The School of the Unfortunate: Greek education spending and EU fiscal rules

 

Europe's fiscal rules now name which public shortfalls are exceptional. In Greece, the classroom is not one of them.

Joanna Demopoulou*

* Joanna Demopoulou holds a PhD in International Affairs and is a former Executive in Residence at the Geneva Centre for Security Policy (GCSP). Her research examines the legal and institutional architecture of public authority and governance, focusing on how law structures, distributes, and constrains public authority across European and international institutions.

Photo: Greek parliament; photo credit A. Sivan, via Wikimedia Commons

 

On 9 October, EU finance ministers agreed that Greece may let its spending exceed the agreed path in order to strengthen energy security. Among the measures Greece intends to finance on that basis is a €45 million energy upgrade of 246 schools and childcare centres. The upgrade of a school building may count among the costs allowed to take spending above that path. The teacher who works inside it may not.

The decision landed in a school year already short of teachers, with pupils, students and teachers filling Athens' Syntagma Square and the steps of the Propylaea to say so. What the protests reject is a particular way of reading that shortage: that an empty desk belongs to the family that has to live with it, not to the state that was supposed to fill it.

Greek and European law answer that question from two different places. The Constitution gives the state a duty to provide education. The fiscal framework Greece operates under grants special treatment to other needs and leaves the shortfalls of education to ordinary budget management. What follows asks how those two answers coexist, and who ends up paying for the space between them.

The building and the teacher

Several weeks into the school year, the secondary teachers' federation OLME estimates that at least 15,000 teaching posts remain unfilled. In practice that means lost teaching hours, merged classes, more pupils in every room. The figure is the federation's estimate, not an official count, and official figures exist only for a smaller part of the system. For Greece's model and experimental schools, their own governing committee identified 473 staffing needs for 2026-27, to be met by substitute teachers in the hiring rounds: 199 in general primary, 190 in general secondary, the rest in special education. These are needs recorded at the start of the year, not posts shown to be empty today.

Energy upgrades are one kind of building need, and not the most urgent one everywhere. The music school of Vartholomio, in the western Peloponnese, has taught its pupils in prefabricated classrooms since 2008. Work on a new building began in 2023 and has not been finished. This autumn pupils and teachers began going home with bites from insects nobody has yet identified, with redness and itching; some were taken to hospital, and in-person teaching was suspended and moved online. According to a parliamentary question tabled by KKE MPs, the region's health directorate has since found the classrooms unfit. No European rule is involved here. Nor is Vartholomio alone: in three municipalities of Attica, a local teachers' association reported in September 2025 that 73 of 132 primary schools and kindergartens had at least some prefabricated classrooms.

Who pays

The gaps reach the subjects that decide university entry. At the general lyceum of Kythira, the Piraeus secondary teachers' union announced on 6 October uncovered hours in three nationally examined subjects, mathematics, chemistry and economics; a month into the year, final-year pupils had not been taught them. On Tilos, the mayor wrote to the ministry on 29 September that the five final-year pupils on the economics track had no economics teacher. Both accounts describe the situation as it stood when they were written; the posts may since have been filled.

Two cases do not make a statistic. According to ministry data reported by the Athens News Agency in early October, after two rounds of substitute hiring, 11 of 13 regions had covered close to all vacancies in exam-subject specialties. That is real progress. What Kythira and Tilos show is something the national figure cannot: what an uncovered post means to the pupils who are living through it, whose exam date does not move to accommodate them.

The Tilos letter adds a further detail. The island, the mayor noted, has no tutoring schools and none of the alternatives available elsewhere. It does not say what individual families did about that. What it establishes is narrower: the usual Greek remedy for a missing lesson, paying for it privately, presupposes a local market that does not exist everywhere. On Tilos, a family willing to pay would have to look beyond the island, or online.

Where that market does exist, its scale can be measured. An analysis by KANEP, the research institute of the Greek General Confederation of Labour, of ELSTAT's household budget survey found that Greek households spent almost €614 million on private tutoring in 2023, about 95% of it for secondary school pupils. In real terms the sum rose by about 36% between 2021 and 2023. Households in the three highest income groups devote 3.5 times as large a share of their consumption spending to tutoring as those in the three lowest. That is a comparison of shares, not amounts, and the gap has widened over the decade.

An unquantified part of that spending has nothing to do with missing teachers. Much of it answers the long-standing pressure of the entrance exams, preparation a family would buy regardless, on top of a school that works. Tutoring that replaces teaching the school owed is a different matter, and the figures cannot separate the two; nor does every vacancy become a private bill. What the figures do show is the channel that opens once a family decides to buy substitute teaching. That channel is open to those who can pay, and the gap in what households spend on it has been widening.

In late September, on live television, Greece's deputy foreign minister, Haris Theocharis, divided the country's parties into those that represent the kakomoirides, the unfortunate, and those, like his own New Democracy, that represent the noikokyrides, householders who manage their affairs responsibly. A few days later he said the phrase had been a mistake. The remark is party rhetoric, and it is read here as nothing more. The setting matters more than the remark. In a country that the Union's own surveillance found this spring no longer faces macroeconomic imbalances, continuing hardship is easily recoded as a private failing. The phrase gave that recoding a name. It did not invent it.

