Migration in South and Southern Africa: Dynamics and Determinants¶
Pieter Kok; Derik Gelderblom; John O. Oucho; Johan van Zyl (2006). Migration in South and Southern Africa: Dynamics and Determinants. HSRC Press.
| Type | book |
| Scope | Wider apartheid literature |
| Method | mixed |
| Period studied | 1840s–2002 |
| Regions | South Africa, Southern Africa (Lesotho, Mozambique, Malawi, Botswana, Swaziland, Zimbabwe), sub-Saharan Africa |
| Themes | mining migrant labour, bantustans removals influx, demography urbanisation, labour markets wages, origins segregation, transition post apartheid, methods data sources, comparative theory, agriculture land |
| Citation key | kok2006 |
Argument¶
This edited HSRC volume analyses internal and cross-border migration in South and southern Africa, asking how much the end of apartheid changed patterns shaped by influx control, removals and homelands. Part I gives continental, regional, historical and comparative context; Part II uses the 2001–02 HSRC Migration Survey and a 2000 health survey to examine reasons for moving, networks and intentions; Part III proposes a synthetic migration model. For this book the historical chapters are most useful: they show contract mine migrancy and influx control as state- and capital-serving devices that kept African labour temporary and cheap. They also suggest that low post-1986 urbanisation and little change in mobility between 1975–80 and 1992–96 reflect apartheid's lasting spatial legacy.
Evidence and method¶
Secondary literature (Van der Horst, Jeeves, Lemon, Crush, Gelderblom & Kok 1994); UN, SAIRR and ILO statistics; Census '96 analysis (Kok et al. 2003); the 2000 SA Migration and Health Survey (2,331 individuals); the 2001–02 HSRC Migration Survey (3,618 households); the October Household Surveys; SAMP surveys; mine employment tables from Crush & James 1995.
Place in the debates¶
- Why the state and capital maintained migrant labour and influx control (see Influx control, pass laws and African urbanisation). Chapter 4 says oscillating migration kept mine labour cheap and externalised social costs, and influx control reflected employer demands for state regulation of the labour market plus the state's unwillingness to grant political rights. This is broadly a cheap-labour view.
- Effectiveness of apartheid urbanisation controls (influx control) (see Influx control, pass laws and African urbanisation). Wentzel & Tlabela say labour bureau controls were largely ignored and urbanisation escalated. Anderson says rural–urban migration was far below what it would otherwise have been until the 1990s. Together they say controls were only partly successful but still suppressed formal urban growth.
- Neoclassical economic migration theory versus structural and behavioural theories (see Influx control, pass laws and African urbanisation). Gelderblom and De Jong & Steinmetz criticise income-maximisation models; they emphasise immobilising structures, networks, norms and life satisfaction, and seek a synthesis.
- Did the end of influx control produce an urban surge? (see Influx control, pass laws and African urbanisation). Anderson and Wentzel & Tlabela find no immediate jump after 1986 and little change in aggregate mobility, partly because of displaced urbanisation and peri-urban commuting.
Findings¶
-
Compounds on the diamond fields created a controlled cheap labour force and began a system lasting over a century (p. 73)
This pattern of large-scale utilisation of cheap labour, and the control of labour by means of pass laws and the compounds, was the beginning of a system that dominated migrant labour in South Africa for more than a century
-
Oscillating migration kept mine labour cheap; the state externalised the social costs of foreign workers (p. 76)
Oscillating migration served the interests of the mining industry, as it meant that labour was cheap. Normally the employer only had to pay the worker enough for his upkeep and to have a small surplus in order to attract him back to the mines.
-
Foreign share of the mine workforce was at least 40% throughout the century and peaked in the early 1970s (p. 77)
Throughout the twentieth century at least 40 per cent of the South African mine workforce was non-South African. This figure peaked in the early 1970s at over 80 per cent
-
Mozambican mine recruitment fell sharply when South Africa adopted preferential employment of South Africans in the mid-1970s (p. 74)
-
Total mine employment declined sharply 1986–1992, with South African and foreign labour both shrinking (p. 58)
Total 477 397 476 127 456 625 432 885 396 968 354 453 324 441
-
Between 1936 and 1951 an estimated 280,000 Africans left white farms permanently despite immobilisation measures (p. 84)
Between 1936 and 1951 an estimated 280 000 Africans left the farms permanently (Nattrass, cited in Gelderblom & Kok 1994).
