poor in Lima, likewise, live in three great peripheral conos radiating from the central city; such huge spatial concentrations of urban poverty are also common in Africa and the Middle East. In South Asia, on the other hand, the urban poor tend to live in a much larger number of distinct slums more widely dispersed throughout the urban fabric in patterns with an almost fractal complexity. In Kolkata, for instance, thousands of thika bustees – 9 hutments of five huts each, with 45-square-meter rooms shared, on average, by an incredible 13.4 people – are intermixed with a variety of other residential statuses and landuses.29 In Dhaka, it probably makes more sense to consider the nonslum areas as enclaves in an overwhelming matrix of extreme poverty.
Although some slums have long histories – Rio de Janeiro’s first favela, Morro de Providencia, was founded in the 1880s – most megaslums have grown up since the 1960s. Ciudad Nezahualcóyotl, for example, had barely 10,000 residents in 1957; today this poor suburb of Mexico City has three million inhabitants. Sprawling Manshiet Nasr, outside Cairo, originated as a camp for construction workers building the suburb of Nasr City in the 1960s; while Karachi’s vast hill slum of Orangi/Baldia, with its mixed population of Moslem refugees from India and Pathans from the Afghan border, was founded in 1965. Villa El Salvador – one of Lima’s biggest barriadas – was established in 1971 under the sponsorship of Peru’s military government, and within a few years had a population of more than 300,000.
Everywhere in the Third World, housing choice is a hard calculus of confusing trade-offs. As the anarchist architect John Turner famously pointed out, “Housing is a verb.” The urban poor have to solve a complex equation as they try to optimize housing cost, tenure security, quality of shelter, journey to work, and sometimes, personal safety. For some people, including many pavement-dwellers, a location near a job – say, in a produce market or train station – is even more important than a roof. For others, free or nearly free land is worth epic commutes from the edge to the center. And for everyone the worst situation is a bad, expensive location without municipal services or security of tenure. In Turner’s celebrated model, based on his work in Peru in the 1960s, rural migrants first move from the province to the city center – location at any price – to find jobs; then, with employment security, they move to the periphery, where ownership is attainable. This progress from (in his terminology) “bridgeheader” to “consolidator” is, of course, an idealization that may only reflect a historically transient situation in one continent or country.36
Figure 7. 30 Largest Megaslums (2005)30
| (Millions) | |
| 1. Neza/Chalco/Izta (Mexico City)31 | 4.0 |
| 2. Libertador (Caracas) | 2.2 |
| 3. El Sur/Ciudad Bolivar (Bogotá) | 2.0 |
| 4. San Juan de Lurigancho (Lima)32 | 1.5 |
| 5. Cono Sur (Lima)33 | 1.5 |
| 6. Ajegunle (Lagos) | 1.5 |
| 7. Sadr City (Baghdad) | 1.5 |
| 8. Soweto (Gauteng) | 1.5 |
| 9. Gaza (Palestine) | 1.3 |
| 10. Orangi Township (Karachi) | 1.2 |
| 11. Cape Flats (Cape Town)34 | 1.2 |
| 12. Pikine (Dakar) | 1.2 |
| 13. Imbaba (Cairo) | 1.0 |
| 14. Ezbet El-Haggana (Cairo) | 1.0 |
| 15. Cazenga (Luanda) | 0.8 |
| 16. Dharavi (Mumbai) | 0.8 |
| 17. Kiberi (Nairobi) | 0.8 |
| 18. El Alto (La Paz) | 0.8 |
| 19. City of the Dead (Cairo) | 0.8 |
| 20. Sucre (Caracas) | 0.6 |
| 21. Islamshahr (Tehran)35 | 0.6 |
| 22. Tlalpan (Mexico City) | 0.6 |
| 23. Inanda INK (Durban | 0.5 |
| 24. Manshiyet Nasr (Cairo) | 0.5 |
| 25. Altindag (Ankara) | 0.5 |
| 26. Mathare (Nairobi) | 0.5 |
| 27. Aguas Blancas (Cali) | 0.5 |
| 28. Agege (Lagos) | 0.5 |
| 29. Cité-Soleil (Port-au-Prince) | 0.5 |
| 30. Masina (Kinshasa) | 0.5 |
In a more sophisticated analysis, housing expert Ahmed Soliman discusses four basic shelter strategies for the poor in Cairo. First, if access to central job markets is paramount, the household can consider renting an apartment; the rental tenements offer centrality and security of tenure, but are expensive and hold out no hope of eventual ownership. The second option is centrally located but informal shelter: a situation described by Soliman as “a very small room or rooftop with a location with a poor quality environment and a cheap rent, or no rent at all, with good access to job opportunities but with no hope of secure tenure. Such illegal dwellers will eventually be forced to move to squatter camps or semi-informal housing.”37
The third and cheapest housing solution is to squat on publicly owned land, usually on Cairo’s desert outskirts and almost always downwind of pollution; negative trade-offs include the very high cost of commuting to work and the government’s neglect of infrastructure. “For example, the squatter area in El Dekhila district has been a settlement for 40 years with no public action or intervention from the local authority.” The fourth solution, eventually preferred by most poor Cairenes, is to buy a house site in one of the vast semi-informal developments (often on land purchased from Bedouins or peasant villages) with legal tenure but without official building authorization. Although far from jobs, such sites are secure and, after considerable community mobilization and political negotiation, are usually provided with basic municipal services.38
Similar rational-choice models can be specified for all cities, generating a huge array of locally specific tenure and settlement types. The typology displayed in Figure 8 is an analytic simplification that abstracts from locally important features for
