CSIRO’s nuclear costings have recklessly ignored construction risk, and the taxpayer will pay

On Monday, the CSIRO released updated estimates for the cost of nuclear as well as a range of other power generation technologies.

According to the CSIRO’s analysis, producing power from a conventional, large-scale reactors would cost between one and half and two and half times more than from a 90% renewables system backed up by batteries and gas.

Unfortunately, the CSIRO’s costing for nuclear power was not a particularly comprehensive one. It doesn’t adequately take account of the complexity involved in nuclear power plant construction and high risk of budget cost blow-outs with this technology.

The CSIRO’s Gencost publication assumes the cost of building a nuclear reactor in 2030 will be around $8.5 billion for a one gigawatt unit. Yet the experience from real world projects across Europe and the United States indicates the lower bound cost is $14.9 billion and the upper end is $27.5 billion.

The nature of the CSIRO’s Gencost publication is that it needs to provide single point estimates for the cost of constructing different power generation technologies which only allows for slightly varying levels of technological progress over time.

It isn’t intended to capture the varying risk of construction cost blow outs across technologies.

However, this is a major problem with nuclear that would be reckless to ignore. As I’ve touched upon previously,  Oxford University’s professor Bent Flyvberg has built up a large database on the track record of major construction projects across the globe in managing to achieve their planned budgets and time-frames.

That database indicates nuclear projects are characterised by extremely high risk of construction cost blow-outs, with budgets typically blowing out by 120%. Solar and wind projects by contrast have some of the lowest risks of budget cost blow-outs.

CSIRO’s costing also accepted the nuclear industry’s claim that they can realise a more than halving in their costs relative to recent experience through what they term “Nth of a kind” efficiencies.

The “Nth of a kind” costing is based on a theory that the nuclear industry will be able to dramatically reduce their costs as they build more of given type of reactor until they reach the Nth number.  

Exactly what number N might be however is usually left rather rubbery. While learning by doing cost reductions have been proven to unfold in renewable energy and batteries, in nuclear power the opposite has occurred.

Instead costs for nuclear have increased over time as the industry has been required to make modifications to address vulnerabilities in reactor safety.

In the last few months we obtained a concrete example of why CSIRO’s costing is unrealistic for the Australian context.

The CSIRO have indicated that they based their capital costs on Korean experience, which is the only developed country that has had much success in building nuclear on time and budget.  

However, in a tender held by the Czechia Government this year, Korean Hydro and Nuclear Power offered to build two reactors for a cost of $14.9 billion per gigawatt, not the $8.5 billion CSIRO estimates.

These would be built on the site of a pre-existing nuclear power plant with all the pre-existing infrastructure and cost savings that provides.

No doubt the Liberal-National Coalition will dismiss all of this by pointing to a range of nuclear “experts” to claim nuclear will deliver us lower costs. Maybe they’ll be right and the nuclear industry will finally get their act together to deliver projects on time and on budget.

But here’s the ultimate test – are those experts prepared to sink their own money into the project rather than Australian taxpayers?

If nuclear technology is so cost effective then why can’t Electricite De France (owned by the French Government); or Westinghouse (owned by overseas pension holders via Brookfield), or Korean Hydro and Nuclear Power (majority owned by the Korean Government) stump up their own cash to build reactors around the country?

Why should Australian taxpayers be required to take on this risk?

In reality Peter Dutton and Ted O’Brien don’t actually need to get involved in evaluating power generation technologies in which they have no qualifications or experience.

Instead, if they are genuinely committed to achieving net zero emissions why not convert this into a legislated requirement for electricity generators to achieve a steadily reducing cap on carbon emissions over time.

They can then overturn the ban on nuclear power and leave it up to people with vastly more power sector experience than them to work out which technology is cheapest.

Back in 2006, the Howard Government-commissioned Switkowski nuclear inquiry came to precisely this conclusion. In response the Howard Government reluctantly agreed that they were right.

Almost twenty years later one is left wondering why is it now so hard for the Coalition to back their emissions target with legislation and put their faith in the private sector to deliver?