The grid already prices location; zonal markets refuse to read the signal.

Executive summary

What zonal pricing suppresses, and what it costs to keep the compass off

Three findings from a US validation lab, a distribution-level evidence pass, and the SSI Index anti-maladaptation gate.

A locational marginal price — the price a nodal electricity market publishes at every point on the network — is the shadow price of the grid: the marginal system value of one more megawatt-hour injected or withdrawn there, given every binding constraint. It is the mechanism by which a decentralised system discovers what it collectively knows about where energy is scarce and where it is abundant. Europe's wholesale markets, and Great Britain's, price electricity by zone instead, paying a single average across areas that can span a thousand kilometres and burying the intra-zonal scarcity in out-of-market redispatch.¹ This brief asks the calibrated question the economics literature poses — not “should Europe go nodal,” but how much locational granularity is worth it, and which zones and nodes benefit most — and it demonstrates the answer with the asset whose value most tightly is the locational signal: the battery.

First, the suppressed signal is measurable and learnable. A transfer function trained only on day-ahead drivers — load, renewable forecast, calendar, hub price — predicts the cross-node price dispersion around a reference hub with out-of-sample skill that is robust across two independent US markets: cross-node spread at R² 0.65 in California over two years of five-fold walk-forward validation, and the charge-side price tail at R² 0.97. Measured T1 The structural width of the locational signal transfers from California to PJM; only the spiky discharge-side tail does not — and that tail was always the trader's piece, never the planner's. Second, the value is larger and far more place-specific at distribution than at transmission — roughly an order of magnitude in published US avoided-cost studies — and its purest component, voltage and power quality, is local, physically un-diversifiable, worth an estimated €150 billion a year of avoided cost across the EU, and almost entirely unpriced.² Probable T1.5 Third, a correctly-sited battery converts that signal into the EU's own objectives — curtailment relief, non-wires capital deferral, and locational resilience — and passes the SSI Index W1–W10 anti-maladaptation gate on every axis, whereas the same capital sited by the flat zonal price degrades reliability and supply-security at the constrained node. Under the Index's own governance test, zonal mis-siting is maladaptation. Opinable

Introduction

A compass switched off, in the decade Europe most needs it

Past, present, and a disciplined three-to-five-year forecast, opened on the recent shock.

ON 28 APRIL 2025, a voltage disturbance on the Iberian peninsula propagated in seconds into the continent's largest blackout in two decades, dropping tens of millions of connections across Spain and Portugal before the grid could island and recover.³ The forensic questions that followed — where did the disturbance originate, which nodes held, which cascaded — are locational questions. They are the questions a nodal price answers continuously and in advance, and that a zonal price cannot, because it does not carry a node.

Europe is committing on the order of €584 billion to its grids this decade under the European Grids Action Plan, and has bound itself to continuous progress in adaptive capacity and resilience under Article 5 of the European Climate Law and to critical-entity resilience for substation operators under the CER Directive.⁴ It is doing so while renewable build races ahead of the wires, while interconnection queues lengthen, and while curtailment — clean electricity discarded because it cannot be moved — rises in exactly the pockets a locational price would flag. The 2024 reform of the EU electricity-market design reaffirmed the intent to sharpen locational and flexibility signals; it did not switch the compass back on.⁵

This is not an argument that Europe should abandon zonal pricing. The peer-reviewed literature is emphatic and calibrated: full nodal pricing yields real efficiency gains, but partial zone-splits are mostly distributional with small welfare deltas, more zones can even reduce welfare if the split is mis-specified, and merely anticipating a better locational signal captures much of the available gain.⁶ Probable The honest object of study is therefore the magnitude and the geography of the suppressed signal — how large it is, where it concentrates, and what it would unlock if it were read — and the honest method is to measure it where nodal prices actually clear (the United States), validate that the measurement is a transferable structural property rather than a local accident, and then reason carefully, tier by tier, toward the European distribution grid where the SSI Index operates and where the value is largest.

The brief is built as a war game, the SSI Index's standard narrative architecture: a briefing on initial conditions, the actors and their objectives, the decision points each faces, the consequences the methodology traces, a surprise event, and a hotwash. The reader is not told what to conclude. The reader plays through and arrives.

Key takeaways

Six findings a busy desk can carry away

Each carries its confidence register; each links to the section that earns it.

• A locational marginal price is the grid's shadow price and the efficient allocation signal for storage, generation, demand and wires; zonal pricing pays the average and hides intra-zonal scarcity in redispatch. §1 Probable
• The suppressed signal is learnable: a day-ahead-driver transfer function predicts cross-node dispersion at R² 0.65 (California, two years, five-fold) and the charge-side tail at R² 0.97; the structural width transfers to PJM. §4 Measured T1
• Only the spiky discharge-side tail fails to transfer across markets — skilful-but-noisy in California (R² 0.33), no out-of-sample skill in PJM over a full year (R² −0.05). That tail is the trader's edge, not the planner's signal. §4 Measured T1
• Locational value is roughly ten times larger and far more place-specific at distribution than transmission; its purest part — voltage / power quality — is local, un-diversifiable, ~€150 bn/yr, and unpriced. §4 Probable T1.5
• A correctly-sited battery advances RePowerEU and resilience simultaneously; a zonal-mis-sited one under-delivers on both because the compass that would place it is off. §5 Opinable
• Under the SSI Index W1–W10 anti-maladaptation gate, correctly-sited storage leaves no community axis worse off and mis-sited storage fails on reliability and supply security — so mis-siting is, formally, maladaptation. §6 Opinable

Phase 1 · Briefing

Initial conditions: a network that knows a price no market will publish

Year zero is a continent pricing electricity by the average of areas that can span a thousand kilometres.

