¹ Epstein, L. et al., The Behavior of Federal Judges: A Theoretical and Empirical Study of Rational Choice (Harvard University Press 2013)Google Scholar.

² Cameron, C.M. and Kornhauser, L., ‘Rational Choice Attitudinalism? A Review of Epstein, Landes and Posner’s The Behavior of Federal Judges: A Theoretical and Empirical Study of Rational Choice, <wws.princeton.edu/faculty-research/research/item/rational-choice-attitudinalism-review-epstein-landes-and-posner%E2%80%99s>, visited 22 March 2017 Google Scholar.

³ Ferejohn, J. and Pasquino, P., ‘Constitutional Adjudication: Lessons from Europe’, 82(7) Texas Law Review (2004) p. 1671-1704 Google Scholar.

⁴ Furlong, P., ‘The Constitutional Court in Italian Politics’, 11(3) West European Politics (1988) p. 7-23 Google Scholar and Pasquino, P., ‘Constitutional Adjudication and Democracy. Comparative Perspectives: USA, France, Italy’, 1(1) Ratio Juris (1998) p. 38-50 CrossRefGoogle Scholar.

⁵ Brouard, S., ‘The Politics of Constitutional Veto in France: Constitutional Council, Legislative Majority and Electoral Competition’, 32(2) West European Politics (2009) p. 384-403 Google Scholar; Hönnige, C., ‘The Electoral Connection: How the Pivotal Judge Affects Oppositional Success at European Constitutional Courts’, 32(5) West European Politics (2009) p. 963-984 Google Scholar; Hanretty, C., ‘Dissent in Iberia: The Ideal Points of Justices on the Spanish and Portuguese Constitutional Tribunals’, 51(5) European Journal of Political Research (2012) p. 671-692 Google Scholar; Garoupa, N. et al., ‘Political Influence and Career Judges: An Empirical Analysis of Administrative Review by the Spanish Supreme Court’, 9(4) Journal of Empirical Legal Studies (2012) p. 795-826 Google Scholar; Garoupa, N. et al., ‘Judging under Political Pressure: An Empirical Analysis of Constitutional Review Voting in the Spanish Constitutional Court’, 29(3) Journal of Law, Economics and Organization (2013) p. 513-534 Google Scholar.

⁶ Dalla Pellegrina, L. and Garoupa, N., ‘Choosing between the Government and the Regions: An Empirical Analysis of the Italian Constitutional Court Decisions’, 52(4) European Journal of Political Research (2013) p. 431-480 Google Scholar; Garoupa, et al. 2013, supra n. 5 Google Scholar.

⁷ Such as social security, defence, justice, the federal police, monetary policy and state-owned companies. It also holds responsibility for the obligations of Belgium and its federalised institutions towards the EU and NATO and controls substantial parts of public health, home affairs and foreign affairs.

⁸ On Belgian politics, see e.g. Deschouwer, K., ‘And the Peace Goes On? Consociational Democracy and Belgian Politics in the Twenty-first Century’, 29(5) West European Politics (2006) p. 895-911 Google Scholar and Swenden, W. et al., ‘The Politics of Belgium: Institutions and Policy under Bipolar and Centrifugal Federalism’, 29(5) West European Politics (2006) p. 863-873 Google Scholar.

⁹ Vanberg, G., The Politics of Constitutional Review in Germany (Cambridge University Press 2005)Google Scholar; Amaral Garcia, S. et al., ‘Judicial Independence and Party Politics in the Kelsenian Constitutional Courts: The Case of Portugal’, 6(2) Journal of Empirical Legal Studies (2009) p. 381-404 Google Scholar; Franck, R., ‘Judicial Independence under a Divided Polity: A Study of the Rulings of the French Constitutional Court, 1959-2006’, 25(1) Journal of Law, Economics and Organization (2009) p. 262-284 CrossRefGoogle Scholar; Franck, R., ‘Judicial Independence and the Validity of Controverted Elections’, 12(2) American Law and Economics Review (2010) p. 394-422 Google Scholar; Dalla Pellegrina, and Garoupa, , supra Google Scholar n. 6; Garoupa, et al. 2013, supra Google Scholar n. 5.

