Skip to content

A Guide to South African Social and Economic Statistics from the Middle of the Nineteenth Century to the End of the Twentieth Century

Literature › Works

Charles Simkins (2003). A Guide to South African Social and Economic Statistics from the Middle of the Nineteenth Century to the End of the Twentieth Century.
Type report
Scope Comparative and methods
Method synthesis review
Period studied 1850–2000
Regions South Africa, Cape Colony, Natal, Transvaal, Orange Free State, Bantustans (TBVC and self-governing territories)
Themes methods data sources, demography urbanisation, labour markets wages, inequality poverty unemployment, bantustans removals influx, macro fiscal finance debt, agriculture land, mining migrant labour, manufacturing industrial policy, education human capital
Citation key simkins2003

Argument

The guide is an annotated bibliography of official and non-governmental statistical sources on South Africa from c.1850 to 2000. It argues that state form, state capacity and state interest shaped what numbers were collected, for whom, and how. Racial exclusions (white-only censuses, missing African income data), the fragmentation of statistics after homeland 'independence', and weak documentation and quality control mean the data require detective-like caution. The bulk of the text lists sources by theme (censuses, production censuses, national accounts, public finance, labour, household surveys, health, education, commissions), with brief warnings about reliability.

Evidence and method

Catalogue of official publications (censuses, production censuses, Reserve Bank national accounts, labour and household surveys), commission reports, bibliographies, and NGO/university sources (SAIRR, Carnegie inquiries, SALDRU, Development Bank). Commentary is drawn from the author's knowledge of the sources, not new data.

Findings

  • Racial and territorial structures of the state produced gaps and asymmetries in the statistical record, including missing African data and homeland-era gaps. (p. 2)

    These territories became responsible for their own statistical systems, which never developed beyond a rudimentary stage. Gaps therefore emerged in statistical information about South Africa between 1976 and 1994

  • Several censuses covered Whites only, and as late as 1970 African household income was not collected. (p. 2)

    the 1918, 1926, 1931 and 1941 censuses were for Europeans only, being constitutionally prescribed for electoral delimitation purposes.

  • The state had little interest in recording the demographic effects of removals, so information on them comes mainly from non-governmental work (Desmond, Surplus People Project). (p. 4)

    The government had little interest in revealing or even collecting information on the demographic consequences of its policies, many of which occurred in the rural areas

  • The 1991 census had to estimate many urban African settlements from aerial photographs and sample surveys amid unrest. (p. 2)

    The 1991 Population Census found it necessary to estimate populations in many urban African settlements by means of dwelling counts from aerial photographs and sample surveys rather than direct enumeration

  • National accounts begin only in 1946 and are subject to revision. (p. 46)

    The South African Reserve Bank was founded in 1921. South African national accounts date from 1946.

  • The Current Population Survey (1978-86), a key source on black unemployment, is unreliable because of non-random sample attrition. (p. 52)

    the sample suffered non-random attrition and the survey results became very unreliable over time. So this source has to be approached with more than the usual caution.

  • The 1995 Income and Expenditure Survey gives the most reliable estimate of the household income Gini, with caveats about survey quality. (p. 53)

    From it emerged the most reliable estimate of the Gini coefficient for the distribution of household income in South Africa at 0.59.

  • Sadie's widely used population estimates inflate the white population and ignore AIDS mortality. (p. 29)

    His assumptions about the pattern of white migration to and from South Africa lead to projections of net immigration which are too high, inflating the size of the white population

  • International migration statistics are unreliable: official emigration counts were about a third of destination countries' immigration figures. (p. 32)

    its emigration statistics (derived from Department of Home Affairs data) to the five main destinations (United Kingdom, United States, Australia, Canada, New Zealand) were about one-third of those countries immigration estimates

  • The Chamber of Mines supplied mining employment statistics to the government, so mine labour data depend on an industry body. (p. 38)

    The government has relied on the Chamber for collection of employment statistics in mining.