Where Has All the Growth Gone? South African Manufacturing Industry 1970-2000¶
Trevor Bell; Nkosi Madula (2001). Where Has All the Growth Gone? South African Manufacturing Industry 1970-2000. NIEP 2001 Annual Forum, Misty Hills, Muldersdrift, 10-12 September 2001. National Institute for Economic Policy (NIEP).
| Type | conference paper |
| Scope | Economics and economic history of apartheid |
| Method | descriptive quantitative |
| Period studied | 1911–2000 |
| Regions | South Africa |
| Themes | manufacturing industrial policy, macro fiscal finance debt, labour markets wages, business and state, comparative theory, methods data sources, education human capital |
| Citation key | bell2001 |
Argument¶
The paper's central fact is that South African manufacturing growth deteriorated from the early 1970s and was especially poor from the early 1980s. The authors argue the 1970s gold-led commodity boom masked real forces slowing growth: falling physical gold output, an OECD/world slowdown, and a real rand appreciation that hurt downstream durable-goods exports while the investment boom sustained their output. After the 1985 debt crisis, South Africa shifted abruptly to export-oriented industrialisation, but exports rose while output stagnated. They attribute this to a foreign-exchange constraint, falling output-capital ratios and persistent excess capacity, with the capital-goods sectors acting as a barometer of the investment rate. They doubt that a skills shortage is the key cause, weigh global explanations (Brenner, Gordon, Easterly) against South African ones, and conclude that manufacturing still matters.
Evidence and method¶
South African Reserve Bank and WEFA national accounts, trade and capital-stock data; T A du Plessis (1965) for 1916-57; Bell and Farrell (1997) and Bell et al (1999); Bhorat and Edwards employment data; international growth comparisons from Pritchett, Easterly, Brenner, Gordon and Nordhaus. Capacity output is estimated with an adapted Panic/Christiano method and checked against Reserve Bank capacity-utilisation indices.
Place in the debates¶
- Causes of manufacturing stagnation: skills shortage vs demand-side and structural factors (see Did the apartheid economy fail? Growth, productivity, investment and the post-1970s crisis). 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.
- Foreign exchange constraint on growth (see Did the apartheid economy fail? Growth, productivity, investment and the post-1970s crisis). 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.
- Sasol investments as crowding out other investment (see Industrialisation, protection and industrial structure). Questions the view that Sasol II and III distorted allocation and slowed MVA growth; they raised the overall investment/GDP ratio rather than reallocating a fixed pool.
- Global slowdown explanations (Brenner falling profit rate, Gordon's 'one big wave', Easterly's skill-biased technical change) (see Did the apartheid economy fail? Growth, productivity, investment and the post-1970s crisis). Treats these as partial; they may explain world decline since about 1973 but not the sharper developing-country decline since the early 1980s.
- Does manufacturing still matter in the information age? (see Industrialisation, protection and industrial structure). Argues manufacturing remains vital because services growth in South Africa came from job losses elsewhere, and growth cannot revive without durable-goods recovery.
- Resource abundance as curse or blessing (see Industrialisation, protection and industrial structure). Suggests natural-resource dependence may now be 'more of a blessing and less of a curse', following Wood and Mayer, while still needing diversification.
Findings¶
-
Manufacturing grew very fast under import-substituting industrialisation, with MVA growth peaking in 1960-65 and averaging 8.6% in the 1960s. (p. 6)
Both the MVA and GDP growth rates, however, reached their peaks at 9.9 per cent and 6.3 per cent during 1960-65, with average annual growth rates during the 1960s as a whole of 8.6 per cent and 5.7 per cent respectively.
-
The commodity boom caused a large real appreciation of the rand, which hurt the competitiveness of downstream manufactures. (p. 10)
Relative to its level in 1970-72, the real effective exchange rate (REER) was 9.4 per cent higher in 1974-78, 24.8 per cent higher in 1979-81, and 28.2 per cent higher in 1982-83.
-
The Sasol II and III investment surge in the 1970s was driven by high oil prices, low real interest rates and a strong rand. (p. 13)
the share of the chemicals sector in manufacturing gross fixed capital formation increased from 7.2 per cent in 1974 to 55.7 per cent in 1979
-
The debt crisis forced an abrupt, involuntary shift to export-oriented industrialisation, with investment falling sharply. (p. 22)
investment, which fell by 20 per cent between 1984 and 1986
-
Rising export orientation reflects stagnant output rather than manufacturing success. (p. 23)
The increases in the degree of export orientation of manufacturing industry, as measured by the ratio of exports to total sales, noted above, thus, have been largely due to the stagnation of manufacturing output.
