Labour Supply in the South African Economy and its Implications for Agriculture¶
J.B. Knight (1976). Labour Supply in the South African Economy and its Implications for Agriculture. SALDRU Farm Labour Conference, September 1976, Paper No. 18. SALDRU, School of Economics, University of Cape Town.
| Type | conference paper |
| Scope | Economics and economic history of apartheid |
| Method | mixed |
| Period studied | 1946–1980 |
| Regions | South Africa, Bantu Homelands/reserves, White farming areas, Natal sugar belt, Ciskei and Transkei |
| Themes | labour markets wages, agriculture land, mining migrant labour, inequality poverty unemployment, bantustans removals influx, methods data sources, manufacturing industrial policy |
| Citation key | knight1976a |
Argument¶
Knight asks whether South Africa is running out of relatively unskilled labour, and so whether national labour-market forces will bid up farm wages. He assembles and criticises census, employment-statistics and agricultural-census data for 1946-1973 and finds that none supports firm conclusions. His tentative result is that 'residual labour' and underemployment, mostly in the Homelands, rose in absolute numbers after 1951. Mine recruitment problems, he argues, reflect a shortage of young fit men and the political risk of foreign labour, not a general shortage. Mining and manufacturing wage rises, especially since 1971, plus pass laws, a fragmented labour market and case studies of mechanisation in mining, sugar and timber, suggest competition for labour will be felt in farming. He judges that specific shortages will coexist with continuing general surplus.
Evidence and method¶
Population censuses 1946-1970, including unpublished tabulations. Department of Statistics employment statistics. Sadie's demographic revisions and the Economic Development Programme 1974-79 projections. Agricultural censuses, the 1969 South African Agricultural Union survey, and the Tomlinson Commission, Native Economic Commission and Marais Commission reports. Nattrass's migrant estimates, Wilson on gold-mine labour, the Maree and de Vos Ciskei/Transkei survey, an ex-miner attitude survey, and aerial-photograph population checks. Interviews with Department of Statistics officials and case studies of one mining group, a sugar estate and a timber company.
Place in the debates¶
- Labour surplus vs labour scarcity (is South Africa running out of unskilled labour?) (see Unemployment: causes, measurement and labour market institutions). Argues the evidence points to a large and rising surplus of unskilled labour. The mining industry's view of a drying-up pool is reconciled with this only by a shortage of physically fit young men.
- Why mine wages rose after 1971 / cause of foreign-labour dependence (see The mine migrant labour system: origins, recruitment, foreign labour and stabilisation). Says the rises were not initially a response to current shortage. The gold price made them possible but did not cause them. The main motive was reducing dependence on risky foreign labour by recruiting South African nationals.
- Backward-sloping supply curve and migrant labour behaviour (see The mine migrant labour system: origins, recruitment, foreign labour and stabilisation). Notes the influential belief in backward-sloping supply curves. Prefers Todaro-type probabilistic migration reasoning to explain falling recruitment of nationals alongside rising Homeland underemployment.
- Causes of farm mechanization and its employment effects (see Farm labour: mechanisation, labour shedding, wages and labour control). Holds that post-war mechanization (tractors) raised output but did not reduce absolute farm employment. Sugar and timber show large labour-shedding responses to wage rises, so wage-induced mechanization is likely if farm wages rise.
- Reliability of official data on unemployment and employment (see Data reliability and econometric method). Contends that census unemployment figures and the Economic Development Programme's methods are misleading, because of definitional changes and undercounting. Only wide ranges of estimates are defensible.
Findings¶
-
The proportion of South African nationals in gold-mine labour fell steadily after the war, as mines relied on foreign recruitment. (p. 57)
The proportion fell from over half before the war to 22.4% at its lowest, in 1972.
