Economic Growth in South Africa since the Late Nineteenth Century¶
Johannes Fedderke; Charles Simkins (2004). Economic Growth in South Africa since the Late Nineteenth Century.
| Type | book chapter |
| Scope | Economics and economic history of apartheid |
| Method | mixed |
| Period studied | 1870–2000 |
| Regions | South Africa |
| Themes | origins segregation, mining migrant labour, manufacturing industrial policy, bantustans removals influx, labour markets wages, macro fiscal finance debt, politics ideology elections, education human capital, methods data sources, comparative theory, inequality poverty unemployment, sanctions divestment |
| Citation key | fedderke2004 |
Argument¶
The chapter surveys modern growth theory and then applies it to South African growth in four international monetary/trade regimes (1870-1914, 1914-45, 1945-70, 1970-2000). The narrative sections treat state intervention before 1945 as wasteful but not ruinous for growth. After 1948 apartheid's heavier efficiency costs fell on human capital, property rights and international isolation. The econometric section (mostly the authors' own earlier work) finds that growth was driven by capital and labour accumulation until the 1990s, when TFP became dominant. It finds that political uncertainty depresses investment and provokes capital outflows, that financial deepening helps growth via investment, and that rising real labour costs ('mispricing') explain poor employment creation. It concludes that the evidence favours the liberal over the Marxist reading of apartheid and growth, and that a poor-quality education system limits technological progress.
Evidence and method¶
Frankel, Union Year Books, Union Statistics for Fifty Years, censuses (1921, 1936, 1946, 1960, 1996), national accounts from 1946 (Stats SA), SARB Quarterly Bulletin series, sectoral panel data 1970-97, a political instability index (Fedderke, De Kadt and Luiz), and the authors' VECM, ARDL and dynamic heterogeneous panel estimates.
Place in the debates¶
- Liberal vs Marxist/revisionist interpretations of apartheid and growth (see Liberal vs revisionist: was apartheid functional or dysfunctional for capitalist growth?). Sides with the liberal view: political instability harmed and property rights helped capital accumulation, and political rights show no growth-supporting effect. It calls the wider debate 'tired and hackneyed'.
- Cause of poor employment creation (labour 'mispricing'/wage rigidity vs structural or demand explanations) (see Unemployment: causes, measurement and labour market institutions). Blames real wages rising faster than productivity and labour-market rigidity; reports strongly negative wage elasticities and argues that trade liberalisation did not cause labour shedding.
- Efficiency costs of segregation before 1948 and of apartheid after it (see Did the apartheid economy fail? Growth, productivity, investment and the post-1970s crisis). 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.
- Exogenous (Solow-Swan) vs endogenous (Romer, Lucas, Schumpeterian) growth (see Did the apartheid economy fail? Growth, productivity, investment and the post-1970s crisis). Finds Lucas-type human capital spillovers and Schumpeterian R&D effects in South African manufacturing, and rejects Romer-type capital spillovers.
Findings¶
-
Real per capita income peaked in 1981 and fell sharply thereafter, which the authors attribute ultimately to political instability. (p. 36)
From then until 1994, the trend was downwards with real per capita income in 1994 18% lower than it had been thirteen years earlier.
-
Pre-1945 labour-market regulation (civilised labour policy and relief work) transferred unskilled jobs from Africans to Whites at taxpayer expense. (p. 30)
Taken as a whole, labour market regulation had the distributive effect of reallocating unskilled wage labour from Africans to Whites.
-
Reliance on foreign ('tropical') mine labour allowed African mine wages to stay stable in real terms for decades. (p. 27)
The increasing reliance on foreign labour was what made possible the stabilisation in real terms of African wages
-
Influx control was effective after 1955 in slowing African urbanisation, and homeland policy greatly shifted the African rural population distribution. (p. 33)
Influx control was then strengthened, and for a period of about twenty years from 1955, it slowed African urbanization substantially.
