1 Alter, K.J., ‘The Global Spread of European Style International Courts’, (2012) 35 West European Politics 135–54 (mapping the diffusion of ECJ-style international judicial bodies)CrossRefGoogle Scholar.

2 Madsen, M.R., Cebulak, P. and Wiebusch, M., ‘Backlash Against International Courts: Explaining the Forms and Patterns of Resistance to International Courts’, (2018) 14 International Journal of Law in Context 197–220CrossRefGoogle Scholar; Soley, X. and Steininger, S., ‘Parting Ways or Lashing Back? Withdrawals, Backlash and the Inter-American Court of Human Rights’, (2018) 14 International Journal of Law in Context 237–57CrossRefGoogle Scholar; Caserta, S. and Cebulak, P., ‘The Limits of International Adjudication: Authority and Resistance of Regional Economic Courts in Times of Crisis’, (2018) 14 International Journal of Law in Context 275–93CrossRefGoogle Scholar.

3 See Alter, supra note 1, at 135–54.

4 This holds true for the CJEU despite the doctrines of supremacy and direct effect. Formally, only domestic courts have the power to set aside national legislation.

5 Keohane, R.O., Moravcsik, A. and Slaughter, A.-M., ‘Legalized Dispute Resolution: Interstate and Transnational’, (2000) 54 International Organization 457–88CrossRefGoogle Scholar.

6 Carrubba, C.J., Gabel, M. and Hankla, C., ‘Judicial Behavior under Political Constraints: Evidence from the European Court of Justice’, (2008) 102 American Political Science Review 435–52CrossRefGoogle Scholar (analyzing the effect of the direction and frequency of member state observations on ECJ decision-making).

7 Huneeus, A., ‘Courts Resisting Courts: Lessons from the Inter-American Court’s Struggle to Enforce Human Rights’, (2011) 44 Cornell International Law Journal 493–531Google Scholar; Alter, K.J. and Helfer, L.R., ‘Nature or Nurture? Judicial Lawmaking in the European Court of Justice and the Andean Tribunal of Justice’, (2010) 64 International Organization 563–92CrossRefGoogle Scholar.

8 Lupu, Y. and Voeten, E., ‘Precedent in International Courts: A Network Analysis of Case Citations by the European Court of Human Rights’, (2012) 42 British Journal of Political Science 413–39CrossRefGoogle Scholar; Larsson, O., Naurin, D., Derlén, M. and Lindholm, J., ‘Speaking Law to Power: The Strategic Use of Precedent of the Court of Justice of the European Union’, (2016) 50 Comparative Political Studies 879–907CrossRefGoogle Scholar.

9 It bears emphasis that the focus of the present contribution is on ‘defiance’ rather on the broader issue of ‘noncompliance’. Although the two concepts largely overlap, ‘defiance’ presupposes that the state’s behaviour is deliberate rather than a result of managerial failure. Defiance and what we may call, for want of a better term, non-deliberate noncompliance, thus, raise distinct sets of problems and call for distinct solutions.

10 See Dothan, S., ‘Judicial Tactics in the European Court of Human Rights’, (2011) 12 Chicago Journal of International Law 115–42Google Scholar (examining how variations in the institutional costs associated with noncompliance affects decision-making on the ECtHR); Carrubba, C.J., ‘Courts and Compliance in International Regulatory Regimes’, (2005) 67 Journal of Politics 669–89CrossRefGoogle Scholar (arguing that international adjudicators only punish states when there is a chance that these will comply); Garrett, G., Kelemen, R.D. and Schulz, H., ‘The European Court of Justice, National Governments, and Legal Integration in the European Union’, (1998) 52 International Organization 149–76CrossRefGoogle Scholar (using insights from game theory to analyze interactions between CJEU and member state governments in the European Union). Relatedly, scholars have debated the extent to which legalization by adding or reducing flexibility can prevent or, on the contrary, precipitate regime breakdown, see Rosendorff, B.P., ‘Stability and Rigidity: Politics and Design of the WTO’s Dispute Settlement Procedure’, (2005) 99 The American Political Science Review 389–400CrossRefGoogle Scholar; Helfer, L.R., ‘Overlegalizing Human Rights: International Relations Theory and the Commonwealth Caribbean Backlash against Human Rights Regimes’, (2002) 102 Columbia Law Review 1832–1911CrossRefGoogle Scholar; Johns, L., ‘Depth versus Rigidity in the Design of International Trade Agreements’, (2014) 26 Journal of Theoretical Politics 468–95CrossRefGoogle Scholar.

