The Referee Can't Keep Score

The question I keep coming back to after the first three issues of Grid Signal is a simple one: if the verification gap is real, why hasn’t someone filled it?

Not a regulation. Not a working group. Someone. A company, a consortium, a standards body, anyone. The gap has been visible for years. Flexibility markets are growing fast — ACER submitted a proposed Network Code on Demand Response to the European Commission in March 2025, smartEn’s 2024 Market Monitor tracks steady growth across member states — and yet the neutral verification layer that would resolve settlement disputes does not exist.

I spent some time trying to work out why — not as a researcher, but as someone who has spent a long time watching how digital infrastructure gets built and how it doesn’t. What I found was less surprising than I expected. Every candidate for the role has a structural reason not to play it.

Who you’d expect to build it — and why it doesn’t quite work

The obvious starting point is the TSO. They run the market, they dispatch the flexibility, they settle the payments. If anyone has the data and the authority to run a neutral verification layer, it’s them.

Except that’s exactly the problem. The TSO is not just a market operator — it is also a settlement counterparty. When a flexibility event is disputed, the TSO has a position. That doesn’t make TSOs bad actors. It just means they can’t be neutral on their own outcomes. The referee can’t also keep score.

You’d think the DSO might fill part of the gap — they host the assets, after all. But the DSO’s visibility into a flexibility event is partial. They see what happens at the meter. They don’t see the commitment the aggregator made, or the full settlement logic that determines whether it was honoured. The verification layer needs all three pieces together. The DSO has one.

Aggregators are the most motivated parties to solve this. They’re the ones absorbing disputed settlements. But self-verification is a contradiction in terms — you can’t be the counterparty to every dispute and also the neutral voice on whether the dispute was resolved fairly.

Metering companies look like a candidate from the outside. They’re already embedded in the measurement stack. But they’re commercial vendors, contracted and renewed by the network operators they’d be auditing. That relationship is hard to reconcile with genuine independence.

ENTSO-E and national regulators write the rules. They don’t run infrastructure. There’s a reason the European Central Bank doesn’t operate SWIFT.

A new vendor could build the technology. But technology is the easy part. The hard part is neutrality — and a private company can’t confer that on itself by claiming it.

I don’t think anyone designed this situation deliberately. It’s more that flexibility markets were built without this seat in mind, and now that the markets are growing, the absence is becoming visible.

What I noticed when I looked at how other markets solved this

When I started looking at how other industries ended up with neutral market infrastructure, I expected to find regulatory mandates. I mostly found the opposite.

The DTCC — the central depository that settles US securities trades — exists because the securities industry drowned in a paperwork crisis in the late 1960s. NYSE volumes tripled to around 15 million shares a day by April 1968, and trades were failing to settle because the back offices could not physically reconcile the certificates. The SEC Historical Society’s account describes how the fix emerged from the industry itself — a Central Certificate Service inside the NYSE that eventually became the Depository Trust Company. The SEC formalised it years later. The regulator ratified; it didn’t initiate.

SWIFT started in 1973 as a cooperative between 239 banks across 15 countries. They needed a better way to communicate payment instructions than bilateral telex, which was expensive and error-prone. No regulator asked for it. Banks built it because the alternative was worse. National and supranational regulators adopted SWIFT as de facto standard in the decades that followed.

Let’s Encrypt is more recent. The project was announced publicly in November 2014 by the EFF, Mozilla, the University of Michigan, and a small group of academic collaborators, and began issuing certificates in late 2015. Free, automated TLS certificates. No regulator required it. But once adoption crossed a threshold, browsers started treating unencrypted connections as warnings, and HTTPS moved from optional to effectively mandatory. The political pressure came after the working reference existed.

The pattern that keeps emerging is something like: a credible group builds a working reference because the pain of not having it is high enough. Early users adopt it because it’s better than the alternative. Once enough parties are running it, the cost of non-adoption flips — it becomes easier to join than to stay out. Regulators formalise what the market has already converged on.

None of these were planned. None started with a mandate. They started with a concrete problem and a group willing to build before anyone asked them to.

I’m not sure how far that pattern extends. But I find it hard to look at the flexibility market and not see something structurally similar waiting to happen.

Why it hasn’t happened yet in energy

The thing that’s different about energy — so far — is where the pain is landing.

In 1968, the paperwork crisis was killing large US broker-dealers. They had the money, the lobbying reach, and the political access to do something about it.