-
The Surplus People Project estimate of forced removals 1960–1983 (p. 88)
The Surplus People Project (SPP) claims that between 1960 and 1983 some 3.5 million people were forced to move out of metropolitan areas and smaller towns into peripheral areas or to remote rural homelands
-
Homelands offered families only a place to live; income came from transfers and remittances (p. 15)
Except for those areas where there was some factory work as a result of industrial decentralisation policies, most of these areas provided no jobs. The only thing they offered to the families arriving there in large numbers was a place to live.
-
Net rural–urban migration was below every comparator group until the 1990s, with no surge after influx control ended (p. 110)
According to Figure 5.7, the estimated rate of net rural–urban migration in South Africa was below that of any other grouping shown in Figure 5.7 until the 1990s.
-
Aggregate mobility barely changed despite abolition of influx control and the Group Areas Act (p. 92)
the overall level of migration between the 1970s and early 1990s declined from about 13 to 11 per cent
-
Lesotho's dependence on mine remittances in the 1980s–90s was very high (p. 77)
By 1994 about 40 per cent of the Basotho male labour force was employed in South Africa, and their remittances accounted for a third of the Lesotho GNP.
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| 2001–02 HSRC Migration Survey (national stratified cluster household survey, 3 618 completed households, 711 EAs) | author survey or fieldwork | 2001–02 | Main micro-level data for Chapters 8–11 on migration streams, motivations, networks and intentions; its field implementation and quality problems are described in Appendix A. |
| HSRC initial survey on causes of internal migration (2000), 911 valid questionnaires | author survey or fieldwork | 2000 | Pilot-type survey used to test scale items; its social-network/social-capital questions are used in Chapter 10 to compare support in previous and current locations. |
| South African Migration and Health Survey (SAMHS) 1999–2000 | author survey or fieldwork | 1999–2000 | Source of descriptive tables and logistic regressions on migrant/non-migrant health in Chapter 6. |
| Census ’96 (South Africa) and its migration questions, as analysed by Kok et al. (2003) | population census | 1996 (with 1975–80 and 1992–96 migration intervals) | Establishes post-apartheid internal migration levels and migrant/non-migrant profiles, and provides the sampling frame for the HSRC surveys. |
| 1980 census internal migration data (1975–1980), excluding Transkei, Bophuthatswana and Venda | population census | 1975–1980 | Baseline for comparing 1970s and 1990s internal migration levels. |
| Census 2001 community profile / Statistics South Africa Census in Brief 2001 | population census | 1991–2001 migration intervals | Basis for GIS maps of inter- and intra-provincial migration flows in Appendix B. |
| United Nations World Population Policies 2003 / International migrant stock series | international trade or finance data | 1976–2003 | Supplies migrant stock, urbanisation and policy-view data for Tables 1.1–1.2 and Figures 1.3–1.8. |
| UN Population Division migration statistics (1995, 2002 wall charts; National Population Policies 2001) | international trade or finance data | 1990–2000 | Used for migrant stock, net migration and policy stance in Table 3.2 of Chapter 3. |
| UN World Population Prospects 2000 revision and World Urbanization Prospects 2001 revision | international trade or finance data | 1950–2000 | Underlies Anderson's residual-method estimates of net migration, urban growth and rural–urban migration rates. |
| SAIRR South Africa Survey 2000/01 (legal immigrants and emigrants 1950–2000) | official statistical series | 1950–2000 | Official count of legal immigrants and emigrants used as second estimate of net migration in Figures 5.3–5.4. |
| October Household Survey 1994–1999 | household or labour survey microdata | 1994–1999 | Used for urban share of Africans and non-Africans and access to basic amenities (Figures 5.8–5.9). |
| SAMP surveys and reports (Migration Policy Series; Tevera & Crush 2003; Rogerson 1999; Dodson 1998; Sechaba 1997/2002; De Vletter 1998) | secondary literature compilation | 1997–2003 | Supplies survey evidence on migrant intentions, gender, skills, miners' amnesty and attitudes to immigrants. |
