Mineral Resources and Energy Minister Gwede Mantashe has reopened the debate on whether or not South Africa requires nuclear power. He has suggested that, slowly but surely, a new plant should be commissioned by 2040 to replace the Koeberg nuclear power plant, which reaches the end of its life by 2047.

The year 2040 is a long way away and monumental shifts in how electricity is delivered will take place between now and then.

One hears from promoters of nuclear power fairy-tale stories of it being cheap power. They claim that nuclear power will create numerous jobs and reignite a new nuclear-based industrial paradise. At best, these are mere public relations exercises. You hear the same type of claims for oil and gas. It all depends on how you stack up the numbers and what you are willing to believe.

Granted, nuclear power plants will create jobs during the construction phase, but the jobs curve rapidly plummets once the construction phase is over. One must make a hard-nosed assessment of all claims from whichever technology provider.

We need additional clean power because, when South Africa’s economy expands, the reserve margins of old coal-fired plants and a smuttering of renewables generation will not be sufficient to meet demand. This is partly because the clunky old coal machines have been consistently out of service and are not operating at full capacity.

Massive shifts are taking place not only in the generation of electricity but also in terms of how to supply it. Government should be watching that space more closely than it is doing currently. What we know for certain is that the level of control that government may have over energy supply and distribution, as well as the fundamental economics of the energy sector, will shift out of government control.

Transformative shifts in supply and demand in a country like South Africa will be significant and, over time, perhaps end the link between economic growth and energy supply.

But why is a large nuclear power plant such a troublesome prospect?

Firstly, modular technologies are more conducive to cost reductions over time if one chooses the auction model of procuring such power. We have enough experience in the procurement of renewable power to know that.

This is very different for bulk infrastructure. Paper claims of cheap power often do not withstand robustness once construction commences. We have good overseas experience and data for not only nuclear but all sorts of bulk infrastructure around the world that show consistency of cost overruns in megaprojects.

In South Africa, we need not look too far. Consider the enormous cost overruns for two of the largest supercritical coal-fired power plants in the southern hemisphere. The record is abysmal, if not scandalous. Petrochemicals group Sasol’s Lake Charles project is running into problems because of significant cost overruns. If we go for nuclear, we are bound to repeat the same mistake.

Secondly, nuclear power always involves geopolitical intrigue. It is subject to not only corporate capture but also sovereign capture. If you are stuck with a foreign power for technology and other dependences, given the bulky nature and complexity of nuclear power, you can bet that you have built yourself a bridge for fiscal and political influence over sovereign decision-making powers.

We may domesticate the nuclear generation capacity, but we will have outsourced our dependences. If you are interested in geopolitical intrigue, just follow the stories that are emerging about the Middle East Marshall Plan involving disgraced former US National Security adviser General Michael Flynn and the secret meetings between the Russians and Donald Trump’s administration about supplying 40 civilian nuclear plants to Saudi Arabia and the United Arab Emirates. This deal is unlikely to go through because it falls foul of the US’s Atomic Energy Act 138, which allows the US Congress to restrict who the US can supply nuclear technology to.

Thirdly, nuclear must always be the exception rather than the rule, especially if there are cheaper sources of power that meet three key criteria: low technological complexity, price competitiveness and quick installation. These criteria should be stretched as far as possible before deciding on a technology option that inevitably involves long-run dependences on a foreign supplier or government.

Fourthly, we may dismiss Fukushima and Chernobyl accidents as minor blips in the flotilla of nuclear plants that are currently operational. The temptation is to rate nuclear hazard as being very low. This may be true – nuclear disasters are indeed rare, but when they occur, the impact is significant. When disasters happen, the scandal is not why the disaster happened but how ill-prepared the authorities are to deal with the disaster.

A nuclear disaster can be overwhelming to a national fiscus, the economy and the psyche of the populace. One incident can lead to many plants being shut down. Even if they are subsequently brought back on line, the cost of safety measures will go up dramatically. You will never know the consequences until a disaster hits you, and when it does, the effects on the fiscus and the economy can be eviscerating, costing much more than what it took to build the plants in the first place.

These small but significant aspects should be understood, not ignored. This is why nuclear power as an option should be gauged against sovereign capacity to deal with significant risks, not risks that can be managed.

For this and many other reasons, one should continue to maintain a certain apostasy towards nuclear power, especially if you can avoid an adventure into the unknown.