Private remedies make the recoding easy. Once a lesson has to be bought, a difference in results can be read as a difference in effort, or in how well a household manages its affairs. A pupil who falls behind after months without a teacher in an exam subject may come to be seen not as someone a public service failed, but as the unfortunate one in the class. Nothing here shows that this has happened to any particular pupil; what the arrangement does is make that reading available. There is a further possibility, though the figures cannot establish it: when a missing lesson becomes a private bill, a failure of public provision may also become less visible as a political problem.

Elsewhere I have described an informal constitution of access, in which rights are formally equal but worth different amounts in practice. In education the mechanism is more mundane, and for that reason more widespread. The right to free education belongs to everyone. When the state does not deliver it, what remains of the right depends on the family's purse, and on where the family lives.

Time that does not come back

A right can be lost through timing as surely as through denial. Writing in EU Law Live earlier this year about a Greek SLAPP case, I argued that delay can itself become the violation: a remedy that arrives after the harm is done protects very little. The analogy has limits; the standards that govern a defamation case do not transfer to schooling. What carries over is the question of effectiveness over time. A post filled in November does not return the weeks of syllabus lost in September and October, and on Kythira the union described a full month lost for the final year. For those pupils the national exams fall on a fixed date, and the time between an empty desk and that date is all there is. Teaching delivered late, for an examination that will not wait, delivers the right only in part.

The answer

On 21 October, the day of the nationwide education protest, Education Minister Sofia Zacharaki told the broadcaster ANT1 that her ministry hears the "real demands" of pupils and that occupying schools is no solution. She cited renovation work in 2,000 schools and €733 million for new student halls through public-private partnerships, along with 53,000 permanent teacher appointments made since 2019, which she said outnumber the retirements over the same period. She also conceded that more resources do not mean every need has been met.

The figure holds up. The Education Ministry's own Strategic Plan for Primary and Secondary Education 2025-2029 records 38,409 permanent appointments for 2019-2024 and a further 10,000 in 2025, and a later report from ERT puts the running total above 53,000. What the same plan shows, that the broadcast figure does not, is why a national total answers a different question. It separates organic vacancies, posts left empty, from functional gaps, which open when the teacher who holds a post is on leave, seconded elsewhere, or otherwise absent. Filling the first does nothing for the second. That is part of why the plan records 51,149 substitute hires in the single school year 2024-2025, alongside that year's permanent appointments, though not all of them replaced an absent teacher, since the total also covers programmes such as parallel support. A national hiring total, however large, cannot show whether a particular post, in a particular school, was filled when a particular class needed it. The ministry's planning document and the cases of Kythira and Tilos describe the same system from opposite ends, and only the second tells a pupil whether there will be a teacher in the room.

On 5 November the Commission presented the EU Teachers Agenda (COM(2026) 795), a proposal for a Council recommendation on teaching careers and working conditions. A July preparatory document for the wider package described such actions as "not legally binding". The October Council recommendation on Greece's expenditure path does something else: it permits specified energy-security spending to exceed that path. For the ordinary cost of teachers, no comparable exception exists.

The numbers

Greece spends less on education than almost anywhere in Europe. According to Eurostat, general government expenditure on education in 2024 stood at 3.9% of GDP in Greece, against 4.8% for the EU. As a share of total public spending, education took 8.2% in Greece, the second-lowest share in the Union after Italy. The OECD's Education at a Glance 2026 puts total expenditure on educational institutions, from public and private sources, at $8,600 per lower secondary pupil in Greece in 2023, in purchasing-power terms, against an OECD average of $15,305 (Table C3.1).

None of this proves that every vacancy comes down to money. It does describe a system that has run below the European average for a long time, and has grown used to families making up the difference.

An exception with a list

In July 2025 the Council of the EU activated the national escape clause of the Stability and Growth Pact for 15 member states, Greece among them. The clause runs for four years and allows a deviation of up to 1.5% of GDP from the agreed net expenditure path, provided the excess is due to higher defence spending. By February 2026 it had been activated for 17 states.

For Greece, the list has since grown. A Council recommendation, proposed by the Commission in September and agreed by finance ministers on 9 October, extends Greece's clause to eligible measures that strengthen energy security, capped at 0.3% of GDP a year and 0.6% in total, within the overall 1.5% ceiling. Among the envisaged measures, the recommendation mentions energy upgrades of public buildings.

Three things should be kept apart. The clause permits a deviation; it funds nothing, and whether a given measure qualifies is for the Commission to assess. The school upgrades are paid for from national resources and would be lawful spending with or without the clause. And nothing in the clause prevents Greece from paying for teachers out of its regular budget. What the clause decides is narrower, and for that reason easy to miss: which costs may push spending above the agreed path without consequences under the fiscal rules.

None of this is an argument against defence or energy spending. In a week of Turkish navigational warnings south of Rhodes and east of Crete, that argument would be naive. Nor is there evidence that the clause took teachers out of classrooms; Greece's budget is decided in Athens, not in Brussels.

