Guarantees of Origin

Why support at all

Two arguments have carried every scheme. The first is the learning curve: deployment drives costs down, so early capacity is worth paying for beyond its energy value. The second is revenue: in an energy-only market a technology with zero marginal cost erodes its own price as it scales, so the more of it there is the less each unit earns, and the missing margin has to come from somewhere. Every mechanism below is a way of supplying that margin. What distinguishes them is not whether they pay but who carries the price risk, how the level of support is discovered, and what they do to the plant's incentive to respond to the market once it is built.

Price-based instruments

A feed-in tariff pays a fixed price for every megawatt-hour and takes the plant out of the market altogether: bankable, simple, and blind — the operator has no reason to care when it produces, and the tariff-setter carries all the risk of getting the level wrong. A feed-in premium puts the plant back in the market and adds a fixed amount per megawatt-hour on top; the operator now sees prices, but a fixed premium also means it profits from producing until the price falls below minus the premium, which is the mechanism behind the deepest negative bids. A sliding premium, or one-way contract for difference, tops the market price up to a strike and lets the operator keep any upside; a two-way contract pays the shortfall below the strike and claws back the excess above it, which is cheapest to finance because the revenue is nearly fixed. The EU has now legislated for that last form: since the 2024 market-design reform, direct price support for new wind, solar, geothermal, reservoir-less hydro and nuclear must take the form of a two-way contract for difference or an equivalent from 17 July 2027, 2029 for offshore, with Commission guidance on their design published in December 2025. State-aid rules since 2022 add a condition of their own: no support is paid for hours in which the price is negative.

Quantity-based instruments: green certificates

The alternative is to fix the quantity and let the market find the price. A quota obliges suppliers to source a rising share of their sales from renewables and to prove it by surrendering certificates, one per megawatt-hour, issued to eligible generators and freely traded; a supplier that falls short pays a buy-out price, which caps the certificate's value. The support level is whatever the certificate fetches, so it is set by competition among generators rather than by an administrator — the scheme's central promise — and, in principle, capped by the buy-out. In practice the price swings with the balance between the quota and the build rate, and that volatility feeds straight into the cost of capital, which is why certificate markets produced dearer renewables than the tariffs they were meant to improve on. Technology-neutrality also proved unwanted: the UK's Renewables Obligation had to be banded by technology in 2009 to build anything other than the cheapest option. The UK closed the RO to new capacity in 2017 in favour of contracts for difference; Sweden and Norway's joint certificate market stopped admitting new plant after 2021 and winds down to 2035; Poland moved to auctions in 2016; Italy retired its certificati verdi in 2015. The regional Belgian schemes and the American renewable portfolio standards, with their tradable RECs, remain the largest survivors.

How the level is set: auctions

Whatever instrument pays, something has to set its level. Administrative setting gave way, across the 2010s, to competitive auctions in which developers bid the premium or strike they need, and EU law has required market-based, competitive allocation since 2021. Auctions discover cost and stop overcompensation; they also invite underbidding, so their realisation rules — deposits, deadlines, penalties for projects that never appear — are as much a part of the design as the price they produce.

Where guarantees of origin fit

A guarantee of origin is not a support scheme, and the standard error in this field is to treat it as one. Both a GO and a green certificate say "one megawatt-hour of renewable electricity was produced"; the difference is what the saying is for. A green certificate is surrendered to meet a legal obligation, and its value is the cost of that obligation. A GO is cancelled to substantiate a claim — a supplier's disclosure, a corporate's consumption — and its value is what a buyer will pay for the claim. Most countries keep them in separate registries under separate rules. Support design nonetheless shapes GO supply directly, because states decide who receives the GOs of a plant they have paid for: Germany issues none for output under its feed-in scheme, to prevent the same electricity being sold as green twice; France retains them and auctions them for the state; the Netherlands lets supported plant sell them. As support migrates toward contracts for difference and corporate power purchase agreements, the GO is becoming the instrument through which renewable attributes are actually priced by demand rather than by a scheme — which is what the rest of this page is about.

Issuance

National issuing bodies grant one GO per MWh of certified production, carrying attributes: technology, installation, commissioning date, country, production period and support-scheme status. The attribute set is the product — everything the market later prices must be recorded here or is lost.

