ABSTRACT
The main labour market survey in South Africa is the Quarterly Labour Force Survey (QLFS) but the General Household Survey (GHS) also measures employment. We show that despite often identical sample designs and questions about employment, the GHS produces higher estimates of total employment from around 2012 than the QLFS, which finds more economic inactivity. The discrepancy in employment totals worsens substantially after Covid and the GHS also starts to report many more unemployed while the QLFS reports many more economically inactive. We investigate some possible explanations for this result. Raw sample composition does not contribute substantially to these differences and weights make a small difference.