Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

1. The Latin American mega-city: An introduction

What is a mega-city?
Are mega-cities different from smaller cities?
Should anything be done to slow down the growth of Latin America's big cities?
The future for Latin America's cities
Structure of the book

Alan Gilbert

By the year 2000, the world will contain 28 mega-cities with more than eight million people each (UNDIESA and UNU, 1991: 6). Twenty-two of these giants will be in less developed countries and five in Latin America. In 1990, Latin America already contained four such cities (Mexico City, São Paulo, Buenos Aires and Rio de Janeiro) with a combined population of around 50 million (table 1.1, figure 1.1). Approximately one in nine of all Latin Americans lived in these four cities; one in six of the region's urban population. By the year 2000, the region may well contain the world's two largest cities, Mexico City and São Paulo.

The sheer number of people living in Latin America's mega-cities is not the only reason for looking at them carefully. Unfortunately, they also demonstrate many of the worst symptoms of the region's underdevelopment: vast areas of shanty towns, huge numbers of poor people, high concentrations of air and water pollution, and serious levels of traffic congestion. Many observers in Latin America have little confidence that such enormous conglomerations are manageable; many worry about the future. Not untypical are the fears expressed by Sanchez-Leon (1992: 201) about one of the aspirant mega-cities:

Table 1.1 Latin America's giant cities, 1995

City Population (millions)
São Paulo 16.42
Mexico City 15.64
Buenos Aires 10.99
Rio de Janeiro 9.89
Lima 7.45
Bogotá 5.61
Santiago (Chile) 5.07
Belo Horizonte 3.90
Pôrto Alegre 3.35
Recife 3.1 7
Guadalajara 3.1 6
Caracas 2.96
Salvador 2.82
Monterrey 2.81
Fortaleza 2.66
Santo Domingo 2.58
Curitiba 2.27
Havana 2.24

Source: United Nations, Department of International Economic and Social Affairs, 1995:132 9.

The Lima of today, with its population of five million, has changed Peru and these changes are also its problems - because Lima is a problem difficult to unravel. A large part of the population, particularly the children and young people, lives in poverty. These are the children of chaos, of poverty, and of urban violence. This vast population exists in a city that does not offer them any chance of personal and social growth, though it has allowed, and even encouraged, a number of creative, organized initiatives.

What is a mega-city?

Mega-cities are "cities that are expected to have populations of at least eight million inhabitants by the year 2000" (UNDIESA, 1986: iii). Like so many threshold figures used in the social sciences, this minimum size seems to have been plucked from the air. There is no theoretical basis for asserting that a city with eight million people is qualitatively different from one with rather fewer inhabitants. Perhaps for this reason different authorities use widely differing definitions of a mega-city. Although Richardson (1993) and UNDIESA and UNU (1991) use the eight million benchmark, UNCHS (1987: 29), Ward (199O: xvii), and the World Bank (1991: 16) use ten million, while Dogan and Kasarda (1988: 18) implicitly use four million.

Figure 1.1 Latin America's giant cities

The term "mega-city" is frequently used as a synonym of words such as "super-city," "giant city," "conurbation," "megalopolis," "world city," and so on. Unfortunately, there is little agreement about what any of those terms means either. Entry to the world of giant cities or metropolises is granted by Dogan and Kasarda (1988: 18) at four million, to the rank of super city by Lowder at five million (1987: 5), and the term "megalopolis" is attributed by Mayhew and Penny (1992) to "any continuous built up area of more than ten million inhabitants." Another term, "world city," seems to be only loosely linked to size at all (Hall, 1977; Friedmann, 1986). World city status, whatever that actually means, is more closely linked to function: "several modestly sized cities, e.g. Washington DC and Geneva, do fulfil true global roles whereas many far larger conurbations patently do not" (Simon, 1992: 185). That there is no adequate definition of a mega-city was recognized by a United Nations seminar which concluded that "there is a need to work out some sort of definition" (UNDIESA and UNU, 1991). While we wait for the formulation of an adequate definition, this book will work with the figure of eight million.

Are mega-cities different from smaller cities?

