Labour-Related Structural Trends in South African Maize Production¶
J. Van Zyl; N. Vink; T.I. Fényes (1987). Labour-Related Structural Trends in South African Maize Production. Agricultural Economics, 1. Elsevier Science Publishers B.V.
| Type | journal article |
| Scope | Economics and economic history of apartheid |
| Method | descriptive quantitative |
| Period studied | 1945–1985 |
| Regions | North-Western Free State, Western Transvaal, Transvaal Highveld, South Africa (commercial agriculture) |
| Themes | agriculture land, labour markets wages, demography urbanisation, inequality poverty unemployment |
| Citation key | zyl1987 |
Argument¶
The authors trace farm labour trends in the summer-rainfall maize-producing areas from 1945 to 1985. Farm employment per unit of cultivated land rose to 1970 and fell thereafter, as capital moved from being a complement to labour (tractors raised yields and so harvest labour demand) to a substitute for it once combine harvesters spread. They attribute the post-1970 substitution partly to policies that cheapened capital: 100% tax write-offs, subsidised interest rates and general agricultural subsidies. Farm labour remuneration rose in real terms, yet the labour cost share fell, which they read as over-mechanization against the theoretical optimum. They conclude that capital is scarce and unskilled labour abundant in South Africa, so factor-price distortions should be reversed to protect rural jobs.
Evidence and method¶
Maize production cost surveys by the Department of Agriculture and Water Supply (rotating samples of c.80 farmers with >100 ha of maize) for three regions, 1945-1985; Abstract of Agricultural Statistics; De Klerk's (1983) survey of harvesting methods in the Western Transvaal; Joubert and Van Wyk (1984); Fényes (1983); an elasticity of substitution estimate from Van Zyl (1986b).
Place in the debates¶
- causes of farm mechanization and labour shedding (policy-induced factor-price distortion vs. other drivers) (see Farm labour: mechanisation, labour shedding, wages and labour control). Stresses capital subsidies (tax write-offs, subsidised interest) and mechanization of harvesting, with urbanization as a contributing factor. Argues some farms are over-mechanized relative to South Africa's factor endowments.
- Habakkuk-style capital-labour substitution in commercial agriculture (see Farm labour: mechanisation, labour shedding, wages and labour control). Shows an early complementarity phase (1945-70) followed by substitution after 1970, using an Ishikawa-curve framework and elasticities of substitution.
- farm wages and 'neo-feudal' payment in kind (see Farm labour: mechanisation, labour shedding, wages and labour control). Finds real cash and in-kind pay rising, with cash growing faster. Says farm pay growth compares favourably with national and black-worker income growth, and cautions about comparing farm and other wages.
Findings¶
-
Capital formation in commercial agriculture grew far faster than farm employment over 1950-1980. (p. 1)
capital formation has increased by 4.0% per annum between 1950 and 1980, compared with an increase of0.71% per annum in the number of farm employees in the same period
-
Farm employment grew to 1970 and then fell, while real capital formation kept growing. (p. 3)
The total number of farm employees increased from 1950 to 1970, but decreased from 1970 to 1980.
-
Harvest mechanization in the Western Transvaal spread very rapidly between 1968 and 1981. (p. 6)
These adoption rates show clearly that the majority of farmers in the Western Transvaal have already changed to mechanized harvesting, and this process must have contributed to a decline in the quantity of labour demanded, especially seasonal labour.
-
Capital and labour were complements in 1945-70 but became substitutes after 1970. (p. 6)
illustrate a substitution of capital for labour in crop production since 1970 in all three regions.
-
Real farm pay rose faster than average real income for all workers and for black workers. (p. 9)
These figures compare favourably with the increase of only 1.92% in real total income per worker for the whole of the R.S.A. and the 2.54% for black workers over the same period.
-
Tax and interest policy encouraged over-mechanization. (p. 12)
Tax provisions which enabled farmers to write off 100% of capital costs against taxes in the year of purchase, as well as subsidized interest rates have contributed to this over-reaction
-
The estimated elasticity of substitution flips sign between periods, but remains inelastic. (p. 13)
Labour and machinery were thus highly significantly inelastic complements during the period 1945-1970, but became highly significantly inelastic substitutes during 1970-1985.
-
The authors attribute the post-1970 fall in farm labour per cultivated area to mechanization, aided by policy and urbanization. (p. 13)
The accelerated rate of substitution of capital for labour in agriculture was probably partially due to tax concessions on new capital improvements and machinery, and also subsidization of agricultural production in general.
-
The labour cost share in maize farming is small and declining. (p. 12)
labour costs expressed as a percentage oftotal costs per farm unit are respectively 13.8%, 10.2% and 10.3% for the North-Western Free State, the Transvaal Highveld and the Western Transvaal.
-
Policy measures lowered the cost of capital and should be reviewed to protect rural jobs. (p. 16)
The major measures that kept the cost ofcapital relatively low were the control of interest rates in general, the subsidization of interest rates in agriculture, and tax concessions on capital investment.
