Urban Segregation and Metropolitics in Latin America: The Case of Bogotá, Colombia


  • The first author gratefully acknowledges the financial support of the Trudeau Foundation. Both authors also acknowledge the help and constructive criticism provided by Oscar A. Alfonso and Gerardo Ardila Calderon, both of whom made this collaborative endeavor possible.


The patterns of spatial socioeconomic segregation in Latin American cities are changing rapidly as a result of suburbanization and metropolization. However, the political consequences of these urban spatial processes are not well understood. This article uses Orfield's framework of analysis to test the hypothesis that spatial segregation at the metropolitan level is driving political polarization between Latin American cities and their suburbs. With Bogotá as a testing ground, we look for evidence that the mechanisms described by Orfield are at play. We conclude that metropolitan spatial segregation does not drive metropolitan politics in Bogotá and explore some of the theoretical implications thereof.


Spatial segregation in Latin American cities is a fact; it has long been and continues to be a defining characteristic of the Latin American metropolis, albeit to varying degrees. This feature manifests itself both in terms of the segregation of ethnic and social groups and in terms of the segregation of uses and functions within the city. In some cases, as in Bogotá, spatial segregation is reinforced by a spatial socioeconomic stratification system used to target subsidies towards the poor, which effectively identifies certain areas of the city as ‘poor’, ‘middle-class’ or ‘rich’.1

However, evidence shows that patterns of segregation in Latin American cities have evolved over time and continue to evolve rapidly. Indeed, there has been much discussion in the literature of the impacts of the neoliberal reforms of the 1980s and 1990s on income polarization and spatial segregation (Portes and Johns, 1989; Portes et al., 1994; Ribeiro and Lago, 1995; Roberts, 2005). More specifically, scholars have described recent trends, such as the consolidation of ‘land invasions’ and the appearance of ‘closed condominiums’ (conjuntos cerrados) or gated communities on the urban periphery, in relation to the deregulation of land markets and the rapid pace of suburbanization (Carter, 2003; Roberts, 2005; Coy, 2006). In short, urban spatial segregation in Latin America is omnipresent but it takes on many different forms and cannot be treated as a homogeneous phenomenon.

Despite this heterogeneity in the spatial evolution of Latin American agglomerations, a clear pattern emerges from this literature: Latin American cities and their metropolitan regions are becoming both more mixed at the macro level and more segregated at the micro level. As pointed out by de Duren (2006), this trend of ‘cellular segregation’ raises a number of important theoretical and policy questions, especially in regard to the potential long-term social consequences of this new mode of territorial occupation. Most importantly, one might ask: is this pattern of segregation likely to diminish social inequalities, or does it serve to further entrench them?

In order to answer this question, we argue that it is necessary to look not only at the direct immediate social and other consequences of this type of segregation, but also — importantly — at the long-term political consequences of this shift in population, and of the elites in particular, from the city center to the urban periphery (i.e. rural areas in transition beyond the city limits). It is noteworthy that urban spatial segregation and social inequality in Latin America are quite well-researched and well-theorized, but that the political shifts associated with the phenomenon of ‘metropolization’ (whether actual or potential) are not. For instance, we could ask whether the ‘secession of the rich’ into gated communities beyond the city limits, as described, for example, by Janoschka and Borsdorf (2004), Alfonso (2005), Coy (2006) and de Duren (2006), could potentially shift power away from core cities and towards richer outlying municipalities (or other units of government) — which could have important distributional consequences.

In the absence of a clear theory of metropolitan politics in the Latin American context, it is interesting to ask whether other descriptive and theoretical work about the functioning of metropolitan regions might allow us to make progress on this question. One interesting attempt at describing the power dynamics of metropolitan regions is that of Myron Orfield in his book American Metropolitics (2002). Orfield describes the phenomenon of income polarization and power diffusion in North American urban regions and explains how and why the interests of core cities and suburban municipalities diverge (or why they are perceived to be diverging). In his account of metropolitan politics, suburbs rich and poor tend to ally against core cities and resist regional redistribution because they perceive their interests to be aligned. However, Orfield makes the case that poor and rapidly expanding middle-class suburbs both have it in their interest to promote regional redistribution as they, too, face a mismatch between fiscal need and capacity. Whatever the case may be, the key inference in Orfield's account is that metropolitan politics are driven by spatial segregation — whether by age, race or class.

More specifically, he identifies the following mechanism: (1) the urban elites ‘vote with their feet’ and relocate to suburban municipalities with a lower ‘social burden’, where taxes are lower and public services better; (2) central cities try to recapture this wealth through political means (annexation, consolidation of specific competencies or tax-base sharing, all of which have to be instated by the state government in the US context); (3) suburban municipalities resist annexation or consolidation and become politically aligned against the central city.

Clearly, Latin American agglomerations differ from those of North America on several important counts, and the explanatory power of Orfield's theory in the Latin American context is necessarily limited. Yet several of the trends observed in Latin America over the past three or four decades, including the outward expansion of urban regions and the migration of the rich from the center to the periphery, are analogous to those that have been observed in North America since the early twentieth century. The patterns and mechanisms described by Orfield may therefore provide us with a ‘theoretical starting point’ to explore the patterns of suburbanization in Latin America.

The thrust of this article, then, is twofold: first, to analyze the current patterns of segregation and suburbanization in Latin America to see whether and how they resemble those described by Orfield and to note in particular where they do not; and secondly, to explore the implications of segregation at the metropolitan scale for the political process in a particular setting: that of Bogotá, Colombia. The case study of Bogotá serves to illustrate both the processes at work in Latin American metropolitan regions and the ways in which suburbanization and spatial segregation play out politically. Our main theoretical interest, therefore, is to see whether the mechanism described by Orfield can shed light on recent political developments in Latin American agglomerations, and Bogotá in particular. More generally, we aim to advance the theory of metropolitan politics in Latin America by using Orfield's theory of Metropolitics as a point of departure.

The first section of the article describes the evolution of urban segregation and suburbanization in Latin America and highlights some of the spatial trends that have emerged there over the past three decades. The second section looks specifically at the case of Bogotá in light of the general patterns and trends found in the region. Finally, the third section provides an initial assessment of the applicability of Metropolitics to the Latin American case.

Segregation, suburbanization and emerging spatial patterns

Neoliberalism and spatial segregation in Latin America

The transformation of urban economies as a result of globalization and neoliberal reforms has been described at length in the literature (Graham and Marvin, 2001; Sassen, 2002; 2006; Taylor, 2007; World Bank, 2009). Yet the spatial transformations associated with these changes in cities of the South are not well understood (Michelutti, 2010). Indeed, as argued by Roberts (2005), cities of the developing world tend to be under-emphasized and under-theorized in scholarly work about globalization and its impacts on spatial urban processes.