In what ways do cities with eight million people or more differ from smaller cities? A major difficulty in answering that question is that there is very little firm evidence to support the case one way or the other. The principal difficulty is to separate the effects of size from those of other variables (Richardson, 1973). Certainly, attempts to examine the relationship between size and the benefits and pathologies of urban growth have produced very little in the way of reliable results. There is seldom any clear relationship between size and any single variable of welfare or illfare. For example, while it is very easy to identify Los Angeles, New York, and Rio as giant cities which have terrible crime rates, other mega-cities such as Tokyo and Shanghai do not suffer from a great deal of crime. Nor is there any clear relationship within countries between city size and the quality of life. To continue with the example of crime, violence is worse in Rio than in the much larger São Paulo, in Detroit when compared with its giant neighbour New York, and in six Colombian cities much smaller than Bogotá (Richardson, 1973; Coyuntura Social, August 1993: 32).

Equally problematic is the fact that some large cities suffer from different problems than other cities of similar size. Most of the differences can be attributed to intervening variables. Air pollution is worst in cities with a great deal of manufacturing industry (Shanghai, Seoul, and São Paulo), in cities which use coal as a domestic and industrial fuel (Shanghai and Beijing), and in those which suffer regularly from temperature inversions (Los Angeles, Mexico City, and São Paulo).

Other large cities suffer much less seriously from air pollution (UNEP, 1992). Certainly the debate about optimum city size suggests that urban problems are not generally worse in giant cities, except possibly with respect to traffic congestion, land prices, and crime. Very large cities also have certain advantages particularly with respect to economic performance and service provision (Richardson, 1973).

Nevertheless, it is possible to argue that size of city does make a difference to certain kinds of problem (Gilbert, 1976; Johnston, 1976). Even if size effects are irregular and unsystematic, that does not mean that size makes no difference. Certainly UNDIESA and UNU (1991: 7) concluded that "there was general agreement that a mega-city is different from an ordinary city." White and Whitney (1992) also believe that city size affects environmental conditions; their argument is summarized in table 1.2.

White and Whitney argue that the poor living in mega-cities in less developed countries face worse problems than those living in smaller cities. The poor are particularly disadvantaged in terms of job opportunities, the cost of food, transportation, the environment, crime, and housing. By contrast, the rich in poor countries face additional difficulties in large cities - higher crime rates, worse housing, and environmental problems - but also benefit, notably through better job opportunities, education, and health facilities.

Table 1.2 Pressure points in cities in less developed and more developed countries by size of centre


Less developed countries

More developed countries

Largea Small/medium Largeb Small/medium
Job opportunities H M L M
Food H M L L
Water H (L) M L L
Energy H H L L
Education H H L L
Transportation H M L L
Recreation L L M H
Health H H L M
Environment H L H M
Crime H L H L
Housing H L H L

Source: White and Whitney, 1992:16.
H = High pressure point
M = Medium pressure point
L = Low pressure point
(L) = Different strength of pressure point for low-income and MDC populations
a. Plus low-income populations in more developed country cities.
b. Plus upper-income populations in less developed country cities.

To what extent can we agree with this argument? A first response to the table is that it is far too generalized. Large cities in less developed countries vary dramatically in their socio-economic characteristics: Buenos Aires has little in common with Calcutta despite their similarity in size. A second response is to query the lack of dimensions. White and Whitney define neither less developed countries nor large cities. It is difficult to know therefore whether they are referring to cities with more than 100,000 people or to those with more than 10 million. How large is a large city? At what point does a country move from the less-developed to the more-developed category? A third response is to question whether the urban poor in developed countries really live lives like those in poor countries. Is inner-city life for the poor in Washington DC really like that of the poor in Calcutta?

To what extent does White and Whitney's estimate of the environmental effects of city size match the evidence from Latin America? Are Buenos Aires, Mexico City, Rio, and São Paulo different from smaller Latin American cities? Unfortunately, there is little in the way of reliable data with which we can compare the quality of life in cities of different sizes. Certainly, no study has calculated an overall indicator of the urban quality of life on which such a comparison could be based. Therefore, we are forced to use more qualitative information, even anecdotal material. On this basis it is possible to argue that the general quality of life in the largest cities is seldom worse than that in smaller cities. Indeed, in many respects it is frequently much better; that is precisely the reason why so many people have moved to Latin America's largest cities over the years. However, as White and Whitney suggest, size is likely to affect different variables and different income groups in different ways. Consequently, it is worth considering each of the variables on their list and to observe how size affects each indicator in Latin American cities.