Begin with the object itself. A locational marginal price (LMP) is what an optimally-dispatched power system would pay for one more megawatt-hour at a specific node, and charge for one more withdrawn there, once every transmission constraint that binds is respected. Formally it decomposes into three parts: a system-wide energy component, a congestion component that reflects the shadow price of the binding lines between that node and the reference, and a loss component. The congestion component is the interesting one. It is zero where the network is unconstrained and large where it is not, and it is exactly the quantity a zonal market cannot express — because a zone has no node, only an average.⁷ Probable

The intellectual lineage is settled. Schweppe and colleagues formalised spot pricing of electricity in the 1980s; Hogan showed how a system of nodal prices supports a coherent market in transmission rights.⁸ The United States, PJM and CAISO among them, runs on this basis today — tens of thousands of pricing nodes, cleared twice daily. Europe and Great Britain, by contrast, chose zonal configurations under Regulation (EU) 2019/943 and its predecessors, for reasons that were substantially about liquidity, hedging simplicity and distributional politics rather than a claim that intra-zonal congestion is small.⁹ When congestion nonetheless binds inside a zone — as it increasingly does, wherever renewable generation has been built into a locally-saturated pocket — the system operator resolves it after the market clears, through out-of-market redispatch whose cost is socialised across all consumers and whose locational information is discarded rather than published.

The consequence for capital allocation is the heart of the matter. A developer siting a battery or a solar farm chases the price it can see, which is the zone price. It therefore builds where the zone looks attractive, not where the node is genuinely scarce — and the two diverge precisely in the congested pockets that most need investment. The gap between the flat zonal price and the true distribution of nodal prices underneath it is the suppressed price-discovery information. Storage is where this shows most starkly, because a battery earns the spread, and the spread is the congestion the zonal price hides.

How large is the intra-zonal gradient? The SSI Enhanced Neural Network's zonal-to-nodal transition model carries per-zone congestion factors for Italy as a modelled input — a working assumption, not a measurement, and flagged as such — and even as an assumption it describes a fivefold gradient from the uncongested north to the island zones.

Read this chart for what it is: a modelled stand-in for the intra-zonal congestion that each zone's single price still hides. Stage 2 of this work has since measured the related inter-zonal gradient directly, on a full year of Italy's seven observed zonal prices (§4.2b) — confirming that a single national price would hide a large, structured locational signal, though the measured gradient is episodic and signed (North dear; South and islands cheap; Sicily the volatility epicentre) rather than the smooth monotone the model assumes. The claim a single zonal price is definitionally unable to make — that location has a price — is, at the inter-zonal level, now observed rather than assumed.

Phase 2 · Actors

Six actors, six objectives, one missing signal

Each actor is rational within the information it can see; the pathology is in what none of them can see.

A walkthrough of this kind is a multi-actor exercise, and the locational-pricing question is illuminating precisely because every actor is behaving reasonably given the price it observes. The dysfunction is not anyone's error; it is the absence of a shared signal.

The transmission and distribution system operators hold the network model and therefore know, better than anyone, where the constraints bind. They resolve intra-zonal congestion through redispatch and see its cost accumulate; but the market design gives them no mechanism to publish that locational knowledge as a price that would attract private capital to relieve it. They carry the information and cannot monetise it into investment signal.

The regulators — ACER at EU level and the national authorities — are charged with efficient, secure and affordable supply, and with the market-design reform's stated intent to sharpen locational and flexibility signals. They face a genuine trade-off between the efficiency of sharper signals and the liquidity, hedgeability and distributional smoothness that zonal pricing was chosen to protect. Their decision problem is not whether locational value exists but whether the institutional cost of revealing it is worth paying, and where.

The Member States weigh cohesion and the distributional map: sharper locational prices create winners (uncongested, well-connected regions) and losers (congested or peripheral ones), and the losers are often already the periphery. This is the real seat of the hesitation, and no efficiency argument dissolves it. It can only be addressed honestly, by naming who bears the adjustment and pairing the signal with the transfer.

The Commission holds the umbrella objectives — RePowerEU's security-and-transition acceleration, the Grids Action Plan's €584 billion, the Climate Law's binding adaptive-capacity duty — every one of which is advanced by capital flowing to the right node and retarded by capital flowing to the wrong one. Its interest is in the allocation outcome, whatever the pricing instrument.

The developers and investors are the actors the signal is supposed to steer. They will build flexibility wherever the observable price rewards it. Give them a zone price and they build to the zone; give them a node price and they build to the node. They are not the problem; they are the transmission mechanism, faithfully transmitting whatever signal they are given.

