Release 19 07 May 2025 Status: stable
Any pan‑European energy information agency based on open data and crafted on the United States Energy Information Administration (EIA) will face a number of legal challenges.
The initial proposal for such an agency is contained in McWilliams et al (2025), published by the Bruegel AISBL think tank based in Brussels. Their proposal will be debated at a hybrid panel discussion on 8 May 2025 entitled “Europe’s energy information problem”.
This posting outlines some of the intellectual property issues that I believe will need to be confronted if and as this proposal moves forward.
Key considerations
These legal issues arise primarily because intellectual property law in Europe — and note that, when necessary, I will fall back to German law as a reference — is very different from that in the United States. Specifically in Europe:
-
the presence of copyright in datasets cannot be excluded — whereas in the US, the landmark 1991 Fiest telephone book ruling clarified that collections of observational data cannot attract copyright
-
so‑called sui‑generis database rights, designed to protect substantial investments in database infrastructure from substantial extractions, can attach automatically under both EU member state and United Kingdom law — while the USA has no equivalent provision
-
unlicensed public sector information (PSI) in the European Union in general (although some member states, Italy for example, have specific national open data provisions) remains legally encumbered and falls short of the requirements of open science, in fact the current PSI legislation allows only for “use”, a concept not elaborated in statute (and also not a common term of art within intellectual property law) — whereas in the United States, outputs from US federal employees are accordingly public domain
-
European energy sector information published under statutory reporting mandates often remains legally encumbered or of uncertain legal status by default — the opposite being true in the US
-
the US is a single country with one federal law covering intellectual property, whereas the European Union is a set of member states with different laws and traditions (for instance, ENTSO‑E tells me Swedish law prevents some of their national energy data from being published under Creative Commons licensing)
-
the use of Creative Commons
CC‑BY‑4.0to license information made public is becoming more common in Europe but is far from universal — while the European Commission did elect to endorse this instrument in a 2019 decision (JRC 2019a, 2019b) -
the development and application of semantic standards that span both data and code will require specialized licensing (perhaps a task for the Creative Commons organization)
Data copyright, database rights, and the related legal uncertainties hinder or prevent the public dissemination of data and thereby frustrate efforts to utilize this information collaboratively.
The best way forward, in most cases, is to push for CC‑BY‑4.0 licensing on primary data and CC0‑1.0 licensing on metadata — and thereby gain the legal surety that these instruments afford. Corner cases in which national law and/or member state policies prevent such licensing will need to be carefully worked around — and law changes might be necessary.
I am fully in favor of the idea of a pan‑European agency for energy information. But its enabling legislation will probably need to carve out some special rights and exceptions in relation to intellectual property law.
As I understand it, the goal of the new agency is to provide comprehensive energy sector data held in trust for the common good. One clear use‑case is to foster a more transparent, flexible, and extensive search for public policies that fulfill both rapid decarbonization ambitions and nominated social objectives such as energy access and affordability. In short, the goal is an expanding information commons.
The solution space for these societal challenges is (hopefully still) large — and having any number of modeling teams working on solutions and trajectories can only help. The status quo is highly problematic in this regard — often based on one or two commissioned modeling teams working with closed datasets, black‑box software, tight instructions, and even prescribed outcomes.
Photo: A derelict power pole reminiscent of the state of usable and reusable energy system data in Europe
United Kingdom
The United Kingdom is included here because the UK participates in the European electricity market. The UK is now also free to diverge from European law on database protection but this seems unlikely.
The UK regulator Ofgem has a policy of presuming “open by default”, albeit duly constrained by security and privacy considerations — and more specifically Ofgem mandates the following (Clark 2023:4):
-
use of the Dublin Core common metadata standard (which includes legal metadata)
-
use of either the most recent version of the Creative Commons License (CC‑BY‑4.0) or the UK Open Government License (OGL‑UK‑3.0) — while the Creative Commons BY variant is not strictly specified in the document, that license is the only sensible choice (and as of mid‑2025, about a 50:50 split of each instrument in practice)
The McWilliams et al (2025) proposal does not include the United Kingdom, but any model of the European electricity system will necessarily need to utilize UK data.
