Contents - Previous - Next

Migrants no more and escaping the city?

It has already been asserted in the introduction to this section on rural-urban linkages that, despite a prevalent view to the contrary, African urbanization rates have fallen in response to the economic crisis (Potts, 1994). This is due mainly to two factors: rates of immigration have declined, and rates of "return" migration have tended to increase; thus net in-migration has often fallen significantly. These are of course quite rational responses, in line with what one might expect given the narrowing or reversal of the rural: urban income gap.

A range of census data on urbanization for selected African countries is presented in table 13.2, with comparisons with World Bank estimates. The choice of countries has been very largely determined by the availability of census data for relevant periods; unfortunately very few countries meet even this simple criterion, and in some cases (e.g. Sierra Leone) the data were deemed too unreliable to be worth analysing. Although it is acknowledged (see also chaps. 2 and 3 in this volume) that African census data should be treated with some caution, it is contended that the evidence of a significant slackening in the rate of population increase in a number of cities in a range of countries, combined with the fact that those countries that have suffered the worst urban crises have recorded the greatest fall in urbanization rates, suggests that a real trend is being identified.

Ghana was one of the first countries to suffer serious economic decline, and it is evident from table 13.2 (see also chap. 3) that it also experienced very limited urbanization during its second post-independence intercensal period of 1970-1984. Accra's growth, which had generally been thought to be very high, was only 20 per cent higher than national population growth rates (compared with more than double between 1960 and 1970). It is difficult to know what natural increase rates are in most African cities, but it is here assumed that they are not very different from national rates. This is because, although there does tend to be lower fertility in African urban areas, death rates are lower and birth rates are often higher¹⁴ because they are boosted by the more youthful composition of the urban areas, with a higher proportion of people being in childbearing age groups. Unbalanced sex ratios used to reduce urban births, but by the 1970s sex ratios in many major cities were often about parity. On this basis, therefore, it is argued that the Ghana census data in table 13.2 show that net rural-urban migration slowed very significantly. Indeed, for Kumasi and Sekondi-Takoradi, the second and third cities, this period must have seen a net out-migration of urban people to rural areas.

It is also clear that Uganda experienced significant reverse migration from major towns in the 1970s, and even the capital city did not grow much faster than the national population. The 1980s, however, saw a return to net in-migration and higher urban growth rates.

Table 13.2 Comparisons of average annual growth rates of national, urban, and selected city populations for selected African countries, 1960s-1990s (census data and other estimates)^a

Sources: Ghana census data - Ghana, 1984 Census Provisional Results, Accra, 1984; Tanzanian census data - United Republic of Tanzania, Bureau of Statistics, President's Office, Planning Commission, 1988 Population Census: National profile, basic demographic and economic characteristics, Dar es Salaam, 1992, Population Census: Regional profile: Dar es Salaam, Dar es Salaam, 1992, and other regional profiles; Barke and Sowden (1992), table 2; Holm (1992), table 1; Uganda census data - Republic of Uganda, Report on the 1980 Population Census: Volume I: Provisional results by administrative area, Census Office, Kampala, 1982; and Republic of Uganda, Statistics Department, Final Results of 1991 Population and Housing Census (Prerelease), Kampala, 1992; Zambia census data - Republic of Zambia, Central Statistical Office, 1990 Census of Population, Housing and Agriculture: Preliminary report, Lusaka, December 1990; World Bank data - 1991 World Development Report, cited in Jamal and Weeks (1993).

a. Intercensal periods indicated by superscript on national population growth rates (e.g. 60-70= 1960-1970). Estimates of growth rates are in italics, with periods indicated in superscript.

b. Estimates taken from Peil (1991), p. 19.

c. "Cleaned" census data from Holm (1992), table 1.