The communities and their adversaries close the set. A community experiences the grid as reliability, voltage quality, local employment and resilience under stress — all local, all nodal in character. And an adversary conducting grey-zone disruption chooses targets by their locational criticality, which is again a nodal property. Resilience, like congestion, is something a zone cannot see.¹⁰

Held together, the actor map yields the exercise's first insight before any move is played: the locational signal is the one piece of information that would make every actor's problem easier, and it is the one piece the prevailing market design withholds. The following phases test what happens when someone tries to read it anyway.

Phase 3 · Decision points

Four doors, and the one the literature says is nearly as good as the best

The choice is not binary between zonal and nodal; it is a graded question of how much locational granularity to reveal.

At the decision node, the regulator-planner faces four doors, not two.

Door one — stay zonal. Preserve liquidity and hedging simplicity; keep resolving congestion through socialised redispatch; accept that capital will keep landing in locally-saturated pockets and that curtailment and queue congestion will keep rising where the signal is darkest. Door two — split zones further. Redraw a few boundaries so the average is taken over smaller areas. The literature's verdict here is sobering: partial zone-splits are mostly distributional rather than efficiency-improving, deliver small welfare deltas, and can even reduce welfare if the split is mis-specified against the true constraint pattern.¹¹ Probable Door three — anticipate the signal. Keep zonal clearing but let planning, connection and flexibility procurement act as if the nodal signal were visible — using locational value estimates to steer where storage, flexibility and reinforcement are sited and rewarded. Door four — go fully nodal. Publish the node price and capture the full efficiency gain, at the institutional cost of rebuilding hedging and confronting the distributional map head-on.

The striking result in the peer-reviewed work — Neuhoff and colleagues, Ambrosius and colleagues, and the German-market modelling of Egerer, Grimm and Trepper — is that door three captures a large share of what door four would deliver.¹² Probable The gain is not primarily in the act of publishing the price; it is in acting on the locational information. If planners and flexibility markets can be given a credible, empirically-grounded map of where the locational value is, they can steer capital most of the way to the efficient allocation without paying the full institutional cost of a nodal rebuild. That is a hopeful finding, and it is the one that makes this brief's method worth the trouble: a good locational-value map is not a second-best consolation for those who cannot go nodal; it is most of the prize.

The decision the brief recommends attending to is therefore not “which door” in the abstract, but “how good is the map that door three depends on.” If the locational-value map is credible, transferable and empirically anchored, door three becomes available to every jurisdiction that finds door four institutionally out of reach — which today is all of Europe. The next phase asks whether such a map can actually be built.

Phase 4 · Consequences

Can the suppressed signal be measured? A US validation lab says yes — for the part that matters

Measure where nodal prices clear, prove the measurement is structural not local, then reason tier by tier toward the distribution grid.

4.1 — The transfer function, and what it learns

To measure the suppressed signal without a European nodal market to observe, the method borrows the two markets that publish node prices at scale — California's CAISO and PJM — and asks a disciplined question: using only information a European planner already has the morning before delivery (day-ahead load forecast, renewable forecast, calendar structure, and the zonal or hub reference price), how much of the cross-node price dispersion can be predicted? Dispersion, not the individual node: the width of the fan of nodal prices around the reference, and the shape of its tails. If that fan is predictable from day-ahead drivers, then the locational signal a zonal market suppresses is recoverable by a planner who wants to act on it — door three becomes operational.

The instrument is a gradient-boosted transfer function, trained and validated out-of-sample under expanding-window walk-forward cross-validation — the discipline that asks whether the model predicts future dispersion from past data, fold after fold, rather than merely fitting history. In California, over two years and five folds, it predicts the cross-node spread with a mean out-of-sample R² of 0.65 (mean absolute error ~$5/MWh) and the charge-side price tail — the p10 basis, which governs when a battery should buy — at R² 0.97. Measured T1 These are not fitted-in-sample numbers; they are the model's skill on data it did not see.

4.2 — The honesty test: what transfers, and what does not

A single-market result could be a California accident. The decisive test is whether the same transfer function, re-estimated on PJM with matched features over a full year under the same five-fold discipline, exhibits the same structure. It does — for the part that matters. The width of the locational fan (R² 0.59) and the charge-side tail (R² 0.62) transfer robustly to PJM, skilful in every fold. The one component that does not transfer is the discharge-side tail — the p90 basis, the spiky upper extreme that governs the single most profitable hours to sell. It is skilful-but-noisy in California (R² 0.33) and carries no out-of-sample skill at all in PJM (R² −0.05, unstable across folds): the scarcity spikes it tracks are idiosyncratic to each network, so the California skill does not replicate.¹¹ Measured T1

This result is now closed at full-year scale rather than pending. The PJM estimate rests on a full year of day-ahead prices under the same five-fold walk-forward discipline — 364 of 365 days retrieved — and the finding holds and sharpens: fan width (R² 0.59) and the charge-side tail (R² 0.62) are robustly skilful in every fold, while the discharge tail carries no skill (R² −0.05). Measured T1 The 120-day pilot had read the discharge tail at R² −0.41; over a full year it settles near zero — the same conclusion, now on a stable base. The brief leans its locational-value argument entirely on the two robust components and disclaims the tail.

