Let me come straight to the point: even though it’s ubiquitous in the literature, I’m not a big fan of levelised cost of energy (or levelized cost, if you prefer your spelling US-style). It’s not its existence as a metric that concerns me; it’s how it gets used. Too often I see it dropped into reports, presentations and news articles and stated as a fact, to justify one or other economic conclusion.
Yet like any calculation, there’s the issue of “rubbish in, rubbish out”. Even based on sound data, LCOE is a metric that doesn’t tell you much if you just take it at face value, without being told – or asking – the assumptions behind its derivation.
- What was the person who calculated it using it for?
- Do the assumptions seem relevant to their case?
- Did they calculate the number themselves, or just repeat one reported elsewhere?
- How sensitive is the conclusion to the assumptions?
- Rather than just the number quoted, what is the possible range of values?
This isn’t going to be an article about methodology, other than to say:
The basic concept is quite clear and universal:
Add up all the costs of a power generation project over its lifetime (capital, operating, fuel etc.). Do the same for all the energy production by that same project over its lifetime. When adding up, use a discount rate to discount future value in both cases. Now divide the lifetime (discounted) costs by the lifetime (discounted) energy generated. The result: a nice, average “cost-per-unit-of-energy” (e.g. units of $/MWh) metric.
However the details of different people’s calculation methods vary:
This is mainly a matter of what they choose to include and exclude from the cost part of the equation. I’ve written an accompanying reference post, with links to some examples and discussions of LCOE elsewhere. If you find that reading them turns something you previously thought of as simple into a topic of uncertainty and concern, then that’s no bad thing. (I usually aim for the opposite, but then the whole purpose here is to persuade you to question a metric that is too often presented as straightforward fact).
Comment 1: Before you use it, know how a number was derived.
Comment 2: LCOE numbers are useful for comparing energy costs derived in a consistent manner in a single study; less so (and often not at all) for comparisons based on numbers picked from different studies.
The Where and the When
There are various location-dependent inputs into a LCOE calculation.
The energy output from a solar farm or wind farm can obviously be very different from place to place, depending on climate. This will certainly be true between different countries (Riyadh vs. Glasgow) but can also be the case within individual ones (Boston vs. Las Vegas). Aside from some assumed degradation, this is unlikely to change over the lifetime of the plant.
What about conventional power plants?
The levelised cost of “gas power generation” is very different depending whether we calculate for one which runs rarely as a “peaker” (so just producing a small amount of energy) or one which runs on a much more regular basis (producing a lot). Not only can the appropriate assumption vary from place to place (depending on the energy mix, grid layout and economics of dispatch), there is also potential for major change over the lifetime of a project (as those same factors evolve).
On the cost side, everything from equipment to labour, finance to fuel and tax to insurance will vary depending where your project is. So building and operating the same plant in Denver vs. Dhaka will likely sum up a very different list of costs.
Looking ahead, conventional power plants gain because future fuel costs are discounted in the calculation; compared to renewable power projects where cost is mostly front-loaded.
However a gas plant (for example) has variability on both the cost side (in particular the gas price, but potentially any or all of the others too) and – as mentioned above – its energy output. What assumptions have been made about all these changes, over the project lifetime, in calculating that nice, simple LCOE number?
On the other hand, since the costs of and energy production from a solar farm are pretty much known from the start, so too is its LCOE. Though one thing to be careful of in any study is when that “start” is: is it based on costs today or costs projected (on what basis?) to some future date.
Comment 3: Don’t use an LCOE number calculated for one location to make assumptions about energy costs elsewhere.
Comment 4: Conventional power plant LCOEs are more sensitive to variation caused by future assumptions than are renewable-power ones. (This is not just because of fuel costs, but because of changes in market dispatch economics too, e.g. by introducing more renewables into the mix!)
Cost not Price
Levelised cost is… well, the clue’s in the name!
What it does not say is anything about is the value of energy over the lifetime of a project, which depends not on the cost of generation but the price for which it can be sold. So, just because one source has a higher levelised cost than another, that doesn’t mean it can’t find a profitable role in a market.
