The role of rural urban migration in the growth of informal settlements in South Africa

Transcription

1 iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii_ URBAN a SATa 0 N Derik Gelderblom The role of rural urban migration in the growth of informal settlements in South Africa This article looks at trends in migration and how these affect urbanisation. In recent years urbanisation has been transforming the face of South Africa. While the numbers of people moving to town at present may be lower than previously anticipated, the process is still taking place fast enough to place considerable stress on the capacity of our institutions to cope with this influx. In addition, there are a great number of potential rural-urban migrants who may migrate in the near future. Social scientists typically distinguish between urbanisation Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2011) Derik Gelderblom is a senior lecturer in Sociology at Unisa. He obtained BA, BA Hons (cum laude) and MA (cum laude) degrees in Sociology from the University of Stellenbosch. He is currently registered for a DPhil degree at the University of Durban Westville, with the topic 'Migration and social inequality in South Africa'. He is the author and co-author of a number of books and articles in the field of urbanisation in South Africa. High density of shacks situated in the floodplain in Alexandra Township, Johannesburg and migration. Urbanisation refers to an increase in the proportion of the population of a country living in urban areas. If the proportion increases, the population of that country is becoming more urbanised. There are two reasons why a population may become more urbanised. It can be due to the natural increase of the population in the urban areas being higher than that of the rural areas, or it can be due to the migration of people from the rural to the urban areas. Natural increase occurs if the births in a population are higher than the number of deaths. It is an important component of urban growth, but for a number of reasons that we shall not go into here it is usually not very important as a factor in increasing the proportion of the population living in urban as opposed to rural areas. A much more important factor in this respect is the migration of people from the rural to the urban areas. It is consequently to this phenomenon that we shall devote most of the rest of this article.,siviele INGENIEURSWESE. SEPTEMBER

2 Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2011) Levels of urbanisation of the South African population According to official census figures, not much urbanisation has taken place in South Africa over the last decade. In the period between 1980 and 1993, the level of urbanisation increased by a mere two per cent from 46 per cent of the population to 48 per cent of the population (see Table 1). However, owing to a number of factors, censuses in South Africa have never produced very reliable information. The last census, that of 1991, was particularly problematic in this regard. It occurred at a time when there was a climate of political violence (which made access to many areas difficult) and distrust of the census authorities. The AI\JC also called on its followers to boycott the census. Table 1: Levels of urbanisation in South Africa % urban % functionally urban Source: Calltz and Grove, 1991 In an effort to improve the accuracy of the census figures the Bureau of Market Research at Unisa was commissioned to verify the figures. Their findings were that the census was most complete in the urban areas, but that substantial undercounting of the African population took place, even in urban areas. As a result the African population of the country was adjusted upwards by 20 per cent, from 18 million people to 22 million (Hindson and Parekh, 1993: 6). Such a large adjustment does not increase one's confidence in the census results and it may well be that the urbanisation levels of the census are incorrect. Another reason for doubting the accuracy of the 1991 urbanisation levels is that the Department of Statistics and the HSRC (which was responsible for the sample surveys that formed an important part of the census) apparently did not use the same criteria to judge whether a particular area is an urban area or not (Hindson and Parekh 1993:7). The census has in addition a very strict (and very unrealistic) definition of an urban area. Only those areas with local authorities have qualified in the past to be enumerated as 'Figures suggest that the scrapping of influx control in 1986, which seemingly was the major obstacle to migration to urban areas outside the homelands, did not have much effect on the rate of migration to those urban areas' urban areas. There are, however, many dense settlements, some with populations in the hundreds of thousands, that do not have local authorities. These could, with a few qualifications, be regarded as urban areas. Most of these are in the former homelands. In order to account for these, the Development Bank of Southern Africa (DBSA) (Calitz and Grove, 1991) has developed a new category of urbanisation. They call this 'functionally