Who decides what is exceptional

The legal basis is Article 26 of Regulation (EU) 2024/1263. It allows a member state to deviate from its expenditure path where exceptional circumstances outside its control have a major impact on its public finances, provided fiscal sustainability over the medium term is not endangered. The Council decides, on a recommendation from the Commission, at the request of the member state itself.

The strongest objection to this argument follows directly from that text, and as far as it goes, it is correct. A war on the Union's border or an energy shock can be described as outside a state's control. A chronic shortage of teachers cannot. It is the kind of problem a government is expected to solve through ordinary budgeting, and the clause is designed to keep it there. Defence and a teacher shortage are not legally comparable circumstances.

Take the objection at its strongest, and the question still stands. Two legal texts now speak about the same school. The Greek Constitution, in Article 16, makes education a basic mission of the state and gives every Greek the right to free education at every level, in state institutions. The European fiscal framework recognises defence and, for Greece, energy security as needs that may lawfully exceed the ceiling. The two texts regulate different things. Treating one need differently from another under the fiscal rules does not, by itself, create a legal contradiction with the Constitution, and the Constitution does not entitle education to a fiscal exception.

The difference matters all the same, because three separate questions are in play, and each is answered in a different place. The first is legal obligation: whose duty is it to teach? The Constitution answers that it is the state's. The second is fiscal eligibility: which spending may exceed the agreed limits? Article 26 sets the general conditions, and the Council's individual recommendations apply them. Ordinary staffing falls outside. The third is actual access: what happens to a pupil when the lesson is not provided? No text answers that. It is settled in practice, by what families can afford and by what is available where they live, and for some, as on Tilos, private replacement is hardly available at all.

That is an institutional asymmetry, not a breach. The need the national constitution names as a basic mission has no fiscal standing comparable to the needs the Union has agreed to treat as exceptional, and every teacher hired has to fit under a ceiling that two other needs may lawfully exceed.

Nor is the list drawn by Brussels alone. The clause is activated at the request of the member state. Greece asked for flexibility for defence in April 2025 and for energy security in August 2026. Both requests were political choices about which needs to present as exceptional, made within the categories the Union was prepared to accept. An exception for teachers was never part of that conversation, and under Article 26 as it stands, it is doubtful that a request for one could succeed. The rules set no formal hierarchy of public needs. In practice, as Europe rearms for the rest of this decade, the needs allowed to exceed the spending limits are being named one decision at a time: defence in 17 member states, energy security in Greece. The classroom is not among them.

What the students refuse

What a right is worth depends not only on its wording but on how institutions classify its shortfalls: as exceptional or ordinary, as the state's failure or the user's misfortune. The pupils and students who camped in Syntagma refuse the second term of each pair. They are asking that a teacher's empty desk and a crumbling building be treated as gaps left by the state in meeting a constitutional obligation, not as the misfortune of families that could not cover them.

A state can hold a constitutional duty to educate and operate under a fiscal framework that grants exceptions to other needs without any contradiction appearing in the texts. The strain shows elsewhere: in a final-year class on an island with no economics teacher and no tutoring school, in a music school that has spent eighteen years in prefabricated classrooms, and in the household budgets of families who can buy what the school did not provide. On paper, the missing lesson remains the state's to provide. In practice it has become the family's to buy, where a family can buy it, and it is simply absent where it cannot. The classroom does not need a place on Europe's list of exceptions; it needs the ordinary budget to reach it. Until it does, Greek law will go on assigning the duty to the state while the cost settles somewhere else.

Bigger Frontex, Smaller Fundamental Rights: Digitalisation and Externalisation in EU Border Governance

Mariana Gkliati, Tilburg University

Marloes Streppel, Leiden University

Photo credit: Francesco Placco, via Wikimedia Commons

Introduction
On 27 October 2026, the European Commission proposed two new Regulations expanding the mandate of the European Border and Coast Guard (EBCG) Agency, Frontex: an amendment to the EBCG Regulation and a new Regulation on the Digitalisation of Returns. The Commission envisages a substantial budget increase for the agency, allocating €11.8 billion for 2028–2034, almost double its current allocation. Its operational capacity would expand through a gradual increase of the Standing Corps to 30,000 officers by 2041 (6,000 Frontex staff, 7,000 nationally seconded officers, and a new 17,000-strong European Border and Return Reserve). Frontex would also have easier access to state-of-the-art equipment and technology, becoming the central purchasing body also for advanced digital and AI-enabled technologies.

The proposal’s opening paragraph begins by highlighting the general achievements of EU border management in past years: the number of irregular border crossings has declined steadily since 2023, reaching its lowest levels since 2021; returns have increased to more than 30%; with the assistance of Frontex, the EU’s “collective capacity to manage its external borders has been progressively strengthened” (Preamble, p. 1).  This opening statement seems to confirm the Commission's own 2024 evaluation of Frontex’s current mandate: “no immediate need for a revision of the EBCG Regulation or its annexes” (Evaluation of Regulation 2019/1896, p. 11). Nonetheless, the Commission continues in its second opening paragraph: “migratory pressures are likely to persist,” loading this prediction with securitarian concerns about geopolitical developments, migrant smuggling, cross-border crime, and hybrid threats (Preamble, p.1).