For Zambia a similar picture emerges of significantly reduced urban growth by the 1970s, and the level of urbanization increased a mere 2 per cent from 40 per cent in 1980 to 42 per cent in 1990. Although the capital city, Lusaka, has maintained a high growth rate, experiencing significant net in-migration for the whole post-independence period, the Copperbelt towns have tended to grow more slowly than the national population since the 1960s; in fact their experience could be typified as counter-urbanization. Their reduced growth accords with Zambia's economic decline, which began in the early 1970s with the usual exigencies of a debt crisis, exacerbated by a dramatic collapse in copper prices and expensive transport problems caused by regional political instability and its principled opposition to white majority rule in neighbouring countries. The reductions in growth in many Zambian cities are obvious from table 13.2.

The situation in Tanzania is much less clear cut than in Uganda and Zambia, for, although there is clear evidence of a reduction in the growth rate of Dar es Salaam and many other individual cities in the 1980s compared with the 1960s and most of the 1970s, the census data indicate that net in-migration is still contributing significantly to growth in most urban areas (see also chap. 3). Table 13.2 shows that for Tanzania there has been a fall in the rate of urban growth, at least for the major centres including the capital city, but there is no evidence of net out-migration, as in some other countries. There are, however, a number of problems in interpreting the census data for Tanzania.¹⁵ It is unclear how the census defined and enumerated the total urban population. The growth rates of the major urban areas shown in table 13.2, however, suggest a lower overall growth rate than 10.7 per cent. Holm has managed to manipulate the census data to establish realistic growth rates for some other urban size categories for five specific regions: Iringa, Kilimanjaro, Arusha, Morogoro, and Mbeya, which are also shown in table 13.2. Although high (between 5.6 and 7.0 per cent), these still do not suggest a national rate of urbanization of over 10 per cent per year. It may be that the establishment of ujamaa nucleated villages in the 1970s led to the reclassification of very large numbers of new settlements as "urban" by 1988, explaining this seeming anomaly. Holm's study indicates that the scope for successful urban agriculture in intermediate and small towns in Tanzania may mitigate the downward pressure on rural-urban migration exerted by the impact of recession on wages and food prices. This helps to explain how the pressures for returning to rural areas are moderated in the specific context of these smaller towns, and points to the importance of detailed case-study material for interpreting census data, as well as the need to be aware that adaptations to economic decline may vary across the urban hierarchy. In Nigeria, for example, further dramatic declines in urban living standards in 1994 are reported to have encouraged Lagos residents to consider moving to second-order cities where the cost of living might be cheaper, or to return to home areas (Abiodun, personal communication, 1994). Thus reductions in the growth of large cities might in some cases be balanced by somewhat enhanced growth in some smaller centres.

Even in Zimbabwe, which experienced rapid urban growth in the 1980s (although lower than that projected), the new economic policies of the 1990s, which follow IMF prescriptions very closely, may be having a dampening effect on rural-urban migration. Research on recent migrants to Harare in 1994, for example, indicates that they almost universally perceive the impact of current economic policies to be far worse in urban than in rural areas (Potts with Mutambirwa, forthcoming).

In the light of the above discussion, it is clear that rates of rural-urban migration have often been lower than projected. In a number of settlements in various countries, it is also evident that there has actually been net out-migration. Such "return" migration has in fact been reported in a variety of sources, although it is usually in passing and rarely are any data produced to support the contention. Examples include Nigeria, where Collier (1988, p. 778) states that the relative decline in urban incomes between 1980 and 1985 led to a reversal of net migration to towns, with many return migrants cutting short their migrant careers (see also Mosley, 1992).

In Uganda, Jamal and Weeks (1993) accept that urban residents have migrated to rural areas as one coping strategy, despite their belief that, in general, rural-urban migration has not abated. They state that: "wage-earner households re-established their links with the countryside, reversing the labour force stabilization process of the 1960s.... Unlike in other African countries, urban migration actually slowed down or even changed its net flow up to 1985" (Jamal and Weeks, 1993, p. 39). Urban-rural migration in Uganda is also noted by Bigsten and Kayizzi-Mugerwa (1992, p. 1425), who state that reverse migration was "quite substantial" in the 1970s and early 1980s.

