African Manufacturing Wages: A survey of new data sources and estimates from African homelands during the late apartheid period (1975 – 1992)¶
Edward M. Kerby (2014). African Manufacturing Wages: A survey of new data sources and estimates from African homelands during the late apartheid period (1975 – 1992). Preliminary draft prepared for the LSE African Economic History Workshop, 25-26 October 2014. London School of Economics and Political Science, Department of Economic History.
| Type | working paper |
| Scope | Economics and economic history of apartheid |
| Method | descriptive quantitative |
| Period studied | 1970–1992 |
| Regions | South Africa, Homelands/Bantustans (Transkei, Ciskei, KwaZulu, QwaQwa, Bophuthatswana, Venda etc.), RIDP decentralisation zones A-H, PWV, Durban, Port Elizabeth, Cape Town metropolitan complexes |
| Themes | manufacturing industrial policy, bantustans removals influx, labour markets wages, methods data sources, business and state, inequality poverty unemployment |
| Citation key | kerby2014 |
Argument¶
Official statistics barely record African manufacturing wages outside the main metropolitan areas, where most Africans lived. The paper builds a new wage dataset for the Regional Industrial Decentralisation Programme (RIDP) zones in the homelands. It triangulates three sources: SALDRU newspaper clippings, RIDP/homeland trade almanacs and magazines, and official or Carnegie Commission reports. Preliminary results show urban African semi-skilled wages well above homeland wages, with wages rising faster in RIDP zones from the late 1970s but unevenly across zones. The paper reads this as manufacturing being displaced from the cities to the periphery, with a widening disparity between peri-urban and rural homeland zones. It closes by proposing ways to extend the work into a fuller post-WWII African wage series.
Evidence and method¶
SALDRU Clippings Collection (c.1975-2000; articles and advertisements citing unskilled wages at 49 of 63 zones); RIDP almanacs and homeland trade periodicals (Transkei Topics, Development, Transkei Today, 1978-1994); Bureau for Economic Research re Bantu Development (1976) report; Gottschalk (1977); five Second Carnegie Inquiry working papers (Thomas, Abedian, Ardington, Bank, McCarthy); Manpower Survey data for urban wages (1970, 1980, 1985); National Archives Pretoria maps; secondary quotations (Hirsch, Tomlinson and Krige, Addleson et al., Pickles and Woods).
Place in the debates¶
- Marxist vs liberal interpretations of low African wages and apartheid (see Liberal vs revisionist: was apartheid functional or dysfunctional for capitalist growth?). Summarises both sides (capital benefited from low wages vs low wages as a by-product of protecting white workers). Argues the low homeland wages in the traditional view of the RIDP are consistent with both, and calls for a more empirical, non-ideological approach.
- What drove the transformation of the apartheid labour market (white preferences vs economic incentives) (see The job colour bar, white labour and its erosion). Builds on Mariotti (2012), who finds economic incentives rather than changed white preferences. Extends the geography to the periphery and suggests some industries relocated to the homelands rather than lobbying to reform job-reservation laws.
- Why industry moved to decentralisation zones (cheap labour vs wider cost and competition pressures) (see Homelands, separate development, decentralisation and removals). Cites Bell (1986, 1987) that international competition, rising costs and political pressure also drove relocation. Hypothesises that lower unskilled wages raise a zone's attractiveness for labour-intensive industry.
Findings¶
-
The RIDP was large in scale, with thousands of factories in dozens of homeland zones, but was never covered by wage statistics. (p. 2)
The industrial decentralization programme witnessed the rollout of 3500 factories across 67 manufacturing zones in the racially segregated homelands over 40 years.
-
Official sources recorded no African earnings in the 1970-85 censuses, so African wage data are the key gap. (p. 4)
The 1970, 1980 and 1985 census did not record African salaries despite recording those of Whites, Coloured and Indian.
-
The mining industry's wage statistics also lost racial breakdown after 1985, compounding data problems. (p. 10)
the mining industry ceased to provide a breakdown of wages according to racial classification from 1985.
-
Mariotti's finding is that job reservation declined without evidence of rising real African wages. (p. 5)
there was little evidence of an increase in African real wages.
