The Role of Regional Policy in South Africa¶
Trevor Bell (1986). The Role of Regional Policy in South Africa. Journal of Southern African Studies, Vol. 12, No. 2. Taylor & Francis (Oxford University Press).
| Type | journal article |
| Scope | Economics and economic history of apartheid |
| Method | descriptive quantitative |
| Period studied | 1956–1985 |
| Regions | PWV (Pretoria-Witwatersrand-Vereeniging), Durban-Pinetown, Western Cape, Port Elizabeth-Uitenhage, Border areas and homelands, KwaZulu, Ciskei, Natal |
| Themes | manufacturing industrial policy, bantustans removals influx, labour markets wages, unions strikes, business and state, demography urbanisation, politics ideology elections |
| Citation key | bell1986 |
Argument¶
Bell uses manufacturing census data to show that the share of manufacturing employment in the four main industrial regions rose until 1965/6 and then fell without interruption to 1979. Border areas and homelands gained share over the same period. He argues that this deconcentration began before the Physical Planning Act (1967) and before large financial incentives existed. It therefore owes more to market and competitive pressures than is commonly thought: racial and regional labour-market segmentation, wage differentials, rising Black education and productivity, deskilling, and growing union power in the metropoles. State incentives and direct labour controls reinforced the trend but did not cause it. The absolute number of jobs created in the periphery is small, so the state's persistence with the policy cannot be explained by stemming Black urbanisation or legitimising homeland governments. Bell suggests the aim is a federal or regional political structure, plus possibly containing Black union power through competition at the margin.
Evidence and method¶
Industrial and manufacturing censuses of 1956/7, 1961/2, 1965/6, 1967/8, 1970, 1972, 1976 and 1979 (Bureau/Department of Statistics, Central Statistical Services), re-aggregated by magisterial district into consistent regions. Also his own earlier studies, Barker on the clothing industry, Board of Trade and Industries reports, conference papers on KwaZulu and Ciskei, and secondary literature (Hirsch, Lacey, Legassick, Davies, Zille, Tomlinson and Hyslop).
Place in the debates¶
- Purpose of industrial decentralisation policy (legitimising bantustan governments vs economic motives vs stemming urbanisation) (see Homelands, separate development, decentralisation and removals). Rejects the Tomlinson-era aim of stemming urbanisation and Hirsch's legitimation thesis as the main explanation of persistence, since the job numbers are too small. Suggests the primary current aim is a federal political structure, with containing Black unions as a possible secondary effect. Gives more weight to Lacey's economic interpretation than Hirsch does.
- State intervention vs market forces as the cause of deconcentration (see Homelands, separate development, decentralisation and removals). Argues that market and competitive forces, including labour-market segmentation, wage differentials and rising Black productivity, were the main cause. Incentives and controls played a smaller role than commonly thought, although the post-1982 incentive package probably matters.
- Role of the racial division of labour and capital restructuring (Legassick, Davies) (see Homelands, separate development, decentralisation and removals). Close to Davies in stressing competitive pressures as the fundamental force changing the racial division of labour, but adds emphasis on the changing inter-regional structure of production.
- Effectiveness and sustainability of regional financial incentives (see Homelands, separate development, decentralisation and removals). Doubts their explanatory power for the 1960s-70s trend and their long-term sustainability. Treats them as a pure income transfer from taxpayers to rural capitalists, citing the Ciskei cost of over R30,000 per job.
Findings¶
-
Decentralisation failed initially: the metropolitan share of manufacturing employment rose in the early 1960s boom. (p. 4)
The proportion of manufacturing employment in the main centres increased from 79.6 to 81.9 per cent, and no headway was made in dealing with the dangers emphasized by the Tomlinson Commission.
-
From 1965/6 the share of the main industrial areas fell continuously, indicating a structural break in the late 1960s. (p. 6)
between 1965/6 and 1979 it fell without interruption from 81.9 per cent to 76.9 per cent.