Practice diverges sharply from principle: in 2023 roughly 90% of European hydro output received GOs, but only ~54% of wind and ~34% of solar — capped by states that deny issuance to subsidised plants and by administrative burden that prices small rooftop PV out entirely (registry fees can differ seventy-fold per certificate between a small wind farm and a large hydro plant). Nuclear GOs are the fastest-moving frontier: issuance nearly doubled to 60 TWh in 2022 as Finland and Slovenia joined, and France began issuing against its ~400 TWh fleet in 2023.

Transfer and trade

GOs move between registry accounts independently of the physical electricity — the deliberate design choice that makes the system work across borders, and the feature critics call 'unbundling'. Trade happens bilaterally, through brokers, and bundled inside power purchase agreements.

The trade map has a stable shape: Norway is by far the largest exporter (hydro), Sweden second; Germany the largest importer — its Doppelvermarktungsverbot denies GOs to subsidised plants, so only ~10% of German renewable output is ever certified — with Ireland's data-centre demand close behind. Italy flipped from exporter to importer after a syngas mis-issuance scandal and weak hydro years. Almost all volume moves over the counter through brokers and traders; about 60% of participants reach the market only through such intermediaries, which is where market power quietly concentrates.

Venues, clearing and delivery

The route a trade takes decides who bears the risk between agreement and delivery, not what delivery is. Most volume is bilateral and brokered, or bundled inside power purchase agreements, and settles when the seller instructs its registry to move the certificates — until then each side carries the other's credit. EPEX SPOT runs pan-European GO spot auctions twice a month, differentiated by country, technology and support regime, and EEX lists GO futures in lots of 1,000 certificates on six January maturities; both are cleared by ECC, which steps in as counterparty to each side, margins both, and on the settlement day instructs the registry operator to move the certificates from the seller's account through its own to the buyer's, at home or in another EECS registry via the AIB hub. State-held GOs from supported production enter through auctions of their own, France's being the largest. Whatever the venue, the certificate only moves when the registry says so: clearing interposes a guarantor, it does not create a second kind of delivery.

Cancellation and disclosure

Value is realised by destruction: a supplier cancels GOs to substantiate the renewable share disclosed to consumers; a corporate cancels them for its consumption claims. A GO that is never cancelled never proved anything — cancellation statistics, not issuance, measure real demand.

Demand is strikingly price-inelastic: November 2022 set a monthly cancellation record of 36 TWh at the very peak of prices, because disclosure deadlines and supply contracts do not renegotiate themselves. Who may cancel is unharmonised — seven domains restrict cancellation to suppliers, while Portugal, Lithuania and Latvia let end-consumers cancel directly — a quiet design fork that decides how traceable a corporate claim can be.

Expiry

Certificates expire (the EU framework works on a twelve-month usability horizon after production, with registry-specific implementation details). Expiry forces temporal honesty at annual scale and creates the characteristic year-end cancellation rhythm of the market.

In practice 5–8% of issued GOs simply expire unused each year — trader buffer stocks, tactical withholding, or fees exceeding residual value. Most states enforce twelve months; Ireland, Italy, Lithuania and Portugal allow eighteen, enabling vintage arbitrage. Italy’s issuer showed what state inventory can do: it withheld auction volumes through the 2023 price peak, then released 14 TWh in March 2024 — near-expiry lots clearing at 0.15 €/MWh — helping crash the market back to pre-crisis levels.

Residual mix

Whoever claims renewable attributes removes them from everyone else. The residual mix — the attribute composition left for consumers without cancelled GOs — is the accounting device that prevents the same green megawatt-hour being claimed twice. Its annual calculation is the system's integrity check.

The result is counter-intuitive wherever exports are large, which is why the selector below is worth moving a few times. Norway generates almost entirely hydro, yet roughly three-quarters of Norwegian consumption is untracked, and the mix assigned to it once the certificates have been sold abroad is 62% fossil at 431 gCO₂/kWh. France, with 82% untracked, keeps a residual mix that is 86% nuclear at 17 gCO₂/kWh. Germany is the opposite case for a different reason: its subsidised EEG volumes are disclosed to consumers through national rules rather than through certificates, so only 3% of consumption falls to the residual pool at all — and what remains in that pool is 1.6% renewable at 701 gCO₂/kWh.

None of those numbers describes the electrons on any wire. They describe what is left to claim once the claims have been made, which is exactly the job the calculation exists to do. It is also where double counting becomes visible in a single figure: a Norwegian company reporting its physical hydro mix under Scope 2, rather than the 431 grams the residual calculation assigns it, is claiming a megawatt-hour that has already been sold to someone else — the seam the defences below are aimed at.

