Why apartheid ended: reform, transition and elite bargaining¶
Was late-apartheid reform genuine or a restructuring, and was the transition driven by internal struggle, economic decline, external pressure or elite calculation?
Works engaging the debate: 46
Positions¶
- Reform as restructuring or co-optation, not dismantling
- Internal resistance decisive (Zunes, Hirson)
- Economic decline and cost-benefit elite calculation (Van Wyk, Lowenberg, Inman and Rubinfeld)
- Overdetermined, with the end of the Cold War as a factor (Levy); business as reform agent vs sceptics
Evidence base¶
The debate has mostly been argued from secondary syntheses, contemporary press and official policy texts rather than from assembled primary series. The sanctions and economic-decline literature rests on a small set of recycled compilations. Levy (1999a) and Levy (1999b) use Jones and Müller for GDP growth and current accounts, Lipton (1989), Lewis (1990), Waldmeir (1997) and Hufbauer, Schott and Elliott's cost estimate of about 0.5 per cent of GNP. Lowenberg (1997) adds Lundahl, Davis, Mohr and T. Moll's cross-country growth comparison. Van Wyk (2005) draws on the SARB balance-of-payments supplement 1946-2000, McCarthy and Moll, alongside memoirs, ANC documents, Leadership SA and HSRC/Schlemmer surveys. Bowman (n.d.) adds national-level Penn World Table 6.2 comparisons with the US and Ireland. Steward (2004b) gives unsourced macro figures from a former insider. The reform-as-restructuring position is built from close readings of the Riekert and Wiehahn reports, White Papers, the Good Hope Plan and the SECOSAF incentive manual (Pickles (1988a)) (Pickles (1988b)). It also uses Afrikaans business periodicals (Volkshandel), AHI memoranda and un-sampled interviews with labour-bureau and policy officials (Greenberg (1983)). Greenberg (1987) adds interviews at TEBA and bantustan labour offices in 1982, set against compiled unemployment estimates (Simkins, Loots, van der Merwe's use of the 1970 census). The internal-resistance case (Zunes (1999)) relies on movement histories (Frederikse, Smuts and Westcott, Price) and illustrative strike figures. Swilling (1985a) and Swilling (1985b) contribute the only original survey material on mass action: Labour Monitoring Group employer surveys of 71 unionised Transvaal factories and of Port Elizabeth-Uitenhage firms, at the establishment level, reported as participation percentages.
Works with systematic microdata or archives are few. Mariotti (2008) and Mariotti (2009a) use 1970 and 1980 population-census micro-data at the individual level. They combine these with biennial Manpower Surveys 1969-1985 covering 271 occupations by race and with an industry panel from the Manufacturing Census 1950-1985. Methods include SEO-ranked occupational indices, OLS and an Oaxaca-Blinder decomposition. Stemmet (2002) is the main archival political history. It draws on the P.W. Botha, Le Grange and Heunis collections at INCH, the ANC Lusaka and Tambo papers at Fort Hare, Hansard 1985-87, government unrest statistics and elite interviews. Charney (1984) cross-tabulates Rapport and Star Retail Data Library polls of 1982 with general and by-election returns for 1977-83. Sisk (1989) and Engel (1999) compile Official Yearbook election series, and Spence (1981a) gives 1977-81 HNP and PFP vote totals. Torchia (1988) reads FCI annual reports and Industry and Trade over 1960-86, and Verhoef (2009) uses Sanlam, FVB and Sankorp board minutes. The Inman and Rubinfeld papers (Inman and Rubinfeld (2005)) (Inman and Rubinfeld (2008)) (Inman and Rubinfeld (2013)) are the only formal causal work. They calibrate a federal hostage game to Financial and Fiscal Commission and Division of Revenue provincial allocations, and run regime-dummy regressions on annual PWT GDP per capita for 1950-2008, about 59 observations.
