How Location Is Priced

PJM computes a price at roughly 11,000 points on its network, every five minutes. The whole of continental Europe computes about 40. One system serves around 65 million people and the other around 450 million.

That comparison is the sharpest single difference between the two largest electricity markets on earth, and everything in this chapter follows from it. Chapter 10 established why a price varies by location: the wires are finite, constraints bind, and the cost of serving one more megawatt-hour genuinely differs from one substation to the next. The design question is whether the market is allowed to see that.

The market that sees the network, and the one that does not

A nodal market solves the dispatch against a model of the actual grid, with every relevant line limit in it, and publishes the resulting price at every point. Generators settle where they inject. The price at a location carries the full information about what it costs to serve that location.

A zonal market draws a boundary, usually a national border, and clears as though everything inside were a copper plate. Any generator inside the zone can serve any load inside the zone, at one price, and the market never considers the lines between them. Only the links between zones are modelled as constraints.

Inside a small, well-meshed country with no serious internal bottlenecks, those two designs produce almost the same answer. Germany is neither small nor unconstrained. Its wind is in the north and on the coast, its heavy industry and its largest cities are in the south and west, and the transmission between them has been inadequate for over a decade. Germany remains a single bidding zone.

What happens when the market ignores a wire

The clearing produces a schedule. Sometimes that schedule cannot be delivered, because moving the scheduled quantity of northern wind to southern demand would overload lines the market was never shown. Physics does not accept the schedule.

So the transmission operator intervenes after the market has cleared. It instructs generation in the north to reduce output and generation in the south to increase it, compensating both: paying the southern plant to run when the market did not select it, and paying the northern plant for output it was scheduled to produce and is no longer allowed to. That process is redispatch, and the cost is recovered through network charges rather than appearing in any price.

Germany spent roughly 3.2 billion euros on redispatch in 2023 and close to 2.9 billion in 2024. In 2023 grid bottlenecks meant around 19 terawatt hours of electricity could not be delivered where it was wanted, and in 2024 about 3.5% of German renewable generation could not reach consumers. Transmission operators expect the volumes to rise again as more wind connects ahead of the lines being built to move it.

Three designs

Australia sits between the two. The National Electricity Market has five regions, one per state, and every asset in a region receives the same spot price. Losses are then handled through marginal loss factors, recalculated annually, which scale a generator's revenue according to how far it sits from the regional reference point. The scheme gives a locational signal of a sort, though delivered as an annual number rather than a price. Loss factors have moved sharply from year to year in weak parts of the network where large amounts of remote renewable generation connected quickly, which produces exactly the revenue uncertainty a long-term contract exists to remove. New Zealand, next door, runs a nodal market.

The German argument, and a number that complicates it

Whether Germany should be split into several bidding zones has been argued for more than a decade, and it is the live version of this chapter's question. The European regulator ACER set out the alternative configurations to be studied, and transmission operators published their assessment in April 2025.

The modelling favours splitting. A configuration dividing Germany and Luxembourg into five zones scored highest, at an estimated benefit of around 339 million euros for the 2025 target year, with the various split options ranging from roughly 251 million upward. ACER considered the operators' study conservative, putting the gain from splitting Germany alongside a Dutch reconfiguration at 450 to 540 million euros a year, around 70% above the operators' figure.

Set those numbers against redispatch spending of nearly three billion euros a year and the relationship between them repays attention. Splitting the zone does not recover the redispatch bill. Much of that spending is a transfer to generators rather than waste, and the physical shortage of wires between north and south survives any decision about pricing. What a split changes is the quality of the signals: where new generation chooses to connect, where industrial load chooses to sit, which projects look attractive, and how much of the cost is visible to the people making those decisions. The gain is measured in better decisions over time rather than in an immediate saving.

The resistance is equally real. A single German price is a political object as much as an economic one, and splitting it would raise prices in the industrial south while lowering them in the windy north, creating identifiable losers in the regions with the most electoral weight. A single zone also supports a deep forward market, since traders can hedge German power as one product. Five zones would fragment that liquidity into five thinner books, and a forward curve that nobody trades is of limited use to anyone trying to finance a plant.

The one system that changed its mind

Texas provides the closest thing to a controlled experiment, because ERCOT ran a zonal market and then replaced it with a nodal one.

The nodal market went live on 1 December 2010, four years later than originally intended and after roughly 500 million dollars of systems investment, along with about fifty additional staff to run the more complex market. Studies of the period after the switch found the daily operating cost of thermal generation fell by around 3.9%, a saving that recovered the investment within a few years.

The transition is cited by both sides of the argument, which suits it. Half a billion dollars and a four-year delay is a genuine barrier, and the efficiency gain arrived and persisted. The relevant difference for Europe is that ERCOT rebuilt one market inside one state under one regulator. Redrawing bidding zones in Europe means renegotiating boundaries between sovereign states whose citizens pay the resulting prices.

What the choice actually settles

Reading this as a contest about accuracy would miss most of it. Nodal pricing is more accurate, and nobody serious disputes that. The argument continues because accuracy is only one of the things a market is being asked to deliver.

A nodal market puts the cost of congestion in the price, where the people whose decisions create and relieve congestion can see it, and hands them a continuous signal about where to build. It also produces thousands of illiquid prices, exposes individual generators to a node nobody can hedge without the instruments in Chapter 22, and creates locations where a single plant behind a constraint holds obvious market power.

A zonal market produces one liquid, tradeable price that a whole country can hedge against, at the cost of making the network invisible to everyone making decisions about it, and moving several billion euros a year into a charge that no market participant responds to.

Which is why this is one of the three axes rather than a solved question. The choice determines where congestion cost surfaces, who sees it, and whether anyone can act on it. Chapter 19 takes up what Europe has done with that choice, and Chapter 22 covers the instruments a nodal market has to invent so that participants can live with the prices it produces.