Capital Intensity in South African Manufacturing and Unemployment, 1972-90¶
Raphael Kaplinsky (1995). Capital Intensity in South African Manufacturing and Unemployment, 1972-90. World Development, Vol. 23, No. 2. Pergamon / Elsevier Science.
| Type | journal article |
| Scope | Economics and economic history of apartheid |
| Method | descriptive quantitative |
| Period studied | 1972–1990 |
| Regions | South Africa |
| Themes | manufacturing industrial policy, labour markets wages, inequality poverty unemployment, sanctions divestment, business and state, transition post apartheid, unions strikes, methods data sources |
| Citation key | kaplinsky1995 |
Argument¶
Kaplinsky tests two explanations of South Africa's poor manufacturing employment and high capital intensity: overexpansion of capital-intensive sectors, and distorted factor prices (mainly wages that are too high). He rejects both. Sectoral investment data show that aggregate capital intensity reflects underinvestment in labour-intensive sectors, not overinvestment in capital-intensive ones. Technique choice within sectors shows little response to factor prices, with the large synfuel investments the main exception and those driven by oil sanctions. He argues that political factors explain the poor employment record: reluctant private investment, sanctions and export barriers, labour conflict and low productivity, a stunted informal sector, and weak FDI. He concludes that the data do not support wage reductions or a withdrawal of the state from allocative decisions.
Evidence and method¶
Aggregate and comparative data from UNIDO (1991), World Bank Tables, Levy (1992) and Industrial Development Corporation (IDC) sectoral data on capital stock, investment and capital-labour ratios. Also Syrquin-Chenery normality comparisons, Moll's TFP estimates, and secondary studies of the informal sector (Manning and Mashigo, Simon and Birch) and FDI (Garner). South Africa is compared with first- and second-tier NICs and developed market economies. The paper itself says its analysis is aggregative and relies on secondary databases.
Place in the debates¶
- Factor-price distortions (high wages) vs political explanations of capital intensity and unemployment (see Unemployment: causes, measurement and labour market institutions). Rejects the World Bank (Fallon), SACOB and Normative Economic Model view that high black wages cause capital intensity and unemployment. Factor prices are a relatively unimportant factor, and the explanation is mainly political.
- Crowding-out and overinvestment in capital-intensive sectors (Levy, World Bank) (see Industrialisation, protection and industrial structure). Argues the crowding-out hypothesis is difficult to sustain. The evidence points to underinvestment in labour-intensive sectors, not overinvestment in capital-intensive ones.
- Manufacturing vs the Minerals-Energy Complex (Rustomjee, MERG) as the analytical unit (see Industrialisation, protection and industrial structure). Acknowledges the MEC argument and agrees the MEC's heavy industries explain much of the high aggregate capital intensity, while using the conventional manufacturing categories.
- Role of the state and privatisation in post-apartheid industrial policy (see Industrialisation, protection and industrial structure). Argues against the assumption that state involvement crowds out investment. In the 1980s investment held up only where the state played a central role, so the state may need to crowd in investment.
Findings¶
-
Total factor productivity in manufacturing fell over the period, among the worst performances internationally. (p. 1)
During 1972-90 total factor productivity declined at an annual rate of 1.02%.
-
Manufacturing employment growth collapsed in the 1980s after growing strongly in the 1970s. (p. 3)
But during the 1980s this expansion of employment came to a virtual halt, with an annual growth rate during 1980-90 of 0.22% p.a.
-
South Africa's employment growth ranking in the 1980s was far better than its output ranking, which Kaplinsky treats as a paradox of high unemployment alongside relatively good employment growth. (p. 4)
whereas South Africa ranked 35 out of 44 countries in the growth of output during the 1980s its ranking with respect to employment growth was 14.
-
Capital intensity is concentrated in the "other basic chemicals" subsector (which includes synfuels), which rose sharply in capital intensity and share of capital stock. (p. 6)
the other basic chemicals subsector increasing its capital intensity by more than three times and its share of total capital stock rose from 11 to 29%.
-
Investment collapsed in the 1980s, and labour-intensive sectors received minimal investment, which undermines the crowding-out explanation. (p. 7)
-
High aggregate capital intensity reflects underinvestment in labour-intensive sectors, not overinvestment in capital-intensive ones. (p. 7)
instead of high levels of capital intensity being driven by "over-investment" in the capital-intensive sectors, a more likely explanation is that these high levels of capital intensity are explained by the absence of investment in labor-intensive sectors.