-
Real African earnings in mining were flat for decades, which Knight links to migrancy, recruiting organisation and monopsony. (p. 56)
Real earnings in mining remained roughly constant from 1951 to 1971, as indeed they had since 1911
-
Mines under the fixed gold price chose foreign recruitment over raising real wages. (p. 57)
Given the fixed gold price between 1949 and 1969, monopsonistic collusion among mines in not competing for labour, and a belief in the backward-sloping supply curve, real wage increases were resisted
-
The gold mines depended heavily on Malawi and Mozambique, and political events in 1974 made that risky. (p. 59)
Malawi and Mozambique together accounted for no less than 53% of African employment in the gold. mines.
-
Post-1971 mine wage rises were aimed at drawing South African nationals back into mining. (p. 59)
The increase in mine wages since 1971 must be seen primarily as an attempt to recruit more South African nationals.
-
Population absorbed by the rural Homelands grew far faster than agricultural output, so productive employment there stagnated. (p. 44)
Homelands far outstripped the growth of agricultural production
-
Unemployment is widely shared as rest periods between wage contracts and early retirement, so it appears as underemployment in the Homelands. (p. 47)
implications of these findings are that un- and underemployment may be widely shared
-
Even with rest-period adjustments a sizeable pool of Homeland male labour remains for mining, though it shrinks sharply with a long minimum rest. (p. 63)
if a minimum rest period is 3 months, the male labour force aged 20-39 available for wage employment becomes 185,000 instead of 351,0000
-
Real farm wages for regular African workers rose slowly through 1956/7-1972/3, though this is only tentatively established. (p. 66)
the increase in their mean real earnings averaged 2.8% per annum
-
White farmers report labour shortages in 1969 that far exceed reported surpluses of African workers. (p. 68)
The shortfall of 105,000 African men far exceeded the reported surpluses of 7,000 African male, 11,500 female and 25,000 child workers
-
Post-war farm mechanization (tractors) increased output without cutting aggregate farm employment. (p. 79)
However, mechanisation caused an increase in production but not an absolute decrease in employment.
-
Case studies show strong employment responses to recent wage rises in sugar and timber. (p. 80)
evidence of a large employment response in sugar and timber, where wages have risen recently
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| Population Census 1970, Report 02-02-02 and Report 02-02-02 Tables 1, 3, 4-6; Report 02-05-04; Single Ages 1941 to 1970 Report 02-05-05 | population census | 1946-1970 (esp. 1970) | Basis for African population, age-specific activity, economically active, employment by sector, unemployment and foreign African counts, adjusted for definitional change and under-enumeration. |
| J.A. Vermaak, Die Vraag na en Aanbod van Mannekrag in die Republiek van Suid-Afrika, Deel 1, HSRC 1974 (from unpublished 1970 census data) | secondary literature compilation | 1970 | Source of 1970 African age-specific activity rates applied to other census years to standardise the labour force. |
| Aerial photograph / hut-count checks of census undercount: Smit (1965 thesis), Schulze (1969), Hattingh (1973) | secondary literature compilation | 1951-1970 | Documents serious census under-enumeration in rural Homelands, mostly before 1970. |
| J.L. Sadie demographic estimates and Projections of the South African Population 1970-2020 (IDC) | secondary literature compilation | 1946-1980 | Provides corrected African population estimates and projections used to adjust census under-reporting (about 10% pre-1970, 3.9% in 1970). |
| Economic Development Programme for the Republic of South Africa 1974-1979 (Prime Minister's office) | commission or official report | 1970-1980 projections | Official projections of population, labour force, non-farm and farm employment and the unemployment method, which the author uses and criticises. |
| Employment statistics collected from employers (South African Statistics 1970/1974; Bulletin of Statistics) and Statistics of Houses and Domestic Servants 1970, Report 11-03-06 | official statistical series | 1946-1973 | Compared with census employment by sector to gauge undercount and trend in African non-farm employment; domestic servant estimate checked against the census. |
| Agricultural Census / Report on Agricultural and Pastoral Production (1953/4-1972/3; 1958-59; 1968-69; 1970-71) | agricultural census | 1953/4-1972/3 | Series of regular and casual African farm employees and earnings, used to trace employment in White agriculture and shown to be erratic and unreliable. |
| Department of Agriculture, 1975 Abstract of Agricultural Statistics (Tables 80, 89) | official statistical series | 1958-1975 | Agricultural production volume and producer prices used to gauge Homeland agricultural growth and farm employment fluctuation (Figure 1). |