urbanised'. Functionally urbanised places are those that are regarded by the census as urbanised, as well as peri-urban settlements that are not, and other settlements with more than inhabitants that are similarly excluded by the census. The peri-urban settlements that are included by this mea- 38 SIVIELE INGENIEURSWESE, SEPTEMBER 1995 surement are of two types: a) informal settlements located outside a former homeland but on the perimeter of a metropole and b) informal settlements that are located in a former homeland and that are on the perimeter of a metropole. The other type of settlement included here is c) 'closer settlements' in the former homelands, ie dense settlements that are relatively large (more than inhabitants) and that are located more deeply in the homelands. If we include these, the level of urbanisation of the South African population has increased from 60 per cent in 1980 to 66 per cent in 1993, according to the Bank (see Table 1). The DBSA figures do help one to overcome the problems inherent in the census definition of an urban area, but still have the drawback that they derived their basic information from the census statistics, which is not very reliable. If we look at the DBSA's disaggregated figures (for the various provinces), it becomes clear that almost all of the difference between the officially urbanised and the functionally urbanised population is accounted for by settlements in the former homelands (types b and c) and very little by informal settlements outside a former homeland. It thus seems as if most recent rural-urban migration was directed to areas within the former homelands and very little to the informal settlements (such as Khayelitsha in Cape Town and Orange Farm south of Johannesburg) within our metropolitan areas that we have seen growing in the recent past. If their data are correct, much of the growth of these settlements must then come from people already living in the urban areas, and now merely moving from overcrowded townships to these settlements. Their figures also suggest that the scrapping of influx control in 1986, which seemingly was the major obstacle to migration to urban areas outside the homelands, did not have much effect on the rate of migration to those urban areas. This conclusion sounds puzzling and it is therefore necessary to explore the process of rural-urban migration in South Africa in more detail. The migration histories of the inhabitants of informal settlements: Recent surveys A number of surveys have recently been carried out that studied the migratory behaviour of residents of informal settlements on the peripheries of metropoli tan areas (Crankshaw, nd, and Sapire, 1992, in Gauteng; Cross et ai, 1992a, 1992b, 1993 and 1994, as well as Hindson and Byerley, 1993a and 1993b, in the Durban functional region [DFRj; Seekings et ai, 1990, in Khayelitsha). Although there were differences between the informal settlements, which will be discussed later, it was generally found that, although most of the respondents were born in a rural area and are therefore rural-urban migrants, most of those who moved into the various urban areas had in fact done so before 1986 (when influx control was abolished). The authors cited above also established that most of the respondents did not move directly from a rural area into these informal settlements, but approached them indirectly. It seems that most of the respondents first lived in another part of the metropoles as lodgers, hostel residents and backyard shack dwellers before moving into the informal settlements. They then generally conclude that intra-urban migration is a more important phenomenon among their respondents than rural-urban migration (see for example Hindson and Byerley, 1993a: 17, and Cross et ai, 1994: 85).

3 Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2011) The finding that a large part of the growth of informal settlements in metropolitan areas derives from intra-urban migration from other parts of the metropole rather than rural-urban migration is strengthened by the declining household sizes observed over time, as well as the relatively small household size in informal settlements. Cross et al (1992b: 16, 32) have observed for example that during the 1980s household sizes dropped markedly for Africans in the DFR, with the result that the average household size in 1991 was just 7, after falling from over 10 in Even with a stable population, the number of households would have increased markedly. If one furthermore assumes that there is one dwelling for each household, it is clear that merely to house the existing population, the number of dwellings had to increase dramatically. This would partly Tolweni in KwaZulu-Natal. Centre of growth in a rural area. explain the mushrooming informal settlements observed on the peripherindicated by the following example. The peri-urban settleies of metropolitan regions. The lifting of influx control in ments to the south of Johannesburg tend to attract exjune 1986 and the new urbanisation policy that accompafarmworkers, especially farmworkers born on farms within nied it were indeed associated with the growth of