By Andy May
Lazard’s levelized cost of energy (LCOE) is cited on the internet all the time as the source for “solar and wind are cheaper than fossil fuels.” They don’t really mean “energy,” they mean “electricity.” The world consumed only 18% of its energy in 2021 as electricity, so LCOE is just the cost of 18% of our total consumption, a fact often lost in these discussions.
However, just a quick look at their data shows that solar and wind are clearly not cheaper. Even within their April 2023 report they are not consistent in their numbers. To make matters worse, they bury critical details in the fine print and do not define their terms. I doubt some of their numbers, but for this discussion I only use the numbers in their report.
Figure 1 is based upon Lazard’s chart on page 8. This is the chart where they try and include the costs of backing up the intermittent nature of solar and wind power generation. The chart is complicated and poorly explained, so I’ve added some clarifying annotations. Solar does not work at night or on cloudy days and wind obviously doesn’t work if there is no wind, some sort of backup (“firming”) is needed for when the Sun isn’t shining, and wind isn’t blowing.
They examine four scenarios, labeled “MISO,” “CAISO,” “SPP,” and “PJM.” They do not explain what these cases are, but I assume they are from specific electric utility companies. The graph shows the cost of Lazard’s unsubsidized LCOE in black, their subsidized cost is in light blue, and the backup or “firming” cost is in beige. They do not specify the backup duration for natural gas, but the specified outage time planned for the Lithium-Ion battery case (CAISO) is only four hours. Windless periods are usually at night and in the winter, when it is dark over 12 hours a day, so I have no idea where “4 hours” came from.
The units on the chart are $/MWh (U.S. dollars per Megawatt-hour). The units for the assumed backup cases in the fine print (see note #1) are kW-mo (kilowatt-months). This is probably to confuse the reader; I can think of no other reason. There are 730 hours in a month and 1,000 kW in a MW, so the conversion is x730 and then divide by a thousand. The figure does the conversion for you.
Notice the Lazard figure specifies a range of LCOE from $39-$101 per Megawatt-hour for Natural Gas Combined Cycle power generation. Yet, in the fine print, they specify that the “Natural Gas CT” backup assumptions for solar and wind (used in MISO, SPP, and PJM) are from $6 to $7.45 per Megawatt-hour. They do not define “Natural Gas CT,” but I interpret it as conventional natural gas power generation. Conventional natural gas plants are cheaper than combined cycle, but less efficient.
How is it that the cost of “Natural Gas CT” is a sixth or less of the cost of natural gas LCOE, when used to back up solar and wind? Can anyone clarify this? I’m no expert, but this looks like disinformation to me. In figure 1 the Natural Gas CT cost, only when used to backup solar and wind, is shown as a red line. It is much, much cheaper than the solar and wind costs provided, whether they are subsidized or not. Since wintertime is when solar and wind fail most often, and it can be for days at a time, batteries are clearly ineffective as a backup. Besides batteries fail most often in wintertime also, as anyone who drives a car knows. If we are using this magical and mysterious “Natural Gas CT” power plant at $6/MWh to back up much more expensive solar and wind, why bother with the solar and wind?
For a more in-depth look at the relative costs of solar, wind, coal, natural gas and nuclear see here.