11 In the remainder of the article, I use ‘State’ (with uppercase ‘S’) to refer to the player in the game-theoretic model, which is an abstract theoretical construct, and ‘state’ (with lowercase ‘s’) to refer to real-world state actors.

12 Conversely, when overly pessimistic, these beliefs can result in lost justice. This occurs when the IC refrains from ruling against a State that would have complied with an unfavourable ruling. Note, though, that because states will invariably welcome a favourable ruling irrespective of their degree of support for the regime, state reactions to favourable rulings are unlikely to provide international judges with new information about state preferences.

13 See Chen, D., Moskowitz, T.J. and Shue, K., ‘Decision-Making Under the Gambler’s Fallacy: Evidence from Asylum Judges, Loan Officers, and Baseball Umpires’, (2016) The Quarterly Journal of Economics 1181–1242CrossRefGoogle Scholar; See Danziger, S., Levav, J. and Avnaim-Pesso, L., ‘Extraneous Factors in Judicial Decisions’, (2011) 108 Proceedings of the National Academy of Sciences 6889–92CrossRefGoogle ScholarPubMed.

14 Ibid.

15 For a discussion of recent advances in judicial behaviour theory and the preferences of judges see Epstein, L., Landes, W.M. and Posner, R.A., The Behavior of Federal Judges: A Theoretical and Empirical Study of Rational Choice (2013)Google Scholar; Epstein, L. and Knight, J., ‘Reconsidering Judicial Preferences’, (2013) 16 Annual Review of Political Science 11–31CrossRefGoogle Scholar.

16 One way of conceptualizing compliance dynamics in integrated regimes is as exploiting differences among domestic actors to render opposition to the regime more difficult. When international adjudicators can rely on multiple domestic constituencies to implement their rulings, it becomes harder for the state to speak with one voice. Inasmuch as all compliance constituencies must agree in order to prevent compliance, the mechanism is thus akin to hedging. Hence, as compliance constituencies get more numerous and differentiated, the joint probability that they will all agree to resist the international regime decreases.

17 One may object that this dichotomous choice fails to capture the full range of options available to international adjudicators. An international court may, for example, elect to declare only certain aspects of a state’s conduct contrary to the regime so as to make its decision more acceptable to the state or dismiss the case on procedural grounds to avoid considerations of its merits. However, although it is true that variations exist between outright win and outright lose holdings, no insight is gained by adding this complication to the model.

18 In that sense, we are not assuming that international adjudicators always prefer ruling against compliant states. In many cases state and court preferences converge so that judges receive their highest payoff when they rule in favour of the state. Yet because such situations, by definition, do not raise a compliance dilemma, they are outside the scope of the present analysis.

19 To this payoff ordering it may be objected that an international court may suffer an institutional loss vis-à-vis other states when it defers to a state in a case seen as a clear violation. Whether my analysis also accommodates this situation depends on how the institutional loss from not ruling against a clear violator compares to the institutional loss from non-compliance. If the institutional cost associated with the latter is greater, the problem is analytically identical. If the two costs are the same or the non-compliance cost is smaller, breach becomes a dominant strategy for the IC (i.e., it is always its best strategy regardless of the State’s strategy). This effectively assumes away the compliance dilemma, which is the focus of this article.

20 I do not mean to suggest that this is necessarily the language that real-world actors will use to designate these two states. Non-supportive states may occasionally be characterized as ‘sovereignist’ or ‘nationalist’ and supportive ones as ‘multilateralist’.

