For much of the apartheid period the government omitted to compile data on the number of Africans without work, and when a regular survey was finally initiated an unduly narrow definition of unemployment was adopted. As it became apparent during the 1970s and 1980s that the number out of work was rising, the issue became the subject of heated debate. There were disputes over both theoretical aspects of the nature and possibility of voluntary unemployment in a market economy, and over the validity of various attempts at empirical measurement. Such measures were typically either indirect estimates that were necessarily highly sensitive to their underlying assumptions and definitions or depended on one-off surveys, often with quite small samples. Both provided useful information, but neither is of much help to historians trying to trace the emergence and changing scale of unemployment over time. For the purpose of the present study an attempt was made to do this by using the one other available source, information on employment and unemployment collected at periodic censuses of the population.

The quality of the data available is very poor, and the results are, at best, rough orders of magnitude. The original census enumerations suffered from changing levels of undercounting of the black population, from inconsistencies in the definition of employment, and from exclusion of the notionally independent states of the Transkei, Bophuthatswana, Venda, and Ciskei (the TBVC states) from the censuses of 1980, 1985, and 1991.