This opinion piece was published on Wednesday 18th December in the Australian Financial Review here.
Danny Price is the Managing Director of Frontier Economics.
I stand by the analysis that shows on a system-wide average economic cost basis, including nuclear remains cheaper than a renewables only system.
On Monday [16 December 2024], The Australian Financial Review (AFR) published criticism by Steven Hamilton of the analysis I did on including nuclear power into the National Electricity Market. Hamilton has made strong but erroneous points about the analysis.
It seems Hamilton’s big point is that the modelling must be doing something wrong because the annual average utilisation of renewables does not change with the introduction of nuclear and there must be a lot of spilled renewable electricity – that is, where renewables are turned down because there is too much generation. Hamilton concluded that I must have ignored times when there was surplus renewable electricity.
There is no shortage of horror stories about nuclear plant cost blowouts but there are also excellent recent examples of projects that are far less costly. The first point to note is that we get spilling of excess renewable energy already – that is, where renewables are turned down because there is too much generation. Given the high cost of storage, it is usually cheaper to spill this surplus electricity than store it.
In a renewables-only system, this spill gets larger and larger as the share of renewables increases. This is because you have to build vast amounts of renewables to produce enough electricity to meet demand and since you never know whether they will produce at the same time or at different times, inevitably you end up at times with too much electricity. Given the high cost of storage, it is usually cheaper to spill this surplus electricity than store it. Hamilton, therefore, better hold on tight if he is worried about spill.
When we did this modelling we forced nuclear power into the system and then the model was re-run to re-optimise everything around the nuclear power stations to find the lowest cost solution. When nuclear is included, the model obviously builds less renewable capacity. Of course, there will be some spill, but the model optimisation explicitly minimises this spill. So, contrary to Hamilton’s baseless claim that I have ignored spill in determining the total system costs it is, in fact, a critical aspect of the optimisation approach I use.
Another point made by Hamilton is that he says it is invalid to assume that nuclear costs will improve from now onwards by an assumed 1 per cent per annum and instead implies that Australians will only start learning from the date of the first megawatt-hour generated by a nuclear power station in 2036. How absurd. Engineers, technical advisers, technology suppliers and construction companies are learning all the time from the experiences of other projects around the world. Learning and productivity improvements are going on all the time.
A further point Hamilton makes is that using $10,000 per kW for the capacity cost of nuclear is optimistic because he could find an example in the US where the costs were higher. There is no shortage of horror stories about nuclear plant cost blowouts, but there are also excellent recent examples of projects that are far less costly than what I have used and even lower than recommended by the CSIRO and with learning rates well in excess of what I assumed. It seems certain to me that if you want to make nuclear power expensive in Australia, you would hand the responsibility to the bureaucrats that turned up to the recent Commonwealth Senate Committee on Nuclear Energy who explained how they were going to bind the nuclear power industry in red tape for years.
Hamilton was also suspicious of the reason for the modelling period going to 2051 and implied that the conclusions about the lower costs of including nuclear power might be reversed if the modelling period was extended. Firstly, the modelling period is determined by the modelling period adopted by the Australian Energy Market Operator’s Integrated System Plan our modelling replicates. Equally importantly, it is clear in the report that the last nuclear generator begins operation in 2049. This means the cost structure of the sector reflects the highest quantity of nuclear capacity as a share of output a few years before the end of the modelling period. Given the costs shown for each year is the annualised cost, the costs of nuclear doesn’t go any higher.
On a system-wide average economic cost basis, including nuclear remains cheaper than a renewables-only system in equilibrium and in transition. Over time, the economic costs are lower, including because we are making use of the existing coal plant, which makes economic sense. I clearly explained in the report.
One final point on Hamilton’s critique is that I did not include a cost of emissions based on the regulator’s value of emissions reduction (VER) that I reported in Report 1. The VER is designed to make otherwise uneconomic abatement policies appear economic, it does not reflect the economic cost of emissions. I note that Minister Bowen, rightly, does not include the VER when he reports his $122 billion present value transition cost. More importantly, the additional emissions are measured against AEMO’s unrealistic early closure schedule, which clearly is not going to happen. Coal generators will likely close much later than AEMO’s assumptions. This means that the additional emissions I reported will be overstated, which I also explain in the report, but this was ignored by Hamilton.
Please contact the Frontier Economics media team for more information.
