🍻 Selber Bier Brauen

With the increasing proliferation of deepfakes and rapid development of artificial intelligence, Denmark has become the first country in the European Union to specifically protect one’s image and voice with a new legislative initiative. As of 31 March 2026, a new intellectual property right is expected to enter into force, modelled as a neighbouring right to copyright law and specifically designed to protect a person’s voice and physical appearance. Traditionally, voice and image have been protected as personality rights that are not transferable, primarily defensive in nature, and only partially commercial. Denmark’s new legislation marks a significant departure from this tradition by reconceptualising voice and image as intellectual property rights, making them potentially transferable and commercially exploitable.

While this step is important, particularly for performers, it cannot replace personality rights, which remain indispensable in cases such as deepfake pornography. Furthermore, given the inherently cross-border nature of deepfake cases, a purely national solution will not suffice.

Deepfakes and the right to one’s own image

Deepfakes are digital manipulations of image, video, or audio material and give rise to a range of risks. They are frequently used to produce so-called deepfake pornography, can be employed for political misrepresentation or the spread of misinformation and, moreover, make it possible to imitate artists, for instance by generating new songs or films with their voice and appearance.

What these scenarios have in common is that virtually any deepfake depicting a natural person interferes with an individual’s right to their own image, irrespective of whether the person concerned is a public figure or a private individual. What distinguishes them, however, is the nature of the harm involved and, consequently, the protective aim pursued. Deepfake pornography primarily harms the individual depicted, interfering with their intimate sphere and personal dignity. Deepfake music and film production similarly affect the individual but shift the focus towards the economic value associated with their identity. Political deepfakes, by contrast, pose a threat not so much to the politician depicted as to the public at large, undermining informed debate and democratic discourse. The protective aim is here less about safeguarding the individual and more about protecting society from manipulation and misinformation.

At the European level, the right to one’s own image is derived from Article 8 of the European Convention on Human Rights, which guarantees the right to respect for private and family life. According to the case law of the European Court of Human Rights, this right encompasses an individual’s control over the use of their image. A comparable approach can be found in French law. Article 9 of the Civil Code protects privacy (vie privée) and thus forms the basis of the French right to one’s image (droit à l’image), which, although not codified, has been developed through judicial decisions. In Germany, the right to one’s own image is recognised both at the level of statutory law and constitutional law. At the statutory level, it is primarily governed by sections 22 and 23 of the Kunsturhebergesetz (Copyright Act for Artistic Works). At the constitutional level, it is understood as an aspect of the general right of personality, derived from Article 1(1) in conjunction with Article 2(1) of the Grundgesetz (Basic Law).

By contrast, the position in the Anglo-American legal sphere is markedly different. The United Kingdom does not recognise a freestanding general right of personality, nor a specific right to one’s own image. Historically, this approach has been justified, inter alia, on the basis that images are generally regarded as carriers of information, the dissemination of which may serve a legitimate public interest and for the freedom of expression. In the United States a distinction is drawn between the Right of Privacy and the Right of Publicity. Whereas the former protects an individual’s private sphere, the latter enables, in particular public figures, to control the commercial exploitation of their name, image, or identity. In several U.S. states this right is expressly conceptualised as a form of intellectual property right.

A regulatory gap

Deepfakes present a novel legal challenge in a number of respects and expose a potential regulatory gap. In response, several Member States of the European Union, most recently Denmark, have begun to pursue legal approaches that increasingly resemble those found in the United States. This development suggests a possible shift in the conceptualisation of the right to one’s own image: traditionally framed as a personality right in most continental European countries, it may gradually be recast along the lines of an intellectual property right. This shift, however, has attracted criticism. Accordingly, the right to one’s own image is intrinsically and closely linked to the personality of the individual depicted and should therefore not be an economic intellectual property right.

However, what distinguishes deepfakes from earlier technological developments, such as photography, which historically prompted the emergence of image rights as a response to the unprecedented ability to reproduce a person’s image without their consent, is that deepfakes merely generate avatars rather than reproducing real moments from a private or public life. Existing control mechanisms in this area are therefore largely limited to injunctive relief and defensive claims. Furthermore, the notion that images inherently convey a form of “truth” must now be regarded as outdated. Deepfakes do not represent the underlying personality, but rather the body as a mere external form. Whereas economic value was traditionally created through an individual’s artistic works, such as songs or performances, that value is now increasingly attached to the body itself, to the “shell”, which, although it cannot be entirely separated from the personality, can nevertheless be distinguished from it far more clearly than in the past, when images were assumed to depict the real person and an authentic moment. Licencing one’s appearance remains problematic in most European jurisdictions, as the human body cannot be “sold”. However, the issue here is not the commodification of the body itself, but rather the use of a person’s digital representation.

