Structural Development of South African Manufacturing Industry - A Policy Perspective¶
C. L. McCarthy (1988). Structural Development of South African Manufacturing Industry - A Policy Perspective. South African Journal of Economics, vol. 56(1).
| Type | journal article |
| Scope | Economics and economic history of apartheid |
| Method | descriptive quantitative |
| Period studied | 1925–1986 |
| Regions | South Africa, Southern Transvaal, homeland decentralization points |
| Themes | manufacturing industrial policy, business and state, bantustans removals influx, inequality poverty unemployment, sanctions divestment, comparative theory, labour markets wages |
| Citation key | mccarthy1988 |
Argument¶
McCarthy reviews South African manufacturing's structural development within the evolution of industrial and trade policy. He argues that the sector, after rapid post-war growth, stagnated from the mid-1970s: its share of GDP and per capita income decoupled, capital intensity rose sharply, and productivity fell. Import-substituting industrialization changed the composition of output and imports (light to heavy industry) but did not reduce the economy's import propensity. Protection kept capital productivity low, while export promotion and 'inward industrialization' are each limited by sanctions, capital shortage and the size of the domestic market. He concludes that industrial development has reached an impasse that only political change can cure.
Evidence and method¶
Central Statistical Services manufacturing censuses, input-output tables and labour statistics; National Productivity Institute data on capital and productivity; World Bank and IMF data; commission and government reports (Kleu, Reynders, Technical Committee, White Paper); and secondary literature on industrialization (Chenery, Hoffman, Balassa, Krueger, Kleu, Marais, Viljoen).
Place in the debates¶
- Import substitution vs export promotion as the engine of industrialization (see Industrialisation, protection and industrial structure). Holds that import substitution has run its course and has not lowered the import coefficient. Exports are the orthodox alternative, but a massive export re-orientation is doubtful under sanctions, so he favours a dual approach of selective protection plus export incentives.
- Is manufacturing the answer to employment and poverty (labour-intensive light industry vs capital-intensive heavy industry)? (see Industrialisation, protection and industrial structure). Sceptical. Employment elasticities are mostly below one, heavy industries such as chemicals and metals created as many jobs as light industry, and manufacturing growth alone is no 'magic wand' for unemployment.
- Is South Africa a mature industrial economy (Viljoen, Marais) or still a developing economy? (see Industrialisation, protection and industrial structure). Rejects the 'maturity' reading. Since the mid-1970s the expected S-curve relationship between manufacturing share and income has broken down, and South Africa is a developing economy at risk of slipping down the income ladder.
- Neo-classical equilibrium vs structural-imbalance justification for industrial policy (see Industrialisation, protection and industrial structure). Accepts the case for deliberate industrial policy (including selective, 'made-to-measure' export assistance) to shift resources to higher-productivity uses.
- Industrial decentralization under separate development (see Homelands, separate development, decentralisation and removals). Critical on economic grounds. Heavy infrastructure spending at remote points, while metropolitan infrastructure such as the Western Cape's is under-utilized, makes job creation costly.
Findings¶
-
Manufacturing's share of GDP rose after the war to a peak in 1981, then declined, and the link with per capita income broke down after about 1975. (p. 2)
During the post-war period the share of manufacturing in GDP increased from 13 per cent in 1946 to a maximum of 22,2 per cent in 1981, thereafter declining consistently to 19,4 per cent in 1985.
-
Aggregate growth slowed steadily from the 1960s to the early 1980s, so per capita income fell. (p. 3)
average GDP growth fall consistently from the high average of 5,8 per cent during the sixties to an average of 3,4 per cent during the seventies and 1,1 per cent during the first half of the eighties.
-
Manufacturing became much more capital-intensive in 1970-86. (p. 3)
In manufacturing the capital-labour ratio increased at an average annual rate of 4,8 per cent during 1970-86, thus also doubling
-
Rising capital per worker did not raise productivity; multi-factor productivity fell, the 'productivity paradox'. (p. 3)
In all commodity producing sectors multi-factor productivity declined by 7,4 per cent during 1970-86 and in manufacturing by 2,4 per cent.
