Unemployment in South Africa: Trends, Causes and Cures¶
Trevor Bell; Vishnu Padayachee (1984). Unemployment in South Africa: Trends, Causes and Cures. Second Carnegie Inquiry into Poverty and Development in Southern Africa, Carnegie Conference Paper No. 119. Cape Town (Second Carnegie Inquiry); Institute for Social and Economic Research, University of Durban-Westville.
| Type | conference paper |
| Scope | Economics and economic history of apartheid |
| Method | descriptive quantitative |
| Period studied | 1946–1983 |
| Regions | South Africa, metropolitan areas, decentralisation/subsidised regions, United Kingdom and OECD (comparison) |
| Themes | inequality poverty unemployment, labour markets wages, manufacturing industrial policy, macro fiscal finance debt, methods data sources, bantustans removals influx |
| Citation key | bell1984 |
Argument¶
The paper defends Simkins' estimates of Black unemployment against criticism that they are cyclically insensitive and falsely imply a rising long-term trend. It argues that Black unemployment rose structurally from about 1967 and non-Black unemployment from about 1974. The main cause is a slowdown in the growth of labour demand, tied to falling economic growth, rather than labour supply shifts or capital-labour substitution driven by low real interest rates. The slowdown is traced to global supply-side shocks (raw material prices, slower technology diffusion, competition from newly industrialising countries) and to gold booms that raised the rand. On policy, the authors play down interest-rate, inflation, UIF and privatisation remedies. They give regional policy a limited role and stress foreign exchange and trade policy.
Evidence and method¶
Simkins' unemployment rate estimates for 1960-81 (from his 1982 revision), registered unemployment for Asians, Coloureds and Whites 1948-83, UK unemployment rates, and OECD trends. Also South African Reserve Bank national accounts (1946-80) and South African Statistics/Quarterly Bulletin series on non-agricultural employment, population growth, real government bond rates, and manufacturing employment, output and capital stock.
Place in the debates¶
- Measurement and trend of Black unemployment (Simkins vs critics such as the Current Population Survey-based estimates) (see Unemployment: causes, measurement and labour market institutions). Defends Simkins' series as cyclically sensitive and indicating a genuine structural upward trend; argues absolute levels depend on arbitrary definitions and matter less than movements over time.
- Capital-labour substitution and factor-price distortions (low real interest rates, overvalued exchange rate) as a cause of unemployment (see Unemployment: causes, measurement and labour market institutions). Rejects it as the main cause: low interest rates also have a scale effect, and manufacturing data show employment slowed mainly because output and capital stock growth fell, not because capital intensity rose.
- Structural versus cyclical, voluntary versus involuntary unemployment (see Unemployment: causes, measurement and labour market institutions). Holds that long-term involuntary structural unemployment is real and explicable by global and domestic growth slowdown.
- Policy remedies: privatisation, restrictive monetary policy, regional decentralisation, informal sector (see Unemployment: causes, measurement and labour market institutions). Sceptical of privatisation, restrictive monetary and fiscal policy, and reliance on the informal sector. Gives regional policy only limited value, and sees foreign exchange and trade policy as crucial.
Findings¶
-
Simkins' series shows Black unemployment at its lowest in 1967 and rising strongly thereafter, implying a structural break in the late 1960s. (p. 4)
the Black unemployment rate fell to its lowest level in 1967, and that since that date there has been a strong upward trend in the Black unemployment rate
-
The Simkins series implies a structural break rather than a mere cyclical artefact. (p. 4)
Simkins' figures thus suggest a structural break in the late 1960s.
-
The close parallel between South African Black unemployment and UK unemployment in 1968-77 is used to validate Simkins' method. (p. 6)
What is particularly noteworthy is that from 1968 to 1977 the South African and British graphs move up and down together in more or less perfect unison.
-
Rising unemployment is attributed mainly to slower growth in labour demand, not labour supply. (p. 10)
the rising trend in unemployment rates has to do primarily with a slowdown in the rate of growth
-
Since 1975-80 the non-Black population has grown faster than non-agricultural employment, ending the earlier decline in non-Black unemployment. (p. 10)
the table shows that in the period 1975-80, indeed for the first five year interval since 1955-60, the non-Black population grew more rapidly than non-agricultural employment
-
Over the cycle the scale effect of lower interest rates dominates, so employment growth moves inversely with real interest rates. (p. 13)
Cyclically, therefore, it seems that the scale effect has predominated
-
The fall in manufacturing employment growth between 1960-65 and 1975-80 reflects slower overall growth rather than capital-labour substitution. (p. 13)
This however, primarily reflects a reduction in the rate of growth of the general scale of the economy rather than a substitution of capital for labour.
-
Predominant cause of rising unemployment is a fall in labour demand growth driven by slower general economic growth, not rising capital intensity. (p. 15)
the predominant cause of rising Black and Non-Black unemployment rates has been a decline in the rate of growth of the demand for labour
-
Global supply-side shocks depressed manufacturing profitability, including in South Africa, explaining the late-1960s/1970s growth slowdown. (p. 16)
These factors depressed profitability in manufacturing industry, and produced a severe problem of adjustment in the advanced industrial countries
-
Gold booms cushioned demand but appreciated the rand and hurt manufactured exports. (p. 17)
by raising the foreign exchange value of the Rand these gold booms have worsened the competitive position of South African goods, especially of manufactures
-
The role of low real interest rates and factor-price distortions in causing unemployment is judged overstated. (p. 18)
it appears that the role of relative factor-price distortions and in particular of low real interest rates, as a cause of unemployment, has been overplayed.