4.2b — The European measurement (Italy, Stage 2)

The US result validates the method; Stage 2 applies it to the one EU market that already clears distinct zonal prices — Italy's seven bidding zones — over a full year (8,760 hours), measuring the European number directly rather than transporting the US one. Three observed findings, all Tier 1. Magnitude: a single national Italian price would hide €0/MWh on the median hour (the zones couple) but €23.5/MWh on the top decile, €114 on the top percentile, with the gap material (>€5) on 25.6% of hours — the suppressed signal is large but episodic, concentrated in congestion events. Measured T1 Geography: the gradient is signed — North dear (NORD +€1.8, CNOR +€2.5/MWh), South and islands cheap (SARD −€3.3, CALA −€2.0, SUD −€2.0), Sicily the volatility epicentre (large deviations on 16.9% of hours) — the renewable-rich, export-constrained South spilling surplus that depresses its price while the demand-heavy, import-constrained North firms. Measured T1 Recoverability: here the honesty of §4.2 pays out. At zone granularity the inter-zonal spread is a coupling/decoupling switch — the same episodic object as the US discharge tail — and day-ahead drivers, even with spatial South-versus-North imbalance features, recover it only weakly (fan-width R² 0.31, against 0.65 for the continuous US node-level fan). Measured T1 The reading is therefore precise: the thesis's core — that a large, geographically-structured locational signal exists and that zonal pricing averages it away — is now directly observed on European ground truth; the anticipation mechanism (door three) is strong at node granularity and weak at zone granularity, because a rare price switch is intrinsically harder to forecast than a continuous fan. The map Europe needs is therefore best built at finer-than-zone granularity, where the signal is continuous rather than a switch.

4.2c — The map, and the granularity paradox (Stage 3)

Italy is the only EU market that clears distinct zonal prices, so a pan-European map of where the suppressed value concentrates cannot be one homogeneous observed surface — it must be a three-tier composite, every entry tagged by how it is known: the observed inter-zonal basis where a market splits zones (Italy and the Nordics now measured on the same engine — nineteen zones between them); the observed redispatch cost where a country runs a single national zone; and, finer than either, the SSI Index’s per-substation surface. The tiers are never blended into one cross-unit number — a €/MWh basis, a €/year redispatch bill and an index score are not commensurable — but each location carries a composite class from a stated rule, and the pattern that falls out is the map’s central finding.²³ Measured T1

The finding is a paradox: the countries with the least locational price granularity carry the most suppressed value. Germany and Great Britain each clear a single national price — the locational signal is invisible in their markets by construction — yet each resolves the congestion that price hides through socialised redispatch running into the billions a year — both now primary figures: Germany’s bill €3.08 billion in 2025 (BNetzA/SMARD; €3.34 bn in 2023, €2.95 bn in 2024), Britain’s £2.21 billion in FY2025-26 (NESO; ≈€2.6 bn, and almost entirely thermal — £1.96 bn of it) — the observable after-the-fact cost of the information a locational price would have published. Italy, with seven zones, at least surfaces part of its gradient in-market and pays a smaller residual — also now primary, €1.48 billion of MSD dispatching cost in 2024 (Terna consuntivo; €0.80 bn on the narrow ancillary-and-congestion definition alone). The compass is switched off hardest exactly where the bill for keeping it off is largest.

Two honesties bound the figure. Redispatch is an annual socialised cost, not a per-MWh locational price — it ranks where a national zone hides the most value, and — now that all three are primary regulator figures — is reported on each regulator’s own basis, which differ at the margin, so the figures are order-appropriate rather than scope-identical, and never a cleared price. And Italy’s own seven-zone split still averages over an intra-zonal residual it does not resolve; the map’s next move is therefore finer, not merely wider — toward the per-substation surface the distribution tier below now motivates.²³

Size is the obvious objection — Germany is simply the biggest — so it is worth normalising the same socialised bill several ways at once, and asking which ranking survives. It does not much matter which yardstick one picks: measured against the size of the economy, the population it serves, or the electricity it consumes, the hidden cost lands at nearly the same place, and the seven-zone system carries the least of it.

4.3 — Down a tier: the value is larger, and more local, at distribution

The transmission-node result is the validated floor of the argument. The larger prize is a tier below it, at distribution, where the SSI Index itself operates — and here the brief switches evidence tier explicitly and does not pretend otherwise. There is no market-cleared distribution price anywhere on earth; distribution locational marginal prices exist only as the output of engineering optimisation on utility-private network models.^12b But the distribution layer is anchorable against observed physical and administrative data — California's downloadable Integration Capacity Analysis maps hosting capacity at feeder and substation granularity; New York's Value of Distributed Energy Resources and California's Avoided Cost Calculator publish administrative locational values — which is honestly short of a price but well above pure proxy.¹³ T1.5

