Energy Economics: Capacity, Capacity Factor, Marginal Cost, Merit Order
The Two Quantities
Chapter 02 introduced power (watts) and energy (watt-hours). In economics, these map to two different things that get built and sold:
Capacity what the generator can produce at peak (MW)
Energy what the generator actually produces over time (MWh)
A 100 MW plant has 100 MW of capacity. Its energy output depends on how often it runs.
Markets pay for both, but differently:
- Capacity payments: paid for being available, whether or not you run
- Energy payments: paid for actually producing energy when called
Different markets structure these differently. Some have explicit capacity markets (paying generators to keep old plants available for reliability). Some rely on energy markets alone and let scarcity pricing handle reliability.
Understanding the distinction saves you from a lot of confused arguments.
Capacity Factor
Capacity factor is the most useful concept in energy economics literacy.
Capacity factor = actual energy produced / (capacity × hours in period)
If a 100 MW plant produces 400,000 MWh in a year (8,760 hours × 100 MW = 876,000 MWh at theoretical max), its capacity factor is 400,000 / 876,000 = 46%.
Different sources have very different capacity factors:
Nuclear 85-95% (runs near-continuously)
Geothermal 70-95%
Coal 40-70% (declining in markets with lots of renewables)
Natural gas CC 30-60% (varies, flexible role)
Hydro 30-60% (depends on water availability)
Offshore wind 40-60%
Onshore wind 30-45%
Utility solar PV 15-27% (no sun at night; varies with latitude and weather)
Gas peakers 5-20% (only run during peak demand)
Why it matters
Capacity factor determines how much capacity you need to provide a given amount of energy.
To produce 1 TWh per year from various sources (roughly):
Nuclear (90% CF) 130 MW of capacity
Wind (35% CF) 330 MW of capacity
Solar (22% CF) 520 MW of capacity
So a 1 GW nuclear plant produces about as much annual energy as 2.6 GW of wind or 4.1 GW of solar. This isn't a criticism of wind or solar; their capacity is just cheaper per MW. It's an accounting distinction.
Changes over time
Capacity factors aren't fixed. They change because of:
- Technology: newer wind turbines are taller with bigger blades, reaching steadier winds. Modern wind capacity factors exceed older ones
- Market conditions: coal's CF has fallen as cheaper solar and wind displace it during the day
- Grid constraints: a solar plant that has to curtail (shut off) because the grid can't absorb its output has a lower realised CF
Headline capacity factors are often year-old averages; the marginal new plant may differ.
Levelized Cost of Energy (LCOE)
LCOE is the total lifetime cost of generation divided by the total lifetime energy produced. It's the standard way to compare sources.
LCOE = (capital cost + lifetime operating costs) / lifetime energy produced
Expressed in $ per MWh (or € per MWh, etc.).
Recent approximate values (2024):
Utility solar PV $30-50/MWh
Onshore wind $30-50/MWh
Offshore wind $60-90/MWh
Natural gas CC $40-75/MWh (fuel-cost dependent)
Coal $70-150/MWh
Nuclear new $80-180/MWh (very project-dependent)
Existing hydro $10-30/MWh
Existing nuclear $30-40/MWh
What LCOE misses
LCOE is imperfect:
- Ignores timing: energy produced when grid doesn't need it is less valuable
- Ignores firmness: intermittent sources need backup; backup costs aren't in LCOE
- Ignores transmission: a cheap source far from load needs lines to deliver
- Ignores grid services: inertia, frequency response, voltage support; some sources provide these, some don't
- Subsidies, taxes, and externalities vary
For new build decisions, LCOE is a reasonable starting point; system-level analysis is needed for full picture. Chapter 09 on the transition goes into more of this.
Marginal Cost and Merit Order
In electricity markets, generators are ranked by their short-run marginal cost of production (fuel and variable O&M). This is the merit order:
Lowest marginal cost first:
1. Solar PV, wind marginal cost ≈ $0
2. Nuclear marginal cost ≈ $5-15/MWh
3. Coal marginal cost ≈ $20-40/MWh
4. Combined-cycle gas marginal cost ≈ $30-60/MWh
5. Simple-cycle gas peakers marginal cost ≈ $60-150/MWh
The grid operator dispatches generators in order of marginal cost until demand is met. The most expensive dispatched generator sets the market price for all generators (marginal-cost pricing).
The intuition
If demand is low, only cheap generators run. Price is low. If demand is high, expensive peakers have to run, and the price is high (because all generators get paid the price of the most expensive one needed).
Renewables have near-zero marginal cost, so they're always dispatched first when available. They push more expensive generation off the margin, reducing prices during sunny/windy periods.
The merit-order effect
A real consequence: more renewables, lower average prices, and lower revenues for other generators. Good for consumers; bad for other generators; creates "missing money" problem for keeping firm capacity available. Markets are evolving to handle this (capacity markets, ancillary services markets, storage participation).
Negative Prices
In markets with lots of renewables, electricity prices sometimes go negative: generators pay to produce.