-
Industrial decentralisation cannot offset slowing employment growth in the metropolitan areas because subsidised-region industry is too small. (p. 20)
the volume of industrial employment in the subsidised regions is still small and rapid employment growth in those regions therefore cannot compensate for the current decline
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| Simkins' estimates of the unemployment rate, 1960-1981 (Simkins 1982, with earlier 1960-69 estimates adjusted per his 1982 suggestions) | secondary literature compilation | 1960-1981 | Supplies the Black unemployment rate series (Fig. 1) used as an index of secular and cyclical movement and compared with UK and non-Black series. |
| Current Population Survey (as used in Simkins 1982) and 1980 Population Census | household or labour survey microdata | to 1980 | Only mentioned in the title of Simkins' 1982 revision; the CPS gives lower unemployment levels than Simkins' estimates, a point noted but not analysed. |
| Registered unemployment of Asians, Coloureds and Whites, 1948-1983 | official statistical series | 1948-1983 | Builds the non-Black registered unemployment rate (Fig. 3) to show a long decline to 1974 and a rise thereafter, which signals structural change. |
| United Kingdom unemployment rate, 1965-1983 | international trade or finance data | 1965-1983 | A comparison series plotted against Simkins' Black rate (Fig. 2) to support the validity of his method and the global character of rising unemployment. |
| OECD countries' unemployment rates (general reference) | international trade or finance data | late 1960s-1980s | Cited in a footnote to show that the rise in unemployment since the late 1960s is general among advanced industrial countries. |
| South African Statistics and Quarterly Bulletin of Statistics (various issues): non-agricultural employment and population growth by race | official statistical series | 1946-1983 | Source for Table 1 growth rates of non-agricultural employment (mining, manufacturing, construction, electricity) and of population by race, which show that the demand for labour slowed. |
| South African Reserve Bank, A Statistical Presentation of South Africa's National Accounts for the Period 1946 to 1980, Supplement to Quarterly Bulletin, September 1981 | national accounts macro | 1946-1980 | Provides the real interest rate and manufacturing employment, output and capital stock series in Table 2, used to test the capital-labour substitution hypothesis. |
| Government bond rate deflated by the South African consumer price index (base 1975) | official statistical series | 1959-1982 | Used to construct the real interest rate series whose cyclical and long-term relationship with employment growth is examined. |
| Physical Planning Act, Section 3 (Black labour controls in metropolitan areas) | legislation gazette | 1970s | Cited in a footnote as a possible cause of rising capital intensity, with no supporting data. |
| Two recent government reports on trade and exchange rate policy (unnamed) | commission or official report | early 1980s | Cited to indicate that official policy is shifting toward an open, outward-oriented economy; the reports are neither named nor analysed. |
Inference¶
Quantification: Descriptive tables.
Descriptive time-series comparison. Simkins' unemployment series is compared with the UK series and with non-Black registered unemployment, and employment growth is compared with population growth by race over five-year intervals. Period averages of real interest rates and manufacturing K, L, Y and K/L are used to judge the scale versus proportions effects of factor prices. These are combined with an argument about global forces; no regression or formal test is used.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| Simkins' Black unemployment series shows a structural upward break in the late 1960s but remains cyclically sensitive, so it is not biased upward. (p. 2) | Simkins' 1982 estimates plotted in Fig. 1 and read off by eye, with yearly rises and falls enumerated. | Partial data |
| The close co-movement of the South African Black rate and the UK rate from 1968 to 1977 supports the validity of Simkins' method. (p. 4) | Visual comparison of Simkins' series and the UK unemployment rate in Fig. 2, with no statistical test. | Illustrative or anecdotal |
| Non-Black registered unemployment fell from 1961 to 1974 and has risen since, indicating structural change in the economy. (p. 7) | Registered unemployment of Asians, Coloureds and Whites divided by their economically active population, 1948-1983 (Fig. 3). | Partial data |
| Rising unemployment stems mainly from a slowdown in the growth of labour demand, not from shifts in labour supply. (p. 8) | Table 1 comparison of average annual growth rates of non-agricultural employment and of population by race, from South African Statistics and the Quarterly Bulletin. | Partial data |
| Low real interest rates did not drive the long-run fall in employment growth through capital-labour substitution; the fall mainly reflects slower output and capital accumulation. (p. 11) | Table 2 five-year averages of real interest rate and manufacturing employment, output, capital stock, K/L and K/Y from Reserve Bank national accounts. | Partial data |
| The decline in growth that lies behind rising unemployment derives largely from global forces: raw material prices, slower technology diffusion and competition from newly industrializing countries. (p. 14) | Qualitative argument and an OECD analogy; no data presented for South Africa. | Asserted |
| The unemployment insurance system, inflation and state enterprise are not fundamental causes of unemployment, and decentralization offers only limited relief. (p. 17) | Argument and brief analogies (ISCOR, British Steel); no data are presented. | Asserted |
The author's caveats¶
The authors note that no unequivocal true measure of unemployment exists and that absolute levels depend on arbitrary definitions. They acknowledge that Simkins' series covers all four race groups combined and so understates the Black rate, which they treat as irrelevant since only movements over time matter. They note that the Black and non-Black series are very differently measured.
Gaps in the evidence¶
The paper relies on one contested estimate series and offers no formal test of cyclical sensitivity or of the capital-substitution hypothesis; regressions of employment on output and factor prices, with sectoral breakdown, would be needed. The global-shock explanation is asserted without South African data on terms of trade, profitability or investment, although national accounts and Reserve Bank series then existed. Evidence on labour supply, such as migration controls and homeland populations, and independent microdata such as the CPS or 1980 census crosschecks, would test the claims better.
Replicability: Data poorly documented.