The LCOE and RCP8.5 One-Two Punch
How they work together to mislead and what we should do about it.
In the energy blogosphere, there are more “Debunking Levelized Cost of Electricity (LCOE)” posts than you can shake a stick at. Like watching videos of people walking into glass doors, reading these articles never gets old. However, in our review, each previous take failed to ask an essential question:
Why do people keep using LCOE if it’s clearly misleading?
Before we answer that question, let’s be a good lil’ energy blog and fulfill our LCOE debunking duties.
Lazard, a financial services firm, created the LCOE metric in 2007 when they set out to answer the question, “How do you compare the cost of different sources of electricity?”
The LCOE calculation is satisfyingly simple — just take the plant’s forecasted total cost (installation, fuel, maintenance) and divide it by the forecasted electricity output over its lifetime. There’s some net present value finance-y voodoo involved, but nothing too crazy.
The result of the simple calculation is a cost per kWh metric for the simple-minded that can be used to compare the cost of electricity generation from different sources — wind, solar, coal, nuclear, natural gas, etc.
If you’ve ever heard “solar and wind are the cheapest sources of electricity,” it comes from the LCOE analysis displayed in the chart above. It’s a classic example of drawing incorrect conclusions from correct numbers — a phenomenon all too common in today’s energy and climate conversations.
Besides everything, the problem with the LCOE metric is that it ignores the grid’s golden rule:
Electricity supply must match electricity demand, in real-time, always.
If the golden rule isn’t followed, grid infrastructure fails and we suffer blackouts.
Integrating renewables into the grid requires incurring extra costs to handle their intermittent and uncontrollable nature. The LCOE metric doesn’t account for this additional, system-level cost. Regions with a high penetration of renewables maintain a “spinning reserve” fleet of natural gas power plants that step in immediately when renewable electricity production inevitably drops. The cost of operating underutilized natural gas plants is ultimately passed onto consumers.
To make matters worse, renewables must be sited in places with ample land and (ideally) plentiful sun or wind. These places typically aren’t close to large population centers where electricity demand is highest, so renewable generation also requires a large amount of transmission infrastructure to connect supply to demand.
These factors explain why grids with a high degree of solar/wind penetration (eg California and Germany) have higher electricity prices than grids that don’t.
Isn’t it funny how LCOE shows renewables are the cheapest but they actually make electricity more expensive? Once you see how misleading LCOE is, you can’t unsee it.
All of this doesn’t mean that LCOE is a worthless metric. It still provides interesting information like showing the incredible cost decline of solar. The problem is that LCOE is misused to perpetuate the myth that renewables are the cheapest way to power the grid. Again, just because solar’s LCOE is cheap, doesn’t mean that running a grid on solar would be cheap. If this is true, then why is it so difficult to stop people from using LCOE in this manner?
That’s where the abuse of representative concentration pathway 8.5 (RCP8.5) in climate science comes into play.
Modeling the climate is a herculean task. To develop a consensus view on how emissions impact climate, modelers need an agreed-upon set of model inputs. Debating outputs from models that have different inputs would be chaotic and unproductive.
This was the rationale behind “representative concentration pathway” scenarios. The climate science community was given a variety of different emissions scenarios to model, but there was no agreement on the real-world likelihood of any individual scenario. The creators of the scenarios even emphasized that “no likelihood or preference is attached” to any specific emissions scenario. This is easy to understand because modeling how emissions evolve is an entirely different exercise from modeling how emissions impact the climate. One involves modeling economic development, the other involves modeling the climate.
Yet, somehow the results of RCP8.5 modeling became the “business-as-usual” scenario for society despite the fact that it has an extremely high level of emissions widely considered to be implausible. It’s a worthwhile academic exercise to put extreme emissions scenarios through climate models. It becomes a problem, though, when the scenario is being used as the baseline for policy-making and news reporting.
has documented the origins and continued abuse of RCP8.5 extensively.Why do people keep misusing LCOE? Because people keep misusing RCP8.5. The abuse of this climate science scenario makes people catastrophize the future which then creates bias when analyzing our energy systems. RCP8.5 convinced us that the world needs saving and the LCOE analysis shows that renewables are our savior.
What should we do about it? In our view, people are able and willing to change their mind — it just takes time. Even though faulty thinking in energy leads to a giant misallocation of capital at best and poverty at worse, we shouldn’t cancel people from spreading misinformation (knowingly or unknowingly) to prevent this from happening.
If you limit the information available to people, you limit debate. If you limit debate, you limit progress. Just take it from this timeless
piece:Even if you think the people doing the critiquing are being disingenuous, or they are funded by nefarious characters seeking to exploit ambiguity for monetary gain, or you convince yourself that the mere airing of such critiques is a danger to society, the moment you give in to the temptation to censor counterarguments – to label them as misinformation, for example – you’ve lost. Either you are willing to outlast your opponents in an extended debate by patiently and calmly rebutting all critiques, or the soundness of your hypothesis must be considered suspect.
We’ve convinced ourselves that winning hearts and minds is easy. It’s not, it’s hard work. We shouldn’t be surprised that change takes time. LCOE and RCP8.5 are case in point.
Our solution isn’t to cancel people to speed things up, it’s to write more.
Thanks for reading!
In Colorado our new Energy Office has a “Net Zero 2040 Energy Pathway” model. Ascend Analytics used an engine with cost-only criteria to design a generation system for our state. The Energy Office was pleasantly surprised that we would reach 98.5% clean energy by 2040 at no additional cost! They were a bit apologetic that the plan was less than 100%…
Magic!
Nuclear power was priced at 10 to 1 over solar using LCOE from NREL.
I can see how these two punches could mislead a lot of stuff! Thanks for clearing this up.
I'm a bit guilty of abusing RCP8.5 myself — while assessing the risks that climate change might pose to the assets of a major European company, I and my team used RCP4.5 (which assumes a more-or-less plausible level of emissions) and RCP8.5. However, most of our focus was on the latter scenario because we hoped that this would help the client company understand what dramatic events might hit it in the future. (This, in turn, led to talks about buying a climate insurance policy which, incidentally, the company I was working for was selling...)