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| Maize production cost surveys, Department of Agriculture and Water Supply (rotation basis, 1945-1985), three maize regions | household or labour survey microdata | 1945-1985 | Core data for growth rates in employees per farm and per 1000 ha, machinery investment, remuneration, costs, income and productivity (Tables 2, 4-8). |
| Unpublished reports of the Department of Agriculture (annual levels of parameters) | commission or official report | 1945-1985 | Cited as the location of exact yearly levels of the parameters; the paper reports only trends. |
| Abstract of Agricultural Statistics (Anon., 1986) | official statistical series | 1950-1985 | Source for national farm employment and real gross capital formation growth rates (Table 1), farm numbers and sizes, employment, and the area under maize. |
| Joubert and Van Wyk (1984), Labour in South African Agriculture (AGROCON 1984) | secondary literature compilation | 1968-1980s | Supplies seasonal worker numbers per farm and per cultivated hectare and the decline in seasonal labour, used to support the mechanization interpretation. |
| De Klerk (1983), MA dissertation on maize harvesting in the Western Transvaal, 1968-1981 | secondary literature compilation | 1968-1981 | Sample-based adoption rates of mechanized versus hand harvesting (Table 3), taken as evidence that mechanization reduced seasonal labour demand. |
| Fényes (1983), The Use and Compensation of Farm Labour in South African Agriculture | secondary literature compilation | 1945-1980s | Source for level data and for corroborating the shift from in-kind to cash remuneration. |
| Fényes and Van Rooyen (1985), South African agriculture and migrant labour | secondary literature compilation | 1970, 1982 | Source for the composition of the commercial farm workforce in 1982 and the agricultural share of the economically active population in 1970. |
| South African Reserve Bank Quarterly Bulletin, March 1986 | national accounts macro | 1986 | Gives commercial agriculture's share of total employment and of GNP. |
| South African Statistics 1980, Department of Statistics (R.S.A., 1980) | official statistical series | 1980 | Source for urban population shares and the inference that commercial agriculture employs over half of the rural economically active population. |
| Van Zyl (1986b), Duality and elasticities of substitution II: An empirical application | secondary literature compilation | 1945-1970, 1970-1985 | Source of the Shadow-measure elasticities of substitution between labour and capital showing a shift from complements to substitutes. |
| Antrobus (1984) dissertation on farm wages in the Albany District, 1957-1977 | secondary literature compilation | 1957-1977 | Corroborates the trend towards cash remuneration. |
| Wilson, Kooy and Hendrie (1977), Farm Labour in South Africa | secondary literature compilation | 1970s | Cited for a caution about comparing farm and non-farm remuneration and for conditions of farm workers. |
Inference¶
Quantification: Statistical analysis.
Time-series trend analysis: exponential growth curves fitted with time as the independent variable, estimated separately for 1945-1970, 1970-1985 and 1945-1985 in three maize regions. Capital-labour complementarity or substitution is inferred from the signs of the trends, a CES-type elasticity from another study, and a theoretical equilibrium condition. Causal explanations (mechanization, tax concessions, subsidies, urbanization) are asserted from trends and secondary studies, not tested.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| Farm employees per 1000 ha under cultivation rose until 1970 and fell thereafter, probably because of mechanization, especially in harvesting. (p. 243) | Table 2 growth rates from the Department of Agriculture cost surveys, plus Joubert and Van Wyk's seasonal labour decline and De Klerk's harvesting adoption data. | Partial data |
| Western Transvaal farmers shifted rapidly from hand harvesting to mechanized harvesting between 1968 and 1981. (p. 246) | De Klerk's (1983) sample of Western Transvaal farmers (Table 3). | Partial data |
| Capital and labour were complements in 1945-1970 but became substitutes after 1970 (capital substituting for labour). (p. 246) | Signs of the Table 2 and Table 4 growth rates; elasticities of substitution from Van Zyl (1986b). | Partial data |
| Real cash and in-kind remuneration per farm worker rose in 1970-1985, cash faster, and total real remuneration grew faster than real income per worker nationally. (p. 248) | Table 5 regression growth rates from the cost surveys, compared with national income per worker figures. | Systematic data |
| Labour's share of total farm costs declined, contrary to the theoretical optimum, which suggests over-mechanization encouraged by tax write-offs and subsidized interest. (p. 252) | Table 6 and 7 growth rates in costs, with the equilibrium condition. Over-mechanization is supported by citing Van Rooyen, Brotherton and Groenewald, and Biggs, not by testing. | Partial data |
| Capital formation in commercial agriculture grew much faster than farm employment over 1950-1980 (4.0% against 0.71% per annum), implying policy-distorted factor prices. (p. 243) | Table 1 from the Abstract of Agricultural Statistics (Anon., 1986). | Systematic data |
| Labour productivity grew faster than machinery productivity but slower than real remuneration per worker, and the Ishikawa two-phase curve holds for South African maize. (p. 253) | Table 8 productivity ratios from the cost surveys; the Ishikawa curve is borrowed from Asian rice experience and applied qualitatively. | Partial data |
The author's caveats¶
The authors say they present only trends and annual changes, not levels, and refer readers to Fényes (1983) and unpublished Department of Agriculture reports for yearly values. They note that comparing farm and non-farm remuneration, especially payment in kind, needs caution because of differences in how remuneration is calculated. Most growth rates are only significant at the 10% level.
Gaps in the evidence¶
The causal attribution to tax concessions, subsidies and urbanization rests on assertion and secondary citations, with no farm-level or regression evidence linking policy to machinery adoption. The data cover only larger maize farms (over 100 ha) in three regions, so nothing is said about smaller farms, the homelands or other sectors. Farm-level panel data from the cost surveys, Agricultural Census records, and tax and interest-subsidy records would allow better tests. Worker-side evidence on displacement, such as labour migration and eviction, is absent, and some of it exists in the Agricultural and Population Censuses.
Replicability: Data poorly documented.