Institutions, Wage Differentiation and the Structure of Employment in South Africa¶
Nicoli Nattrass; Jeremy Seekings (2012). Institutions, Wage Differentiation and the Structure of Employment in South Africa. CSSR Working Paper No. 309, Centre for Social Science Research, Sustainable Societies Unit. University of Cape Town.
| Type | working paper |
| Scope | Wider apartheid literature |
| Method | synthesis review |
| Period studied | 1924–2011 |
| Regions | South Africa, Cape Town, Johannesburg, Newcastle/northern KwaZulu-Natal, bantustan industrial areas |
| Themes | labour markets wages, unions strikes, manufacturing industrial policy, origins segregation, bantustans removals influx, inequality poverty unemployment, transition post apartheid |
| Citation key | nattrass2012 |
Argument¶
Bargaining councils, which descend from the 1924 industrial councils designed to protect 'civilised' white labour, keep excluding low-wage workers from regulated employment even though they are now deracialised. The authors argue that national councils, by extending uniform minimum wages, erode the regional wage differentiation on which labour-intensive industries such as clothing depended. This pushes firms toward capital- and skill-intensive production or out of business, so low-wage jobs are destroyed. They claim that aggregate wage and employment regressions and coverage estimates (Godfrey et al., Bhorat et al.) miss this effect. Well-organised employers and unions can pursue their own interests at the expense of the less organised, justified by an ideology of 'decent work'. The argument is made through a history of the councils, a clothing-industry case study, and a critique of defenders' arguments and the econometric literature.
Evidence and method¶
Secondary literature and government commissions (Economic and Wage Commission 1926; Barker 1962; Altman 1994; Godfrey 1997; Godfrey et al. 2010; Anstey 2004), Labour Force Survey figures, bargaining council compliance data, union membership figures, and minimum-wage comparisons across regions. It also uses econometric studies by Butcher and Rouse and by Bhorat et al., and Pollin et al. on labour intensity.
Place in the debates¶
- Effect of bargaining councils and extensions on wages and employment (post-apartheid labour market flexibility) (see Unemployment: causes, measurement and labour market institutions). Nattrass and Seekings argue that councils and extensions destroy low-wage, labour-intensive jobs and that aggregate studies understate this. They reject Godfrey et al. (negligible employment effect), Altman and Valodia, and Roskam.
- Whether high labour costs explain mass unemployment (see Unemployment: causes, measurement and labour market institutions). They criticise Pollin et al. for dismissing the link, arguing that reform would create new low-wage jobs rather than cut existing wages. They also hold that high wages for less skilled workers exacerbate poverty and inequality.
- Role of institutional continuity from apartheid-era racial wage protection (see Unemployment: causes, measurement and labour market institutions). They see continuity: the industrial council system of 'civilised wages' was reincarnated by COSATU unions as 'living wage' and 'decent work' bargaining.
Findings¶
-
Industrial councils were set up to protect high-wage skilled white workers from undercutting by lower-wage workers, and the 1930 amendment extended agreements to African workers to prevent their replacing white workers. (p. 9)
In 1930, the Act was amended to allow collective agreements and their extensions to cover African workers also, explicitly to prevent high-wage white (and coloured) workers being replaced by low-wage African workers.
-
Skilled white South African wages were high by international standards, while unskilled African wages were comparable to those of labourers in parts of Europe. (p. 10)
Skilled (white) workers and artisans in South Africa were better paid than their British, American or Australian counterparts, and much better paid than in most of Europe, even taking the cost of living into account.
-
Clothing employment grew in the apartheid decades because wage differentiation allowed a low-wage sub-sector to coexist with a higher-wage one, with bantustan and uncontrolled areas used as low-wage locations. (p. 13)
The industry grew to employ as many as 170,000 workers through combining a higher-wage, higher-productivity sub-sector, located first in Johannesburg and later in Cape Town, with a lower-wage, lower-productivity sub-sector
-
From the 1960s the apartheid state used lower wages and subsidies to attract firms to bantustan industrial areas, exempting them from Wage Board determinations. (p. 12)
the apartheid state sought to promote industrialisation and employment in the bantustans (Glaser, 1987), and used lower wages as an incentive to firms to relocate to industrial areas within the bantustans
-
In the 1980s the new independent (FOSATU/COSATU) unions moved from hostility to embracing industrial councils as vehicles for industry-wide minimum wages and a 'living wage'. (p. 14)
The discourse of a „civilised‟ wage was thus reincarnated via the very same institutions through which it had been effected in a racialised form half a century earlier.