In certain respects, therefore, deepfakes nonetheless exhibit parallels with copyright law. In jurisdictions such as Germany and France, copyright follows the model of Author’s Rights, encompassing not only economic rights but also moral rights that protect the personal bond between author and work. Accordingly, copyright in continental Europe maintains a strong personality-based dimension. By contrast, the United Kingdom and the United States adhere more closely to a copyright model that primarily emphasises economic exploitation rights.

In both German and French law, an intellectual property-based approach to image rights is generally rejected so far. However, economic rights in relation to one’s image are already recognised, but whether these must be explicitly regulated under intellectual property law remains an open question. For example, the Bundesgerichtshof (German Federal Court of Justice) has not yet ruled definitively on this matter.

New intellectual property right

Denmark is currently pursuing an approach that is also being actively debated for adoption in the Netherlands. The Danish approach is particularly noteworthy in that it distinguishes between two categories of individuals: ordinary persons and performers. On the one hand, performers, understood as performing artists who do not hold copyright in the underlying work but instead enjoy rights in their performance, are to be granted specific rights over their voice and appearance. On the other hand, the proposed framework seeks to extend protection to individuals who are not performers. Denmark’s initiative does not extend copyright to cover voice and image, but rather creates an entirely new intellectual property right, modelled on the neighbouring rights already familiar from copyright law, such as those enjoyed by performers or broadcasters.

However, there remain strong voices arguing that the right to one’s own image should continue to be conceptualised as a personality right. One major advantage of this approach lies in its close connection to the individual. The human body and personality are inherently intertwined; it is difficult, if not impossible, to separate a person’s physical appearance from their identity. In this sense, the right to one’s own image is fundamentally an expression of personality. This understanding is particularly relevant in light of the reality of deepfakes. A significant proportion of deepfake content, often cited as around 96 per cent, consists of pornographic material, disproportionately affecting women and children. Such uses clearly implicate personal dignity and identity, reinforcing the view that protection should be grounded in personality rights. In this context, P. Bernt Hugenholtz argues that different harms associated with deepfakes should be addressed within their respective legal frameworks. In particular, misinformation primarily concerns the protection of the public rather than the individual depicted. In such cases, regulatory measures, such as transparency obligations, may be sufficient. This approach is reflected, for instance, in the Article 50 EU AI Act, which introduces transparency requirements for AI-generated content, while deepfake pornography is more appropriately addressed within the scope of personality rights.

However, there are also disadvantages to this approach when applied to deepfake usage. Personality rights primarily function as defensive rights, which may lead to gaps in protection. A significant gap arises from the imbalance between the stronger economic protection afforded to creative works and the comparatively weak protection afforded to a person’s voice and physical appearance. While copyright holders benefit from a well-developed system of economic rights, including licensing, transferability and enforcement mechanisms, no equivalent comprehensive framework exists for personal attributes such as likeness and voice. This asymmetry becomes particularly evident in deepfake cases and complicates, among other things, the assessment of damages due to the absence of established valuation standards. In practice, image and voice are increasingly subject to commercial exploitation through contractual and licensing arrangements. The economic value of the right to one’s own image is therefore likely to continue to rise in the future. Moreover, robust protection of image and voice rights may help to preserve the incentive structure underlying copyright, which is already under pressure from AI, by encouraging the creation of art and new works without fear of unauthorised exploitation.

This reflects a broader shift in which voice and appearance are treated as economic assets. From this perspective, recognising an intellectual property dimension in the right to one’s own image aligns with current market realities.

Conclusion and outlook

Reconceptualising the right to one’s own image, particularly for performers, as an intellectual property right is necessary to address the growing commercial exploitation of voice and appearance in the digital age. Such a framework, must, however, incorporate moral rights at least as strong as those found in continental European copyright law, if not stronger, given the uniquely personal nature of the attributes at stake. Unlike a mere economic right, as seen in the United States, this would ensure that the personality interests underlying voice and image remain protected even where commercial exploitation is permitted.