For Zambia, increased rates of return migration as urban economic conditions deteriorated in the 1980s have been noted by a number of analysts (Ferguson, 1990; Purbrick, 1990; Mulenga, 1991; Macmillan, 1993; Rakodi, 1994a). Pottier (1988) had earlier signalled that rural-urban migration patterns were changing in his anthropological study Migrants No More. Purbrick (1990, pp. 145-146) notes that changes were already occurring in the 1970s:

[The 1980 census] indicated a significant increase of urban-rural migration. Urban-rural migrants were identified as urban-born people with links in the rural areas, dependents of rural-urban migrants and retiring or unemployed migrants. The migration of urban-born people to the rural areas was a fundamental departure from trends in the colonial period and the first years of independence. It indicated that urban distress had mounted throughout the 1970's and 1980's. Increasingly people were looking to the rural areas for a viable survival strategy... the long-term nature of the 1970's economic decline has meant that a return to the land is the only option for many in the Copperbelt. On the basis of surveys conducted on the Copperbelt in the mid-1980s, Ferguson (1990) argues that unemployed and retired urban-dwellers were increasingly adopting strategies that involved them resuming to rural "homes." Records from one mine indicated that 90 per cent of ex-miners intended to return "home," even if they had never actually lived there.

Access to land is a key variable in determining the extent to which urban-rural migration can operate as a successful coping strategy. This will vary not only between countries but also between different groups of residents within any one city, because they will be drawn from or associated with regions with differing land resources and allocation processes. In addition, there may be variations in the ease of access to land according to length of stay in the city and how successfully the urban-dweller has maintained his or her links with the "home" village. Female-headed households will frequently be much less able to "return" because their access to rural land is generally far more constrained than a man's (and may be the reason why they left for town in the first place).

The significance of land tenure in affecting return migration processes is emphasized by Andrae (1992), who studied textile workers in the Nigerian towns of Kano and Kaduna. They were living with their wives and children in town, but many expected to have to go (reluctantly) to "home" rural areas if they were made redundant, where 70 per cent had access to adequate land. Andrae also emphasizes that commercialization of land would very much affect return migration. Freehold title to land is now the norm for parts of Kenya, for example, and here Andrae believes that the "high level of commercialisation of land ownership has closed this escape route for large sections of the urban working class" (1992, p. 219). The very severe land shortages that characterize the rural sector of a country such as Egypt also constrain such options for the urban poor (see chap. 4 in this volume).

Further evidence of urban-rural migration comes from Ghana and Tanzania. In Ghana, Jeffries (1992) reports that many redundant public sector employees invested in farming, because the drastic narrowing of the rural: urban income gap made cash-cropping appear more advantageous than for many years. The census data for Ghana (see above) suggest that reductions in net migration have been occurring over a considerable period for some centres. For Tanzania, work by von Troil (1992) suggests that here return migration is perhaps a phenomenon of the 1990s rather than the 1980s In a paper that mainly focuses on the causes and effects of rural-urban migration in 1984/85, she indicates that by 1990, after structural adjustment had been operating for some time, there had been some changes in urban survival strategies, specifically including a return to the village. Thus the increasing gap between wages and food costs, together with improved farm prices, meant that "many of the young men who had left earlier to try their luck in towns had been persuaded to return and take up farming" (von Troil, 1992, p. 235).

In this volume, Piermay's analysis of contemporary Kinshasa (chap. 7) also considers whether that city has experienced a decrease in population, owing to the remarkable severity (even by sub-Saharan African standards) of Zaire's political and economic climate. Dubresson (chap. 8 in this volume) notes that migration to Abidjan had slowed by the end of the 1980s. Many of the migration and rural linkage strategies discussed here have apparently been adopted in that city, and in the case of Côte d'Ivoire there is some evidence that "return" migration is boosting the growth of smaller towns, at the expense of Abidjan, possibly because land is easier to obtain. Remarkably, Yousry and Aboul Atta (chap. 4 in this volume) report that, in Egypt, rural population growth rates have exceeded urban rates for the past few years, although this appears to be due to spill-over effects just beyond urban boundaries.