-
Extending minimum wages across whole sectors blocked small low-wage labour-intensive firms and thus raised capital intensity, feeding the unemployment that threatened wages. (p. 15)
The extension of minimum wages across whole sectors had the effect of preventing the survival or emergence of smaller, low-wage, labour-intensive enterprises, and thereby contributed to the rising capital-intensity of production
-
Regional clothing minimum-wage differentials narrowed sharply after the national council was established. (p. 17)
Through the second half of the twentieth century, minimum wages in low-wage areas were about one-third of the minimum wages in the highest-wage area (initially Johannesburg, and later Cape Town).
-
Econometric studies find a bargaining council wage premium, larger for unionised African workers, but these capture only surviving workers, not those who lose jobs. (p. 21)
The premium associated with being covered by a bargaining council was 9-10 percent, and this rose to 25-30 percent for workers who were also members of a trade union
-
Godfrey et al.'s estimate that extensions cost only about 50,000 jobs is criticised for understating coverage and ignoring the counterfactual. (p. 23)
Godfrey et al. underestimate the scope of extensions, and they use estimates of the employment elasticity of wages that cannot take into account the employment that would have existed if lower wages had been possible.
-
Labour intensity of formal production has declined over the long term (citing Pollin et al.). (p. 24)
Between 1994-2001, the number of workers utilized per unit of output – i.e. a basic measure of labor intensity – fell by an average of nearly four percent per year
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| Statistics South Africa, Manufacturing Industry: Financial Report, no. 30-20-30 (2008, published 2010) | official statistical series | 2008 | Establishes the capital:labour ratios per job in manufacturing, in textiles/clothing/footwear and in apparel. |
| Altman (1994), An Industrial Strategy for the Clothing Sector (employment figures) | secondary literature compilation | 1950s-1991 | Provides clothing employment totals over time and the 1989 living-wage comparison with the machinist minimum wage. |
| Barker (1962), The Economics of the Wholesale Clothing Industry of South Africa, 1907-1957 | secondary literature compilation | 1907-1957 | Source for regional wage differentials, industrial council concessions, and the characterisation of lower-wage firms' product markets. |
| Report of the Economic and Wage Commission (1925), U.G. 14 of 1926 | commission or official report | 1925-1926 | Documents high skilled white wages, the 'civilised wage' rationale, job losses from extending wages to small towns, and criticism of extensions. |
| Industrial Conciliation Acts of 1924, 1930 and 1937, 1925 Wage Act, 1923 Natives (Urban Areas) Act, 1911 Native Labour Regulation Act | legislation gazette | 1911-1937 | Establishes the legal origins of industrial councils, extensions, and exclusion/inclusion of African workers. |
| Wage Board wage determinations (WD 221 of 1961, WD 471) and ECC sectoral determination no. 4 (2000) | legislation gazette | 1961-2000 | Documents the regulation of clothing wages outside industrial council areas. |
| Godfrey (1997), unpublished MA thesis on Western Cape clothing employers (interviews 1991-1995) | interviews oral history | 1991-1995 | Evidence of employers' ambivalence about industrial councils and that only a quarter favoured national bargaining. |
| Anstey (2004), account of NBC establishment and compliance (incl. NBC compliance officer investigations) | secondary literature compilation | 2001-2004 | Source for the negotiation of the NBC, and firm non-compliance and investigation numbers in 2002-04. |
| Godfrey et al. (2010), Collective Bargaining in South Africa: Past, Present and Future? and Godfrey et al. (2006) bargaining council studies | secondary literature compilation | 2004-2010 | Uses their estimates of bargaining council registration, extension coverage, compliance and employment effects, which are then critiqued. |
| Labour Force Survey (LFS), Statistics South Africa, 2004 | household or labour survey microdata | 2004 | Used to show at least 130,000 clothing workers should have been registered, versus under 100,000 registered, so extension coverage is understated. |
| Bhorat, Goga and van der Westhuizen (2009, 2011) and Butcher and Rouse (2001) wage premium regressions | secondary literature compilation | 1995 and 2005 data | Provides estimates of 9-10 percent bargaining council wage premia (25-30 percent with union membership), then critiqued for aggregation. |
| Pollin, Epstein, Heintz and Ndikumana (2006), An Employment-Targeted Economic Programme for South Africa | secondary literature compilation | 1967-2001 | Supplies data on declining labour intensity (employment per unit output) in the formal economy. |
| Union and NBC records on clothing union membership and bargaining council registration (SACTWU, NBC website) | union or movement records | 1991-2011 | Evidence for membership decline and the number of workers registered with the NBC. |
| NBC clothing industry exemption applications 2008 and collective agreements/minimum wage rates | employer association records | 1998-2010 | Underpins claims about minimum wage rates by region and that wage-related exemptions are almost never granted. |
| Nattrass and Seekings (2012), Differentiation within the South African Clothing Industry (CSSR working paper) | author survey or fieldwork | c. 2012 | Companion study cited for the claims that party firms are more capital/skill-intensive and that retailers import cheap clothing. |
Inference¶
Quantification: Descriptive tables.
