South African Trade Union Growth 1979 to 1991¶
Frost, [first name not given in text] (1993). South African Trade Union Growth 1979 to 1991. Honours dissertation, Department of Sociology, University of Cape Town (research project initiated by Ian Macun). University of Cape Town.
| Type | thesis |
| Scope | Economics and economic history of apartheid |
| Method | mixed |
| Period studied | 1979–1991 |
| Regions | South Africa |
| Themes | unions strikes, labour markets wages, methods data sources, inequality poverty unemployment, manufacturing industrial policy, resistance social movements, comparative theory |
| Citation key | frost1993 |
Argument¶
The thesis asks why registered trade union membership in South Africa, especially in the Cosatu manufacturing unions, grew so fast after the 1979 Wiehahn reforms. It reviews international union-growth theory (business cycle, prosperity, price, structuralist and agency approaches) and the South African literature, which it finds narrow and focused on organisation and legislation. It audits the official and union membership statistics (Department of Manpower, CSS, union and federation sources) and shows they are unreliable. It then assembles density figures and argues that the 1979 legal reforms and organisational strength were 'facilitative' (parametric) factors, while worker demand for union protection, strike successes and wage aspirations were 'causal' ones. The author admits the explanatory analysis is tentative and incomplete.
Evidence and method¶
Department of Manpower annual reports, CSS South African Labour Statistics, NMC annual reports, Cosatu congress figures, SALDRU and other union directories, Andrew Levy reports, the CASE Cosatu shop steward survey, the LRS 'Organising Potential for Cosatu' document, and interviews (Oct 1992-Jan 1993) with officials of Cosatu, CWIU, FAWU, Nactu, SACOB, SEIFSA, the DOM Industrial Registrar and the CSS.
Place in the debates¶
- Causes of post-1979 union growth: legal reform vs organisation vs economic/political factors (see Black trade unions: growth, wage effects, Wiehahn and strategy). Holds that the 1979 legal reforms were necessary but not sufficient (facilitative). Organisation, worker need for protection and political struggle are also treated as important, and no single factor is prioritised.
- Applicability of international business-cycle models (Bain and Elsheikh) to South Africa (see Black trade unions: growth, wage effects, Wiehahn and strategy). Finds the cycle-based price and prosperity theories cannot explain growth amid structural unemployment and prolonged recession. Regression was dropped because the data are unreliable and short-run.
- Structure vs agency in union growth (see Black trade unions: growth, wage effects, Wiehahn and strategy). Favours a multi-causal, eclectic approach. Admits its own analysis is mostly structuralist, and treats agency and leadership tentatively.
- Economic ('bread-and-butter') vs political motives for joining unions (see Black trade unions: growth, wage effects, Wiehahn and strategy). Reports the debate (Baskin, Morris and Erwin, Maller, the CASE survey) as unresolved and ambiguous.
- Does union density measure union strength (Erwin and Morris)? (see Black trade unions: growth, wage effects, Wiehahn and strategy). Accepts density as a useful measure but notes the caveat that union power depends on strategic location, not just density.
Findings¶
-
Registered union membership rose sharply, with the 1979-91 increase quantified. (p. 53)
In 1991 the figured had leaped to 2 718 970. All in all an increase of 568 percent. Between 1979, the year of the Wiehahn reforms, and 1991 there was a 287.5 percent growth.
-
Registered union density rose steeply even as non-agricultural employment growth slowed. (p. 54)
the total (registered) union density rose dramatically from 15.37% to 52.92%.
-
Cosatu's share of registered membership and its density in non-agricultural employment rose. (p. 54)
Its share of the total registered 'cake' was 35% in 1985 and 46% in 1991
-
Cosatu manufacturing affiliates grew rapidly, with density among black manufacturing workers above 60% by 1991. (p. 55)
-
Official DOM union figures are not reliably verified and conflict with federation data. (p. 40)
There are clear problems with the validity of DOM figures for trade unions in South Africa.
-
Even the best-organised union (Numsa) struggled to prove its membership in the 1992 Seifsa litigation, which undermines confidence in union counts. (p. 43)
Numsa was forced to prove the accuracy of its membership records.
-
Cosatu officials said growth came from workers seeking out the unions, so organisation is facilitative, not directly causal. (p. 66)
the primary reason for the swelling of membership was a result of workers "knocking on the unions' doors" wanting to sign up.
