Merit Order and Marginal Pricing
The merit order is the queue of power stations sorted by how much it costs to run them for one more hour. The market fills demand from the cheapest end of that queue upwards, and the price everyone pays is set by the last plant needed to meet the final megawatt. Plants further down the queue receive the same price even though they cost far less to run, and this is deliberate, not an accident of design.
The stack
What matters for the ordering is the short run marginal cost: the cost of producing one more unit of electricity now, with the plant already built. Capital costs, debt and past investment are irrelevant to this decision, because they are paid whether or not the plant runs.
That produces a fairly consistent ordering.
- Wind and solar sit at the bottom. The fuel is free, so the cost of another megawatt hour is close to zero, and can be effectively negative where output is paid per unit produced.
- Nuclear is next. Fuel is a small part of its costs, and the plants are built to run continuously.
- Hydro is a special case. The water itself is free, but a reservoir can only be used once, so its operator prices it against what the same water would earn in a more expensive hour later.
- Coal comes higher, with fuel plus the cost of emitting carbon where a carbon price applies.
- Gas is usually higher still, though the ordering between coal and gas swaps whenever fuel or carbon prices move far enough.
- Oil and older peaking plants sit at the top, expensive and rarely needed.
Stack these from cheapest to dearest and you have a rising supply curve. Draw demand for the hour as a vertical line, because most consumption does not respond to price within the hour, and the point where the two meet is the market outcome. Our fuel pages show how much each of these technologies actually produces in each country.
The last plant sets the price
At the crossing point sits one plant, part loaded, that is just worth running. It is called the marginal plant, and its running cost becomes the market price for that hour.
Everything below it in the stack receives that same price. A wind farm whose marginal cost was zero and a coal plant whose marginal cost was 60 both receive it if the marginal plant needed 85 to be worth running.
The reason is that the market is trying to produce two things at once: the cheapest way to meet demand, and a price that tells everyone the truth about what the next unit of electricity costs. Uniform pricing does both. Because each seller knows the price will be set by someone else’s offer, the safest strategy is to offer at your own running cost, which is exactly the information the auction needs to sort the stack correctly.
The gap between the price and a cheaper plant’s own costs is not a mistake either. It is how plants that are expensive to build and cheap to run recover their construction costs, since a plant paid only its fuel bill would never pay back the money borrowed to build it.
Why gas so often sets the price
Gas plants have two properties that put them at the margin more often than their share of generation suggests.
They are flexible, so they are the plants most naturally used to fill the gap between whatever the weather has provided and whatever demand happens to be. And they sit in the upper middle of the stack in most systems, which is the part of the curve that demand crosses on a normal day. Cheaper plants are already running at full output, and more expensive plants are not needed.
The result is that gas can set the price in a large share of hours while producing a much smaller share of the electricity. It also means the wholesale price tracks the gas price closely. When gas became expensive across Europe in 2021 and 2022, electricity prices rose with it even in countries where gas produced a small minority of the power, because gas plants were still the ones at the crossing point.
Where gas is also the largest fuel outright, the link is even more direct. Ember’s statistics show gas producing about 40% of all electricity in the United States in 2025.
What changes as wind and solar grow
Adding generation with near zero running cost does not change the mechanism. It pushes the crossing point further down the stack, which is why heavy wind or solar output lowers prices in the hours when it occurs. This effect is large enough to be visible in any market with substantial renewable capacity.
Three things follow.
First, prices become more variable. A calm evening and a windy midday now clear at very different points on the stack, so the spread between hours widens.
Second, the marginal plant is less often a conventional generator. A battery deciding whether it is worth discharging now or holding charge for later, an interconnector importing from a neighbouring market, or a large consumer choosing to pause, can all be the unit at the margin.
Third, in the hours when zero cost generation exceeds demand, the crossing point falls off the bottom of the stack and prices can go to zero or below. That case has its own explainer: why electricity prices go negative.
Where the simple picture stops
The textbook stack is a good model of how prices form, but real dispatch is more complicated in ways that matter.
Plants have start up costs and minimum run times, so a unit may run at a loss for an hour to avoid a costly restart. Some plants must run for reasons other than energy, such as heat supply or local voltage support. The transmission network constrains what can be delivered where, so the cheapest plant available is not always usable. And operators hold reserves back from the energy market, which removes capacity from the stack.
None of this changes the central idea. Meet demand from the cheapest available sources upward, and let the last one needed tell you what the next unit is worth.