Contested evidence¶
The sharpest disagreement is over the timing of economic decline relative to sanctions. Levy (1999a) (Levy (1999b)) and Lowenberg (1997) read the Jones and Müller growth series as breaking in 1974, well before the 1985-87 sanctions, and so treat sanctions as consequence rather than cause. Bowman (n.d.) uses PWT relative-income ratios to argue that South Africa's divergence from a global boom happened after 1985 and that the 1970s labour reforms cannot explain it. Inman and Rubinfeld (2008) and Inman and Rubinfeld (2013) attribute a late-apartheid growth rate of about -0.8 per cent to sanctions and COSATU dummies. That estimate is identified from a handful of annual observations that coincide with each other, so sanctions and union power cannot be separated.
The August 1985 Chase Manhattan rollover refusal is read three ways. For Levy (1999a) (Lowenberg (1997)) (Steward (2004b)) it was a private market reaction to domestic risk and debt structure, with external debt of about 35 per cent of GNP. For Bowman (n.d.) it was induced by the June 1985 House vote. For Stemmet (2002) it was a response to repression and the Rubicon speech. All three readings rest on the same Waldmeir quotations and chronology, not on lender records.
Labour-market reform is attributed to incompatible sources. In Mariotti (2008) and Mariotti (2009a), census and Manpower Survey evidence shows semi-skilled desegregation under way from the late 1960s, before the 1973 strikes. Mariotti reads this as white workers protecting skill premia, but the political mechanism is inferred, not documented. Zunes (1999) and Hirschsohn (1996) treat black labour militancy and post-Wiehahn bargaining as causal. Torchia (1988) reads FCI statements as cost-driven manufacturer reformism. Freund (1987a) and Freund (1987b) counter Lipton's case for business reformism with Webster's 1980 foundry-manager poll and Bernstein's Urban Foundation finding that 45 per cent of managers preferred some influx control.
Whether reform was restructuring or dismantling turns on how the same documents are read. Pickles (1988a) and Greenberg (1983) read Riekert and the orderly-urbanisation White Paper as retaining control by new means. Pickles's one outcome statistic, a decentralised manufacturing share rising from 12.9 to 19.3 per cent over 1978-84, is borrowed from Wellings and Black. The white electoral evidence is also read divergently. Spence (1981a) and Sisk (1989) treat the HNP rise from 34,161 to 192,214 votes, and the fall in the Afrikaner NP vote from 85 to 63 per cent, as a constraint on reform. Charney (1984) uses polls to argue that class rather than language structured the split. Strauss (1993) and Van Wyk (2005) read the 1992 referendum as white consent to settlement, but without constituency returns or exit polls.
Underused and missing sources¶
Several primary series that could settle the timing disputes are almost never used. These include monthly or quarterly SARB balance-of-payments, capital-account, financial-rand and debt-standstill records, and bank-level lending exposure for 1984-86. Commodity-level trade flows that would test circumvention of sanctions are also missing. Others are Department of Manpower strike and stoppage series, influx-control arrest and prosecution statistics, and labour-bureau registrations. Every evidence ledger flags these gaps, and none supplies them except via secondary quotation; even (Stemmet (2002)) lacks the macro series. Cabinet, State Security Council and National Party caucus records are effectively absent, apart from Botha's private papers in Stemmet (2002) and the one Broederbond document in Giliomee (1995). Van Vuuren (2006) documents the destruction of 44 tonnes of NIS records, which limits recovery. Industrial Council and Industrial Tribunal determinations, which Mariotti defers, are also unused, as are FCI, Assocom and Chamber of Mines internal minutes, IMSSA case files and the ELTSA archive noted in Effect of Sanctions on South... (n.d.).
Strength of key claims¶
| Basis | Share of key claims |
|---|---|
| Systematic data | 6% |
| Partial data | 42% |
| Illustrative or anecdotal | 24% |
| Asserted | 28% |
Comparative figures for all debates are on the Literature overview.
