Articles

How do you calculate mark-to-market on a power forward?

A short baseload strip valued month by month, with the hour counts, the discounting and the two shortcuts that move the answer.

The mark-to-market of a power forward is, for each delivery period, MW × hours × (contract price − current forward price), discounted to the payment date, for a seller; reverse the sign for a buyer. An illustrative 100 MW baseload sale of Q1 2027 at 92.00, against monthly forwards of 98.40, 95.10 and 84.60, is worth −135,477 today: −134,660 before discounting.

Worked in full in Energy Trading by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →

Three things make a power MTM different from the textbook forward. The volume is not a fixed quantity but megawatts times the hours in the period, and the hours differ from month to month. A quarter or a calendar is a strip of months that do not trade at the same price. And an OTC forward is paid after delivery, so the value has to be discounted, while a cleared future is settled daily and is not.

The assumptions

An illustrative short baseload strip, valued on 12 October 2026. Prices per MWh.
InputJanuary 2027February 2027March 2027
Volume sold100 MW100 MW100 MW
Hours (central European time)744672743
Contract price92.0092.0092.00
Current forward price98.4095.1084.60
Payment date20 Feb 202720 Mar 202720 Apr 2027
Discount rate, annual (illustrative)3.50%3.50%3.50%

March has 743 hours, not 744, because the clocks go forward on its last Sunday. The quarter was sold as one product at 92.00; it is valued as three months because the market now prices them differently.

The calculation, step by step

Mark-to-market of the short strip, month by month.
MonthMWhContract less marketUndiscountedDays to paymentDiscount factorMTM
January74,400−6.40−476,1601310.9877−470,317
February67,200−3.10−208,3201590.9851−205,221
March74,3007.40549,8201900.9823540,062
Q1 2027215,900−134,660−135,477

The result

The strip is worth −135,477. Expressed as a quarterly price, the market is the hour-weighted average of the months, 92.62, and the undiscounted MTM is simply 215,900 × (92.00 − 92.62) = −134,660. The quarter is close to flat, but the months are not: the position is losing 684,480 on January and February and gaining 549,820 on March. A report that shows only the quarterly figure hides a winter exposure that a cold January would reveal.

Discounting made the value worse here, not better, by 817. That is not an error: the gain sits in the latest month and is discounted the most, while the losses come earlier. With mixed signs across the strip the direction of the discounting effect is not obvious and has to be computed.

What if: the curve moves

Mark-to-market after a parallel shift of the three months.
Shift per MWhMTMChange
−101,991,2112,126,688
−5927,8671,063,344
0−135,4770
+5−1,198,821−1,063,344
+10−2,262,165−2,126,688

Each 1.00 per MWh is worth 212,669, the discounted volume. For a seller, rising prices are losses. If the same position were a cleared future, the change would not wait for delivery: it would be paid in variation margin the next morning, as worked in how to calculate variation margin on power futures.

From mark-to-market to realised P&L

The MTM is an unrealised number that changes every day the curve moves. It turns into cash month by month. When January delivers, the seller receives 92.00 on 74,400 MWh from the counterparty and, if the strip hedges nothing, buys the power back in the spot market at whatever January's hours actually average. The realised result for January is then the contract price less the spot average, not less the 98.40 forward. The difference between the two is the part of the MTM that the market took back, or added, between today and delivery.

That is why the valuation date, the price source and the cut-off time belong on the same line as the number. A forward curve taken from the broker screen at noon and one taken from the exchange settlement at close can differ by enough to move a strip of this size by tens of thousands, and two desks that disagree about the MTM are often only disagreeing about which curve they used.

The common mistakes

Averaging the months without their hours. The simple average of 98.40, 95.10 and 84.60 is 92.70, not 92.62. Applied to the whole quarter it gives −151,130, overstating the loss by 16,470, because February, the shortest month, carries a high price.

Standard months of 720 hours. Using 720 hours a month gives −151,200, 16,540 out. Count the hours of each product from its calendar, with the clock changes; peak products have their own count, as in the off-peak price calculation.

Reporting the hedge MTM alone. If this strip hedges a plant or a customer book, its MTM is half of a position. A gain on the forward is only meaningful beside the change in value of what it hedges.

Takeaway

The hours, the curve and the cash that a mark-to-market turns into are modelled in the free workbook for this case.

Questions readers ask

Why value a quarterly power forward month by month?

Because the months have different prices and different hours. January has 744 hours, February 672 and March 743 in central Europe, where the clocks change. Weighting the three illustrative prices by hours gives a market of 92.62; a simple average gives 92.70, and applying it to the whole quarter overstates the loss by 16,470.

Does discounting matter for a power forward MTM?

Less than the hours, but it is not optional for an OTC forward paid after delivery. On the Q1 strip, discounting at 3.50 per cent to payment dates 131 to 190 days away moves the value from -134,660 to -135,477. Exchange futures are settled daily through variation margin and are not discounted.

How sensitive is the MTM to a parallel price move?

By the discounted volume: 212,669 per 1.00 per MWh on 215,900 MWh. A 5.00 rise takes the short strip from -135,477 to -1,198,821; a 10.00 fall takes it to a gain of 1,991,211. For a short position, rising prices are losses.

Read the whole case

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

Also on this site

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.