In other situations, especially cases involving deepfake pornography, stricter measures are required, many of which are more appropriately rooted in personality rights. Deepfake misinformation cases, however, present a particular challenge. They raise more direct conflicts with freedom of expression and pursue a somewhat different protective aim, as they are concerned not only with safeguarding the individual depicted but also with protecting the public from deception and manipulation.

It appears unlikely that countries such as France or Germany with strong personality rights will take immediate legislative action in this area. However, the European legislator may be better positioned to address these issues at a supranational level, not least in order to prevent forum shopping. Harmonisation at the European level is particularly important in light of digitalisation and the inherently cross-border nature of these cases. The European Union has already begun to respond to these challenges, most notably through the EU AI Act.

Nevertheless, significant questions remain, particularly with regard to implementation and enforcement. Violations of personality and intellectual property rights on the internet are likely to persist, and the cross-border dimension will often make it difficult, if not impossible, to identify and hold perpetrators accountable. European states have attempted to address this issue, in part through criminal law measures, as reflected for example in provisions such as a newly proposed § 201 b of the German Criminal Code in Germany and the Article 226-8 Code Pénal in France.

The introduction of a new intellectual property right in respect of voice and appearance is a meaningful step forward, and one that extends protection to both performers and ordinary people, giving both categories a robust right over their digital likeness. It can, however, only address certain categories of deepfake cases. The deeper challenge remains one of implementation and enforcement, and it is here that the European Union has the opportunity to take meaningful steps towards advancing coherent protection across the European legal space.

The post Copyrighting Voice and Image appeared first on Verfassungsblog.

Over the weekend, Italians resoundingly rejected the Meloni government’s constitutional reform on the overhaul of the judiciary via referendum. With the “No” side receiving 53.2% of the popular support, with an unexpectedly high turnout at 55.7%, this is Meloni’s first political defeat since becoming Prime Minister in 2022. The consequences of the referendum show that Italian checks and balances are stronger than one might have feared, despite the fact that, on the eve of the referendum, V-Dem listed Italy as an autocratizer for the first time amid a series of illiberal policies implemented by the sitting government. However, this is not the last challenge for Italian democracy. As Meloni is likely to adjust her strategy, and with new electoral reforms on the horizon, Italian democratic resilience will soon face another real test.

Background on the reform

In broad strokes, the reform proposed by the government (and passed through the two houses of Parliament without a single amendment), can be broken down into three parts. Firstly, the separation of career paths within the judiciary, going from the more general umbrella career path for “magistrates”, to one distinguishing between “judges” and “prosecutors”. Secondly, and connected with the former, the separation of the unitary self-governance body, the High Judicial Council (Consiglio Superiore della Magistratura or CSM), into three distinct bodies: a CSM for judges; a CSM for prosecutors; and a new High Disciplinary Court, with rules to be defined via ordinary legislation. The third and final change pertained to the appointment to the three bodies, switching from election by peers, to sortition (for more commentary on its merits, see here).

Supporters of the “Yes” vote welcomed these changes as heralds of greater judicial independence, especially in light of scandals related to political “currents” or “factions” within the magistracy that had suggested lack of impartiality within the CSM. Moreover, they saw this as an opportunity to reduce phenomenon of judges leaning in favour of prosecutors, a problem that criminal defence lawyers have long decried, and the reason why a majority of them campaigned in favour of the reform.

The “No” side, on the other hand, highlighted the broader rule of law consequences of the reform, including a greater politicisation of the judiciary by fundamentally weakening the CSM through the new election process, but also potentially drawing the prosecution service closer to the executive. Furthermore, the lack of clarity on the legislation needed to actualise the reform – e.g. defining the parameters for compiling the lists from which to draw lots, new rules on disciplinary proceedings, etc – raised fears about potential abuse by the political class.

Lessons from the vote

As many commentators pointed out, the judicial overhaul proposed by the government bore an undeniable resemblance to illiberal moves witnessed in countries like Poland or Romania, introducing a new disciplinary regime for judges and new rules that would undermine judicial self-governance potentially running afoul of European standards (see the new Venice Commission rule of law checklist) and EU values (as per established CJEU case law). Therefore, as an immediate outcome of the vote, it would appear that the spectre of autocratic decay has been vanquished, at least momentarily. Nonetheless, I would argue that this referendum offers more opportunities for reflection beyond the obvious.

