The state and capital: instrumentalism, autonomy and fractions of capital¶
Was the South African state an instrument of capital (mining or national), symbiotic with it, or relatively autonomous and internally divided?
Works engaging the debate: 36
Positions¶
- Fractions-of-capital and hegemony analysis (Davies, Kaplan, O'Meara)
- State-mining symbiosis (Yudelman)
- Relative autonomy and divided bureaucracy (Duncan, Dubow)
- Pluralist interest-group and class-struggle critiques
Evidence base¶
The debate has been argued from three largely separate evidentiary traditions. The first is theory-led, and the second and third are archival. The Poulantzian periodisation of Davies et al. (1976) rests on aggregate and often unsourced figures. These include Chamber of Mines estimates of tariff and railway costs per ton milled, the Board of Trade and Industries report of 1945 on the 1925 Customs Tariff Act, poor-white estimates for 1916-33, white employment in private industry 1923/4-1932/3, wartime sectoral national income and 1948 seat outcomes, read through Hansard and legislation. Its critics argue from equally thin material. Innes and Plaut (1978) use unsourced union membership, strike counts and mining wage averages for 1919-32. Kantor and Kenny (1976) use no data, only secondary illustration (van der Horst, Wilson, the Keiskammahoek survey). Hackland (1980) extends the fractions model to the 1970s through SAIRR survey series, Progressive Party minutes and elite interviews with Oppenheimer, Eglin and Godsell. The relative-autonomy position is built from departmental files in the Central and Transvaal Archives Depots. Duncan (1989) reconstructs the 1944 milling strike from NTS 199/332, NTS 138/332, ARB LC1052/409 and police reports. Duncan (1990a) and Duncan (1990b) compare three inspectorates through MNW, GNLB, NTS 697/408C and commission evidence (Low Grade Mines, Mining Regulations, Mills, Native Economic). Dubow (1986) uses NTS, SDK (Public Service Commission), JUS and PM files plus Year Book budget shares of 1-2 per cent. Evans (1997) uses NTS, Chief Native Commissioners' conferences, NBRI housing-cost data and DNA annual reports for 1944-64. Posel (1990b) substitutes ASSOCOM and FCI non-European affairs files, Brakpan and WRAB municipal records and interviews for the then-closed BAD files. Morrell (1990) and Bradford (1990) work at district level from NTS, GNLB, JUS and LDB files, farmers' association minutes and NTS 2233 recruit schedules.
Contested evidence¶
The symbiosis tradition draws on Chamber of Mines, WNLA, NRC and TEBA records, Colonial Office files and the Zimbabwe and Botswana archives. Jeeves (1975a) and Jeeves (1975b) use Chamber Annual Reports for 1890-1906, SNA files and HC 17/86 employment returns (African mine employment of 107,756 in April 1905 and 90,420 in July 1906). Jeeves (1987) and Yudelman and Jeeves (1986a)/(Yudelman and Jeeves (1986b)) add an appended Chamber/WNLA series of labour origins for 1920-83, in which the South African share runs 58 per cent in 1936, 22 per cent in 1973 and about 60 per cent in 1985. Crush (1986b) uses TEBA aggregates showing the foreign share falling from 79.6 to 49.3 per cent between 1973 and 1977. Crush (1993) uses NRC district recruit counts for 1973-79. Quantitative tests are few. Nattrass (1990) decomposes manufacturing wage share, profit share and output-capital ratio from CSS and SARB national accounts for 1948-55 and 1955-60. Fine and Rustomjee (1996) reclassifies the 1984 and 1988 input-output tables, for example finding that 58.3 per cent of MEC inputs come from within the MEC. Bonner and Shapiro (1987) builds a TGME profit series for 1895-1932 from company annual reports. Sitas et al. (1984) combines the 1976 Manufacturing Census with a purposive survey of 14 Durban firms.