Despite this general trend, the patterns of urban spatial transformation in Latin America have received quite a bit of attention, which makes Latin America an interesting terrain for theory testing. For instance, many authors have described the transition from import-substitution industrialization (ISI) policies in the 1950s, 1960s and 1970s to neoliberalism and export-oriented industrialization in the 1980s and 1990s, and highlighted the impacts of market deregulation, privatization and export-oriented policies more generally on urban economies. In what follows, we describe the patterns of spatial segregation under each economic regime and give an overview of some of the emerging spatial trends in Latin American cities. It should be noted that the aim here is not to give an exhaustive account of the spatial processes at work, but rather to focus specifically on those spatial processes and patterns that could (or did) have implications for metropolitan politics.

Spatial segregation under import-substitution industrialization (ISI)

ISI refers to an economic development theory (and subsequently, policy) developed by Raúl Prebish in the late 1950s. His theory mainly advocated the ‘replacement’ of imported goods with goods produced domestically, but ISI policies in most countries were accompanied by other interventionist policies and constraints on free-market mechanisms — including land markets.

One way in which ISI policies affected patterns of spatial segregation was by preserving some of the traditional features of Latin American cities, as the movement of people within urban regions and between regions was generally constrained. For instance, Butterworth and Chance (1981) note that Iberians (and Latin American elites, by extension) have traditionally valued compact urban centers and that the rich have tended to dominate central locations — a pattern that has endured in many cities throughout the ISI period.

Large cities in Latin America have also historically had a high primacy within their country, which is a reflection of the region's ‘extraordinarily centralized governance’, accompanied by weak and highly dependent local governments (Rodriguez-Acosta and Rosenbaum, 2005). It is interesting to note, for instance, that until the late 1980s, most countries in Latin America did not have elections at the municipal level. What is more, most municipal administrations in Latin America rely primarily on transfer payments from higher levels of government — and not on property-tax collection. As a result, those municipalities that happen to also be the seat of power (whether at the departmental, provincial or central level) have traditionally received more than their share of resources (see, for example, Gouëset, 1998: Chapter 6). All these things combined underscore the importance of centrality in understanding the spatial organization of cities on the continent.

The second way in which ISI has impacted on the patterns of spatial segregation in Latin America is through government intervention in land markets. Carter (2003: 47) notes that ‘land-use controls and resulting lack of tenure on the periphery under ISI focused speculation toward the center of Latin American cities’. This, combined with the relative lack of transportation and other infrastructure in outlying areas, created a situation in which land and housing generally became much more expensive in central neighborhoods, thereby forcing low-income households to move further away — which further reinforced existing patterns of segregation.

However, increased congestion in central areas eventually ‘forced holes in the urban limits’ as both the rich and the poor started to settle the periphery — albeit in different parts and under different conditions of land tenure. Whereas the rich were able to circumvent land-use controls, the poor often settled rural lands illegally. It is interesting to note that Bähr and Mertins (1982: 25) were writing about ‘cellular urban expansion in the periphery’, referring precisely to this phenomenon, as early as 1982 — before the so-called ‘neoliberal shift’ of the early 1980s could have taken effect. Hence, there is evidence that ISI policies initially served to reinforce the centrifugal forces at play in Latin American urban regions, but eventually pushed both poor and rich segments of the population beyond the city limits (Ward, 1990).

Spatial segregation under neoliberalism

As described by Roberts (2005), the so-called shift from ISI to neoliberalism in Latin America was gradual and did not operate in the same way everywhere. For instance, while some researchers initially found that income polarization in the 1980s was closely tied to increased spatial segregation (Portes and Johns, 1989), subsequent studies of the same phenomenon are much more nuanced (Portes, 1989; Ribeiro and Lago, 1995). Caldeira (1996), for instance, purports that the rise of gated communities is more directly linked to rising urban insecurity — which may or may not be explained by income polarization.

Yet there is agreement that neoliberalism and globalization impacted on spatial segregation in particular ways. The first mechanism through which neoliberalism affected spatial segregation was indirectly through increased income polarization. As described by Roberts (2005) and Portes and Roberts (2005), the growth of what is known as the ‘producer-service’ sector in Latin American cities does not seem to be linked to the appearance of new global functions; however, it does appear to be associated with class and income inequality as in other ‘global cities’. In effect, rapid economic growth in the 1990s has not diminished the gap between rich and poor that was left by the ‘lost decade’ of the 1980s (except perhaps in Brazil).2 This is confirmed by recent evidence collected by UN-Habitat (2009), according to which Latin American and Caribbean cities are among the most unequal in the world, with urban inequalities increasing and becoming more entrenched. This combination of high inequality and low social mobility, in turn, may exacerbate existing patterns of segregation.

The second mechanism that is described in the literature is the deregulation of land markets. Carter (2003: 48) explains that ‘neoliberalism shifts investment and speculation more toward the periphery by allowing increased suburban development by the upper and middle classes and more rapid consolidation by the peripheral poor’. This shift from land speculation in the center to land speculation on the periphery, in turn, facilitates both high-end commercial and residential development beyond city limits — often through leapfrogging (Mertins, 2004).3

A third mechanism that has been identified, and which is closely connected to the deregulation of land (as well as other) markets in several metropolitan regions, is the development of infrastructures and transportation infrastructure, in particular. As pointed out by Carter (2003: 223), ‘while private automobiles allow people to live in more homogeneous suburban communities, infrastructure development allows these communities to be built’. It is important to remark that highway development does not dramatically change the situation of the poor: they still cannot afford to live in central areas, but cannot afford to live too far from their workplace either — as the cost of transportation for them remains more or less the same (Borsdorf and Hidalgo, 2010). The development of modern public transit systems, however, does facilitate the establishment of the poor further away from centers of employment (see for instance Cariola and Lacabana, 2003, about Caracas).

In summary, the shift towards neoliberalism is thought to have affected patterns of spatial segregation in mainly three ways: (1) through an increase of inequalities (which reduces the options available to the poorest); (2) through the deregulation of land markets (which facilitates commercial and residential development on the periphery); and (3) through the development of infrastructure (which increases the mobility of certain population segments and facilitates development).

The contemporary Latin American metropolis: mixed and fragmented

It is important to restate that North American and Latin American cities had very different starting conditions at the onset of their respective ‘outward expansion’ in the twentieth century.4 However, as alluded to earlier, it is the trajectory and not the starting point that concerns us here. Hence, the patterns described below do not tell the whole story; they are highlighted precisely because they are indicative of a particular trajectory.

Gated communities, consumer tastes and the fortified city

The advent of neoliberalism has significantly increased the ability of the rich to segregate themselves. Yet globalization and neoliberalism did not merely increase the supply of suburban development; they also increased the demand for the suburban way of life (Gilbert, 1998). In Gilbert's account of these processes, there is a clear link between the internationalization of consumer tastes in Latin America, the appearance of suburban malls, the dramatic increase in car ownership and the rapid pace of suburbanization on the continent during the 1980s and 1990s. In other words, the process of suburbanization appears to be driven by both endogenous and exogenous factors.