Best-evidenced works¶
Mariotti (2008), Mariotti (2009a), Stemmet (2002), Charney (1984), Inman and Rubinfeld (2013), Swilling (1985b)
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 |
|---|---|
| Bowman (n.d.) | Rejects the Lowenberg-Levy 'internal forces' thesis as too reductive and economistic. It gives primacy to international finance and the political isolation of the country. |
| Bowman (n.d.) | Treats the fall of the USSR as the 'final factor' pushing the government to the table, following Levy and Waldmeir, but subordinate to sanctions. |
| Charney (1984) | Revisionist. Economic liberalisation serves capital's need for skilled, organised black labour, but the Afrikaner bourgeoisie's reliance on the state rules out political liberalisation. |
| Charney (1984) | Denies it. The NP offers authoritarian reform with white control of the state, not majority rule. |
| Charney (1984) | Rejects it. The split freed the NP to pursue capital's restructuring agenda. |
| Cronin (1994) | Rejects elite pacting and 'low intensity democracy' as demobilising. Argues that mass action and negotiation reinforced each other and that the transition should be mass-driven. |
| Engel (1999) | Describes the removal of apartheid as an inter-linked process of internal revolt, failed controlled reform from above and international pressure, without privileging any single factor. |
| Fine (1992) | He welcomes civil society theory as a counterweight to etatism and warns against elite-pacting. He argues it is deficient without democratised political parties, and he opposes a government of 'national unity'. |
| Fine (1992) | He calls it a myth that reform was a clear-cut victory. The state survived and quelled the uprising, and De Klerk revived a corporatist reform strategy. |
| Freund (1987a) | Sceptical. He thinks most capitalists backed the Botha reforms, fear mass democratisation and majority rule, and will not break with the existing political order. |
| Freund (1987b) | Sceptical. Most business backed Botha until 1985, was equivocal on the Emergency, feared majoritarianism, and liked trade unions only when the alternative was disruption. |
| Giliomee (1995) | Mainly an elite initiative. The regime was not forced to negotiate by revolutionary pressure, since the security forces remained coherent and the 1986 State of Emergency set back the extraparliamentary movement. Economic, demographic and external pressures shaped the elite's choice. This is set against Price's view of a state forced by struggle. |
| Giliomee (1995) | Treats them as complementary, not exclusive, following Lipset. Pacts are made within structural constraints. |
| Greenberg (1983) | Builds on Saul and Gelb's 'organic crisis' reading but shifts attention to the content and breakdown of dominant ideology, and to struggles among state actors and their outside connections. |
| Greenberg (1983) | Treats market-liberal discourse as partly an ideological mystification that obscures state-market entanglement, not a real retreat of the state. Officials continue to expect more regulation. |
| Greenberg (1983) | Sees employer and business pressure, especially Afrikaner business, as the main impetus for the Wiehahn and Riekert reforms. Lower-level bureaucrats are isolated from this alignment. |
| Greenberg (1987) | Argues that capitalist farmers, mining houses and manufacturers, once champions of control, have withdrawn support and pushed for reform as their labour requirements changed. |
| Hepple (2006) | He holds that the AAM and IDAF played a secondary or complementary role, rejecting the apartheid state's claim that they instigated rebellion. |
| Hirschsohn (1996) | Rejects a story centred only on Mandela and De Klerk; locates the origins in the unintended consequences of the 1979 labour reforms and the negotiation culture of industrial relations, while granting that leadership mattered. |
| Hirson (1988) | Hirson holds that, since Stadler concedes reform aimed to control and divert pressure from below, the forces working for change must be central to any analysis. Stadler sidelines them. |
| Inman and Rubinfeld (2005) | The elite would only concede if protected against a unitary, majority-dominated state. Credible provincial fiscal powers and the threat of local punishment gave that protection, making the settlement self-enforcing. |
| Inman and Rubinfeld (2005) | Constitutional and fiscal design, not only external pressure or insurgency, made the elite's consent credible. ANC leaders saw continued military resistance as too costly. |
| Inman and Rubinfeld (2008) | Offers a peaceful, institutional alternative to Acemoglu and Robinson's coup-threat account: federalism lets the majority credibly promise not to exploit the elite. |