Significant adaptations have therefore occurred in the nature of rural-urban and urban-rural migration in Africa. These changes have often not been recognized by the World Bank or IMF or even some academic observers, although they often began before structural adjustment policies were imposed. Given that "migration and migrants are a sensitive barometer of the totality of change taking place" (Gutkind, 1974, p. 54), this has serious implications. The IMF policies towards urban areas are partly predicated on the assumption that these areas are privileged and that continued rapid growth indicates the persistence of urban bias in government policies; yet the evidence is that the supposed privileged starting point had long been eroded and that urban growth was greatly exaggerated.

It should be noted that generalizations about migration trends for the whole of Africa will rarely be valid for all countries. Factors affecting the sorts of adaptations detailed above include the severity and timing of economic decline and the nature and timing of structural adjustment. As noted by Simon in chapter 3, the impact of war and/or drought is another vital variable, because these tend to increase rural-urban migration dramatically,¹⁶ overwhelming the negative impact of structural adjustment policies. On the other hand, the wars in Somalia and Rwanda have led to an emptying of urban centres, and it is as yet too early to know what will happen in these countries if, and when, political stability is re-established. As emphasized by Aina in chapter 12 in this volume, some countries in Africa have experienced a "simultaneous collapse of state structures and the disintegration of social order." In such circumstances the impact on urban processes is bound to be locally specific. A number of factors related to land have also been discussed above. These include the availability of rural land for migrants; the commercialization of land; and the nature of urban settlement patterns and land tenure and its implications for access to land for housing and agriculture. Of critical significance as well is the nature of government policy towards the rural sector.

Having detailed the importance of rural-urban linkages and their changing nature, this chapter now turns to consideration of strategies adopted by those who must, or choose to, stay in town.

Informal sector and second jobs

The informal sector is not a new phenomenon in African cities. Its importance for urban employment was highlighted in the 1970s (see O'Connor, 1983), and its more recent growth is detailed in this volume in chapter 10 by Rogerson (see also chap. 6 on Lagos by Abiodun). The huge number of jobs created in this sector often involve self-employment and include a wide range of activities from food vending, shoe shining, and hair braiding, to repair and maintenance services, "illegal" transport services, and small-scale metal fabrication. Because open unemployment is difficult to maintain for long periods of time in the African urban context, in the absence of any institutionalized welfare systems, the jobs developed in the informal sector, albeit usually low paid, have played a central role in accommodating rapid labour force growth.

There is no doubt that the informal sector has become even more important in the era of debt crisis and structural adjustment because of the massive and precipitate decline in formal wages and the significant absolute fall in formal sector jobs in many African cities since the 1980s (see Rogerson, chap. 10 in this volume). Large-scale redundancies have resulted largely from structural adjustment policies: in the public sector, donor conditionality often insists on contracting the civil service payroll in order to cut budget deficits, and, in the private sector, trade liberalization has made many previously protected enterprises uncompetitive, while reductions in demand (e.g. for new clothes) owing to increasing poverty have forced others to close or retrench.

Prior to the introduction of trade liberalization policies, controls on foreign exchange, imports, and prices of many domestically produced products created illegal job opportunities in parallel markets. Thousands of urban residents became involved in trading scarce goods at prices well above the official and this was another important coping strategy in the late 1970s and much of the 1980s (see, for example, Maliyamkono and Bagachwa, 1990, for Tanzania; and McGaffey, 1991, and Piermay, chap. 7 in this volume, for Zaire). This trade was often very lucrative, and many formal sector workers ameliorated their impossible "wages puzzle" because they had access to officially priced goods (or stole them) and sold them at the "market-determined" price. Thus, in Kampala at the end of the 1980s, the explanation for how people got enough money to survive was "each eats from his place of work," meaning illicit trade or "magendo" (Mutasa, 1989).