Expansive Trends in EU Migration Management

Over the past decade, the area of migration management, including asylum and border management, has become one of the fastest-growing areas of EU law, with a rapid expansion of legislative acts. More importantly, the area has undergone a historical regulatory intensification, with an exceptional expansion in the scope, complexity, and institutional reach of EU regulation.

This legislative overgrowth in EU migration is reflected in the New Pact on Migration and Asylum, which restructured virtually the entire EU asylum and migration framework through the adoption of 10 legislative acts (mostly Regulations) in 2024 (entered into force in 2026) (link). Shortly after this widespread legislative overhaul, the regulatory activity continued: at lightning speed, the EU legislature adopted two amendments to the new Asylum Procedures Regulation (link, link) and a new Returns Regulation (link). Moreover, the Commission has presented an ‘anti-smuggling package,’ composed of two proposals, one amending the EU rules on preventing and countering the facilitation of unauthorised entry, transit and stay, and one enhancing police cooperation on migrant smuggling and trafficking and the role of Europol in this regard.

 This legislative expansion has gone hand in hand with institutional expansion, increasing the powers, competences, and operational responsibilities of EU agencies, particularly Frontex, the EU Agency for Asylum, and Europol. Two further obvious trends in this regulatory activity are the technological and territorial expansion well beyond the EU’s borders. All three trends of digitalisation, externalisation and agenification are compounded in the two proposed Regulations.

Digitalisation

Frontex has progressively evolved into an important information hub within the EU's border and security architecture. Its work has a significant intelligence dimension. In particular, the agency is tasked with developing and operating common and interoperable systems which enable swift and reliable information-sharing regarding emerging risks at the external borders, irregular migration, and returns (Art. 10(1)(ab and ac), EBCG Regulation) and operates under the obligation to share information with Member States and other EU Agencies (Arts. 11 and 12 ECCG Regulation). 

Since its establishment in 2004, Frontex has been tasked with conducting risk analyses, drawing on information supplied by Member States and other sources. Its intelligence function has since expanded considerably, notably through the Common Integrated Risk Analysis Model (CIRAM), which collects and analyses statistical and operational data provided by Member States and other Agencies, but also media and other sources, including those produced by the Agency’s own work (e.g. risk analyses, vulnerability assessments).

The 2011 amendment of its mandate introduced limited competences to collect and process personal data concerning irregular migration and cross-border criminal activities. The 2016 amendment further expanded these competences, including powers to process information obtained during screening and debriefing activities and to transmit relevant information to Europol and national law enforcement authorities for criminal investigations. Frontex’s 2016 mandate permitted it to collect, process and share personal data not only for purposes of migration management but also for law enforcement purposes, including combating terrorism, human smuggling and trafficking, and document fraud.

The 2019 amendment expanded the role of Frontex in the digitalisation of returns. The agency was then tasked with developing and operating a centralised return management platform to process and share return-related information, including personal data, biographic data, and passenger lists. Frontex also became responsible for setting up and managing the Central Unit of the European Travel Information and Authorisation System (ETIAS). Today, Frontex has access to a range of (largely interoperable) EU and international information systems and databases, particularly those relating to border management, migration, asylum, identity verification, travel movements and law enforcement (e.g. Schengen Information System (SIS)), while also having its own information infrastructure including EUROSUR, JORA (Joint Operations Reporting Application), Frontex Risk Analysis Network (FRAN) and Frontex's ETIAS Central Unit.

New measures

Developments around Frontex’s expanding digital powers are deeply tied to returns. The 2026 proposal turns Frontex into a fully-fledged European Return Agency, the bloc’s operational arm. Even though the proposal stresses that Member States remain responsible for issuing return decisions, the new powers represent quite a step-up from the coordinating role that the agency has now. Within the backdrop of the Pact and the Return Regulation, which foresee a stronger, central coordinating and operational role for the agency, Frontex is well positioned to take on a much more dominant role, becoming “the Union entity responsible for the planning, coordination and implementation of returns.” (Preamble, para. 98).

Fulfilling this promise, the proposal, firstly, introduces a Return Coordination Centre (RCC), containing a SIRENE Office, similar to the one of Europol. The RCC is intended to collect and analyse information on Member States’ return needs to facilitate early identification, return counselling, return operations, and reintegration assistance. The information capacity of the agency is a central component in the proposal, especially in relation to returns. Despite Frontex’s broad access to information held by national authorities, EU databases, other EU agencies, and third-country partners, and its ability to aggregate such information from all different sources to conduct risk analysis, identify patterns, and produce intelligence, the Commission identifies a gap in the agency’s information availability. This gap concerns supplementary information on secondary movements by people subject to return decisions and attempts by people subject to entry bans to enter the Union, which Member States exchange bilaterally through national SIRENE (Supplementary Information Request at the National Entries) Offices. Such supplementary information does not form part of the SIS alert and is not stored in the SIS database itself. It constitutes an additional layer of information exchange, including information concerning identity verification, the circumstances underlying an alert and the operational follow-up of alerts, hits and actions taken or not taken, consultations between authorities, data quality, incompatible alerts, and requests concerning data subjects' rights (eu-LISA SIS Technical Report 2023-2024).