| Inman and Rubinfeld (2008) | Argues both matter: federal institutions enabled the transition, and the elite's human-capital endowment gave the institutions their content. |
| Inman and Rubinfeld (2013) | Proposes democratic federalism as a fourth credible-commitment institution and treats South Africa as a test of Weingast's theory. |
| Inman and Rubinfeld (2013) | Estimates that both the majority and nearly all of the elite gain in present-value terms, with an elite internal rate of return of about 28% under federalism. |
| Levy (1999a) | Overdetermined: labour-market distortions, black political opposition and the fall of the Soviet bloc are given priority over sanctions. |
| Levy (1999b) | Stresses internal economic costs of apartheid labour-market distortions, black political opposition and the end of the Cold War and Soviet communist threat. He admits that the end of apartheid was overdetermined and that sanctions cannot be definitively ruled out. |
| Lowenberg (1997) | Cost-benefit account: rising costs of repression, shrinking white incomes relative to black, and a lower cost of redistribution made handover preferable. The end of the Cold War and the Namibia settlement hastened it. |
| Malherbe and Segal (2001a) | Credits large firms with an active reform role after 1976 (Urban Foundation, support for Wiehahn, early ANC talks, the Consultative Business Movement secretariat). Notes that socially progressive firms were not governance pioneers. |
| Mariotti (2008) | Argues that internal labour-market change preceded and was independent of the 1973 unrest and 1970s sanctions. These later events may still have contributed to apartheid's ultimate collapse. |
| Mariotti (2008) | Finds no evidence that fear of revolution motivated the relaxation of job reservation, though it may have mattered for the later collapse. |
| Mariotti (2009a) | Challenges the conventional view that mid-1970s and 1980s pressure explains the unravelling. It argues that labour-market desegregation began earlier, at the height of apartheid, for economic reasons. |
| Miller (2003) | Concludes that internal struggle played the more significant role, with external pressure supporting it. |
| Pickles (1988a) | Restructuring: a 'facelift' for petty apartheid and a reworking of grand apartheid that keeps its core social relations. |
| Pickles (1988b) | Treats reform as restructuring: a 'facelift' that changes the visibility of apartheid and the spatial basis of control while safeguarding the basic structures of apartheid social relations. |
| Rich (1990) | Rejects Giliomee's 'third way' as unlikely to win ANC support and as a recipe for balkanisation. He allows it might serve temporarily to sell a settlement to the Afrikaner platteland. |
| Sisk (1989) | Credits both Soweto-style 'revolutionary violence' and the economy's need for skilled black labour and the embourgeoisement of Afrikaners (following Giliomee, Adam and Baker) as pushing the NP toward an 'ideology of survival'. |
| Sisk (1989) | Argues that the NP's decline may open the way for multiracial coalitions and a nonracial democracy, citing Huntington's point that violence works by dividing the dominant group. |
| Spence (1981a) | Sceptical: Spence argues that Afrikaner unity, Verkrampte/Herstigte pressure and external support for Pretoria limit reform to piecemeal measures. |
| Spence (1981a) | Doubts the 'generals favour reform' argument, since military success against SWAPO and FAPLA may encourage faith in a military solution. |
| Spence (1981a) | Describes the government's expectation that Section 10 urban blacks would form a moderating middle class, and treats this as a modest alteration of the status quo that black opinion rejects as grossly insufficient. |
| Spence (1981a) | Argues that Reagan-era support eased external pressure but encouraged right-wing resistance to reform, although future US pressure over Namibia and Cuban troops might force evidence of reform. |
| Spence (1981b) | Sceptical: the election outcome and the constraints of Afrikaner unity make major reform unlikely; expected change is modest (urban Section 10 'middle class', softened influx control, wider union scope). |
| Spence (1981b) | Doubts it; argues that military success against SWAPO and FAPLA may encourage belief in a military solution. |
| Spence (1981b) | Argues that US 'constructive engagement' gave Pretoria flexibility and emboldened the right, while future US pressure over Namibia/Cubans might eventually force reform. |
| Spence (1981c) | Sceptical. Spence argues the election strengthened constraints on Botha, that verligtheid is not liberalism, and that only modest, controlled changes are likely. |