| Unpublished Department of Statistics tabulations of Homeland African population (1970 boundaries) and Census 1951/1960/1970 age by magisterial district; P.J. van der Merwe mimeo | population census | 1946-1970 | Tables 12-13 on Homeland population, density, masculinity and adult male numbers. |
| Jill Nattrass, 'The migrant labour system and South African economic development', SAJE March 1976 | secondary literature compilation | 1946-1970 | Estimates of temporary male migrants from rural areas (635,000 in 1946 to 1,357,000 in 1970), used for rest-period calculations. |
| Summary of the Report of the Commission for the Socio-economic Development of the Bantu Areas (Tomlinson Commission), 1955 | commission or official report | 1930-1955 | Estimates of reserve income, livestock, carrying capacity and number of farming families; basis for productive labour in Homeland agriculture. |
| Official Yearbook of the Republic of South Africa | official statistical series | 1960-1970 | Homeland GDP from agriculture, forestry and fishing rose 33% 1960-70, compared with price rise to infer real growth under 1% per year. |
| Report of the Native Economic Commission 1930-32 (UG 22-1932) | commission or official report | 1930-32 | Qualitative evidence of growing impoverishment, land shortage and landlessness in the reserves. |
| Social and Economic Planning Council report (1946) | commission or official report | 1946 | Cited as concluding that many reserves are overpopulated, corroborating the statistical picture. |
| Maree and de Vos survey of four districts of Ciskei and Transkei, 1965 and 1968 | author survey or fieldwork | 1965, 1968 | Only sample survey of Homeland unemployment and underemployment; yields rates over 20%. |
| Department of Statistics, Bulletin of Statistics, South African Statistics and Reports on Agricultural and Pastoral Production (earnings series) | official statistical series | 1940s-1975 | Average real African earnings in mining, manufacturing and farming, deflated by the urban White CPI (Figure 2). |
| Francis Wilson, Labour in the South African Gold Mines 1911-1969 (1972), Tables 5 and 8 | secondary literature compilation | 1911-1969 | Mining real wage series and gold-mine African employment numbers by nationality 1936-56. |
| Mine Labour Organisations (N.R.C.) Ltd., Reports and Financial Statements (1966-) | chamber of mines or recruiting records | 1966-1972 | Table 16 on South African versus foreign African labour in gold mining. |
| Chamber of Mines Human Sciences Laboratory attitude survey of ex-miners (M. Masipa, mimeo) | author survey or fieldwork | c.1973-74 | Survey of about 240 Africans not employed in mines shows pay was the main dissatisfaction and improved pay the main inducement to return. |
| Redelinghuys, HSRC 1974 study of Tswana work-seekers at labour bureaux, 1971/2 | author survey or fieldwork | 1971/2 | Evidence on factors determining length of unemployment between contracts. |
| Natal Agricultural Union, 'Service Conditions of Bantu on Farms in Natal' (mimeo, 1973) and survey of Natal farms 1971/2 | employer association records | 1971-1973 | Poverty datum line comparison for farm families and cash wage recommendations. |
| Second Report of the Commission of Inquiry into Agriculture (Marais Commission), RP 84/1970 | commission or official report | 1966/7-1970 | Benefit-in-kind ratio and conclusions on mechanisation and labour use in agriculture. |
| South African Agricultural Union, 'Bantu Labour Policy' (mimeo, October 1974) | employer association records | 1974 | Source on farmers' views of wage policy, agreement labour and competition with mining. |
| Froneman committee (1963) report on employment of foreign Africans, as summarised by Ken Owen (SAIRR 1963) | commission or official report | 1960-1963 | Estimates of foreign African workers in 1960 and recommendation to replace them with indigenous labour. |
| Interviews and company information from a mining group, a sugar estate and a timber company | interviews oral history | 1965-1975 | Case studies of employer responses to wage increases through mechanisation and reduced African employment. |
| Plewman and Bleazard papers in Journal of the South African Institute of Mining and Metallurgy (1974) | secondary literature compilation | 1974 | Age distribution of mineworkers and occupational structure of mechanised vs unmechanised mines. |
| Donald Pursell, 'Bantu real wages and employment opportunities' SAJE 1968; P.J. van der Merwe, SAJE 1969 | secondary literature compilation | 1958-1969 | Explanations for manufacturing wage rises and evidence on labour bureau system and urban evasion. |
New data¶
Author-constructed series: standardised and under-count-adjusted African labour force and non-farm/farm employment, 'residual labour' and underemployment estimates 1946-1970 and projections to 1980, full-time-equivalent casual farm labour, and Homeland labour-pool estimates, all derived from published and unpublished official data.