-
The durable-goods sectors with the fastest export growth after 1985 had falling real value added, while natural-resource-based manufacturing kept growing. (p. 25)
Value-added in these sectors fell at an average annual rate of 2.1 per cent in 1980-90 and at 1.1 per cent in 1990-98.
-
Falling physical gold output explains part of the early-1970s growth slowdown. (p. 27)
the output of gold in physical terms began falling after 1970, and in the year 2000 was only 43 per cent of its 1970 level.
-
In 1985-90 export growth in constant US dollars overstated the real foreign-exchange gain; trade-weighted total exports actually fell as gold exports declined. (p. 30)
In constant price trade-weighted currency units, a more appropriate measure, South Africa's total exports, in 1985-90, fell at 0.8 per cent a year, owing to an 8.2 per cent a year decline in gold exports
-
Manufacturing has persistent excess capacity because its output-capital ratio kept declining even as its capital stock grew fastest of any sector. (p. 33)
though the real fixed capital stock grew at 2.89 per cent a year, in manufacturing industry, faster than in any other main economic sector in 1992-2000, and at only 1.25 per cent a year in the economy as a whole
-
The skills-shortage thesis is challenged by the fast growth of the more skills-intensive FIRE sector. (p. 36)
how come that the output and employment growth performance of the FIRE sector, a far more skills-intensive sector, was so vastly superior to that of manufacturing, especially in the 1990s?
-
Employment fell in both manufacturing and services in the 1990s, so services did not offset manufacturing job losses as in the US. (p. 45)
Between 1990 and 1998, employment fell in both manufacturing (by 161827 or 10.6 per cent), and services as a whole (by 71241 or 3.2 per cent).
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| South African Reserve Bank database (national accounts: GDP, sectoral value added, MVA, exports/imports, fixed capital stock, capacity utilisation index) 1946-2000 | national accounts macro | 1946-2000 | Underlies the growth-rate calculations for GDP, MVA, mining value added, exports/imports and capital stock in Tables 1, 4, A1, A5 and Figures 1, 2, A1-A3. |
| WEFA database (sectoral value added, manufactured exports, employment, 3-digit SIC) | secondary literature compilation | 1970-1998 | Source of sectoral growth rates and shares for manufacturing value added and exports (Tables 2, 3, A2-A4, A6) and employment after 1970. |
| T A du Plessis (1965), The Industrial Growth Pattern and the Real Forces of Economic Expansion in South Africa, 1916/17-1956/57 (unpublished D.Comm. thesis, UNISA) | secondary literature compilation | 1916/17-1956/57 | Provides early-period GDP and MVA growth rates and manufacturing share of GDP and of exports used to describe pre-1970 import-substituting industrialisation. |
| South African Statistics 1990 | official statistical series | 1920-1970 | Source of manufacturing employment growth rates for 1920-1970 in Table 1. |
| Bell and Farrell (1997) 'The Minerals-energy Complex and South African Industrialisation' (import/domestic supply ratios, import substitution, sectoral MVA shares) | secondary literature compilation | 1916/7-1980 | Supplies import/domestic supply ratios, import-substitution contributions (incl. motor vehicles) and sectoral structure of MVA for pre-1980 periods. |
| Bell, Farrell and Cassim (1999) 'Competitiveness, International Trade and Finance in a Minerals-Rich Economy' | secondary literature compilation | 1960-1998 | Source of export growth rates 1960-70, real exchange rate estimates for sector groups, evidence of foreign-exchange constraint, and imported-capital-goods ratio. |
| Reynders Commission (1972), Report of the Commission of Inquiry into the Export Trade of the Republic of South Africa, R.P. 69/1972 | commission or official report | 1969-1972 | Cited as evidence that policy attention turned to export diversification in 1969-72 before the commodity boom overtook it. |
| South African Reserve Bank (2001), South Africa's Balance of Payments 1946-2000 (Quarterly Bulletin Supplement) incl. nominal effective exchange rates | official statistical series | 1946-2000 | Source of nominal effective exchange rates used to convert rand exports into constant-price trade-weighted foreign currency values. |
| Reserve Bank index of capacity utilisation in manufacturing (originally from Statistics South Africa surveys) | official statistical series | 1970-2000 | Used to validate the authors' own estimated capacity utilisation series. |
| Nattrass (1990) Table 7.1 profit rates in manufacturing | secondary literature compilation | 1964-1975 | Provides declining manufacturing profit rates to assess Brenner's falling-profit thesis for South Africa. |
| Edwards (2000) 'Globalisation and the Skill Bias of Occupational Employment in South Africa' (occupational employment data 1984-1997) | secondary literature compilation | 1984-1997 | Evidence of rising skilled employment share across sectors, used in the skills-shortage discussion. |
| Bhorat (2001) Employment Trends in South Africa, and Bhorat (2000) wage differentials | secondary literature compilation | 1993-1997 | Skill-level employment changes by sector (manufacturing vs FIRE) and wage differentials used to test the skills-shortage explanation. |
| International growth comparisons: Pritchett (2000) and Easterly (2001); Wood and Mayer (1998) schooling data | secondary literature compilation | 1960-1998 | Benchmarks South Africa's per capita growth slowdown against developed and developing country averages. |
| US productivity studies: Jorgenson and Stiroh (2000), Nordhaus (2001), Gordon (2000), Brenner (1998), DeLong (2001) | secondary literature compilation | 1958-2000 | Used as comparators for the role of IT-related manufacturing sectors (SIC 35, 36) and for global long-downswing hypotheses. |
| McCarthy (1999) 'The Manufacturing Sector during the 1970s'; World Bank (1987) Korea data | secondary literature compilation | 1970s | McCarthy supports ISCOR expansion timing; World Bank supplies South Korean MVA structure for comparison. |
New data¶
Author-estimated series of potential (capacity) manufacturing value added, potential output-capital ratios and capacity utilisation for 1946-2000, and constant-price trade-weighted foreign currency export series, constructed from Reserve Bank and WEFA data.
