The economics of isolation, trade and investment: case studies from Taiwan & apartheid South Africa¶
Kerby, Edward M. (2016). The economics of isolation, trade and investment: case studies from Taiwan & apartheid South Africa. PhD thesis, Department of Economic History, London School of Economics and Political Science. London School of Economics and Political Science.
| Type | thesis |
| Scope | Economics and economic history of apartheid |
| Method | mixed |
| Period studied | 1948–1996 |
| Regions | South Africa, Bantustans/homelands (Transkei, Ciskei, KwaZulu, QwaQwa, Bophuthatswana etc.), Taiwan (ROC), Free State (Botshabelo-Thaba'Nchu), Newcastle-Madadeni, Eastern Cape (Dimbaza, Butterworth), PWV |
| Themes | manufacturing industrial policy, bantustans removals influx, labour markets wages, business and state, sanctions divestment, macro fiscal finance debt, origins segregation, methods data sources, inequality poverty unemployment |
| Citation key | kerby2016a |
Argument¶
The thesis uses South Africa's Regional Industrial Development Programme (RIDP) and the arrival of Taiwanese investors as a lens on late-apartheid economic history. It reconstructs three phases of decentralisation (Grand Apartheid 1948-69, Growth and Benevolence 1970-79, Desperation 1980-92) and argues that the programme largely failed to create homeland employment or domestic clustering. Taiwanese entry was not just a product of pariah-state diplomacy and subsidies: Taiwanese offshoring pressures (currency appreciation, Multi-Fibre Agreement quotas), improved property protection and 'honorary white' status, a weak rand, and low-wage female labour all mattered. Taiwanese migrants also raised and diversified exports to Asia via a 'migrant-trade effect'. Firm clustering in a few zones is explained by the interaction of economic geography, incentive variation and co-ethnic production networks, with new archival data on wages, incentives, firms, trade and migration.
Evidence and method¶
Declassified archives in Pretoria (Board for the Decentralisation of Industry annual reports, business licence applications), Cape Town (Library of Parliament, UCT Government Publications), IDC Sandton, Kew, homeland archives (Transkei, KwaZulu, QwaQwa), Hansard, SARB foreign liabilities registers, Foreign Trade Statistics, SARS data, Garner (1994), IMF DOTS, ATCC almanac, SALDRU newspaper clippings, Carnegie working papers, census enumeration areas, GIS-geocoded RIDP zones, and interview transcripts (PARI/Eddie Sun, Ming Wang).
Place in the debates¶
- Liberal vs radical/revisionist interpretations of apartheid and capitalism (decentralisation as political racial engineering vs economic competitiveness) (see Homelands, separate development, decentralisation and removals). Argues some truth lies on both sides; the RIDP was a political tool to anchor Africans in the homelands but later phases were also driven by economic pressures, and Taiwanese investment exposes both logics.
- Incentives vs spontaneous decentralisation (Bell) as drivers of industrial dispersal (see Homelands, separate development, decentralisation and removals). Sides with Bell that incentives were less decisive than commonly thought in Phase 2 (incentive outlays flat, infrastructure expanded), while in Phase 3 incentives rose sharply; for Taiwanese firms he argues geography and networks interacted with incentives rather than incentives alone explaining clustering.
- Why Taiwanese investment came to South Africa: diplomatic/pariah-state alliance vs economic motives (Geldenhuys, Pickles and Woods) (see Homelands, separate development, decentralisation and removals). Challenges the diplomatic-only reading; stresses Taiwanese offshoring, MFA quotas, exchange rates, financial liberalisation, property-rights protection and low-cost labour as complementary economic drivers.
- Sanctions, trade diversion and sanction-busting (Garner 1994) (see Homelands, separate development, decentralisation and removals). Finds exports to Taiwan growing fastest, partly consistent with re-export/conduit suspicions for homogeneous minerals, but differentiated goods growth cannot be explained by this and is associated with Taiwanese investors (not causal).
- Determinants of foreign firm agglomeration (economic geography vs incentives vs co-ethnic networks) (see Homelands, separate development, decentralisation and removals). Argues all three interacted; first-mover Taiwanese production networks, market access and female labour pools explain clustering in four of sixty-three zones, with incentives amplifying rather than explaining it.