The sharpest disagreement concerns the low-grade ore argument. Jeeves (1975b) and Van-Helten (1980) accept the Chamber's case for cheap labour, citing Reyersbach's yield evidence, Frankel's ore-grade and return series, and Paish. Yudelman and Jeeves (1986a) argue that the Chamber's evidence to the 1919-20, 1922, 1930-32 and 1943 commissions was self-serving and misled historians. Duncan (1990b) asserts that dividends refute the cost constraint, but presents figures only for Simmer and Jack. The same Chamber/TEBA series supports competing readings of internalisation. Crush (1986b) and Yudelman and Jeeves (1986a) attribute the 1974-77 shift to the Malawi and Mozambique withdrawals rather than to state policy, against Clarke's internalisation thesis. On the 1975 farm-labour episode, (Yudelman and Jeeves (1986a)) reads the trilateral meetings as state-Chamber symbiosis. Crush (1993), using NRC files, finds the DBA and the farmers' unions blocking the mines and the state's public liberalisation contradicted in practice. Both work from Chamber-side records without DBA or Cabinet files. The NTS series also yields opposite pictures of the Native Affairs Department. It appears as protectionist and resistant to coercive tasks in Dubow (1986), as declining to serve as an instrument of farmers in Morrell (1990), and as allied with mineowners in Jeeves (1975a) (13 assault and 28 wage cases against thousands of disciplinary actions). In Bradford (1990) it is complicit in a pre-1948 farm labour 'slave trade', with about 75 per cent of over 18,000 northern entrants channelled to farms by 1945. On national-capital hegemony, (Davies et al. (1976))'s 1924 break is undercut by Bozzoli (1975), whose manufacturers' journals show protectionist lobbying two decades earlier. Freund (2012)'s continuity from the Smuts era and Fine and Rustomjee (1996)'s claim that MEC sectors industrialised early also cut against it. Innes and Plaut (1978) reads the same 1920s labour figures as evidence of white workers' defeat rather than alliance. On business and the apartheid state, (Posel (1990b)) shows organised business defeating repeal of section 10 but not the quota controls. Freund (1987a) and Freund (1987b) rebut Lipton with Webster's foundry poll and Bernstein's finding that 45 per cent of managers favoured some form of influx control. Nattrass (1990) finds the profit share rising in 1948-55 while the manufacturing profit rate fell throughout 1948-60, which complicates both functionalist and liberal readings.
Underused and missing sources¶
Several kinds of evidence that would settle these disputes are cited but seldom exploited systematically. Mine-level cost, grade, wage and dividend records would test the low-grade thesis and Duncan's dividend claim directly. These include the Barlow Rand/Rand Mines archives (Eckstein, CMIC), Chamber of Mines internal minutes and WNLA statistics. Only (Van-Helten (1980)), (Jeeves (1987)) and Bonner and Shapiro (1987) touch them, and selectively. The state side of the 1960s-70s is largely missing: BAD, DBA, Cabinet and Mines files, which (Posel (1990b)) and Crush (1993) could not use. The full minutes of evidence of the Low Grade Mines Commission of 1919-20 (used by 6 works), the Lansdown Commission (3 works) and the Fagan Commission (4 works) are mostly mined for quotations rather than series. Others are rarely read as continuous series: Board of Trade and Industries tariff application records (3 works), Hansard division lists (Hansard is used by 9 works, but only for speeches), FCI, AHI and SAAU submissions, and Broederbond membership records. These are what (Davies et al. (1976))'s hegemony claims and the fractions debate require. Systematic inspectorate prosecution and accident series, labour-bureau placement registers, and DNL files after 1935 would test the relative-autonomy claims beyond episodes.
Strength of key claims¶
| Basis | Share of key claims |
|---|---|
| Systematic data | 8% |
| Partial data | 52% |
| Illustrative or anecdotal | 28% |
| Asserted | 12% |
Comparative figures for all debates are on the Literature overview.
Best-evidenced works¶
Yudelman and Jeeves (1986a), Nattrass (1990), Posel (1990b), Dubow (1986), Jeeves (1975b), Fine and Rustomjee (1996)
Most-used primary sources¶
Holdings are matched by machine against the archive's file manifest; a dash means no match was found, not that the project certainly lacks the source.