In their analysis of recent trends in urban spatial segregation in Latin America, Borsdorf and Hidalgo (2010: 25) come to similar conclusions: ‘As the state did not intervene in urban planning processes, and private investment to a large extent neglected public means of transportation, a new, car-based lifestyle has flourished, causing an even greater degree of fragmentation and spatial segregation. The changed lifestyle has also become the motor of urban expansion’. Numbers confirm these general trends: the vehicle fleet in Latin America grew by about 250% between 1970 and 1990, reaching 37 million vehicles (World Bank, 1997). Meanwhile, vehicle ownership rates in Latin America, at about 90 vehicles per 1,000 people, are much higher than in other developing regions, including the Middle East, and are still increasing at a steady pace (Schipper et al., 2010). The affirmation that there is a flourishing ‘car culture’ in Latin America therefore has a factual basis.5

This so-called ‘internationalization of consumer tastes’ manifests itself in at least one other important way in terms of spatial segregation: the rapid increase in the number of gated communities in Latin American metropolitan regions. Janoschka (cited in Coy, 2006) estimates that 450 closed residential ‘communities’ (or barrios cerrados) were developed in Buenos Aires during the last decade of the twentieth century, housing more or less 500,000 people. It is also estimated that about 700 such communities were created in and around Mexico during the same period (Kanitscheider, cited in Coy, 2006).

In a line of argument similar to that of Borsdorf and Hidalgo (2010), Coy (2006: 122) argues that the initial appeal of gated communities came from the ‘cultural influences’ of globalization, which he alleges ‘reinforced the change of locally or regionally defined traditional lifestyles towards a more homogeneous, globalized lifestyle of the privileged’. Interestingly, Lacarrieu and Thuillier (2001) note that social distances are considerably reduced within gated communities, as these communities often become microcosms of society, with their own set of imposed rules and norms. However, as pointed out by Vidal-Koppmann (2007), the social distance between those within and without is considerably increased as a result of this fragmentation and fortification of urban space.

In addition to the search for exclusivity, Coy also mentions security (along with status and lifestyle) as one of the main concerns of the elites. The security hypothesis is, of course, fraught with the problem of reverse causality — as gated living could very well be the cause of perceived insecurity. Whatever the case may be, there is evidence that fear is, indeed, a major factor explaining the increase in private security — which is closely tied to the rise of gated living and ‘closed’ living and, ipso facto, to the process of spatial segregation.6

The shift from polarization to fragmentation

Research suggests that the processes of income polarization and urban segregation are not always correlated in Latin America. However, there is evidence that income polarization and urban restructuring may be associated with an increase in urban fragmentation — or the appearance of small units of wealth and poverty that are spatially contiguous but socially isolated from one another (for example, when the maids and gardeners of the rich live in the next-door shantytown and only have formal interactions with those living in gated communities). Borsdorf and Hidalgo (2010) remark that this tendency has led to a number of scholars concluding that Latin American cities were becoming more income-mixed; however, as they correctly point out, increased spatial proximity does not necessarily lead to increased social mobility. Hence, the Latin American metropolis is becoming more socially mixed at the neighborhood scale, but also more spatially and socially fragmented at the block and street level.

Indeed, as noted by Roberts (1989: 675) more than 20 years ago, the consequences of neoliberalism are such that ‘an impoverished middle class seeks cheap accommodation in low-income neighborhoods, and low-income residents cling to the niches they have developed, close to centers of work’. Coupled with the fact that urban land markets are now largely deregulated, the result is a complex patchwork of rich, poor and middle-class neighborhoods, which are spatially proximate but socially distant (Rojas, 2008). What is more, this patchwork is found both within and outside city limits.

How did this patchwork come about? According to de Duren (2006: 324), it is no coincidence that the majority of gated communities built during the 1990s ‘ended up disproportionately concentrated in municipalities characterized by a high percentage of poor households and a large area of under-serviced land’. She remarks that the localization of gated communities within poor exurban municipalities outside of Buenos Aires was purposeful, as ‘these municipalities were increasingly likely to change their regulation of land usage to accommodate gated communities, which they value as service and tax providers’ (ibid.). Vidal-Koppmann (2007), who specifically investigated the effect of these developments on modes of territorial occupation on the periphery of Buenos Aires, also presented evidence that private developers are indeed choosing these particular locations, whether because of land price or because planning exceptions are more easily obtained in these localities.

Whatever the mechanisms at play, this structuring process appears to be at the heart of metropolitan dynamics of Latin American cities, as described by Borsdorf and Hidalgo (2010). Very similar trends have been described in the past decade by a number of scholars, for Caracas (Lacabana and Cariola, 2003), Bogotá (Alfonso, 2009; Osorio, 2011), Lima and Quito (Borsdorf, 2002), as well as for cities in Chile (Sellés and Stambuk, 2004), Mexico (Rosas, 2006; Aparicio et al., 2011) and Brazil (Coy, 2006). The result of this process is often (although not necessarily) a form of cohabitation without integration or ‘splintering urbanism’;7 as described by de Duren (2007: 36): ‘The physical proximity of these distinct spaces not only produces social contrasts, but also promotes an uneven distribution of resources among locals that furthers these initial contrasts’.

To summarize, increased spatial proximity of rich and poor does not appear to be indicative of greater social tolerance — although it might, in some cases, result in greater social proximity (Cáceres and Sabatini, 2004). Rather, it seems to be the result of the interplay of (1) competition for housing in central areas among members of the middle and lower middle classes; (2) consumer demand for gated and closed living throughout the metropolitan region; and (3) the supply of such forms of housing, which is greater where land is more readily available — often in poor or rural municipalities.

Enduring and emerging spatial trends in Latin American cities

Carter (2003: 225) aptly summarizes the spatial transformation of Latin American cities over the last few decades in this way:

Land markets should lead to segregation region-wide. Depending on the degree and types of neoliberal land reforms, and the level of enforcement of zoning laws, this segregation can take different forms — with upper income groups concentrated near the center and limited sectors and the poor on the periphery, as in older urban models, or with upper income groups in new suburban communities and the poor on the other side of town, as seen in models from the 1990s.

For the purpose of identifying those spatial trends that are likely to have an impact on metropolitan politics, we can rephrase Carter's assessment and highlight three trends:

  1. The first spatial trend of interest is the suburbanization of the rich and, increasingly, the middle class, which has considerably altered the traditional core–periphery socio-spatial structure of Latin American cities;
  2. The second notable trend is the increase in physical proximity of the rich and poor, as new ‘closed’ residential developments are developed in both the center and on the periphery, in close proximity to poorer areas;
  3. The third spatial trend, which may be indicative of a trajectory, is the increase in the number of ‘no-entrance’ and ‘no-go’ areas' in Latin American cities; the byproduct of the city's ‘fortification’ in some parts appears to be the lack of both public and private security in others.

The case of Bogotá, Colombia

Why Bogotá?

