The mystery of South Africa's ghost workers in 1996: measurement and mismeasurement in the manufacturing census, population census and October Household Surveys

This paper compares estimates of total manufacturing employment from South Africa's 1996 manufacturing census, the 1996 population census and the October Household Surveys from 1995, 1996 and 1997. Findings show that there are 300 000 too few manufacturing workers recorded in the 1996 population census. Furthermore there are other inconsistencies between these data sources. Several possible explanations for the deficit in manufacturing workers are put forward, but none are compelling. It is concluded that aggregate employment series constructed from household surveys should be treated with caution.