Double-counting defences

One production device, one registry, one certificate per MWh; support-scheme flags where states retain or auction GOs for subsidised production; audited issuance data. The system's credibility is exactly as strong as its weakest issuing body.

The registry architecture itself holds; the live leak is at the accounting seam. Most Norwegian companies report location-based Scope 2 figures — claiming the physical hydro mix — while Norway simultaneously exports those same attributes as GOs: one green megawatt-hour, effectively claimed twice, entirely legally. The reactive defences — Germany’s 2023 ban on Icelandic GOs, Flanders’ refusal of Swedish ones, Italy’s syngas clean-up — show enforcement working case by case rather than by design.

AIB and the EECS rules

The Association of Issuing Bodies maintains the European Energy Certificate System — the common rulebook that harmonises issuance criteria, registry operation and certificate content across member countries. EECS is what turns two dozen national schemes into one market.

The AIB hub

Cross-border transfers settle through the AIB's central hub connecting national registries. Membership of the hub — and occasionally suspension from it, as past import controversies have shown — is the practical boundary of the integrated European GO market.

National issuing bodies

Each country designates an issuer and registry operator (regulator, TSO or agency). Implementation differences — auctioning of GOs from supported production, issuance for all generation versus renewables only, fees and account rules — are where the harmonised system stays stubbornly national.

The legal foundation

The Renewable Energy Directive establishes GOs as the exclusive instrument for proving renewable origin to final consumers, with mutual recognition between member states as the default and refusals requiring justification. Each directive revision has tightened standardisation and expanded scope (gas and hydrogen GOs alongside electricity).

Beyond the EU perimeter

Certificate systems exist worldwide (I-RECs, national schemes); recognition into EU disclosure is restricted. The interaction of GO markets with hydrogen certification and carbon accounting frameworks is pulling the institutional architecture into new territory.

Brexit ran the controlled experiment: after mutual recognition ended, the undersupplied British REGO market decoupled and spiked from under 1 €/MWh to 24.40 €/MWh by October 2023 — the clearest demonstration on record of what structural undersupply does to certificate prices, and the mirror image of the oversupplied continent. Meanwhile the Energy Community is extending the system southeast: Serbia already issues far more than it cancels and exports the surplus — enlargement, on current design, adds supply faster than demand.

Gas, hydrogen and the Union Database

Gaseous GOs are a different regime wearing the same name. Biomethane production reached 49 TWh in 2023 but only 2–3% of certificates trade across borders, and prices near 18 €/MWh track the gas market rather than the roughly 110 €/MWh cost of production; electrolytic hydrogen volumes are still negligible. For renewable fuels the architecture changes outright: RFNBO certificates must travel bundled with proofs of sustainability inside the Union Database — mandatory for gaseous fuels since late 2024 — and are not tradeable outside it, a certification design closer to fuel-compliance accounting than to the free-floating electricity GO.

Verification and audit

Registries reconcile issuance against metered production; auditors test device registration and measurement chains. As corporate claims based on GOs face growing scrutiny, audit depth is shifting from formality to front line.

Supply drivers

Issuance follows renewable build-out and weather — hydro reservoir years move Nordic supply materially. Policy choices matter as much: whether states issue, retain or auction GOs for subsidised fleets decides how much certified volume ever reaches the market.

The arithmetic is structural: issuance has grown ~74 TWh a year against ~59 TWh of cancellations, so the surplus widens with every renewables auction. Weather moves the residual — the 2022 drought removed 55 TWh of hydro GOs, almost exactly offset by new wind and solar — and policy moves it more: state auctions of subsidised-fleet GOs supply 5–15% of the market in discrete lumps, 16 TWh in March 2024 alone.

Demand drivers

Supplier disclosure obligations set the floor; voluntary corporate demand — renewable targets, Scope 2 accounting, customer pressure — sets the price. Demand concentrated in specific technologies, vintages and countries has fragmented what was once a single price into a quality spectrum.

The demand stack is regulatory sediment: supplier disclosure at the base, then the EU Taxonomy, CSRD sustainability reporting (market-based accounting for tens of thousands of companies), state-aid conditions in seventeen member states, data-centre reporting duties, renewable-hydrogen rules that require cancelling matching GOs, and green tariffs covering a third of German and over half of Dutch households. The crisis revealed its texture: when renewable GOs got expensive, large buyers quietly downgraded claims from “renewable” to “carbon-free” and absorbed cheap nuclear certificates — nuclear cancellations jumped 36 TWh in a year.

