Clean spark spread versus clean dark spread: worked examples for reading generation margins
Clean spark spread and clean dark spread are common shorthand figures in power market commentary, and for good reason: they represent a gas or coal plant's theoretical margin per megawatt hour after fuel and carbon costs. However, they don't show actual earnings, as they omit factors like operating, maintenance, start-up costs, and technical constraints unique to each plant. Consequently, these figures should be seen as indicators of the relative competitiveness of different technologies rather than precise measures of profitability.
The examples below use deliberately round, hypothetical figures to walk through the mechanics. Any live comparison should be built on current power, fuel and carbon prices rather than the numbers used here.
What the spark spread and dark spread measure
The spark spread is the difference between the power price and the cost of the gas required to generate that power. The dark spread is the same concept applied to coal. Both convert a fuel price into an equivalent electricity cost using the plant's efficiency, or heat rate, since a less efficient plant requires more fuel input to produce the same unit of output.
Spark and dark spreads serve as a quick indicator of a plant's potential profitability before considering additional costs. A positive spread indicates the plant covers its fuel costs at the current power price, while a negative spread shows it does not, based solely on this measure.
Different users rely on these figures for various reasons. Traders monitor the spread as a real-time indicator of relative competitiveness. Analysts and investors use it as an initial check on plant economics before developing a more detailed model. Utilities involved in dispatch and hedging use it to verify if a plant should be generating or holding back. Additionally, it serves as a straightforward entry point for newcomers to power market fundamentals, since the calculation is simple even if the market itself is complex.
Why the clean versions include carbon cost
Clean spark and clean dark spreads take the same logic a step further by also subtracting the cost of the carbon allowances a plant must buy to cover its emissions. That cost depends on two factors: how much the plant emits per unit of output (its emissions intensity) and the prevailing carbon price.
Since coal releases significantly more emissions per unit of output than gas, a set carbon price impacts a coal plant's profit margin much more than it does a gas plant's. This is why clean spreads generally provide a more accurate reflection of relative competitiveness in markets with significant carbon pricing: the raw, non-clean spreads may underestimate how much existing carbon costs influence which technology is likely to operate.
Worked example: clean spark spread for a gas plant
Consider a hypothetical gas plant with 50% thermal efficiency, a gas price of €30 per MWh of thermal input, a carbon price of €70 per tonne of CO2, and a power price of €90 per MWh.
Converting the gas price into a cost per unit of electricity output requires dividing by the plant's efficiency: €30 divided by 0.50 yields a fuel cost of €60 per MWh of power generated.
Carbon cost follows a similar logic. A typical gas emissions factor of around 0.20 tonnes of CO2 per MWh of thermal input, when divided by the 50% efficiency, yields roughly 0.40 tonnes of CO2 per MWh of electricity output. At a €70 carbon price, that equates to a carbon cost of around €28 per MWh.
Subtracting both fuel and carbon costs from the €90 power price leaves a clean spark spread of roughly €2 per MWh, a thin but positive margin under these illustrative assumptions.
It's worth noting how sensitive that figure is to each input. A gas price a few euros higher, or a carbon price a few euros higher, would be enough to turn this hypothetical plant's clean spark spread negative. That sensitivity is part of why the calculation is worth redoing with live inputs rather than relying on a spread calculated even a few days earlier, particularly for a plant already close to breakeven.
Worked example: clean dark spread for a coal plant
Now consider a hypothetical coal plant with a lower efficiency of 38%, a coal price of €12 per MWh of thermal input, and the same €70 carbon price and €90 power price as used above.
Coal's fuel cost per MWh of output is €12 divided by 0.38, or approximately €31.60 per MWh. Coal's emissions factor is considerably higher than gas's, typically around 0.34 tonnes of CO2 per MWh of thermal input. Dividing by the plant's 38% efficiency gives roughly 0.89 tonnes of CO2 per MWh of electricity output. At the same €70 carbon price, that's a carbon cost of around €63 per MWh, more than double the gas plant's carbon cost under these assumptions.
Subtracting fuel and carbon costs from the same €90 power price leaves a clean dark spread of roughly minus €4 per MWh: a negative margin, meaning this hypothetical coal plant wouldn't cover its fuel and carbon costs at all at this power price.
Strip the carbon cost out of that calculation, though, and the picture flips. On fuel cost alone, this hypothetical coal plant's €31.60 per MWh compares favourably with the gas plant's €60 per MWh, which is exactly the point: without a carbon price, or with a low one, coal's fuel cost advantage can be substantial. It's the carbon cost, driven by coal's much higher emissions factor, that erodes and here reverses that advantage entirely. That's the mechanism worth understanding, not just the final spread figure.
Comparing the two and what the comparison signals
Set side by side, at the same €90 power price and €70 carbon price, the two hypothetical plants land in very different places:
- Gas plant: clean spark spread of roughly +€2 per MWh
- Coal plant: clean dark spread of roughly minus €4 per MWh
That's exactly the kind of relative signal these spreads are designed to surface: under these assumptions, gas appears more likely to be dispatched.
Because carbon costs fall much more heavily on coal than on gas, rising carbon prices tend to narrow, or in cases like this one, even reverse, any fuel-price-only cost advantage coal might otherwise hold. Tracking how the relative spread between gas and coal shifts over time, rather than looking at either spread in isolation, is often a more useful signal of where the generation mix and the merit order are heading.
It's worth being clear about what real dispatch decisions still depend on beyond this comparison: must-run constraints, plant availability, start-up costs and a range of other operating costs that clean spreads deliberately leave out. A plant with an attractive clean spread on paper can still sit idle for reasons the spread itself doesn't capture, and one with a thin or negative spread can still run to meet contractual or system obligations.
Clean spreads also directly shape how a broader generation stack is read. Ranking plants by their spreads, rather than by fuel type alone, is essentially the same exercise as building a merit order from short-run marginal costs, and the two are worth considering together rather than in isolation: a shift in relative spreads between gas and coal is, in effect, a shift in where each technology sits in the stack.
Conclusion
Clean spark and clean dark spreads offer a simplified yet genuinely useful way to compare gas and coal margins once fuel and carbon costs are accounted for, and the comparison between them indicates which technology appears more competitive under current market conditions. They remain theoretical margins, not a measure of actual plant profitability, and the worked figures above are illustrative rather than a market call. Any live comparison should be built from current power, fuel and carbon price data rather than assumptions carried over from an earlier one.



.jpeg)




.webp)