Works and their positions¶
| Work | Position |
|---|---|
| Beinart (1982) | Argues the state must be disaggregated. Local Transkeian officials sometimes diverged from central departments that served mining and agrarian capital. |
| Bond (1998) | Works within a radical political-economy and Harvey-style framework. He draws on Innes and O'Meara to portray centralisation and monopoly finance capital, and the rise of Afrikaner financial capital, as shaping the state and nationalist politics. |
| Bonner and Shapiro (1987) | Argues the debate neglects the diversity of mining. Coal, Rhodesian and eastern Transvaal gold producers had sharper profitability problems and conflicts of their own, partly created by the pull of the Rand. |
| Bozzoli (1975) | Argues that analysis must go beyond economic determinants. Ideologies are the means by which capitalists relate to and gain access to the state, and the political and economic processes should be kept distinct. |
| Bozzoli (1975) | Mining and finance opposed protection before the war. The decline of financial dominance in British imperialism during and after the war removed this opposition, though she calls the explanation partial and tentative. |
| Bradford (1990) | Rejects the dichotomy. The NAD depressed foreign workers' wages in the 1920s and sanctioned a 'slave trade' in the 1930s and 40s, so paternalism is presented as a thin veneer. |
| Bradford (1990) | Holds that the state's 'relative autonomy' was continually redefined under crisis. Endorses Yudelman that the South African state must be periodised differently, with the rhythms of accumulation and struggle as ultimate determinants. |
| Coupe (1990) | Stresses an uneasy relationship between organised industry and the Nationalist government under Strydom and Verwoerd. Industry objected to influx control, job reservation and decentralisation as impediments to productivity. This friction faded after Sharpeville, as industry retreated into discreet lobbying. |
| Crush (1986b) | Challenges the presumption that the South African state and the mining industry shared identical interests on foreign labour. |
| Crush (1993) | The state was 'hamstrung' by farmer pressure on this issue. Internal bureaucratic divisions were shallower than on foreign labour, where Foreign Affairs and Native Affairs split. |
| Davies et al. (1976) | Argues that national (agricultural and manufacturing) capital held hegemony from 1924, and that the state taxed gold to subsidise it, in contrast to accounts of a state subordinate to mining capital. |
| Davies et al. (1976) | Rejects it: 'the workers' state (even the white workers' state) was a myth'. White wages did not recover and the gains were mainly employment under the civilised labour policy. |
| Dubow (1986) | Argues against a static functionalism in which the state acts as 'a sentient being organising capital's hegemony'. The NAD, though receptive to capital, had its own internal ideology and interests, and segregation was modified by struggles within the state. The paper claims to be a complement to, not a replacement for, existing explanations. |
| Dubow (1986) | Doubts Kallaway's view that Malan sustained Cape liberal policy in the NAD after 1910. Dubow locates the continuity within the Department's own dynamics and the Transkeian tradition. |
| Duncan (1989) | Criticises Yudelman, Davies and Johnstone for treating the state as overly unified; stresses departmental divisions and sectional interests, while accepting that the state broadly acted to protect the long-term interests of capital. |
| Duncan (1989) | Accepts Moodie's point on departmental differences but argues he overstates state unity; the coal and milling strikes show departments at odds with each other, employers and unions. |
| Duncan (1990a) | Accepts that the state generally served capital's long-term interests but argues officials had their own interests and perceptions. Civil servants could side with white workers, with a sub-section of the state, or with African workers without 'bringing down the whole edifice'. |
| Duncan (1990a) | Endorses Yudelman's account of mine politics and the role of Kotze and Pirow. Criticises his tendency to judge officials by how well they grasped his 'symbiotic' theory, and adds the social background and attitudes of administrators as explanatory factors. |
| Duncan (1990b) | Accepts that the state served capital's long-term interests but argues bureaucrats had relative autonomy. Their own institutional interests, professional backgrounds and racial attitudes shaped policy, so branches of the state could respond to other interests without the state 'bringing down the whole edifice'. |
| Duncan (1990b) | Agrees with Yudelman's account of Kotze and Pirow's role in mine politics, but says Yudelman judges them by how well they fit his symbiosis theory. Duncan adds social background and perceptions of constituency to explain why mines and factory inspectors behaved differently. |
| Evans (1997) | Adopts a 'weaker' thesis of state autonomy (Block, Skocpol, Polanyi): state managers had leeway but within constraints from farmers, miners and industrialists, with bureaucracy instrumental rather than self-referential, unlike Nazi Germany. |
| Fine and Rustomjee (1996) | Rejects a simple zero-sum reading of Afrikaner national capital against imperial/mining capital. Argues for interdependence, a disjuncture between English economic and Afrikaner political power that blocked diversification out of mining, and sector-by-sector variation. |
| Freund (1987a) | Criticises Lipton for lacking a theory of the state. He sees business-state friction as secondary to a basic Nationalist commitment to capitalist accumulation and business loyalty to the state. |