-
Industrialization shifted from light (consumer) to heavy (capital and intermediate goods) industry between 1925 and 1982. (p. 4)
Between 1925 and 1982 the share of light industry in total manufacturing value added declined from 63,2 per cent to 36,8 per cent, with heavy industry experiencing a commensurate increase.
-
Employment elasticities show that light consumer industries are not the best job creators, contrary to conventional wisdom. (p. 7)
it appears justifiable to conclude that the conventional wisdom which holds that consumer-orientated light industries such as food, clothing, textiles and furniture are the ideal vehicles for the promotion of direct employment opportunities may be doubted.
-
Import substitution changed the composition of imports but did not reduce the economy's import dependence (coefficients of 0,25 in 1920-24 and 0,24 in 1980-84). (p. 7)
It is clear that South Africa has not experienced a decline in its average propensity to import.
-
Homeland-oriented industrial decentralization is costly, since infrastructure is lumpy and agglomeration economies are absent. (p. 8)
the average cost of jobs created at many of these points is bound to be higher than may be thought wise in a less than prosperous developing country.
-
Manufactured exports showed no clear rise as a share of total exports in 1960-84. (p. 9)
the manufactured component varied between 30 and 45 per cent of total nongold exports and between 17 and 28 per cent of total exports during the period 1960-84.
-
Export assistance policy moved from uniform across-the-board incentives to selective, targeted support. (p. 10)
This is a fundamental deviation from the principle of uniformity contained in the current system.
-
'Inward industrialization' based on black urban demand is no panacea. (p. 11)
the only viewpoint that can be accepted with certainty is that inward industrialization is not a panacea
-
The author concludes that political change is the precondition for industrial recovery. (p. 11)
the malaise of South African industry can only be cured if political wisdom prevails.
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| Central Statistical Services national accounts series on GDP, per capita GDP and manufacturing share of GDP, 1946-1985 (Figure 1) | official statistical series | 1946-1985 | Establishes the rise and fall of the manufacturing share of GDP and its relationship to per capita GDP, used to test the S-curve. |
| Central Statistical Services data on manufacturing value added (net output) and employment by industry group, 1925-1982 (Table 1) | industrial mining census | 1925, 1940, 1946, 1960, 1970, 1982 | Shows the shift from light to heavy industry and computes relative productivity (value added share divided by employment share). |
| Manufacturing Censuses, Central Statistical Services (Table 2: real value added at constant 1980 prices and employment, 1970-1982) | industrial mining census | 1970-1982 | Used to compute sectoral contributions to growth and employment elasticities by manufacturing industry. |
| National Productivity Institute data on capital-labour ratios, labour productivity and multi-factor productivity, 1970-86 | official statistical series | 1970-1986 | Source for the growth rates of capital intensity, output per worker, capital output and multi-factor productivity in commodity sectors and manufacturing. |
| World Bank, World Development Report 1987 (income classification, labour force growth, GDP and population Tables 3 and 27) | international trade or finance data | 1980-1985 and projections to 2000 | Used for South Africa's per capita GNP classification, labour force growth, and projections of per capita GDP under different growth scenarios. |
| Central Statistical Services, Input-Output Tables 1985 and various input-output tables (import shares in supply) | official statistical series | 1960-1985 | Used to compute import shares in domestic supply by industry and import composition in 1985. |
| Import coefficients (imports/GDP) for South Africa for five-year periods 1920-24 to 1980-84, and IMF International Financial Statistics Yearbook 1986 for Latin American comparators | international trade or finance data | 1920-1984 | Shows South Africa's import propensity did not decline, so import substitution in the ratio sense did not occur, and compares it with Latin America. |
| South African export statistics by SITC category, 1960-84 (manufactured exports share) | official statistical series | 1960-1984 | Used to show manufactured exports remained 30-45% of non-gold exports with no clear increase, and that basic metals dominate. |
| Van Zyl (1984), 'South Africa in World Trade', SAJE, data on fully manufactured exports | secondary literature compilation | 1970-1982 | Supplies an alternative raw-material-content definition of manufactured exports and their share in non-gold exports. |
| Report of the Commission of Inquiry into the Export Trade (Reynders Commission), RP 69/72 | commission or official report | 1972 | Documents the official shift from import replacement to exports as the engine of growth. |
