8. Electricity

Teaching rewrite of electricity markets: generation, the supply chain, market design and regulation, price drivers, spark/dark spreads, trading conventions, locational pricing, and power derivatives. Original wording; worked numbers recomputed. Prices are teaching values. Not investment advice.

Introduction

When England lost to Germany in a 1990 World Cup semi-final, UK demand jumped by about 2,800 MW in minutes — over a million kettles switched on at once. Electricity is the commodity that can’t be stored at scale: supply must match demand every second, or the lights go out. That single fact shapes everything — from how prices are set to why they can spike to thousands or fall below zero.

This lesson covers:

• How electricity is made and measured (MW vs MWh, thermal efficiency)
• The supply chain: generation, transmission, distribution, supply
• Market design: pools vs bilateral markets; EU, UK, and US regulation
• Price drivers: load, weather, capacity, fuel, renewables, negative prices
• Spark and dark spreads, heat rates
• Trading: load shapes, contract volumes, merit order, imbalances, UK EFA calendar, US locational marginal pricing and FTRs
• Derivatives: forwards, swaps, CfDs, swaptions, spread options, embedded optionality, ratio swaps, heat rate options

Lesson 7

Lesson 7 (natural gas) is reserved for the next chapter of the source; this lesson uses gas only as a generating fuel.

---

8.1 What electricity is

Electrons moving between atoms create a current. Electricity is a secondary energy source: it must be transformed from a primary source and transferred to where it’s used, then transformed again into heat, light, sound, or motion.

flowchart LR
  S[Primary source<br/>fuel, wind, water, sun] -- transfer --> P[Power station] -- transformation --> E[Electrical energy]
  E -- transfer --> U[Homes, factories] -- transformation --> O[Heat, light, motion]

• Thermal plants (gas, coal, nuclear, oil): boil water → steam → turbine → generator (electromagnetic induction, like a bicycle dynamo).
• Wind, wave, hydro, tidal: movement drives the generator directly.
• Solar: photovoltaic panels convert light directly; solar-thermal plants make steam.

Fuel mix

Fuel	USA 2003	USA 2018	UK 2004	UK 2016
Natural gas	16%	35%	40%	42%
Coal	51%	27%	33%	9%
Nuclear	20%	19%	19%	21%
Renewables	7%	17%	1%	25%

Shale gas pushed US generation from coal to gas; the UK has moved to gas and renewables — and closed its last coal plant in 2024. Norway is almost 100% hydro: exports in wet years, imports in dry ones.

Thermal efficiency

Electrical output ÷ energy input. Combined-cycle gas ≈ 47–49%; nuclear ≈ 38%; coal ≈ 36%. The rest leaves as heat (the plumes from cooling towers).

Plant types

Plant	How	Role
CHP (combined heat & power)	Gas-fired; captures waste steam/hot water for heating	Efficient where heat is needed
CCGT (combined cycle gas turbine)	Gas turbine + steam turbine from exhaust heat	Efficient baseload/mid-merit
OCGT (open cycle)	Gas turbine only	Less efficient but ramps fast — peaking/reserve

Measuring electricity

Unit	Meaning
W, kW, MW, GW	Power — an instantaneous rate (a 100 W bulb draws 100 W at any moment)
Wh, kWh, MWh, GWh, TWh	Energy — power × time (ten 100 W bulbs for 1 hour = 1 kWh)
Volt	“Pressure” pushing electrons

Buying a 100 MW plant = capacity; using 100 MWh = 100 MW for one hour. US homes used about 1,462 bn kWh in 2018.

---

8.2 The supply chain

Link	Role
Generators	Build and run power stations — portfolios or independents
Transmission system operator (TSO)	High-voltage grid: schedules flows, balances supply and demand in real time, manages reserves and emergencies, plans new lines
Distribution (DSO)	Substations step down voltage; local networks to homes
Suppliers	Sell to and bill end users
End users	Domestic, commercial, industrial

In Great Britain, National Grid owns transmission, while system balancing moved to the publicly owned National Energy System Operator (NESO) in 2024.

