Did the apartheid economy fail? Growth, productivity, investment and the post-1970s crisis¶
How well did the apartheid economy perform, and what caused the slowdown and accumulation crisis from the 1970s?
Works engaging the debate: 35
Positions¶
- Mediocre growth and low TFP; no 'apartheid boom' (Moll, Feinstein)
- Strong growth until an organic crisis or a crisis of racial Fordism (Saul and Gelb, Gelb)
- Profit squeeze vs political instability as the driver of falling investment (Nattrass, Fedderke)
- Balance-of-payments and gold constraints, plus public infrastructure decline
Evidence base¶
The debate has been argued overwhelmingly from published aggregate series rather than from archival or micro data. The core is the South African Reserve Bank national accounts. These comprise the Quarterly Bulletin real GDP, investment, saving and balance-of-payments series, the fixed capital stock estimates (Prinsloo and Smith, 1960-93) and the public-sector infrastructure gross fixed capital formation series. Behind them sit the official compendia: South African Statistics, the Official Year Book and Union Statistics for Fifty Years. The 'mediocre growth' case (Moll (1991b); Moll (1991c); Moll (1991d)) rests on four tests. The first compares peak-to-peak real GDP growth before and after 1948, using South African Statistics 1976/1982 (1963 and 1975 prices) and Du Plessis's series, with gold output re-measured from Chamber of Mines ore-milled tonnages. The second ranks South Africa against 20 middle-income countries selected via Summers-Heston and measured from World Bank World Tables, 1950-85, where it ties 13th. The third uses manufactured-export shares from UNCTAD Handbooks (SITC 6-9 less 67-68, 1955-85). The fourth sets near-zero 1950-73 TFP from Moll's unpublished Cambridge thesis against Maddison's figures for seven countries. The opposing 'rapid growth' reading (Sources of Growth in the... (1990)) uses Nattrass's and Nedbank compilations benchmarked against Kuznets's historical cross-country rates (about 5% total, 2% per capita). Growth accounting for the post-1970 period comes from Fedderke's decompositions for 1970-97 and his manufacturing estimates for 1920-97 (Fedderke and Simkins (2004); Fedderke (2006)). Bell's manufacturing work (Bell and Madula (2001); Bell (1995)) uses SARB, WEFA 3-digit sectoral data, T.A. du Plessis's 1916/17-1956/57 constant-price series, IDC export values for 1972-93, and Wright's capacity-adjusted TFP.
On the causes of the post-1970s accumulation collapse, the evidence is econometric work on short annual series. Heintz (Heintz (2000); Heintz (2002)) runs fixed-effects panels of seven sectors' net investment rates, 1970-93. His regressors are after-tax profit rates, the accelerator, the Fallon-Lucas user cost of capital, and a principal-components unrest index built from prison population, detentions without trial and strike counts drawn largely from the Survey of Race Relations. Fedderke's work (Fedderke (2006); Fedderke and Luiz (2008)) uses pooled mean group panels of manufacturing investment and ARDL capital-flight estimates. It also includes a Johansen VECM on about 39 annual observations (1954-92), all built on the Fedderke-De Kadt-Luiz political instability, property-rights and political-rights indices. The profit-squeeze side rests on Nattrass's wage-share and rate-of-profit decompositions from CSS/SARB data, 1948-60, read alongside Wage Board and Department of Labour reports (Nattrass (1990)). Her 1970-75 racial wage and income-share reconstructions draw on Manpower Surveys, Bureau of Market Research household studies and a 1973 survey of 282 manufacturing firms (Nattrass (1977)). The infrastructure strand (Perkins et al. (2005a); Perkins et al. (2005b); Fedderke et al. (2006)) assembles 1870/1875-2002 series from SARB investment data, railway administration reports, CSIR/SANRA road data, Eskom and Telkom. It documents a 72% per capita fall in public infrastructure investment, from 8.1% to 2.4% of GDP, over 1976-2002. A structural VAR on SARB quarterly GDP, government consumption and real interest rates, 1960Q2-2006Q4, puts the fall in potential output at about 30% relative to trend from the mid-1970s to the mid-1990s (Du Plessis et al. (2008a)). The sanctions sub-debate draws on several sources: secondary compilations (Hufbauer et al., Jones and Müller, Lipton, Lundahl) (Levy (1999b); Lowenberg (1997)); event studies on CRSP returns and hand-collected Financial Times JSE and rand data (Teoh et al. (1999)); and Penn World Table growth regressions with regime dummies (Inman and Rubinfeld (2008); Inman and Rubinfeld (2013)).