The data emerging from the referendum itself paints an interesting picture. Only three mainstream (opposition) parties campaigned for “No”, representing a minority of voters at the previous parliamentary as well as at the last European Parliament elections, and this despite the political turn that the “Yes” campaign took in the final days. The result nevertheless suggests a sharp inversion in the trend of support Giorgia Meloni has enjoyed since the beginning of her mandate. The fact that voters did not follow party lines speaks to a continued trust and belief in the robustness of Italy’s post-fascist Constitution, and the solidity of the checks and balances it created. According to YouTrend data, 61% of “No” voters were motivated by a desire to not amend, and thus preserve, the Constitution, presumably out of concern that the reform would overall weaken the text.

At the same time, it is worth noting that this result is broadly in line with other referendums of this kind. Out of five constitutional referendums in the history of the Republic, only two have succeeded – in 2001 and in 2020. Moreover, an abrogative referendum to overhaul the judiciary was attempted as recently as 2022, but it failed to pass the turnout threshold with only 20.5% of eligible citizens casting a vote. Arguably, the lack of a turnout threshold for the 2026 referendum should have benefited the proponents of the reform, as mobilisation was not an imperative. However, the final turnout registered was the second highest this century, only below the even more controversial and politically charged referendum on the overhaul of the political system proposed by the Renzi government in 2016. As a whole, this can be seen as a good sign for Italian democracy, and a strong response against the populist rhetoric promoted by the government – including high-ranking officials within the Ministry for Justice stating that their objective was to “get rid of the judiciary”, equating it to “firing squads”.

Now what? Wider consequences of the vote

While checks and balances have proven solid in this instance, expecting the referendum to fold future attempts to skew the separation of powers in favour of the executive would be naïve. The Meloni government has already eroded the rule of law significantly outside the realm of judicial independence, including by undermining the public sphere, restricting the rights of marginalised groups, and weakening the anti-corruption framework (see here for a comprehensive overview). Despite enjoying widespread popular support and having achieved nearly unprecedented stability, what the government lacks is a strong enough majority in Parliament to amend the Constitution freely (2/3 majority required). This reform, beyond its content, was the government’s first opportunity to test its consensus, before proposing more far-reaching reforms that would have a much costlier political impact.

Unlike Renzi in 2016, or David Cameron with Brexit, Meloni was careful not to link the referendum to her political survival, never suggesting she would resign in the event of a loss. Only in the last days of the campaign she appeared more visible in the media, to no avail. This experience will provide a good indication of whether there is scope to attempt a referendum again in the last year of her mandate, before the new elections expected in 2027 may redraw the Parliament’s composition.

Perhaps it is no coincidence that “the mother of all reforms”, a constitutional amendment introducing the direct election of the Prime Minister and revolutionising the system of government accordingly, has been effectively put on hold. Instead, last month, the government presented a bill to amend ordinary electoral legislation. The main objective would be that of granting an automatic parliamentary majority (potentially up to 60%) to any coalition winning more than 40% of the vote, in order to increase stability and a government’s ability to legislate. This measure, if approved by the legislature and if indeed in line with previous jurisprudence of the Italian Constitutional Court, would open new political scenarios. A more powerful executive could perhaps amend the Constitution more easily and more frequently, with potential consequences for democracy and the rule of law. This is especially evident now, as the latest referendum showed that the electorate is not always aligned with their representatives on systemic reforms changing the constitutional text.

Lastly, it is important to acknowledge that although the electorate has demonstrated its support for and trust in the magistracy as an institution, the Italian judicial system is far from flawless or above reform. As the 2025 Rule of Law Report highlights, persistent staffing shortages and the length of judicial proceedings remain serious problems. Moreover, complaints put forward by criminal defence lawyers, including regarding the high number of unfounded criminal prosecutions and requests for precautionary measures, low level of scrutiny in preliminary and pre-trial hearings, and subsequently the high number of appeals, should not be ignored. Nevertheless, this reform has not effectively addressed these particular issues, despite the claims in the referendum campaign suggesting otherwise. Indeed, even the government itself admitted that the reform would not allocate much needed resources or otherwise address these concerns. It rather obfuscated these problems with populist claims that the reform would stop judges from fostering immigration and that, with a “No” win at the referendum, magistrates would “free rapists and paedophiles”. Therefore, as badly as a reform of the justice system is needed, this proposal was far from the answer.