Job opportunities

Until 1980, there is little doubt that there were more jobs, and more better paid jobs, in the largest cities. While the situation undoubtedly changed during the economic crisis of the 1980s, particularly in Mexico, it is difficult to conclude that the employment situation is more difficult in large cities than in smaller ones. Table 1.3 certainly suggests that there is no clear pattern. In Brazil, unemployment rates are lower in São Paulo and Rio de Janeiro than in the six largest metropolitan areas of the country combined. In Argentina, unemployment is higher in Buenos Aires than in the other urban areas combined, although it is currently higher in Córdoba and Rosario than in the capital. In Colombia, during the 1990s, Bogotá had consistently lower levels of unemployment than Barranquilla, Cali, and Medelín. In Peru, Lima has a higher level of unemployment than the rest of the country; in Venezuela, Caracas has a lower rate. In Mexico, unemployment is higher than average in the capital, but some smaller cities have higher rates of unemployment. In sum, there is no clear conclusion. Differences in occupational structures are clearly more important determinants of unemployment rates than size of city.

Table 1.3 Unemployment levels by city and country (percentages)

  1992 1993
Urban areas 7.0 9.6
Buenos Aires 6.7 10.6
Brazil 4.8 n.a.
Main 6 cities 5.8 5.3
Rio de Janeiro 5.4 n.a.
São Paulo 3.4 n.a.
Chile n.a. 4.6
Santiago 4.9 4 0
Main 4 cities 10.0 8.5
Bogotá 8.4 7.4
35 cities 2.8 3.3a
Mexico City 3.4 3.9a
Monterrey 3.2 4.9a
Tampico 5.0 5.4a
Peru 6.7 n.a.
Lima 9.4 9.9
Venezuela 8.0 6.6
Caracas 5.7 n.a.

Sources: ECLAC, 1994; Revista del Banco de la República, September 1993; Webb and Baca de Valdez, 1992; INEGI, 1993: table 2.2.
a. First six months of year.

Nor do poverty levels seem to be higher in large cities. As Tolosa shows in chapter 9, Rio de Janeiro has a higher proportion of poor people than Belo Horizonte, but São Paulo has a much smaller proportion of poor than either Rio or Belo. Studies comparing living conditions in Buenos Aires and the rest of Argentina show that structural poverty is much less marked in the metropolitan area than in the country as a whole (CEPA, 1992 and 1993). In the cities of Colombia, the incidence of poverty among households in Bogotá in 1985 (18 per cent) was half the rate found in Medellín and was lower than the average of thirteen large Colombian cities combined (27 per cent) (Muñoz, 1991: 286). A United Nations study of poverty levels in the middle 1980s found that levels of poverty in Bogotá, San Jose, Panama City, Lima, Montevideo, and Caracas were all lower than those found in other urban areas of their respective countries (Fresneda, 1991: 164). Most large cities, therefore, compare very favourably on poverty indicators (Bolvinik, 1991).


Unlike the situation in African cities, the urban poor in Latin America produce little food in their backyards. In this respect there is little difference between living in a small or a large city; there is too little spare land for a productive garden in most Latin American cities. Many people obtain food outside the market, but mainly by retaining links with kin in the countryside and returning frequently for visits. When they come back to the city, the buses are full of chickens, vegetables, fruit, and other kinds of produce. The cost of food in large cities is probably higher than in smaller cities, but much depends on where cities are located. Food costs are much higher in Ciudad Guayana or Brasilia than in Caracas or São Paulo. In any case, urban subsidies reduce food costs - at least they did until 1980, when many subsidies were cut. If recent cuts to subsidies have made life more expensive for the poor in large cities, the same is true for the poor in every city. Evidence on malnutrition is scarce but there is little to suggest that it is more rampant in the largest cities; ironically it is usually higher in the rural areas. In the late 1970s, for example, 62 per cent of rural families were malnourished compared to 26 per cent of urban families (Wilkie and Perkal, 1984).


As in most third world countries, water networks are better in most large Latin American cities than in most smaller cities (Richardson, 1993: 50). Chapter 4 shows that service conditions in large cities are generally superior to the same country. This is a consequence of huge investments in infrastructure provision during the 1960s and 1970s. The major cities have always received more resources from central government because of their political importance. They have also better organized servicing agencies both because of the pressure exerted by the large number of powerful companies located there and because of their higher per capita tax revenues.

