Skip to content

Agricultural Production in the African Reserves of South Africa, 1918-1969

Literature › Works

Charles Simkins (1981). Agricultural Production in the African Reserves of South Africa, 1918-1969. Journal of Southern African Studies, Vol. 7, No. 2. Taylor & Francis.
Type journal article
Scope Economics and economic history of apartheid
Method econometric
Period studied 1918–1969
Regions African reserves / homelands, Transkei, Ciskei, Natal/Zululand, Northern Transvaal, Western Transvaal/Orange Free State
Themes agriculture land, bantustans removals influx, demography urbanisation, inequality poverty unemployment, methods data sources, mining migrant labour
Citation key simkins1981

Argument

Simkins tests the consensus (Wilson, Wolpe) that reserve agricultural output collapsed from early in the twentieth century. Using Agricultural Censuses to build a constant-price production series, he finds total output roughly constant from 1918 to 1965, so the fall in output per head was driven by population growth. The steep per-capita decline came after 1948/1955, not before. Before 1955 emigration (labour migration) acted as a 'productivity maintenance mechanism'. Strengthened influx control and relocation then ended that safety valve and drove population density, per-head output and subsistence cover sharply down. Cross-sectional Cobb-Douglas regressions and cattle-ownership distributions support and qualify these conclusions.

Evidence and method

Union Agricultural Censuses 1918-69 (reserve figures partly estimated), Population Censuses, Union/SA Statistics price series, Social and Economic Planning Council Report 9 (1946), Tomlinson Commission, Keiskamahoek survey, Fox & Back 1937, BENBO 1976; the 1943 Witwatersrand Mine Natives' Wages Commission subsistence bundle; comparison with other estimates (Lehfeldt, NUC, Thirion, Retief, BENBO). Published sources only.

Place in the debates

Findings

  • Total constant-price reserve agricultural output did not fall between 1918 and 1965; it fluctuated around a stable index. (p. 8)

    Total agricultural production (valued at constant prices) did not fall over the period 1918-1965. On the contrary, one may detect a slight rise after the war

  • Per capita output fell only modestly before 1948 and dramatically afterwards. (p. 8)

    One must locate the really dramatic decline in production per capita in the period after 1948 rather than in the period before that date.

  • Reserve production met well under half of food needs even in 1918, but the share stayed stable until the mid-1950s. (p. 10)

    the proportion of requirements they were able to meet remained substantially constant between 1918 and 1955, declining rapidly only after that date.

  • Subsistence cover fell from 45-50% of food requirements to 26% by 1967. (p. 10)

    this average exceeded 45 percent again once in 1955 and then declined continuously to 26 percent in 1967.

  • Terms of trade improving for reserve agriculture masked the post-war per-head decline until the mid-1950s. (p. 11)

    This improvement offset the decline in product per head up until 1955.

  • Emigration was the main mechanism maintaining reserve productivity before 1955. (p. 16)

    This could only be achieved by a high rate of emigration from the reserves which can be regarded as the primary 'productivity maintenance mechanism.'

  • Influx control ended emigration as a counter-force; density almost doubled from 1955 to 1969. (p. 17)

    An immediate consequence of the modernisation and extension of influx control in the early fifties was the end of counteraction by emigration of tendencies towards decline.

  • Reserve populations grew very fast in the 1960s, largely attributed to state relocation. (p. 17)

    The relocation policy of the state must have been responsible for this in large measure.

  • The urban wage to reserve income ratio more than doubled, producing growing dependency on remittances. (p. 17)

    the private industry wage/agricultural product ratio rose from 5 in 1955 to 11 in 1965 ushering in the contemporary disequilibrium between urban and rural African incomes.

  • Cross-section regressions never show zero marginal labour productivity; Transkei and Zululand show no density effect, Ciskei and Natal do. (p. 16)

    In no case, did a marginal productivity of labour of zero emerge at any time.