Copyright
Copyright in collections (also known as compilations in the United States) is the relevant provision here. For member states, article 3 of the 1996 96/9/EC database directive (European Parliament and European Council 1996:25) says that databases can be protected under such copyright thus: “by reason of the selection or arrangement of their contents, constitute the author’s own intellectual creation”.
European Commission (2017:10) argues that a mere alphabetical sorting will suffice in this context (odd because the selection is mandated in law and the database itself does not require ordered or sorted inputs):
For the avoidance of doubt, we want to emphasize that the systematic or methodical arrangement does not need to be creative. Even simple structuring as an alphabetical order suffices. The required publication of energy data on transparency platforms as ENTSO-E is structured in a systematic way and meets the criterion of the definition of a database.
Furthermore, any copyright in the database is separate from and without prejudice to the copyright in the entries. In our use‑case, of course, the entries are too trivial to trigger this provision.
For copyright to take hold in other contexts — photography, for instance — a modicum of creativity is required (Gervais 2002), so this database provision is a departure in terms of legal tradition.
The concept of a database is also broad and even paper maps can fall under the directive (Schweizer et al 2015).
Because such copyright is automatic (no claim for copyright is now required), it is the responsibility of downstream users to determine whether or not copyright protection is present. This assessment is often highly conservative because the institutions involved are risk‑averse — leading to the suboptimal social use of available public interest information.
European Commission (2017:10) is again very clear that this form of data copyright could well attach to datasets provided by the ENTSO‑E Transparency Platform (that portal is discussed later under statutory reporting) — while noting that their legal opinion is likely to err on conservative and does not constitute a judicial ruling.
Database rights
The so-called sui‑generis database rights, stemming from article 7 of the 1996 96/9/EC database directive, is especially problematic because the amount of information that can be retrieved by third parties from public portals can easily exceed the threshold of “substantial extraction” and thus be in violation of this right — given a sufficiently exposed investment in infrastructure by the database maker. These trip points are usually not public knowledge and can be manipulated in any case (Davidson 2008). My earlier comment about risk‑averse behavior and suboptimal social usage applies again.
Recent litigation in Bavaria, Germany between the German Society for Civil Liberties (GFF) and the Bavarian state government over energy sector data (including fine‑grained information on building energy performance and on potential sites for renewables) is an example of how database rights can impede public interest usage. Some of the hosted material is paradoxically under CC‑BY‑4.0 licensing, although not enough to override the extant database protection. A German court has ruled that the state of Bavaria must offer individual bilateral licenses upon request — so a legal victory in a sense but not one that leads to open data.
I suspect that there will be more civil litigation over sui‑generis database rights in both directions, going forward.
Semantic standards
The term “semantic standards” is used here to describe casual or formal standards used to inform both data collection and system model design. This terminology is somewhat unusual, but there is no other phrase (that I know of) that includes meanings and excludes technical considerations. Semantic standards may range from relatively simple controlled vocabularies to highly‑formalized ontologies (Hoyer-Klick et al 2022).
The licensing of semantic standards can be problematic, although semantic standards (unlike technical standards) are unlikely to contain patented material. My preferred instrument is (using SPDX syntax) MIT OR CC0‑1.0 — which then respectively defeats patents and database rights, as well as any copyrights in the code and data, to the degree necessary.
Statutory reporting
Statutory reporting refers to public disclosures specified by law. In the energy sector, such disclosures were initially motivated by a desire to improve market transparency and address competition deficits. More recently, that same information is seen as providing new social and commercial opportunities as digitalization takes hold.
An important portal in Europe is the ENTSO‑E Transparency Platform (Hirth et al 2018, ENTSO‑E 2019, 2022). The 2013 legislation establishing the platform (European Commission 2013:3) states in article 2(23) that the: “‘primary owner of the data’ means the entity which creates the data” — but (curiously) does not mention which intellectual property rights might pertain and under what circumstances.