Conclusion

The popular rejection of Giorgia Meloni’s judicial overhaul is ultimately a good sign for Italian checks and balances, as it, perhaps surprisingly, constitutes her first set back since 2022. This is good news for Italian checks and balances in the short term. However, the government may become better in the future at executing its autocratic playbook that seeks to undermine the judiciary. Meloni continues to enjoy a large majority to pass ordinary legislation, which she has consistently used to undermine the rule of law, and the momentum is unlikely to stop in the final year before the general elections. If anything, I would argue, this referendum was more of a test for the Italian people (and civil society), than for the government. The electorate responded, but this just means that Meloni will change her strategy and act accordingly from now on (see: the various referendums called and lost by Orban, before he simply stopped using them as a tool to test consensus). With new electoral legislation on the horizon, the true resilience of Italian democracy will soon be tested yet again.

The post No, Grazie appeared first on Verfassungsblog.

Veto threats are ordinary currency in Brussels. A veto against an agreed compromise, used to force concessions on an unrelated dispute and to stage a domestic election campaign, is not. The events of 19 March 2026 were serious not only because Viktor Orbán blocked money for Ukraine, but because he did so after having promised in December 2025 not to stand in the way. This time Orbán went too far – if the other leaders fail to respond effectively, they will be teaching everyone that the most profitable strategy is blackmail.

Consensus is more than the veto rule

The European Council is an unusual institution. It does not legislate in the ordinary sense. Under Article 15 of the TEU, it sets the Union’s general political directions and priorities. Yet because it gathers the heads of state or government, its conclusions carry exceptional political weight. Formally, the European Council decides mostly by consensus. Practically, that means outcomes are built through negotiation, bilateral contacts, and brokerage by the President rather than by visible votes.

Veto threats are therefore not an aberration. They are part of the system. National leaders use them to signal domestic red lines, to demand clarifications, or to improve their bargaining position. Sometimes a government threatens to block sanctions until a technical exemption is secured. Sometimes it resists a paragraph in summit conclusions until wording is softened. Much of this belongs to the normal repertoire of intergovernmental politics. The mere use of a veto threat is not yet sabotage.

What makes the system work is something more fragile than treaty text. The European Council relies on informal norms: trust that commitments made in the room will still bind tomorrow; reciprocity among leaders who know they will need one another on the next file; and a compromise culture that discourages pressing every formal right to its absolute limit. Those norms allow unanimity to function without turning every summit into institutional paralysis.

That is also why there is a line between hard bargaining within a file and holding unrelated files hostage. A veto can be legitimate even when it is politically costly. But it remains legitimate only if it is embedded in self-restraint. Once a member treats consensus not as a method of joint problem-solving but as a resource to be monetized again and again, the institution’s informal foundations begin to erode. The problem is then not only delay. It is the loss of confidence that compromise will stick.

Orban’s long practice of illegitimate issue linkage

Seen against that background, Orbán’s conduct on Ukraine is not an isolated outburst. It is the latest episode in a longer strategy of illegitimate issue linkage. Linkage as such is not unusual in European Council politics. Package deals are often necessary, and some issues are genuinely connected. The difficulty arises when consent on one matter is made conditional on concessions in another matter that is only weakly related or entirely unrelated, and when that tactic is used not to solve a policy problem but to exploit unanimity for leverage.

Hungary has repeatedly moved in that direction. In 2022, Orbán blocked an EU loan for Ukraine while fighting Brussels over frozen funds tied to rule-of-law concerns, and it also held up the EU’s implementation of the global minimum corporate tax. In June 2023, both Hungary and Poland refused to sign off on European Council language on migration after being outvoted on migration rules elsewhere. In late 2023 and early 2024, Orban again blocked a major Ukraine package while arguments raged over the release of EU funds to Budapest. And in February and March 2026, Hungary tied both a new sanctions package on Russia and the implementation of the €90bn loan for Ukraine to the restoration of Russian oil flows through the Druzhba pipeline.

This matters because it imposes real institutional costs. It creates blockade. It consumes summit time and diplomatic energy. It distracts the European Council from its strategic role by forcing it into permanent damage control. And it harms the institution’s reputation. European Council conclusions matter because they signal reliability to partners, markets, and adversaries. If solemn summit bargains can be reopened at will for unrelated side-payments, the credibility of the Union’s highest political forum suffers.

