Why the fuel hedge belongs on the dispatch, not on the availability, and what the surplus gas does to the year and to the margin account.
The gas hedge of a gas-fired plant is expected generation times the heat rate, not capacity times available hours. An illustrative 620 MW plant with a heat rate of 7.2 MMBtu per MWh that runs 3,900 hours burns 17,409,600 MMBtu. Hedged on its 8,000 available hours it buys 35,712,000, and is long 18,302,400 MMBtu of gas it will never burn.
Worked in full in Energy Trading by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
Most generators size the power hedge first and copy the convention to the fuel. If the power block is sold on availability, the gas block is bought on availability, and the position report shows both legs "fully hedged". The power leg of that mistake, a hidden short in the hours the plant does not run, is worked in how to calculate the hedge ratio of a power plant. This page counts the other leg, which usually sits on a different clearer and in a different commodity, and so is easier still to miss.
| Input | Value |
|---|---|
| Capacity | 620 MW |
| Heat rate | 7.2 MMBtu/MWh |
| Gas price, hedged flat for the year | 8.60 per MMBtu |
| Variable operating cost | 3.40 per MWh |
| Available hours (760-hour outage in the cheapest hours) | 8,000 |
| Running hours, power above marginal cost | 3,900 |
| Gas move tested, power curve unchanged | ±1.00 per MMBtu |
The running hours come from the year's price duration curve read against the plant's marginal cost, the method set out in how to build a price duration curve.
=MW*RunHours*HeatRate.The right hedge is 48.75 per cent of the availability hedge: the plant's output over its availability. Bought at 8.60, the gas on the right hedge is worth 149,722,560; the availability hedge commits 307,123,200, more than twice the fuel the plant needs.
The idle hours therefore hold two positions, not one: short 2,542,000 MWh of power and long the 18,302,400 MMBtu that would have made it. Together that is a short spark spread on 2,542,000 MWh. A report that shows the power short alone is showing half the trade.
A 1.00 fall in gas loses 18,302,400 on the surplus at unchanged dispatch. The plant does not stay unchanged: at 7.60 gas its marginal cost drops to 58.12, the 2,400 hours priced at 63.00 come into the money, and running them earns back 7,261,440. The year still ends 11,040,960 worse.
Each row moves gas by 1.00 with the power curve unchanged and lets the plant re-dispatch on the new gas price. The change is against the 60,792,240 gross margin at 8.60.
| Hedge sized on | Gas hedged (MMBtu) | Long gas (MMBtu) | Gas −1.00 | Gas +1.00 |
|---|---|---|---|---|
| Capacity, 8,760 h | 39,104,640 | 21,695,040 | −14,433,600 | +22,307,600 |
| Availability, 8,000 h | 35,712,000 | 18,302,400 | −11,040,960 | +18,914,960 |
| 75% of availability | 26,784,000 | 9,374,400 | −2,112,960 | +9,986,960 |
| Expected dispatch, 3,900 h | 17,409,600 | 0 | +7,261,440 | +612,560 |
| Unhedged | 0 | −17,409,600 | +24,671,040 | −16,797,040 |
Only the dispatch-sized hedge gains in both directions. That is the plant's switch showing through: with its expected fuel locked, a fall lets it run more hours and a rise lets it stop the 72.00 band, which no longer pays at a 72.52 marginal cost, saving 612,560. Every row above it adds a long gas position on top of the option; the unhedged row is short the fuel instead.
The dispatch itself is fragile. At 8.2778 the 63.00 band comes into the money, a fall of only 0.3222; at 9.5278 the 72.00 band drops out, a rise of 0.9278. Outside that range the running hours, and therefore the right hedge volume, change. The gas and power prices that set those thresholds are linked through the implied heat rate.
Once the gas is bought, it is tempting to dispatch on 8.60 for the rest of the year. But surplus gas can be sold back, so burning one more MMBtu costs what it would have fetched, not what it cost. At 7.60 the plant should dispatch on 58.12, not 65.32. A desk that keeps the hedged price in the dispatch model leaves the 7,261,440 of extra running hours on the table and books the full 18,302,400 loss.
The second mistake is cash. If the gas hedge is cleared, a 1.00 fall calls 35,712,000 of variation margin on the whole availability volume the next morning, 22.6 per cent of an illustrative 158,000,000 of cash and undrawn facility, while the dispatch gain arrives hour by hour. The arithmetic of the call is the same as for power, set out in how to calculate variation margin on power futures.
The gas leg, the two thresholds and the first-morning call are worked as case one in the free workbooks for this case, with the share hedged and both gas moves as inputs.
Its heat rate, in MMBtu per MWh. A combined-cycle unit at an illustrative 7.2 burns 7.2 MMBtu for every MWh it sends out, so 620 MW at full load burns 4,464 MMBtu an hour. At 8.60 per MMBtu the fuel costs 61.92 per MWh, and with 3.40 of variable operating cost the plant's marginal cost is 65.32.
Because it bought more gas than it burns. Hedged on availability, the illustrative plant is long 18,302,400 MMBtu. A 1.00 fall loses 18.30 million on that surplus; cheaper fuel brings 2,400 more hours into the money and earns back 7.26 million, so the year ends 11.04 million worse. Hedged on dispatch, the same fall gains 7.26 million.
Yes, on the same volume. Sell the MWh the plant expects to generate and buy that volume times the heat rate in gas, plus the carbon allowances. Sized on availability, the idle hours carry a short power position of 2,542,000 MWh and a long gas position of 18,302,400 MMBtu: a short spark spread the position report usually shows only half of.
This article is one calculation from Energy Trading. The book takes the same case from first principles to the decision, chapter by chapter, and every figure it prints is a live formula in the free companion workbooks.
Get the book on Amazon →Free companion files
Also on Amazon UK · Amazon Germany · Amazon France · Amazon Canada
Reading guide: infrastructure, data centres and energy → · All 1340 articles →
If this book helped, or didn’t, a few lines on Amazon are worth more than they look: they are what the next reader goes on. Write a review. The workbook stays free either way.