-
Resettlement and removals moved the African population away from commercial farms and towards homelands. (p. 33)
In 1950, there were eleven Africans in reserves for every ten on commercial farms. By 1996, there seventy-five Africans in tribal rural areas for every ten on commercial farms.
-
Post-Sharpeville capital controls and the dual exchange rate fostered capital flight and inefficiency in the later apartheid era. (p. 33)
Capital controls and a dual exchange rate also created incentives for capital flight, especially at times of instability, which had reached substantial proportions by the end of apartheid.
-
Uncertainty, especially systemic political uncertainty, strongly lowers manufacturing investment. (p. 42)
The striking finding is that uncertainty exercises a statistically significant and strong effect on investment expenditure in South African manufacturing industry
-
Capital flows are sensitive to political risk; both instability and political liberalisation stimulated outflows. (p. 46)
Higher instability, and political liberalization in South Africa both served to stimulate capital outflows.
-
Gold mining employment fell dramatically between 1985 and 1997. (p. 53)
for Gold & Uranium Mining the decline was from 526839 to 241352 employees over the same period
-
Mining labour demand was elastic with respect to real wages, notably in diamond and other mining. (p. 56)
the real wage elasticity for the Coal, Gold & Uranium and Diamond & Other Mining sectors was -0.44, -0.69, and -1.45
-
The authors conclude that wage elasticities of labour demand are strongly negative, so wage moderation was insufficient. (p. 61)
First, wage elasticities of labour demand are persistently negative in the South African economy, and they are strongly negative.
-
Their econometric results favour the liberal over the Marxist interpretation of South African economic history. (p. 51)
Fourth, note that the evidence favours the liberal rather than the Marxist interpretation of South African economic history.
-
Post-1948 apartheid imposed heavier efficiency costs than earlier segregation. (p. 68)
The efficiency costs imposed after 1948 were heavier and they stored up trouble for later growth.
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| Union Statistics for Fifty Years (Bureau of Census and Statistics, 1960) | official statistical series | 1910-1945 | Main statistical basis for the 1914-1945 narrative: prices, money supply, exports, gold output, agricultural indices and employment. |
| Union Year Book 1910-1916 and later editions (1916-1945) | official statistical series | 1910-1945 | Gives public debt in 1910, banks and life assurance companies, and relief-work and civilised-labour employment figures. |
| Statistics South Africa, South African Statistics 2002 (tables 19.3-19.4), gross value added by industry | national accounts macro | 1946, 1970, 1994 | Table 1 shows the industrial structure of the economy and its shift towards manufacturing, with a falling mining share. |
| 1960 Population Census (educational attainment by birth cohort) | population census | cohorts born 1885-1970 | Establishes average years of education by race for birth cohorts and the educational attainment gap. |
| 1921, 1936 and 1946 population censuses (employment data) | population census | 1921-1946 | Used to compute labour absorption rates for Whites, Coloureds and Asians in 1921 and 1946, and the economically active white population in 1936. |
| Agricultural censuses 1918, 1925, 1930, 1937 and 1945 | agricultural census | 1918-1945 | Source of agricultural employment data, with the caveat that months of collection differed. |
| 1951 Census (African labour force participation rate) | population census | 1951 (applied to 1938) | The participation rate is applied to the 1938 population to show that the 1937 committee underestimated the size of the African labour force. |
| 1904 Transvaal Census | population census | 1904 | Shows twice as many children in school as in 1898, taken as evidence that the colonial administration governed more in line with growth. |
| 1897 Industrial Commission of Enquiry into the mining industry (South African Republic), via Smith (1996) | commission or official report | 1897 | Illustrates the clash of interest between the mines and the state over concessions, rail rates and duties. |
| Cullinan Commission on Trade and Industries 1910-11 and Board of Trade and Industries (1945) report | commission or official report | 1910-1945 | Source of the tariff principles guiding industrial protection and the 1939 import duty rate of 14.25%. |