Contested evidence¶
The growth question turns less on different data than on different benchmarks. Moll and uncertain1990 broadly accept similar aggregate growth rates for 1948-73. Moll judges them against contemporaneous middle-income peers and the pre-1948 South African record with gold re-measured (Moll (1991d)). The other side judges them against Kuznets's long-run historical rates for advanced economies (Sources of Growth in the... (1990)). That reading also asserts, without a growth-accounting test, that growth came mainly from factor accumulation. TFP is contested on measurement. Moll's near-zero 1950-73 estimate comes from an unpublished thesis and is compared with Maddison's differently constructed figures. Bell (Bell (1995)) uses Page, Nishimizu-Robinson and Wright to argue that negative manufacturing TFP was common among developing countries. He adds that measured 1980s TFP decline largely reflects excess capacity, making productivity a dependent variable. Fedderke's decompositions (Fedderke (2006); Fedderke and Simkins (2004)) instead show TFP becoming the main source of growth after 1970 only 'by default', as factor contributions shrank.
The dating of the break is read differently from the same SARB aggregates. Bell finds a 1960-65 peak and a 1970-75 drop driven partly by falling physical gold output and real appreciation (Bell and Madula (2001)). Levy dates the break to 1974, well before the 1985-87 sanctions (Levy (1999b)). Du Plessis and Smit match negative supply shocks from 1973 to events only by eye (Du Plessis et al. (2008a)).
On investment, Heintz's results cut both ways. His standardized coefficients make the after-tax profit rate the largest single determinant. Yet his between-cycle decomposition attributes -4.54 points of the fall to unrest against -1.50 to profits (Heintz (2002)). His rebuttal of Nattrass's profitability findings is asserted without data. The unrest index itself is ambiguous. Prison and detention counts may measure repression or the business cycle rather than investor uncertainty, and its validation against the long-short bond spread rests on 23 observations significant at 10% (Heintz (2000)). Fedderke's VECM finds instability endogenous to property rights, crime and output (Fedderke and Luiz (2008)). That finding undercuts the treatment of instability as exogenous uncertainty in both Heintz and Fedderke's own earlier panels, and all of these instability measures are author-constructed subjective indices.
The balance-of-payments reading treats the post-1984 fall in I/GDP as a symptom of foreign-exchange and debt shocks (Bell (1995); Bell and Madula (2001)). This sits against the uncertainty reading of the same SARB investment series (Fedderke (2006)) and the World Bank's asserted parastatal demand shock (World Bank (1994)). On infrastructure, bivariate bounds tests show GDP mostly driving physical infrastructure while public investment forces GDP (Perkins et al. (2005b)). The VECM finds a robust infrastructure-to-output effect but unstable feedback (Fedderke et al. (2006)). Neither controls for gold cycles or sanctions. On sanctions, the evidence is incompatible: PWT regressions attribute roughly -0.8% late-apartheid growth to sanctions and union power (Inman and Rubinfeld (2013)), against Hufbauer's 0.5% of GNP cost estimate (Levy (1999b)) and null financial-market reactions (Teoh et al. (1999)). The racial-Fordism and organic-crisis position has no quantitative test of its own in these evidence ledgers. It is carried by historiographical essays (Freund (2009); Freund (2012); Iliffe (1999)) and invoked in labour studies (Macun and Wood (2001); Poswa (1994)). Posel's re-reading of commission evidence disputes the crisis periodisation of its foundational text (Posel (1983)).
Underused and missing sources¶
Several sources that could settle disputed mechanisms are rarely or never used. Chamber of Mines records appear only for ore-milled tonnages (Moll (1991d)) and skills composition (Fedderke (2006)). Nattrass's mining profitability analysis ignores Chamber wage and recruiting records (Nattrass (1990)). Manufacturing and industrial census returns are not exploited for firm- or industry-level productivity or profitability, though mariotti2014 points to 1950-85 manufacturing data for substitution elasticities. Several other records also go unused:
Whole-economy coverage is compromised because the homelands were excluded from censuses and macro data until 1993-96 (Mariotti and Fourie (2014)).
The 2011 IEC election file bears only on the post-apartheid fiscal modelling (Inman and Rubinfeld (2013)), not on apartheid-era performance.
Strength of key claims¶
| Basis | Share of key claims |
|---|---|
| Systematic data | 19% |
| Partial data | 47% |
| Illustrative or anecdotal | 21% |
| Asserted | 14% |
Comparative figures for all debates are on the Literature overview.