Trade liberalization and currency devaluation in many African countries have now significantly reduced the opportunities in parallel markets. However, many formal sector workers retain their jobs, even though it might seem that they would be better off concentrating on their informal sector activities. There are a number of reasons for this: keeping a foot in the door of the formal sector may be a good idea in case the economy picks up; the job may give the worker access to important, wealthier individuals; and many jobs give access to services that may be difficult to obtain "on the outside." This gives workers the opportunity to sell access to the service to others (e.g. telephones) or to use it to further a second, informal sector job (Maliyamkono and Bagachwa, 1990).

The rate of growth in the informal sector increased in the 1980s as the formal economy declined. For example, research in Dar es Salaam by Tripp in 1987/88 found that 64 per cent of surveyed informal enterprises had begun in the previous five years, compared with only 15 per cent in the five years before that (Tripp, 1988, cited in Maliyamkono and Bagachwa, 1990). A separate survey of the informal sector in Dar es Salaam and Arusha confirmed this trend: four-fifths and three-quarters, respectively, of the informal sector activities sampled had been established between 1980 and 1987 (Maliyamkono and Bagachwa, 1990).¹⁷ Furthermore, Tripp found evidence that the low level of formal wages was increasingly driving workers out of the formal and into the informal sector (including full-time urban agriculture): only 30 per cent of respondents who had transferred between sectors had done so during the pre-crisis period of the 1970s, whereas 70 per cent had moved after 1980.

That the necessity of taking on second jobs is not confined to the poorest is also evident from Tanzania. The ratio between the take-home pay of a worker on the minimum wage and a top salary-earner fell from 1:19 in 1962 to 1:6 by 1985. Between 1980 and 1985, top-level public sector employees' earnings fell 66 per cent in real terms. By 1987 it was estimated that most senior executives in large parastatals could cover the necessary costs of maintaining their family in town for only two weeks of the month. Thus top-level "moonlighting" became necessary. Amongst poorer urban workers the second-job syndrome was rife: a survey in 1985/86 of poor, lowermiddle-income and middle-income wage-workers in Dar found that 54 per cent had secondary activities (Maliyamkono and Bagachwa, 1990). Another survey of Dar household heads in 1986/87, in which only 17 per cent of respondents were in a low-income bracket and 56 per cent were owner-occupiers, found that 72 per cent had supplementary incomes. Furthermore, half the households had more than one person working, and 70 per cent of these jobs were in the informal sector (Kulaba, 1989).

A rather different form of second-job strategy is to send a household member abroad and use their remittances as a supplement to urban incomes. El Sammani et al. (1989) have documented the increasing importance of this strategy for poor households in unplanned areas in Khartoum during the 1980s. In 1985, remittance contributions from workers who had migrated to Arab countries were vital to maintaining urban families, making up 67 per cent, 71 per cent, and 100 per cent of household budgets in three different neighbourhoods, respectively. Such remittances are also vital to an understanding of the urban economy in Somalia (Jamal and Weeks, 1993). As an urban strategy this option seems to be largely confined to north-eastern African countries with access to Arab job markets, including Egypt (see chap. 4 in this volume).

Growing your own food: The expansion of urban agriculture

Food looms large in the household budgets of the urban poor, and came to dominate budgets in the 1980s. Strategies for reducing the impact of food purchases on meagre cash resources involve obtaining food from sources other than the market. As already seen, rural linkages play an important part in this, but growing one's own food within the urban area (or just outside the boundary), on plots sufficiently accessible to allow the "farmer" to remain an urban resident, has played a major part in survival strategies. The scale of the operation ranges from intensive cultivation of garden plots around one's home (a common enough practice before the onset of serious economic decline), to small plots of staple food crops on undeveloped urban land or pert-urban land, to quite large-scale, formal sector enterprises on the urban periphery. Not only crops but livestock raising may be involved.