¹⁰ A few exceptions are De Jaegere, J. et al., ‘Exploring the Deliberative Performance of a Constitutional Court in a Consociational Political System. A Theoretical and Empirical Analysis of the Belgian Constitutional Court’, <ecpr.eu/Filestore/PaperProposal/1c16e505-e983-44c9-9935-8c573f521ab0.pdf>, visted 22 March 2017 Google Scholar; De Jaegere, J., ‘Inclusiviteit als delibera-tieve bouwsteen van legitimiteit: een empirische analyse van het Belgische Grondwettelijk Hof’, 4 Tijdschrift voor Bestuurs-wetenschappen en Publiekrecht (2015) p. 194-215 Google Scholar.

¹¹ This section provides a brief overview of the Belgian Constitutional Court. For more details, we refer to the website of the Court (see <www.const-court.be>, also available in English and French); see also Bocken, H. and De Bondt, W., Introduction to Belgian Law (Kluwer and Bruylant 2001)Google Scholar; Popelier, P., Procederen voor het Grondwettelijk Hof [Litigating in the Constitutional Court] (Intersentia 2008)Google Scholar; Velaers, J., Van Arbitragehof tot Grondwettelijk Hof [From Court of Arbitration to Constitutional Court] (Maklu Uitgevers 1990)Google Scholar and, Rigaux, M.-F. and Renauld, B., La Cour constitutionnelle (Bruylant 2008)Google Scholar.

¹² Referring more particularly to the fact that it was meant to be a referee between the central and different decentralised levels of government, see Peeters, P., ‘Expanding Constitutional Review by the Belgian “Court of Arbitration”’, 11 European Public Law (2005) p. 475-479 Google Scholar; Adams, M. and Van der Schyff, G., ‘Grondwettigheidstoetsing door de rechter als “list van de rijke”? Methodologische en andere vragen bij processen van rechtsverandering’, Tijdschrift voor Privaatrecht (2008) p. 913-977 Google Scholar at 928. For an historical account see Velaers, , supra n. 11, p. 1-85 Google Scholar.

¹³ See further Velaers, , supra n. 11, p. 545-557 Google Scholar.

¹⁴ Commentators hold that nominations take place on the basis of a political majority in the Chamber of Representatives and the Senate Moonen, T., ‘Graag meer aandacht voor besluitvorming Grondwettelijk Hof’, 284 De Juristenkrant, 26 February 2014, p. 12). ‘The twelve seats in the Constitutional Court are divided among the political parties’ (E. Maes, ‘Waarom geen hoorzitting voor nieuwe rechters Grondwettelijk Hof’, 279 Juristenkrant Google Scholar, 4 December 2013, p. 10).

¹⁵ Velaers, , supra n. 11, p. 604-606 Google Scholar.

¹⁶ Velaers, , supra n. 11, p. 607 Google Scholar.

¹⁷ Velaers, , supra n. 11 Google Scholar.

¹⁸ One of the judges reports on the case at a public hearing. The second reporting judge, from the other linguistic group, may file a supplementary report.

¹⁹ De Geyter, L., ‘De werkwijze van het Arbitragehof’, 27 OAPR, 21 September 2009, Comments on Articles 54-61 Google Scholar.

²⁰ De Geyter, , supra n. 19 Google Scholar.

²¹ Velaers, , supra n. 11, p. 454 Google Scholar.

²² There is evidence from the behavioural literature showing that the President can have a strong influence on the decision of a panel. On the one hand, low-effort judges may free-ride on the work of their colleagues; on the other hand, charismatic or powerful Presidents may push the decision in the direction they desire. Eisenberg and co-authors showed the influence of the President: see Eisenberg, T. et al., ‘Group Decision Making on Appellate Panels: Presiding Justice and Opinion Justice Influence in the Israel Supreme Court’, 19 Psychology, Public Policy and Law (2013) p. 282-296 Google Scholar. See also Biard, A., Judges and Mass Litigation. A (Behavioural) Law and Economics Perspective (dissertation, Erasmus University Rotterdam, December 2014) p. 275-276 Google Scholar.

²³ With the journal De Tijd, 27 September 2004, p. 5.

²⁴ Moonen, supra n. 14.