Best-evidenced works¶
Moll (1991d), Fedderke (2006), Heintz (2000), Bell and Madula (2001), Fedderke et al. (2006), Nattrass (1990)
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 |
|---|---|
| Bell (1995) | Argues TFP decline was largely a consequence of falling output, excess capacity and low investment, so it is a dependent variable driven by macro conditions. |
| Bell (1995) | Rejects the ISI/underconsumption explanation. ISI was derailed by the 1980s foreign-exchange shocks and debt crisis, not the cause of them. |
| Bell and Madula (2001) | Rejects skills scarcity as the fundamental cause; the skills-intensive FIRE sector grew far faster, so manufacturing's problems lie 'more deepseatedly' within manufacturing itself. |
| Bell and Madula (2001) | Argues that since the early 1980s a balance-of-payments constraint, driven by falling gold exports and rising import intensity, has capped investment, output and capacity utilisation. |
| Bell and Madula (2001) | Treats these as partial; they may explain world decline since about 1973 but not the sharper developing-country decline since the early 1980s. |
| Coupe (1995) | Endorses Moll's view of 'disastrous' productivity and argues that the industrial relations system of racial job protection contributed substantially to it, though it may be impossible to quantify. |
| Du Plessis et al. (2008a) | Provides a quantified estimate (about 30% fall in potential GDP relative to trend) that the authors say is consistent with historical accounts by Moll (1991) and Feinstein (2005). It treats the late-apartheid years as a period of cumulative negative supply shocks. |
| Du Plessis et al. (2008a) | Finds that supply shocks dominate GDP over the long run. Fiscal shocks dominate monetary shocks in the short-to-medium run only in the longer sample, and the ranking reverses after 1983, which the authors link to the De Kock-era monetary regime change. |
| Fedderke (2006) | Holds that sectoral and systemic uncertainty depress manufacturing investment, and that apartheid's discretionary state power raised uncertainty. |
| Fedderke (2006) | Argues that only quality indicators (maths, natural/engineering science degrees) raise TFP; apartheid's duplicated university system was wasteful. |
| Fedderke and Luiz (2008) | Finds strongest (weak-form) support for North's institutions/property-rights primacy, some support for modernization, and little for political-rights primacy. |
| Fedderke and Luiz (2008) | Confirms that political instability lowers aggregate investment in South Africa, but argues instability is endogenous rather than exogenous uncertainty. |
| Fedderke and Luiz (2008) | Treats apartheid as a dysfunctional institutional system whose unequal property rights triggered distributional conflict; no support for a purely political-aspiration driver. |
| Fedderke and Simkins (2004) | Segregation-era intervention (civilised labour policy, protection, farm support) imposed static costs but did not seriously undermine growth. Post-1948 social engineering imposed heavier costs that stored up later trouble. |
| Fedderke and Simkins (2004) | Finds Lucas-type human capital spillovers and Schumpeterian R&D effects in South African manufacturing, and rejects Romer-type capital spillovers. |
| Fedderke et al. (2006) | Takes the cointegration/VECM route to avoid spurious OLS results. Finds a significant but moderate direct output elasticity (about 0.1–0.2, up to 0.5 with institutions) and a strong crowding-in effect on private investment. |
| Fedderke et al. (2006) | Infrastructure leads growth robustly. Growth leading infrastructure receives some support but is not robust to alternative infrastructure measures. |
| Fedderke et al. (2006) | Deliberately sets social infrastructure aside as a separate modelling problem, since apartheid created differential access to social rather than physical infrastructure. Suggests the growth payoff of economic infrastructure could help finance social spending. |
| Fedderke et al. (2006) | Points to collapsing public infrastructure investment and political instability and weak property rights as depressing private investment. Notes that after 1980 government consumption rose at the expense of investment. |
| Freund (2009) | Cites Moll, Gelb and Saul and Gelb as pioneers in identifying crisis. He sees the MEC as explaining industrial weakness through business clustering around profitable raw-material options. |
| Freund (2012) | Accepts racial Fordism and the minerals-energy complex as partial descriptions, with qualifications. Finds ISI the best fit and rejects corporatism (following Pretorius). |
| Heintz (2000) | Finds that the user cost of capital is insignificant; after-tax profit rates, sectoral growth and social unrest drive investment, siding with the post-Keynesian/Kaleckian profitability tradition. |
| Heintz (2000) | Provides South African time-series and panel support that unrest lowers investment independently of profit outcomes, while noting broader claims need more empirical work. |
| Heintz (2002) | Defends the SSA reading. Concedes Nattrass's point that manufacturing profit trends were weak, but argues the critique is too narrow and untested, and that political instability explains the investment crisis. |
| Heintz (2002) | Political unrest, independent of profitability, accounts for most of the fall in the rate of accumulation. Profit rates matter but are secondary. |
| Iliffe (1999) | Sets out the three views without choosing one. He observes that post-1994 ANC policy conformed largely to the third position, and says the crisis will be a central problem for future historians. |
| Inman and Rubinfeld (2008) | Treats sanctions (from 1976) and COSATU (from 1985) as having depressed growth through reduced trade openness and investment, giving the elite a growth dividend from ending apartheid. |