According to the Commission, Frontex’s lack of access to this information keeps the agency from identifying emerging and operational trends, Member States’ needs, and prioritising return-related resource deployment. Access to such information, which moves beyond the identification of individual border crossers and relates to “investigative knowledge” produced about the activities of the individual (Art. 3(2) SIS Regulation), extends the agency’s role beyond coordinating return operations towards monitoring cross-border movements and informing EU-wide return policies.

Secondly, the proposal strengthens the role of EUROSUR as an instrument of early warning, situational awareness, and operational coordination to facilitate rapid response in ‘crisis’ situations. EUROSUR Fusion Services combine the monitoring capabilities of the Member States and also include information from third countries. Additionally, the proposal envisages the use of new and any future EU space systems, such as Galileo, LEO PNT, GOVSATCOM, and IRIS2 (proposed recital 45). In the proposal, the scope of EUROSUR is substantially expanded beyond migration, cross-border crime and saving lives at sea also to include overall security at the external borders (proposed Art. 22). This would include hybrid threats, explicitly including the instrumentalisation of migration, disinformation and cross-border crime (proposed Art. 31). The new rules provide for a reporting obligation by Member States within 24 hours of any (likely) event or incident that may have an impact at the external borders whether this is related to migration, cross-border crime, or hybrid threats and the overall security of the external borders (p. 87 of the Proposal). Moreover, the proposed Art. 32 establishes an advance-warning mechanism with a more immediate reporting obligation for more impactful events, which would trigger an automatic escalation towards EU-wide alerts and the activation of the agency’s surveillance capabilities (EUROSUR Fusion Services). Furthermore, Frontex would acquire stronger monitoring and supervisory functions over the reporting by Member States, including with respect to delays or insufficient quality of the reporting (proposed Art. 26(4)). As a result of such monitoring, the Agency can recommend that a Member State request the initiation of a joint operation or border intervention, with the Member State obliged to respond within six working days and justify any negative response (proposed Art. 46).

Thirdly, Frontex’s situational picture for the purpose of risk analysis would be enriched with aggregated and anonymised statistical information derived from activities outside the EUROSUR framework, including debriefing interviews and online-platform monitoring (proposed Arts. 29(2)(f) and 45(5)(b)). Article 33 of the proposal establishes an Online Monitoring and Response Centre, which would allow the agency to monitor online content that facilitates, promotes, or provides instructions for unauthorised crossing in a manner likely to endanger border security or the safety of persons crossing, or intending to cross, them. This can include content on platforms such as Facebook, Instagram, X, TikTok, and public discussion forums, calling for, organising or providing instructions for such unauthorised crossing, inciting violence or the disruption of border infrastructure, content relating to cross-border crime, or content providing information relevant for instrumentalisation, disinformation, cyber incidents, and cross-border criminal activities as part of a ‘hybrid campaign’ (Article 33(3)). Through this centre, Frontex can act as a trusted flagger under the Digital Services Act (DSA), notify platforms of illegal content on their sites, take action determined in EU codes of conduct or crisis protocols in implementation of DSA, and request national authorities to take measures against illegal content (Article 33(1)).

Finally, the Centre must regularly report to the Commission on any relevant requests, any content requiring development of counternarratives by the Commission and the Agency, and general information on the European Situational Picture (proposed Art. 33 (9)), thus coordinating responses to online narratives and prompting policy and implementation responses by the Commission.

Regulation on the Digitalisation of Returns

The Frontex Recast proposal is accompanied by a proposal for a Regulation on the Digitalisation of Case Management in the Area of Returns, Readmission and Reintegration, which develops an interconnected return case management infrastructure. According to the Commission, the current system for return-related data management and exchange is insufficient and fragmented across Member States, with different IT systems hindering interoperability and information exchange and, therefore,  the effectiveness of returns as a whole (proposal p. 10). Therefore, the Returns Digitalisation proposal is intended to set up the digital infrastructure that facilitates swift information exchange between Frontex, national authorities of Member States, other agencies (especially Europol and eu-LISA), and third countries.

In this system, Frontex is a key actor; it is responsible for developing and operating key EU systems, while Member States remain responsible for national return case management. The innovations of the technical infrastructure mainly include the development of: a) RECAMAS, a model for the national management of return cases (proposed Art. 11), b) iRCMS, the EU system for the management of readmission information (proposed Art. 12) and c) RIAT, the Reintegration Assistance Tool (proposed Art. 13).