A legal opinion published by European Commission (2017:10) shows that both copyright in collections and sui‑generis database rights can easily apply to the ENTSO‑E Transparency Platform. The more recent CV‑Online Latvia ruling suggests otherwise (Bond 2021, CJEU 2021, Hamilton 2021, Jacobacci et al 2021, Synodinou 2021). An updated legal opinion from the Commission would therefore help.
ENTSO‑E (2022) has made good progress on CC‑BY‑4.0 licensing on the Transparency Platform but it is member states who decide which public terms, if any, should accompany their contributions. Without naming member states, not all agree to Creative Commons licensing — which will then pose significant problems for any new pan‑European agency based on open data and also for open modeling in general.
European Commission commitment to CC-BY-4.0
On 22 February 2019, the European Commission committed to favoring Creative Commons CC‑BY‑4.0 licensing to facilitate reuse (JRC 2019a, 2019b). That decision was titled “Commission Decision adopting Creative Commons as an open licence under the European Commission’s reuse policy”. At the time, the Creative Commons organization reported this commitment as Vollmer (2019).
Choice of public license and license interoperability
In my view (and without running through the arguments directly), the clear choice for a pan‑European open data portal is CC‑BY‑4.0 for primary data and CC0‑1.0 for metadata.
However, the CC‑BY‑4.0 license is immiscible with the share‑alike ODbL‑1.0 license used by OpenStreetMap (Poole 2017) — meaning that data mixing can only be done locally and system modeling teams are not legally able to modify and combine their input data and republish under a single unified license (setting aside the question of whether the OSM project would litigate such infringements).
The UK government states that their OGL‑UK‑3.0 license is (inbound) compatible with the CC‑BY‑4.0 attribution license, but I would prefer to see independent legal confirmation for this claim (the choice‑of‑law provisions in the OGL alone may well prevent this interoperability).
I have previously written about license‑triggered databank siloing here. And this is a very real problem.
Note that if no intellectual property present, applying a CC‑BY‑4.0 or ODbL‑1.0 license is not, in of itself, legally problematic.
Legal metadata
The Dublin Core provides a reasonable and widely adopted standard for legal metadata, such as recording authorship and public license notices.
If copyright is present and the data arises in Germany (or a similar civil law jurisdiction), then the attribution of the originator must be tracked. German copyright law is unlike United Kingdom or United States law, for instance, in that German copyright law expressly prevents the author or creator from being reassigned to a third party, say a data hosting organization.
As a result, the legal metadata associated with datasets must be assiduously tracked through an ever‑growing and evolving graph structure of datasets. Ideally, however, that databank graph should never be more than about three deep, because corrections and additions should be propagated upstream — and not further downstream to create a combinatorial explosion.
Techniques for tracking source code and the associated license compatibilities in complex software (using so‑called SBOMs) can probably be refined and replicated for complex data repositories — indeed, the days of copy/pasting everything into a single spreadsheet and circulating the resulting file by email are probably (thankfully) drawing to a close.
Open definitions
The widely accepted definition for open data from the Open Knowledge Foundation (OKF) is known as The Open Definition. I am using this statement as the benchmark for “open data” in this posting. There is no equivalent statement for semantic standards — and that lapse in coverage needs to be addressed.
Standardized scenarios
The concept of standardized scenarios is important to include in this discussion. Scenarios are the lifeblood of system modeling and having different teams working from a pooled set of scenarios is useful for obvious reasons.
The European Commission has been working towards the release of a standardized energy sector reference scenario. This reference scenario must be fully complete (and not 90% as currently indicated) and not be legally encumbered for use and reuse. The Commission says it is considering CC‑BY‑4.0 licensing but no final announcement has been made.
Update on the European Commission reference scenario
Commission staff advised me on 5 May 2025 that:
I understand that you are referring to the spreadsheet containing the assumptions on technology costs to be used in the Reference scenario.
