Growth pole theory and regional development in South Africa: Results of a survey in selected growth areas¶
Alison Todes; Vanessa Watson (1984). Growth pole theory and regional development in South Africa: Results of a survey in selected growth areas. Second Carnegie Inquiry into Poverty and Development in Southern Africa, Carnegie Conference Paper No. 245. SALDRU, University of Cape Town.
| Type | conference paper |
| Scope | Economics and economic history of apartheid |
| Method | survey fieldwork |
| Period studied | 1940s–1984 |
| Regions | Western Cape (Atlantis, George), Eastern Cape (Butterworth, Kingwilliamstown, Transkei, Ciskei), Transvaal (Brits, Babelegi, Potgietersrus, Seshego, Lebowakgomo), Northern Cape (Kimberley), Orange Free State (Selosesha) |
| Themes | manufacturing industrial policy, bantustans removals influx, labour markets wages, inequality poverty unemployment, demography urbanisation, comparative theory, business and state |
| Citation key | todes1984 |
Argument¶
The paper tests South African industrial decentralisation policy against classical growth pole theory, using a postal survey of 13 growth centres conducted in April 1983. It argues that the centres rarely attract propulsive industries, have weak local linkages, and lack agglomeration economies. They remain dependent on metropolitan inputs, markets and head offices, and on heavy, indiscriminate and long-lasting incentives plus cheap labour. Benefits do not spread to hinterlands or the settlement hierarchy, and jobs are mostly unskilled, low-wage and costly per job, especially in homeland and border centres. The authors conclude that the policy was never simply developmental. It served contradictory political imperatives (influx control, homeland legitimation, constituency politics) and may come to depend on coercive measures to keep conditions attractive to industrialists.
Evidence and method¶
Postal survey of industrialists in 13 growth centres (April 1983), a small sample with low response rates that over-represents homeland towns, new towns and capital-intensive sectors. Also the 1976 Census of Manufacturing regional statistics, Decentralisation Board reports (1982/83), the Financial Mail, secondary studies (Bell, Davies et al., McCartan, Gottschalk, Moseley, Gilbert, Robinson and Salih) and interviews in Butterworth.
Place in the debates¶
- Applicability of growth pole theory/policy to South Africa's periphery (see Homelands, separate development, decentralisation and removals). Argues the growth centre model has failed on all three counts (self-sustaining growth, quality of life in the centre, spread to the region), as in studies elsewhere, and that failure is not a teething problem.
- Was decentralisation largely spontaneous, and would it have occurred without incentives (Bell)? (see Homelands, separate development, decentralisation and removals). Partly sceptical of Bell. The authors say industries were sometimes forced out of metropolitan areas and that incentives were a reason for location in almost half of cases, though labour supply matters. They hold that incentives mainly distort location and encourage inefficient location.
- Purpose of decentralisation policy: developmental vs political/coercive control of urbanisation and migration (see Homelands, separate development, decentralisation and removals). Holds the policy was never simply developmental. It was governed by contradictory imperatives (curbing migration, bolstering homelands, white constituency politics), and homeland unemployment is the direct result of state policy.
- Do multi-plant branch firms harm regional development? (see Homelands, separate development, decentralisation and removals). Inconclusive. Non-independence shows no clear negative or positive effect in the data, because local linkage is low everywhere.
Findings¶
-
Propulsive industries were almost entirely absent from the surveyed growth centres, and the one example was subsidised and inefficient. (p. 10)
In fact in only one instance had a propulsive industry (a) been attracted to a growth centre
-
Backward linkages to the growth centre itself are negligible: most inputs come from metropolitan areas. (p. 13)
In a smaller sample (f) it was found that only 3,1% of the Rand value of inputs was bought in the growth centre.
-
Forward linkages are also weak in value terms, so the local multiplier is limited. (p. 15)
Table 11 (in Appendix 2) shows that the Rand value of local output is only 4,6% of the total output.
-
Most growth-centre firms are branches or subsidiaries of larger companies, i.e. tied to metropolitan centres. (p. 16)
60% of firms sampled in the survey are part of larger companies (j).
-
Incentives were long-lasting rather than temporary, undermining the growth pole logic. (p. 25)
40% of industries which were established in growth centres before 1960, still receive incentives
-
Incentive dependence is substantial, with a significant share of centre employment at risk if incentives were withdrawn. (p. 24)
the survey showed that if incentives were removed, some 19,2% of jobs in growth centres would be lost while a further 7,8% would be in doubt.
-
Incentives were spread thinly across many places, preventing agglomeration economies from building up at any one point. (p. 26)
applications for assistance were approved at no less than 122 places.