UK household bill (source era): wholesale ≈ 32%, network ≈ 23%, environmental/social obligations ≈ 20%, operating ≈ 17%, VAT ≈ 5%, supplier margin < 1%.

Rooftop solar, batteries, smart homes, and EVs are reshaping this chain.

---

8.3 Market structure and regulation

Pools vs bilateral markets

Pool: generators bid into a central auction run by the system operator; everyone is paid the highest accepted bid — the system marginal price.

Worked example: demand 1,000 MW; four generators each offer 250 MW at 60, 65, 70, 75.

Demand	In merit	Clearing price
1,000 MW	I, II, III, IV	75
750 MW	I, II, III (IV out of merit)	70

Bilateral (e.g. Great Britain): buyers and sellers contract directly (OTC or on exchange), notify the system operator of net positions by gate closure, and sellers self-dispatch; the operator balances the residual.

Europe

• Regulations apply directly; directives set goals each state implements.
• 1996 First Electricity Directive: large customers choose suppliers; third-party network access; unbundling of networks (regulated) from generation and supply (competitive). 2003 and 2009 packages extended this toward an integrated, price-convergent market.
• ACER (2011) coordinates regulators. REMIT bans market abuse (manipulation, insider trading), requires disclosure of inside information and reporting of suspicious trades.

Great Britain

NETA (2001) introduced bilateral trading, a balancing mechanism, a single notional delivery point, short-term power exchanges, OTC derivatives, and imbalance settlement; extended to Scotland as BETTA (2005).

United States

A patchwork: competitive markets in some regions, vertically integrated utilities in others.

• Suppliers: investor-owned utilities (~51% of sales), power marketers (~22%), public utilities (~14%), co-ops (~11%), federal agencies (e.g. Tennessee Valley Authority).
• FERC regulates wholesale markets; states regulate retail.
• ISOs/RTOs run grids and markets: CAISO, ERCOT, MISO, ISO-NE, NYISO, PJM, SPP.
• NERC sets reliability standards (adequacy and security) via regional entities.

---

Checkpoint 1: basics and market design

Quiz

Checkpoint — sections 8.1 to 8.3

1 / 5

Twenty 50 W bulbs run for 3 hours. Energy used?

Pick an answer to see if you are right.

---

8.4 Price drivers

Because power can’t be stored, today’s high price says little about tomorrow’s. Shortfalls cause blackouts (or brownouts — dimming).

Load

• Baseload: the steady minimum, served by baseload plants.
• Peak: the highest demand, served by peakers; intermediate in between.
• Load shape depends on user type, time of day, season, region (US peaks in summer except the north and Florida).

Demand drivers

Driver	Example
Economic activity	More output → more power
Weather	Temperature, humidity, cloud, rain; heating/cooling degree days relative to ~21 °C (70 °F)
Time and calendar	Peak (working hours) vs off-peak, weekends, seasons (winter Oct–Mar, summer Apr–Sep)
User type	Households weigh weekends; industry weekdays
Special events	Britain’s “TV pick-up” at the end of matches and soaps
Efficiency	Better appliances reduce demand
EVs	See below

EV demand model: EVs × km driven × kWh/km + grid losses (7–8%). Example: 1 m EVs × 12,000 km × 0.18 kWh/km = 2.16 TWh; +7.5% losses → ≈ 2.32 TWh to generate. Unmanaged evening charging collides with peak demand — hence “managed” overnight charging.

Supply drivers

• Capacity: plant types, outages, maintenance, transmission limits.
• Reserve margin = (installed capacity − peak demand) / installed capacity. Typical targets ~15–17% (US); below ~10–13%, scarcity premiums climb steeply.
• Interconnectors (GB links to France, Netherlands, Ireland and more) move power toward higher prices.
• Marginal fuel: cheapest plants run first (wind, hydro, nuclear), then gas or coal; the marginal plant’s fuel sets the price. Gas prices matter most when gas is marginal.
• Renewables and bidding behaviour.