Frontex’s RECAMAS Model is a blueprint for managing return cases developed by the agency, to which Member States must adapt to ensure consistent handling of returns across the EU. National RECAMAS systems would be integrated with the case management systems in the area of readmission and reintegration, iRCMS and RIAT (proposed Art. 3) and connect to the Communication Portal, developed by eu-LISA to facilitate the exchange of information between Member States, which is not available on SIS or Eurodac (proposed Art. 14). The communication portal should also connect all national RECAMAS as well as Frontex through the Integrated Return Management Application ('IRMA') already operated by Frontex (proposed recital 19). Moreover, national RECAMAS, iRCMS, and RIAT would use the European Search Portal (‘ESP’) to access EU information systems and would then feed information into the Central Repository for Reporting and Statistics (CRRS) to enable eu-LISA to generate more detailed and harmonised statistical analysis.

Externalisation

Through status agreements, working arrangements, and informal cooperation with non-EU countries, the agency has gradually extended its reach well beyond the EU’s borders and taken an active role in exporting EU borders, mainly to the Balkans and to West and North Africa.

Frontex had been operating in third countries even before it acquired a mandate to do so in 2011. In fact, it has been participating in such operations beyond EU borders since its foundation under bilateral agreements between EU Member States and third countries. For example, Operation HERA, launched in 2006, involved assets from member states and, under bilateral agreements with Spain, assets and staff from Senegal and Mauritania in patrolling the coastal areas of Senegal, Mauritania, Cape Verde and the Canary Islands. These patrols intercepted migrants attempting to leave the coasts of these states.

The 2016 European Border and Coast Guard Regulation marked a turning point for the agency’s role in border externalisation, since the agency acquired the competence to launch operations under status agreements in non-EU neighbouring countries (Art. 54(3) EBCG Regulation 2016). Under this competence, Frontex launched the first EU border surveillance operations outside the EU borders in the Western Balkans, and is today operationally active in Moldova, Serbia, Montenegro and Albania. In 2019, the agency's mandate to launch border surveillance operations expanded beyond neighbouring countries, and it may now launch such operations anywhere in the world (Art. 54(2) EBCG Regulation 2019). Besides joint operations, the agency may also launch and finance technical assistance projects in third countries and provide other operational and technical assistance relevant to returns (Art. 4 (6-7)). It may further deploy its statutory staff and other experts as liaison officers (Art. 77) and exchange information with the non-EU countries in the framework of EUROSUR, the pan-European border surveillance system aimed at improving situational awareness and the reaction capabilities of the border and coast guard authorities of the EU member states (Art. 75).

The initial proposal to amend the 2019 Regulation empowered Frontex to support third countries with returns by organising and coordinating operations from one third country to another (Art. 75(4) of COM(2018)631). These powers did not eventually survive the trilateral negotiations, and the relevant provision was removed only to re-emerge in the 2026 legislative proposal.

The groundwork for Frontex's proposed new powers in third countries has already been laid out by the two Regulations adopted earlier in the same year. First, the 2026 amendment to the Asylum Procedures Regulation (APR), which significantly broadened the application of the safe third country concept, allowing applications to be declared inadmissible where an applicant has transited through a third country or based on an agreement with the third country allowing for the examination of their asylum claim there, without requiring an individual connection with that country. Second, the new Return Regulation formally introduced the possibility of transferring individuals subject to return decisions to return hubs in third countries, including arrangements providing for onward return to countries of origin or other third countries (see here)

New measures

The Frontex Recast proposal would significantly expand the agency’s mandate in third countries in relation to returns, most notably an “entirely new set of tasks and responsibilities” to carry out return activities from one third country to another (proposal p. 17). Strengthening cooperation with third countries is part of the EU’s “whole-of-route” approach to migration management, which aims to prevent arrivals and facilitate returns (proposal p. 70).

In particular, the agency would be able to provide technical assistance to third countries, including vulnerability assessments, tailored training or risk assessment products, and other capacity building and support in returns and reintegration (proposed Art. 63, p. 5). It would also support them in implementing returns, exercising executive powers on third-country territory (proposed Art. 64). The proposal defines technical assistance as any technical, logistical, or financial support that does not involve the exercise of executive powers (proposed Art. 2(39)). Executive powers, in turn, are defined as: the power and authority necessary to perform border management or return which entail the exercise of public authority or direct enforcement of public power by the agency (proposed Art. 2(39), (40)).

The proposal establishes two distinct frameworks for Frontex’s cooperation with third countries in the area of return. Where more structural or long-term cooperation is envisaged, Frontex may provide technical assistance based on working arrangements concluded with the relevant third-country authorities. However, such arrangements are not made mandatory. Alternatively, technical assistance may be provided under existing status agreements, including those concluded before Frontex acquired the relevant mandate. Importantly, however, the proposal does not make working arrangements and status agreements the exclusive legal frameworks for such cooperation, thereby leaving room for more informal, ad hoc forms of engagement (proposed recital 110).

Frontex is allowed to provide operational assistance exclusively based on an international agreement concluded by the EU (status agreement). This covers both the implementation of Member States' return decisions from third countries, potentially including onward returns from third-country return hubs, and assistance with returns based on return decisions issued under national law.