| Freund (1987b) | Argues Lipton has a weak concept of the capitalist state. Hegemony (Gramsci) better explains state-business tensions, which were secondary to the Nationalist government's basic sympathy for capitalist accumulation. |
| Freund (2012) | The state consulted English and Afrikaner business but did not plan with it; business was divided, and by the 1980s wanted release from the security state. |
| Hackland (1980) | Argues that parties express alliances of capitalists. The NP supplied the conditions for accumulation and monopoly capital's rise, but monopoly capital hedged by backing the PFP when the NP's ability to manage the crisis came into doubt. |
| Hirson (1988) | Hirson criticises Stadler for asserting that the state-capital relationship is crucial without defining the state or demonstrating the claim. He implicitly calls for explicit conceptual grounding. |
| Innes and Plaut (1978) | Rejects the fractions/hegemony framework as pluralist and economistic. Insists the state must be located in the capital-labour struggle with black workers at the centre. |
| Innes and Plaut (1978) | Denies a class alliance. Sees the Pact as an assault on and neutralisation of white workers combined with co-option of their leaders. |
| Jeeves (1975a) | Takes a middle course. The state and the mines formed a close working alliance, yet the mines were dependent on government and divided among themselves rather than dominant. He argues Lagden's department often deferred to the Chamber and WNLA. |
| Jeeves (1975b) | Rejects both simple capitalist domination (Denoon) and Mawby's view that Milner dominated the mines. Argues for a complex, variable alliance in which the mines depended on the state. |
| Jeeves (1975b) | Holds that the industrial policies produced by this collaboration were of demonstrable importance for South Africa's subsequent history. |
| Jeeves (1987) | Argues that bureaucratic officials like Gemmill, with long tenure and close ties to the state bureaucracy, decisively shaped policy. The Chamber's power was real but limited by farmers, the Union government and neighbouring governments. Extends Yudelman's focus on white labour to black labour strategy. |
| Kantor and Kenny (1976) | The state mediates a plurality of interests, mostly white, and pursues political goals. Class analysis is therefore of limited use. This is a pluralist critique of the instrumentalist view of the state. |
| Morrell (1988) | Rejects the Afrikaner nationalist, liberal and revisionist portrayals of mining and farming as simply antagonistic. Argues that progressive maize farmers were allied with the mines, and that the alliance later dissolved for economic reasons. |
| Morrell (1988) | Notes the complexity of agriculture's relationship to the state. Cites Bozzoli's view that national capital captured hegemony after the gold standard crisis, and argues that the alliance's end was gradual, not a product of the 1924 election alone. |
| Morrell (1990) | Argues the state was responsive to white farmer pressure and incapacity-ridden, so it drew farmers into its apparatus. Yet the NAD retained some autonomy and often frustrated farmers. |
| Nattrass (1990) | Rejects this. Government macro policy (deflation, import controls, exchange-rate parity with sterling, hostility to foreign capital) harmed profitability, so the interests of capitalists and of capitalism diverge. |
| Pirie (1993) | Qualifies the view of smooth collaboration: state-owned railways had interests not wholly aligned with the mines, and conflict ran over tariffs, scheduling and cost recovery. |
| Posel (1990b) | Rejects simplistic formulations; the power of manufacturing and commercial capital was uneven, and policy reflected continuing conflict and compromise rather than one set of interests. |
| Posel (1990b) | Stresses the BAD bureaucracy's agenda-setting power and Broederbond influence, but shows these were checked by business and local authorities. |
| Schneider (2000) | Rejects the view that Union meant simple triumph of local over imperial interests (Ehrensaft, Davies et al.) and describes a negotiated compromise in which mining kept cheap labour, low tariffs and cheap electricity. |
| Sitas et al. (1984) | Radical: the state created the conditions for capitalism through racial labour controls, and state corporations, tariffs and the IDC promoted concentration. The liberal free-market invisible hand is contested. |
| Stott (2008) | Largely preservative: the formula tax, lease system and later subsidies prolonged mine lives and the industry still prospered. Feinstein's view that taxes were discriminatory and exploitative is acknowledged, and some policy (e.g. Hertzog's gold standard stance) is treated as self-serving. |
| Stott (2008) | Sides with Yudelman and Fine and Rustomjee: a symbiotic relationship in which the state viewed mining as its instrument, and mining houses had strong influence via the Chamber of Mines and the allocation of property rights. |
| Stott (2008) | Rejects both halves. Mining houses were profit-maximising, 'perfectly competitive' firms who were both exploitative and preservative, and the state lacked information and sometimes served voters or itself. |
| Van-Helten (1980) | Holds that the mines were confident, assertive lobbyists who needed state support for labour supply and cost reduction. He rejects Kubicek's picture of an industry vulnerable or unsure in dealing with government. |
| Yudelman and Jeeves (1986a) | Bureaucratic divisions (Native/Bantu Administration versus Foreign Affairs) existed, but state and capital were fundamentally symbiotic and mutually dependent in a crisis. |
| Yudelman and Jeeves (1986b) | Describes a 'symbiotic' relationship with internal bureaucratic divisions (Bantu Administration vs foreign affairs), with the state reliant on industry data and deferring to the Chamber. |