| Deportation statistics (South African Home Affairs; Botswana immigration) as reported by Oucho 2000, Crush, Tevera & Zinyama | official statistical series | 1988–1997 | Used as an indirect indicator of the scale of undocumented migration. |
| Statistics SA Documented Migration 2001 (Report No. 03-51-03) and earlier Tourism and Migration reports (RSA 1976, 1977; CSS 1988, 1991) | official statistical series | 1974–2001 | Provides counts of documented and skilled immigrants by origin and of Mozambican refugees accepted. |
| Presidential Commission to Investigate Labour Market Policy (1996), Restructuring the South African Labour Market | commission or official report | 1996 (data 1986–1994) | Source of numbers of economically active and professional immigrants by origin country and of documented Mozambican refugees. |
| ILO (1998) Labour Migration to South Africa in the 1990s | commission or official report | 1982–1998 | Used for mine workforce composition, cross-border traffic and costs of migrant labour system. |
| Apartheid-era commissions: Stallard Commission (1922), Holloway Commission (1932), Fagan Commission (1948), Riekert Commission, President's Council report (1985) | commission or official report | 1922–1985 | Cited via Lemon, Unterhalter and Gelderblom & Kok to trace the evolution of influx control policy. |
| Legislation: Natives (Urban Areas) Act 1923, Native Law Amendment Act 1937, Consolidation Act 1945, Natives Laws Amendment Act 1952, Group Areas Act 1950, Abolition of Influx Control Act 1986, Land Acts, Immigration Act 2002, Refugees Act 1998 | legislation gazette | 1913–2002 | Statutes are described to show how the state controlled African urbanisation and foreign labour. |
| Bilateral labour agreements and recruitment records (ZAR–Portuguese East Africa 1897, Modus Vivendi 1901, Nyasaland 1938, Botswana/Lesotho 1973) | chamber of mines or recruiting records | 1896–1977 | Described via Jeeves, Van der Horst, Breytenbach to show historical mine labour supply and recruitment. |
| Mine employment figures by country 1965–2003 (Crush 2000; Cobbe 2004; De Vletter 1998; Van Aswegen & Verhoef 1982; Department of Labour 2003) | secondary literature compilation | 1965–2003 | Documents rise and decline in numbers of Basotho, Mozambican, Swazi and Batswana mineworkers. |
| UNHCR Global Report 2001 and 1993 estimates | international trade or finance data | 1993–2001 | Source for refugee and asylum-seeker numbers by host and origin country (Table 3.5) and Mozambican refugees in South Africa. |
| Development Bank of Southern Africa migration surveys (Cross et al. 1996, 1999) and Cross & Harwin analyses of 1995 OHS | secondary literature compilation | 1995–1999 | Comparison benchmark for lifetime rural-to-rural and rural-to-metro migration shares. |
| Gelderblom & Kok (1994), Urbanisation: South Africa’s Challenge; Lemon (1991) Homes Apart | secondary literature compilation | 1840s–1990 | Provides urbanisation estimates (Table 4.1) and the historical narrative of African urbanisation and removals. |
| Surplus People Project estimate of forced removals 1960–1983 | secondary literature compilation | 1960–1983 | Quantifies the scale of forced removals under apartheid. |
| SAMP attitude surveys (1998) on South African views of immigration | secondary literature compilation | 1998 | Shows hostility to immigrants supporting restrictive policy. |
| HSRC study of causes of cross-border migration from Mozambique and Zimbabwe (Wentzel & Bosman 2001, unpublished) | author survey or fieldwork | 1998–2001 | Supplements the HSRC Migration Survey on cross-border migrants. |
| Qualitative case studies on networks (James 1991; Sharp & Spiegel 1985; Sharp 1994; Spiegel et al. 1996; Meagher 1997; Kotzé 1993; Wilpert 1992) | secondary literature compilation | 1970s–1990s | Evidence that migrant networks can erode, exclude the poor or disengage. |
| Cross-country intention surveys (Taiwan, Thailand, Philippines, Egypt, Kenya, Romania) | secondary literature compilation | 1980s–2001 | Comparison of South African intention prevalence and determinants with other countries. |
New data¶
The 2001–02 HSRC Migration Survey: 3 618 completed households in 711 enumerator areas nationwide (one randomly selected 18–69-year-old respondent each), plus a 2000 initial survey of 911 households and the 1999–2000 SAMHS of 2 371 African households; fieldwork was partly redone because of poor quality.