Structural oversupply and the price record

The market’s defining fact is surplus: issuance exceeds cancellation by roughly 100 TWh a year, and the cumulative overhang reached 464 TWh by 2024. That is why GOs spent 2019–2021 below 1 €/MWh — and why the one episode of perceived scarcity was so violent. Through 2022 the drought, hydro producers buying back forward-sold certificates and crisis-era willingness to pay drove monthly averages to 8 €/MWh by November 2022 (daily prints near 10), briefly making GOs 3–4% of the power price instead of a rounding error. The overhang then reasserted itself: prices decayed back below 1 €/MWh by late 2024. The lesson generalises — against inelastic, deadline-driven demand, certificate prices are either negligible or explosive, with little in between; Britain’s post-Brexit REGO spike to 24 €/MWh is the same physics run in reverse.

Price differentiation

The market prices attributes: Nordic hydro versus continental solar, new-build versus depreciated assets, domestic versus imported. Differentiation is the market saying that not all green megawatt-hours support the same claim — the design question is how much of that the certificate should encode.

In practice differentiation is thinner than the theory: technology spreads on the wide market run below 0.16 €/MWh, because hydro — still more than half of issuance, its capital long depreciated — anchors the whole complex the way Brent anchors crude. The premia that do appear are local and demand-driven: Dutch wind traded at five times the European price in 2021 on domestic-claim preference. The structural consequence: as long as depreciated hydro sets the margin, GO prices cannot say anything about the cost of new wind or solar.

Temporal matching

Annual matching lets a solar GO cover midnight consumption. The granular-certificate agenda — hourly matching, pioneered for 24/7 procurement and hardened by the EU's renewable-hydrogen rules requiring temporal correlation — would transform GOs from annual accounting instruments into something approaching a delivery obligation. It is the sharpest live debate in the field.

France already runs the experiment at monthly granularity — its energy code requires month-of-production matching for disclosure, and its price volatility is measurably higher for it. The hydrogen rules go further, importing hourly correlation into the certificate world for RFNBO producers. The report’s caution is worth keeping: granular certificates prove ex post that consumption matched production, but an after-the-fact certificate cannot by itself call flexible investment into existence up front.

Support-scheme interaction

A GO auctioned by the state alongside a subsidised asset raises revenue but severs the certificate from any additionality narrative. Whether GO revenue should flow to producers as an investment signal or to states as subsidy recovery is a quiet fight over what the instrument is for.

GOs inside PPAs

Corporate PPAs bundle certificates with energy, making the GO the legal carrier of the buyer's renewable claim. As PPA volumes grow, GO design stops being a disclosure technicality and becomes part of the investment case for merchant renewables.

The numbers say the two markets are siblings, not twins: European PPAs covered ~217 TWh of generation in 2024 and effectively all off-site PPA volume carries GOs — yet the merchant GO market is at least three times PPA volume, and GO and PPA prices show no correlation. The PPA is priced by the electricity market; the GO rides along as the claim’s legal carrier, not its value.

Two families, one instrument

Every certificate scheme in the world falls into one of two families, and confusing them is the commonest error in this field. In the first, the certificate is a disclosure instrument: it proves an attribute so that a consumer can make a claim, demand is voluntary, and the price is whatever corporate buyers will pay — near zero until they arrive, and volatile once they have. European guarantees of origin, British REGOs, Turkish YEK-G and the international I-REC all sit here. In the second, the certificate is a compliance instrument: someone is legally obliged to surrender it, demand is created by statute, and the price is a policy variable rather than a market discovery. American renewable portfolio standards, Korean RECs, Indian RECs, Australian large-scale generation certificates and now China's green electricity certificates all sit here.

The distinction decides how the market behaves, which is why the price-versus-quantity diagram above applies to the second family and not the first. A compliance certificate is a quota instrument: the state fixes volume and the certificate price floats to whatever clears, so the certificate is the subsidy and its collapse is a support-scheme failure. A disclosure certificate carries no obligation, so its price measures corporate willingness to pay for a claim — and if that price goes to zero, nothing about the renewable plant's economics changes. Europe runs the largest disclosure market and the smallest compliance one. Most of the rest of the world does the opposite.

I-REC, and the limits of a private standard

The I-REC electricity code did something states had not: it created a tracking system where none existed. Issuance runs in more than sixty countries, buyers sit in more than a hundred and fifty, and the code is administered by Evident against the I-TRACK Foundation's standard — a private rulebook doing a public job, and doing it well enough that corporate procurement in emerging markets became possible at all. Volumes tell the story of that success and of its ceiling: 283 TWh issued in 2023, 42% up on the year before; 188 TWh in the first half of 2024, growth down to 17%; roughly 227 million certificates in the first seven months of 2025, growth down again to 12%.

