Men looking for piece jobs in Meredale, Johannesburg. Picture: KATHERINE MUICK
Men looking for piece jobs in Meredale, Johannesburg. Picture: KATHERINE MUICK

The SA national lockdown has brought a substantial part of economic activity to a halt for five weeks and the economic effects are already being felt, including job losses and large increases in the number of people going hungry. One way of estimating which workers are vulnerable to job loss is to determine which workers can and cannot continue to work during the lockdown.

There are two groups of individuals who can continue work: workers in essential services or producing essential goods, and those whose work allows them to work from home. To estimate who is more likely to be able to work and who is not we use Statistics SA’s Quarterly Labour Force Surveys (QLFSs), which ask workers detailed questions to then classify the industry they work in (“Retail trade in food, beverages and tobacco” or “Growing of crops”) and their occupation (“Spaza shop owner” or “Nursing and midwifery professionals”). We use the Post-Apartheid Labour Market Series (Palms) version of the Quarterly Labour Force Survey data from 2017, 2018 and the first two quarters of 2019.

Unfortunately, Statistics SA has had to stop in-person fieldwork, and has also stopped the quarterly employment statistics firm survey, so actual job losses will be difficult to estimate until these surveys resume. One obvious answer is for the department of employment and labour to report new UIF claimant numbers, but these will only cover those whose employers pay UIF on their behalf — that is, informal and low-earning workers will not be counted, and these workers are likely to be hardest hit, as discussed below.

To estimate how many workers in SA are classified as essential, we use the regulations stipulating which industries, workers and services are essential to classify workers’ industries as essential or not. We also used occupation data to identify some essential workers such as security guards and spaza shop owners. The occupation data from the QLFS can also be used to classify occupations into those that can be done from home and those that cannot.

The classification we use comes from a recent paper by Jonathan Dingel and Brent Neiman at the University of Chicago, who estimated that 37% of Americans could work from home. The authors classified occupations as feasible to work from home or not based on their occupation and the tasks required in each occupation, using data from the US Bureau of Labour Statistics’ Occupational Information Network (O*NET). The ability to work from home depends on the degree to which an occupation requires working outdoors, operating vehicles, or using mechanised equipment, among other examples.

Some jobs may feasibly be done at home in the US, but not in SA. Teachers, for example, are classified as able to work from home by Dingel and Neiman, but in SA most primary and secondary teachers cannot work from home for reasons of access to internet on the parts of both teachers and students. We therefore adjust the Dingel and Neiman classification based on our own judgment for SA’s context.

We estimate that 26.7% of the employed before lockdown began would have been working in essential industries or occupations, or about 4.5-million workers. This is an overestimate of the number of essential workers currently working under lockdown, since some essential industries are not running at full capacity. Our methodology is useful because it is a way of estimating which workers are more likely to suffer job loss, based on whether they are essential workers or can work from home.

We estimate that there are about 750,000 agricultural workers, 650,000 health workers and 600,000 security guards who are all classified as essential. There are another 400,000 essential workers in food and beverage manufacturing, 300,000 in food retail and petrol stations, 250,000 mineworkers (assuming 50% go back to work), 200,000 minibus taxi drivers and 100,000 spaza shop owners. Police officers, the SA National Defence Force, correctional services workers, and workers in banking and insurance make up other substantial groups, with another 500,000 are in other smaller industries.

We estimate that 13.8% of the employed in SA could feasibly work at home, or just over two-million people. Those whose work involves tasks that could be done at home are all in more highly skilled occupations. We estimate that 65% of senior managers and 56% of professionals could work from home. But no workers in low skilled occupations could work from home, since their jobs involve tasks that require them to be at their workplace (security guards, for example). Sixty-three percent of workers, or 10.5-million workers, are neither essential nor could work from home. The most severe job losses are likely to be concentrated among this group of workers.

In the bottom half of the earnings distribution, only 28% of workers are either essential or could work from home. By contrast, 61% of workers in the top 10% of the earnings distribution could work at home or are considered essential, meaning that low-earnings workers face much higher probabilities of job loss.

Job losses resulting from lockdown will be much more likely among workers who cannot work from home and are not in essential services. We estimate that this group of workers constitutes just under two thirds of those employed before the lockdown and is overrepresented in the bottom half of the earnings distribution, where only 28% of the employed could work from home or are considered essential.

Our methods could be used to estimate the number of extra workers who could go back to work if the lockdown is further eased, and which types of workers would be affected by these changes.

• Thornton is a researcher at the Development Policy Research Unit, and Kerr is an economist and chief research officer at DataFirst, University of Cape Town.

Would you like to comment on this article or view other readers' comments?
Register (it’s quick and free) or sign in now.

Speech Bubbles

Please read our Comment Policy before commenting.