21 The framework presented here can easily be extended to account for the situation where an international court prefers ruling against a state even if compliance is unlikely or the situation where international adjudicators have an intrinsic preference for judicial restraint. In these two hypotheses, however, information and beliefs about state preferences are far less relevant to judicial decision-making.

22 Drawing on the language of medical diagnostics, defiance can be thought of as a false positive: believing that the State is supportive when in fact it is not. Similarly, lost justice can be compared to a false negative: believing that the State will not comply when in fact it will. For reasons discussed above (see supra note 9), developing strategies to minimize false negatives is more difficult than for false positives (which are the focus of the present article).

23 The two nodes are, to employ the terminology of game theory, part of the same information set.

24 For the technically minded, the assumed solution concept here is known in game theory as ‘Perfect Bayesian Equilibrium’. ‘Perfect’ stands for subgame perfection while ‘Bayesian’ indicates that players have incomplete information about the world (such as the other player’s true type). Because choice made at a given node is assumed to reflect anticipated choices made at all subsequent nodes if these were reached (that is what subgame perfection is about), a complete statement of the equilibrium requires that we indicate not only what happens on the path of play (the predicted choices in the game tree) but also off-the-path of play (the paths not taken in the game tree). In this example it would be:

25 In that case, the expected payoff of breach, the riskier bet, is the same as the expected payoff of ∼breach. Note, though, that this presumes risk neutrality. Results differ when a different risk profile is assumed. If we assume that judges are risk averse, as human decision-makers in general tend to be, they will, in such a situation, prefer the sure thing, which is to defer to the State.

26 See Alter and Helfer, supra note 7; Alter, supra note 1.

27 See Neuman, G.L., ‘Import, Export, and Regional Consent in the Inter-American Court of Human Rights’, (2008) 19 European Journal of International Law 101–23CrossRefGoogle Scholar; Sweet, A.S and Brunell, T.L., ‘Trustee Courts and the Judicialization of International Regimes’, (2013) 1 Journal of Law and Courts 61–88CrossRefGoogle Scholar.

28 It bears underscoring that overconfidence here is overconfidence with hindsight. Based on the judges’ subjective prior beliefs, their choice is still a rational one, even if it fails to produce the expected result.

29 Campbell v. Zimbabwe, Judgment, Case No. SADCT: 2/07 (SADC, 28 November 2008). Text of the judgment available at www.worldcourts.com/sadct/eng/decisions/2007.11.28_Campbell_v_Zimbabwe.htm.

30 Ibid.

31 Ibid. See Hemel, D. and Schalkwyk, A., ‘Tyranny on Trial: Regional Courts Crack down on Mugabe’s “Land Reform”’, (2010) 35 Yale Journal of International Law 517–23, at 520Google Scholar.

32 Nathan, L., ‘The Disbanding of the SADC Tribunal: A Cautionary Tale’, (2013) 35 Human Rights Quarterly 870–92CrossRefGoogle Scholar.

33 As we shall see below, a strong record of compliance is often likely to be the principal reason judges underestimate p in the first place.

34 de Londras, F. and Dzehtsiarou, K., ‘Managing Judicial Innovation in the European Court of Human Rights’, (2015) 15 Human Rights Law Review 523–47, at 541CrossRefGoogle Scholar (contrasting the widely differing weight of the UK and Macedonia in shoring up the authority of the ECtHR); see also S. Dothan, supra note 10 (distinguishing costlier noncompliance by ‘high-reputation’ states and less costly noncompliance by ‘low-reputation’ states).

35 Economists have conceptualized the role of legal counsels in court proceedings as that of an ‘enfranchised lobby’ providing judicial policy-makers with decision-relevant information, see Dewatripont, M. and Tirole, J., ‘Advocates’, (1999) 107 Journal of Political Economy 1–39CrossRefGoogle Scholar.

36 See Steinberg, R.H., ‘Judicial Lawmaking at the WTO: Discursive, Constitutional, and Political Constraints’, (2004) 98 The American Journal of International Law 247–75, at 273CrossRefGoogle Scholar (highlighting in the WTO context rules and practices that assist the Appellate Body in generating information on state preferences).