Spot vs forward

• Short-dated: plant availability, fuel prices, rain, wind, interconnectors, temperature, carbon.
• Long-dated: forward fuel, new/retired capacity, reservoirs, weather trends, carbon, growth.
• Forwards can’t be cash-and-carried (no storage) → forward ≈ expected spot ± risk premium (usually a premium); curves show seasonality (winter dearer in Europe).

Negative prices

Occur when demand is low, wind/solar abundant, and inflexible plants (nuclear, coal) prefer paying to run over costly shutdown/restart; also from transmission constraints and subsidies (generators still earn the subsidy). Examples: Germany intraday −EUR 500/MWh (2012); US Northwest hydro, California solar, PJM/Texas night wind; Great Britain in 2019 (as low as −GBP 70/MWh). They have become more frequent as renewables grow.

Price spikes too

In February 2021, Winter Storm Uri drove ERCOT (Texas) prices to the USD 9,000/MWh cap for days — the mirror image of negative prices.

---

8.5 Spark and dark spreads, heat rates

Spark spread = power − gas cost; dark spread = power − coal cost — the generator’s gross margin. Adjusted for carbon = clean spread.

Raw spread = power − fuel cost per MWh ÷ efficiency Clean spread = raw spread − (t CO₂ per MWh × carbon price)

Conventions: gas efficiency 49.13%, CO₂ factor 0.42 t/MWh; coal efficiency 36%, 0.85 t/MWh.

Worked example: clean spark spread (GB)

Power GBP 34.50/MWh; gas 36p/therm; carbon GBP 10.75/t.

Step	Calculation	GBP/MWh
Gas in GBP/therm	36p ÷ 100	0.36
Gas per MWh (thermal)	0.36 ÷ 0.0293071	12.28
Fuel per MWh of power	12.28 ÷ 0.4913	25.00
Raw spark	34.50 − 25.00	9.50
Carbon cost	0.42 × 10.75	4.52
Clean spark	9.50 − 4.52	4.98

Worked example: clean dark spread

Coal USD 71.15/short ton, GBP 1 = USD 1.90 → GBP 37.45; at 7 MWh/t → GBP 5.35/MWh.

Raw dark = 34.50 − 5.35 / 0.36 = 19.64; clean dark = 19.64 − 0.85 × 10.75 = 10.50.

Coal burns dirtier, so carbon prices hit coal harder (bigger effect in coal-heavy Germany than in Britain). Generators sell the spread (sell power, buy fuel forward) to lock in margin — mostly OTC, since power futures are less liquid.

Heat rates

Heat rate = fuel input per unit of power (MMBtu/MWh) = the inverse of efficiency (1 kWh = 3,412 Btu):

Efficiency	Heat rate (Btu/kWh)
100%	3,412
47% (gas)	≈ 7,260
36% (coal)	≈ 9,478

• Technological heat rate = input ÷ output (e.g. 3,500 MMBtu ÷ 500 MW = 7 MMBtu/MWh).
• Market (implied) heat rate = power price ÷ gas price.
• Spark spread with heat rates = power − gas × heat rate.

Example: power USD 39.75/MWh, gas USD 2.89/MMBtu → market heat rate ≈ 13.75; spark at heat rate 7 = 39.75 − 2.89 × 7 = USD 19.52/MWh.

Correction

The source quotes 19.49; 39.75 − 20.23 = 19.52.

---

Checkpoint 2: drivers and spreads

Quiz

Checkpoint — sections 8.4 and 8.5

1 / 6

Installed capacity 60 GW; peak demand 52 GW. Reserve margin?

Pick an answer to see if you are right.

---

8.6 Trading electricity

Participants: utilities, generators, suppliers, industrial users, banks, hedge funds, brokers, trading houses. Largest European markets: Germany, France, GB, Italy, Spain.

