The Levelized Cost of Horseshit

Taking apart the one statistic that muddles every energy debate

Clicking “like” helps new people find One Percent Brighter.

The Levelized Cost of Electricity. In the perpetual nerd wars between the nuclear bros and the wind-and-solar bros, it’s a live force. LCOE, as it is known, is the one number that’s supposed to settle the key question: which is cheaper?

LCOE has a nice technocratic ring to it — a no-nonsense aura that covers up, it turns out, an immense mound of nonsense, all smothered in enough complexity to befuddle most people.

But make no mistake: LCOE lies to you.

It’s a subtle, insidious falsehood that has left climate activists convinced that wind-and-solar are now the cheapest sources of energy. They’re not – at least not in any sense that matters in the real world. LCOE just makes it seem like they are, resulting in a lot of confusion as well as terrible policy.

So what’s the problem with LCOE, exactly?

To get at it, you need to know where the concept comes from.

Back in the 1950s, utilities faced a problem: how do you compare the costs of very different sources of energy? Say you’re trying to decide if you should build a hydroelectric dam or a coal-fired power station — which one’s cheaper? Wouldn’t it be nice to have a simple formula to settle it?

What if you just calculated the cost of building and running each of them over their entire lifecycle, then divided the total amount of power they’ll generate by that total cost?

Do that and you get a nice, neat, tractable number, expressed in dollars per megawatt-hour (the standard measure of power output). That one number lets you compare the dam and the coal plant on a like-for-like basis. That’s LCOE.

So far, so logical.

Conceived long before weather-dependent renewables like wind and solar were a thing, LCOE doesn’t account for the intermittency that is inherent to them. It wasn’t designed to do that. It works by averaging out all the power a solar panel or a wind turbine produces. That means it treats each megawatt-hour produced as equal to all the others.

This makes wind-and-solar look extremely cheap, because, yeah, when the sun is shining it costs almost nothing to generate power with a photovoltaic array. And when it’s good and windy the marginal cost of wind power is basically zero.

But, of course, all megawatt-hours are not equal: megawatt-hours generated at 7 p.m., when people are coming home after work, turning up the heater and cooking dinner, is far more valuable to a grid operator than a megawatt-hours generated at 2 p.m., when the sun is shining bright but demand on the grid is low.

In the real world, grids need to cover peak demand reliably — and power sources that depend on the weather can’t do that.

When you serve actual humans who expect the lights to come on regardless of the weather, you quickly realize that you’re going to need a second set of generators to backstop wind-and-solar when it’s gloomy, dark, and/or calm.

Usually, this is done with fossil fuel-powered plants, which are good at two things: ramping output up-and-down quickly to match demand, and killing people by putting particulate gunk into their lungs. (To be fair, they’re pretty good at spewing greenhouse gases too.)

What’s really perverse is that the cost of this fossil-fueled back-up system isn’t counted in wind-and-solar’s LCOE!

Oh no, that we count separately – in the LCOE account for the fossil fuel plant.

It gets worse.

In many jurisdictions, environmental regulations force grid operators to give wind-and-solar preferential access to the grid, letting them offload all their power first, and only bringing other sources as needed to cover when renewables flake out.

So now you have to keep that big old dirty power station always ready-to-go, but idle most of the time. Or, if not idle, running far below capacity.

Run this by your accountant.

He’ll tell you that, to stay in business, you’re going to have to charge a lot when the plant is turned on. Because now you now have to recoup all the capital investment it took to build the plant on what you make during those unpredictable periods when renewables are out to lunch.

This means the gas plant’s LCOE numbers look terrible, while the supposedly ‘levelized’ numbers for wind-and-solar don’t reflect any of the extra cost and instability they bring into the system.

Of course, environmentalists then jump on these numbers and say, “See! wind and solar are much cheaper than gas or coal, we should build more wind-and-solar!” Pretty soon, they’re lobbying politicians to subsidize new wind-and-solar either with tax credits, or with even more favorable market rules.

So more and more wind-and-solar gets built.

With every new wind turbine built or solar array installed, the grid goes farther out of whack. Prices get more volatile, crashing when it’s sunny and windy then spiking when it isn’t. Industrial facilities start to shut down, because there’s no way to plan production when your power bill is a crap-shoot. They relocate, usually to places still heavily reliant on fossil fuels. This turns up in the statistics of the place they left as diminished carbon emissions. Then environmentalists say: “See? It’s working!”

It’s messed up. But shockingly few people understand it.

This amazing piece from the Guardian a few weeks ago shows this perverse dynamic in all its glory: it reports in horror that British gas-fired power stations are demanding incredibly high prices on days when the weather isn’t cooperating.

The journo behind the piece apparently never stops to consider that somebody has to keep the back-up barely used gas plants operational 24/7. Ideology is a hell of a drug, though. She has all the elements of the story right there in front of her. She’s reporting them, ably. But she can’t put two and two together. Not in The Guardian, anyway.

We need to step outside green ideology to actually grasp how strange the LCOE debate has gotten. Maybe an analogy will help. The way one reader put it, wind-and-solar are cheap in the same way a trailer is fuel-efficient. It may not use any fuel, but without a truck, a trailer is useless. LCOE is an absurd metric in that it computes the trailer’s fuel efficiency separately from the truck’s!

Like a truck and trailer, in the real world renewables and fossil fuels are part of one system. It would make more sense to add up all of the capital costs and all of the running costs of both together, then divide them by the total number of megawatt-hours produced. If you did that, you could report a kind of omnibus LCOE for the combined wind-and-solar-and-gas system, which is the one you actually depend on to keep the lights on.

Lazard, the financial advisory firm that did most to popularize the LCOE concept, has recently begun to address the problem. Their newer reports try to measure the cost of “firming” renewables with back-up power sources. Their results vary a lot from place to place, but firming is never cheap.

To give one extreme example, they find that in California, the LCOE of solar power is just $49 per megawatt-hour, but the cost of a usable megawatt-hour of solar, backed with fossil fuels, climbs all the way to $153. That’s more than the LCOE of building a new nuclear reactor in the U.S., an avenue greens routinely derided as far too expensive.

Of course, you won’t find many green NGOs quoting those firming numbers. They can get away with it, because the subject is obscure enough to befuddle all but the nerdiest of the power bros. And so they’ll keep obscuring the issue, because their business model depends on you not understanding this post.

Deep down, though, they must know LCOE is horseshit.