There is also a deeper asymmetry here. Consensus only works if accommodation runs in all directions over time. Orbán’s method reverses that logic. The 26 national leaders accommodate, while one government repeatedly converts accommodation into leverage for the next fight. If that tactic is normalized, others might learn from it. The result is not simply one more difficult European Council summit. It is a degraded decision-making environment in which free-riding and grandstanding become rational strategies.

Why March 2026 crossed a red line

The events of 19 March 2026 were serious not only because Orbán blocked money for Ukraine, but because he did so after having promised in December 2025 not to stand in the way. The European Council had already agreed in December to provide a €90bn loan to Ukraine for 2026 and 2027. The design of that package mattered politically. It was based on EU borrowing backed by budget headroom, and the December agreement explicitly ensured that the mobilization of EU budget resources would not affect the financial obligations of Hungary, Czechia, and Slovakia. In other words, Orbán had secured an opt-out from financial liability. Hungary was not being asked to shoulder the cost.

By February 2026, the EU had advanced the legal framework for implementing that December agreement. Orbán then used the next European Council to reopen a deal he had already accepted. His public justification was a dispute over the Druzhba pipeline. After a Russian attack damaged pipeline infrastructure in January, Hungary demanded guarantees that oil flows through Ukraine would resume and would not be interrupted again. But whatever the merits of Hungarian energy concerns, this was not an objection to the design of the Ukraine loan itself. It was an unrelated dispute pressed into service as leverage.

That is what makes this episode qualitatively different from ordinary veto politics. Orbán did not discover a hidden burden on Hungary. He did not point to a newly revealed legal defect in the loan instrument. He took a settled bargain, one already tailored to shield Hungary from direct liability, and converted it into a pressure tool in a separate conflict. The move was therefore not simply hard bargaining. It was bad-faith reneging on his own word.

The domestic motive is not hard to infer. Orbán faces national elections in Hungary on 12 April 2026. For years, his political method has relied on the construction of external adversaries – “Brussels”, migrants, and more recently Ukraine – in order to mobilize domestic voters around a sovereignty narrative. The March veto fits that script perfectly. It lets him present himself as the defender of Hungarian households against foreign pressure and turns the European Council into a stage for domestic campaign politics. That is why António Costa, the European Council president, is right to say that a deal is a deal and that nobody should be allowed to blackmail this institution. Germany’s chancellor Friedrich Merz was right, too, to call Orban’s conduct a grave disloyalty among EU member states. Trust is a governance resource in this institution. If leaders no longer believe that a December compromise still means something in March, the willingness to compromise at all begins to collapse.

What the other leaders should do now

The other national leaders should now stop treating this as one more nuisance to be managed through patient appeasement. That approach has been tried for years, and it has taught Orbán that the short-term costs of obstruction are low while the domestic political rewards can be high. If that incentive structure remains unchanged, repetition is likely.

First, national leaders should escalate rhetorically and openly call out what happened. Members of the European Council usually do not antagonize one of their own in public. Under ordinary conditions, that instinct is understandable. But norm enforcement sometimes requires visible disapproval. Public naming changes the reputational calculus and clarifies that this is not a legitimate disagreement over policy design but a breach of trust.

Second, they should do what they normally avoid doing: build intergovernmental solutions outside the usual EU framework when one member abuses unanimity to block urgent action. Enhanced cooperation is one route within the treaties; ad hoc intergovernmental arrangements among willing states are another. The crucial point is that delivery to Ukraine cannot remain hostage to one individual domestic campaign strategy.

Third, leaders should increase pressure on Orbán both financially and politically. That means stricter use of existing rule-of-law conditionality, refusal to release funds as the price of tactical cooperation, and renewed seriousness about Article 7 TEU to eventually suspend Hungary’s voting rights. It also means refusing to normalize issue linkage by rewarding it. The broader lesson is stark. Unanimity can coexist with frequent veto threats. It cannot coexist with systematic bad faith. On 19 March 2026, Orbán crossed that line. If the other leaders fail to respond more confrontationally now, they will not be defending European unity. They will be teaching everyone that the most profitable strategy in the European Council is blackmail.

The post A Deal Is a Deal appeared first on Verfassungsblog.