South Africa is going through a complicated, disputed change in the structure of its electricity generation process. It’s become fiercely political and, controversially, even Eskom executives have questioned the process. It might seem a novelty to South Africans, and there is a legitimate question about whether it’s worth it. But actually, the changes happening in the country now are pretty bog-standard around the world, where many jurisdictions long ago made this change.
Essentially, the change follows the recognition that while transmission and local distribution are natural monopolies — it makes no sense to build competing sets of pylons or power lines down your street — electricity generation isn’t. It can and probably should be competitive.
That, in essence, is what South Africa is doing. Unlike what the unions falsely claim, the current reform is correctly characterised as competition without privatising the grid. The presidency’s July 2026 plan is for an independent, state-owned transmission system operator outside Eskom, owning the transmission assets and providing nondiscriminatory grid access. The UK began this process in 1990 and New Zealand in the late 1980s and 1990s. Australia’s national electricity market started in 1998.
Because South Africa is treading such a well-worn path, it’s worth asking how it worked out elsewhere. Did electricity get cheaper? Did reliability improve? And, as South Africa’s trade unions fear, did employment decline?
In the age of AI these questions are embarrassingly easy to investigate, so forgive the derivative nature of the exercise. But the answers are illuminating.
First, why are we doing this?
It isn’t just about load-shedding. South Africa needs huge amounts of new generation and transmission investment over the coming decades. The state’s balance sheet, after years of supporting Eskom and paying for its enormous building programme, cannot sensibly finance all of it. The point of separating an independent, state-owned transmission operator from Eskom and creating a wholesale market is to allow lots of generators, with lots of different balance sheets, to build power stations and compete to sell electricity.
There is also a productivity problem.
In 2008 Eskom had 35,404 employees and sold 224,366GWh of electricity — about 6.3GWh per employee. In 2025 it had 42,030 employees, while energy sent out was only around 191,000GWh — about 4.5GWh per employee. The measures aren’t perfectly identical, but the direction is hardly encouraging. An independent calculation using Eskom tariff data and Stats SA CPI puts the cumulative Eskom wholesale electricity price increase from 2007 to 2026 at 1,172%, against cumulative consumer inflation of about 174%.
Employee costs tell a similar story. Eskom’s total employee-benefit expense worked out at about R321,000 per employee in 2008. In 2025 it was about R1.08m per employee, nominally. The narrower salary line works out at about R703,000 per employee. These are averages, not what the typical Eskom worker earns, but they provide some perspective. Household electricity now costs more in South Africa than it does in New Zealand and the US and about the same as in Australia — much richer countries.
So, what happened abroad?
Initially, it went rather well. In Britain domestic electricity prices fell about 15% in real terms in the six years after privatisation. Australia experienced falling real prices too, while productivity rose sharply.
But here comes the first warning: cheap generation did not necessarily mean cheap electricity bills. Over time household prices rose again. New Zealand’s residential electricity prices were 79% higher in real terms in 2018 than in 1990. Britain today has some of the world’s most expensive household electricity.
Why? Because generation is only part of the bill. Transmission, distribution, environmental costs, taxes, reserve capacity and bad regulation don’t disappear because generators compete. Reliability, broadly, survived the transition perfectly well. But the employment fears are much better founded.
Britain lost about 60,000 electricity sector jobs between 1990 and 1997 while productivity increased roughly 2½ times. Australian electricity employment almost halved from the late 1980s as output per employee more than doubled. Some jobs were outsourced; many disappeared. But of course, electricity forms the base of the industrial production pyramid; productivity in the electricity sector is more than compensated for in the industrial sector. And you can impute that in reverse in South Africa, because as electricity has become more expensive, the industrial sector has shrunk dramatically. Electricity expense is not the only reason, but it’s one of the main causes.
There is the bargain in a nutshell.
Competitive electricity markets don’t guarantee cheaper electricity forever. They don’t abolish regulation and they certainly don’t protect every incumbent job. What they do rather well is force productivity, attract different sources of capital and remove the absurdity of asking one enormously indebted state company to finance every new electron the economy will need.
South Africa’s electricity reform is therefore neither novel nor especially ideological. The evidence suggests it is probably necessary.
The more interesting argument is whether we do it well.
