Last week, I wrote about some of the silly errors and questionable ideas in Saul Griffith’s Quarterly Essay. I could write about why I think some of his other ideas, like covering 38,000km of freight rail in Australia with solar panels, are poorly thought through.
But I won’t.*
Instead I’ll be constructive and focus my energy on what I thought was the most promising idea in the essay. The opportunity to electrify everything.
One particular chart, which demonstrates the potential opportunity, stood out to me.
This chart came from a study Griffith’s organisation, Rewiring Australia, undertook in 2021.
The chart is labelled and structured slightly differently in the essay but it conveys the same information. In the essay it also has some text overlays which say that “$12b in total subsidies up to 2027” will help lead to “$302 billion in home savings by 2035".
Source: Castles and Cars Technical Study 2021 (cited in Quarterly Essay)
At first glance this chart seems pretty damn compelling.
For every $1 in subsidies that the government provides, households will save $25. That seems like a pretty slam-dunk case for providing subsidies. Let’s get cracking!
But when I looked closer, I noticed this chart has 4 major issues.
The chart does not show the amount of electrification that subsidies would produce
Electrifying all households by 2030 by providing only $1,200 in subsidies per household seems very unlikely
Griffith’s suggested subsidy for EVs is less generous than the Labor government’s current policy
The chart does not show a realistic prediction for electrification (and therefore savings) by 2030
The chart does not show the amount of electrification that subsidies would produce
My first issue with this chart is that it doesn’t actually show the difference (or delta) in the amount of savings or emissions reduction we could expect to see with or without subsidies.
Ideally, you would have two versions of the chart. The first demonstrating the savings when a policy of electrification subsidies is offered and a second where they are not.
More subsidies should translate into more electrification and therefore translate into greater savings. Without the comparison, we cannot tell how much additional savings the policy would drive.
One simplistic way to interpret this chart would be to assume that the $12b in subsidies generates the entire $302b in savings. However this would greatly overstate its impact.
To generate this graph, Rewiring Australia used a future cost profile of EVs based on data from Bloomberg New Energy Finance.
Source: Castles and Cars Technical Study 2021 (page 102)
This cost profile suggests that EVs will become cheaper than Internal Combustion Engine (ICE) vehicles in 2027 regardless of whether or not subsidies are provided. We would therefore expect to see some amount of vehicle electrification regardless of whether or not subsidies are provided.
As a result the graph doesn’t actually show the incremental savings for households.
Additionally, Griffith does not provide any specific details on how the subsidies would work and in fact the Technical Study does not specify how the subsidies are applied in their modelling.
As a result, we don’t know what specific subsidy policy they are advocating for or its impact.
Electrifying all households by 2030 by providing only $1,200 in subsidies per household seems very unlikely
This chart suggests that with $12b in subsidies, 100% of households could be electrified by 2030 (this includes switching to electric cars).
Given that there are roughly 10 million households in Australia, this would suggest that this can be done by providing a $1,200 subsidy per household.
Ignoring all the other machines to electrify, if each household has 1.8 cars (as Griffith’s says), then he is assuming that a subsidy of just over $650 per car will be enough to entice people to switch to an EV.
I don’t think this is enough.
Let’s consider a hypothetical situation.
Cassie is a 19-year old first-year university student. She’s just purchased a 2012 Toyota Yaris for $6,000. She will use this car for the next 4 years until 2027 at which point it will have done over 140,000km and will require costly repairs to continue running. Cassie decides that she should sell it and buy another car. At this point Cassie is earning $30,000 a year from her hospitality job and is in her second last year of study.
She can buy a regular 2017 Toyota Yaris for $9,000. The cheapest second-hand EV on the market is $28,000. She is annoyed that her old car has broken down and she doesn’t want to spend much money on a new car as she is saving up for an overseas trip before her final year of study.
In this scenario, do we truly believe a $650 subsidy would be sufficient to make her switch to an EV?
I don’t think it will.
Griffith’s suggested subsidy for EVs is less generous than the Labor government’s current policy
The Labor government has enacted a policy which means that EVs are exempt from the Fringe Benefits Tax (FBT). This translates into a subsidy of $10,000 or more per vehicle.