Participant	Activities
Producers	Long-term supply contracts; exchange trades (e.g. EPEX); OTC hedges; futures via exchange members
Suppliers	Contracts with producers and end users; exchange balancing; risk solutions
Consumers	Fixed-price tariffs (backed by suppliers’ swaps); large users hedge directly
Intermediaries	Views, cross-market arbitrage, market making, brokerage, structured products

Load shapes

Shape	Typical definition
Baseload	24 hours a day
Peak	e.g. 07:00–19:00 weekdays (GB)
Off-peak	All other hours
Overnight	e.g. 23:00–07:00

Contract volumes and values

Traded in MW (rate), settled in MWh (energy). GB has 48 half-hour settlement periods per day.

Trade	Total volume	Daily	Per half-hour	Value
Winter peak 20 MW (12 h × 5 d × 26 wk)	20 × 12 × 5 × 26 = 31,200 MWh	240	10	at GBP 14 → GBP 436,800
Summer base 100 MW (24 h × 7 d × 26 wk)	100 × 24 × 7 × 26 = 436,800 MWh	2,400	50	at GBP 15 → GBP 6,552,000

Price formation: the merit order

Price tends toward short-run marginal cost. A producer stacks plants by operating cost:

Fuel	Capacity (MW)	Cost (USD/MWh)
Nuclear	360	70
Renewable	120	80
Natural gas	350	110
Coal	130	120
Diesel	40	150

Demand 500 MW = nuclear 360 + renewable 120 + gas 20 → price set by gas: USD 110/MWh. Spare capacity 500 MW. In reality the stack curves sharply upward near full capacity (“the devil’s elbow”), where scarcity premiums appear.

Optimising production and timing

Producers forecast demand (contracts, weather, events, history) and choose to generate, buy from others, or hedge forward — refining positions via annual, seasonal, monthly, and shape-specific contracts, then day-ahead and intraday auctions. Traded volume far exceeds physical output (e.g. GB 2018: ~333 TWh generated, ~884 TWh traded OTC; Germany ~540 TWh vs ~5,330 TWh).

Imbalances: after gate closure (e.g. one hour ahead), parties submit final physical notifications; the operator resolves deviations with bids/offers and charges those who were out of balance. Causes include trading mismatches, generation over/under-delivery, demand forecast errors, plant failures, scheduling errors, and transmission losses (~10% rule of thumb).

Timing mismatch: generators want long-dated sales (to service plant loans); consumers buy shorter-dated — intermediaries bridge the gap.

GB conventions: the EFA calendar

Pre-2014 GB trading used the Electricity Forward Agreement calendar: seasons (summer EFA Apr–Sep, winter Oct–Mar), quarters, EFA months, numbered weeks, EFA days running 23:00 to 23:00, and six 4-hour blocks:

Block	1	2	3	4	5	6
Hours	23–03	03–07	07–11	11–15	15–19	19–23

“WD3” = weekday 07:00–11:00. Gregorian contracts were added in 2014, but block indices persist.

US: locational marginal pricing (LMP)

US power is regional — main markets CAISO, MISO, ISO-NE, NYISO, PJM, SPP, ERCOT, with many trading hubs (e.g. PJM West, SP-15, NP-15). When transmission is congested, prices differ by node.

Worked example: node A demand 50 MWh, generator cost USD 10; node B demand 100 MWh, cost USD 15; line A→B limited to 80 MWh.

• Dispatch: A produces 50 + 80 = 130 MWh; B produces 20 MWh.
• LMPs: A 10, B 15 (all B load pays 15).

	Calculation	USD
Collected	50 × 10 + 100 × 15	2,000
Paid to generators	130 × 10 + 20 × 15	1,600
Congestion revenue		400

Correction

The source’s totals (1,500 collected, 1,300 paid) don’t match its own lines; the figures above add up.

Financial transmission rights (FTRs): auctioned contracts paying the price difference between two nodes in the stated direction (and charging if it reverses). Generator A selling into B with an A→B FTR receives USD 5/MWh here, restoring the B price.

---

Checkpoint 3: trading

Quiz

Checkpoint — section 8.6

1 / 5

A 15 MW winter peak contract (12 h, 5 days, 26 weeks). Total MWh?