-
Some industries relocated to the periphery instead of pressing for reform of reservation laws. (p. 5)
certain manufacturing industries chose to move to peripheral areas rather than pressurize government to reform reservation laws.
-
The new newspaper-based series covers most zones. (p. 12)
49 out of the 63 development sites have unskilled wages cited in articles or advertised in a financial publication.
-
Reported unskilled wages in zones varied widely by gender and zone, with a Transkei de facto minimum of about R56 a month in 1985. (p. 14)
unofficial de facto minimum wage is roughly R0.35 per hours in 1985.
-
Urban African semi-skilled wages were substantially higher than homeland or periphery wages, with the Transvaal highest. (p. 17)
On average, wages across this complex were 50% higher than the other urban areas.
-
The author's main preliminary finding is relocation of manufacturing to RIDP zones with emerging disparity between zones. (p. 17)
Manufacturing was being displaced from urban centres to regional RIDP zones. However a disparity between zones was emerging.
-
The Marxist-liberal debate is summarised as turning on whether capital or white-worker protection explains low wages. (p. 8)
The Marxists contended that capital benefited from low wages whereas liberals argued that although capital did benefit from low wages, low wages were simply the outcome of a policy
-
Homeland labour law suppressed worker voice in the Transkei zones: strikes were banned and a low minimum wage was set. (p. 14)
Strike action is not permitted. Minimum wages have been set down for various trades and industries. The minimum wage is normally 35c an hour
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| Manpower Survey 1980 (Statistics South Africa / Manpower Commission), National Archives Pretoria, with Bell (1987) data | official statistical series | 1980 (and series 1969-1985) | Shows share of manufacturing employment outside metropolitan areas rising while the main industrial complexes stagnated or declined (Figure 1). |
| Man Power Surveys 1970, 1980 and 1985 (urban African manufacturing wages) | official statistical series | 1970, 1980, 1985 | Supplies the urban African semi-skilled wage benchmarks against which homeland RIDP wages are compared in Figure 3. |
| Population censuses 1970, 1980 and 1985 (27 enumeration areas) | population census | 1970-1985 | Used to demonstrate the limited coverage of official wage data (no African salaries, metropolitan enumeration areas only) which motivates the new data. |
| Regional manufacturing census | industrial mining census | Apartheid era (unspecified) | Described as containing wages but lacking racial classification, forcing researchers to assume African worker ratios; a limitation, not used directly. |
| SALDRU Clippings Collection 1975-2000 (English-language newspaper and financial periodical clippings), University of Cape Town | newspapers periodicals | 1975-1991 (collection 1975-2000) | Primary source of unofficial publicly cited African unskilled wage estimates per zone, compiled into a series for 8 development regions. |
| Homeland trade periodicals: 'Transkei Topics', KwaZulu 'Development', Transkei Today (1978-1994) | newspapers periodicals | 1978-1994 | Used to test and substitute for newspaper wage estimates, e.g. a de facto minimum wage of about R0.35 per hour in 1985, while treated cautiously as propaganda. |
| Transkei Development Corporation 1983 Almanac and Transkei Today Vol. 3 No. 2 (April-June 1985), Mthatha Archives | company records or reports | 1983-1985 | Provides the statutory minimum wage (35c/hour) and bans on strikes and unions used to corroborate estimates of homeland wages. |
| Bureau for Economic Research re Bantu Development (BENSO), Black development in South Africa (1976), Statistical Survey of Black Development 1980 and 1981 | commission or official report | 1971-1975 wages (reports 1976-1981) | Official minimum and maximum unskilled male and female wages 1971-75 used as a proxy for RIDP zones and to corroborate estimates. |
| Gottschalk (1977) wage survey of eighteen RIDP zones and peri-urban growth points | secondary literature compilation | 1974-1975 | Used to corroborate the Bureau for Economic Research wage estimates for the mid-1970s. |
| Second Carnegie Inquiry into Poverty and Development in Southern Africa: five SALDRU working papers (Thomas 1984, Abedian 1983, Ardington 1984, Bank 1984, McCarthy 1984) | commission or official report | early 1980s | Field and survey-based wage estimates for early 1980s, covering both TBVC and non-independent homelands. |