Bogotá is the capital city of Colombia and by far the largest city, with approximately 6.8 million inhabitants in 2005 in the city proper and approximately 7.7 million in the ‘metropolitan area’ (Medellín, the second largest metropolitan area, is less than half its size, with approximately 3.3 million inhabitants, according to the 2005 census). It is interesting to note that its population in 1938 was only approximately 330,000 — less than 5% of its current size (Gilbert and Davila, 2002).

The case of Bogotá, Colombia, is particularly interesting in so far as the process of ‘metropolization’ is concerned, as seen in Figure 1. First, unlike some of the other major cities of Latin America, but similar to cities of North America, the city of Bogotá has grown rapidly during the twentieth century through migration and natural increase but also through amalgamation — the towns of Usme, Bosa, Fontibón, Engativá, Suba and Usaquén were annexed in 1954 when the ‘Capital District’ was formed. As explained by Gouëset (1998), this seemingly anodyne detail is far from anodyne: Bogotá was given ‘special’ prerogatives from very early on in its history (it first became a ‘Capital District’ in 1905) and its growth was not impeded by strict growth controls as was the case in other cities of Latin America. According to Gouëset, this is in part a result of the influence of Harvard and World Bank economist Lauchlin Currie, who came to Colombia in 1949 and convinced then President Rojas Pinilla that the country's economic growth depended on the growth of its cities — and that of Bogotá in particular. The city of Bogotá, under this planning regime, did not enjoy all the prerogatives associated with ‘home rule’ in the United States, but it had much more autonomy than other cities in Colombia — or the rest of Latin America. One could argue, therefore, that Bogotá was already evolving under a liberal regime prior to the continent-wide ‘neoliberal shift’ of the 1980s (much like cities of North America). Another notable similarity between Bogotá and cities of North America is the widespread use of tax incentives in periurban municipalities to attract industry. Indeed, as described by Alfonso (2012: 45–46), ten of the most important municipalities surrounding Bogotá, including some of the poorest, have adopted such policies.8

Figure 1.

Note: The red line designates those parts of the region that are located above 2,600 m in altitude, which are de facto part of the metropolitan area.

Map of the municipalities located within the Bogotá metropolitan area (source: CEDE, 1999)

However, despite the fact that Bogotá historically shared some characteristics with cities of the US, such as greater political autonomy and a pro-growth land-use planning regime, it is in many ways a typical Latin American capital city, where ‘long dominant national political institutions, such as the president, centralized bureaucracies, and national political parties, still remain important, and traditional elites still influence public policy’ (Myers, 2002: 1). For instance, much as in the rest of Latin America, the mayor of Bogotá was appointed by the president until 1988 and the city still concentrates within its area the lion's share of functionaries and bureaucrats as well as most of the country's economic and political elite. Hence, when the city's mayor attempted to form a metropolitan planning commission in the late 1990s, national politicians took an active part in the debate (El Tiempo online, 2000).

As in other Latin American cities, there has historically been in Bogotá a ‘core–periphery’ pattern of segregation, where the status of a household was directly proportional to its proximity to the center (Jaramillo, 2007; Brewer-Carias, 2008).9 However, as early as the late nineteenth century, rich families in Bogotá started moving northward and the first ‘country residences’ appeared on the close northern periphery (Jaramillo, 2007). In the early 1970s, the pace of urban–rural migration accelerated quickly; as an illustration of this trend, Dureau (2010) cites the case of Chía, which already had a growth rate of 8% by the mid-1970s, with more than half of its new residents coming from Bogotá. Interestingly, the spatial processes at play during the first half of the twentieth century and the first phase of ISI (in-migration of the poor from the countryside, out-migration of the rich) were largely unmediated by the state; indeed, the authorities became more involved in spatial planning after the city started receiving foreign capital in the 1950s (Sáenz, 2003). Nevertheless, the reforms of the 1970s had similar effects as on other Latin America cities: first, the gradual deregulation of land markets with the decentralization of planning functions contributed to the movement (or displacement) of the urban poor towards the periphery through land speculation and redevelopment; secondly, as noted earlier, it led to a ‘tax war’ between surrounding municipalities to attract industry; finally, it fuelled the construction of ‘apartment compounds’ in all parts of the metropolitan region, including in middle- and lower-income areas. In effect, as described by Chois (2006), the reforms initiated by President Pinilla under the guidance of Lachlan Currie were intended to strengthen the state's economic planning functions but ultimately had the effect of weakening its land-use planning functions; in particular, the 1968 Decreto 1119 and the 1974 Decreto 159 introduced individual lot zoning, thereby replacing the comprehensive planning approach of the previous decades, which imposed strict density restrictions.10 This change, in combination with the appearance of new modes of financing and the devolution of certain planning functions to the municipal level in 1978 with the passage of the Ley Orgánica de Desarrollo Urbano, are said to be partly responsible for the rapid increase in the number of new apartment compounds in both rich and middle-class neighborhoods, as density restrictions were relaxed in many areas (Cortes, 2008). Further, it is important to note that in 1987 and 1988, land markets were further deregulated as 40% of the conservation zone around the city was slated for development.

In short, the Capital District of Bogotá started to expand into its hinterland as early as the mid- to late 1970s, but it has retained a very centralized mode of governance and is still highly dependent on the central government of Colombia — similarly to other capital cities of the continent.11 These characteristics make it an ideal case for testing a theory of spatial segregation and metropolitan politics, as the process of metropolization in the region is quite far along, which means that suburban municipalities have had time to gain demographic, economic and political weight relative to the city.12 Yet the mode of governance of the metropolitan region has not significantly changed in the past few decades, which places Bogotá more or less on an equal footing with the typical Latin American metropolis.

Spatial trends and segregation in Bogotá

Suburbanization of the rich and the middle class, and gentrification at the core

The urbanization of Bogotá took place over several decades, but the city grew at a particularly rapid pace during the decades from 1940 to 1970, with growth rates of approximately 6% a year on average. As a consequence of this rapid increase in population throughout the 1950s and 1960s, and because of the growth of industry in outlying areas, the city's surrounding municipalities also started growing rapidly as early as the 1960s. By 1985, it is estimated that 50% of inhabitants of the towns located immediately north of Bogotá (such as Chía, Tabio and Cota) were not born there (Mertins, 2004). Notably, many of these migrants came from the countryside to work in industry; this is especially true in the western suburbs of Funza, Madrid and Mosquera, where the flower industry flourished from the 1960s (Montañez Gómez et al., 1994).

However, it is estimated that by the mid-1980s at least a quarter of the inhabitants of Chía, the municipality immediately north of Bogotá, were commuters working in the core city — most of them belonging to the upper middle class (Dureau, 2003a). According to Alfonso (2010), a total of approximately 246,000 city dwellers moved from Bogotá to the suburbs (whether from higher or lower estratos) between 1993 and 2005, and growth projections suggest that this trend will continue.