-
The one detailed micro study found little evidence that technique choice responded to relative factor prices. (p. 9)
The only detailed microeconomic study of the determinants of technical choice in South African manufacturing found little evidence that the choice of technology was affected by relative factor prices
-
Synfuel investments driven by oil sanctions were a major, economically unviable, capital-intensive choice not caused by the factor-price ratio. (p. 10)
It is significant, though, that this inappropriate choice of technique was not driven by the capital/labor price ratio.
-
Apartheid-era legal restrictions and a hostile environment stunted informal manufacturing. (p. 10)
In part this is because for many years it was illegal for black South Africans to head their own enterprises or to engage in manufacturing activities
-
Poor employment performance is attributed to political factors, with private capitalists reluctant to invest in labour-intensive sectors. (p. 11)
this underinvestment has been a consequence of the prevailing political climate in the dying years of apartheid in which private sector capitalists have been reluctant to invest in these labor-intensive sectors.
-
Policy implication: the state may need to crowd in investment rather than withdraw, since investment held up only where the state was central. (p. 12)
in a context of low investor confidence, the only sectors where manufacturing investment held up during the 1980s were those sectors where the state played a central role in resource allocation.
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| UNIDO industrial statistics (Industry and Development: Global Report 1991/92; Handbook of Industrial Statistics 1990) – manufacturing value added, employment and subsector data | international trade or finance data | 1980-89 (some series to 1990) | Provides comparative MVA and employment growth (Tables 1-2) and subsectoral shares of MVA (Figures 4-7), including value-added per employee as a capital-intensity proxy. |
| Industrial Development Corporation (IDC) database, 76-sector breakdown of South African manufacturing | official statistical series | 1972-90 | Source of sectoral capital-labor ratios, capital stock shares and net investment by sector (Tables 5 and 6), underpinning the underinvestment argument. |
| Levy (1992), World Bank Informal Discussion Paper No. 1, and World Bank Tables (diskette) | secondary literature compilation | 1961-89 | Supplies the economy-wide capital-labor ratios (Table 3) and manufacturing incremental capital-output ratios (Table 4) used to show high comparative capital intensity. |
| World Bank Tables (diskette) – investment/GDP and capacity-to-import series | international trade or finance data | 1972-90 | Used for investment as a share of GDP (Figure 3) and capacity-to-import comparisons showing falling investment in the 1980s. |
| Manufacturing share of GDP series: Houghton (1967), Nattrass (1988) and IDC database (1992) | national accounts macro | 1950-90 | Charts the share of manufacturing in GDP (Figure 1), showing stabilisation at 23-25% after the mid-1970s. |
| Syrquin-Chenery normality comparisons for the manufacturing sector | secondary literature compilation | 1972-88 | Used to show manufacturing's GDP share is close to what would be expected for South Africa's per capita income and population. |
| South African Reserve Bank, Annual Economic Report 1992 | official statistical series | 1988-1992 | Source for the estimate that 46% of the labour force was unemployed in mid-1992, up from 39% in 1988. |
| Moll (1990) D.Phil dissertation on output and productivity trends (TFP and MVA growth estimates) | secondary literature compilation | 1960s-1984 | Supplies comparative TFP growth rates in manufacturing 1973-84, linking South Africa to Latin American rather than East Asian patterns. |
| Rustomjee (1993) on the Minerals-Energy Complex and MERG (1993) report | secondary literature compilation | 1960-90 | Provides the MEC share-of-GDP figures (15-17% non-MEC, 25-27% MEC) used to reframe the manufacturing sector. |
| Nattrass and Brown (1977) microeconomic study of technical choice; Black (1991) | secondary literature compilation | 1970s-1991 | Used as evidence that choice of technique was little affected by relative factor prices, especially for unskilled black labour. |
| Crompton (1993) research report on synfuel investments; SASOL Facts (1992) | company records or reports | 1984-91 | Supplies estimates of SASOL and MOSSGAS investment (R6bn and R12bn; R31bn by SASOL), used to show synfuels were about half of manufacturing investment growth. |
| Manning and Mashigo (1993) microenterprise research report; Simon and Birch (1992); Mashigo (forthcoming) | secondary literature compilation | 1984-1992 | Evidence on the small size, low incomes and hostile environment of informal manufacturing, including comparative manufacturing shares of the informal sector. |
| 1962 Regulation for the Administration and Control of Townships in Bantu Areas | legislation gazette | 1962-1980s | Cited as an example of legal restrictions that barred black-run enterprises and thereby stunted the informal sector. |
| Garner (1993) LSE research paper on determinants of foreign direct investment flows | secondary literature compilation | 1989-93 | Supports the claim that political developments, not economic factors, determined timing of inward FDI. |
| Policy documents: SACOB (1991), Nationalist Government Normative Economic Model (CEAS 1993), World Bank (1993) An Economic Perspective on South Africa | commission or official report | 1991-1993 | Represent the factor-price-distortion view that the paper tests and rejects, particularly the call for wage restraint. |
Inference¶
Quantification: Descriptive tables.