| Secondary wage quotes: Addleson et al. (1985), Tomlinson and Krige (1997), Hirsch (1986), Pickles and Woods (1989) | secondary literature compilation | 1980s | Anecdotal wage figures (R55/R45 per month; R0.50/hour) used to externally validate the new wage series. |
| Regional Industrial Development Programme maps and zone application reports, National Archives Pretoria (RSA TAB7678 01/79) | maps spatial data | circa 1978-79 | Digitised and geocoded to map RIDP zones and the eight development regions. |
| Bureau for Economic Research re Bantu Development (1976) archival maps of homeland boundaries | maps spatial data | 1976 boundaries (map circa 1955) | Source for digitised map of Bantustan boundaries (Map 1). |
| Industrial Development Archive, Sandton, Johannesburg | state archives | apartheid era (unspecified) | Cited for the description of decentralisation and deconcentration zones as a compromise between business and government. |
| Email from Martine Mariotti, November 2013 | interviews oral history | 2013 | Supports the claim that researchers rely on highly inaccurate ratios of African workers. |
| Mariotti (2012) and Leibbrandt (2011) analyses of census and homeland data | secondary literature compilation | 1969-1985; 1980s | Benchmark studies with which the new results are said to be consistent and complementary. |
New data¶
A new dataset of nominal African semi-skilled and unskilled manufacturing wages (1978, 1982, 1984, 1991) for about 49-63 RIDP zones, grouped in 8 regions (Table 1), plus homeland-level annual wages for 1970, 1980 and 1985 (Figure 3), built from newspaper clippings, trade publications and official or Carnegie reports; also a zone-by-sector count of factories (Appendix D).
Inference¶
Quantification: Descriptive tables.
Descriptive comparison across regions and time of a triangulated wage compilation (newspapers, trade publications, official and Carnegie reports) against urban Manpower Survey wages; preliminary trends read from summary tables and maps, with no statistical testing.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| Semi-skilled African manufacturing wages were substantially higher in urban areas than in homelands or the urban periphery. (p. 17) | Comparison of Manpower Survey urban wages with the compiled RIDP and homeland wage estimates in Figure 3. | Partial data |
| Manufacturing was displaced from urban centres to RIDP zones, with widening disparity between zones. (p. 17) | Figure 1 employment shares and zone wage tables; Natal/Durban to Zone E illustrated. | Partial data |
| The share of manufacturing jobs outside the metropolitan areas rose while main industrial complexes stagnated or declined. (p. 3) | Figure 1 from Bell (1987) and the 1980 Manpower Survey. | Partial data |
| RIDP homeland wages escalated from the late 1970s, faster than urban wages but not uniformly, and then fell toward 1990. (p. 17) | Table 1 and Figure 3 nominal wage compilations; explanation (end of decentralisation or influx control) is speculative. | Partial data |
| The new unofficial wage estimates are externally valid because they are consistent with Hirsch, Tomlinson and Krige, the RIDP almanac and Carnegie estimates. (p. 13) | Cross-checking SALDRU clippings against a few quoted secondary figures and the Transkei almanac minimum wage. | Illustrative or anecdotal |
| Official sources contain little or no African wage data for homelands and periphery, because of excluded enumeration areas, independent homelands and missing racial classification. (p. 10) | Census and Manpower Survey coverage as described, plus quotations from Hofmeyr, Leibbrandt and Fourie and Schirmer. | Partial data |
The author's caveats¶
The author acknowledges that the newspaper data are not exhaustive (49 of 63 sites), that the empirical evidence is neither free of bias nor misrepresentation, and that homeland trade magazines were propaganda with a clear agenda and must be treated with caution. Results are labelled preliminary and require further statistical analysis; the data are described as imperfect.
Gaps in the evidence¶
The wage series rests on scattered press quotations, promotional magazines and a handful of anecdotal secondary figures, with no description of sampling, deflation, skill grade or how multiple conflicting observations were reconciled. Figure 3 and Table 1 are not clearly linked to their underlying sources, and the comparison with urban Manpower Survey wages is not like for like, so claims of widening disparities are fragile. Firm-level wage records, Carnegie survey microdata and homeland development corporation files exist in archives and could be used to build a documented, deflated panel.
Replicability: Data poorly documented.