites show an increasing international orientation and are becoming less dependent on the services of the lower status groups in neighbourhoods. There is no need to live in close proximity to them and if they live in the same neighbourhoods, the life world of the wealthy is clearly larger than their living neighbourhood. Webber’s (1964) old idea of ‘communities without propinquity’ seems to become more important for those at the upper end of the economic spectrum today. For the very poor, by the same token, their spatially defined neighbourhoods become more and more irrelevant to the functioning of the mainstream economy. The location of either with relation to the other recedes dramatically in importance. A logical result is an urban society that is increasingly socially and spatially disconnected, fragmented and polarised.Footnote 6

The second line of reasoning focuses on globalisation as leading to a kind of socioeconomic symbiosis within an increasingly polarised society, which can be seen in a growing number of highly educated, wealthy persons and households, but also in an increasing number of people in the lower segments of the economy (those in dead-end jobs and the chronically unemployed). Sassen (1984, 1986, 1988, 1991) can be seen as the main proponent of this argument. The crux in her line of reasoning is that rich and poor, those included in and those excluded from the (formal) economy are dependent on each other. One group has the money for products and services that the other group can provide. The emphasis on symbiotic relationships might end up with a society that is both more polarised and more interdependent and with spatial patterns characterised by a spatial mix of different groups.

5 A contested role of globalisation?

Although probably nobody van deny the existence of globalisation, we do have important questions concerning the effects of this process. We do not believe that globalisation always and automatically results in the same spatial patterns. Even when globalisation leads to a situation of increasing social polarisation, symbiosis between groups might lead to urban areas where people with different incomes, ethnicities, skills and education all live together. In other cities, the same social process might very well lead to different consequences for different groups, leading some to form enclaves and others to be confined to ghettos. Any automatic reasoning from megatrends through social processes to spatial patterns should be watched with great care.

The main reason behind the idea that globalisation does not explain everything is the fact that other developments are also at play, either independently or mediating its influence. Each of these developments is directly or more indirectly linked to global processes, but it is useful not to treat globalisation as a unitary and all-explaining process. It is better to examine its relevant components together with other, not necessarily global, changes also taking place today. To our opinion, we should at least take seven contingencies into account:

  • The physical setting of a city

  • History

  • Economic development

  • Inequality

  • Race and racism

  • Political power

  • Governance

Some of these contingencies are directly related to the “older” theories mentioned above. Below we will elaborate on these contingencies.

The development of a city will be heavily influenced by the first contingency, its physical setting. In the Dutch city of The Hague, for example, it is virtually impossible to build westward, because it borders on the sea, while the same holds for Barcelona in the other direction. Many cities are at least on one side bordered by mountains. For many cities their relation to their harbour (Liverpool, Rotterdam, Antwerp, Hamburg, etc.) is crucial. On a lower spatial scale, the existing location of major buildings, residential settlements, transportation infrastructure and utilities will guide future urban developments: they can heavily constrain the impact of forces of change. These constraints are ultimately essential for a detailed understanding of developments in any particular city. We have formulated this elsewhere as the canvas on which new shapes may be painted, the physical bounds within which social and economic changes must take place (Marcuse & van Kempen, 2000c).

History can be seen as the second contingency. History embraces everything that has developed in the past. In this sense it is really the major determinant of physical form, spatial pattern and urban development in general. Urban geographers should always have an eye for the rootedness of political, social, economic, and ideological events in their own past development and interrelatedness (path dependency) (see also Friedman, 1995).

The much longer history of European cities compared to North American ones, their background in a feudal system absent in the United States, the non-capitalist history of eastern European cities, the impact of colonialism both on imperial and colonised countries, the variations in local autonomy, are all critical in understanding the differences in development of the cities discussed in this volume. Chakravorty (2000) stresses this point eloquently in his discussion of Calcutta. A few studies have disaggregated these historical factors to examine their separate impact on city development and they reveal the importance of looking at political, institutional and cultural factors as a whole, rather than focusing on, say, the level of centralisation as a separate factor.