Following a consultation with expert stakeholders, we are preparing an updated version of the document. This version will be published together with the results of the Reference scenario (scheduled for early 2026).
There is no decision about the public licence yet.
A networked science paradigm
Regarding open science, the legal context and the social paradigm in play are inseparable. The open source software movement has certainly taught us this.
There is currently considerable interest and effort in developing data brokerage systems that use contract law to mediate the relationships between data providers and consumers. While data brokerage has clear and obvious roles, especially for private data, the information commons operates on a very different paradigm. Indeed, networked science (Nielson 2013) is a good expression of this paradigm in our context. Data brokers (Icebreaker One in the UK, for instance) sometimes view CC‑BY‑4.0 as simply another kind of contractual relationship, while failing to recognize that the social context for networked science is radically different from the transactional model they are deploying. These differences should always be kept in mind. Züger (2024), for example, asks how agencies, like the one being proposed here, can prioritize and facilitate public interests.
A new pan-European energy information agency will also need to be aware of the wider energy data ecosystem that is already developing in Europe. This includes research data generated through open science and the associated projects to manage, store, and redistribute this information. With a key concept being that of linked open data (LOD).
DeCarolis et al (2025) — and note that the lead author is a panelist in the upcoming debate in May — argue strongly for a networked science approach to public interest systems modeling — albeit one designed to operate under United States intellectual property law (I encourage people to read this remarkable 20‑page document).
Robustness
One reason for building an information commons is to protect scientific data from rogue government activity. Gilmour (2025), referring to the current United States administration, writes:
A common refrain of the saboteurs is that if these functions that they are targeting are important enough, the states or the private sector will step in to fill the gap. While some of these functions of the federal government are replicable outside of government, privatization will render them less accessible, more expensive and subject to the whims of the markets.
Closure
I see this proposal for a new pan‑European agency as part of the effort to build a functioning information commons, suitably licensed to allow for republication and use in original or modified form. While access and transparency are necessary conditions for an information commons, they are not sufficient. Only suitable public licensing of the data, metadata, and informing semantic standards can ensure the twin aims of open science and public accountability with any legal certainty.
The development of open source software provides a number of lessons that might be heeded. One is that licensing is paramount and another is that there needs to be a set of advocates who both push for the common good and weed out false claims of openness.
The Open Energy Modeling Initiative (openmod) began a decade ago by arguing that data under statutory reporting should be licensed CC‑BY‑4.0 (Brown 2017, Schmid et al 2019). There have been clear successes (for instance, the German BNetzA SMARD portal) and also good but incomplete progress in other areas (for instance, the ENTSO‑E Transparency Platform).
I look forward to the continuing debate for an integrated European energy sector information portal based squarely on principles from open science and public accountability.
Also to keep in mind that what will work legally in the United States, will not necessarily work in Europe.
The enabling legislation for a pan‑European energy information agency may need to grant the new agency additional privileges and rights. And current legislation covering statutory reporting may also need to be amended to facilitate data openness.
The GFF Bavarian database rights case (described earlier) provides a broad indication of what could be in store if the various legal issues raised here are not resolved. While the outcome in that case was not strictly a loss, neither was it a win for open data.
The European Parliament might consider amending the database directive to freeze all new formations of data copyrights and sui‑generis database rights. Indeed, some have argued that the directive does more social harm than good, especially in today’s data climate.
Failing that, legislation to prevent the database directive from applying to energy sector information mandated by statute should be contemplated.
My questions for the 8 May 2025 event
When panelists debate an EIA‑style energy information agency for Europe, they will need to grasp the legal intricacies articulated in this posting. So I am flagging the legal context in Europe as a necessarily central issue for the upcoming discussion on 8 May 2025. My question for that event is therefore:
How best to grasp the interconnected legal nettles covering the public licensing of non‑private datasets, metadata, and semantic standards? Could the combination of compromised legal reusability and immiscible licensing be a showstopper for this new agency?