Evidence ledger

Sources used

Source Kind Period How used
Union Agricultural Censuses 1918-1969 (reserve output and livestock returns) agricultural census 1918-1969 (gaps 1919-20, 1922, 1931-3, 1939-45) Main source of quantities of crops and livestock produced in the reserves, used to build the value-of-production series in Table 1.
Population Censuses 1936, 1946, 1951, 1960, 1970 (reserve population; Report 02-05-12 for 1970) population census 1936-1970 Supplies reserve population totals for per-head output, density, and the 1960-70 growth rate.
Sadie's evaluation of census undercount of blacks (SAJE 1970; IDC Projections 1973) secondary literature compilation 1936-1970 Used to gauge the undercount of the African population in censuses, though no adjustment was made.
Union Statistics for Fifty Years (1960); South African Statistics 1974 and 1978; Handbook of Agricultural Statistics 1904-50; Statistical Yearbook 1964 official statistical series 1904-1978 Source of producer, wholesale and retail price series, livestock numbers and national output used to value reserve production and the subsistence bundle.
Crops and Markets (wholesale prices of cowpeas, beans, peas) official statistical series 1950-1960 Wholesale prices used to value pulses.
Social and Economic Planning Council, Report No. 9, The Native Reserves and their Place in the Economy of the Union of South Africa, UG 32/46 commission or official report 1946 (data for 1916, 1936, 1939) Supplies district populations and land areas for the cross-section regressions and cattle distribution data.
Beaumont Commission Report (1916 population) commission or official report 1916 Source of the 1916 reserve population figure used in the per-head series.
Tomlinson Commission Report (Chapter 9 Appendix 1 production estimates; Chapter 19 cattle distribution) commission or official report 1937-1951 Alternative production valuations for comparison with Simkins's series, and cattle-ownership distribution data.
Native Economic Commission Report 1930-2 and Report of the Commission into Native Reserves 1946 (hide and skin prices) commission or official report 1930-1946 Cited via Wolpe's account as part of the received view; the 1946 report supplies the 1939 hide price.
BENBO, Black Development in South Africa (Swart Ontwikkeling in Suid-Afrika), 1976, tables B.7.2, B.9.4 commission or official report 1960-1970 Homeland GDP shares, agricultural production for 1968-9 for comparison, and an index for extrapolating post-1965 values.
Earlier estimates of reserve production: Lehfeldt 1922; NUC Dept of Economics 1949; Thirion 1954; Retief 1960 secondary literature compilation 1918-1959 Used to cross-check Simkins's valuation series in Table 2.
Stadler thesis on gross domestic product 1911-1959 secondary literature compilation 1911-1959 Cited on the underenumeration of agricultural production in the censuses.
Witwatersrand Mine Natives' Wages Commission 1943 subsistence bundle commission or official report 1943 Defines the fixed family-of-five subsistence bundle against which reserve production is compared.
Fox and Back 1937 preliminary survey of Ciskei and Transkei agriculture and nutrition author survey or fieldwork 1937 One of four sources of herd-size distribution data for the cattle ownership analysis.
Hobart Houghton and Walton, Keiskamahoek Rural Survey, Vol. II (1952) author survey or fieldwork 1948 Survey data on household cattle holdings used in the herd-size distribution.
Wage bill and employment series for African workers in private industry (Union Statistics for Fifty Years; South African Statistics 1974) official statistical series 1925-1968 Used to compute the ratio of average African industrial wage to reserve agricultural product per five persons.
Cooper, How Urban Workers in Botswana Manage their Cattle and Lands (1980) secondary literature compilation 1980 Provides herd-size thresholds used to define middle and rich peasantry in the reserves.

New data

Constructed annual (with interpolated gaps) series of the value of reserve agricultural and pastoral production 1918-1969 in current and 1946 prices, per-head output, subsistence-bundle costs, and district-level output per head and density estimates for 1927, 1937, 1950 and 1960. Also built estimated distributions of cattle ownership by household in 1936 and 1951.

Inference

Quantification: Statistical analysis.

Time-series reconstruction of reserve output from published census data valued at constant prices and compared with subsistence needs. Complemented by cross-section Cobb-Douglas regressions across magisterial districts with covariance analysis. The cattle distribution is estimated by fitting exponential distributions to regional data, and periodisation is based on the ratios computed.

Key claims

Claim Basis Strength
Total reserve agricultural production at constant prices did not fall over 1918-1965 and fluctuated around a stable level. (p. 262) Table 3 built from Agricultural Censuses quantities valued at 1946 prices from official price series. Partial data
The dramatic decline in production per capita came after 1948, not before. (p. 262) Per-head output computed from Table 3 and interpolated population from the Beaumont report and Population Censuses. Partial data
Reserve inhabitants could meet only a minority of subsistence needs from agriculture as early as 1918, with the proportion roughly constant to 1955 and falling rapidly after. (p. 264) Production compared to cost of the 1943 Wages Commission subsistence bundle priced from official retail and wholesale series (Figure 1). Partial data
Cross-sectional regressions show land degradation or population density effects in Ciskei and Natal but no zero marginal productivity of labour in any region. (p. 270) Cobb-Douglas regressions across magisterial districts for 1927, 1937, 1950 and 1960 using SEPC Report 9 population and land data with interpolations. Partial data
Emigration maintained productivity until 1954; after influx control and relocation, density rose and the wage/reserve income ratio jumped. (p. 271) Population density series, wage-to-reserve-product ratio from official wage and employment series, and census growth rates for homeland districts. Partial data
Cattle ownership was unequal; in 1951 about 6.6% of reserve households were middle peasants (20+ cattle) and 0.5% rich peasants (50+). (p. 276) Exponential distributions fitted to herd-size data from Fox and Back, SEPC, Keiskamahoek and Tomlinson, with assumed household size and Botswana-based thresholds. Illustrative or anecdotal

The author's caveats

Notes the study rests on published sources only, that the Agricultural Censuses are reticent about methods and probably underestimate reserve output, and that omitted items would add about 5%. Also flags the constant quality assumption, uncounted population undercount, interpolated years, and the unreliability of herd ownership data. Production estimates differ from other researchers' valuations because of methodological inconsistencies.

Gaps in the evidence

Unpublished Native Affairs and magisterial district records, and the original enumeration or estimation sheets, would test the reliability of the reserve output figures. Household or village-level survey data on production, trade and landlessness would replace the assumed per-capita subsistence bundle and the fixed quality and terms-of-trade assumptions. Such sources partly existed in archives then, and are accessible now.

Replicability: Data described but not public.