²⁵ Moonen, , supra n. 14 Google Scholar; Moonen, T., ‘De invloed van het Grondwettelijk Hof op de uitlegging van de Grondwet door andere rechters en door de wetgever’ [The influence of the Constitutional Court on the interpretation of the Constitution by other judges and the legislator], 4-5 Tijdschrift voor Bestuurswetenschappen en Publiekrecht (2015) p. 216-227 Google Scholar; Moonen, T., ‘De keuzes van het Grondwettelijk Hof: argumenten bij de interpretatie van de Grondwet’, 26 Rechtskundig Weekblad (2015-2016) p. 1003 Google Scholar and Moonen, T., De keuzes van het Grondwettelijk Hof (die Keure 2016)Google Scholar.

²⁶ Alen, A., ‘Wij zijn geen juristen in een ivoren toren’, 284 De Juristenkrant, 26 February 2014, p. 8 Google Scholar.

²⁷ De Jaegere, , supra n. 10 Google Scholar.

²⁸ Maes, E., De rol van een Grondwettelijk Hof in een rechtstatelijk perspectief (dissertation, Catholic University of Leuven, 29 January 2016)Google Scholar and Maes, E., ‘Een diverser Grondwettelijk Hof, voor meer legitimiteit en kwaliteit’, 284 De Juristenkrant, 24 February 2016, p. 8-9 Google Scholar.

²⁹ Vandormael, K.-J., Het Grondwettelijk Hof: rechter of regelgever. Analyse van de draagwijdte van de rechtspraak van het Grondwettelijk Hof (Larcier 2015) p. 38-68 Google Scholar; Popelier, , supra n. 11, p. 353-403 Google Scholar.

³⁰ Note that the Belgian Constitutional Court sometimes decides to moderate the retroactive effect of an annulment, thus giving the legislator some time to repair unconstitutional legislation.

³¹ Once again within six months of the publication of the disputed regulation in the Official Journal.

³² The effects of these rulings are somewhat different. The court that referred the preliminary question, and any other court passing judgment in the same case (for instance an appeal court), must comply with the ruling given by the Constitutional Court on the preliminary point of law. Where the Court finds a violation, the legislative act will remain part of the legal system. However, a new six-month term commences during which an action for annulment of the legislative act concerned can be brought forward. See Popelier, , supra n. 11, p. 230-268 Google Scholar; Alen, A., ‘Les questions préjudicielles posées à la Cour d’arbitrage. Règles générales, exceptions, etc. (on se comprit la situation du justiciable au recours de la procédure et l’autorité de la chose jugée de la décision de renvoi)’, in A. Arts et al. (eds.), Les rapports entre la Cour d’arbitrage, le pouvoir judiciaire et le conseil d’état, acte du symposium du 21 octobre 2005 (La Charte 2006) p. 153-194 at 164-172, andGoogle Scholar Sautois, J., ‘Saisir la cour constitutionnelle d’une demande de suspension’, in P. Martens (ed.), Saisir la Cour constitutionnelle et la Cour de justice de l’Union Européenne (Anthemis 2012) p. 55-84 Google Scholar.

³³ De Jaegere, et al., supra n. 10 Google Scholar; De Jaegere, , supra n. 10 Google Scholar.

³⁴ Since there is no formal hierarchy between the legislative acts of the different legislators in the Belgian federal system, it was necessary to create a court to resolve potential conflicts, see Adams, and Van der Schyff, , supra n. 12, p. 928 Google Scholar.

³⁵ See interview with the former President of the Court of Arbitration [now Constitutional] Alex Arts, De Tijd, 27 September 2004, p. 5: ‘Votes in our country are not made public… You can assume that consensus is the rule and that decisions based on a simple majority are the exception confirming the rule.’

³⁶ As explained before, the composition of the Panel is variable but exogenous to the decisions taken by the Court.

³⁷ It is sometimes alleged that the legal secretaries assisting the Rapporteurs in preparing the reports also play an important role. These secretaries have a university degree in law and are selected on the basis of an open competition, the terms and conditions of which are determined by the Court. Obviously, their political preferences would be more difficult to ascertain. Furthermore, the cases do not mention which legal secretary was involved in making the report.

³⁸ The court term starts on 1 September each year. The rotation of the presidency (between the different language groups) is considered a guarantee that one language group cannot dominate the other, see De Geyter, L., Article 142 Constitution in: Public Procedural Law. Comments per Article with an Overview of Case Law and Legal Doctrine, I, The Constitution (2011) p. 1-48.Google Scholar

³⁹ Shapiro, M., ‘Judicial Review in Developed Democracies’, 10(4) Democratization (2003) p. 7-26 Google Scholar, argues that constitutional courts tend to serve as agent of the central government, ‘policing’ the regional governments, while only rarely limiting the competences and powers of the central government.