-
The border areas and homelands steadily gained share of manufacturing employment after 1967/8. (p. 6)
substantially increased their proportionate share of manufacturing employment without interruption from 11.75 percent in 1967/8 to 15.54 per cent in 1979
-
Growth rates in the metropolitan core collapsed while the periphery's held up. (p. 8)
Whereas the average annual rate of growth of manufacturing employment in the main industrial areas was 5.9 per cent in 1961/2 to 1967/8 and fell to 2.1 per cent in 1967/8 to 1979
-
The absolute number of jobs created in the periphery is too small to have altered the geography of the Black population. (p. 9)
these numbers of extra jobs clearly cannot be expected to have had any significant impact on the geographical distribution of the black population
-
Deconcentration in clothing and textiles predates both the incentives and the controls, pointing to market forces. (p. 11)
It appears, therefore, that there was in operation a tendency towards industrial deconcentration in certain industries going as far back as the 1950s and even the 1940s
-
Racial and regional labour-market segmentation gave firms an incentive to relocate to cheaper-labour areas. (p. 11)
The South African labour market, clearly, is segmented both on a racial basis within metropolitan areas by the job labour bar, and on a regional basis by influx control measures of various kinds.
-
The incentive package was escalated sharply: the wage-bill subsidy rose from 40 to 95 per cent in the homelands in 1982. (p. 13)
From April 1982 this subsidy in the Black homelands was raised to 95 per cent of the wage bill
-
Incentives cannot explain the timing of the 1960s trend, since they were negligible then. (p. 14)
there is no apparent reason why the inducements should have become more effective from the mid to late 1960s, when the trend begins, than they had been previously.
-
The Ciskei incentive scheme was fiscally unsustainable, with a payback of over thirty years. (p. 15)
average the Ciskei government will recoup the cost of concessions paid in just over thirty years time, and this does not allow for any discount rate to cover the cost of capital to the Ciskei.
-
Government influence on the trend operated mainly through labour controls rather than financial incentives. (p. 15)
in so far as Government has had a significant influence on the trend towards industrial decentralisation it has probably been mainly through controls on population movements and on the use of Black labour in the metropolitan areas
-
The current aim of regional policy is political: a federal structure based on the Good Hope development regions. (p. 17)
The primary aim of state policy in the current phase of development appears to be the creation of a federal political structure, possibly based on the development regions A to H announced at the Good Hope Conference
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| Industrial Census 1956/7, Report No. 240 (Bureau of Census and Statistics) | industrial mining census | 1956/7 | Baseline regional manufacturing employment for 1956/7, summed from magisterial districts to build Table 1. |
| Industrial Census 1961/2, Report No. 304 (Bureau of Statistics) | industrial mining census | 1961/2 | Regional manufacturing employment for 1961/2 for Tables 1 and 2. |
| Census of Manufacturing 1967/8, 1965/6 and 1963/4, Report No. 10-21-18 (Department of Statistics) | industrial mining census | 1963/4-1967/8 | Regional manufacturing employment for 1965/6 and 1967/8, the turning point in the concentration trend. |
| Census of Manufacturing 1970, Report No. 10-21-27 (Department of Statistics) | industrial mining census | 1970 | Regional manufacturing employment for 1970 in Tables 1 and 2. |
| Census of Manufacturing 1972, Report No. 10-21-29 (Department of Statistics) | industrial mining census | 1972 | Regional manufacturing employment for 1972 in Tables 1 and 2. |
| Census of Manufacturing 1976, Report No. 10-21-33 (Department of Statistics) | industrial mining census | 1976 | 1976 regional employment used for the 1976-79 comparison, including Natal and Kwazulu detail in Table 3. |
| Census of Manufacturing 1979, Report No. 10-21-40 (Central Statistical Services) | industrial mining census | 1979 | 1979 endpoint showing the main areas' share falling to 76.9% and the rise of homeland and border employment. |
| Tomlinson Commission Report, U.G. 61/1955 | commission or official report | 1955/56 | Frames the original political aims of industrial decentralisation, including the recommendation on white investment in the homelands. |
| Physical Planning and Utilization of Resources Act No. 88 of 1967 | legislation gazette | 1967 | Cited as a direct control on the Black:White labour ratio in metropolitan manufacturing, tested as a cause of deconcentration. |
| Wiehahn Committee report 1979 and Homelands citizenship/constitution legislation, influx-control statutes (1952 Native Laws Amendment Act, 1964 Bantu Laws Amendment Act) | legislation gazette | 1949-1979 | Institutional background on labour-market segmentation, influx control and industrial-relations reform used to weigh state controls against market forces. |