Pick an answer to see if you are right.

---

8.7 Electricity derivatives

Forwards

Priced from expected supply/demand (no arbitrage to spot). Uses: price certainty, avoiding volatile spot, securing supply, separating supply from risk (financial forwards), speculation, financing requirements.

Terms: trade date, buyer, seller, governing contract, price, volume, load shape, delivery schedule and point.

• French example: 50 MW for 12 hours (08:00–20:00) on one day → 600 MWh × EUR 60 = EUR 36,000.
• GB example: 50 MW baseload for one day → 1,200 MWh at GBP 32.
• Floating (index) forwards: price set at delivery from an index.

Cross-border trade: France EUR 55, Germany EUR 60, interconnector capacity EUR 3/MWh, 100 MW annual baseload → (60 − 55 − 3) × 24 × 365 × 100 = EUR 1,752,000.

Swaps

Fixed vs floating, cash-settled — e.g. against an index of day-ahead baseload prices.

Example: a hedge fund pays fixed GBP 45.50 on 30 MW winter baseload; a 31-day month averages GBP 40.00:

• Pays 30 × 45.50 × 24 × 31 = 1,015,560; receives 30 × 40 × 744 = 892,800 → net pays GBP 122,760.

The fixed price ≈ a weighted average of the monthly forwards — the fixed payer must “beat the forward”. Suppliers pay fixed to fix costs; generators receive fixed to fix revenue. Continental swaps: monthly calculation periods, baseload/peak/off-peak, payment 5 or 20 days in arrears, hourly floating prices (e.g. 92-day baseload at 6 MWh/h = 13,248 MWh).

Contracts for difference (GB low-carbon CfDs)

The government-owned Low Carbon Contracts Company pays or receives the difference between a strike and a reference price for 15 years.

Onshore wind strike GBP 79.23:

Market price	Generator sells at	CfD	Net
70.00	70.00	+9.23	79.23
85.00	85.00	−5.77	79.23

(Strike prices are set in competitive allocation rounds and indexed to inflation.)

Swaptions and collars

A manufacturer could pay fixed GBP 50/MWh on half its 25 GWh/month use for 12 months. At an average of 55 it receives 5; at 40 it pays 10 — net 50 either way.

A payer swaption gives the right to pay fixed — suits a buyer who expects prices to fall but wants protection. Most corporates avoid premiums, so banks offer a min-max (zero-cost collar): buy a payer swaption (cap) and sell one at a lower strike (floor).

Spread options

Options on spark/dark spreads, or on cross-border differences (e.g. a UK chemicals maker protected if UK power rises relative to France).

Monetising embedded optionality

An aluminium smelter with a flexible supply contract can stop smelting and resell its power when power prices spike. An indifference curve marks power/aluminium price pairs where it doesn’t care. It can sell OTM power calls (e.g. strike USD 100, premium USD 3, power at 80) — exercise leads to a forward sale, giving time to shut down. Higher implied vol → more premium. Risk: if aluminium rises, the curve shifts — a two-asset barrier can help.

Ratio swap: power and aluminium

The smelter swaps floating aluminium revenue to fixed and floating power cost to fixed, in a ratio (e.g. 14 MWh per tonne), locking its smelting margin in one structure.

US monthly and daily “power swaps” and swaptions

US “power swaps” are effectively physically settled forwards for a region and month (or day), settling just before delivery.

Monthly swaption example: European call, PJM West peak (16 h × 23 weekdays), 100 MW → 36,800 MWh, strike USD 50, premium USD 5/MWh, cash-settled vs CME futures. Power implied vols above 100% are common.

Daily swaptions = a strip of daily options — dearer than one monthly option (up to 23 chances to pay vs one).

Generator selling put swaptions (strike 50, premium 5):

Market price	Physical	Premium	Option	Net
60	+60	+5	0	65
50	+50	+5	0	55
45	+45	+5	−5	45
30	+30	+5	−20	15

Income boost in normal markets, double hit in a slump.