Of course, there are exceptions to this general tendency. Buenos Aires has relatively poor water provision because of the extensive use of wells in large swathes of the city. Similarly, Caracas is served worse than the rest of urban Venezuela. In general, however, the largest cities have better water provision.


Electricity is available in most homes. There are few reliable data about provision by city but it is certain that most large cities and most industrial centres have the best services. The worst problems are to be found in small cities. In Colombia, indeed, there were numerous civic strikes during the 1980s as a result of such service deficiencies (Santana, 1989).


Literacy rates often show marked differences between cities of different size, although the pattern is not always quite that expected. In Colombia, for example, there is remarkably little difference in the level of education of the population in different cities except at the university level, where Bogotá has far more graduates than most smaller cities (Coyuntura Social, May 1990: 43). Similarly, in Brazil, whereas 16 per cent cannot read or write in the intermediate cities of the State of São Paulo, the proportion in Greater São Paulo is 20 per cent (see chapter 10). In general, however, education provision is better in larger than in smaller cities.


Figueroa argues in chapter 5 that large cities suffer from worse traffic congestion than smaller cities, that rush-hour flows into central areas are higher, and that average journey times tend to be much longer. Similarly, Richardson (1973: 29) recognizes that "increasing city size probably does generate higher traffic congestion costs." The reason is that distances between home and workplace are usually longer. It is also because large cities are more prosperous and have higher levels of car ownership. However, even if traffic problems become more acute as cities grow in size, the level of traffic congestion is determined more by a city's shape and its land-use structure. Some small cities suffer badly from traffic congestion.


There can be little doubt that the major cities have a much wider array of leisure facilities. There are more cinemas and theatres, more varied and higher-quality shopping facilities, and a wider range of clubs of all types. It is unlikely, of course, that the poor gain access to most of these facilities; even attending a football match is beyond the budget of many. Nevertheless, the range of recreation opportunities is still higher for the poor in the larger cities.


Health care in any Latin American metropolitan areas is generally much better than in the rest of the country. There are more hospitals and doctors per head than in smaller cities because there is a much higher level of private healthcare provision. In Brazil, the ratio of doctors, nurses, and hospital beds to inhabitants is superior in Rio and São Paulo compared to the situation in most other metropolitan areas (see chapters 9 and 10). However, health conditions are not better in São Paulo than in the medium-sized cities of the same state. In 1985, life expectancy was one year lower in Greater São Paulo and there was a vast difference in infant mortality rates: 31 babies out of every thousand died in the intermediate cities compared to 54 in the metropolitan area. Greater São Paulo only does better when compared with smaller cities in the north-east of the country, where health conditions are appalling (Wood and Carvalho, 1988).


Several major Latin American cities, notably Mexico City, Santiago, and São Paulo, suffer badly from air pollution. However, air pollution is not only a function of size. Mexico City suffers because it is an industrial city, has vast numbers of cars, is in a relatively dry part of the country, and suffers from temperature inversions in winter (see chapter 8). Rio de Janeiro suffers less from air pollution because it has sea breezes to clear the fumes. Despite their size, many small cities suffer very badly from air pollution; until action was taken in 1984, the city of Cubatão in Brazil was notorious for its noxious air, the consequence of uncontrolled emissions by heavy industries in a location prone to temperature inversions.


Pinheiro (1993: 3) argues that "it is clear that the rapid growth of large cities and the cramming of their increasingly impoverished inhabitants into restricted areas has undermined sociability and increased the level of violence occurring in conflicts." Richardson (1973: 102) agrees, although with an important proviso: "In conclusion, the evidence is clear that the incidence of crime is higher in big cities than in smaller towns, and in urban than in rural areas. The social and economic costs of crime are greater for big city residents than for others ... [nevertheless] it is much less convincing to argue that more crime is a direct result of city size." Evidence from Latin America shows that certain large cities have lower crime rates than some smaller cities. In Colombia, the murder rate in Medellín is much higher than that in Bogotá.