The third contingency is formed by economic developments. They can also be decisive in shaping social and spatial change. Type as well as stage of development can be important, although we also have to recognise that development never proceeds in a linear fashion, or that, for example, (cities in) less developed countries eventually follow the patterns established by (cities in) more developed countries. Not only are alternate paths of development conceivable, but also the very definition of development may vary. But certainly the extent of the resources available in the society as a whole has an impact on the shape of that society. Consider that the poor are pushed to the outskirts of cities in many Third World countries, for instance, while similar locations are desirable for the middle class in others. This fact has something to do with the ability to provide infrastructure over large territories in the one and the scarcity of resources for the purpose in the other. The level of urbanisation itself, including for instance the importance of rural-to-urban migration, will similarly affect how other trends play themselves out (Marcuse & van Kempen, 2000b).

Levels of inequality differ widely among countries and can be seen as the fourth contingency. Waley (2000) places Japan among those with the least inequality, while the United States is among those with the highest. The level of inequality is itself a function of the other contingencies we have discussed above. However, having been produced, it then exercises an independent influence on the divisions of cities. One of the most significant general outcomes of differential economic and political developments is the overall level of social inequality and social polarisation within countries and cities. It is possible that the degree of polarisation and inequality (for a sensible discussion of the distinction between the two: see Hamnett & Cross, 1998), on the one hand and the extent of spatial separation on the other hand would be roughly parallel, but it would be a mistake to see this relation as an automatic one.

The importance of race and racism for the development of the spatial structure of cities is not always clear but can definitely be seen as an important contingency. Their importance in the United States experience, however, cannot be exaggerated. A number of works make this point eloquently (see Goldsmith & Blakely, 1992; Massey & Denton, 1993). Goldsmith (1997, 2000) sees racial exclusion and residential segregation as essential, perhaps the essential qualities of the US metropolis in the last half of the present century. The fact that most African Americans in the United States usually live in neighbourhoods that are racially homogeneous illustrates this point, especially when we realise that few of them live there because they truly want to. According to Goldsmith, numerous surveys have indicated that African Americans do not wish to be segregated any more than others. Moreover, statistics show that income differences explain very little of black segregation in US cities: “... all evidence points to the primacy of overwhelming and effective racial discrimination in the development of cities and suburbs in the United States” (Goldsmith, 1997, p. 302).

Indications of similar racist tendencies exist in many other countries, where opposition to immigration and immigrants has led to xenophobic political policies and even more visible direct hostilities, including physical attacks on immigrants. Turkish residents in Western Europe, for instance, have often met hostility based variously on immigrant status, culture, religion, and language (see Özüekren & van Kempen, 1997; van Kempen & Bolt, 1997). Until now, however, this has generally not led to the relegation of those who are not accepted in European urban societies to only a very small number of neighbourhoods that can be considered as the worst places in town where nobody else lives.Footnote 7

Living in separate neighbourhoods has many disadvantages (Bolt, Burgers, & van Kempen, 1998). One specific point should be mentioned here. When people live in separate neighbourhoods groups learn of each other by means of distant, indirect experience, through reading or through what they hear from their neighbours or see on television. Their knowledge is then based on shallow understanding, prejudices, and superficialities. More directly: radical segregation causes white suburban residents and inhabitants of “gated communities” to be ignorant and even afraid of the population elsewhere, especially in the cities (see Goldsmith, 1997; Sennett, 1970). This is not quite a good basis for tolerance, understanding, and living together.

The sixth contingency to be mentioned is that of the balance of political power, which of course is a primary determinant of national policy issues. In most nations of the world in the immediate post-war period, a relatively stable, essentially social-democratic regime dominated policy-making. And most were able to provide some improvement in living conditions for the large majority and an adequate level of profits for the business community. With the election of Margaret Thatcher in Britain, Ronald Reagan in the United States, and Helmut Kohl in Germany, a long-term political turn to the right was a fact. The shift to the right paralleled the shift of the balance of power in the economic sphere from labour to capital. Periodic protests throughout the period, particularly if they rose to the level of civil unrest, as in the mid-1960s, produced different constellations of power, but generally only briefly. These shifts have obvious consequences for national policy, which in turn has consequences for spatial structure (Levine, 1995, pp. 106/7; he cites Logan & Swanstrom, 1990).