Findings¶
- The Tomlinson plan was drastically underfunded by the state and received no private capital. (p. 45)
-
Phase 2 decentralisation incentives were roughly flat, so infrastructure rather than rising incentives drove dispersal. (p. 52)
Table 2.2 demonstrates that disbursements of incentives during the second phase were approximately flat.109 Instead of expanding incentives (i.e Pull Factors), growth in decentralisation may have been driven by an expansion of infrastructure
-
Infrastructure costs for the homeland estates were large relative to GDP. (p. 54)
the costs of constructing this infrastructure equated to 1.5 % of GDP in each year from 1972 - 1979.
-
Only a small share of decentralisation jobs in Phase 2 were in homelands or border areas. (p. 56)
approximately 10 % of the decentralisation employment (i.e. 110,597 jobs) was in the homelands or adjacent zones (i.e. border areas).
-
Domestic manufacturers showed little interest in relocating to decentralised zones despite higher incentives. (p. 66)
only 35 % of the sample indicating they would consider relocating between 1980 - 1985.
-
Taiwan dominated foreign RIDP applications, rising from a majority to nearly nine-tenths. (p. 68)
The results show that 57 % of the RIDP applicants before 1985 were Taiwanese industrialists. These results are demonstrable of Taiwan’s importance, growing to 88 % of the applicants by 1992.
-
The RIDP overall failed on employment and cost, with high incentive cost per job. (p. 69)
based on the employment levels, each new job cost the national budget R 8000 in incentives.
-
By 1992 the RIDP had created far fewer jobs than projected. (p. 70)
the RIDP appears to have been a failure with only 214,394 additional jobs created by 1992. This is far below the estimated 300,000 that the RIDP had projected for the Taiwanese cohort alone.
-
Taiwan held the majority of Asian foreign liabilities in South Africa in 1992, mostly in long-term assets. (p. 97)
Taiwan is found to posses R 3,113 million or 58 % of all the Asian investment in South Africa.
-
Taiwanese firms clustered heavily in a few of the sixty-three RIDP sites. (p. 151)
only 11 of the 63 RIDP sites attracted 84 % of all Taiwanese investment
-
Wage-related incentives and geography vary across zones, with female labour exceeding male in most homeland zones. (p. 159)
In most zones, the ratio of females to males exceeded two to one.
-
Differentiated exports to Taiwan grew sharply during the sanctions years, alongside Taiwanese migration. (p. 115)
the share of total differentiated products in South Africa’s exports escalated from 0.9 % to 4.93 % in favour of Taiwan.
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| Board for the Decentralisation of Industry (BDI) Annual Reports, 1970-1992, incl. incentive and concession disbursement series (National Archive Pretoria, TAB 7869-ZA etc.) | commission or official report | 1970-1992 | Builds the time series of incentives and concessions (Tables 2.2, 5.3, 5.5; Fig 2.7) used to show flat Phase 2 incentives and the Good Hope surge, and the incentive mix in Chapter 5. |
| BDI business licence applications and declassified BDI firm records (National Archive Pretoria, TAB 9879/987; TAB 39149-ZA) | state archives | 1979-1992 | Supplies foreign and Taiwanese licence applications (Tables 1.1, 2.9), the annual Taiwanese migration proxy, and firm location, sector and start-date data for the agglomeration analysis. |
| SALDRU newspaper clippings collection (UCT), 1975-2000, incl. Financial Mail, Star, Natal Witness, Mercury, Rand Daily Mail articles | newspapers periodicals | 1975-2000 | Principal source for reported unskilled wage estimates in 57 of 63 RIDP zones, forming the new wage series. |
| Homeland trade periodicals and development corporation almanacs (Transkei Today, KwaZulu 'Development', KFC, TDC, Free State DC publications 1978-1994) | company records or reports | 1978-1994 | Used for wage data and qualitative accounts of Taiwanese firms, incentives and location choice, treated with caution as propaganda. |
| Bureau for Economic Research: Bantu Development (1976) wage report and Gottschalk (1977) wage survey | commission or official report | 1971-1975 | Official and survey wage benchmarks used to validate the newspaper and almanac wage data and fill the 1970 baseline. |
| Second Carnegie Inquiry into Poverty working papers (Thomas 1984, Abedian 1983, Ardington 1984, Bank 1984, McCarthy 1984) and Wilson & Ramphele (1989) | commission or official report | 1980s | Five archived working reports supply early-1980s wage estimates and employment effects for the wage series. |
