I was invited by the editors of the echr Law Review to consider how the game-theoretic analysis that I have applied to courts,1 in particular constitutional courts,2 can find an application in the system of the European Convention on Human Rights (echr) and in the case law of the European Court of Human Rights (ECtHR, the Court), that is, in the case law of a quasi-constitutional human rights court that operates at the international level. I will start by explaining the key tenets of the arguments to explore, before turning to their effects in the echr regime.
There are two arguments in game theoretic literature about the judicial discretion and authority of an international court like the ECtHR. The first replicates arguments made about national courts, while the second deals with conditions prevailing in international institutions. I will present these arguments sequentially.
1 Judicial Discretion: The Domestic Argument and its Transferrability to the echr System
It has been argued that national courts’ judicial discretion increases when the number of veto players (that is, actors whose agreement is necessary3 for a change in the legislative status quo) in the corresponding legal order increases.4 I will explain the argument, and then transpose it to international courts. Let us assume that a political system has three veto players. Think, for instance, of a three-party coalition, where each one of the parties can block a change in the legislative status quo. Figure 1 presents the ideal points (preferences) of these three parties in a two-dimensional space.5 If each one of these actors prefer points closer to their own preference over points further away, they cannot change any policy that is located inside the triangle (‘123’). Indeed, for any point inside this triangle, any movement of the status quo to the North will be objected to by veto player 3, any movement to the South will be objected to by veto player 1, and any movement to the East or West will be objected to by either veto player 2 or 3. So, a legislative change from point ‘L1’ to ‘L2’ is impossible, because it will be objected to by veto players 1 and 2, who will find the final outcome further away from their preferences. Similarly, a change from ‘L2’ to ‘L1’ will find veto player 3 objecting.
This analysis can be used to explain judicial discretion, since any court decision inside the triangle cannot be overruled by the political system and its lawmakers. If the judiciary in the corresponding country prefers ‘L1’ or ‘L2’, it can interpret the law accordingly, without any fear of being overruled by the political system and its law-makers. In addition, the courts could depart from their previous case law, that is, modify their decisions (albeit this poses a delicate question in light of stare decisis) from ‘L1’ to ‘L2’, without any interference from the political system and its law-makers. However, if the judiciary prefers points ‘J’ or ‘K’, it will have to select points ‘J΄’ and ‘K΄’ in order to avoid new legislation overruling its interpretation. So, as long as judicial interpretation remains within the political core (the triangle ‘123’), no reaction of the political system is possible. Therefore, the size of the legislative core is an appropriate proxy for the discretion of the judiciary.
Let us now add an additional veto player. This could happen if the political game changes, either because the existing veto players lose power and need an additional coalition partner, or because we are considering the interpretation by a court of a constitutional rule, which, to be overruled by the political system and its law-makers, usually requires larger coalitions.6 We would then have to consider a larger core.
Figure 2 gives a visual representation of this situation. I have added one more veto player to Figure 1, and the quadrilateral (‘1234’) represents the new core. Consider that we are discussing a constitutional amendment that requires the consent of four players. As a result, changes of the constitutional status quo inside the quadrilateral ‘1234’ are impossible as there is no possible constitutional amendment that will achieve the agreement of all veto players. Yet, any constitutional interpretation inside this area becomes possible. The reader can verify from Figure 2 that, while a judicial decision ‘J’ would be overruled by the political system and its law-makers (whether the decision concerned the interpretation of ordinary or constitutional rules), a decision ‘K’ would be overruled if it concerned an ordinary rule, the amendment of which by the law-maker does not depend on a big number of veto players, but would be valid (that is, it could not be overruled) if it concerned the interpretation of a constitutional rule, as overruling such a rule would require a constitutional amendment. In other words, a judicial decision over ‘K’ would be overruled if it concerned the interpretation of an ordinary rule, but it could not be overruled if judicial interpretation concerned a constitutional rule.
So, in our hypothetical example, if a court had based its decision on the interpretation of a constitutional rule (or of any rule that requires more players to consent for its amendment), an overrule of the judicial decision by the lawmakers by means of ordinary law would have been irrelevant, as an ordinary rule cannot prevail over a constitutional rule, whereas a constitutional amendment would have been impossible. So, the larger the difference between the constitutional and the legislative core (the shaded area in Figure 2), the wider the discretion of courts to make constitutional interpretations (as opposed to the interpretation of ordinary rules). Assuming that the constitutional and supreme courts do not want to be overruled by the other standard-setting authorities (represented above by the legislative or constitutional core), they will exercise discretion proportionally to the size of the corresponding core. It follows that, when considering discretion with respect to constitutional matters, it is appropriate to use the size of the constitutional core as a measure of discretion in the interpretation of a constitutional rule. For the interpretation by any court of an ordinary rule (the change of which depends on a smaller number of players), the determinant factor will be the legislative core.