Continuing the “whole-of-route” approach, third-country cooperation is also part of the digital systems feeding into the EU Situational Picture to ensure not only pre-frontier, but broadly extra-territorial situational awareness. While information acquired through EUROSUR can already be shared with third countries, under the proposed recast the detailed rules for such cooperation become part of the agreement with the third country, making access to EUROSUR information provisional for the third country upon broader cooperation with the agency (proposed Art. 90). Moreover, the Online Monitoring and Response Centre would also cooperate with third countries subject to relevant arrangements (no exclusive mention of working arrangement and status agreements), and independent external stakeholders, from which non-governmental and private actors are not excluded and without any mention of prior formal agreements (proposed Art. 33(6-8, 10)).

Finally, on the digitalisation of returns management, iRCMS and RIAT provide for the processing and exchange of information between Frontex and third-country authorities. Under Art. 17(2) of the digitalisation proposal, working arrangements would establish the framework for such cooperation, while the data transmitted would be limited to what is strictly necessary for the particular readmission case. Given the broader system integration described above, national RECAMAS systems would integrate information generated in third countries on readmission and reintegration, further feeding into EU situational awareness.

 

Fundamental Rights and Weakened Institutional Scrutiny

Risks

The mandate expansion again raises new concerns for the agency's fundamental rights-sensitive work on returns and data management. Early NGO analyses already sound the alarm regarding returns from one third country to another and social media surveillance (e.g. Amnesty International, Front-LEX, PICUM). Indeed, the involvement of Frontex in third countries raises a series of human rights issues, especially when the partner countries have questionable human rights records and do not follow the standards of the European Convention for Human Rights (ECHR).

Engagement in third-country return operations increases the risk of (chain) refoulement, especially since it also covers returns based on return decisions issued under national law. This option allows for the circumvention of EU procedural safeguards. Frontex would thus be mandated to support the implementation of return decisions originating exclusively through the national asylum, return, and judicial systems of third countries, without the safeguards of the EU Asylum Procedures or the EU Returns Regulation. Notably, this option is not limited to EU accession candidates, countries bound by the European Convention on Human Rights, or countries designated as safe third countries at Union level but also includes any other country provided that there is clear EU interest so long as it respects international human rights standards (proposal p. 18, Art. 64). Relevant safeguards, concerning due process, non-refoulement, incident reporting and independent monitoring, vulnerability and the use of force are left to be specified in the status agreements. It remains unclear how Frontex will verify compliance in practice (e.g. regarding effective access to asylum procedures, legal assistance, interpretation and suspensive remedies), and what the exact consequences for non-compliance will entail (e.g. partial or whole suspension or termination of the relevant operation or broader cooperation). Thus, involvement in third-country return operations leaves Frontex vulnerable to complicity in violations committed by partner countries, particularly where the agency's operational involvement contributes to a chain of removals across multiple jurisdictions.

Moreover, the prospect of Frontex’s involvement in social media monitoring has drawn comparisons with the invasive tactics of ICE American federal agency. The Chair of the OSCE Parliamentary Assembly’s Ad Hoc Committee on Migration has raised serious concerns about privacy, freedom of expression, and data protection, and the result is a chilling effect on public debate around migration, especially given the widespread use of these tools at the EU’s external borders.

This expansion of Frontex’s data-collection tools comes against the backdrop of longstanding criticism of the agency’s poor reputation for data protection and management, including its unlawful collection and exchange of migrant data with Europol. The proposed architecture further raises several privacy and data-related concerns. Firstly, Frontex's information-gathering activities increasingly extend beyond border management and migration control to cross-border crime, terrorism, hybrid threats, disinformation and internal security. This raises questions about purpose limitation and potential function creep. What enhances the complexity in this regard is the expanding interconnection between RECAMAS, iRCMS, RIAT, the Communication Portal, IRMA, the European Search Portal and the Central Repository for Reporting and Statistics, which allows for information collected in one context and for a certain purpose to be used in another context potentially beyond its original purpose. Moreover, such interoperability and cross-linking of multiple databases and information systems allows for an increasingly intrusive picture of individuals. This substantial aggregation of information from different sources, including debriefing interviews, operational reports, social media monitoring and third-country authorities, may produce discriminatory risk profiles of individuals, nationalities or groups. Finally, information supplied by third-country authorities, operational partners, online platforms and other external sources may differ significantly in quality and reliability. Still, it may also not be collected and processed in accordance with the EU privacy and data protection standards.

Responsibility

Despite the increased risks for the engagement of Frontex in human rights violations, and the uncertainties regarding the allocation of responsibility when multiple actors are involved, the new Regulation increases the level of complexity and opacity by allowing for different levels of often informal cooperation with the partner countries. Formal status agreements are not always required. Several actions may be taken through informal working arrangements concluded directly by the agency, evading scrutiny by the European Parliament. Moreover, ad hoc assistance, in the absence of structural cooperation, lacks any clear legal and procedural basis. The combination of formal status agreements, working arrangements, operational arrangements and other informal cooperation risks producing a fragmented framework in which responsibilities are dispersed, and the applicable safeguards differ depending on the form of cooperation.