Griffith is suggesting, at best, a $650 subsidy per vehicle and so this would imply that he is suggesting that we should significantly reduce the incentive to purchase an EV.
This contradicts his entire mission of speeding up the electrification of everything.
I don’t think that this is his intention and instead this is just a mistake and contradiction that he is unaware of.
The chart does not show a realistic prediction for electrification (and therefore savings) by 2030
This chart does not show a realistic prediction for electrification.
The chart suggests that 100% of households will be electrified by 2030. It’s not made totally clear what that means. By some definitions (i.e does the house have electricity?) we have already achieved 100% electrification. Of course, this is unlikely to be the definition they intend.
Given that Griffith suggests that cars have the largest impact, it’s reasonable to assume that 100% electrification means that 100% of cars are EVs.
The chart below, which is reproduced in the essay, shows that swapping to an EV is responsible for the majority of the expected reduction in a household’s energy use.
Source: Castles and Cars Technical Study 2021 (page 9)
The problem is that even with some heroic assumptions, 100% of the car fleet is unlikely to be EVs in 2030.
Consider that EVs currently account for only 3% of new car sales. If we assume that this rapidly increases to 100% of new sales within 4 years, that would mean that there would be only 3-4 years prior to 2030 in which all new cars purchased are EVs.
This would not be enough to replace the existing stock of cars.
Put another way, there are roughly 1 million cars purchased each year in Australia and roughly 20 million cars registered. Even if all the cars purchased between now and 2030 were EVs (roughly ~7m), there would not be enough cars to entirely replace the stock of ICE cars.
So despite seeming compelling at first glance, this chart does not provide us with much actionable information
Now that we’ve looked closer at this chart and the modelling that sits behind it, what are we left with?
We’re left with
No concrete policy proposal or subsidy structure
No accurate estimate of the savings or emissions reduction that $12b in subsidies would generate
No realistic forecast of the electrification of vehicles
We’ve gained very little insight from their analysis.
Of course, I haven’t looked deeply at the other sections of their analysis but based on what we’ve seen so far, I wouldn’t be surprised if it contained some equally implausible assumptions.
Conclusion: Unfortunately, this essay is quite bad
I read this essay hoping to hear some interesting and well-thought out policy ideas. Unfortunately, this was not the case.
From sloppy editing (there are times when Griffith refers to Figure 8 when he means Figure 9 and vice versa) to the substantive ideas, this essay is quite bad.
Consider the two big ideas that I’ve mentioned in these posts; Building a 15,000km bike path covered in solar and covering 38,000km of freight rail with solar panels. Just these two ideas alone would likely cost in excess of $50 billion.
You could probably build 4x the amount of renewables if you focused instead on building transmission and generation assets closer to existing transmission lines and population centres.
$50 billion, or something in that vicinity, is a lot of money. And to be suggesting it be deployed in such an ineffective way undermines Australia’s efforts to reduce our emissions.
The essay does not engage at all with the Australian Energy Market Operator (AEMO) plan for the future of the grid in the National Electricity Market (NEM). If Griffith had made a critique of AEMO’s plan that would have been a constructive contribution to the public debate. Instead, we’re left wondering whether Griffith’s is even aware of the Integrated System Plan (ISP) that AEMO has produced.
We are also left wondering whether he believes we should prioritise the “round-the-nation” bike path ahead of the infrastructure investments suggested by AEMO.
Given the many flaws in this essay, I am puzzled as to how someone like Ketan Joshi can describe it as a “very good essay” and write an overwhelmingly positive review.
Accurately quantifying the benefits of electrification policies is a valuable task. This task however, seems beyond Griffith’s capabilities.
It would be nice if someone took the time to do it properly.
Climate change is an important problem and if we’re going to address it sufficiently, we need to do better than this.
*It would be a bad idea because a large chunk of the solar panels will be 1,000km+ from any population centre. You would therefore need to build a ton of transmission. All of this would cost a boatload of money.