Inference¶
Quantification: Statistical analysis.
A multi-author edited volume: cross-tabulations and logistic regressions on HSRC surveys (health, intentions), descriptive tables from UN, SAIRR and census data, comparative analysis using residual-method estimates of net migration, historical narrative from secondary literature and legislation, and theoretical synthesis. Historical and apartheid-era claims rest mainly on secondary literature, not original archival work.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| Net migration to South Africa turned negative by the late 1990s mainly because immigration fell, not because emigration rose. (p. 105) | Anderson compares residual-method net migration from UN population figures with SAIRR official counts of legal immigrants and emigrants; both show similar patterns. | Partial data |
| Under apartheid, net rural–urban migration rates were extremely low relative to other countries and did not surge after influx control was removed in 1986; only in the 1990s did they rise to LDR levels. (p. 110) | UN urban population and growth estimates (2001, 2002d) turned into residual net rural–urban migration estimates, compared across world regions. | Partial data |
| Overall internal migration levels did not rise after apartheid controls were abolished: the share of migrants fell from about 13 to 11 per cent between 1975–80 and 1992–96. (p. 92) | Kok et al.'s analysis of the 1980 census and Census '96 migration questions; the 1975–80 data exclude Transkei, Bophuthatswana and Venda. | Partial data |
| Rural-to-rural moves account for most lifetime internal migration among Africans (62%), with rural-to-metro only 8%, though metro flows are probably under-reported. (p. 211) | Preliminary 2001–02 HSRC Migration Survey birthplace-to-residence tabulations, compared with DBSA surveys; author flags under-reporting in the Western Cape. | Partial data |
| Migrant networks matter for destination choice and information: most prospective migrants (about 70%) intend to go where family/friends live, except the affluent. (p. 242) | 2001–02 HSRC survey cross-tabulations by income and gender, with Johannesburg/Cape Town as control for chance. | Systematic data |
| Migration intentions are driven by life dissatisfaction, young age, spouse/mother pressure, divorce/separation and poor electricity; household income is not a significant determinant. (p. 260) | Logistic regressions on 3 400 cases from the 2001–02 HSRC survey, total, rural and urban samples. | Systematic data |
| Apartheid influx control and the migrant labour system were designed to keep Africans, and foreign mineworkers, temporary in urban areas, serving the interests of capital and the state; foreign workers made up over 80% of the mine workforce at the early-1970s peak. (p. 77) | Secondary literature (Lemon, Gelderblom & Kok, Crush, ILO 1998), legislation and recruitment history; no primary archival data. | Illustrative or anecdotal |
| Migrants and non-migrants report similar self-rated health, and rural-to-urban migrants perceive the largest improvements in medical access, though not in health status. (p. 128) | SAMHS 1999–2000 tabulations and logistic regressions on African respondents. | Systematic data |
The author's caveats¶
Authors acknowledge that the 2001–02 HSRC survey fieldwork was poor and large parts redone, with excessive non-response, women over-represented, within-household selection rules not always followed, and delayed data. Cross-country comparisons suffer from definitional differences; urban definitions vary; the 1975–80 census excluded some homelands; Western Cape under-reporting of origins; illegal migrants are not counted in official statistics; small sample sizes for minor streams; intentions data are not validated against actual behaviour; Chapter 9 results are labelled preliminary.
Gaps in the evidence¶
The volume says little about the apartheid-era migration itself using original sources: no Chamber of Mines/WNLA/TEBA recruiting records, pass-law or influx-control administrative data, or archival files are analysed, so historical claims are second-hand. Longitudinal or panel data linking intentions to actual moves (planned for 2006–07) would better test the intentions-based causal claims, and the 1980 and 1991 census microdata including homelands would permit a true comparison of migration before and after influx control. Better measures of undocumented migration and of under-reported origins in the metros are also needed.
Replicability: Data described but not public.