Who Dispatches

Everything in Part Two assumed an operator. Somebody receives the offers, runs the optimisation, decides which machines start, publishes the price and instructs plants to move. That entity has appeared in every chapter so far without being described, and its identity turns out to shape what the price can mean.

Three arrangements exist. In one, the operator owns the generation it is dispatching. In another, it owns nothing and runs both the grid and the market. In the third, it runs the grid while somebody else entirely runs the market.

The utility that owns everything

The original arrangement, and still the arrangement for roughly 30% of US electricity demand, is the vertically integrated utility. One company owns the power stations, the transmission lines and the distribution wires, holds an exclusive franchise to serve customers in a territory, and dispatches its own fleet to serve its own load at least cost.

No wholesale price emerges from this, because nothing is being bought or sold. The utility decides internally which of its plants to run. Its customers pay a regulated tariff, and the tariff is set in a proceeding before a state commission, which reviews the utility's costs, decides which were prudently incurred, and permits a rate of return on the capital deployed.

Readers arriving from a market background tend to treat this as the unreformed version. It survives across the US Southeast, most of the Mountain West and much of the Northwest, and it has a coherent logic. All three questions in this part of the book still have to be answered under it. How much capacity to build is answered by an integrated resource plan approved by a regulator. Whether a plant was worth building is answered by a prudency review. The difference is that the answers arrive through a legal proceeding rather than through a price, and the risk of getting them wrong sits with customers rather than with investors.

The operator that owns nothing

The American alternative arrived through FERC Orders 888 and 889 in 1996, which required transmission owners to offer other people access to their wires on the same terms they gave themselves, and Order 2000 in 1999, which encouraged the formation of regional transmission organisations. Seven independent system operators and regional transmission organisations now serve around 70% of US demand.

The defining feature is that the operator owns no generation. It has no commercial interest in which machine runs, so it can be trusted with a decision worth billions a year to the parties on either side of it. A dispatcher that owned a coal plant would face an obvious temptation every time the coal plant sat just outside the money.

The second feature matters more than it usually gets credit for. An American ISO runs the grid and the market together, in one optimisation. The same calculation that decides which units are committed also produces the prices at which everyone settles, which is why a locational price in PJM is a physical statement about the network rather than the outcome of a separate trading venue.

The operator that runs the wires but not the market

Europe separated the same functions differently, and the difference explains a good deal about how the two continents behave.

A European transmission system operator, such as RTE in France or Amprion and TenneT in Germany, owns and operates the high-voltage network, runs balancing in real time, and carries out the redispatch described in Chapter 15. It does not run the day-ahead market. That is done by power exchanges, EPEX SPOT and Nord Pool among them, which are commercial businesses competing for volume, clearing an auction across coupled bidding zones.

So the market clears at an exchange, against a simplified picture of the network, and the transmission operator afterwards makes the physical adjustments the exchange result requires. Chapter 15 described the cost of that arrangement. Here is the institutional reason it exists: the entity computing the price and the entity responsible for the wires are different organisations, and only one of them has the network model.

Those three European structures come from the Third Energy Package of 2009. Ownership unbundling bars a generation or supply business from owning the transmission network at all. The ISO model leaves ownership with the parent while handing operation to an independent body. The independent transmission operator model, the lightest of the three, leaves both ownership and operation inside the group under ring-fencing rules intended to secure non-discriminatory access. Twenty-six independent transmission operators have been certified across ten member states, including France, Germany and Italy, so the lightest option is in wide use.

Three arrangements, side by side

Why the American reform stopped where it did

The restructuring wave of the 1990s was expected to continue until it covered the country. It stopped, and the reason has a date. California restructured, ran into the crisis of 2000 and 2001, and produced rolling blackouts, a bankrupt utility and a set of federal investigations. States that had been preparing to follow watched what happened and declined.

Chapter 2 covers what actually went wrong there, and the short version is that the failure had specific causes in a specific design rather than being an indictment of markets in general. What matters here is the institutional consequence. The map of who dispatches American electricity was frozen for two decades by a single episode, which left a third of the country outside organised markets and produced the awkward situation Chapter 17 has to describe.

The map is moving again

For the first time in twenty years, that boundary is being redrawn, and the pressure comes from the same load growth driving Chapter 23.

The Western Energy Imbalance Market, a real-time-only arrangement run by CAISO, has delivered around 5.6 billion dollars of benefits since 2014 by letting western utilities share resources across balancing areas five minutes at a time. That success made a day-ahead version attractive, and CAISO launched its Extended Day-Ahead Market on 1 May 2026, with PacifiCorp joining at launch and Portland General Electric following in October.

SPP moved on the same territory from the other direction. On 1 April 2026 it extended its regional transmission organisation into the Western Interconnection, becoming the first such organisation to operate across two interconnections, covering utilities in seven states from Arizona to Montana. Its separate Markets+ offering has commitments from further utilities for 2028.

A western region that spent two decades as the clearest example of vertically integrated utilities coexisting without a market is now being competed over by two organisations offering to dispatch it. The pattern is familiar from Chapter 2: the arrangement changes when the technical case becomes overwhelming enough to overcome the institutional inertia, and rarely before.

The three axes together

Part Three is complete, and the three questions can be stated in one place. Who dispatches. How location is priced. How capacity is paid for.

The combinations that exist in the world are not random. A vertically integrated utility has no need of locational prices, because it has no counterparties to send a signal to, and no need of a capacity market, because its regulator approves construction directly. An operator that owns nothing must produce a price, because a price is the only instrument it has for telling anyone anything. The further a system moves from ownership toward coordination, the more of its decisions have to be expressed as numbers that strangers will act on.

Part Four takes the systems one at a time. Every one of them is a set of answers to these three questions, chosen at a particular moment, for reasons that made sense to the people writing the rules.