Issue attention on international courts: Evidence from the European Court of Justice

For a discussion of case selection on the US Supreme Court see Perry (2009), Caldeira et al. (1999) and Kastellec and Lax (2008).

Supplementary Materialstogether with replication files are available on the website of the Review of International Organizations.

This rule knows, of course, a number of exceptions. In parliamentary systems, legislative procedures may afford the executive branch substantive control over the agenda of the legislature (Huber 1992). In the EU, the treaties severely restrict the agenda-setting powers of the European Parliament, the democratic world’s largest supranational assembly.

That policy questions are raised within the context of dispute resolution severely restricts the scope for issue creation. This is because policy determinations are expected to bear some relevance to the dispute being resolved (Cameron and Kornhauser 2013). Judges cannot easily escape this institutional constraint.

The actio popularis, which allows any person to bring action in the interest of the public, can be viewed as an extreme variant of such an institutional arrangement. Such a procedure existed in Hungary between 1990 and 2011 before the Constitutional Court. The Hungarian actio popularis led to an explosion in constitutional litigation and allowed the Hungarian Constitutional Court to become one of the world’s most activist judicial constitutional tribunals during this period (see Scheppele 2005).

The ability to dismiss cases that are “manifestly inadmissible” or references that pertain to “settled case law” affords the ECJ a limited measure of negative agenda control (Craig and Burca2015, 484). There has been no empirical research of note on issue suppression on the ECJ. We point it out as an interesting area for future research.

The coding protocol for the Court of Justice Cases dataset compiled by Clifford Carrubba advises against using issue area codes from the ECJ case books and refers to correspondence with the Court as indication that the choice of terms is not based on a well-defined coding scheme from the Court. See http://​polisci.​emory.​edu/​home/​people/​carrubba_​ecjd/​ECJ_​Access_​Data_​Codebook.​pdf(Accessed 5 July 2018).

Compared to topic modelling, corpus comparison is a less developed research field. For a survey of the literature see Kilgarriff (2001) and Remus and Bank (2012).

More recent years were still incomplete at the time of writing.

Only about ten references were either withdrawn or dismissed between 1961 and 1995.

For referrals, these are the words: “court”, “preliminary”, “refer”, “question”, “main”, “proceedings”, “proceeding”, “judgment”, “referring”, “referred”, “reference”, “meaning” and “ruling”. For infringements: “obligation”,“fail”, “failure”, “fulfil”, “state”, “opinion”, “obligations” and “failed”. For annulments: “parliament”, “commission”, “european”, “government”, “member”, “annulment”, “council”, “decision”, “state”, “contested” and “adopted”.

The case with the highest 𝜃_monarchies is Kingdom of Spain v United Kingdom of Great Britain and Northern Ireland over Gibraltar, one of the rare infringement actions initiated by a national government. Unsurprisingly, less than a handful of decisions have a 𝜃_monarchies larger than 0.5.

The classification of landmark constitutional rulings amakes intuitive sense. Simmenthal, a case arising from a dispute over meat imports in which the ECJ held that every domestic court had the power to set aside domestic legislation contrary to EU law, is classified as 63% constitutional. Costa v ENEL, in which the ECJ first spelled out the principle of supremacy of EU law, is 48% constitutional. Factortame I, the first time an act of the British Parliament was declared contrary to EU law, shows a similar topic proportion (47 percent). So too does Van Gend en Loos, the 1963 ruling that gave birth to the doctrine of direct effect (43% constitutional). At 30% the topic proportion for Francovich, in which the ECJ established the principle of state liability for failure to transpose EU directives, is somewhat lower, although still plausible.

Ordered by year the court became operational these are: the Benelux Court (became operational in 1974), the Andean Tribunal of Justice (1984), the Central American Court of Justice (1992), the European Free Trade Area Court (1992), the West African Economic and Monetary Union (1995), the Common Market for East African States (1998), the Central African Monetary Community (2000), the East African Community Court (2001), the Caribbean Court of Justice (2001), the Court of Justice of the Community of West African States (2001) and the Southern African Development Community (2005).