What that observed physical and administrative evidence shows is that locational value at distribution is both larger and far more place-specific than at transmission. In published US utility avoided-cost work, distribution-primary avoided capacity value spans roughly $13.6–$74 per kilowatt-year across locations, against a representative transmission figure near $7.7 — an order of magnitude larger at the top and, more importantly for the allocation argument, spread over a five-fold-plus locational range.¹⁴ Probable T1.5

And the purest component of distribution locational value is the one that no market anywhere prices at all: voltage and power quality. Reactive power cannot be transported over distance — it is consumed locally by the physics of the network — so voltage support is irreducibly local and un-diversifiable, unlike frequency, which is a system-wide property already remunerated through balancing markets.¹⁵ The Leonardo Power Quality Initiative put the cost of poor power quality to the EU economy at roughly €150 billion a year, of which the large majority is voltage-driven; and it is met today almost entirely through mandatory grid-code obligation rather than through any price.¹⁶ Probable A battery's inverter can supply that voltage support — including in reactive-only mode, drawing no energy — at exactly the node that needs it. Here the locational signal is not merely suppressed; it is absent from every market, and the value is left entirely to grid-code compulsion.

4.4 — The battery as the place the signal becomes value

Assemble the streams and the demonstrator writes itself. A correctly-sited battery — sited at a node the locational-value map reveals as high-value — earns energy arbitrage (the spread the transfer function predicts), relieves curtailment by absorbing renewable surplus that the constrained node would otherwise spill, defers or avoids network reinforcement as a non-wires alternative, supplies the local voltage and power-quality support no market prices, and holds a resilience option that is worth most exactly where the grid is most stressed. The ordering of these streams by locality is the whole argument: energy arbitrage can in principle be earned anywhere on the network and is barely locational; voltage and power quality can be earned only at the node and cannot be solved remotely. The SSI Enhanced Neural Network's valuation methodology already encodes this ordering as a locality factor running from about 0.05 for arbitrage to 0.95 for power-quality-as-a-service, and prices the whole stack, including the zonal-to-nodal regime shift itself, as a real option.¹⁷ T2

The magnitudes on this bridge are illustrative and carry no client or portfolio numbers — the hard wall against proprietary data is observed throughout — but the proportions carry the argument: the streams a zonal price can partly signal (arbitrage) are the small end of the bar, and the streams it cannot signal at all (voltage, resilience, the locational part of deferral) are the large end. A market that reads only the arbitrage is reading only the sliver of value that is least locational. The same battery, sited to the zone rather than the node, keeps only the low-λ sliver and forfeits the rest — which is the demonstrator's counterfactual, and the subject of the surprise phase.

4.5 — The demonstrator: what even the arbitrage sliver is worth (Stage 4)

The bridge’s smallest, least-locational stream — energy arbitrage — is also the only one that lives entirely in observed day-ahead prices, so it is the one the demonstrator measures rather than models — here on a full year of Italy’s seven observed zonal prices. Stage 4 runs a transparent price-taker battery (charge the cheapest hours, discharge the dearest; four-hour, one cycle a day, round-trip 0.85) twice: located, dispatching against and settling at the local zonal price, and national, dispatching against and settling at the single cross-zone average a single-zone market would show. The difference is the arbitrage a storage asset earns — or forgoes — purely because it settles at the local price rather than the national average: the storage-relevant, and signed, face of the §4.2b information loss.²⁴ It is emphatically not the basis of §4.2b: the basis is a cross-sectional level gap, this is an intertemporal spread the local price’s hour-to-hour shape offers over the average’s; the two are different components of locational value and are never added.

On the observed year the result is not merely large but signed, and the sign is the finding. In all four renewable-surplus, export-constrained southern and island zones a battery earns more than the national-price operator — Sicily and Sardinia by roughly €11,700/MW-year each, the South by €5,600, Calabria by €4,500 — because their cheap, congestion-driven local prices are the most volatile, and volatility is what a battery monetises. In all three dear, import-constrained northern zones it earns less — the North by €10,800/MW-year — because the national average is itself pulled around by the volatile South, so a northern battery settling at its own firmer local price forgoes arbitrage the average would have shown it. The sign of the uplift matches the sign of the §4.2b basis in every one of the seven zones: cheap-and-volatile pays the storage operator to read location, dear-and-firm does not. A single national price does not merely mis-state this — it makes the southern value uncapturable, because there is no local price to dispatch against, and it flatters a northern siting the local reality would not support. And this is still only the low-λ sliver: the large, un-signalled streams of the §4.4 bridge sit above it. Whether a high-uplift node should actually receive the battery is then not a value question at all but a governance one — the anti-maladaptation test the hotwash runs on this very node.

The same engine, pointed at a second observed market — the Nordics’ twelve zones — both confirms the finding and sharpens it. The signed locational uplift is there again, larger and wider (from +€54,900/MW-year in East Denmark to −€36,900 in southern Norway), and again a single national price would hide all of it. But the geography is the mirror image of Italy’s, and that mirror is the point.