Why this happens:
- Generators receive subsidies per MWh produced (even at negative prices, they make money)
- Shutting down and restarting (thermal plants) is expensive; they'd rather keep running at a loss briefly
- Storage and demand response haven't caught up
Negative prices are signals. They say: too much supply, not enough flexible demand or storage. Over time, markets develop responses (more batteries, more industrial demand flexibility).
Negative prices are most common in windy regions on low-demand days: California with midday solar, Germany with windy nights, Texas with wind.
Capacity Markets
Some markets explicitly pay generators for being available, separate from energy payments. The idea: energy markets alone may not pay enough to keep firm capacity (nuclear, coal, gas) online as renewables grow. Capacity markets fill the gap.
Arrangements vary:
- PJM (mid-Atlantic US): formal capacity market, three-year forward
- UK: national capacity market auctions
- Texas (ERCOT): no capacity market; relies on scarcity pricing
- Many places: utility or regulator-run resource adequacy processes
Capacity markets are controversial; critics say they over-pay incumbent fossil generators. Defenders say they ensure reliability without blackouts.
The Cost of Intermittency
A much-argued topic. Wind and solar are cheap on LCOE. System cost at high penetration is higher because:
- Sometimes neither produces; you need backup (batteries, gas, hydro, nuclear)
- Transmission needs expansion to move power from windy/sunny regions to demand centres
- Ancillary services (voltage, frequency) need provision from other sources
The question isn't "can renewables work" (they can; Denmark, Germany, California demonstrate) but "what's the total system cost at X% penetration". Estimates:
- Up to ~50% renewable: modest extra cost above LCOE; mostly integration and transmission
- 50-80% renewable: larger extra costs; storage and flexibility become significant
- 80-100% renewable: costs depend heavily on storage, long-duration storage, transmission, demand flexibility
Honest analyses land somewhere around "deep decarbonisation is achievable and will cost more than just building the cheapest generation, but less than doing nothing about climate." Specific numbers vary by model and assumption.
Natural Monopolies and Regulation
Some parts of the energy system are natural monopolies: one large provider is cheaper than multiple competing ones.
- Transmission and distribution: duplicate wires in the same street would be wasteful
- Large hydropower: site-specific; scale matters
These are typically regulated: either publicly owned or private but with returns capped by regulators.
Other parts are competitive:
- Generation (in liberalised markets)
- Retail (in some markets; consumers can choose their electricity provider)
- Energy services (energy efficiency, demand response, distributed solar)
The mix of regulated and competitive parts varies by country and state. It affects who pays for what and who profits from what.
Vertical Integration
Historically, utilities were vertically integrated: they owned generation, transmission, distribution, and retailed to customers. This model is still dominant in many places.
Liberalisation (1980s-2000s in many countries) unbundled these:
- Generation becomes competitive
- Transmission and distribution stay regulated
- Retail can be competitive
The result: more actors, more market signals, sometimes better outcomes, sometimes coordination problems.
The current trend in some places is partial re-integration or at least more coordinated planning, especially as transmission expansion becomes a bottleneck for the transition.
Value of Firm Capacity
A subtle point: not all energy is equally valuable.
- Solar at noon on a sunny day: low value (abundant supply)
- Gas generation at 6 PM on a hot summer day: high value (peak demand)
- Baseload nuclear running 24/7: high average value (always available)
Markets that only pay energy prices (no capacity value) can underpay firm sources. This is one reason nuclear struggles economically in some liberalised markets despite its reliability.
The "value of firm capacity" argument says we should pay more for generation that's there when we need it, not just for generation that's cheap on average. Capacity markets and reliability tariffs try to do this.
Why Energy News Is Confusing
Energy arguments usually mix:
- LCOE (averaged lifetime cost)
- Wholesale market prices (short-term)
- Retail rates (end-user price including everything)
- Capacity markets (availability payments)
- Externalities (climate, air pollution, land use)
Different frames produce different conclusions. A source that wins on LCOE may lose on system cost. A source with high LCOE may win on value per MWh. A cheap source at the wholesale level may cost more at retail because of distribution charges.
Reading energy news well means noticing which frame the source is using and whether it's complete.
Common Pitfalls
"LCOE says solar is cheapest; we should build only solar." System cost matters. LCOE is necessary but not sufficient
"Gas is cheap." Gas's marginal cost depends on fuel price, which varies enormously. Gas LCOE in 2022 was much worse than in 2019; 2024 prices are somewhere in between
"Nuclear is too expensive." New-build nuclear is expensive and slow in most places. Existing nuclear operating costs are low. These are different questions
"Capacity factor should be 100%." It can't; weather, maintenance, demand limit every source. Target isn't 100%; it's the right mix of sources with complementary patterns
"Negative prices prove we have too much renewable." Or they prove we have too little flexibility. Same fact, different conclusion. Build storage, build transmission, electrify more demand
Next Steps
Continue to 09-the-energy-transition.md for how the current system is changing.