Inference¶
Quantification: Descriptive tables.
Reconstructs census and employment series under multiple explicit assumptions to bracket residual labour and underemployment, supplemented by indirect Homeland indicators, regional regression of farm wages on shortage, wage series comparisons, and employer case studies.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| Census definitions of economic activity changed, so inter-census comparisons of the economically active are worthless, and half of the 1960-70 rise in economically active women is definitional. (p. 8) | Comparison of census coding instructions and Department of Statistics interviews across 1951, 1960 and 1970. | Systematic data |
| African non-farm employment grew faster than the African labour force in the 1960s, and is projected to continue doing so. (p. 20) | Census, employment statistics and the Economic Development Programme, adjusted for undercount. | Partial data |
| Statistics on farm employment are unreliable, with agricultural census, SAAU survey and population census giving inconsistent figures, and recorded farm employment fell over time. (p. 26) | Comparison of agricultural census series, the SAAU 1969 survey and the 1970 census. | Partial data |
| Residual labour and African underemployment rose in absolute numbers since 1951, but trends in percentage terms are ambiguous. (p. 38) | Table 11 and Table 14 combining adjusted labour force, employment series and Tomlinson-based Homeland farm employment. | Partial data |
| Homeland agricultural production grew far more slowly than Homeland population, so absorption of people outstripped production and productive employment in Homeland agriculture was roughly constant. (p. 44) | Official Yearbook GDP, Abstract of Agricultural Statistics prices, Tomlinson Commission, Homeland population from the unpublished census compilation. | Partial data |
| The decline in South African nationals on the gold mines reflects wage policy, monopsony and foreign recruitment rather than exhaustion of the labour pool; only certain young fit men may be scarce. (p. 57) | Wilson and NRC employment series, wage series, Chamber attitude survey, Nattrass migrant estimates and rest-period calculations. | Partial data |
| Farm labour shortages persisted despite pass laws; regional farm wages show no clear link to shortage, and mining competition affected sugar areas. (p. 68) | SAAU 1969 survey regression against agricultural census regional earnings for 34 regions; sugar industry wage data. | Partial data |
| Wage increases in the 1970s induced mechanisation and reduced African employment in sugar, timber and, more slowly, mining. (p. 78) | Interviews and company information from a mining group, a sugar estate, a timber company; Marais report. | Illustrative or anecdotal |
The author's caveats¶
The author states that the data do not permit clear-cut answers, that census undercount and changing definitions are serious, that the residual labour series is built from doctored series prone to error, that the CPI is inappropriate for Africans, and that the Maree-de Vos survey cannot be extrapolated.
Gaps in the evidence¶
The argument lacks reliable age- and sector-specific employment and enumeration data for Homelands and illegal urban residents, and national sample surveys of Homeland underemployment and farm wages. Farm employment and earnings series were erratic and pass-law effects unquantified. The company case studies are anonymous and few; later labour force surveys and mine and company records could test them.
Replicability: Data described but not public.