Descriptive, comparative argument: manufacturing growth, employment and capital intensity are compared across sectors, time (1972-90) and country groups using UNIDO, IDC and World Bank aggregates. Competing hypotheses (crowding-out, overinvestment, distorted factor prices) are rejected by inference from sectoral investment patterns and secondary studies, and the political-explanation alternative is argued by assertion and selected literature rather than tested.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| South African manufacturing employment growth was almost nil in the 1980s (0.22% p.a.), yet relatively good in international ranking given weak output growth. (p. 3) | UNIDO (1991) employment and MVA data, Table 2 and rankings across 44 countries. | Systematic data |
| Capital intensity in South Africa is high compared with Latin American and Asian NICs, but exaggerated by the high Rand and by the mining and agriculture sectors in economy-wide figures. (p. 4) | Levy (1992) and World Bank Tables for capital-labor ratios and ICORs; exchange-rate adjusted capital-labor ratios. | Partial data |
| Capital intensity is concentrated in other basic chemicals (synfuels) and basic metals, which hold most capital stock but little employment, while labour-intensive sectors such as clothing and leather saw falling capital stock. (p. 6) | IDC database sector-level capital-labor ratios, capital stock shares and net investment (Tables 5 and 6). | Systematic data |
| High aggregate capital intensity reflects underinvestment in labour-intensive sectors rather than overinvestment or crowding-out by capital-intensive sectors. (p. 7) | IDC sectoral net investment (Table 6), World Bank investment/GDP (Figure 3) and UNIDO subsector shares (Figures 4-7). | Partial data |
| Factor-price distortion is not the main cause of capital-intensive technique choice; the only clear exception, synfuels, was driven by oil sanctions rather than the capital/labor price ratio. (p. 10) | Nattrass and Brown (1977), Black (1991), and Crompton (1993)/SASOL cost estimates; author states evidence is inadequate for a firm judgement. | Partial data |
| The informal manufacturing sector is small, low-income and constrained by politically determined factors (skills, infrastructure, insecurity, legal restrictions). (p. 10) | Manning and Mashigo (1993), Simon and Birch (1992), Mashigo (forthcoming), and the 1962 township regulation. | Illustrative or anecdotal |
| Poor employment performance is attributable mainly to political factors (private-sector reluctance to invest, sanctions, labour militancy, FDI behaviour) rather than factor prices. (p. 11) | Inference from the sectoral investment patterns above plus Garner (1993) on FDI; no direct test of the political mechanism. | Asserted |
The author's caveats¶
The author notes that research on manufacturing is at an early phase, so analysis is aggregative and reliant on secondary databases; that capital-output figures 'wobble' and ICORs are overstated by including replacement investment; that no uniform comparative database exists (IDC 76-sector vs UNIDO 23/28-sector); that UNIDO and ILO employment data differ and make comparisons of labour-force shares unreliable; that UNIDO employment probably includes the TBVC states; that South African population and per capita income figures are uncertain; and that the informal sector is hard to measure and the evidence on technique choice is inadequate.
Gaps in the evidence¶
The political explanation is not tested directly: there is no firm-level investment, profitability or survey data on investor behaviour, labour militancy, or sanctions effects, nor any counterfactual or regression separating political from factor-price effects. Plant-level census microdata, capital stock series by sector deflated consistently, and user-cost-of-capital and wage series by race would allow a proper test of the factor-price hypothesis; these existed in part (Central Statistical Service manufacturing censuses, IDC, SARB) but were not used. Informal-sector evidence rests on a few small studies; later household and labour-force surveys would give better evidence.
Replicability: Data poorly documented.