these settlethe PWV area, rather than people who have arrived from the ments, not only because of new in-migration, but also beformer homelands (Crankshaw, nd: 6). It seems as if the locause it made land invasions and the founding of new settlecation of these settlements to the south of Johannesburg ments easier. makes them more likely to receive migrants who originate in Household sizes in informal settlements are also smaller rural areas further to the south (such as the Free State) rather than in other African urban areas. Cross et al (1992b: 16) than from the former homelands to the north of found an average household size of 5 in the informal settlejohannesburg. Because ex-farmworkers are so poor, they ments of Inanda, compared with the DFR average of 7. A are in addition less likely to find accommodation in formal similar finding was made by Seekings et al (1990: 15) for townships such as Soweto and consequently more likely to Khayelitsha compared with the Cape Peninsula. In a comgravitate towards peri-urban settlements where accommoparison between a part of the formal township of Umlazi and dation is cheaper (Crankshaw et ai, 1992). an informal settlement close by, it was observed that the latter had an average household size of 5,9 while the household size of the formal township was 6,6 (Hindson and Byerley, 1993a - household sizes calculated on the basis of information supplied in Table 1 on p 9 of their report). Another observation is that household members in informal settlements are typically younger than those in formal townships (Hindson and Byerley, 1993a: 13). This finding, together with information that household sizes are decreasing and that the smallest households are observed in informal settlements, is consistent with the observation that the growth of informal settlements is associated with the fission of larger households in the formal townships and the migration of the newly created households to informal settlements (Hindson and McCarthy, 1994: 8). These new households would then consist of parts of previously extended families (the middle generation in three generation households and their children). lodgers and backyard shack dwellers previously forming part of the bigger formal township households. Besides the general trends I have sketched above, there are also differences between individual informal settlements in terms of the migration histories of their inhabitants. This indicates that local level factors are also important in determining the form of the migration process (Sapire, 1992: 677; Cross et ai, 1992b: 4). In Durban, for example, the younger settlements are more likely to have residents who recently (within the last five years) migrated to the DFR (Cross et al; 1994: 86). The geographical location of a settlement, and the social networks built up around this, can also be a factor in determining the kind of migrants that it will receive. This is The migration histories of the inhabitants of informal settlements: A problem of interpretation Up to now I have merely reported the findings of the relevant surveys. There is, however, a methodological problem in the design of most, if not all, of these surveys, which makes it difficult to interpret the information they provide and which casts some doubt on their findings. This problem has to do with their strategy of deducing the migration histories of the whole household from that of the respondent (the respondent is the person in the household whom the researcher interviews). This may be an acceptable strategy in the case of middle-class people, but in the case of poor African people it is definitely not acceptable. In the case of the former, the whole household generally follows the household head if he or she migrates. In the case of the latter, and in terms of the migrant labour system, only the breadwinner would typically migrate at first. This trend was partly the result of the influx control system, which made the migration of the dependants of breadwinners illegal, but it was also a rational strategy of risk minimisation for the household whereby they would attempt to maintain access to land in a former homeland area to serve as something to fall back on if things go wrong in the city for the breadwinner (unemployment, etc). The dependants of the breadwinner would then typically remain in the rural area in order to maintain the rural base (see the discussion of this issue in Gelderblom and Kok, 1994: 37-47,52-57). SIVIElE INGENIEURSWESE, SEPTEMBER