| Technical Committee of Inquiry into the Practicality of the Van Huyssteen Group Proposals, Export Incentives in the Republic of South Africa (1978) | commission or official report | 1978 | Shows the origin of the uniform Categories A to D export assistance system. |
| Kleu Report: Report of the Study Group on Industrial Development Strategy (1977 appointment; report 1983) | commission or official report | 1977-1983 | Source on factor intensity of light versus heavy industry, effective protection rates, and the moderate selective protection policy. |
| White Paper following the Kleu Report (1985) | commission or official report | 1985 | Shows the government's stated dual approach of moderate protection plus export promotion. |
| Interim report of the Kleu committee on export assistance (1986) and Ministerial press release of 18 March 1986 | commission or official report | 1985-1986 | Documents the move to selectivity ('made-to-measure') in export assistance. |
| Board of Trade and Industries, 'Guide to the policy and procedure relative to customs tariff protection and tariff relief' (undated brochure) | commission or official report | undated | Source of the three tenets of South African tariff protection policy. |
| Dreyer and Brand (1986) simulation of development scenarios | secondary literature compilation | 1986 | Used to support the claim that food, clothing and leather industries would benefit from more equal income distribution and urbanization. |
| Central Statistical Services, South African Labour Statistics and South African Statistics (1986) | official statistical series | to 1986 | Listed as general statistical sources for employment and economic series behind the growth, employment and non-agricultural employment figures. |
Inference¶
Quantification: Descriptive tables.
Descriptive time-series and cross-period comparison of official aggregate statistics (manufacturing shares, productivity, import coefficients, export shares), set against stylised theory (Chenery S-curve, import-substitution phases) and a chronological review of industrial policy reports; no regression or causal identification, except a simple elasticity calculation and growth-scenario extrapolations.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| The manufacturing share of GDP rose to a peak in 1981 and then declined, and the positive relation between manufacturing share and per capita income vanished after about 1975, so South Africa does not fit the S-curve. (p. 2) | Central Statistical Services GDP and manufacturing series plotted in Figure 1. | Systematic data |
| Capital intensity rose sharply in 1970-86 while labour productivity and multi-factor productivity stagnated or fell, indicating underutilization of capital. (p. 3) | National Productivity Institute series on capital-labour ratios, output per worker and multi-factor productivity. | Partial data |
| Industrialization followed the standard shift from light to heavy (consumer to capital and intermediate goods) industry between 1925 and 1982, with light industry's share of value added falling from 63,2 to 36,8 per cent. (p. 4) | Table 1 from Central Statistical Services manufacturing value added and employment data, using the Chenery-Syrquin light/heavy classification. | Systematic data |
| Import substitution changed the composition of imports and manufacturing structure but did not reduce the economy's overall import propensity. (p. 7) | Import/GDP coefficients for five-year periods 1920-84 plus input-output based import shares by industry. | Partial data |
| Manufactured exports show no clear rising share in 1960-84 and manufacturing is generally not export-oriented except basic metals. (p. 9) | SITC-based export shares 1960-84 (data held by the author) and Van Zyl's raw-material-content definition. | Partial data |
| Growth of 1975-85 or 1980-85 extended forward would cut South Africa's per capita GDP by 2000, so the country would fall down the income ladder. (p. 3) | Extrapolation from World Bank GDP and population data (Tables 3 and 27). | Illustrative or anecdotal |
| Protection is a probable final cause of low factor productivity, and inward industrialization is not a panacea, claims resting on reasoning rather than tested evidence. (p. 11) | Argument from trade theory and policy review; no direct test. | Asserted |
The author's caveats¶
Notes that classification into light and heavy industry involves an arbitrary element; that footnoted rounding means columns do not sum to 100; and that the employment elasticity is sensitive to method (0,67 versus 0,48 with exponential fitting). Detailed import and export share data are available only from the author.
Gaps in the evidence¶
The argument that protection caused low productivity and that import substitution is exhausted is not tested with firm- or industry-level data on effective protection, productivity and trade, which the Board of Trade and Industries and the manufacturing censuses could have supplied. The geographic and decentralization claims are asserted without regional output or employment data, which the censuses contain. Disaggregated import and export series and the National Productivity Institute inputs are not reproduced, so the productivity and import-share findings cannot be checked from the paper.
Replicability: Data described but not public.