Chapter 4 shows that size of city influences residential tenure. While the evidence is hardly conclusive, La Paz, São Paulo, Bogotá, Mexico City, and Caracas all have distinctly lower rates of home ownership than their respective national urban averages. Of course, the relationship between size and ownership is not close because too many other variables influence the level of ownership (Gilbert and Varley, 1991). Nevertheless, there is a systematic tendency for larger cities to have higher levels of rental tenure. Richardson (1993: 39) attempts to explain this finding in terms of the cost of land and infrastructure: "Land values are an exponential function of city size ... For most households, accommodation costs in large cities are usually higher than in smaller cities."

Overall, therefore, the case that size of city makes a real difference to living standards in Latin America is unproven. Size probably causes additional problems in terms of transport, crime, and housing, but produces clear advantages in terms of work, infrastructure, and services. Poverty indicators also tend to be lower in large cities. Unfortunately, this kind of analysis does not take us very far, given that it takes no account of government policy. For many years, Latin American governments tended to pamper their largest cities. Not only did such cities contain a significant proportion of active voters, they also had most of the gossipping classes. In so far as the largest cities were also capital cities, governments were particularly sensitive to protest in their backyards. Mega-cities needed to be placated and, consequently, large cities did well in terms of government spending. Large-city bias meant that both rich and poor were treated rather better than their cousins elsewhere.

Despite this analytical problem I would suggest that mega-cities do face special problems in three important areas which have yet to be addressed. The first is in the area of public administration. Running a large city is complex and is made more difficult by the way so many giant cities have spread into neighbouring administrative areas. Of course, such problems are not peculiar to large cities. As Davey (1993: 3) points out: "Porto Alegre, a city of 1.4 million people, is only the largest of 22 municipalities in a metropolitan region of 2.5 million [and] Recife, with 1.3 million people, is one of 12 municipalities in a metropolitan region of 2.3 million." Nevertheless, the problems are bound to be more serious as cities grow and spread into neighbouring areas.

A second problem relates to the issue of local democracy. It is difficult for governments to consult their populations in smaller cities but it is particularly difficult in a city of, say, 15 million people. This is perhaps a contributory factor in explaining why so few capital cities have popularly elected local administrations (see chapter 3).

Finally, the issue of social equity is more complicated in large cities. Chapters 6 to 11 are consistent in deploring the high level of inequality in Latin America's largest cities and in arguing that the distribution of income has deteriorated in recent years. And, although there is little evidence that levels of income inequality are higher in large cities than in smaller cities, the poor are more likely to be segregated in distant ghettoes in the mega-cities. Because of the larger distances and the larger settlements in giant cities, rich and poor are almost bound to live further from one another (but see chapter 4). While this is hardly likely to affect their social lives - rich and poor would not mix anyway - many would argue that healthy cities should not be so starkly divided. People in different income groups should come into contact on a regular basis.

The rich certainly gain more from living in large cities than the poor. The rich are more likely to get jobs, the poor less likely; service provision for the rich is good, that for the poor problematic. Critically, the rich are able both to benefit from the advantages of large cities and to escape most of the diseconomies. Thus, even if average conditions are better in the mega-cities, the poor may gain little advantage. If the poor gain from superior infrastructure, such as education and water supply, other apparently superior facilities may not help them. Large cities may have excellent hospitals, clubs, restaurants, and universities, but most of these are open only to those with money. The poor may as well be living in a different city as far as these kinds of facility are concerned. Similarly, the range of better-quality jobs available in the largest cities benefits the poor only in so far as such jobs create more work in lower-paid activities. The poor, lacking university education and appropriate social skills, will not gain access to highly paid forms of employment. Many of the advantages of large cities, therefore, are not on offer to the masses.

The poor also reap more of the disadvantages. In so far as levels of car ownership are much higher in large cities, traffic congestion is worse. Of course the rich are also delayed by traffic jams, but the affluent can listen to their car stereos, whereas the poor are held up in crowded buses. Similarly, air and water pollution is likely to affect the poor much more than the rich. Indeed, those with enough money can buy accommodation in the least polluted areas. Pollution does reach the high-income suburbs in some Latin American cities, but the rich always live in the least affected areas. The rich can also escape more easily from noise and pollution. They belong to country clubs, and they can also get out of the city at weekends because they are more likely to own a car and even a second home. In sum, urban problems in large cities affect rich and poor in very different ways.

Contents - Previous - Next