Mertins (2004) notes that this process of ‘suburbanization’ of the rich, which started early in the twentieth century but really took off in the 1980s and 1990s, was accompanied, and sometimes preceded, by the migration of private schools, colleges and universities along with sports clubs, shopping malls, medical clinics and other institutions to those same northern towns. This migration, compounded by a high demand for housing (both for the middle class and the rich, defined as estratos three to six), physical limits to expansion to the east and west and a very lenient land-use regulatory framework, led to a reconfiguration of the metropolitan region. For example, the rural municipality of Chía attracted in the 1980s and 1990s a large number of middle- and higher-income earners who were seeking a ‘return to the countryside’ or a life away from congestion (Alfonso, 2009). According to Dureau (2003b), a third of the municipality's households in 1993 had been there for less than 5 years, half of the inhabitants were from the wealthier northern part of Bogotá and 52% worked outside of Chía.

However, Dureau (2003c) also notes that the process of suburbanization did not only involve the rich (defined as estratos five and six) seeking the ‘calm of the countryside’. Indeed, the poor and the very poor (from estratos one and two) also left the core city (those neighborhoods that had already been part of Bogotá before the 1954 amalgamation) in large numbers to establish themselves in outlying municipalities. Figure 2 shows the distribution of estratos within the borders of the Capital District. What is more, there is evidence that the relatively poor suburban municipality of Soacha also started attracting middle-class families (of the third socioeconomic category out of six) from the late 1980s. According to Dureau, these families were attracted to Soacha mainly because of the appeal of homeownership, which could be more easily achieved in Soacha (83% of households living in compounds in Soacha are owner-occupiers). To a lesser extent, car ownership made living in Soacha possible for members of the middle class working in the central business district or other parts of Bogotá, although the commute from Soacha to the center is no doubt a disincentive to moving there. Another important trend that was observed from the 1970s and 1980s is the migration of poor rural households to the industrial towns of Mosquera and Fusagasuga to the west of Bogotá.

Figure 2.

Note: The estratos in Colombia correspond to socioeconomically defined geographical areas, which form the basis of the spatially differentiated tariffs that Colombians pay for public services such as water, electricity and waste removal. The determination of these strata is based on the characteristics of housing units and the nearby infrastructure, and not on household income.

Map of estratos within the city of Bogotá (source: Alcaldía Mayor de Bogotá, 2004)

In addition, another process unfolded in the decades from 1980 to 2000, parallel to the process of suburbanization: that of densification and gentrification in central parts of the city. Dureau (2003b; 2010) and Sáenz (2003) note that congestion and higher transportation costs in the 1980s attracted a segment of the rich back to the central business district (for example, to the neighborhood of La Macarena) and other employment areas, which fuelled speculation in these areas and promoted the construction of exclusive higher-density housing. The process of wealth concentration in suburban municipalities was therefore accompanied by a similar concentration of wealth in the neighborhoods immediately north of the downtown core, as well as the migration of some middle-class families further south and west.

Fragmentation in an already fragmented society

One could argue that the city of Bogotá was both segregated and fragmented prior to the advent of neoliberalism in the 1980s. Indeed, a large number of gated conjuntos (condominium communities) were built in the western and northern part of the city in the 1970s and, according to Dureau (2003c), this form of housing appealed mainly to middle-class households interested in homeownership. As mentioned above, these gated condominium communities had started to appear in the southern part of the city as well as in the poor southern and western suburban municipalities — which suggests that gated living in Bogotá preceded the so-called ‘internationalization of consumer tastes’. Rather, it seems to have been initially a response to the concerns of middle-class families over safety — whether or not these concerns were based on fact or perception (Alfonso, 2009). However, as in other cities of Latin America, it has now become a sign of status to live in an apartment compound (Dureau, 2010).

According to Medina et al. (2008), the city of Bogotá is still ‘highly segregated’, with large concentrations of wealth and poverty in a few areas. Yet there is evidence that the city is also becoming more mixed as a result of the numerous population movements within the metropolitan region (to and from the core city; to and from the countryside; to and from the outlying municipalities) and gentrification in the city's central neighborhoods. Fragmentation within the city of Bogotá proper is also increasing, but not as dramatically as in its northern suburbs (Dureau and Salas, 2010).

In effect, one of the most radical transformations of the metropolitan landscape is taking place in the municipality of Chía, where closed and gated condominium communities are being built in close proximity to rural villages or hamlets (veredas) surrounding the town, but without any connection to their surroundings (Osorio, 2011). Based on a number of interviews conducted by Osorio with both newcomers and poorer ‘traditional inhabitants’ of Chía, it appears that the level of interaction between the ‘new’ and the ‘old’ population is very low or close to non-existent. In fact, Osorio reports that newcomers do not generally become involved in local affairs. Moreover, these newcomers often keep ties to Bogotá (as noted by Dureau, 2003c), and as Colombians vote in the jurisdiction where they are ‘registered’, which is not necessarily the jurisdiction where they live, it is likely that a significant share of Chía's new inhabitants do not vote there (although, to our knowledge, there exist no data charting who votes where). In other words, the construction of gated condominium communities has altered the social dynamics of Chía, but not necessarily its local political dynamics.

Different suburban realities, differing local interests

So far we have seen that the processes of suburbanization, segregation and fragmentation in the Bogotá metropolitan region are quite variegated and multifaceted. Although it is beyond the scope of this article to provide a detailed account of these processes in each suburban municipality of the Bogotá region, the contrast observed between the southern conurbation (poor suburban municipalities receiving middle-class households), the northern conurbation (traditional rural communities outside the Capital District receiving rich, upper-middle-class and middle-class households) and the western conurbation (industrial suburbs outside the Capital District receiving rural migrants, middle-class and lower-income households) is instructive: not all suburbs have the same social makeup and socioeconomic trajectory.

Similarly, Gouëset (2005) identifies different ‘ideal-type’ suburban communities in the Bogotá region based on their avowed and unavowed interests in regional cooperation and redistribution. He bases his analysis on the diverse reactions of the 17 municipalities surrounding Bogotá to the mayor of Bogotá's proposal in 1999 to form a ‘metropolitan area’ so as to better plan the growth and development of the region.13 Mayor Peñalosa's proposal was met with fierce opposition from a number of suburban mayors, while others were indifferent and still others were interested in sharing resources with the central city.

In effect, despite the fact that the Capital District of Bogotá clearly does not have the power to annex its neighbors,14 several suburban mayors denounced the plan proposed by Peñalosa as a thinly veiled attempt to ‘take away’ their autonomy and the plan ultimately failed, in part because of the opposition from these municipalities as well as from the Department of Cundinamarca (which surrounds the city of Bogotá, but has no direct authority over the city's affairs). If we set aside what the mayor's real intentions were, it is interesting to note that the rich municipalities located immediately north of the Capital District, despite the fact that they are quite different from one another, were the most staunchly opposed to annexation, while the rich municipalities further away from the Capital District were less concerned with being subsumed or denatured by becoming a part of the metropolitan area. The poorer western and southern suburban municipalities, by contrast, were very keen to become part of the metropolitan area to gain access to the comparatively abundant resources at the disposal of the Capital District for providing affordable housing and social services to populations in need (Gouëset, 2005). However, they were not willing to ‘import’ poor households from Bogotá or have polluting industrial plants relocate onto their territory. More recently, in 2008, the mayor of Bogotá and the governor of Cundinamarca signed an agreement that was meant to form the basis of a new ‘Capital Region’ comprising all the municipalities of the Department as well as Bogotá. Once again, the plan failed — not because of public mobilization against it, but rather because of the staunch opposition of individual mayors (Instituto de Estudios Urbanos and Lincoln Institute of Land Policy, 2010).