| Frankel (1938) Capital Investment in Africa and Frankel (1944) national income analysis | secondary literature compilation | 1887-1944 | Used for foreign capital in diamonds and gold, gold yields, national income growth periods, state spending on agriculture and manufacturing's import dependence. |
| Hobart Houghton (1971) Economic Development 1865-1965 (wool prices, balance-of-payments episodes) | secondary literature compilation | 1920-1961 | Source for wool price variation and the four balance-of-payments adjustment episodes. |
| Wilson (1972) Labour in the South African Gold Mines, 1911-1969 | secondary literature compilation | 1911-1969 | Supports the claim that reliance on foreign labour stabilised real African wages. |
| Pearsall (1937), Busschau (1945), Richards (1935), Schumann (1934), Malherbe (1977) (period studies cited for data) | secondary literature compilation | 1911-1945 | Provide data on manufacturing structure, capital intensity, agricultural subsidy costs, business-cycle dating and educational quality. |
| Fedderke and Schirmer (2004) manufacturing growth accounting 1920-1997 | secondary literature compilation | 1920-1969 | Source of Tables 2 and 3 classifying manufacturing growth as capital-led, labour-led or TFP-led. |
| Fedderke (2002a) growth accounting decomposition 1970-97 | secondary literature compilation | 1970-1997 | Figure 2 shows the contributions of labour, capital and technology to aggregate and sectoral output growth. |
| South African Reserve Bank Quarterly Bulletin series S-116, S-127, S-129, S-131 | official statistical series | 1960-2000 | Used to construct the private savings and private investment rates in Figure 3. |
| Fedderke (2001a) investment estimation, South African manufacturing sectors | secondary literature compilation | 1970s-1990s | Source of the standardized coefficients showing that systemic and sectoral uncertainty depress investment. |
| Political instability index, Fedderke, De Kadt and Luiz (2001a,b) | secondary literature compilation | 1935-1997 | Index of eleven indicators of repressive state response, used as the systemic uncertainty and instability measure in investment, capital-flow and institutions estimations. |
| Fedderke and Liu (2002) capital flow and flight estimates | secondary literature compilation | 1960-95 | ARDL estimates showing that capital flows are sensitive to political risk and rights changes. |
| Kularatne (2002) VECM study of financial deepening, 1952-92 | secondary literature compilation | 1952-1992 | Supplies the findings that stock market liquidity and credit extension affect growth via investment. |
| Mariotti (2002) VECM and threshold study of fiscal and monetary policy | secondary literature compilation | post-war | Source of findings on negative long-run effects of government consumption and inflation and of optimal levels. |
| World Bank (2000) large manufacturing firm survey, 1998-99 | household or labour survey microdata | 1998-99 | Cited as evidence of credit rationing in the South African economy. |
| Fedderke, De Kadt and Luiz (2001b) time series study of institutions and growth 1935-97 | secondary literature compilation | 1935-1997 | Source of Figure 6 and the claim that the evidence favours the liberal rather than the Marxist interpretation. |
| Fedderke and Pirouz (2002) mining employment, productivity and real labour cost, 1970-97 | secondary literature compilation | 1970-1997 | Figures 7-10, Table 4 and the wage elasticities link real labour cost to job losses in mining. |
| Fedderke and Mariotti (2002) 48-sector labour market panel 1970-97 | secondary literature compilation | 1970-1997 | Shows that sectors with wages tied to productivity create more jobs, and gives manufacturing wage elasticities. |
| Fedderke, Shin and Vaze (2003) and Fedderke (2004) on trade and skill-specific wage elasticities | secondary literature compilation | 1970-97 | Supports the claims that trade liberalization did not hurt labour and that unskilled labour demand is strongly wage elastic. |
| Fedderke (2001g) dynamic heterogeneous panel on TFP determinants, with education data from Fedderke, De Kadt and Luiz (2000, 2003) | secondary literature compilation | 1970-97 | Table 5 tests Romer, Lucas and Schumpeterian spill-over hypotheses; Figures 12-13 and Table 6 document declines in maths and NES degrees and university research output. |
Inference¶
Quantification: Statistical analysis.