In flexibility markets, the parties absorbing the verification gap most directly are aggregators. Most of them are small. I don’t have hard data on government-affairs headcount across the sector, but from what I’ve seen and heard, politically they are almost invisible compared to incumbent utilities. They absorb disputed settlements as a cost of doing business because withdrawing from the market isn’t a realistic option.

The end users — industrial customers whose flexibility is dispatched but whose settlement is opaque — have even less voice. They negotiate bilaterally and rarely coordinate.

The TSOs and DSOs, who have the political weight to force a change, are not the ones losing. They run the systems whose opacity creates the gap.

That’s not a criticism of anyone specifically. It’s just a structural observation about where the pressure is and isn’t. The political conditions for a neutral to emerge don’t exist yet, because the parties carrying the pain can’t compel the parties with the power to act.

That will probably change. But it’s worth being honest that it hasn’t yet.

What I think the cloud architecture lens adds — cautiously

I want to be careful here, because infrastructure analogies can make things sound more solved than they are.

But there’s something in how internet infrastructure was built that feels relevant. BGP, DNS, the certificate authority system — none of it was legislated into existence. All of it started as a working proposal that enough parties ran in production that it became de facto standard. The governance structures came later, built around the running code.

The reason that matters is that the alternative — waiting for a central authority to specify and mandate a specific technical design — has a poor track record for this kind of problem. The specifications get contested. The deployment is too fragmented. The political consensus needed to mandate a specific design doesn’t form without a reference to point at.

What the flexibility market would need isn’t new technology. The primitives — signed records, reproducible audit logs, verifiable timestamps — are decades old. What it would need is a working reference: something running in production, on real data, that demonstrates the approach is viable and that other parties can implement against a published specification.

I don’t know exactly what that looks like. That’s what I’m trying to work out.

The signals I’m watching

Two things would shift the situation faster than anything else.

The first is a significant aggregator — not a startup, a tier-one player — publicly stepping back from a market because settlement disputes have become commercially unmanageable. That hasn’t happened. When it does, it stops being a small-firm complaint and becomes a visible market participation problem that regulators can’t quietly ignore.

The second is a greenfield flexibility market somewhere outside Europe — the Gulf, Singapore, parts of the US — where third-party verification is built in from the start. If that market outperforms European equivalents on dispute rates or settlement timelines, the pressure reverses. European aggregators start asking why their home market doesn’t work the same way.

Neither has happened yet. But both feel like they’re forming.

The less dramatic signal — the one I find more interesting to watch right now — is at the contract level. Large industrial flexibility providers are starting to demand independently verifiable settlement data in their bilateral contracts. Not because regulators require it. Because the payment volumes are large enough that unverifiable settlement is a commercial risk they won’t accept.

Data centres, electrolysers, cold storage operators. When one of them insists on signed event logs as a condition of signing a flexibility contract, and a vendor ships them to win the deal, and another industrial customer sees it and asks for the same — that’s the SWIFT pattern in embryo. Not a standard. Just something a few large parties needed, that others then had to match.

I don’t know how fast that moves. But it’s the most concrete signal I’ve seen that the conditions for something to emerge are forming at the edges of the market, before the centre has decided to act.

What’s coming in #005

If a working reference is what the market eventually needs, the next question is what it would actually have to do.

Not the technology. The properties. What does a flexibility settlement record need to look like for a regulator, a disputing aggregator, and a sceptical industrial customer to all be able to trust the same number? What can it afford to be uncertain about, and what can’t it? Where are the real edges of the problem?

That’s what I want to sit with next issue. Not a proposal — a specification of the question.

Sources and further reading

- DTCC origin story — SEC Historical Society: The History of Securities Depositories; Depository Trust Company on Wikipedia; Optimizer Online: The Paperwork Crisis

- SWIFT founding — Swift: Our Story; Scott & Zachariadis, LSE: Origins and Development of SWIFT, form 1973

- Let’s Encrypt — Let’s Encrypt; EFF: Let’s Encrypt Brings Free HTTPS to the World (2015)

- EU flexibility regulation — ACER proposal for Network Code on Demand Response (March 2025); European Commission consultation; smartEn Market Monitor for Demand Side Flexibility 2024

---

Grid Signal covers the intersection of grid modernisation, distributed trust, and digital infrastructure for energy. No hype — just structural insight for practitioners.

If a colleague would find this useful, forward it.