⁴⁰ Landes, W. and Posner, R., ‘The Independent Judiciary in an Interest-Group Perspective’, 18(3) Journal of Law and Economics (1975) p. 875-901 Google Scholar.

⁴¹ Epstein, et al., supra n. 1 Google Scholar.

⁴² It is unlikely that smaller parties and other potential private interests are in a position to offer better future career and sinecure opportunities than the main political parties. As a consequence, we expect the interests of the large coalitions to dominate over any other strategic interest.

⁴³ Dalla Pellegrina, and Garoupa, , supra n. 6 Google Scholar; Garoupa, et al. 2013, supra n. 5 Google Scholar.

⁴⁴ We have looked at all actions for annulment, except the ones that have been declared inadmissible or have been revoked before a decision on the merits was made.

⁴⁵ In 40% of the 96 disputes involving the federal government the latter is the Plaintiff.

⁴⁶ <www.const-court.be>. In most of the cases, neither the Plaintiff nor the Defendant have a clear and straightforward political identification (e.g. disputes between private citizens). We chose to concentrate on disputes between the federal government and regional governments as well as on disputes where opposing parties are both regional governments so that the existence of political conflict or alignment can be easily assessed and is less subject to subjective considerations. Consequently, we have not included disputes between private citizens where one or more parties have a clear political identification (for example, the famous Brussels-Halle-Vilvoorde judgment) because that would not be a random sample of private disputes.

⁴⁷ See <www.fed-parl.be/index.html>; <www.vlaamsparlement.be/vp/index.html>; <www.parlement.wallonie.be/content/>; <www.parlbruparl.irisnet.be>.

⁴⁸ The political nature of the nominations in the Belgian Constitutional Court is as such not in dispute. In the words of former President Alex Arts: ‘All former members of Parliament enter with a political party signature on their backs. But to a certain extent that is also the case for the lawyers. Parliament has to approve the proposition of a member of the Court of Arbitration with a two-third majority and their party politics plays a role. The mandates are divided between the parties on the basis of their strength in Parliament. This is a defendable system since it allows all political tendencies to be represented in the Court’ (interview with former President Alex Arts, De Tijd, 27 September 2004, p. 5).

⁴⁹ It is not common for all judges to take part in a decision of the Court; hence the composition of the Panel of judges is different across petitions. In our sample, it ranges from seven to over ten to twelve members.

⁵⁰ In general, the following authorities and persons may bring an action for annulment: the Council of Ministers and the governments of the communities and regions, the Presidents of all legislative assemblies at the request of two-thirds of their members, and natural or legal persons, both in private law and public law, Belgian as well as foreign nationals. The latter category of persons must declare a justifiable interest.

⁵¹ For example, in case 108/2000, the Flemish Executive asked for an annulment of Arts. 190, 191 and 194 of the federal law of 25 January 1999 on social provisions. The Court annulled Arts. 191 and 194, but not Art. 190. In the Tobacco Advertisement Case (102/1999) discussed below, the Court upheld the federal ban on tobacco advertising, except for internationally organised events that took place before 31 July 2003.

⁵² We performed regressions excluding cases questioning articles already regulated by other laws, to avoid perfect prediction of the Petitioner’s failure. In fact, in our data, when the decision of the Court invokes another rule disciplining the same argument the Petitioner always loses.

⁵³ This situation regards 12 petitions, mostly involving two regions as petitioners. In the empirical analysis, we will treat these cases as separate observations (decisions) and use fixed-effects to control for possible elements common to cases that belong to the same petition. We also cluster the standard errors to account for possible unobservable factors common to the observations that belong to the same petition.

⁵⁴ Until recently the Court numbered 11 men and one woman. The newest nominee (Riet Leysen) has raised the number of women to two. Member of Parliament Sabine de Bethune, on the occasion of 100th International Women’s Day, pleaded in favour of the nomination of more women to the Constitutional Court in a lecture to the House of Representatives and the Senate in Washington Het Grondwettelijk Hof zoekt m/v met talent’ [The Constitutional Court looks for M/F with talent], De Morgen, 4 March 2013, p. 23)Google Scholar.