In Italy the uplift tracked the basis sign in every zone because the signal there is renewable-surplus curtailment: cheap, volatile southern zones pay the storage operator, dear firm northern ones do not. In the Nordics the sign does not track the basis at all — it tracks volatility, which here is a hydro-versus-continental story. The Danish and southern-Swedish zones, wind-heavy and coupled to the continental market, carry the most volatile prices and the largest positive uplift; the Norwegian and northern-Swedish hydro zones, whose reservoirs smooth prices flat, offer a battery less arbitrage than the national average and score deeply negative. The magnitudes are far larger than Italy’s (±€25–55k against ±€2–12k), reflecting the much wider zonal divergence the Nordic system has run in recent years, not a steady-state promise. Two observed markets, then, reach the same structural conclusion — locational storage value is large, signed, and invisible under a national price — through opposite mechanisms. That is precisely why the map of §4.2c must be built per market from observed prices rather than extrapolated from one formula: the value is real and locational everywhere, but what drives it is the local grid’s own physics.

Phase 5 · Surprise event

4.6 — The transatlantic mirror: location priced, or location buried

The measurement so far has been European. But the sharpest test of the thesis is transatlantic, because the United States already ran the experiment Europe is now debating — it stopped suppressing the locational signal. The theory is not new: locational (nodal) pricing was set out in full by Schweppe, Caramanis, Tabors and Bohn in Spot Pricing of Electricity in 1988, a decade before any market implemented it.²⁵ What followed was not one reform but a gradual, market-by-market conversion — PJM in 1998, the Midcontinent in 2005, California in 2009, Texas in 2010, the Southwest Power Pool in 2014 — each system moving to nodal pricing as congestion made the zonal approximation untenable (Fig 4c). Nodal pricing was never a whim; it was the answer a market reaches when the cost of the shortcut grows too large to ignore. Observed T1

Comparing the two systems takes one distinction, held with care, because the naive comparison is wrong. In a zonal market the grid’s locational constraints are absent from the price, so the operator relieves them out of market and socialises the cost — the redispatch, constraint and MSD bills of Fig 3b. That is deadweight corrective spend. In a nodal market the same constraints are in the price: they surface as congestion rent, a priced transfer collected by the operator and largely returned to load and transmission owners through financial transmission rights — a number the market can hedge and invest against. Congestion rent is therefore not the same kind of number as redispatch spend, and the two must never be summed. The true like-for-like of Europe’s socialised redispatch is the American uplift, or make-whole payment: the residual the operator still pays out of market once the price has done its work. In a market that prices location well, that residual is small — and its smallness is the whole point.

Figure 4b sets the two kinds side by side as basis points of GDP. Europe’s socialised redispatch bill runs at roughly seven basis points — deadweight. The American out-of-market residual, its make-whole uplift, is an order of magnitude smaller: even PJM, the largest organised market, pays uplift worth about half a per cent of its billing.²⁶ The locational value did not vanish; it moved into the price, where it shows up as some four basis points of congestion rent — a signal the market reads, hedges and builds against, rather than a bill socialised quietly onto consumers. Both continents pay for location. One prices it; the other buries it. Measured T1

And the buried cost is growing. US congestion cost held near six to eight billion dollars a year before 2021 and has since stepped to a ten-to-twenty-billion “new normal” (Fig 4d), lifted by three forces that operate identically on both sides of the Atlantic: renewable generation sited where the resource is rather than where the load is, an interconnection queue swollen to roughly 2.6 terawatts of mostly wind, solar and storage awaiting connection, and transmission build that lags both.²⁷ The same drivers are lifting Europe’s hidden bill — the pan-European remedial-action cost rose from €4.0 billion in 2023 to €4.3 billion in 2024, on some sixty terawatt-hours of redispatched energy.²⁸ The difference between the two continents is not whether the locational cost is rising. It is whether the market can see it. Measured T1

Europe stands at the start of the road the United States travelled, and so far it is hesitating. Great Britain examined locational pricing under its Review of Electricity Market Arrangements and, in July 2025, declined it in favour of a reformed single national price; Germany continues to refuse a split of its single bidding zone; the pan-European bidding-zone review is unresolved; and Italy has refined its long-standing zonal map only at the margin, from six zones to seven in 2021.²⁹ None of that changes the physics the US prices and Europe socialises. It changes only who can see the bill — and therefore who can act on it. That is the gap a substation-granular index is built to close.

A compound shock arrives at the constrained node — and only one battery is there

Resilience is local; the surprise phase is where the locality of the value stops being an abstraction.

Inject the surprise the scenario is built to test: a compound shock at a congested distribution node — a heat-driven demand peak coincident with a local voltage excursion, of the family the Iberian event made concrete, or a flood-plus-cyber concurrence of the kind the SSI Index's compound-risk modifier is designed to capture.¹⁸ The question the game poses is not whether batteries help — it is which battery helps, the one placed by the node signal or the one placed by the zone signal.

The correctly-sited battery is physically present at the stressed node. It can hold voltage in reactive mode, arrest the local excursion before it propagates, and carry the peak — the resilience option converts to realised value exactly when it is needed, which is why the option is worth most where the grid is most stressed. The zonal-mis-sited battery, holding the same nameplate megawatts, sits somewhere else in the same large zone, earning the same average zone price on paper, and does nothing for the node that is failing. Resilience does not travel. Reactive power does not travel. The zone average concealed the very gradient that decided which node would fail, and so it concealed where the battery needed to be.