With a follow‑up question:
Is it time to abandon the 1996 database directive (while preserving existing entitlements)? Or at the very least, remove its provisions from data under statutory reporting?
Remarks
-
The Bruegel think tank should consider open licensing the report (McWilliams et al 2025) that underpins their proposed pan‑European energy information agency.
-
I am now in email contact with the European Commission regarding the status of their forthcoming reference scenario — Commission staff did respond and I have included that information in the main text.
References
Bond, Toby (21 July 2021). CV-Online Latvia: CJEU complicates the enforcement of database rights. Bird and Bird. London, United Kingdom. Legal blog.
Brown, Tom (13 October 2017). ENTSO‑E data availability, quality and licensing — Presentation. Frankfurt, Germany: Frankfurt Institute for Advanced Studies (FIAS), University of Frankfurt. Presentation to 7th Open Energy Modelling Initiative Workshop, Munich, 13th October 2017. Creative Commons CC‑BY‑4.0 license. 
CJEU (3 June 2021). Judgment of the Court on ‘CV-Online Latvia’ SIA v ‘Melons’ SIA, case C‑762/19, document ECLI:EU:C:2021:434. Luxembourg City, Luxembourg: Court of Justice of the European Union (CJEU). 6 pages.
Clark, Charles (7 August 2023). Decision on updates to Data Best Practice Guidance and Digitalisation Strategy and Action Plan Guidance. London, United Kingdom: Office of Gas and Electricity Markets (Ofgem). OGL‑UK‑3.0 license.
Davidson, Mark J (January 2008). The legal protection of databases. Cambridge, United Kingdom: Cambridge University Press. ISBN 978-0-521-04945-0. Paperback edition.
DeCarolis, Joseph F, Sauleh Siddiqui, A LaRose, J Woollacott, Cara Marcy, Chris Namovicz, J Turnure, K Dyl, A Kahan, J Diefenderfer, N Vincent, B Cultice, and A Heisey (October 2024). A new generation of energy-economy modeling at the US Energy Information Administration — Working Paper 204. Washington DC, USA: US Energy Information Administration (EIA). Public domain status.
ENTSO‑E (18 February 2022). List of data available for free re-use — Last modified: 18 February 2022. Brussels, Belgium: European Network of Transmission System Operators for Electricity(ENTSO‑E).
ENTSO-E (1 February 2019). TSOs increase number of open data available through ENTSO-E’s Transparency Platform. ENTSO-E.
European Commission (December 2017). Legal opinion: legal aspects of European energy data — Output 2 of the “Study on the quality of electricity market data”. Brussels, Belgium: European Commission. The legal analysis was undertaken by Till Jaeger from JBB Rechtsanwälte in Berln.
European Commission (15 June 2013). “Commission Regulation (EU) No 543/2013 of 14 June 2013 on submission and publication of data in electricity markets and amending Annex I to Regulation (EC) No 714/2009 of the European Parliament and of the Council (text with EEA relevance)”. Official Journal of the European Union. L 163: 1–12.
European Parliament and European Council (27 March 1996). “Directive 96/9/EC of the European Parliament and of the Council of 11 March 1996 on the legal protection of databases”. Official Journal of the European Union. L 77: 20–28.
Gilmour, Jonathan (21 April 2025). “The Trump administration is sabotaging your scientific data”. The Guardian. London, United Kingdom. ISSN 0261-3077.
Gervais, Daniel J (2002). “Feist goes global: a comparative analysis of the notion of originality in copyright law”. Journal of the Copyright Society of the USA. 49: 949–981.
Hamilton, Chloe (5 July 2021). CJEU ruling narrows the scope of EU database rights. Mishcon de Reya LLP. London, United Kingdom. Concerns case C‑762/19 and ruling ECLI:EU:C:2021:434.
Hirth, Lion, Jonathan Mühlenpfordt, and Marisa Bulkeley (1 September 2018). “The ENTSO-E Transparency Platform: a review of Europe’s most ambitious electricity data platform”. Applied Energy. 225: 1054–1067. ISSN 0306-2619. doi:10.1016/j.apenergy.2018.04.048. Creative Commons CC-BY-4.0 license.
