Samir El Khanza

The core case is the Comprehensive Economic and Trade Agreement (CETA) between the EU and Canada. Concluded in August 2014, it is among the most far-reaching bilateral agreements the Union has negotiated. Because it also touched shared and Member State competences, the question of whether it could be concluded as an EU-only agreement or had to be classified as mixed became the central institutional conflict. The Commission conceded mixity in July 2016, under pressure from several governments, in a decision that Commissioner Malmström described as political rather than legal. That concession mattered structurally, since mixity multiplied the consent arenas and turned every national and, where relevant, subnational parliament into a potential veto player. What followed confirmed the point. The Walloon Parliament nearly derailed signature in October 2016, and the crisis was resolved only through the Joint Interpretative Instrument and its thirty-eight accompanying statements. CETA was signed on 30 October 2016, received European Parliament consent in February 2017, and has been provisionally applied since September 2017, though provisional application deliberately excludes investment protection and the Investment Court System. Almost a decade later it has still not fully entered into force.

The comparison rests on four Member States, selected under a most-similar-systems design with variation on the outcome. All four are founding members of the Community, among the largest Member States by population and economic weight, subject to the same EU legal framework during the CETA process, and confronted by the same agreement with the same contested provisions. These shared traits function as controlled variables. The four diverge, however, on the two dimensions the thesis sets out to explain: domestic institutional structure, which determines legal strength, and the form and intensity of parliamentary contestation, which determines whether that strength is activated. They diverge, too, on the outcome itself, since each followed a different CETA pathway.

Belgium offers the strongest legal architecture of the four, and the most unusual in Europe. Its governing principle, in foro interno, in foro externo, was established by the Special Law of 8 August 1980 and constitutionalised through the 1993 Sint-Michiels reform. Whichever level of government is competent domestically is also competent externally, so a mixed agreement cannot be signed or ratified without the consent of every affected level. Operationalised through Article 167 BC and the Cooperation Agreement of 8 March 1994, this arrangement gives each subnational parliament an effective and independent veto, a configuration with no equivalent among the other cases or, indeed, anywhere else in the Union. The Walloon vote of 14 October 2016 was its visible expression. That Wallonia, the French Community and Brussels-Capital did not not adopt their decrees of assent in more than eight years is the more durable one.

Germany concentrates treaty-making at the federal level under Article 59(2) GG, with ratification passing through the Bundestag and requiring Bundesrat consent wherever Länder competences are affected. Its constraining force, though, runs through two channels the others lack. The first is the depth and enforceability of parliamentary oversight in EU affairs, structured by Article 23 GG, the Wesentlichkeitslehre, and the Bundesrat's collective representation of the Länder. The second, and the more distinctive, is the accessibility and activism of the Bundesverfassungsgericht. Through the Verfassungsbeschwerde, more than 125,000 complainants challenged CETA's compatibility with the Basic Law, and the Court's interim rulings conditioned provisional application before it cleared the way in 2022. Ratification followed only afterwards, in December 2022.

The Italian Parliament looks weak ex ante and strong ex post. The instruments provided by Law No. 234/2012, its information rights and the Article 7 resolution mechanism, are formally articulated but rarely used; the Article 7 mechanism formally binds the government yet carries an escape clause that softens its effect, and there is no ex ante constitutional review of treaties. What Italy retains is a genuine ex post veto. Under Article 80 of the Constitution, and given symmetric bicameralism, both chambers must authorise ratification, so either can block the agreement simply by declining to act. That is what has occurred. No formal ratification procedure has been opened, and the constraint here works through inaction rather than refusal.

The French Parliament distributes its legal strength across three channels: ex post ratification under Article 53 FC, ex ante constitutional review under Article 54 FC, which a minority of sixty deputies or senators can trigger, and the asymmetric bicameralism of Article 45 FC, which allows the government to give the last word to the Assemblée nationale. The decisive variable, however, was political. The fait majoritaire carried ratification through the Assemblée nationale in 2019, by 266 votes to 213. Once the presidential majority was lost in 2022, that guarantee dissolved, and on 21 March 2024 the Sénat rejected the bill by 211 to 44, driven by cross-party agricultural opposition and the farmers' protest movement. The second reading in the Assemblée nationale has not been scheduled since, which leaves CETA in indefinite suspension.
Read side by side, the four cases describe a repertoire rather than a single mechanism.

What accounts for outcomes this varied is not legal strength on its own, but the interaction between each constitutional order and the political dynamics of parliamentary contestation. Comparable powers, activated under different conditions, produce very different results.