This is the point at which the two EU objective-umbrellas the brief opened with converge on the same siting decision. RePowerEU and the transition want the battery where it relieves curtailment and defers reinforcement — the congested node. Security-of-supply and resilience want the battery where it holds the grid under a contingency — the same congested node. The locational signal points both objectives at one place; the zonal signal points them nowhere in particular. A correctly-sited battery advances decarbonisation-with-affordability and security-with-resilience together, and the reason it can is that it was placed by a price that reads location.

The surprise phase thus does what such a walkthrough's surprise is for: it collapses an abstract efficiency argument into a concrete, adversarially-tested claim. The value of reading the locational signal is not a diffuse welfare triangle; it is whether the node that failed had storage on it. The hotwash asks what discipline would have ensured it did — and whether the discipline can tell a genuine adaptation from a false one.

Phase 6 · Hotwash

Mis-siting is maladaptation — and the SSI Index already has the governance test

What the play surfaced: the highest-leverage decision, the honest boundaries, and the governance gate that turns the finding into a discipline.

The exercise surfaces one decision above all others: the locational-value map is the high-leverage instrument, because it is the input door three depends on and door three captures most of the available gain. Everything upstream (which pricing regime) and downstream (which battery, which reinforcement) is steered by the quality of that map. The brief's contribution is to show that the map's core — the structural width and charge-side shape of the locational signal — is empirically measurable and transferable across markets at transmission-node granularity, and — as the Italian Stage-2 measurement showed (§4.2b) — directly observable in magnitude and geography at EU zone granularity, even where the episodic zone-level spread is only weakly anticipable from day-ahead drivers. The bound is stated, not hidden.

6.1 — The governance test the finding plugs into

The SSI Index carries, at v4.2, a ten-axis community-value framework — W1 Reliability, W2 Carbon, W3 Local Economic, W4 Climate Resilience, W5 Digital Inclusion, W6 Sovereignty, W7 Local Value Retention, W8 Communications, W9 Supply Security, W10 Compute Multiplier — that functions as an anti-maladaptation gate: a per-substation, per-axis floor that any intervention must protect, on the discipline that no axis may be left worse off after a battery deployment, a line upgrade or an operational change.¹⁹ The framework's intellectual anchor is the maladaptation literature — adaptation that increases vulnerability, locks in exposure, or shifts burden onto the vulnerable — formalised in the IPCC's Sixth Assessment and the subsequent review work.²⁰ Probable

The gate is exactly the instrument the demonstrator's counterfactual needs. A correctly-sited battery, placed by the locational signal, improves reliability (W1) and supply security (W9) at the constrained node and leaves every other community axis at least as well off — it passes the gate. A zonal-mis-sited battery, placed by the average, leaves the constrained node's reliability and supply-security unimproved or degraded relative to the counterfactual where the capital had gone to the node — it fails the gate on W1 and W9. Under the SSI Index's own published governance test, therefore, zonal mis-siting is a maladaptation: it is an intervention that, judged per-axis and per-node, leaves a community-value axis worse off than the available alternative. The finding is not a rhetorical flourish; it is a per-axis test the framework already runs.

Stage 4 runs that test as a rule rather than a picture, and on the node the map names. On the observed epicentre — Sicily, a renewable-rich, export-constrained island — siting storage clears every axis: it absorbs surplus that would otherwise be curtailed, so W2 Carbon improves; it adds local ride-through, so W4 Climate Resilience and W1 Reliability improve; nothing is left worse off. PASS. But the same high-value node fails the moment a second, physical fact is admitted: if its substation sits in a flood- or heat-exposed location — the SSI Index’s own v4.2 R6c/R6d/R6e physical modifiers — then W4 Climate Resilience degrades and the gate returns FAIL, with the €11,686/MW-year of §4.5 value entirely unchanged. The value does not license the siting. And the sign runs the other way too: in a deficit, fossil-marginal zone a battery that charges on a fossil-marginal grid raises emissions, so the gate fails on W2 Carbon — the map’s own signed geography (§4.2b) driving the carbon axis. The gate is therefore non-trivial in both directions; Fig 6’s direction-only illustration is, in Stage 4, a concrete per-axis computation the demonstrator emits for any candidate node.²⁴ Opinable T2

Two roles compose cleanly here and are worth separating. The SSI Index measures the per-substation community-value floor and runs the anti-maladaptation gate — the public, governance layer. The SSI Enhanced Neural Network measures the monetised uplift a battery creates on top of that floor — the private valuation layer. The Index measures whether an intervention is admissible; the valuation measures how much value an admissible intervention creates. The locational-value map this brief validates is the input to the second; the gate is the discipline on the first; and a deployment is worth doing only when it both creates uplift and protects every floor.

