Load shedding is having a devastating impact on the competitiveness and productivity of the South African economy. It is puzzling that after almost 17 years of load shedding, there are hardly any academic studies into the economic costs associated with load shedding. Even before load shedding, South Africa’s long term productivity performance had been poor compared to other economies. Since productivity is the main driver of per capita income growth over the long-term, the intensification of load shedding augurs ill for the long-term outlook of the South African economy.
The most recent estimates we are aware of are from Nova Economics who estimate that load shedding cost the South African economy nearly R45 billion (in 2022 rands) between 2007 to 2019 (subtracting a cumulative 5% points from GDP growth), and almost R225 billion between 2020Q1 to 2023Q1 (subtracting a cumulative 15% points from GDP growth). Between 2022Q2 and 2023Q2 their estimates imply loadshedding subtracted 1.15 percentage points from annual average GDP growth. The South African Reserve Bank (SARB)’s latest estimate is that load-shedding has reduced economic growth in 2023 by around 1.8 percentage points. By comparison, PWC’s estimate of the impact on GDP is much larger: they estimate an impact of up to 5 percentage points for 2022, implying a potential growth rate for the economy for 2022 of close to 7%.
That is a lot higher than alternative estimates of South Africa’s potential growth. Structural and statistical models tend to produce much lower estimates. One gets a potential estimate below 2% from a production function approach, whereas statistical filters produce estimates below 1% per annum. SARB’s latest estimates have potential at 0.5% in 2022 and 0.1% in 2023, while we estimate that the IMF’s estimates for economic slack for South Africa imply a potential growth rate of below 1% for 2022-2024.
It is worth reiterating that it is difficult to compare estimates of load-shedding impacts because of differences in assumed electricity production losses, assumed economic costs of electricity shortages and the level of GDP at the time that different studies produced their estimates. The table below summarises recent estimates from a variety of sources and the basis of their calculation.
Yet, as far as we are aware, there is no journal published research that has considered the implications of South African load-shedding on total factor productivity and potential growth using frameworks typically used to estimate these concepts in the literature.
| Source | Estimated impact on GDP growth (percentage points) | GDP lost (R million per day) | Description |
| Nova (2023 | -1.15 between 2022Q2 and 2023Q2 | ||
| PWC (2023) | Up to -5 in 2022 | ||
| SARB (2023) | -1.8 in 2023 | R204-R899 | Stage 3 to 6 for 6 to 12 hours per day |
| Intellidex/Krutham (2023) | -1.8 in 2023 | 1bn/day for full day stage 6 | |
| 600mn/day for stage 4 | |||
| 100mn/day for stage 2 | |||
| Alexander Forbes (2022) | 4000 at stage 6 | ||
| PWC (2022) | Up to -3.1 in 2022 | Reduction to real GDP growth, costing the economy up to 400,000 potential jobs. | |
| Western Cape Government (2021) | 500 per stage | ||
| Eskom (Nova, 2021) | -0.4 | 236 in stage 1 to 942 in stage 4 | Growth sensitivity to a 1% Load-shedding increase (as % of electricity sales). |
| Eskom (Nova, 2020) | R35 billion total cost to the economy between 2007 and 2019 in constant 2020 prices. | ||
| CSIR (2020) | -1.1 to -2.2 in 2019 | 700 per stage | Total cost of between 167 to 388 billion between 2007 to 2019, which is equivalent to around 0.3 to 0.7 of cumulative GDP since 2007. Costs in 2019 estimated at between 60 and 120 billion. |
| National Treasury (2019, Budget Review) | -0.2 in 2019 | Estimated decline in GDP growth from expected decline in energy availability in 2019. | |
| SARB (2019, April MPR) | -1.1 in 2019 | Growth impact relative to baseline 2019 forecasts. Employment loss estimated by 125 000 jobs. | |
| Morema et al (2019) | -0.3 in 2019 | 348 in stage 1 to 753 in stage 4 | Assuming load-shedding continued throughout 2019. |
| Goldman Sachs (2019) | -0.9 in 2019 | Estimated decline in GDP growth if load-shedding persisted throughout 2019. | |
| RMB Morgan Stanley (2019) | -0.5 loss to annualised GDP growth at about stage 4 | Growth sensitivity to a 10% power supply cut (annualised). | |
| National Treasury (2015, Budget Review) | -1 in 2015 | Estimated real GDP growth impact of electricity shortages in 2015. |
Footnote
The production function model referred to below is based on a constant-elasticity of substitution approach and the statistical estimates would be the levels of potential implied by a range of statistical filters.
There are some studies that look at the cost of planned outtages in South Africa (see here for a literature review), but we are not aware of studies that look at the costs of loadshedding at the scale South Africa has experienced over the last couple of years.