The deceleration is not market fatigue. It is sovereignty. China had been among the largest I-REC markets, and in July 2023 the National Development and Reform Commission, the finance ministry and the energy administration issued a full-coverage notice extending green electricity certificates to essentially all renewable generation, superseding the 2017 pilot; issuance and trading rules followed in August 2024; and Beijing designated the GEC as the sole instrument for tracking renewable consumption in China, with I-RECs leaving the market. Close to five billion GECs had been issued by the end of 2024. The lesson generalises beyond China: a private standard operates at the pleasure of the state whose electricity it is tracking, and the moment a government decides that tracking is a public function, the private instrument has no standing to object. Certificate infrastructure is only as international as national policy allows.

What most systems lack: a residual mix

The single largest difference between the European system and almost every other is the step Europe takes at the end of the year. Because Europe calculates a residual mix, an attribute that has been claimed is removed from what everyone else may claim. Almost no other jurisdiction does this. Where there is no residual mix, a consumer who buys no certificates falls back on a location-based grid average that still contains the attributes already sold to somebody else — so the same megawatt-hour supports two claims, and not because anyone cheated. Double counting is the arithmetic, not the abuse.

This reframes the Norwegian seam described earlier. Norway looks like the problem case because Europe measures it; a market with no residual calculation has the same leak and no instrument that would show it. Any jurisdiction serious about the integrity of certificate claims has to build the unglamorous half of the system, and the unglamorous half is the annual accounting, not the registry.

The claim standard is the real market rule

What ultimately prices every certificate on earth is not the scheme that issues it but the accounting standard that decides what it lets you say. That standard is under revision now. The GHG Protocol put a redraft of its Scope 2 guidance to public consultation from October 2025 to the end of January 2026, drawing more than four hundred responses, with a revised draft expected later in 2026; among the proposals are hourly matching between generation and consumption, deliverability boundaries confining a claim to a grid the electricity could physically reach, additionality requirements, and quality tiers ranking instruments against one another.

Any one of those would reprice certificate markets everywhere at once. A certificate that cannot be matched to the hour of consumption, or that comes from outside a deliverable boundary, would stop supporting the claim it currently supports — and the certificate's value is entirely derived from that claim. This is the largest single source of demand risk in the whole field, larger than any national policy decision, and it rests with a private standard-setter rather than a legislature. For a European reader that is a useful corrective: the design questions on this page are not parochial. Every jurisdiction that builds a certificate market arrives at the same four, in the same order — what the certificate proves, who is obliged to buy it, what happens to the remainder, and who decides what a claim means.

One certificate, one attribute

A Guarantee of Origin records the attributes of a defined quantity of energy. Issuance is separated from the physical electricity flow, allowing certificates to be transferred independently.

Lifecycle

Issuing bodies create certificates from verified production. GOs may be transferred domestically or through connected registries and are cancelled when used for disclosure or a renewable claim.

Disclosure and residual mix

Cancellation prevents double claiming. Untracked electricity attributes are allocated through the residual mix so that disclosed renewable consumption remains consistent with certificate use.

Price differentiation

Technology, country, production period, plant age, support status and consumer preferences create differentiated products rather than one uniform European GO price.

Trading venues

Primary sales and auctions coexist with bilateral and brokered secondary trading through participants such as STX, ACT and enmacc. EPEX auctions and EEX futures add organised-market reference points.

Market-design question

The central issue is additionality: whether certificate revenue materially changes investment or operating decisions, and how disclosure rules translate consumer willingness to pay into producer value.

The Commission’s 2025 monitoring study renders the verdict in numbers: GOs work as a tracking and reporting instrument, and fail as an investment signal — roughly 23% of the certificate price goes to fees and 12% to administration, leaving producers ~40%, about €140 million a year across all of Europe against tens of billions invested annually in renewables. Its 27 reform options pull in opposite directions — granular certificates, full disclosure, new-build-only tradability, abolishing book-and-claim — which is the report’s real conclusion: the EU must first decide what the instrument is for, tracking or steering, before it can sensibly redesign it.

Figures on this page draw on the European Commission’s GO-system monitoring study (LBST, Öko-Institut & Trinomics for DG ENER, June 2025).