6.2 — The boundaries, stated plainly

Four honest limits bound the claim, and naming them is itself part of the discipline. First, the brief produces the locational signal, not the full equilibrium outcome: it reveals where the suppressed value is and points directionally at curtailment relief, capex deferral and resilience, but a fully-quantified “therefore X gigawatts re-site and €Y is deferred” needs a capacity-expansion layer this brief does not run. Second, resilience is under-monetised even by a perfect nodal arbitrage price — energy arbitrage under-prices the option value of local capacity under contingency, which is why capacity mechanisms exist — so the locational signal is necessary but not sufficient; the brief claims that zonal is strictly worse, not that nodal captures all the value. Third, the distributional politics are real and are the true seat of regulator hesitation; the transition-and-resilience public-good framing helps, but any honest deliverable must carry the distributional truth and name who bears the adjustment. Fourth, the validated measurement is at the transmission node today; the distribution layer, where the value is largest, is anchored to observed physical and administrative data (Tier 1.5) and to engineering models (Tier 2), never yet to a cleared price — and the brief never lets the distribution figures borrow the transmission measurement's credibility.

On that fourth boundary the SSI Index has a specific institutional role. Regulators themselves reach for the index form when a full nodal price is out of reach: the CEER has listed “archetypical” locational network tariffs — classifying locations into representative types — alongside nodal and zonal options, which is conceptually an index.²¹ California's Integration Capacity Analysis maps hosting capacity at the feeder, which is an observed physical index. The SSI Index — substation-granular, open-methodology, peer-reviewed, and already operating across thirty-nine countries — is the Tier-1.5-to-Tier-2 bridge for a European grid that has no public hosting-capacity map, and it is methodologically aligned with the regulators' own archetypical option rather than a workaround to it.²²

The play, then, arrives where it began: the grid keeps a locational price, the prevailing markets decline to read it, and the cost of the silence lands on exactly the decarbonisation and resilience objectives Europe has bound itself to. The instrument to read the signal is buildable and partly built; the governance test to keep the reading honest already exists; and the asset that turns the signal into value is the one whose mis-siting the test can already flag as maladaptation. The reader is left, as a war game intends, to weigh which door to open.

Conclusions

Read the price the grid already keeps

The arc closed end-to-end — measured, mapped, valued, governed — and the forward pathway that does not require going nodal.

The locational marginal price is not an American peculiarity; it is a property of any constrained network, and the European grid is increasingly constrained. Zonal pricing does not make the property go away — it makes it invisible, and resolves it after the fact through socialised redispatch that discards the information a price would have published. This brief has shown that the suppressed information is measurable where nodal prices clear, that its structural core transfers across two independent markets, that it is larger and more place-specific one tier down at distribution, and that its purest component — voltage — is local, un-diversifiable, and priced by no market at all.

The first European calibration is now done (§4.2b): on a full year of Italy's seven observed zonal prices, a single national price is shown to hide a large but episodic locational signal — €23–114/MWh on the congested tail, material on a quarter of hours — with a clear signed geography (North dear, South and islands cheap, Sicily most volatile). That measurement then extends in both the directions the brief promised. Wider, into the map (§4.2c, Stage 3): where a country runs a single national zone, the suppressed value surfaces only as billions a year of socialised redispatch, and the compass is switched off hardest — Germany, Great Britain — exactly where the bill for keeping it off is largest. And into value and governance (§4.5 and §6.1, Stage 4): even the least-locational stream, energy arbitrage, is worth capturing only where the map says so — measured on the observed year in two markets as a large signed gradient (Italy positive in the cheap, volatile South ~€11,700/MW-year and negative in the firm North; the Nordics from +€54,900 in East Denmark to −€36,900 in southern Norway), invisible under a national price in both, and reached by opposite mechanisms — curtailment in Italy, hydro-versus-continental volatility in the Nordics — and whether the battery should actually land there is settled not by the value but by the per-axis anti-maladaptation gate, which the demonstrator now runs as a rule (a surplus epicentre passes; a physically-exposed or fossil-marginal node fails, value unchanged). What still remains is the honest next step the Italian measurement itself argued for: finer still, from seven zones toward the per-substation surface where the switch becomes a continuous fan, and deeper, into the intra-zonal residual a seven-zone split still hides. The magnitude-and-geography leg rests on direct European observation, the predictive leg on the transmission-node evidence, and both point the same way.

The forward pathway does not require Europe to go nodal. It requires Europe to act on the locational signal it can already estimate — to steer where storage, flexibility and reinforcement are sited and rewarded by the best available map of where the value is, and to run every such siting through a per-axis anti-maladaptation gate so that the capital lands where it leaves no community worse off. The correctly-sited battery is the demonstrator because it is the asset whose value most tightly is the locational signal; but the argument is general, and it is the SSI Index's argument: measure the grid at the resolution at which its value actually varies, and let the measurement guide the capital. The grid already prices location. The task is to read it.

Annex

Data sources, methodology pins, and the evidence-tier register

Every quantitative claim in this brief traces to one of the rows below; the machine-readable equivalent is the sidecar manifest.

Register legend. Measured validated out-of-sample on data · Probable peer-reviewed literature applied to the case · Opinable methodologically-disciplined inference where reasonable analysts may differ · Speculative explicitly hypothetical · Modelled a stated engineering assumption, not a measurement.