Related to the widespread retreat of the welfare state in Western countries, the central state has devolved many of its duties to other kinds of governments, like provinces, regions and cities. Within their territory, some cities have shown further decentralisation tendencies by giving (some) power to city districts or even neighbourhoods. Other tasks have been privatised. In relation to all these deregulation, decentralisation and privatisation processes, some decisions concerning urban developments, neighbourhood policies, and neighbourhood regeneration or reconstruction are in many cases not made by (local) government alone, but by a mix of many different organisations and individuals.Footnote 8 According to Elander and Blanc (2001), governments have faced a development towards fragmentation and more differentiated forms of governance: government has become governance. This can be seen as the seventh contingency. Privatisation, deregulation and multi-actor policy-making are key ingredients of this trend. In a local setting, local governments no longer play an exclusive role as the leading policy-maker. They are more than ever before merely one of the many actors in the governance arena (Elander & Blanc, 2001; Healey, Cameron, Davoudi, Graham & Madanipour, 1995).

Strongly related to this development, the innovation that seems to have become commonplace all over Europe lately is the creation of ‘partnerships’. Partnership has been defined as a coalition of interests drawn from more than one sector in order to prepare and oversee an agreed strategy for the regeneration of a defined area (Bailey, Barker, & MacDonald, 1995, cited in Elander & Blanc, 2001).

The theme of governance, and the related theme of partnerships, now shows up in many texts about urban developments in general and about neighbourhood regeneration policies specifically (see Bailey et al., 1995; Friedrichs, 2001; Hastings, 1996; Kristensen, 2001; Walzer & Jacobs, 1998). Although it is probably too early to see the governance approach as a new one that fits in the row of approaches we mentioned before, it is important to mention this focus of analysis. It opens our eyes to the fact that the role of the state is not stable any more, nor defined everywhere in the same manner. Coalitions between parties and stakeholders are formed differently in different cities and the results, in terms of processes and patterns of segregation and concentration, may come out significantly different from each other and from results of, for example, a decade ago.

A few examples may clarify the possible role of urban governance.

  • While housing allocation may have been always the remit of local government or housing associations (or both) in a city, partnerships, comprising local governments, housing associations, neighbourhood organisations, private developers etc. may take over the decision procedure, leading to inclusion of some housing applicants and exclusion of others.

  • While urban and neighbourhood regeneration policies have traditionally been the task of local governments, often with a certain kind of support of the local population (at least in many Western European cities), presently, also private developers, private firms and many different kinds of neighbourhood organisations might become involved. This might lead to powerful coalitions that are successful in generating money (at the local or national, or even European level) or to other coalitions that are not so successful, but do have the same problems in the neighbourhood (see Walliser, 2001, who describes this process for the city of Barcelona).

Our expectation is that governance, coalitions and partnerships will be more influential as explanatory variables of processes of concentration and segregation in the near future.

6 Final remarks

From this paper it has become clear that a number of issues and developments mentioned in older theories of urban change are crucial when we want to explain the emergence, existence and developments of divided cities. It is impossible to look at divided cities without paying attention to individual preferences and individual constraints and opportunities. Of course, these opportunities are shaped by many developments that are far beyond the reach of the individual (including globalisation and other macro-level developments). But in essence the individual, or household, still has to be seen as an important decision maker with respect to housing market behaviour and the place where one resides. Residents of a city “... are not simply puppets dancing to the tune of socioeconomic and political logics...” (Beauregard & Haila, 1997, p. 328). Various institutional arrangements, including the role of individuals within institutions, add to the complexity of the explanation of divided cities. Most local and national states have changed their goals and their way of working, but they do still exist and are still important in determining the future of cities.

How divisions emerge and develop is at least partly determined by a number of contingencies. If we do not recognise these contingencies, which can work out very differently for each place, it is impossible to explain differences between cities. The specific morphology of cities, their histories, their geographical characteristics, the extent of inequality in a society are just a few of these contingencies that determine the present and future of divided cities.

Globalisation is without any doubt an important development in the present era and it pervades every continent, every city, neighbourhood or even individual. But it would be wrong to put too much weight on this development and lose sight of the important urban actors, such as individuals, households and institutions (including governments). One of the most important urban questions is how to get the best out of the interaction between macro-developments on the one hand and local opportunities on the other hand. To achieve this, concrete mechanisms by which these national and international forces produce specific spatial changes within cities should be found (Marcuse, 1997). Because these mechanisms are in most cases still unclear and because they differ between contexts, this might be seen as an important task in the future—a joint effort by urban geographers, urban sociologists, urban economists, urban planners and maybe many others.