37 Larsson, O. and Naurin, D., ‘Judicial Independence and Political Uncertainty: How the Risk of Override Affects the Court of Justice of the EU’, (2016) 70 International Organization 377–408CrossRefGoogle Scholar; see Carrubba, Gabel and Hankla, supra note 6; Granger, M-P.F., ‘When Governments Go to Luxembourg ... : The Influence of Governments on the Court of Justice’, (2004) 29 European Law Review 1–31Google Scholar.

38 See Larsson and Naurin, ibid.

39 See the analysis with respect to domestic courts in Dyevre, A., ‘Domestic Judicial Defiance and the Authority of International Legal Regimes’, (2017) 44 European Journal of Law and Economics 453–81CrossRefGoogle Scholar.

40 It bears observing that my argument here is that these mechanisms can be used to address uncertainty, not that states intentionally designed them for that purpose when they created or reformed the judicial bodies under consideration.

41 Art. 43 ECHR:

For a discussion of the procedure, see de Londras and Dzehtsiarou, supra note 34, at, 533.

42 Of the 1,791 judgments pronounced against Italy between 1959 and 2016, 1,190 involved a violation of the right to a speedy trial. See ECHR Overview 1959–2016, available at www.echr.coe.int/Documents/Overview_19592016_ENG.pdf.

43 Italy’s paradoxical situation of high political commitment to the ECHR and dismal record of violations mirrors its status as founding EU member state but most frequent EU law infringer. See Börzel, T.A., Hofmann, T., Panke, D. and Sprungk, C., ‘Obstinate and Inefficient: Why Member States Do Not Comply with European Law’, (2010) 43 Comparative Political Studies 1363–90CrossRefGoogle Scholar.

44 Lautsi v. Italy, Judgment of 3 November 2009, ECHR, Application no. 30814/06.

45 The unmatched intensity of the public backlash against the decision is reflected in Google searches for the term ‘Corte europea’ on the Italian web. From the plot below, we can see that the day the chamber ruling was announced (3 November 2009) coincided with a sharp and unprecedented surge in internet searches for the term, indicating that the ECtHR was receiving an unusually high level of attention in the Italian peninsula:

(Data source: Google Trends).

Google Searches for ‘Corte Europea’, Monthly Intensity, 2004–2013

46 ‘“Via il crocifisso dalle scuole” Vaticano: “Sentenza miope”’, Corriere della Sera, 3 November 2009, available at www.corriere.it/cronache/09_novembre_03/crocifisso-aule-scolastiche-sentenza-corte-europea-diritti-uomo_e42aa63a-c862-11de-b35b-00144f02aabc.shtml.

47 ‘Il crocifisso resterà nelle aule’, Corriere della Sera, 6 November 2009, available at www.corriere.it/politica/09_novembre_06/berlusocni_crocifisso_9674b4aa-cacc-11de-89f9-00144f02aabc.shtml.

48 Armenia, Bulgaria, Cyprus, Greece, Lithuania, Malta, Monaco, Romania, Russian Federation, and San Marino.

49 These included Albania, Austria, Croatia, Hungary, Macedonia (FYRM), Moldavia, Norway, Poland, Serbia, Slovakia, and Ukraine. See Puppinck, G., ‘The Case of Lautsi v. Italy: A Synthesis’, 2012 Brigham Young University Law Review 873–930, at 873Google Scholar.

50 Lautsi v. Italy, Grand Chamber, Judgment of 18 March 2011, ECHR, Application no.30814/06.

51 See Zucca, L., ‘Lautsi: A Commentary on a Decision by the ECtHR Grand Chamber’, (2013) 11 International Journal of Constitutional Law 218–29CrossRefGoogle Scholar.

52 Al-Khawaja and Tahery v. United Kingdom, Judgment of 20 January 2009, ECHR, Applications nos. 26766/05, 22228/06.

53 R. v. Horncastle, [2009] UKSC 14.

54 Ibid., at 4.

55 Ibid., at 14.

56 The European Convention on Human Rights, 1953 (1950). On the interpretation of this provision see O’Brian, W.E., ‘Confrontation: The Defiance of the English Courts’, (2011) 15 The International Journal of Evidence & Proof 93–116CrossRefGoogle Scholar.