This suggests that municipalities within the ‘functional’ metropolitan area of Bogotá reacted to the mayor's proposal of establishing a metropolitan planning agency according to their own specific interests, which were themselves shaped by the socioeconomic makeup of each municipality. Interestingly, however, civil society in general, and private actors in particular, such as NGOs and private corporations, were largely absent from this debate (Gouëset, 2005). Indeed, the conflict which erupted around the issue of ‘regionalism’ in the Bogotá metropolitan region was not fuelled by public opinion, but rather by the political actors themselves. A search in the archives of the newspaper El Tiempo, the most important print media outlet in Bogotá, confirms this: of the 352 articles mentioning the Bogotá metropolitan area in 1999 (the year it was formally proposed), only a handful report on popular support of — or opposition to — the initiative (El Tiempo online, 2011a).

The municipalities of Soacha, Funza, Mosquera and Madrid agreed to work together with the City of Bogotá to plan, among other things, the development of the Transmilenio transit system. However, the plan failed in 2001 for legal reasons and every attempt at consolidation since then (be it the creation of a City Region or the formation of a Central Region) has also failed because of the opposition of small-town mayors — from wealthy suburbs, traditional rural communities and industrial towns. Once again, the opposition to consolidation came from political elites, and not from civil society.

Significantly, as shown in Table 1, the local election results in Cundinamarca for the years 2003, 2007 and 2011 confirm that voters in the suburbs of Bogotá do not vote very differently from the voters registered in rural communities in the Department — they vote for different parties at different elections, with the Liberal Party, the Conservative Party, the party ‘Unidad Nacional’ and different coalitions of these parties all making strong showings.15 The Bogotá electorate tends to vote further to the left, but by and large there is no discernible ‘coalition’ of suburban and rural voters in opposition to the city. There is also no clear electoral pattern in the ‘richer’ northern municipalities of Bogotá, whether we look at these as a single grouping, or whether we look at the near suburbs of the north (Chía, La Calera, Cota, Cajicá) separately (Registraduría Nacional del Estado Civil, 2011).

Table 1. Local election results for 2003, 2007 and 2011 in the Bogotá metropolitan area
 MunicipalityAverage Population Growth Rate (1993–2005)2003 Local Elections2007 Local Elections2011 Local Elections
Number of Registered VotersVoter TurnoutWinning PartyNumber of Registered VotersVoter TurnoutWinning PartyNumber of Registered VotersVoter TurnoutWinning Party
  1. Source: Author's own calculations, based on data from the Registraduría Nacional del Estado Civil (2011) and the Secretaría Distrital de Planeación (2009)
Central CityBogotá1.95%3,922,81843.51%Partido Polo Democrático Independiente4,378,02647.83%Polo Democrático Alternativo4,904,57247.4%Progresistas
Nearby Northern SuburbsCajicá2.47%n/an/an/a27,15767.82%Cajicá de Pie31,35566.21%Partido Conservador Colombiano
Chía5.27%48,91358.06%Partido Colombia Siempre60,05760.47%Unidos por Chía71,98758.4%Movimiento Autoridades Indigenas de Colombia
Cota3.69%9,85565.78%Para Qu e Florezca Un Nuevo Municipio12,36966.11%Partido Cambio Radical17,88766.76%Alianza Social Amigos por Cota
La Calera1.37%13,78362.77%Movimiento Equipo Colombia15,513 65.79%Partido Social de Unidad Nacional ‘Partido de la U’17,70265.61%Coalicion Cambio Radical y Partido Conservador
Northern ExurbsSopo4.47%9,34572.58%Partido Conservador Colombiano11,29574.56%Partido Social de Unidad Nacional ‘Partido de la U’13,64673.41%Partido Verde
Tabio5.54%9,04672.14%Movimiento Equipo Colombia10,86274.32%Partido Cambio Radical12,80573.73%Partido Cambio Radical
Tenjo0.87%8,89264.86%Partido Liberal Colombiano10,18070.88%Partido Cambio Radical12,04973.35%Partido Social de Unidad Nacional ‘Partido de la U’
Western SuburbsMadrid2.9%33,35766.06%Movimiento Nuevo Liberalismo38,89767.08%Unidos Construyendo Futuro45,59367.76%Partido Social de Unidad Nacional ‘Partido de la U’
Mosquera8.07%n/an/an/a33,01069.71%Partido Social de Unidad Nacional ‘Partido de la U’46,20472.81%Partido Social de Unidad Nacional ‘Partido de la U’
Funza2.92%31,73360.58%Movimiento Nuevo Liberalismo36,84664.9%Partido Cambio Radical49,63368.61%Garantia de Ciudad
Southern SuburbsSoacha3.86%151,35651.11%Movimiento Somos Colombia165,65251.3%Partido Liberal Colombiano183,86552.13%Coalicion Partido de la u y Partido Liberal Colombiano
Sibate2.12%17,23367.04%Partido Liberal Colombiano18,26068.23%Partido Conservador Colombiano21,69671.1%Coalicion Partido Conservador, Cambio Radical y Partido de la U

It is interesting to note that the Colombian law that legally defines the concept of the ‘metropolitan area’ (Ley Orgánica de Ordenamiento Territorial) is currently under review; hence the debate around consolidation in the metropolitan region of Bogotá might yet be revived. Several observers have noted, for instance, that there has been increased collaboration between the governor of Cundinamarca and the mayor of Bogotá in recent years, especially during Samuel Moreno's mayoralty. However, there are still no mechanisms for effective metropolitan governance in Bogotá, as evidenced by the recent failure of the plan to extend the Transmilenio to Soacha (which required Bogotá's mayor, Soacha's mayor, as well as the governor of Cundinamarca to come to an agreement). Therefore, in the absence of strong political will from the central government and lack of public mobilization, there is no reason to expect that the situation will change.

Metropolitics in Latin America: an initial assessment

Are the spatial trends observed in the cities of Latin America indicative of a metropolization of urban politics? To answer this question, we shall consider the three main observable implications of this process, as described by Orfield (2002), in terms of spatial segregation and its effects. We then test whether these implications can be observed in Latin American cities and, more specifically, the Bogotá metropolitan area.

First implication: suburbs will become polarized in terms of income

If metropolization of politics is indeed under way, we can expect to see a ‘specialization’ of suburban municipalities, either as a result of the process described by Tiebout (1956), where municipalities within a metropolitan region cater to specific socioeconomic groups by offering different ‘baskets of services’ priced differently, or simply as a result of rich households' ability and desire to segregate themselves — both physically and politically.

