subscribe Support our award-winning journalism. The Premium package (digital only) is R30 for the first month and thereafter you pay R129 p/m now ad-free for all subscribers.
Subscribe now

Access to electricity — measured by the share of individuals who use electricity as their main energy source — has significantly improved in lower-income countries in recent decades. Thanks to large-scale rural electrification and infrastructure development programmes, 90% of the population in these countries had access by 2021, up from about 74% at the turn of the century.

In many of these countries, however, electricity supply is often unreliable. Generating capacity is inadequate, there is not enough investment in infrastructure, and energy prices are high. Consequently, outages are frequent and long-lasting. By forcing households and firms to use alternatives, such as diesel generators and backup batteries, these outages raise the costs of energy, thus limiting the benefits of access. Those costs come in the form of direct expenses for households and firms and broader social impacts.

Power outages have long been identified as a major constraint to economic development. Many studies document their negative effects on economic growth,  productivity and sales. It is unsurprising then that people are willing to pay a relatively large amount for reliable power.

Because of those negative effects, it’s likely that outages have an effect on the labour market too. This is of particular interest in high-unemployment contexts, like SA, where creating decent jobs is key to alleviating poverty. However, evidence on these effects is scarce.



We analysed the labour market effects of load-shedding in a recently released paper.

We found that power outages have had negative effects on employment, as well as working hours and monthly earnings among those who remained employed. The effect on employment has been larger than the effect on working hours or earnings. This highlights the threat that load-shedding poses to job preservation and job creation efforts.

These effects were not, however, the same for all firms. Workers in the energy-intensive manufacturing industry appear particularly vulnerable to losing their jobs. Also, small and large firms responded differently. Small firms tended to favour reducing working hours rather than introducing layoffs.

Lastly, effects varied by load-shedding intensity. Low levels of load-shedding did not affect the labour market strongly, but high levels did.

SA boasts almost universal access to electricity, but has been subjected to power cuts since the end of 2007.

Load-shedding is primarily a consequence of frequent breakdowns at Eskom. This is due to a combination of poor long-term planning, a lack of financial resources, rampant state capture and corruption, and ageing coal-fired power stations, 80% of which have passed their midlife point.

These outages have become more frequent and severe in recent years. In 2023 — the worst year on record — load-shedding was in place for a total of 289 complete days.

Several studies have found that power outages reduce economic growth. But to our knowledge there has been no empirical evidence of their labour market effects until now.

We modelled these labour market effects using more than 15 years’ worth of nationally representative labour force survey data, covering nearly 3-million individuals. We merged this with macroeconomic data and high-frequency electricity data from 2008 to 2023.

We considered effects in three distinct ways. First, in a binary sense, which compared labour market outcomes during periods of load-shedding with outcomes when there was no load-shedding.

Second, in a continuous sense to account for differences in load-shedding intensity over time, as measured in megawatts of unmet demand. Third, because load-shedding in reality is implemented in stages, we analysed differences in effects between different load-shedding stages. For instance, stage 1 refers to up to 1,000MW of unmet demand and stage 6 up to 6,000MW.

We focused on effects on employment, working hours, hourly wages and monthly earnings. We also considered how effects varied across firms of different sizes and in different industries.

We adjusted all our models to ensure that the measured effects were not driven by other factors. These included labour market dynamics during the Covid-19 pandemic period, seasonality or changes in macroeconomic conditions relating to GDP, the interest rate, exchange rates and investment.



We found that load-shedding was significantly and negatively associated with employment, working hours and monthly earnings. On average, periods of load-shedding were associated with a 2.6% lower chance of being employed, 1.3% fewer working hours a week (equal to about half an hour) and 1.7% lower real monthly earnings. These are large effects. We did not, however, find evidence of a relationship with hourly wages. This suggests that the monthly earning reductions were driven by fewer working hours.

Low levels of load-shedding did not have these associations. But they were markedly worse with higher levels.

We did not find any evidence of negative associations for stages 1 and 2 load-shedding, but from stage 3 upwards, the effect became evident and grew stronger. For instance, stage 3 was associated with 1.9% lower employment, compared with 3.6% for stages 4 and 5 and almost 6% for stage 6.

Not all firms were affected equally. Manufacturing — a relatively energy-intensive industry — was worst off by far. We found that load-shedding was associated with nearly 17% lower manufacturing employment, about 6.5 times larger than the average of all industries.

In most industries, our model suggested that working hours were cut when power was cut.

Workers in large firms were vulnerable to all outcomes. In contrast, those in small firms were vulnerable to reductions only in working hours, but not to job losses or wage cuts. One might expect larger firms to be less vulnerable, as they would have more resources to pay for alternative energy sources. While that is probably true, large firms are more likely to operate in energy-intensive sectors.

Our analysis suggests that small firms have tended to reduce working hours rather than laying off staff, an outcome that is not unique to SA.


These results highlight the negative effects of load-shedding on the real economy. From a policymaking perspective, the primary goal, of course, must be to reduce the frequency and intensity of the outages, and ultimately eliminate them.

Recent events are encouraging. Efforts by the national utility have resulted in fewer breakdowns and improved generation capacity. In addition, deregulation and tax incentives appear to have increased demand from households and firms for alternative energy sources like solar.

Supply is expected to remain constrained in the medium term, however. Longer-term policy decisions revolve around moving faster towards renewable energy sources, encouraging private investment and, overall, building a more resilient and sustainable energy system.

• Bhorat is professor of economics and director of the development policy research unit at the University of Cape Town and Köhler is junior research fellow and PhD candidate in the Development Policy Research Unit in the School of Economics at the University of Cape Town.

• This article was first published on The Conversation.

subscribe Support our award-winning journalism. The Premium package (digital only) is R30 for the first month and thereafter you pay R129 p/m now ad-free for all subscribers.
Subscribe now

Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.

Speech Bubbles

Please read our Comment Policy before commenting.