57 Al-Khawaja and Tahery v. United Kingdom, Grand Chamber, Judgment of 15 December 2011, Application nos. 26766/05, 2228/06.

58 Ibid., at 126–65.

59 See Redmayne, M., ‘Hearsay and Human Rights: Al-Khawaja in the Grand Chamber’, (2012) 75 The Modern Law Review 865–78CrossRefGoogle Scholar.

60 One question raised by the analysis of defiance avoidance mechanism relates to the credibility of the state’s signal. Obviously, the resolve to defy the pronouncements of an international tribunal is credible enough when it is expressed through the final decision of a court of last resort – as with Al-Khawaja. But the answer may seem less evident when it comes to the declaration of politicians which, as such, are not legally binding. Inasmuch as supportive states too would prefer international judges to exert restraint rather than produce an adverse judgment, all states would seem to have an incentive to bluff and feign defiance in an attempt to deter judicial activism. International adjudicators may treat such a signal as inconsequential cheap talk. On the other hand, the failure to follow through on a defiance threat can entail significant reputational costs. First, a state regularly vowing to defy the rules of an international legal regime would no longer enjoy the benefits associated with a reputation for multilateralism. Second, being caught bluffing by international judges, after complying with a judgment that it had initially pledged to defy, would earn the state a reputation for bluff, which in turn would damage its soft power and diplomatic influence. See Sartori, A.E., ‘The Might of the Pen: A Reputational Theory of Communication in International Disputes’, (2002) 56 International Organization 121–49CrossRefGoogle Scholar. In democratic regimes, reneging on a public promise to defy the decision of an international judicial body may also entail electoral costs. The difficulty of evading these costs, by turning the promise into an effective pre-commitment, confers greater credibility on the signal; see Putnam, R.D., ‘Diplomacy and Domestic Politics: The Logic of Two-Level Games’, (1988) 42 International organization 427–60CrossRefGoogle Scholar. A related question is whether reputation is an attribute of the state or an attribute of its leaders; see Brewster, R., ‘Unpacking the State’s Reputation’, (2009) 50 Harvard International Law Journal 231–70Google Scholar.

61 Of course, rulings of the General Court can be appealed to the Court of Justice under Art. 256 TFEU. However, not only does the Court of Justice have little discretion over case selection but cases exclusively pertain to EU regulatory measures (such as Commission decisions over cartels and mergers) which do not directly raise compliance issues.

62 Hartley, T.C., The Foundations of European Union Law: An Introduction to the Constitutional and Administrative Law of the European Union (2014), 75Google Scholar.

63 Ibid., at 176–85.

64 Ibid., at 218–39.

65 Obermaier, A.J., ‘The National Judiciary—Sword of European Court of Justice Rulings: The Example of the Kohll/Decker Jurisprudence’, (2008) 14 European Law Journal 735–52, at 749CrossRefGoogle Scholar (arguing that the ECJ, by excluding some categories of patients from the scope of its patient mobility doctrines, significantly lowered compliance costs, thereby making defiance less likely).

66 See for rights associated with free movement of workers provisions R. v. Saunders [1979] ECHR 1129; Morson and Jhanjan v. Netherlands [1982] ECHR 3723. See also Craig, P.P. and Búrca, G.D., EU Law: Text, Cases, and Materials (2015), 762CrossRefGoogle Scholar.

67 See Arts. 20 and 21 TFEU.

68 See Craig and Búrca, supra note 66, at 865; Elsuwege, P.V., ‘Shifting the Boundaries? European Union Citizenship and the Scope of Application of EU Law’, (2011) 38 Legal Issues of Economic Integration 263–76Google Scholar.