Inference¶
Quantification: Descriptive tables.
Descriptive macro-historical interpretation based on comparison of period average growth rates (sectoral MVA, exports, investment) across sub-periods, supplemented by international comparisons, a simple capacity-output estimate (Panic/Christiano adaptation) and engagement with theories (Brenner, Gordon, Easterly); no formal regression or causal identification.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| South African MVA and GDP growth declined from the 1960-65 peak, falling further in the 1970s and sharply to near stagnation from the early 1980s through 1990s. (p. 9) | Table 1 growth rates calculated from Reserve Bank, du Plessis and WEFA data. | Systematic data |
| The 1970s commodity boom appreciated the real exchange rate and shifted manufactured exports toward natural-resource-based sectors, harming downstream durable goods exports, while spending effects boosted durable goods output via investment. (p. 15) | Table 2 and A2 export growth rates and shares, REER figures, Table 3 value added growth and fixed capital formation shares. | Partial data |
| After 1985 export growth and export orientation rose, particularly in durable goods, but accompanied by output stagnation or decline, reflecting export-oriented stagnation tied to falling investment. (p. 23) | Tables 2 and 3, export/sales and export/output ratios computed from WEFA/Reserve Bank data. | Partial data |
| Declining physical gold output and mining value added in 1970-75 explain part of the fall in GDP growth and the sharper South African per capita growth slowdown relative to other developing countries. (p. 26) | Mining value added series (Figure 1, Table A5), gold output figures and a back-of-envelope calculation on mining share of GDP; comparison with Pritchett's cross-country growth figures. | Partial data |
| Since the early 1980s a foreign exchange constraint and persistent excess capacity in manufacturing (falling output-capital ratio) have limited growth. (p. 33) | Table 4 export/import growth, current account-GDP relationship from Bell et al 1999, estimated potential MVA and capacity utilisation (Figures 2, A2, A3). | Partial data |
| A skills shortage is probably not the fundamental cause of manufacturing stagnation, since the more skills-intensive FIRE sector grew rapidly in output and employment. (p. 36) | Edwards (2000) and Bhorat (2001) employment by skill category; Table A5 value-added growth. | Partial data |
| The slowdown in South African per capita growth since the early 1980s is of about the same magnitude as for developing countries generally, and global factors (Brenner, skill-biased technical change) may be relevant. (p. 25) | Comparison of South African per capita growth with Easterly's median; theoretical discussion of Brenner, Gordon, Nordhaus, Jorgenson and Stiroh. | Illustrative or anecdotal |
The author's caveats¶
Notes that hard data on manufacturing before WWI are hard to come by; acknowledges that profit-rate decline cannot be tied to output prices and comparable services data are unavailable; flags doubts over sustainability of 1990s export growth and the lack of detail in Easterly's skill-bias suggestion; notes that conversion of exports to constant-price trade-weighted currency is needed because of US dollar swings.
Gaps in the evidence¶
The argument leans on aggregate period averages from the Reserve Bank and WEFA databases and on other studies' figures without firm-level, sector-level or microdata to test mechanisms such as investment sensitivity, user cost of capital, or skills constraints. Causal claims about the foreign exchange constraint, the gold-decline effect and global factors are not tested econometrically; Brenner's profit-rate thesis lacks comparable sector profit data. Plant- or firm-level data, input-output tables, tariff and trade-policy series and sectoral profit rates exist from official sources and could now test these claims.
Replicability: Data poorly documented.