| Foreign Trade Statistics 1975-1985 (Department of Customs and Excise/Department of Trade and Industry) and declassified DTI annual returns 1986-1992 | official statistical series | 1975-1992 | Digitised to build the new bilateral trade series by SIC/HS category used in Chapter 4. |
| South African Revenue Service (SARS) balance of trade database, 1992-1995 (HS codes) | official statistical series | 1991-1995 | Supplies post-sanctions monthly/annual trade flows to extend the series to 1995. |
| Garner (1994) South Africa's trade partners 1986-1991 and IMF Direction of Trade Statistics | secondary literature compilation | 1986-1991 | Used to cross-validate and supplement the embargoed 1986-1990 trade years and to establish baselines. |
| South African Reserve Bank Quarterly Bulletin, register of foreign capital liabilities (Appendix S71), 1979-1997 | official statistical series | 1979-1997 | Used to show Asian and Taiwanese inward direct versus non-direct liabilities and permanence of Taiwanese investment. |
| SARB national accounts series: gross fixed capital formation, GDP, exchange rates, gold and mining shares | national accounts macro | 1950-1999 | Supports the infrastructure-versus-incentive argument, phase periodisation and rand/NT$ exchange-rate context. |
| Permanent Committee for the Location of Industry reports (1969, 1971), National Archive Pretoria TAB 62657-ZA87 | commission or official report | 1960-1970 | Source for Phase 1 state assistance totals (Table 2.1), the allocation to state monopolies and employment outcomes. |
| Bantu Investment Corporation (1975) Homelands: the role of the corporations | company records or reports | 1959-1975 | Used for BIC capitalisation, 127 agency factories, and Phase 1 job creation figures. |
| South African Government Information Services (1979) Economics and Finance in South Africa's Black Territories, National Archives Pretoria | official statistical series | 1972-1979 | Source for infrastructure spending (Table 2.4) and Phase 2 employment and capital expenditure (Table 2.5). |
| Industrial Development Corporation archive, Sandton (incl. Federated Chamber of Industries archival reports 1972 and 1985; foreign-owned plant data) | employer association records | 1972-1985 | Evidence on industry opposition to the push policy, voluntary decentralisation, and foreign-owned manufacturing employment (Table 2.6). |
| South African Parliamentary Hansard (1972, 1977, 1978, 1979, 1982, 1986) | parliamentary debates | 1972-1986 | Ministerial and MP statements on decentralisation aims, incentives increase and Taiwan productivity comparisons. |
| Study Group on Industrial Development Strategy (1983) and Good Hope Plan documents; Government Gazette on RIDP incentives | commission or official report | 1979-1983 | Documents the Good Hope Plan design, consultation and incentive schedule; Gazette restrictions on take-overs. |
| RIDP Schedule of Incentives and BDI pamphlet on proposed RIDP incentives (1981), TAB 7869-ZA/09 | legislation gazette | 1981-1992 | Zone-level incentive schedule (wage rebates, rail rebates, interest subsidies) used to compute the Rebate Wage Gap and transport costs. |
| South African Railways and Harbour Gazette (1980) and SAR&H annual reports 1978-1980, Transnet Library | official statistical series | 1978-1980 | Rail tariffs combined with GIS path-distance to estimate container transport costs per zone. |
| Digitised archival maps and GIS data (Board for Decentralisation maps, Manufacturing Census Maps, RIDP maps) | maps spatial data | 1970s-1980s | Geocoding of 63 RIDP zones and least-cost path distance to ports. |
| Declassified UK Whitehall report 'United Kingdom's Interests in Southern Africa' (1980), National Archives Kew | state archives | 1980 | Corroborates the scale of western investment and trade ties with South Africa. |
| Black (1984)/Wellings and Black (1986a) survey of 210 domestic manufacturing firms | secondary literature compilation | 1980-1985 | Only national firm survey, showing domestic firms' reluctance to relocate (Table 2.8). |
| ROC-RSA Bilateral Trade Agreement 1975 and related agreements (Government Gazette series); Taiwanese Liaison Office reports | legislation gazette | 1975-1982 | Shows MFN status and diplomatic protection underpinning Taiwanese investment. |
New data¶
Author-built datasets from archives: bilateral South African trade series 1975-1995 for seven partners by SIC/HS category; annual Taiwanese migration series from BDI licences; 63-zone RIDP wage series 1978-1991 from newspapers, almanacs and Carnegie papers; zone-level demographic imbalance dataset from 1980 census enumeration areas; RIDP incentive time series 1971-1992; geocoded database of 306 Taiwanese firms (1992) with sector and start year; GIS rail-cost estimates for 63 zones.
