There was outrage at energy & mineral resources minister Gwede Mantashe’s suggestion that the state should create another Eskom. But it is a rather better idea than many think, though in a completely different way to what I think the minister means.

The future energy system will have an independent state-owned system operator that procures power from multiple different electricity generators. Eskom will be one, but there will be many others, mostly privately owned renewable electricity generators. Mantashe’s proposal, which was mentioned by President Cyril Ramaphosa on Friday as part of various comments on system reform, is not terrible in that context.

In a competitive future market, private generators will be strongly incentivised to produce electricity as cheaply as possible. A competitive wholesale market, in which there are multiple generators and multiple customers is the best way to find the best equilibrium price of electricity. This is, on the whole, a good thing, certainly if we are t o improve the competitiveness of SA’s industrial sector which requires cheaper and more reliable power to take on global markets. But there are some potential downsides.

Competitive wholesale markets still have shocks. The over-cited California power crisis of the early 2000s is one example: the state suffered multiple outages in some areas due to several factors including market manipulation by private providers. It was less than two years from beginning to end and has never been repeated, but it taught lessons that have been built into energy supply markets around the world.

One of those is that the market can still fail. Extreme weather events, which we will be experiencing more in the future, mean that detailed energy consumption forecasting is becoming more difficult.

An energy system therefore needs to be highly responsive, but wholesale markets tend to have long-term contracts with fixed capacity. They can be complemented by spot markets in which urgent top-up supply can be acquired, but even these will come unstuck in highly stressed energy demand scenarios.

Ultimately there will always be some risk that a demand shock will arise that overwhelms supply. Of course, supply side shocks are also a risk, including weather shocks and failure of large baseload power sources.

Historically, when Eskom was at its best, it coped with this risk by having more than 5GW of energy reserves in its pumped storage facilities and its peaking power plants consisting of hydro and gas turbine stations. These famously have all been running to the furthest extent possible, at huge cost, rather than performing their design function of occasionally topping up the national grid during periods of excess demand.

You might think that the market should be able to solve the problem of unpredicted shocks. Let there be a spike in power prices, in turn reducing consumption, but driving producers to increase supply. But the demand side typically is not elastic — consumption does not move rapidly in response to short-term price spikes.

Some markets reduce this problem by integrating large consumers into the system, who can be called on to reduce consumption if needed. But markets around the world have found it difficult to regulate appropriately for that mechanism to work. They also find it politically impossible to impose short-term price spikes on retail consumers.

Market failures generally arise either when a scenario presents itself that neither regulators (who could manage it) nor producers (who could price for it) predicted. Predictable risks can be managed — but deep uncertainty, the unknown unknowns, can’t be. State ownership in the supply side can help to manage these hard-to-predict risks. The state balance sheet is the largest in the country and the only one that can absorb large but highly uncertain risks.

This reasoning suggests that you should want the state to still play a role in electricity supply. The problem is that role is not where the minister said it should be — baseload, always on supply — but at the other end of the availability spectrum — peaking supply.

The risk is that peaking plants are never used, a risk that the private sector would find impossible to bear. It would be inefficient for the state to take this risk away by providing availability payments to a private operator to simply be ready to supply if needed. It may be more efficient for the state to simply own and operate peaking supply itself. And that can be done in a state-owned entity different to Eskom.

Ramaphosa made the claim that competition between state-owned operators would help overall supply, referring to China where this happens. It is an odd point. You should only want the state to intervene where the market fails, not to have multiple interests in the market itself.

If competition works to deliver our public policy objectives, then there is no reason for the state to intervene. If competition does not deliver it — as in hard-to-predict supply and demand shocks — then the state should operate in the most effective and efficient way possible. A state-owned peaking power producer would need to meet that bar, at least to the extent that it ends up being cheaper than the alternative of availability payments to private producers.

The president said the idea of an Eskom 2.0 was being developed for discussion to the executive. Let us hope it starts by spelling out what the problem is that needs to be solved.

• Theobald is chair of research-led consulting house Intellidex. This article first appeared in Business Day.