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| Department of Manpower (DOM) Annual Reports (registered trade union membership returns, LR2/LR45 forms) | official statistical series | 1956-1991 | Source of registered and unregistered union membership totals, white-union membership, and Cosatu manufacturing-union figures used to compute growth and density. |
| Industrial Registrar membership returns forms LR1, LR2 and LR45 (Labour Relations Act 1956) | legislation gazette | 1989-1991 (current files) | Reproduced as annexures to explain how official membership figures are collected and verified. |
| CSS South African Labour Statistics (incl. monthly manufacturing employment survey and Manufacturing Census), 1992 and various years | official statistical series | 1973-1992 | Employment series (non-agricultural, manufacturing, food/beverage/tobacco, by race) as denominators for union density. |
| CSS South African Statistics and Statistical News Release | official statistical series | 1956-1990s | Supplementary official series (registered union membership pre-1979, CPI, GDP, wages) used in descriptive tables. |
| CSS Current Population Survey (CPS) unemployment and workers-by-industry data | household or labour survey microdata | 1980s | Discussed and rejected as unreliable for unemployment and employment-by-industry; used only as rough guide for unemployment trends. |
| National Manpower Commission (NMC) Annual Reports, incl. NMC 1982 and 1991 reports | commission or official report | 1982-1991 | Source of union membership figures (given primacy for 1990 Cosatu data) and statement on the 1979 extension of union rights. |
| Cosatu Congress membership figures (via Baskin 1992, SALB, Andrew Levy Annual Reports 1990/1991, SALDRU/SALORU Trade Union Directories, CASE Cosatu shop steward survey) | union or movement records | 1985-1992 | Primary basis for Cosatu and affiliate membership series, density and growth calculations, cross-checked against DOM figures. |
| CASE survey of Cosatu Shop Stewards (1992) | household or labour survey microdata | 1992 | Cited for shop steward attitudes to politics and as source of 1991 Cosatu membership figure. |
| Labour Research Service (1992) 'Organising Potential for Cosatu' | secondary literature compilation | 1992 | Shows sectoral overlap of union scopes and identifies poorly organised sectors with growth potential. |
| Author interviews with union, employer and state officials (Rani, Baskin, Mtayi, Bezuidenhout, Shortt, Martell, Skhosana, Nkosi, Arendse, Scheefer, de Jager), Oct 1992-Jan 1993 | interviews oral history | 1992-1993 | Evidence on membership collection methods, DOM procedures, recruitment dynamics and future growth areas. |
| Industrial Council agreements statistics (CSS Labour Statistics 1984, 1991, 1992) | official statistical series | 1970-1991 | Number of councils, agreements and employees covered, used to show decline then jump in coverage. |
| Moll's estimates of unemployment (economically active population without formal work) | secondary literature compilation | 1975-1988 | Unemployment series (25.8% 1975 to 42% 1988) used as rough economic-context variable. |
| Official South African Trade Unions Directory and Industrial Relations Handbook 1991; SALDRU Trade Union Directory | secondary literature compilation | 1985-1992 | Informal sources on union membership and legal definitions, e.g. the exclusion of essential-service workers from registration. |
| DOM statistics on strikes | official statistical series | 1980-1991 | Strike incidence and causes presented as indicators of worker militancy and wage demands. |
| Newspapers: Cape Times (19/10/1992, 27/11/1992) and Finance Week (Oct 1992) | newspapers periodicals | 1992 | Used for retrenchment impacts and regional recession effects as context. |
| Macun (1992) unpublished paper on trade union growth and the wider UCT research project | secondary literature compilation | 1992 | Parent research project from which the thesis data were drawn. |
New data¶
Author-assembled series (as research assistant on the UCT project) of registered union membership 1979-1991, Cosatu and Cosatu manufacturing-union membership 1985-1991, FAWU and CWIU series, and derived density figures against CSS non-agricultural, manufacturing and food-sector employment, with overall figures and unregistered estimates; compiled from DOM, Cosatu and other sources and interviews.
Inference¶
Quantification: Descriptive tables.
Descriptive time series of membership and density from official and union sources, set against international union-growth theory and the South African literature, plus interview impressions; regression modelling (Bain and Elsheikh style) was explicitly abandoned owing to data unreliability, and determinants are discussed qualitatively as 'facilitative' or 'causal'.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| Registered union membership grew from about 702,000 in 1979 to 2.72 million in 1991 (287.5%), and registered density against non-agricultural employment rose from 15% to 53%. (p. 53) | DOM Annual Reports membership returns divided by CSS Labour Statistics non-agricultural employment (Table B). | Systematic data |
| Cosatu manufacturing unions grew from about 148,000 members in 1985 to 673,000 in 1991, with density among black manufacturing workers reaching roughly 60%. (p. 55) | Cosatu congress/NMC/survey figures (Table E) over CSS manufacturing employment. | Partial data |
| Official DOM membership figures and union-reported figures are unreliable, so explanations of union growth can only be tentative. (p. 39) | Comparison of DOM vs Cosatu series (Table A, F), interviews with Scheefer and union officials, Numsa Supreme Court case. | Partial data |
| The 1979 legal reforms were a facilitative foundation but not a direct cause of growth; the organisational dimension is parametric rather than causal. (p. 54) | Literature review plus interviews with Cosatu officials reporting that workers 'knocked on the unions' doors'. | Illustrative or anecdotal |
| Union membership grew despite rising unemployment and stagnant manufacturing employment, contradicting cyclical (price/prosperity) theories. (p. 57) | Juxtaposition of DOM membership series, CSS manufacturing employment and Moll unemployment estimates; interview with Shortt. | Partial data |
| FAWU membership more than doubled while food-sector employment stayed constant, raising density to about 58% overall and 68% of black workers; CWIU grew at a markedly different rate. (p. 55) | FAWU membership (various sources) against CSS food, beverage and tobacco employment (Tables H, I). | Partial data |
| White union membership declined from over 350,000 in 1981-82 to about 230,000 by 1990. (p. 51) | DOM Annual Reports, white registered union membership (Table J). | Systematic data |
The author's caveats¶
Author states official and union membership data are unreliable and untested, TBVC areas are excluded from employment but possibly included in union counts, 1990 Cosatu figures were chosen by an 'entirely unsatisfactory' elimination method, union scopes overlap the SIC so density can only be done for FAWU, unregistered union estimates have unknown accuracy, and requests to federations and others were refused or ignored.
Gaps in the evidence¶
The argument lacks audited union membership records (DOM keeps financial audits closed), union-by-SIC membership breakdowns, and member-level surveys for testing why workers joined. Proper density calculations and the Bain-Elsheikh regressions would require reliable annual union and employment series including the TBVC areas; some union records and later CSS/census benchmarks exist now. The interview base is small and omits many employer and state respondents.
Replicability: Data described but not public.