4 Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2011) Since most of the breadwinners were men, the effect of this was to create a very large gender imbalance between the homelands and the rest of South Africa, with most African men present in the common area and most African women in the homelands (Gelderblom and Kok, 1994: ). In 1985, in fact, two thirds of all African men in the economically active age categories were living in the common area of South Africa, while only one third of all African women in those age categories were doing so. Male respondents in a survey are therefore far more likely to answer that they have been present in an urban area for a long time than female respondents. They are also unlikely to give a proper reflection of the migration histories of their children. This contention is borne out by the results obtained by Seekings et al (1990). In 98 per cent of the cases where a married couple did not move to the Western Cape together, the husband came first (Seekings et ai, 1990: 45). In half of those cases the wife followed a few months later, while in the other half the wife only followed years later or has not yet arrived in the Western Cape. It furthermore seems that it happened often that husband and wife did not arrive together - 86 per cent of respondents moved to the Western Cape on their own, and only eight per cent of respondents arrived in the Western Cape with their spouse (if they had one). Women were also far more likely to have arrived in Khayelitsha directly from an area outside the Western Cape (Seekings et ai, 1990: 35). In addition, 39 per cent of the households in their survey had children living outside the Western Cape, which is an indication that the migration histories of the children are different from those of the adult respondents. 18th Avenue in Alexandra, Johannesburg. A vast influx of people. Given these differences, it is clear that the sex or the age of the respondent will affect the impression that the researcher gets of the migration histories of the household, if the migration history of the respondent is generalised to the whole household. Since it is unlikely that the gender and age composition of the sample of respondents is an accurate reflection of the gender and age composition of the households, it is clear that there is a built-in bias in most of the surveys reported on above. Unfortunately, researchers do not often give information about the composition of their respondent sample, with the result that it is not clear in which direction the bias in their 40 SIVIELE INGENIEURSWESE. SEPTEMBER 1995 research lies. In the two cases where we are supplied with this information, the respondent sample is more likely to contain older female members of the household (See kings et ai, 1990; Cross et ai, 1992b). The predominance of females would have reduced the time households have been said to live in town, while the age factor would have increased this time. What the balance of these two opposing biases would be is difficult to say. After acknowledging that there are problems in generalising from the respondent sample to the household sample, Cross et al (1992b: 27) defend their decision to do so by saying that 'only couples with children or, more rarely, single mothers with children will usually take up separate accommodation as households in their own right'. This might be true when the household eventually settles in the informal settlement they are studying (Inanda in Durban in this case). but it is not necessarily an accurate reflection of what happens when the breadwinner first arrives in town. In many cases the breadwinner arrives first. as we saw above in our reference to the work of See kings et al (1990). While they are on their own, breadwinners are more likely to live as lodgers or backyard shack dwellers. It is only after their families start to arrive in town that they are likely to look for a site in an informal settlement. A significant part of the intra-urban migration of household heads from overcrowded backyard shacks, hostels and lodgings in townships might thus be accompanied by the rural-urban migration of other members of their household. This would make sense if one considers that an important reason for moving provided by respondents is that they were looking for a bigger place to stay for their family (see for example Cross et ai, 1993: 63). Building a shack might make it possible for them to house the rest of their family. Unfortunately, owing to their failure to distinguish between the migration histories of the respondents and those of the other household members, most of these surveys do not allow us to distinguish between this rural-urban migration and the intra-urban migration of the household head. The rate of rural-urban migration in South Africa Given the problems with the census figures and the limitations of the surveys, it might seem as if we are no more nearer to an understanding of how much rural-urban migration is taking place at present. This would be an overstatement of the problems of interpretation we face, however. The information regarding the declining household size over time and the argument that this in itself would necessitate the growth of new settlements seems to be convincing. This would strengthen the observation that a large part of the inhabitants of informal settlements in our major urban areas are not new arrivals. Owing to the limitations of the surveys, we might be unsure about the exact proportions of newcomers and more established migrants we are dealing with, but the general observation seems to be accurate. It should also be noted that in all of these surveys a significant amount of rural-urban migration is indeed picked up. In their survey of Khayelitsha, Seekings et al (1990: 40) found that only 11 per cent of their respondents were born some-