⁵⁵ During and after their membership in the Belgian Constitutional Court, judges have held several types of positions, e.g. member of an international court (e.g. André Alen: ad hoc judge with the European Court of Human Rights since 2010), member of international committees (e.g. Marc Bossuyt: member of the UN Committee on the Elimination of Racial Discrimination from 2000-2003 and 2014-present), member or chairman of the Management Board of public organisations (e.g. Eric Derycke: chairman of the Management Board of the VUB hospital). Some judges have also been ennobled during or after their membership of the Constitutional Court (e.g. Marc Bossuyt in 2009; Etienne Cerexhe in 2009).

⁵⁶ Judges’ political affiliations have also been verified with the support of local experts. The alleged political position has been confirmed by Belgian constitutional law scholars we have contacted. Moreover, as mentioned previously, even the (former) President of the Constitutional Court held that the political preferences of all members of the Court (both the politicians and the lawyers) is known (interview with Alex Arts, De Tijd, 27 September 2004, p. 5) (for list of judges see supplementary materials, Table A4).

⁵⁷ The combination of Petitioners and judges deciding over a case can be observed only after filing, therefore it is not possible for the Petitioner to know exactly which judges will be deciding the case. However, inferences could be drawn based both on the expected time elapsing from filing to decision, and on the rotating system defining President, Rapporteurs, and composition of the Panel. We assume that the Petitioner makes such inferences. Under this hypothesis, political changeover taking place in both the federal and regional governments between filing and decision can alter the alignment observed at these two different moments.

⁵⁸ Wooldridge, J.M., ‘Cluster-sample Methods in Applied Econometrics’, 93(2) American Economic Review (2003) p. 133-138 Google Scholar; Wooldridge, J.M., ‘Cluster-sample Methods in Applied Econometrics: An Extended Analysis’, mimeographGoogle Scholar; Stock, J.H. and Watson, M.W. Google Scholar, ‘Heteroskedasticity-robust Standard Errors for Fixed-effects Panel Data Regression’, NBER Technical Working Paper No. 323, 2006.

⁵⁹ Kezdi, G., ‘Robust Standard Error Estimation in Fixed-effects Panel Models’, 9(1) Hungarian Statistical Review, Special English volume (2004) p. 96-116 Google Scholar.

⁶⁰ Relatively high and significant pairwise correlation between some of the covariates (see supplementary materials, Tables A1 and A2) suggests considering some variables’ contribution to the explanation of the Petitioner’s success rate independently from other regressors (e.g. judges belonging to the French quota).

⁶¹ See, for example Long, J.S., Regression Models for Categorical and Limited Dependent Variables (Sage Publications 1997)Google Scholar.

⁶² The general appointment mechanism is described in Section II.

⁶³ Another way of managing this problem is to perform a partial regression where the contribution of all judges with a legal background is disregarded, as if they never took part in a Panel. Such a criterion, however, has the drawback of discarding a great amount of observation concerning precisely those disputes where either the President or one of the Rapporteurs has a legal background. Given the relatively low number of observations in the dataset, we preferred to rely on random imputation of political preferences, as in Dalla Pellegrina and Garoupa, supra n. 6.

⁶⁴ When we exclude judges with a legal background, the distribution of preferences seems to be even across the seven parties dominating the political scenario in the period under investigation (three judges are allegedly affiliated to CD&V, three to CDH, one to Ecolo, three to MR, two to Open VLD, two to PS, and three to Spa). Taking all judges together, preferences appear to be normally distributed (eight judges are allegedly affiliated to CD&V, three to CDH, one to Ecolo, six to MR, three to Open VLD, seven to PS, and five to Spa). Therefore, we also performed the same exercise drawing from a normal distribution where parties have been allocated according to their frequency, and then rounding to the next integer value. Those results, which are available upon request, do not display remarkable differences when compared to the extraction from a uniform distribution.

⁶⁵ This looks reasonable to the extent that the inclusion of the Petitioner-Defendant political alignment dummy implies that the parameters for partisan alignment between the Petitioner and the Court estimate the effects on the rate of success of the Petitioner in the event both Petitioner and Court have different preferences as compared to the Defendant. Notice, however, that due to the high correlation of the dummies identifying the alignment between the Petitioner and actors of the Court with the Petitioner-Defendant political alignment (respectively 65%, 64% and 88% with the President, the First Rapporteur and median judge in the Panel), we decided to leave the latter variable as a non-interacted covariate in all other regressions.