Narrative and simple tabulation for 1870-1945, using published historical statistics and earlier historians' estimates. Descriptive structural tables for 1946-1970. For the post-1945 period the paper summarizes the authors' own earlier econometric work: growth accounting, investment estimations, ARDL capital-flow models, VECM and threshold models for finance and policy, dynamic heterogeneous panels and labour demand elasticities. Interpretation is by theory-driven reading of regression coefficients.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| The 1914-1945 economy was buffeted by adverse external shocks (post-war inflation and deflation, the return to gold, the Depression) until the 1933 gold price rise permitted rapid growth. (p. 31) | Union Statistics for Fifty Years and Union Year Book series on prices, money supply, exports, gold price and output, plus Frankel's national income periodisation. | Partial data |
| Civilised labour and relief work policies reallocated unskilled wage labour from Africans to Whites at a fiscal cost, but had limited impact on growth before 1945. (p. 30) | Union Year Book counts of relief and subsidised employment, the 1936 Census white economically active population, and the authors' reasoning on productivity and wage differences. | Partial data |
| Growth in 1945-1969 manufacturing was predominantly capital-accumulation led, whereas 1920-45 growth drivers were diverse across industries. (p. 34) | Growth accounting by sector in Fedderke and Schirmer (2004), reported only as summary tables. | Partial data |
| Post-1970 growth was driven increasingly by TFP by default, as labour and capital contributions declined. (p. 37) | Growth accounting decomposition for 1970-97 from Fedderke (2002a), Figure 2. | Systematic data |
| Uncertainty (systemic and sectoral) significantly lowers investment in South African manufacturing and is a key reason for poor investment performance. (p. 42) | Fedderke (2001a) estimation using SARB series and the political instability index; standardized coefficients in Figure 4. | Systematic data |
| Political instability and political liberalization both stimulated capital outflows, and capital flows are sensitive to political risk. (p. 46) | ARDL estimates of capital flow and flight measures, Fedderke and Liu (2002), using the political rights and instability indices. | Systematic data |
| The evidence favours the liberal over the Marxist interpretation: property rights and stability aided capital accumulation while political rights did not. (p. 51) | Fedderke, De Kadt and Luiz (2001b) time series on instability, rights indices and output, 1935-97. | Partial data |
| Rising real labour cost relative to productivity, not trade liberalization, explains employment losses, with strongly negative wage elasticities especially for unskilled labour. (p. 61) | Sectoral panel econometrics and descriptive correlations from Fedderke and Pirouz (2002), Fedderke and Mariotti (2002), Fedderke, Shin and Vaze (2003) and Fedderke (2004). | Systematic data |
The author's caveats¶
The authors state that pre-1945 statistical information is sparse and that full national accounts date only from 1946. The 1911-1919 growth estimate is very uncertain because nominal national income and a GDP deflator are lacking. The African labour force estimates are imprecise. Growth accounting is described as only broadly indicative. The Kularatne estimate for the liquidity effect may be exaggerated through misspecification. Inequality and social capital are omitted for lack of reliable information.
Gaps in the evidence¶
The pre-1945 narrative leans on secondary estimates (Frankel, Pearsall, Richards) and published aggregates rather than archival or firm-level sources. The Chamber of Mines, company records and state archives are not used to test claims about labour and wages. The post-1945 econometric claims about apartheid, institutions and uncertainty rest on a constructed instability index and earlier papers, and are summarised without coefficients or specifications, so they cannot be checked here. Firm-level or micro-level investment, wage and household survey data, and more refined historical capital stock series, would test the causal claims better; much of this exists now.
Replicability: Data described but not public.