Value derives from generating at the times of highest demand, when people most need electricity. So it’s been pointed out by plenty of commentators that in hot countries, where solar power generation often coincides with air-conditioning-driven peak demand, it replaces low capacity-factor (high LCOE), conventional-fuel plants. Solar simply doesn’t need to be cost-competitive with low-LCOE baseload sources, so comparing the two is pointless.
Of course whether “value” is reflected in “price” is a matter of where and how the electricity is being sold!
In some cases electricity is sold on true, auction-based markets, with prices adjusting constantly according to balances of supply and demand. In others, prices are set by governments. Often, where renewable power has been introduced, conventional generators face volatile market prices while its renewable competition has the safety of fixed FiT (feed-in-tariff) prices; be they determined by government directly or following a bid process.
Comment 5: It is price (and policy), not LCOE, which largely decides whether power projects make sense for investors. Is the margin between price and cost paying them back generously enough, or fast enough? (Requirements which may vary from market to market according to their view of risk).
Comment 6: Variations in price/revenue between markets (plus future assumptions of price changes) could outweigh any variations in underlying LCOE.
Beyond the gate
LCOE considers only the costs up to the “gate” of the power plant (or as far along the grid connection as the developer has had to pay for). As an investor in that plant, that’s fine; I just need to make sure that the price at which I can sell electricity on from here gives me a big enough profit.
As a consumer however, the “costs of energy” certainly don’t end at the power plant gate! Here in the UK, only about 50% of my electricity bill comes from wholesale electricity costs (and the profits of those generating it). A big chunk of the other half comes from getting it from the power plant to the plug in my wall: transmission and distribution (“the grid”). So if some entity within “the grid” has to pay to extend or reinforce transmission lines to some distant windy spot, or even offshore, that has to be paid for. Ultimately, by me! Also, hidden within that “wholesale” part of my bill are any costs associated with balancing supply and demand, including any costs that may accrue elsewhere in the system as a result of integrating a new generator into it.
There have been plenty of studies looking at the additional costs (or not) of balancing variable solar or wind power using conventional, “dispatchable” generators. They draw widely different conclusions, but the key point remains…
Comment 7: Not only can LCOE have little or no relation to the wholesale value/price of electricity generation in a market, it’s even further distant from the retail price, that end-users will pay, since it considers no costs beyond the power plant.
Comment 8: “Grid parity” is another term applied too often, in my opinion, rather vaguely. For distributed generation – residential solar PV for example – there is a valid comparison to be made between LCOE (the cost of energy generated by the system on your roof) and the cost of energy provided through the grid (i.e. the price you pay to your utility). But how should we interpret claims of “grid parity” with reference to a large-scale, transmission-connected wind farm? By using its LCOE? Or its full costs, including external ones? By comparing to what? The wholesale price of which other generators? At what time of day?
There are some other metrics you may come across, such as “Levelized Avoided Cost of Energy” (LACE) and even “Societal Cost of Energy” (SCOE). The former tries to take account of the wider system costs of building a new power generator, while the latter looks wider still; including economic impacts (job creation), carbon costs and so on. I’ve written briefly about these in the past and I hope these and other metrics will be further developed as time goes by.
From a policy point of view, surely LCOE is too simplistic a metric on which to base energy choices? Why only consider internal costs, when for your citizens the final price they pay includes at the very least the other system costs; and potentially significant additional impacts, for example on their health, environment and job prospects?
Comment 9: LCOE is a useful metric for comparing “out of the gate”, lifecycle energy costs between different power generating options, when calculated within a closely constrained and consistent study. However, in my opinion, LCOE numbers are thrown around far too often, with far too little context, as a basis for talking about future energy mixes and policy decisions.
Closing comment 10: I’ve nothing against LCOE. No, really I haven’t! 🙂 It’s a metric that’s been around for years, is commonplace in discussions of energy, and is often calculated thoroughly, based on sound assumptions. My point is not that you should ignore LCOE, but to urge that you stop, think and question, before reading – or indeed writing – more into it than it deserves.