As with the informal sector, urban agriculture was part of the African urban scene before the onset of serious economic decline. The difference today is that far more households are involved, far more land is involved, and the role of the food produced is much more critical to household survival. Writing in the mid-1970s, Gutkind (1974, p. 71) commented that the urban population was dependent on food supplies from the rural areas, and "few urban Africans are able or willing to grow their food needs.', Shortly after this, the situation was to change quite markedly - the remarkable expansion of urban agriculture in more recent years is documented in this volume by Rogerson (chap. 10). This section therefore focuses on identifying some factors that influence the extent to which urban agriculture can contribute to household coping strategies in the different types of cities and regions in Africa.

One factor is the density of settlement within the city. Very densely settled cities, with closely packed plots even in unplanned areas and very little open space (e.g. Cairo, some South African conurbations, the older areas of Lagos), clearly do not offer much chance for the poor to grow food. A second factor is the size of the city, especially if densely settled. In large cities, access to land on the edges of the city is more difficult for many residents because of the sheer distances involved. Piermay (chap. 7 in this volume) identifies this as a factor hindering the development of urban agriculture within Kinshasa, for instance. Thirdly, there is the issue of land tenure, both within and on the periphery of the urban area. If land is held on a freehold or leasehold basis and the land market and registration are functioning (e.g. Harare, most South African cities), then illegal farming on vacant land may receive short shrift and the farmers be driven off. A fourth factor is the attitude of city authorities and local politicians: if they are unsympathetic to the needs of the poor and unresponsive to democratic demands, then again urban agriculture may be vulnerable.

In fact, in many African towns and cities the conditions for urban agriculture are quite positive. Small, spread-out settlements with many open areas are typical of much of Africa's urban network, giving relatively easy access to cultivable land. Even many of the larger cities incorporate areas that are not densely settled, and of course, by Asian or Latin American standards, even most African capitals are still fairly small. Access to land within the city boundaries may be through traditional routes, even when there are, officially, formal and bureaucratic procedures (e.g. Kinshasa - see Mbuyi, 1989; Piermay, chap. 7, and Rakodi, chap. 11, in this volume). This may be because the official system is simply too corrupt or overloaded to function. City land may largely be under the control of one ethnic group, giving it priority and security in obtaining plots for urban agriculture. Thus a marked increase in growing food within Kampala's boundaries was particularly a Baganda phenomenon, because they had control over the land and many in-migrants found it difficult to get access to it (Jamal and Weeks, 1993). Land just outside the city boundaries, and quite easily accessible in many towns, is very often under communal tenure. Depending on the density of the rural population, it may be quite simple to obtain permission to use a plot for cultivation at low cost. As Rogerson notes in chapter 10, city planners are rarely keen on unplanned land use. However, local party officials may be far more prepared to lobby for the rights of constituents to use land for food, which greatly increases the security felt by urban farmers. This has been noted by Kulaba (1989) for urban Tanzania, where urban agriculture was encouraged by the ruling party after a 1974 nationwide drought and food shortage, and for urban Zambia when Kaunda's UNIP party was in power (Mulenga, 1991). Urban agriculture not only provides food for cultivating families but may also be a source of supplementary income when surpluses are sold. In this way this strategy contributes to the informal sector coping mechanisms already discussed.

Servicing and shelter strategies

The collapse of urban services in African cities has been thoroughly documented in Stren and White (1989), and further case-study material is provided in Touré and Fadayomi (1992). In most urban areas, particularly in anglophone countries, urban service infrastructure such as urban transport, electricity, water, and sewerage was run by parastatals. Health and education were primarily government functions. The causes of deteriorating urban services and the effects of structural adjustment policies on the service and welfare sectors have been discussed above. Here the scale of the problem will be illustrated with some selected examples, and the types of service adaptations that have emerged to fill the service "gap" discussed.

Many of the unplanned housing areas that have been built since the beginning of the 1980s will have had virtually no services of any type installed - even squatter upgrading programmes, which were quite successful in the late 1970s and early 1980s in a number of countries (e.g. Zambia, Tanzania), are too much for many governments' current austerity programmes. Thus there may be not much difference between the service environment of unplanned areas in Abidjan, Dakar, Kinshasa, or Dar es Salaam and that of rural villages.