Historical-institutional case study of the clothing industry combined with a critical review of existing econometric and descriptive studies of bargaining council effects; claims are argued through narrative, secondary data and logical critique of others' aggregate regressions, with no original statistical estimation.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| Clothing employment declined from about 150-170,000 in the 1980s-90s to around 100,000 in 2010, with only 56,000 registered with the NBC, driven by squeezed margins from imports and rising labour costs. (p. 2) | Altman (1994), Godfrey (1997) figures, SACTWU and NBC membership/registration data; no systematic series presented. | Partial data |
| Regional wage differentiation sustained the clothing industry's growth through the apartheid era; minimum wages in low-wage areas were about one-third of the highest, and this gap narrowed sharply after 2002. (p. 13) | Barker (1962) for 1930s-1950s differentials; NBC minimum wage rates for the post-2002 narrowing; stated without tables. | Partial data |
| Industrial councils and extensions originated as devices to protect high-wage white workers from lower-wage competition, and extending wages to small towns caused job losses. (p. 8) | 1924, 1930 and 1937 Acts, Economic and Wage Commission 1926 report, Van der Horst (1942). | Partial data |
| The NBC and national extensions, driven by SACTWU and Cape-based employers with ministerial support, raised minimum wages in low-wage regions (e.g. Newcastle doubled in real terms) and caused closures and job destruction in labour-intensive production. (p. 21) | NBC wage rates, Anstey (2004) compliance data, Godfrey et al. (2010), press-style account of the 2010-11 compliance drive; no counterfactual or employment regression. | Illustrative or anecdotal |
| Studies finding that bargaining councils have limited reach and negligible employment effects understate the scope of extensions because they omit unregistered employees and ignore the counterfactual. (p. 16) | LFS 2004 clothing employment versus NBC registration; critique of Godfrey et al. (2006, 2010) and Bhorat et al. | Partial data |
| Wage-related exemptions are almost never granted in the clothing NBC, so the exemption system is not an effective safety valve. (p. 19) | Authors' comparison of 2008 exemption applications with NBC reasons; Godfrey et al. (2006) on few employees covered by exemptions; Nattrass (2000). | Partial data |
| Formal production in South Africa has become less labour-intensive, with employers replacing un- and semi-skilled labour by capital and skills, which the authors link to the cost of low-skilled labour. (p. 22) | Pollin et al. (2006) labour intensity data; link to wage-setting institutions is argued rather than tested. | Asserted |
The author's caveats¶
The authors note that the evidence on bargaining council effects was weak and dated until recently, that Godfrey et al.'s estimates rely on registered employees only, that counterfactual employment is probably impossible to establish, and that the clothing NBC is atypical. The text contains unfinished placeholders for references ("Refs: ??", "Ref?", "xxxx").
Gaps in the evidence¶
The argument lacks firm-level or establishment-level panel data on clothing employment, wages, compliance and closures by region, which would permit testing the wage-differentiation and job-destruction mechanism against import competition and exchange rate effects. Disaggregated NBC wage schedules and registration/compliance records, which exist, are cited but not presented systematically. Quantile or sector-level regressions of low-skilled employment on bargaining council coverage using the LFS and QLFS would better test the claim of effects hidden in aggregate analyses.
Replicability: Data poorly documented.