Although there is some indication that a number of suburban municipalities in Latin American metropolitan regions have ‘specialized’ in certain socioeconomic niches (Janoschka and Borsdorf, 2004), the degree of specialization seems to be much lower than in metropolitan regions in North America. According to de Duren (2006), this is partly the result of the outcomes of decentralization and suburbanization being deeply affected by the socioeconomic context in which they are implemented. On the one hand, she remarks, ‘Charles Tiebout's classical analysis (1956) of decentralization in the United States, where people can “vote with their feet”, does not translate easily to a population of limited resources. On the other hand, his valuation of mobility control as a market asset does apply to developers who can choose which municipality offers them the best conditions for investment’ (de Duren, 2006: 324).

As we have seen both in Latin America as a whole and in Bogotá more specifically, the result is a more complex pattern of segregation with more situations of proximity between rich and poor. Therefore, despite the higher concentration of high-income households in northern suburbs, the rich still do not represent the majority of the population in any one municipality — contrary to what has happened in exclusive North American suburbs. In fact, in 2006 the percentage of the population classified as belonging to the three highest estratos (four, five or six) did not exceed 15% in any one municipality of the Department of Cundinamarca.16

Second implication: suburbs will resist regional redistribution

Under a ‘typical’ metropolitan scenario such as that described by Orfield (2002), the outlying suburbs (and their constituents) would generally resist regional redistribution, as they would not want to take responsibility for the core city's social and economic problems. In the Latin American case, we would first expect urban municipalities to try and incorporate surrounding suburban municipalities or to promote tax-base sharing and for surrounding municipalities to resist these schemes.

Several researchers (Gouëset, 2005; Rodriguez-Acosta and Rosenbaum, 2005) have noted that there has indeed been resistance to the annexation of suburban municipalities by ‘core cities’ in Latin America, but that this has not generally been a ‘grassroots’ effort. Rather, resistance seems to have come from the local political class (as well as from provincial and departmental units of governments, in some cases). This is in sharp contrast to the broad-based protest movements against amalgamation in regions such as Montreal, Toronto and Los Angeles (Keil and Boudreau, 2005).

Based on the patterns observed in Latin America, two potential explanations for this general lack of interest of the public in issues of regional distribution suggest themselves.

First, the rich in many Latin American countries pay for the private provision of services that would otherwise be provided by local authorities, namely education and security. As far as utilities are concerned, these have also been privatized in most Latin American countries, such that the quality of service would have little to do with the municipality itself, and more to do with someone's ability to pay (as described by Ungar, 2007). For these reasons, affluent (and politically influential) suburban dwellers may be less sensitive to changes in the boundaries of administrative units.

Secondly, many suburban dwellers do not themselves distinguish between what services are offered where and by which unit of government. Indeed, ‘many of the municipalities adjacent to capital or other large cities in Latin America are already so close to, and so intertwined with, the “principal” city that it becomes almost impossible to identify where one ends and the other starts’ (Rodriguez-Acosta and Rosenbaum, 2005: 296). This is confirmed by Gilbert and Davila (2002) who, writing about the residents of neighboring municipalities in the Bogotá metropolitan region, remark that many already use hospitals and other such facilities in Bogotá (without paying for them through taxation). These residents would therefore not necessarily see how amalgamation could affect them personally as the city and the suburbs are already functionally integrated on many levels.

Third implication: suburbs will align politically against the core city

The last and most important implication of spatial segregation under a political regime such as that described in Orfield's Metropolitics (2002) is the struggle between suburban municipalities and the core city, one that is often framed in terms of regionalism versus localism (and is therefore intimately related to the second implication described above). This implication differs from the previous one in that we should expect to see not only a resistance to regionalism, but also a ‘coalition’ of suburban municipalities against the core city.

Our review of the literature in general and the case study of Bogotá in particular suggest that there are, indeed, conflicts and rivalry between core cities and suburbs in Latin America. However, these conflicts appear to be either mostly issue-specific or else embedded in the national political context. There is little evidence that coherent ‘suburban coalitions’ such as those described by Orfield (2002) and others (see, for example, Weir et al., 2005; Puentes, 2006) have formed in Latin America. In the case of Bogotá, alliances ‘of convenience’ were formed between suburban municipalities during the debate around the formation of the metropolitan area (for example, Chía and Cota), but these alliances were circumstantial, based on the narrow interests of politicians and, more importantly, they did not endure.

Our analysis indicates that the reasons for this might be found in the region's particular urban history. First, it bears repeating that capital cities in Latin America usually have a high primacy — which means that not only are they considerably larger than other large cities, but also much larger than their own suburbs, both in terms of population and in terms of political representation at the level of the state, department, province or at the national level. Bogotá, for instance, has a population of approximately 6.8 million people, whereas the 17 municipalities comprising the rest of the metropolitan region have a combined population of less than 1 million. This means that a coalition of suburbs in Latin America is unlikely to yield the same political power as might a suburban coalition in a North American metropolitan region, where suburbs sometimes represent more than 50% of a metropolitan region's total population.

Secondly, capital cities and their regions in Latin America are often seen as self-serving centers of power and are therefore often pitted against the outlying regions (or departments). Hence, the politically salient form of wealth disparity exists between the capital and the country's periphery — and is not necessarily found between the core city and periphery of the metropolitan region. In the case of Bogotá, for instance, the city ‘suffers from the exaggerated feelings of hostility that all Colombians born outside the city feel toward “their” capital. In a centralized country with strong regions, provincial politicians elected to Congress are expected to do all they can for their regions and do nothing to help Bogotá’ (Gilbert and Davila, 2002: 45). Overall, the urban–suburban struggle seems to be overshadowed by the core–periphery struggle at the national level.

Finally, it is interesting to mention that suburban life in Latin America — or at least in Colombia — is not necessarily seen in the same way as it is North America. In effect, Dureau (2003c; 2010) makes the case that for many ‘commuter’ households living in affluent suburbs such as Chía or working-class suburbs such as Soacha, the suburb is but a ‘step’ in the journey and not necessarily the final destination. She notes that a number of affluent households moved back to Bogotá in the 1990s and reports that many of the commuters she interviewed are planning to go back to the city at a later stage of their lives. In other words, urban dwellers appear to have a strong attachment to their core city, even though many of them decide to ‘shield’ themselves from it.


Latin American urban regions are changing at a rapid pace, and new patterns of segregation and fragmentation have appeared, which do not fit neatly into any theory of urban change. One could argue that the reality of Latin American cities has outpaced the theoretical apparatus for understanding them.