Heat rate options

A heat rate call pays: max(power − heat rate strike × gas price − variable charge, 0) — a call on the spark spread. It mimics owning a gas plant.

Daily example (strike heat rate 7, gas USD 5/MMBtu, ignoring variable charge):

Day	Power	Market heat rate	Payoff per MWh
1	40	8	5
2	35	7	0
3	30	6	0

Typical term sheet: daily peak (08:00–23:00 → 1,500 MWh/day at 100 MW), ~250 days → 375,000 MWh, monthly premium USD 11,000 (132,000/yr), variable charge USD 3.50/MWh, ISO New England day-ahead power vs Henry Hub gas.

Uses: generators buy puts (margin protection) or sell OTM calls (income); traders take correlation views (calls gain as power–gas correlation falls); delta-hedge with forwards, other Greeks with heat rate or single-commodity options.

---

Checkpoint 4: derivatives

Quiz

Checkpoint — section 8.7

1 / 6

Fixed payer at GBP 48 on 20 MW baseload for a 30-day month; average 52. Net cash flow?

Pick an answer to see if you are right.

---

Rank: electricity practice

Quiz

Electricity practice

1 / 2

Rank the merit order from FIRST to LAST dispatched.

OrderDrag the handle to rearrange — top should be first.

1. ⋮⋮1Nuclear (70)
2. ⋮⋮2Natural gas (110)
3. ⋮⋮3Renewable (80)
4. ⋮⋮4Coal (120)
5. ⋮⋮5Diesel (150)

Drag items into the correct order (or use ↑↓), then check.

---

Check: electricity

Quiz

Electricity

1 / 6

MW measures ____; MWh measures ____.

Pick an answer to see if you are right.

---

Calculate: numeric practice

Type the number (no units; minus sign for payments). 1 therm = 0.0293071 MWh.

Formula sheet

Item	Formula
Energy	MW × hours
Contract volume	MW × hours/day × days
Reserve margin	(capacity − peak) ÷ capacity
Raw spark / dark	power − fuel per MWh ÷ efficiency
Clean spread	raw − CO₂ factor × carbon price
Market heat rate	power ÷ gas
Spark (heat rate)	power − gas × heat rate
Swap net (fixed payer)	(floating − fixed) × MW × hours
CfD payment to generator	strike − market price
Heat rate call	max(power − HR × gas − charge, 0)

Compute

Calculate: electricity

1 / 12

Story problem — figure out which numbers are the formula inputs, then compute.

Summer baseload 100 MW for 26 weeks (24 h × 7 days). Total MWh?

Enter a number (at most 0 decimal places).

Randomized practice

Compute

Practice: power drills

1 / 6

Story problem — figure out which numbers are the formula inputs, then compute.

Node A LMP 29, node B LMP 54. 150 MW flows A → B for one hour. Congestion revenue collected by the ISO?

Enter a number (at most 0 decimal places).

---

Takeaways

Remember

• Power can’t be stored: supply = demand every second; MW (rate) vs MWh (energy)
• Chain: generation → transmission (TSO) → distribution → supply; networks regulated, generation/supply competitive
• Pools pay the marginal price; bilateral markets self-dispatch and settle imbalances
• Prices follow load, weather, reserve margin, marginal fuel, renewables — spiking or going negative
• Spark/dark spreads (and clean versions with carbon) and heat rates measure generator margins
• Contracts by load shape; value = MWh × price; the merit order sets price; GB EFA blocks; US LMP and FTRs
• Derivatives: forwards, swaps, CfDs, swaptions/collars, spread options, embedded optionality, ratio swaps, heat rate options

Further Reading

• 6. Crude oil
• 2. Derivative valuation — negative prices and spread options
• Next: 9. Plastics
• Energy Institute Statistical Review of World Energy; US EIA; Ofgem; ACER; NERC
• Background: N. Schofield, Commodity Derivatives: Markets and Applications (Wiley) — used as a study outline; V. Kaminski, Energy Markets (2012)