Discover more from Fresh Economic Thinking
The 'sunk cost' trickery that makes renewables seem cheaper than they are
CSIRO's GenCost report needs some careful scrutiny
This post is based on a Twitter thread originally posted by @QuixoticQuant here.
As always, please like, share, and subscribe.
The importance of supporting infrastructure
It’s become widely acknowledged that solar and wind generation costs have fallen considerably in recent years. By some measures, even those including lifecycle costs such as the levalized cost of electricity, each kilowatt hour from these sources appears much cheaper than traditional fossil fuel generators.
But it’s also a fact beyond dispute that wind and solar generators require considerable supporting infrastructure - including extra transmission, storage, and backup ‘firming’ generators - in order for their intermittent nature to be accommodated, and for those kilowatt-hours of energy to be delivered in the places, and at the times, that actually meet electricity demand.
For a genuine comparison, it is important to factor in the additional cost required for the additional supporting infrastructure that allows intermittent generators to actually displace electrical energy otherwise supplied from ‘dispatchable’ thermal generators like coal or gas. The task isn’t necessarily simple. The supporting infrastructure comes in big, lumpy projects, highly specific to the geography and meteorology of a grid. But this is the task that CSIRO, Australia’s government-funded scientific agency, has set about achieving in their GenCost report.
The headline graph that this report produces is probably the most important graph in Australia’s energy transition. It single-handedly supports the idea that wind and solar are cheap, even accounting for the supporting storage, transmission, and firming generation they require.
However, the numbers presented in this graph are bogus. By use of a bizarre ‘sunk-cost’ assumption in their modelling, CSIRO cleaves the cost of infrastructure built prior to 2030 (when we would supposedly already have reached over 50% renewable penetration) from any solar and wind generators built thereafter that might depend on that infrastructure.
Unpacking the infrastructure exclusions
We can see this on page 52 of the report, which describes their ‘Business As Usual’ scenario, i.e. all the projects which are specifically excluded from the incremental costs of integrating new solar and wind. It’s packed full of mega-projects, massive storage, transmission, and firming generation capacity to support renewables:
Let’s unpack that:
First is Snowy 2.0. A massive pumped-hydro scheme, initially priced at $2 billion, that now looks like it might reach $10 billion.
Then there is the ‘Battery of the Nation’. It’s currently just a concept study, but it’s for another massive pumped-hydro scheme in Tasmania. Similar size to Snowy 2.0. It'll be billions. You’ll need both hands.
And "various transmission expansion projects already flagged by the ISP process to be necessary before 2030" So VNI West, Marinus Link, HumeLink, Sydney Ring... I'm losing count of the billions...
Then there are the gas peaker plants at Kurri-Kurri and Illawarra. Together they’re about a gigawatt of firming capacity and will cost roughly a billion dollars.
Much larger is the 16GWh of storage in NSW, presumably batteries. The trend between Hornsdale (both phases), Torrens, and Kwinana is that prices are approaching one billion dollars per gigawatt hour. But allowing for scale, and a lesser discharge rate, perhaps it’ll be half that. Perhaps $8 billion.
All of these tens of billions of dollars of projects are explicitly excluded from the cost of integrating renewables.
So what on earth did they actually include when assessing the cost of supporting infrastructure? Basically, not the mega-projects. Just some tiny, constant amount of transmission within and between existing Renewable Energy Zones, where wind and solar are mostly located, and a minuscule amount of storage. As anyone building a solar home will know, supporting batteries make the panels look cheap.
In other words, the Integrated System Plan (ISP) from the Australian Energy Market Operator (AEMO) demands that we build an enormous new transmission and storage system to move renewable energy around the country, clock, and calendar. Lots of it is lumpy massive mega-projects, mostly needed by or around 2030 when will be approaching 50% renewable penetration. That’s declared “Business as Usual”, so CSIRO then just prices the much more modest transmission and storage expansions required for generators to connect to that system thereafter.
How CSIRO justifies the exclusions: “Sunk Cost”
But wait, this deception is so brazen and transparent, surely someone else would have raised this in an earlier draft or something? Oh, but they have. CSIRO devotes several pages to exactly such an objection. (Page 94, Appendix D, Section 2.3)
What the authors of CSIRO’s GenCost said in response, is simply staggering. Every economist, politician, and policymaker relying on this report simply must hear about this. They explicitly and clearly defend the idea that all the prior investment must be treated as ‘sunk’. They’re adamant that these investments are like risky private investments and therefore there must be a penalty for bad investments made by people who didn’t adequately study what would most likely be required in the future.
Perhaps most incredibly, they explicitly make this claim:
“The market does not owe the owner a reasonable return on their investment.”
Let’s compare that claim with how things actually work with big transmission projects. It’s the exact opposite. Read the details here.
Put simply, transmission projects are approved to be added to the ‘regulated asset base’ by AER, based on projected economic value. And then we guarantee a rate return to the owners of the ‘regulated asset base’. Funded by electricity users. Through power bills.
The business cases for these projects rely directly on the assumption that they're enabling efficient use of new solar and wind for decades into the future. Here’s an example from AMEO’s Integrated System Plan, Appendix 6, showing how all the benefits accrue in the decades after the projects are built.
And what of the large-scale storage and firming generation? Snowy 2.0, Kurri Kurri? They're just funded up-front by the government, normally as the sole shareholder of a state-owned enterprise. Those investments are also justified on the basis that they’ll facilitate the increased adoption of renewables. If that public owner doesn’t get a reasonable return on investment, all Australian taxpayers are worse off.
With incredible honesty and clarity, at the bottom of a 100-page report, CSIRO admits the unthinkable; their model divorces the costs of earlier infrastructure from the later renewables they facilitate. Their assumption is that only temporally coincident investments with the ‘last entrant’ at a given penetration level should be considered:
If anyone has even a moment’s hesitation in rejecting that logic as entirely absurd and uneconomic, have a careful read of this discussion piece about transmission ‘financeability’ that Energy Networks Australia tweeted at the Energy Minister recently. They describe the extent of the financial difficulty caused by the fact that large transmission projects have to wait so long before enough generators connect, and benefits are delivered, to pay what we deem is owed to the asset owner.
And so this absolutely disastrous circular logic is closed:
Politicians build transmission and storage because they think solar and wind are cheap because science says so.
Science (i.e. CSIRO) says solar and wind are cheap because high transmission and storage costs required to facilitate these renewable generators are an already built 'sunk cost' and ignored in their calculations.
A government 'scientific agency' is giving its authoritative scientific analysis to economists and politicians, which is actually polluted by a catastrophic economic error.
We’ve seen examples before where a government agency, so explicitly devoted to doing such detailed analyses correctly, might have made such a simple yet catastrophic error, setting the nation back years, and billions of dollars. It seems perfectly possible that this might be another.
There are plenty of other technical issues that one could and should raise with the rest of the report. But this particular error seems so clear-cut, and massively consequential, that it demands a singular response.
Why should Australians keep accepting the claim that renewables are cheap, even with supporting infrastructure, when the report that justifies this claim explicitly excludes the infrastructure mega-projects by calling them a sunk cost?
Please consider sharing this post, or the original thread with friends, family, politicians, or journalists that you know. This report has been relied upon by countless news agencies, politicians, government departments, and prominent public figures, as I’ve listed in another thread here. It’s important that people understand the true assumptions that the public narrative is built upon.
Thanks for reading Fresh Economic Thinking! Subscribe for free to receive new posts and support my work.