| Spence (1981c) | Doubtful. He suggests that military success against SWAPO and FAPLA may encourage faith in a military solution, and that Malan's successors may not share his reformist view. |
| Spence (1981c) | Argues that Reagan's approach eased external pressure but encouraged right-wing resistance to reform; real reform might be the price of US support over Namibia. |
| Stemmet (2002) | Argues the regime held the military upper hand, so the economy and isolation, not violence or defeat, forced the shift to negotiation. Cites de Lange, De Klerk, Ackerman and Sunter that economic change and 'conscience' eroded apartheid. |
| Stemmet (2002) | Sees reform as carrot and stick, a tactic to buy time that left white power intact and gave no clear intent or timetable. |
| Steward (2004b) | Sceptical. Sanctions were a factor to be weighed but 'not by any means the main factor'. Internal economic and social change, the failure of partition and the end of the Cold War mattered more. This is the position of an interested former regime insider. |
| Strauss (1993) | Presents both strands. The Yes campaign leaned heavily on sanctions, trade, jobs and investment fears, and Strauss calls sport and economic issues the main pillars of the Yes vote. She also reports Giliomee's view that reform was a choice for survival in justice. |
| Strauss (1993) | Documents big business breaking its tradition of keeping out of politics: it funded a Private Sector Referendum Fund and told employees to vote Yes. The CP counter-charge of employer intimidation is reported rather than adjudicated. |
| Strauss (1993) | Cautions that its importance should not be over-estimated. She treats it as a propaganda triumph and mandate that did not stop violence or the later Codesa deadlock. |
| Swilling (1985a) | He sees the first outcome as more likely given the state's armed strength, but expects the costs of repression to be very high. He describes the situation as an interregnum in which neither the state's nor the movements' hegemonic project has gained the strategic advantage. |
| Swilling (1985b) | Reform (Riekert, Black Local Authorities, Regional Services Councils) is a repressive 'reform from above' that creates new contradictions. Capital and reformers back accelerated reform; the state combines repression with restructuring. |
| The Liberation Struggle in South... (n.d.) | No explicit position. It records the Carlton Conference of government and business leaders, the Wiehahn and Riekert reforms, and the 1985 Lusaka meetings between businessmen and the ANC, which is chronological evidence of a business reform turn. |
| Torchia (1988) | Mixed. Rising economic costs explain business opposition to labour and education policy (the Lipton model), but unrest is needed to explain the shift to demanding political reform (the Greenberg model). |
| Torchia (1988) | Skeptical. Business was politically weak and dependent on the state, and its reformism was aimed at rationalising apartheid rather than ending white political dominance. |
| Van Kessel (1989) | The reviewer reports that the books address the question. She treats Botha-era reform as modernisation aimed at survival rather than dismantling. |
| Van Kessel (1989) | Notes that Lemon and most Kitchen contributors stress internal factors and white agency, while Blumenfeld's contributors stress external pressures (sanctions, regional and military). She is critical of the top-down internalist view that neglects Black society's dynamics. |
| Van Kessel (1989) | Endorses Baynham's point that in Pretoria's view reform must parallel or follow repression. She argues that detentions removed the UDF's local leadership and so fuelled unrestrained township violence. |
| Van Vuuren (2006) | Quotes Ken Owen's 'liberation owed to greed' view and accepts it as 'clearly true in part'. Insists liberation had several causes including internal and external struggle. |
| Van Wyk (2005) | Combines them: Sisk-style rational-choice negotiation and Marxist structural analysis, with economic decline as the decisive underlying driver. |
| Van Wyk (2005) | Economic decline is primary; unrest, sanctions and the Cold War's end are important chiefly through their effect on the economy and the NP's calculation of a window of opportunity. |
| Verhoef (2009) | Presents Sankorp's equal opportunity strategy and BEE deal as stabilising, interest-driven attempts to shape the transition and protect the free-market system. |
| Zunes (1999) | Argues non-violent mass action was decisive; armed struggle was marginal and at times counterproductive, though the threat of it may have helped force compromise. |