| Inman and Rubinfeld (2013) | Finds sanctions had a negative but imprecisely estimated effect, and majority unions a significantly negative effect on growth in 1985-93; both effects disappear under democracy. |
| Innes and Plaut (1978) | Attributes the 'golden period' of accumulation to the defeat of white and black workers in the 1920s, which allowed higher exploitation of all labour. |
| Levy (1999b) | The growth break came around 1974 (oil shock, labour-market distortions, external borrowing), well before the 1985-87 sanctions. GDP growth then recovered after sanctions. |
| Lowenberg (1997) | Apartheid and import-substitution protection produced Latin American-style mediocre growth rather than East Asian-style growth. |
| Macun and Wood (2001) | Adopts Nattrass's account of a crisis of racial fordism (skills shortages, low productivity) that pushed reform and opened space for African workers' bargaining power. |
| Mariotti and Fourie (2014) | They endorse Moll's finding that early rapid growth reflected the post-war global rebound rather than cheap labour, and that growth was slow by comparative standards. They also endorse the view that apartheid collapsed under its own weight (Feinstein): gold profitability, industrial inefficiency and external pressures. |
| Moll (1991b) | Rejects it. The post-1948 growth rate was not higher than before 1948, and South Africa's rank among comparable developing countries was consistently mediocre. |
| Moll (1991b) | Attributes the poor export performance to domestic supply-side and policy factors rather than external demand. Sanctions only mattered from the late 1970s, and the limited home market was mainly a consequence of poor export policy. |
| Moll (1991c) | Rejects it. Growth was no faster after 1948 and was mediocre against similar developing countries. |
| Moll (1991d) | No. Compared with similar developing countries, South Africa ranked around 12th-13th of 21 over 1950-85, and well below the newly industrialising countries. |
| Nattrass (1977) | Survey evidence points to market size as the main past constraint, with skilled labour costs and availability the main future one. She sees no inherent conflict between Black wage growth and growth, and views political constraints as a likely limit. |
| Nattrass (1990) | The 1948-55 reversal in wage trends is said to bolster the claim, but she stresses that much of the turnaround had economic and structural roots, not policy roots, such as demobilisation, women leaving the labour force and urbanisation. |
| Nattrass and Seekings (2010b) | Rejects the strength-to-crisis story (e.g. Saul and Gelb); argues profit rates trended down from 1960 and growth was weak relative to comparable economies, though mining profitability did rise in the 1950s. |
| Nattrass and Seekings (2010b) | Follows Moll and Feinstein: growth was good by north-west European standards but slow against middle-income peers, with very low TFP growth. |
| Perkins et al. (2005a) | Notes that estimates vary widely and OLS results on time series may be spurious. It uses bounds-test methodology suited to non-stationary data and finds bidirectional evidence, with a forcing relationship from public infrastructure investment and capital stock to GDP. |
| Perkins et al. (2005a) | Argues that causality runs both ways. GDP drives most physical infrastructure indicators, but roads and public infrastructure investment and capital stock drive GDP. |
| Perkins et al. (2005a) | Acknowledges the argument and cites overcapacity and lumpiness. Nonetheless judges the long decline relative to GDP probably undesirable, and links it to growing government consumption and deficits in the late apartheid years. |
| Perkins et al. (2005b) | Takes a middle position. Notes wide variation in estimates, with Aschauer high and Holtz-Eakin and Garcia-Milà near zero, and warns that some DBSA results are spurious (OLS on time series). Finds causation running in both directions, with GDP driving most infrastructure and only roads and aggregate public infrastructure investment and capital stock forcing GDP. |
| Perkins et al. (2005b) | Acknowledges the over-expansion argument and overcapacity in some areas, but leans towards seeing the long decline as probably undesirable. Also links it to growing government consumption and fiscal deficits late in apartheid. |
| Posel (1983) | Rejects the 'organic crisis' concept as vacuous and functionalist; the trends cited were endemic contradictions since the 1920s. |
| Poswa (1994) | Cites the over-accumulation explanation dating back to 1974 but does not test it; uses it to justify the need for industrial democracy. |
| Schirmer (2008) | Follows Moll in blaming blanket protectionism and the focus on 'advanced' sectors for post-war underperformance and the collapse of manufactured exports. |
| Schneider (2000) | Multi-causal: lack of technological autonomy, the civilized labour policy, restrictions on black labour, sanctions, separate development costs, and over-capital-intensive industry. |
| Sources of Growth in the... (1990) | Presents these as unresolved policy disputes. It notes that capital widening predominated until the 1970s, limiting productivity gains. |
| Teoh et al. (1999) | Emphasises the falling gold price. It notes that recessions coincided with gold declines, that the JSE rallied during peak pressure, and that capital flight did not come with worsening terms of trade. |
| World Bank (1994) | Locates it in rising capital intensity, capital misallocation towards the public sector, skills shortages, rising wages and unrest, and a post-1985 collapse in parastatal investment, rather than in sanctions or financing alone. |