5 Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2011) where within the Western Cape. The other 89 per cent were born somewhere else and migrated to the Western Cape at some stage of their lives. Forty per cent arrived before 1980, 35 per cent arrived between 1980 and 1985 and only 15 per cent arrived from 1986 to the middle of 1988, that is after the scrapping of influx control (it should be noted that the proportions of earlier migrants were probably underestimated, according to them). What this points to is not a flood of new rural-urban migrants after 1986 and nothing before, but a rate of migration that has been slowly increasing over time - picking up momentum since the beginning of 1980 and then increasing again after the lifting of influx control in This trend is confirmed by Cross et al (1994: 88) in their surveys of the DFR. It should not be too surprising to find that the rate of ruralurban migration already started to increase before the lifting of influx control. The intensity with which influx control was enforced varied over time, and there was a period between 1979 and 1982, coinciding with the beginning of the P W Botha administration, when influx control was not enforced as strictly. This is demonstrated very clearly in Table 2. Between 1978 and 1980 the number of people arrested for influx control offences each year decreased from to After 1980 it started to pick up again, but it did not approach 1978 levels until 1983, when it peaked at The number subsequently declined again. During the early eighties we also had the Rikhoto judgment, which made it easier for some categories of migrants to get permanent residence rights, and the so-called Pretoria experiment, during which the in- Table 2: Number of people arrested for influx control offences Year Number of arrests Source: South African Institute of Race Relations ( ) flux control regulations were relaxed for a period. It should also be remembered that the level of resistance to the state increased during 1984 and that this made the townships, and with it influx control, very much ungovernable. We started our discussion with a consideration of the census figures on urbanisation (which were rejected as unrealistic) and the Development Bank's figures for the functionally urbanised population. It was then stated that the Bank's figures indicate that most urbanisation was channelled to informal settlements within a homeland boundary, and not to informal settlements outside of them. Since most of the surveys studied here were done in the latter kinds of settlements (with the exception of those in Umlazi in the DFR, which were in KwaZulu), and since most of these surveys do indeed report some level of in-migration, the Bank's figures, being themselves mostly based on the census figures, probably underestimate the amount of rural-urban migration to those settlements. 'With changes to the tenure system, and with the start of the house-building programme of the RDP, these circumstances might change and we might yet see the delayed stream of migrants arrive in our cities' Reasons why rural-urban migration is not taking place at a faster pace The question arises why rural-urban migration is apparently taking place at a slower rate than expected in the aftermath of the scrapping of influx control. Most of our theories of migration explain migration only in terms of push and pull forces. All of these forces work in the direction of the city, and the question now arises why so few people out of the huge potential pool of migrants actually migrate. In order to account for this, I would argue, we also need to take into account urban repel factors and rural retard factors. Among the former I would include factors such as violence in the urban areas, lack of housing and a number of factors leading to insecurity among migrants, such as insecure tenure and lack of welfare services. All of these would motivate migrants or potential migrants to keep a rural base (if it is at all possible) while they are working in the urban areas. With the communal tenure system in the former homelands it was possible to do just that, as long as a member of the household stayed behind to maintain the migrant's claim to land. With changes to the tenure system, and with the start of the house-building programme of the RDP, these circumstances might change, and we might yet see the delayed stream of migrants arrive in our cities. Cross et al (1992a: 54) allude to another important urban repel factor. According to them, it often happens that leadership structures (in the form of civics and youth organisations) in informal settlements take responsibility for controlling inmigration into the settlement. This is to ensure that longstanding residents are not swamped by a sudden influx of new migrants, thus losing control over the settlement and the resources (water, shelter, etc) it offers. The arrival of new migrants in fact is often associated with outbreaks of violence between old and new migrants (Cross et ai, 1992a: 51; Sapire, 1992: 687). In order to forestall this, strict control is thus often exercised over