Inference¶
Quantification: Descriptive tables.
Archival case study using descriptive tables, time series and cross-zone comparison (Chapters 2-4); Chapter 5 location-choice analysis rests on ranked tables, firm counts, GIS-based transport costs and a rebate wage gap, combined with qualitative archival accounts and quotations. No regression; author states results are not causal.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| Decentralisation incentive disbursements were approximately flat in Phase 2, so growth in decentralisation was driven more by infrastructure than by incentives; incentives rose steeply only after the 1982 Good Hope Plan. (p. 52) | BDI annual-report series (Table 2.2, Fig 2.7) and SARB gross fixed capital formation index (Fig 2.3) plus IDC cost estimate. | Partial data |
| The RIDP failed to create employment: only 214,394 additional jobs by 1992, far below projections, while Taiwanese firms were about 10% of 4,164 RIDP firms. (p. 70) | BDI 1992 firm and employment summary (Table 2.10) and BDI annual report. | Partial data |
| Taiwanese applicants dominated foreign RIDP applications: 57% before 1985 and 88% by 1992, representing greenfield rather than takeover investment. (p. 68) | BDI foreign business licence applications (Table 2.9) and Government Gazette restrictions. | Partial data |
| Taiwanese investors held longer-term, more permanent assets in South Africa than other Asian or western investors, indicating greater risk tolerance linked to diplomatic protection. (p. 97) | SARB foreign liabilities register, 1992 country breakdown only (Tables 3.4-3.5); country data unavailable before 1992. | Partial data |
| Taiwanese migration and investment were accompanied by diversification of South African exports to Taiwan from homogeneous to differentiated goods (migrant-trade effect), though not causally established. (p. 121) | New trade series by SIC (Figs 4.3-4.4, Tables 4.2, D.5), BDI-derived migration proxy and archival accounts. | Partial data |
| Taiwanese firms clustered heavily: 11 of 63 RIDP sites attracted 84% of Taiwanese firms, with three industrial complexes holding 65%, versus 29% of domestic firms. (p. 151) | Geocoded 1992 firm data from BDI and ATCC almanac (Table 5.6). | Systematic data |
| Phase 1 decentralisation delivered little: the programme received only £48m of the £104m proposed and none from private capital, with most funds going to peri-urban areas and state monopolies. (p. 45) | Permanent Committee reports (Table 2.1) and Tomlinson Commission report. | Partial data |
The author's caveats¶
Author stresses that apartheid-era trade, investment and migration data were embargoed or unreliable, African wages sparse, and homeland censuses inadequate. Wage data are triangulated from newspapers and promotional almanacs treated 'with caution'. Migration figures are a conservative proxy from licence applications; SARB country data exist only from 1992; trade cannot be attributed to specific areas; results are described as 'not causal'. Case-study limits and sparse data prevent robust statistical modelling.
Gaps in the evidence¶
The central network and migrant-trade claims rest on correlations in aggregate trade and firm counts without firm-level export data, controls or a counterfactual, so causal identification is weak. Zone-level wage data are sparse, heterogeneous press and almanac figures with unequal observations. Firm-level export and employment records, homeland development corporation ledgers and Taiwanese-side records (customs, investment approvals) would better test location and trade effects; some of these exist in Taiwanese and South African archives, and a regression with zone and sector fixed effects could be run on the assembled firm panel.
Replicability: Data described but not public.