South Africa has been one of the countries most adversely affected by the COVID-19 pandemic in Africa. Although the pandemic continues to pose important risks to public health, South Africa’s lockdown was always expected to lead to substantial short and long-term economic costs. Official labour force data shows that there were approximately 2.2 million fewer people employed in the second quarter of 2020 relative to the first3 – essentially erasing the last 10 years of job growth in the economy. Only a partial recovery can been observed in data from the third and fourth quarters of the year, with net employment still down 1.4 million relative to pre-pandemic levels. Despite the large amount of important work that has already been done to measure the various socio-economic impacts of South Africa’s lockdown, many of these studies are largely descriptive in nature. In this paper, we use representative labour force data – the Quarterly Labour Force Survey (QLFS) – to provide both a detailed descriptive and econometric account of the effects of the COVID-19 pandemic on employment. In terms of the econometric approach, we use a Propensity Score Matched (PSM) Difference-in-Differences (DiD) approach to measure employment impacts across observationally comparable affected and unaffected workers.

In our descriptive analysis, we show that employment loss was concentrated amongst the youth, those with lower levels of formal education, those living in urban areas, the private sector, non-union members, the secondary sector (particularly manufacturing and construction), and low- and semi-skilled workers. Notably, the lockdown disproportionately affected informal sectors workers who accounted for 1 in every 2 net jobs lost, despite representing just 25% of pre-pandemic employment. From our quasi-experimental findings, we show that the national lockdown decreased the probability of employment for those not permitted to work by 8 percentage points relative to the control group – a finding which holds across several robustness tests. We find larger effects for more stringent lockdown levels and distinct sub-groups – specifically own-account workers (most of which are in the informal sector) who experienced a nearly 3 times larger negative employment effect than the overall average treatment effect. This latter finding is indicative that working in the informal sector seems to be a key determinant of not being employed during the lockdown period.