The analysis presented here suggests that patterns of urban spatial segregation in the Latin American metropolis (1) differ significantly from those found in the North American metropolis (if any generalization is possible) and (2) are not necessarily indicative of a metropolization of politics based on income or class segregation. In other words, suburbs in Latin America are not becoming politically polarized in the same way as they have been in North America, and inter-municipal conflicts seem to be less related to the distribution of resources and have more to do with the distribution of power among members of the local political elite. As a result, the movement against regional redistribution in metropolitan regions of Latin America does not appear to be broad-based nor does it seem to have enlisted many actors from civil society — in sharp contrast to such movements in North America.

As a consequence of the relative lack of interest of urban and suburban residents (and voters) in these issues, the main mechanism proposed by Orfield (2002) to build city-suburb coalitions — that is, the formation of a pro-redistribution political coalition of representatives or congressmen at the state or department level — is unlikely to be successful. Thus, it would seem that Latin American cities are indeed at a different place and on a different trajectory when compared to their North American counterparts. On the one hand, the patterns of segregation and fragmentation in Latin America are distinct from those found in North America. On the other hand, spatial segregation in Latin America seems to have little impact on the course of metropolitan politics — and there seems to be little hope of a city-suburb coalition for redistribution.

Orfield's account of metropolitics may not illuminate the process of metropolization in the Latin American metropolis, but applying his theory to this context helps us identify a different trajectory for cities of the region. This trajectory is not necessarily encouraging for poor municipalities in the hinterland of capital cities, as such municipalities are likely to continue receiving a large number of poor households without having the means to provide them with basic services. Indeed, despite the migration of affluent households to the periphery, there is little evidence for now of redistribution or trickling down — and little interest on the part of politicians to improve the living conditions of poor households in these peripheral municipalities.

De Duren (2006: 325), citing Mollenkopf and Castells (1991) points out that ‘polarized societies do not lack interaction between their extremes; rather they promote a social dynamic dependent on the furthering of these differences’. In light of the trends described here, we might ask whether the Latin American pattern of mixed fragmentation is likely to be self-reinforcing, or whether the greater physical proximity of rich and poor will eventually translate into greater opportunity and social mobility for the latter. We do not have an answer to this question, but the present trajectory could potentially lead both to a more fair or a more unequal city.


  1. 1

    See, for example, Alfonso (2009), who notes the relative lack of social mixing and social interaction between households of different estratos in Bogotá. See also Uribe et al. (2006), who describe the negative effect of Colombia's stratification system on social mobility.

  2. 2

    According to the World Bank (2003), social inequality in Latin America decreased slightly in the 1990s, but only after having rapidly increased during the 1980s. Specifically, the report states that ‘[i]nequality decreased in the 1970s during times of relative economic prosperity and increased in the lost decade of the 1980s … The recovery of the 1990s should have brought significant distributional improvements. However, there is no evidence that this has happened’ (ibid.: 47).

  3. 3

    It is interesting to note that the spatial models proposed by Arreola and Curtis (1993) and Ford (1996), some 15 years after Bähr and Mertins' first attempt (in 1982) reflect these new development patterns quite clearly: regional shopping malls are now described as focal points of suburban residential developments.

  4. 4

    The legal tradition of ‘home rule’ in the United States contrasts sharply with the relative lack of autonomy and fiscal capacity that still characterizes most cities in Latin America.

  5. 5

    It should be noted that governments in several parts of Latin America, including Colombia, have invested massively in transit and busways (see Wright, 2001, in particular); nevertheless, this trend has generally been accompanied by a rapid increase in car ownership. In Bogotá, for instance, it is estimated that the number of cars has more than doubled since the inauguration of the Transmilenio in 2000, from 662,000 to 1,374,000 (El Tiempo online, 2011b).

  6. 6

    Ungar (2007) speaks of the gated community as a part of the fortified city arsenal, which is deployed in response to perceptions of insecurity. He estimates that upward of 300,000 men and women are employed in private security in Argentina, and approximately 900,000 in Mexico and 1,600,000 in Brazil. Although it is not known exactly who hires these private security firms and in what proportions, there is no doubt that gated and closed communities are an important source of employment.

  7. 7

    The processes leading to the appearance of urban spaces worldwide in which physical proximity exists without social proximity are not unique to Latin America. They constitute, in the words of Graham and Marvin (2001) a new global form of ‘splintering’ urbanism characterized, among other things, by ‘glocal bypass’ and the fortification of private space. This aside, the case of Latin America is of particular interest, because these patterns of segregation are relatively new, but becoming widespread.

  8. 8

    The municipalities included in this list are El Rosal, Subachoque, Madrid, Mosquera, Zipaquira, La Calera, Gachancipa, Cajica, Sopo and Tocancipa.

  9. 9

    As Jaramillo (2007) notes, the population of Bogotá was less than 40,000 in the middle of the nineteenth century and its spatial organization was, at that time, no different from that of other Latin American cities. It was only during the period of rapid population growth, which occurred between 1870 and 1930, that the pattern of core–periphery segregation started to change. This interpretation is corroborated by Brewer-Carias (2008), who writes that physical proximity to the central square was for a long time in the history of Bogotá the main marker of social status.

  10. 10

    Cortes (2008) notes that experimentation with the construction of apartment compounds in Bogotá started in the 1950s, but only came to fruition after the departure of President Rojas Pinilla, who considered it ‘indecent’ to force families to live ‘on top of one another’.

  11. 11

    According to official documents, 22.6% of the budget of the city's central administration came from central government in 2011 (Alcaldía Mayor de Bogotá DC., 2011). Moreover, in the absence of a ‘metropolitan’ administrative unit, the land-use planning framework for the Capital District is still determined centrally.

  12. 12

    This has happened in spite of the fact that the Capital District encompasses approximately 88% of the metropolitan region's population. Indeed, there is no contradiction per se between central-city primacy within a metropolitan region and the development of metropolitics. See, for example, Gainsborough (2001) on the politics of regional cooperation in Houston, Texas, and Los Angeles, California.

  13. 13

    According to Article 325 of the 1991 Constitution, ‘the Capital District may form a metropolitan area with the adjacent municipalities and a region with other territorial entities of departmental character’ with the purpose of ‘guaranteeing the performance of the plans and programs of integral development and the timely and efficient provision of the services for which it is responsible’ (see República de Colombia, 1991).

  14. 14

    Article 326 of the 1991 Constitution states that ‘the adjacent municipalities may become incorporated into the Capital District if this is what the citizens who reside in them determine by means of a vote that will be held when the District Council has expressed its approval of such incorporation’ (emphasis added). Hence, the annexation of a suburb by the Capital District would require both municipalities to agree to it (see República de Colombia, 1991).

  15. 15

    It is important to note that voters in a given municipality are not necessarily residents of that municipality; indeed, Colombian citizens are entitled to vote where their identity document (cédula) is registered, which may or may not be where they live.

  16. 16

    In 2006, following the wave of migration to the suburbs that took place in the 1990s, only 12.6% of Chía's population was classified as belonging to one of the three highest estratos (four, five or six). The proportion of the population belonging to these estratos was 10.5% in Tabio, 7% in Sopo and 3.3% in Cota (Secretaría de Planeación de Cundinamarca, 2006).