the arrival of new migrants. This kind of control, called community control by Cross et ai, is more often exercised in older settlements. In these settlements, new sites are often allocated only to relatives of existing inhabitants, or to people who have graduated from being lodgers (and whose acceptability to the community has been demonstrated). The control over site allocation is also a function of the paucity of new sites as the settlement matures and becomes built-up. New migrants, who do not have family connections to older residents, are therefore more likely to be restricted to newer settlements, where these controls are not exercised. Besides the age of the settlement, political factors, such as the kind of informal local government structures (warlords, civics, tribal authorities) and their effectiveness also determines how much of this control is going to be exercised. Besides these urban repel factors, there are also rural retard factors that may slow down the rate of rural-urban migration. These mostly revolve around the costs of migration, such as transport costs, costs associated with political boundaries, information costs and psychic costs (Gelderblom I SIVIELE INGENIEURSWESE. SEPTEMBER

6 and Kok, 1994: 61-62). Owing to these costs the poorest might well be prevented from migration because they are unable to finance it. The potential immobility of the poorest in the face of out-migration by the rest is a potential problem for policy formulation. It is also a reminder that the greatest need still exists in the rural areas and that any consideration of urbanisation cannot afford to ignore rural needs. References 1. Calitz, J M a nd Grove, M J A regional profile ofthe Southern African population and its urban and non urban distribution DBSA: Johannesburg. 2. Crankshaw, O. No date. Squatting, apartheid and urbanisation on the Southern Witwatersrand. 3. Crankshaw, 0 and Hart, T The roots of homelessness: Causes of squaning in the Vlakfontein senlement south of Johannesburg. South African Geographical Journal, 72: Crankshaw, 0, Heron, G and Hart, T The road to 'Egoli': Urbanisation histories from a Johannesburg squatter senlement. In: Smith, 0 (ed). The apartheid city and beyond. Urbanisation and social change in South Africa. Witwatersrand Univ Press: Johannesburg. 5. Cross, C, Bekker, S, Clarke, C and Wilson, C. 1992a.Search ing for stability: Residential migration and community control in Mariannhill. Univ of Natal: Durban Cross, C, Bekker, S, Clarke, C and Richards, R. 1992b. Moving on: Migration streams into and out of Inanda. Natal Town and Regional Planning Commission: Durban. 7. Cross, C, Bekker, S and Clarke, C Fresh starts: Migration streams in the southern informal settlements of the DFR. Natal Town and Regional Planning Commission: Durban. 8. Cross, C, Bekker, S and Clarke, C Migration into DFR informal senlements: An overview of trends. In: Hindson, o and McCarthy, J (eds). Here to stay Informal settlements in KwaZulu-Natal. Indicator Press: Durban. 9. Erasmus, J South Africa's nine provinces: A human development profile. DBSA: Johannesburg. 10. Gelderblom, 0 and Kok, P Urbanisation. South Africa's challenge. Vol 1: Dynamics. HSRC: Pretoria. 11. Hindson, 0 and Byerley, M. 1993a. Class and residential movement - report on a survey of households in the Umlazi Malakuzi area of Durban. Unpublished report. ISER, Univ of Durban-Westville: Durban. 12. Hindson, 0 and Byerley, M. 1993b. Class and residential movement - report on a survey of Clermont and KwaDabeka residents. Unpublished report. ISER, Univ of Durban Westville: Durban. 13. Hindson, 0 and McCarthy, J lnformal settlements in KwaZulu/Natal: Analysis and policy implications. Unpublished report. ISER, Univ of Durban Westville: Durban. 14. Hindson, 0 and Parekh, B Urbanisation, housing and poverty: A nationallitera(ure survey SALDRU: Cape Town. 15. Sapire, H Politics and protest in shack senlements of the Pretoria Witwatersrand-Vereeniging region, South Africa, Journal of Southern African Studies, 18(3): See kings, J, with Graaff, J and Joubert. P Survey of residential and migration histories of the shack areas of Khayelitsha. Research Unit for the Sociology of Develop ment: Stellenbosch. 17. South African Institute of Race Relations (SAIRR).A survey of race relations in South Africa. Johannesburg. Editions consulted: Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2011) We provide a comprehensive range of geotechnical services throughollt South Africa: Ground Anchors and Rock Anchors Soil Nails Gunite and Shotcrete Drilling and Grouting Well point Dewatering Underpinning Piling Pipejacking Thrustboring Drilling Investigations: - Diamond core drilling - Dutch probing - Dynamic Cone Penetrometer tests - Down-the hole shear vane tests Large Diameter Mud Boring - W(l/erwell screen inswll(l/ioll - Gravel packing - Pump testing 42 SIVIElE INGENIEURSWESE. SEPTEMBER 1995 F IR ROTHER GEOTECHNICAL ENGINEERING cc ~ ' (CK85/03/036/23) Stabilis(l/ion oislope at Partridge Point, Cape Peninsula Contact persons: Ian Fairbrother or Adrian Meerburg Address: 4 Estmil Close. Diep Rivier, Cape Town Telephone: (021) Fax: (021)