In 1980 it was estimated that nearly one-half of poor urban households had no services of any kind (e.g. water, electricity, sanitation, refuse disposal) in Côte d'Ivoire's towns; and a mere 7.5 per cent had collective electricity connections (Touré et al., 1992). Water supplies in Côte d'Ivoire's urban areas are provided by a private company, SODECI. According to the World Bank (1983, cited in Stren, 1989, p. 41) this is an efficient supplier because it operates on a full cost-recovery basis and can charge low rates for smaller users, so the poor can afford water connections. Unfortunately this flies directly in the face of the evidence: between 1977 and 1983 the levels of water connections in Abidjan fell from 57 per cent to 47 per cent of the population. This was because the growing population was mainly settling in unplanned settlements, where the level of connections was, "in principle," zero. Instead, water is provided by vendors at about five times the unit price paid by those with connections. During the 1970s, as described by Dubresson in chapter 8 in this volume, there was far more state-planned low-income housing, and the state in fact subsidized water connections to homes in poorer areas. The dramatic fall-off in connections in the 1980s reflected the shortage of government funds available to continue this programme (Stern, 1989). It is evident from Abidjan that a purely cost-recovery, market approach to water supplies cannot help the urban poor.

The capital city of the Central African Republic, Bangui, has no public highway maintenance or garbage collection system at all. Most of the more recent peripheral settlements have virtually no services of any kind, and water supply and electricity networks meet only 20 per cent of demand (Faustin and Takam, 1992). Similarly the southern peripheral extensions of Kinshasa, which have housed most recent population growth, have few services and no electricity (Mbuyi, 1989). In Khartoum, conditions are even worse: many of the new squatter areas do not even have pit latrines (El Sammani et al., 1989). A 1986/87 survey of Tanzanian socio-economic and housing conditions found that in Dar es Salaam there was an average of 15 people per house, and 89 per cent of houses had unimproved pit latrines as their sole sanitation, of which 21 per cent overflowed during the rains. Only 17 per cent of the household heads of these dwelling units (which also housed large numbers of tenant families) were regarded as low income (Kulaba, 1989). The cost of lack of public transport services is also a problem for the urban poor.

The response to the shortage or absence of public or large-scale private provision of services often involves self-provision, usually on an individual but sometimes on a collective basis. In terms of shelter, many of the poor erect houses in unplanned areas and build some sort of latrine. Very often some sort of title to the land may be obtained from a local chief or local family, although its legality is sometimes questionable (Mbuyi, 1989; Tipple, 1994). Many more poor households are then housed in such dwellings as tenants - for the house-owner, who often lives on the premises, this may represent a desperately needed extra source of income. One response to the urban economic crisis has been to adapt existing housing to accommodate tenants by converting kitchens to bedrooms, or covering courtyards to provide an extra room, and adding outhouses, as, for example, in Abidjan in the 1980s (Stern, 1989). Various forms of outhouses for rental occupation on formal, planned low-income housing plots are very common in southern African countries such as Malawi, Zimbabwe, and South Africa, where they are often known as "backyard shacks" (e.g. see Beavon on Johannesburg, chap. 5 in this volume). Despite the lamentations of urban planning authorities, it is now relatively rare for people to be evicted, given the absence of any alternative. Collective political action at a local level often wards off any such threats; local MPs or councillors may find their entire constituency consists of people in unplanned housing. Thus in Dar es Salaam, or in most South African cities, the issue of removing informal settlements is usually too politically hot for administrations to contemplate. However in Harare, Zimbabwe, squatters are never allowed to remain in any one place for long, and the incidence of unplanned settlements is still very minor, involving hundreds, rather than hundreds of thousands, of people, as in most other African cities (Potts with Mutambirwa, 1991). Recent reports of evictions also come from Dakar, Senegal (Sy et al., 1992), and Khartoum, where the ability to organize politically is a key factor in preventing settlement demolition (El Sammani et al., 1989).

