Europe

Every day, shortly after noon, an algorithm called EUPHEMIA clears the largest electricity auction on earth. It takes bids and offers collected by power exchanges across 27 countries, along with the transfer capacities the transmission operators are willing to make available between them, and solves for the set of prices and flows that maximises total welfare across 37 bidding zones for all 24 hours of the following day. In 2024 it cleared around 1,840 terawatt hours. It runs in under seventeen minutes.

The result is a set of prices that are identical across any two zones with spare capacity between them, and diverge exactly when the interconnection between them fills up. Power flows toward the expensive zone automatically, because the algorithm allocates the cross-border capacity as part of the clearing rather than requiring anyone to buy it separately first. Market coupling replaced a system in which traders bought transmission capacity and energy in separate transactions, and frequently moved power in the wrong direction.

That machine is the achievement of European electricity policy. What the same continent has struggled with is everything around it.

The institutional split

Chapter 14 described the arrangement and it shapes what follows. The exchanges that collect the orders are commercial businesses. The transmission operators that own the wires are separate companies, organised nationally, coordinating through ENTSO-E. The algorithm belongs to the exchanges collectively. No single entity holds both the network model and the market clearing.

So the auction clears against a simplified representation of the grid, internal constraints inside each zone are invisible to it, and the transmission operators afterwards redispatch to make the outcome physically deliverable. Chapter 15 covered the cost of that in Germany, close to three billion euros a year. The design produces one liquid price per country that a whole economy can hedge against, and pays for it with a redispatch bill nobody trades against.

2022, and the year a textbook term became a tax

Chapter 8 explained uniform-price clearing and inframarginal rent as matters of market design. In 2022 they became matters of politics, and the speed of that transition is the most instructive thing in this chapter.

Russian pipeline supply fell away through the year while French nuclear availability was unusually poor and hydrology was weak. Gas at the Dutch TTF hub reached about 339 euros per megawatt hour on 26 August 2022, against a 2015 to 2019 average near 17.59. Because gas plants were setting the power price in most hours, the power price followed. German year-ahead power passed 995 euros per megawatt hour and the French equivalent went beyond 1,100.

The mechanism worked exactly as Chapter 8 describes. Every wind farm, solar park, hydro station, lignite plant and nuclear reactor on the continent received the price set by the marginal gas unit, while its own costs had barely moved. Across 2022 average European electricity prices rose from roughly 80 to 272 euros per megawatt hour, with fuel costs accounting for around 85% of the increase.

The response came in two stages, and the distinction between them is worth holding.

The emergency measure was Council Regulation (EU) 2022/1854, adopted on 6 October 2022, which capped the market revenues of inframarginal generators at 180 euros per megawatt hour and directed the excess to member states for return to consumers. The word inframarginal appeared in binding law. A concept that exists only because of a clearing rule became something governments collected.

The structural measure followed. The reformed electricity market design was agreed at the end of 2023, adopted in May 2024 and in force from July 2024, and its centrepiece is the two-way contract for difference. New wind, solar, geothermal, reservoir-free hydro and nuclear built with public support are to be contracted at a strike price. When the market price sits below the strike, the public counterparty pays the generator. When it sits above, the generator pays the difference back, and those receipts flow to customers. The rules apply after a three-year transition.

In roughly two years the same idea travelled from a term in a textbook, to an emergency confiscation, to permanent architecture for how new generation is financed on a continent.

Iberia, and the service somebody finally paid for

The second set piece is more recent and pays off Chapter 6 and Chapter 13 together.

The expert panel found that voltage and reactive power control failed on 28 April 2025, and that conventional plants obliged to hold voltage did not do so. Chapter 13 argued that services procured through markets are verified every settlement interval while services imposed as obligations are checked whenever somebody chooses to check them. Spain had the second kind.

The details of the Spanish arrangement make the point sharper than any general argument could. Renewable facilities were required to hold their power factor inside a band measured on an hourly basis, with reactive capability limited to a fraction of active output. There were penalties for breaching the requirement and no payment at all for supplying dynamic voltage support in the seconds during which system stability is actually decided. The rule asked for an hourly average and the physics needed a second-by-second response.

The instrument that governs this, Operational Procedure 7.4, was about twenty-five years old, and a revision had been sitting with the regulator in draft form for roughly five years. It was updated on 12 June 2025, six weeks after the blackout, and fully implemented on 17 March 2026. The revision lets the system operator draw on reactive capability from generation, demand and storage, and it introduces remuneration for the more demanding real-time setpoint-following service.

Spain started paying for voltage support.

Set that beside Chapter 13's other case. FERC concluded in Order No. 904, effective January 2025, that generators should not be compensated for reactive power within the standard power factor range, reasoning that operating in that band is a condition of being connected and that a separate payment risks paying twice. Within a few months, and on the other side of the Atlantic, Spain reached the opposite conclusion about the same capability, and did so with a blackout as its evidence.

Both decisions are defensible on their own terms, and this book will not pretend one regulator was obviously right. The pair illustrates something more useful: the boundary between what a power system buys and what it merely requires is genuinely unsettled, moves in both directions, and gets moved by events rather than by argument.

The other consequence was equipment. Spanish grid-scale battery capacity went from around 28 megawatts in April 2025 to roughly 193 megawatts a year later. The absolute numbers are small next to Texas or California. The rate of change, in a country that had almost no storage before the blackout, shows how quickly a rule change moves capital once it makes something payable.

Europe on the three axes

Set against Part Three, Europe answers the three questions in a way no American reader would design from scratch, and each answer has a reason.

The through-line is that European market design is made in public, through legislation, and usually after something has gone wrong. The 2022 crisis produced a revenue cap and then a contract regime. The Iberian blackout produced a grid code revision that had been stalled for five years and a payment for a service that had been free. Bidding zone boundaries are being argued about in a formal regulatory process with published numbers, which Chapter 15 covered.

That process is slower than a state regulator changing an ERCOT protocol, and it produces something Texas does not have: rules that thirty countries have agreed to, in writing, which is the only way a single price can exist across a continent that fought two wars in the last century.