Hoyer-Klick, Carsten, Markus Blesl, Lüder von Bremen, Johannes Frey, Ulrich Frey, Anastasis Gainnousakis, Ludwig Hülk, Stefan Kronshage, Patrick Kuckertz, Gerald Lohmann, Christoph Muschner, Michaja Pehl, and Marion Schroedter-Homscheidt (November 2022). FAIR data in energy systems analysis — Poster. Presented at Implementing FAIR in the Domain of Energy Systems Analysis, Forschungsnetzwerk Systemanalyse, Berlin, 8–9 November 2022.
Jacobacci e Associati (5 July 2021). CJEU, C-762/19: the sui generis right of the database maker. Lexology. London, United Kingdom. Legal blog.
JRC (22 February 2019a). COMMISSION DECISION adopting Creative Commons as an open licence under the European Commission’s reuse policy. Brussels, Belgium: European Commission Joint Research Centre. Main document only and not the two accompanying appendices.
JRC (22 February 2019b). COMMISSION DECISION adopting Creative Commons as an open licence under the European Commission’s reuse policy. Brussels, Belgium: European Commission Joint Research Centre. Landing page URL given.
McWilliams, Ben, Simone Tagliapietra, and Georg Zachmann (19 February 2025). Europe’s energy information problem — Policy brief issue 07/25. Brussels, Belgium: Bruegel AISBL.
Nielson, Michael (2013). Reinventing discovery: the new era of networked science. Princeton, New Jersey, USA: Princeton University Press. ISBN 978-069116019-1.
Poole, Simon (17 March 2017). Use of CC BY 4.0 licensed data in OpenStreetMap. OpenStreetMap Blog.
Schmid, Eva, Ingmar Schlecht, Tomas Šumskas, and Florence Melchior (October 2019). Why do we need an open data licence on the ENTSO‑E Transparency Platform? An FAQ. Private document. 
Schweizer, Mark (5 November 2015). C-490/14 — Verlag Esterbauer: Get off my map!. The IPKat. London, United Kingdom. Regarding Case C‑490/14.
Synodinou, Tatiana (22 February 2021). Search engines and databases in the search for a balance: the AG’S opinion in the ‘CV-Online Latvia’ case. Kluwer Copyright Blog. Concerns case C‑762/19.
Vollmer, Timothy (2 April 2019). European Commission adopts CC BY and CC0 for sharing information. Creative Commons.
Züger, Theresa (10 January 2024). How does the data institute become public-interest orientated?. Berlin, Germany: Alexander von Humboldt Institute for Internet and Society. doi:10.5281/zenodo.13221900.
Further reading
Anon (21 November 2017). A review of the ENTSO-E Transparency Platform: Output 1 of the “Study on the quality of electricity market data” commissioned by the European Commission. Brussels, Belgium: European Commission. Publication date from PDF metadata. From the cover, VVA, Copenhagen Economics, Neon, and Deloitte were involved. Asks survey questions on licensing but does not traverse the issues.
McGeever, Mags, Angus Whyte, and Laura Molloy (2015). Five things you need to know about research data management and the law: DCC checklist on legal aspects of RDM. Edinburgh, United Kingdom: Digital Curation Centre (DCC).
Shipley, David E (21 December 2007). “Thin but not anorexic: copyright protection for compilations and other fact works”. Journal of Intellectual Property Law. 15 (1): 91–141. UGA Legal Studies Research Paper 08-001.
Vilgo, Kristjan (30 November 2020). RSC Conference 2020 parallel session “Digital tools”. Video 00:50:26. Filmed on 24 November 2020 at Regional Security Conference 2020. Presenter is from Estonian TSO Elering AS.
External links
- Bruegel AISBL website
Disclaimer
The author has no legal training and the issues presented are clearly complicated and evolving.
Any corrections gratefully received.
▢