69 See Kremzov v. Austria, [1997] ECHR-I 2629; Uecker and Jacquet v. Land Nordrhein-Westfalen, [1997] ECHR I-3171.

70 Case C-34/09, Ruiz Zambrano v. ONEM, [2011] ECR I-1177.

71 Kochenov, D., ‘The Right to Have What Rights? EU Citizenship in Need of Clarification’, (2013) 19 European Law Journal 502–16CrossRefGoogle Scholar; Schmidt, S.K., ‘Judicial Europeanisation: The Case of Zambrano in Ireland’, (2014) 37 West European Politics 769, at 774CrossRefGoogle Scholar.

72 Supra note 70 (Opinion of Advocate General Sharpston, 30 September 2010).

73 See Craig and Búrca, supra note 66, at 868.

74 See Schmidt, supra note 71.

75 Ibid., at 774–75.

76 Case C-434/09, Shirley McCarthy v. Secretary of State for the Home Department [2011] ECR I-3375.

77 Case C-256/11, Murat Dereci and others v. Bundesministerium für Inneres [2011] ECR I-11315.

78 Case C-87/12, Kreshnik Ymeraga and others v. Ministre du Travail, de L’emploi et de L’immigration EU:C:2013:291.

79 Joined Cases C-356 and 357/11, O and S v. Maahanmuuttovirasto (C-356/11) and Maahanmuuttovirasto v. L (C-357/11) EU:C:2012:136.

80 Ibid., at 56.

81 See Schmidt, supra note 71, at 775.

82 Dawson, M., ‘How Does the European Court of Justice Reason? A Review Essay on the Legal Reasoning of the European Court of Justice’, (2014) 20 European Law Journal 423–35, at 428CrossRefGoogle Scholar.

83 Lenaerts, K., ‘EU Citizenship and the European Court of Justice’s “Stone-by-Stone” Approach’, (2015) 1 International Comparative Jurisprudence 1–10CrossRefGoogle Scholar.

84 See Sadl, U., ‘Case-Case-Law-Law: Ruiz Zambrano as an Illustration of How the Court of Justice of the European Union Constructs Its Legal Arguments’, (2013) 9 European Constitutional Law Review 205–29CrossRefGoogle Scholar; Kochenov, supra note 71.

85 See WTO statistics, available at www.wto.org/english/tratop_e/dispu_e/stats_e.htm.

86 For statistics on win rates for complainants and respondents see Maton, J. and Maton, C., ‘Independence Under Fire: Extra-Legal Pressures and Coalition Building in WTO Dispute Settlement’, (2007) 10 Journal of International Economic Law 317–34, at 328CrossRefGoogle Scholar.

87 See Art. 67, IACHR.

88 Art. 68(4) of the Court’s Rules of Procedure stipulates that a request for interpretation does not suspend the effect of the judgment. See www.corteidh.or.cr/sitios/reglamento/nov_2009_ing.pdf.

89 See Neuman, supra note 27, at 108

90 Ibid., (arguing that the IACtHR has not paid sufficient attention to regional consent, thereby threatening the acceptance and effectiveness of the inter-American human rights system).

91 Ibid. The fact that the IACtHR does not consistently behave in a compliance-maximizing manner may reflect the perception that the defiant states are low-reputation states lacking the power to inflict substantive institutional damage on the Court. See supra note 10. This state of affairs can also be characterized as ‘contentious equilibrium’, see Dyevre, supra note 39 (characterizing situations where domestic and international judges ignore each others’ ruling as ‘contentious equlibrium’).

92 Alter and Helfer, supra note 7, at 572–3.

93 Similar to the European Court of Justice, the Tribunal of the Andean Community can hear preliminary references from national courts, see Alter and Helfer, ibid., at 568.

94 ATJ Preliminary Ruling 29-IP-98.

95 Alter and Helfer, supra note 7, at 577.

96 The conjecture that activism would, overall, be more frequent derives from the assumption that international adjudicators tend to be risk averse. Judges who do not like gambling with compliance will prefer to defer to state preferences when the risk of defiance is nontrivial. In a world of complete and perfect information, by contrast, there is no need for gambling.

97 Supra note 2.