| Board of Trade and Industries report No. 303 of 1947 | commission or official report | 1947 | Cited via Barker as early recognition of labour-cost-driven dispersal of clothing plants. |
| Clothing and textiles industry regional employment shares (from Bell's own earlier work, 'Growth and Structure of Manufacturing Employment in Natal', 1983, and Carnegie Paper 243) | secondary literature compilation | 1956/57-1979 | Shows PWV share of clothing and textiles employment falling before 1968, supporting market forces over state controls. |
| H. A. F. Barker, The Economics of the Wholesale Clothing Industry of South Africa 1907-1957 | secondary literature compilation | 1907-1957 | Evidence that clothing plant dispersal to rural and small-town locations began in the late 1940s. |
| Spies, 'The Kwazulu Experience' and Bridgman, 'The Ciskeian Experience' (1984 workshop papers) | secondary literature compilation | 1984 | Substitutes for post-1979 census data on rapid manufacturing employment growth in Kwazulu and Ciskei, and supplies Ciskei cost-per-job and payback figures. |
| Zille interview with Dr Simon Brand (in South African Review I) | interviews oral history | c.1983 | Official statement of the legitimacy objective of decentralisation policy. |
| Decentralisation Board application data, as reported in Tomlinson and Hyslop 1984 | secondary literature compilation | 1981-84 | Shows rising applications for incentives after the 1981 Good Hope Conference, reflecting prospective rather than actual jobs. |
| Statement by Dr P. J. Riekert, January 1970 | other | 1970 | Illustrates the late-1960s shift in official perceptions of homeland development. |
New data¶
Author reconstructed a consistent time series of manufacturing employment for the four main industrial regions, border areas and homelands (excl. Transkei), 1956/7-1979, by harmonising changing census region codes and magisterial districts, with estimates for Inanda and Sasolburg.
Inference¶
Quantification: Descriptive tables.
Comparison of regional shares and growth rates of manufacturing employment across census years, then qualitative causal reasoning on timing and sectoral (clothing/textiles) evidence to weigh market forces, inducements, controls and unions; no formal statistical testing.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| The share of manufacturing employment in the four main industrial areas fell continuously after 1965/6, from 81.9% to 76.9% in 1979, after rising from 1956/7. (p. 6) | Census of Manufacturing series 1956/7-1979 harmonised in Table 1. | Systematic data |
| The border areas and homelands increased their share of manufacturing employment steadily from 11.75% in 1967/8 to 15.54% in 1979 and grew faster than the metropolitan core. (p. 6) | Table 2 from census regional data; growth rates computed 1961/2-1967/8 and 1967/8-1979. | Systematic data |
| The absolute number of jobs created in the border areas and homelands was small, so decentralisation cannot explain state aims of stemming black urbanisation or legitimising homeland governments. (p. 9) | Census employment counts (Table 3, Kwazulu and Ciskei increases) plus argument. | Partial data |
| Deconcentration began before the Physical Planning Act and large financial inducements, driven by market forces and labour-cost differentials, as shown by clothing and textiles shifting away from the PWV before 1968. (p. 11) | Sectoral PWV share data for clothing and textiles from Bell's earlier work, Barker's history, and timing argument. | Partial data |
| Financial inducements before 1982 were too weak and mistimed to explain the long-term deconcentration; the 1982 package is of a different order and its apparent impact is partly due to recession. (p. 14) | Comparison of inducement schedules over time with timing of the census trend; pending application data and Ciskei case reports. | Partial data |
| Intensified international competition and the rise of black union bargaining power may have reinforced deconcentration, though no detailed evidence links these to industry-level trends. (p. 13) | Timing inference and secondary literature (Beenstock, Murray, Davies, Lacey); author admits no industry-level evidence. | Asserted |
| The primary aim of current state policy appears to be creating a federal political structure based on the Good Hope development regions. (p. 17) | Inference from Good Hope Conference regions and secondary literature (Cobbett et al.). | Asserted |
The author's caveats¶
Author notes a paucity of comprehensive data and that census definitions changed, requiring estimates for Inanda and Sasolburg and exclusion of the Transkei and Bophuthatswana. No census data exist after 1979, and application data reflect only prospective jobs. Absolute increases in homelands are from very small bases.
Gaps in the evidence¶
The causal weight given to market forces versus inducements and controls rests on timing and two sectors, with no firm-level location decisions, wage differentials or plant relocation data. Firm surveys, Decentralisation Board project files and actual disbursements, and industry-level import-competition data would test the argument better. Some of these existed in the 1980s in archives and censuses, and post-1979 census and later establishment data exist now.
Replicability: Data described but not public.