Capacity Markets Explained

A capacity market pays generators, storage operators and flexible consumers to be available at times of system stress, separately from any payment for the electricity they actually produce. The product being bought is megawatts of readiness, not megawatt hours of energy. Regions use them because an energy market alone may not give investors enough confidence to build and keep the capacity that will be needed on the worst day of the year.

The problem they are meant to solve

Electricity markets normally pay generators for what they deliver, hour by hour, at a price set by the most expensive plant needed to meet demand. In most hours that price is modest, because demand is comfortably met.

Now think about a plant that exists for the fifty hardest hours of the year: a peaking unit that sits idle through spring and autumn and runs on the coldest evenings or the hottest afternoons. Its whole revenue has to come from those hours. In theory prices in those hours should rise high enough to cover its annual costs, because supply is genuinely scarce. In practice several things get in the way.

Markets usually have price caps, imposed because regulators are reluctant to let prices reach the levels that pure scarcity would produce. Operators also intervene before prices spike, dispatching reserves or calling on emergency measures, which suppresses the very price signal that would have paid for the plant. And investors face the risk that a mild year produces no scarcity at all, leaving them with no revenue against a fixed cost.

The result is known as the missing money problem: the energy market on its own delivers less revenue than the cost of the capacity the system wants to have. A capacity market supplies the difference explicitly, as a payment for availability.

How they usually work

Designs differ by region, but the shape is fairly consistent.

A reliability standard is set. A regulator or system operator decides how reliable the system should be, often expressed as an expected number of hours or events per year in which supply falls short. This is a policy choice, and it drives everything downstream.

A capacity requirement is calculated. The operator forecasts peak demand for a future year and adds a margin, producing a total quantity of capacity to procure.

Resources are derated. Not every megawatt is equally useful during a stress event. Each resource is credited with the capacity it can be expected to deliver at those times, so a wind farm might be credited with a modest fraction of its nameplate rating, a battery with an amount depending on how many hours it can run, and a thermal plant with a figure reflecting its typical forced outage rate.

An auction is held, usually years ahead. Providers bid the annual payment they need. The operator buys from the cheapest upwards until the requirement is met, and the price of the last accepted bid usually sets the payment for everyone. Running the auction several years in advance is deliberate: it gives time for new plants to be built and for the result to influence whether they are.

Obligations and penalties apply. Winning a contract is a commitment. If the system declares a stress event and a contracted resource is not delivering, it is penalised. That is what turns the payment into something more than a subsidy.

Consumers pay. The total cost is charged to electricity suppliers in proportion to their demand, and reaches households inside the network and policy portion of the bill.

How this differs from the energy market

The two run side by side and pay for different things.

• The energy market pays per megawatt hour delivered, clears daily and intraday, and rewards being cheap to run.
• The capacity market pays per megawatt available, clears once for a whole delivery year, and rewards being reliably there.
• Ancillary services markets sit alongside both, paying for fast response, reserves and voltage support.

A plant can earn from all three. A peaking unit might earn most of its income from capacity, a little from ancillary services and almost nothing from energy, while an efficient gas plant running many hours earns mostly from energy.

The alternatives

Capacity markets are not the only way to keep the lights on, and some large systems deliberately do without one.

Energy only markets rely on prices rising far enough during scarcity to pay for capacity, using a high price cap and sometimes an administrative adder that lifts prices as reserves get tight. The ERCOT grid in Texas is the best known example of this design. It is simpler and avoids forecasting demand years ahead, but it accepts sharper price spikes and more uncertainty for investors.

Strategic reserves keep a block of plant outside the market entirely, to be used only in emergencies, with consumers paying for it directly. This is a narrower, cheaper intervention than a full capacity market, and it is easier to remove later.

Reliability options pay providers for capacity but require them to refund the difference whenever the market price exceeds a strike price, which caps consumer exposure to high prices while still paying for availability.

Long term contracts and central procurement take the question out of the market altogether and place it with a planner or a state owned utility, which is how much of the world still operates.

The main criticisms

Capacity mechanisms attract sustained argument, and the criticisms are worth stating on their own terms.

Paying for what would have been built anyway. Some plants would have been available regardless of the payment, so part of the money buys nothing extra. Supporters reply that paying a clearing price to everyone is what makes an auction competitive.

Propping up old plants. Because existing plants have low remaining fixed costs, they can bid low and win, which can keep ageing fossil units in service for years. Several jurisdictions now attach emissions limits to participation in response.

Forecast error. Requirements are set years ahead from demand forecasts, and forecasts are often wrong. Buying too much wastes money, buying too little defeats the purpose.

Bias against newer resources. Derating rules, minimum sizes and complex qualification processes can disadvantage demand response, storage and aggregated small resources, even where they are permitted in principle.

Complexity and cost to consumers. The rules are intricate enough that few outside the industry can evaluate them, while the cost appears in bills with no visible connection to anything a household recognises.

It treats a symptom. If the real problem is price caps and operator intervention suppressing scarcity prices, critics argue the direct fix is to reform those rather than add a second market on top.

Supporters accept much of this and argue that the alternative is worse: relying on rare, extreme price spikes to finance capacity puts reliability at the mercy of investor confidence, and the cost of getting it wrong is measured in outages rather than money. Both positions are seriously held, and the choice a region makes usually reflects its history and its tolerance for price volatility as much as any technical argument. Our grid zone pages show what each US system actually has running, which is the ground truth these arguments are about.