⁶⁶ Since the exact composition of the Panel is not known ex ante, this mainly concerns the alleged political party of the President. In fact, the rotating mechanism of appointment could make it easier to infer who will be the President, while it is, in principle, less likely that the other judges of the Panel, including the Rapporteurs, can be correctly inferred.

⁶⁷ From descriptive statistics in Table 1 it emerges that Wallonia always succeeds when it is Petitioner (regardless of the identity of its counterpart). The same holds for Brussels, although this occurs only when the case involves two regions. For this reason, the fixed-effects for these two regions cannot be estimated.

⁶⁸ See e.g. Claes, E. and Vandaele, A., ‘The Tobacco judgment of the Constitutional Court: a political verdict?’, 1 Jura Falconis (1999-2000) p. 79-94 Google Scholar.

⁶⁹ See De Standaard, 2 and 3 October (pieces by Dirk Achten).

⁷⁰ Moonen, , supra Google Scholar n. 14.

⁷¹ Adams, and Van der Schyff, , supra n. 12, p. 929 Google Scholar.

⁷² Adams, and Van der Schyff, , supra n. 12, p. 930 Google Scholar.

⁷³ Berx, C., Rechtsbescherming van de burger tegen de overhead [Legal Protection of the Citizen against the Government] (Intersentia 2000) p. 246 Google Scholar; Vermeire, L., ‘De oud-politicus als onpartijdig rechter in het Arbitragehof’, Rechtskundig Weekblad (1986-1987) p. 2441 Google Scholar.

⁷⁴ See more particularly Maes, supra n. 14. She criticises the fact that an advertisement for the position of judge in the Constitutional Court was placed in the Official Journal, whereas the most important actual requirement for nomination (having the support of one of the leading political parties) was not mentioned ‘The twelve seats in the Constitutional Court are divided among the political parties’ (Maes, supra n. 14). See also Vuye: ‘Members of political cabinets can become judges in the Council of State or in the Constitutional Court… . There is a kind of interwoveness between certain high magistrates and certain politicians. How is it possible that at many occasions one easily hears whispering – since one is not supposed to speak about this – about the political preference of a particular judge?’ (interview with Hendrik Vuye, De Morgen 19 December 2008, p. 21).

⁷⁵ See the previously mentioned interview with former President Alex Arts in which he mentions that all nominations take place on the basis of party politics.

⁷⁶ In a recent interview, former President André Alen held that ‘The judges in the Constitutional Court are surely no lawyers in an ivory tower. Half of the judges are previous members of Parliament. They therefore are aware of the reality of politics and the court takes that into account. Nor are the other judges removed from reality’ (interview with André Alen, Juristenkrant, 26 February 2014, p. 9).

⁷⁷ In the words of the former President André Alen: ‘The Court should take into account every recommendation it receives, but subsequently, it has to decide in complete independence. The Court has to guarantee the protection of fundamental rights. We should not be blind to social reality, but at the same time we are also forced to make fundamental choices and we have to decide without fear and in full independence’ (interview with André Alen, Juristenkrant, 26 February 2014, p. 9).

⁷⁸ In the words of former President Alex Arts: ‘There are differences between the language groups, diverging visions and ideological differences and opinions. But there are no real tensions, precisely because we do everything we can to remove tension…. No one has an interest in an open conflict since the next day we need each other again… Decisions are indeed sometimes the result of search for a consensus between twelve people with different ideological and language sensitivities. Between the lines of a decision one can read that search for a consensus’ (Interview with former President Alex Arts, De Tijd, 27 September 2004, p. 5).

⁷⁹ Maes, supra n. 14.

⁸⁰ Moonen, supra n. 14.

⁸¹ Leysen was (although not formally politically linked) suggested by the Flemish Party N-VA. That party criticised the fact that now a hearing in the Senate was held necessary although there was no clear legal basis for such a hearing and it did not take place with previous candidates (see Nieuwsblad, www.nieuwsblad.be/article/printarticle.aspx?articleid=DMF20131205_00873829, visited on 24 September 2014).