Frontex’s existing framework of shared responsibility has long been identified as insufficient for attributing legal responsibility for human rights violations within an operational system involving multiple actors, allowing for increased impunity for border violations. The proposal fails to address these concerns. Article 7 on shared responsibility is simply amended (arguably in response to WS vs Frontex) to include the specific responsibilities of the agency in the area of returns, stipulating that Frontex shall be responsible for the strategic planning, prioritisation, coordination and implementation of return operations, for forced and voluntary returns, and other return-related activities. A report commissioned by The Greens/EFA Group provided concrete recommendations on how Frontex can ensure a clearer regime of responsibility allocation, including reforming Article 7 to include a model of responsibility more similar to the Europol Regulation. In particular, Art. 50 of the Europol Regulation stipulates that an individual who has suffered damage may seek compensation either from Europol or from a Member State (see also Kočner v Europol on joint and several liability). Furthermore, the report calls for a protocol on responsibility allocation before any operational deployment and concrete consequences for breaches. Given the agency's ever-increasing data management competencies, failing to include a provision on responsibility for unlawful data processing, similar to the one in the Europol Regulation, would be a missed opportunity.

Monitoring and Accountability Safeguards

Alarmingly, the proposal would weaken existing monitoring and accountability safeguards. Under the current regulation, Fundamental Rights Monitors (FRMs) would ‘constantly’ monitor the agency’s operational activities and at least one of them must be assigned to each operation (Art. 110(1) and (3) of Regulation (EU) 2019/1896). The proposal would replace these requirements with a more discretionary, risk-based approach (proposed Art. 130(1)(4), preamble 134). FRMs would be allocated to an operation based on a risk assessment prepared by the Fundamental Rights Officer in coordination with the Executive Director. Importantly, the number of FRMs, which already do not suffice to cover the agency’s activities, remains unchanged (40) even though the agency’s operations would be increased (proposed Art. 130(7).

The proposal goes further in the context of the Quick Reaction Force, opting for ex-post fundamental rights compliance reviews. Under proposed Art. 72, where a Member State faces a crisis or emergency at its external borders, fundamental rights monitoring may take place up to one month after the operation has been completed, unless the Member State otherwise requests. The concern is particularly acute because it would leave potential violations to be identified only after the fact, which increases evidentiary complications and does not allow for an immediate response.

Moreover, the proposal attempts to limit the scope of the Fundamental Rights Office's operations. According to Art. 129(3)(e), FRMs would only report violations by the agency’s own staff, leaving out incidents that may be related to Frontex-deployed personnel or the national authorities of Member States and third countries. Monitoring in the context of the deployment of the Quick Reaction Force would also focus only on the agency's activities rather than the entire operation (proposed Art. 72(5)). The scope of the Fundamental Rights Officer’s (FRO) activities is also affected. Even though the relevant Art. 129 remains mostly unchanged, the FRO is prompted in the preamble to focus his work on the activities performed by the agency’s own staff, while simply transmitting to the relevant national authorities information about ‘incidents that may concern them’ (proposed preamble 133, and Art. 129(2)(c)).

Finally, the ultimate power (in fact an obligation) of the Executive Director to suspend or terminate an operation in case of continuing serious fundamental rights violations is significantly being watered down. In particular, under the new regulation, the Executive Director would be required to first consult with the agency's Management Board before making such a decision (proposed Art. 52 (8)).

In his presentation of the new Regulations before the LIBE Committee of the European Parliament, Magnus Brunner, the EU Commissioner responsible for internal affairs and migration, stated in unequivocal terms: ‘The fundamental rights framework is actually among the strongest of any EU agencies and that stays the same. We’re not changing anything.’ Despite this strong assertion, the proposal only further weakens the agency’s already flawed accountability framework.

Conclusion

The new legislative proposals consolidate and significantly extend three mutually reinforcing trends in EU migration governance: the institutional expansion of EU agencies, the digitalisation and datafication of asylum, border and return management, and the territorial expansion beyond EU borders. Frontex is no longer envisaged primarily as a coordinating agency supporting Member States at the external borders. It is increasingly positioned as an operational and informational hub at the centre of the EU’s return, border surveillance and even internal security architecture, with responsibilities extending from the collection and analysis of information to the implementation of returns in third countries.

At the same time, the proposal expands the range of actors, technologies, and databases involved in these activities, creating an increasingly complex governance structure in which decisions and actions are covered by several layers of complexity, opacity, and plausible deniability. Yet this expansion of operational and informational power is not matched by an equivalent strengthening of fundamental-rights safeguards. The proposal appears to shift from systematic monitoring to risk-based deployment of FRMs and, in certain emergency operations, ex-post review. It would further constrain the FRO's powers and politicise the decision to suspend or terminate operations by opening it up for debate at the Management Board. The result is a significant asymmetry between the expansion of the agency's operational, informational and territorial reach and the mechanisms available to scrutinise, prevent and answer for violations.

As the proposal is dissected in the coming months, these considerations are expected to feature prominently in both the academic and political debates. As negotiations will take place in the hands of a European Parliament and Council heavily focused on security, the alertness and rigour of scholars and civil society are imperative to ensure that fundamental rights protection receives appropriate weight in the upcoming negotiations.