-
Official projected job figures are likely to be overstated, since realised employment is a fraction of approved project employment. (p. 28)
employment in projects which actually materialise is likely to be between 20% and 74% of employment in approved projects: in most years it is about half.
-
Growth-centre jobs are costly and are offset by jobs lost in metropolitan areas through Physical Planning Act refusals. (p. 35)
The cost of jobs is even higher when set against the 146 000 potential jobs lost in metropolitan areas between 1968 and 1983
-
Jobs in growth centres are overwhelmingly unskilled, especially in homeland centres. (p. 36)
Table 32 indicates that overall 73,6% of jobs in growth centres in the survey are classified as unskilled and in homeland areas this proportion rises to 87% of total jobs.
-
Women, the majority of workers in homeland and border centres, earn very little relative to national counterparts, depressing wages. (p. 37)
while their earnings are some 32% - 34% of their national counterparts.
-
The scale of manufacturing job creation in the Transkei growth points is tiny relative to what would be needed for agricultural transformation. (p. 42)
Given that it has been possible over the last twenty years to create only 11 000 manufacturing jobs in Butterworth and Umtata
-
Homeland unemployment is attributed to state policy, and decentralisation did not add net welfare gains. (p. 49)
It seems likely that if neither decentralisation nor influx control had existed, the same number of jobs would have been created elsewhere, and under better conditions for workers.
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| Postal survey of industrialists in 13 growth centres, April 1983 (Todes and Watson) | author survey or fieldwork | April 1983 | Primary evidence on propulsive industries, input/output linkages, firm status, services, labour, incentive dependence, skills, wages, social infrastructure. |
| Interviews with industrialists in Butterworth | interviews oral history | c.1983 | Supports claim that industrialists prefer female employees and illustrates how larger firms internalise services. |
| Linkage survey format (de Necker, Stellenbosch University) | secondary literature compilation | n/a | Survey instrument format borrowed for the linkage questions; a smaller sample of full answers is used for Rand-value input/output shares. |
| Republic of South Africa, Board for the Decentralisation of Industry, Report on the Activities of the Board 1 April 1982 to 31 March 1983, RP 52/1983 | commission or official report | 1982-83 | Source for number of places receiving assistance, approved-project job figure (65 000), applications and employment in approved projects. |
| 1982 Decentralisation Board Report (cost per job) | commission or official report | 1981-82 | Basis for estimated cost of R15 932 in direct investment per industrial job in growth centres. |
| Census of Manufacturing, Principal Statistics on a Regional Basis (1976) | industrial mining census | 1976 | Used to compare Black manufacturing wages in selected growth centres with the national average (Table 33). |
| Table 23: Growth in manufacturing employment in selected growth centres; Tables 27-29 on employment in approved vs actual projects and non-metropolitan manufacturing output (1976) | official statistical series | 1975-1983; 1976 | Shows drop-out between approved and realised projects, the countercyclical pattern of decentralised employment, and the small share of clothing/textiles outside the metropolitan areas. |
| McCartan, P., 'The Demand for Skilled Labour in the Border, Ciskei, Southern Transkei Regional Economy', ISER Development Studies Working Paper No. 15, 1983 | secondary literature compilation | 1979-1985 (projected) | Source of labour, capital and capital/labour growth figures (Table 31) showing capital intensification. |
| Physical Planning and Utilisation of Resources Act, Section 3 refusals 1968-1983 (via Financial Mail) | legislation gazette | 1968-1983 | Cited for 146 000 potential jobs lost in metropolitan areas through refusals under the Act. |
| Financial Mail articles ('Diesel Engines. Where's the Magic Formula', 18 Feb 1983; 'Physical Planning Act. Maintaining Job Destruction', 10 Feb 1984) | newspapers periodicals | 1983-1984 | Details on the single propulsive industry (IDC-established diesel engines) and on job losses under the Physical Planning Act. |
| 1980 Development Strategy for Transkei (as reported) and Transkei urban plans | commission or official report | 1980-2003 | Source of projection that 60% of rural population must be urbanised by 2000 and plans for 90 rural service centres and 20 market towns. |
| Personal communication, Dr Bikitsha, Transkei Minister of Health | interviews oral history | c.1983-84 | Supports claim of housing restriction and overcrowding in Transkei towns. |
| Davies, Lochner and Wait, Greater East London/Kingwilliamstown Development Region, Institute for Planning Research Report No. 20, 1980 | secondary literature compilation | 1980 | Detailed linkage findings for East London used to corroborate survey linkage patterns. |
| Bell, R.T., 'The Growth and Structure of Manufacturing Employment in Natal', ISER Occasional Paper No. 7, 1983 | secondary literature compilation | 1983 | Counter-argument that decentralisation is largely spontaneous; authors test it against their survey data. |
| Moseley, Robinson and Salih, Gilbert growth pole impact studies (Rennes, Kuala Lumpur, Medellin) | secondary literature compilation | 1971-1976 | International empirical findings used to show spread effects limited to a narrow radius around the growth centre. |
| Gottschalk, 'Industrial Decentralisation, Jobs and Wages', SA Labour Bulletin 1977; Black (1976); Rogerson and Kobben (1982); McCarthy (1983) | secondary literature compilation | 1976-1983 | Used for prior documentation of lower growth-centre wages, purchasing power leakage, clothing factory closures, and 115 000 annual jobs needed in homelands. |
New data¶
Postal survey of industrialists in 13 growth centres (April 1983), roughly 100 responding firms, covering location type, age, incentives, linkages, status, skills, wages and infrastructure. Programmed by SALDRU; appendices available on request.