As indicated by the Harare example, demand for housing cannot always be met by "unplanned" solutions within the city. Strong freehold land markets, high land prices, and high-density occupation are factors that will mitigate against such solutions. In these circumstances, alternative solutions may operate. In Cairo, for example, there has been significant "decentralization" to agricultural land outside the urban boundary. Very high land prices in the city (partly driven up by flows of remittances from temporary migrants to Middle Eastern countries - Yousry and Aboul Atta, chap. 4 in this volume) have driven many people out to become commuters. Strictly speaking it is illegal to build housing on the agriculturally valuable delta land, but the sheer demand for accommodation has overwhelmed planning controls there (El Shakhs, personal communication, 1994). According to Abiodun (chap. 6 in this volume), land scarcity within Lagos metropolis has also driven up prices there to a remarkable degree, and illegal land occupation is now limited because land development controls are being enforced.

Illegal hook-ups to electricity may be one strategy for obtaining power, but so many of the residential areas occupied by the poor have no nearby power sources that this is a much less common strategy than in Latin America or Asia. In many African cities, private minibuses and taxis have become a vital part of the transport system - for example, matatus in Nairobi, dala-dalas in Dar es Salaam (Lee-Smith, 1989; Kulaba, 1989), emergency taxis in Harare, or "Zola Budds" in South African cities. The urban authorities were very often initially reluctant to allow these facilities, but the incapacity of public provision made them necessary, as in Tanzania, where they were permitted after 1983 (Stern, 1989). However, they always cost more than the public buses and are frequently unaffordable to many of the poor. Very early starts and a long walk to work are thus the usual "strategy." The private taxis also often have a notoriously dangerous reputation because they are driven fast in order to increase narrow profit margins or driver commissions - the South African Zola Budds are so-called because of their speedy driving.

A service that it is hard for the poor to provide for themselves is water. In some areas, digging wells is possible, although rarely feasible for every house. Once an area is established, the authorities find it hard not to provide at least some standpipes. However, the adequacy of this solution varies enormously between settlements and depends on the distribution and number of standpipes in relation to the geography and population of the settlement and on the reliability of the water supply. Where there are few water points and/or they are far away, many people have to rely on private water vendors because the time and energy involved in collecting and queuing for water are not feasible. However, as previously discussed, water thus purchased is often expensive in relation to public supplies (see, for example, Stren, 1989, on water supplies in Dar es Salaam, Abidjan, and Nouakchott in Mauritania; El Sammani et al., 1989, on Khartoum; Mbuyi, 1989, on Kinshasa; Jaglin, 1994, on Ouagadougou).

Refuse collection for the poor is rarely provided. Periodic town clean-ups encouraged by government exhortations sometimes occur, as in Tanzania and Nigeria. Recycling waste is common, adding to informal income-generating activities.

For many wealthier households throughout Africa the solution to the service crisis is the private sector, and private refuse collection, septic tank emptying, electricity generators, and even, on occasions and for the rich, telecommunications are all well established (Stern, 1989; Uduku, 1994). A significant increase in the provision of private educational facilities has also occurred as government provision has deteriorated (e.g. Uduku, 1994), but the costs involved may put this option beyond the reach of the poor. Community self-help projects to provide schools are often popular, however, reflecting the great demand for education. Such community projects are very common in Khartoum, for example (El Sammani et al., 1989), and in Ouagadougou water sales at public taps enabled revolutionary committees to finance schools and dispensaries (Jaglin, 1994).

Urban people have, therefore, adopted a range of strategies to meet those service shortages that have the highest priorities. However, it must be said that these are rarely very satisfactory; in particular, many of the practices have negative health implications. Poor sanitation is a major issue (Potts, 1994); expensive and distant water discourages water consumption, with negative hygiene implications; walking to work may adversely affect already undernourished workers; and unregulated transport is often dangerous, as are illegal electricity connections.

Continue

Contents - Previous - Next