Inference¶
Quantification: Descriptive tables.
Survey-based comparison across types of growth centre (deconcentration points, border, homeland, isolated White towns), with cross-tabulations and a few chi-square/t-tests, benchmarked against growth pole theory and supplemented by official data, secondary studies and illustrative cases (Butterworth, Transkei).
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| South African growth centres have almost entirely failed to attract propulsive industries; only one was found (an IDC-established, subsidised plant near a metropolis). (p. 4) | Authors' 1983 postal survey with a size/capital-intensity/linkage definition of propulsive industry. | Partial data |
| Local linkages are very weak: few firms buy inputs from the growth centre and most spin-off accrues to nearby metropolitan areas. (p. 7) | Survey Tables 7-12 (N=80 for inputs, 75 for outputs) plus a smaller sample giving Rand values of inputs and outputs. | Partial data |
| Incentives are not temporary or designed for specific industries; many firms have long depended on them, though only a minority would collapse without them. (p. 18) | Survey Tables 25-26 on dependence on incentives, plus the Decentralisation Board 1983 report on 122 places. | Partial data |
| The official job projection of 65 000 jobs in approved projects is likely to be heavily overstated, since realised jobs are typically about half of approved ones. (p. 22) | Decentralisation Board 1983 report and tables of approved vs actual project employment in earlier years (Table 27), with reference to the drop-out rate. | Partial data |
| Jobs in growth centres are overwhelmingly unskilled and low-wage, especially in homelands and border areas, with a high share of women workers. (p. 30) | Survey skill-level coding (Table 32), 1976 Census of Manufacturing regional wages (Table 33) and survey wage data (Tables 34-35). | Systematic data |
| Growth centres provide inadequate social infrastructure and raise living costs relative to wages; overall dissatisfaction is high. (p. 32) | Survey Table 36 and industrialists' complaints (Table 37), plus illustrative cases (Atlantis, Transkei). | Partial data |
| Growth centres have almost no spread effects on hinterlands: remittances are negligible, absorption of rural population is minimal, and settlement hierarchy linkages are poor (Transkei). (p. 35) | Transkei example using worker numbers, estimated wages and household subsistence levels, the 1980 Development Strategy, and international growth pole studies; the survey could not test diffusion. | Illustrative or anecdotal |
| Decentralisation policy's sustained industry in the deep periphery depends on subsidies and cheap labour, and may rely increasingly on coercive measures such as influx control. (p. 43) | Synthesis of survey findings and argument from the literature; no direct test. | Asserted |
The author's caveats¶
The authors state the sample is small, with very low response rates overall and to certain questions; it over-represents homeland and new towns (and Butterworth) and capital-intensive sectors; no Natal towns and several of the largest growth centres (Newcastle, East London, Bloemfontein) are omitted. Results are said to be treated 'with a great deal of caution' and supported by supplementary evidence where possible. Wage data are for 1976 only, recent data are lacking, and no consolidated current cost-per-job figures exist.
Gaps in the evidence¶
The argument would need a representative, larger firm-level dataset (or administrative Decentralisation Board records and the Census of Manufacturing at town level) with panel data on employment, incentives and linkages over time. Counterfactual evidence on what would have happened without incentives, and direct measures of hinterland impact (household income, remittances, commuting), are missing and rely on illustrative Transkei examples. The Board's project files and Census of Manufacturing returns existed then and now survive in archives, so many of these tests are feasible.
Replicability: Data described but not public.