Let us now transpose this argument to an international court, such as the ECtHR, which, like national courts, interprets human rights rules but, unlike what national courts usually do, its interpretation concerns the text of an international treaty that establishes these human rights rules. International agreements do not bind states against their consent. An eligible state to join the echr can choose not to become a party to it (e.g., Belarus), and thus the echr would not bind it. However, as soon as a state has become a party to the echr, it is bound by it, and, in principle,7 unanimity, that is, the consent of all other parties to the treaty, is required for amendments to the design of the echr system to be made. This turns each individual echr state party into a veto player. For instance, Article 7 of Protocol 15 echr, provides that the Protocol cannot enter into force unless all echr parties -each one of which is thus a veto player- have accepted to be bound by the Protocol. Protocols supplementing the echr by adding protected rights do not require unanimity,8 that is, for these Protocols to enter into force it is not necessary that all echr parties express their consent to be bound. Rather, such Protocols only bind the respective parties to each Protocol.
If, to amend the echr, the consent of each one of the 46 states parties, or the consent of all states parties to a Protocol adding certain human rights rules is necessary, then the echr core is very wide. In the absence of agreement between all states parties there is no possibility of setting an echr standard that would contradict any judgment of the Court. This affords the ECtHR an extraordinary power of discretion. Let us give an example. States that are not happy with the interpretation of the ECtHR on the voting rights of prisoners cannot overrule the Court’s case law by amending the text of the echr in order to limit the voting rights of prisoners (and the ECtHR’s interpretative and standard setting discretion), unless all other echr parties agree.
However, this argument assumes that judicial decisions will be implemented, which could be dubious in domestic politics9 but certainly cannot be assumed with international organisations and their courts. In other words, this analysis presents the highest possible levels of judicial discretion because it assumes the implementation of the ECtHR’s judgments. This assumption may be achieved but certainly not sustained by any international court, in particular when its judgments do not produce direct effects within the domestic legal order of the respondent state. For a standard set by the ECtHR to apply domestically, national authorities – domestic courts, law-makers, and governments – need to allow this to happen. Yet, national authorities can choose to defy the ECtHR’s case law on the voting rights of prisoners, for instance, and apply the standards that they opt for domestically, even if these standards are not aligned with the ECtHR interpretation and thus violate the echr.
Considering this, a more realistic approach to the actual powers of and discretion enjoyed by the echr can be given by the second argument below.
2 The Implementation of Judgments and its Impact on Judicial Discretion
For the judicial decisions of international courts there is a much more complicated implementation problem that both the ECtHR and states parties are faced with: both the ECtHR and the respondent state have to consider what will happen if the Court’s judgment finds the respondent in violation of the echr rules. If a respondent refuses to comply with an ECtHR judgment, it is possible that it faces pressure, which can theoretically escalate to the sanctions of Article 8 of the Council of Europe (CoE) Statute, or that the ECtHR loses its reputation because of the lack of implementation by the insubordinate state. This is only the beginning, because the pressure can be ignored, generating a conflict between the respondent state and the broader system of the organisation, including the Committee of Ministers that oversees the implementation of the ECtHR’s judgments. Similarly, a court that loses part of its reputation as an adjudicator will have consequences down the line with other judgments that it delivers, and so forth. As a result, what happens in one particular case will have consequences for all involved actors for the indefinite future. In other words, the game is iterated, and it involves potentially all actors: the (respondent or not) states, the Court, and possibly other Council of Europe organs, and the Council of Europe itself.
Let us analyse the different steps of this game. Consider the conduct (e.g., regulation, judicial decisions, or acts/omissions of the executive) of an individual state on any particular issue that could be challenged before the ECtHR. Figure 3 presents an abstraction of the set of outcomes that the ECtHR is willing to accept. More generally, a court’s preferences may include elements of political judgment (like the decision of the US Supreme Court on Bush v Gore),10 absolute principles (like the prohibition of slavery; ‘A’ in the Figure), proportionality between such principles (e.g., balancing freedom of expression against other human rights/general interests; ‘P’ in the Figure),11 or any other rule of textual interpretation that one considers in play. The shaded area presents the intersection of the ECtHR’s judicial preferences (e.g., hiv patients can(not) be transferred to a third state that does not offer adequate health care standards) and constraints (no person can be transferred to a third state where they risk being tortured). I call it winset of the status quo [W(sq)], that is, the set of points that the conduct of a state will survive ECtHR scrutiny. In this simplified game, if the state makes a decision, ‘G’, outside the shaded area, it will be overruled by the ECtHR, that is, found to be in breach of the echr. The state will, therefore, make a decision, ‘G’, inside the shaded area if it does not want to challenge the Court. This model produces no judiciary strikes, because they are anticipated by the state, and the proposed solutions are not objectionable by the ECtHR. The only way in which there would be an ECtHR judgment holding the state liable for its conduct is if the state has a dominant strategy to provoke the ECtHR and be found in breach of the echr by it (or at least not care about it). Such a situation could happen if a national authority, such as a government in an echr state party receives more payoffs from another arena (say electoral) from the interaction with the ECtHR and possible penalties that the latter may impose. An example would be a conservative government introducing the death penalty in order to appeal to its supporters regardless of the fact that it will be overruled by the judiciary. Another example would be that of an illiberal government inventing reasons/pretexts to keep political opponents, activists, or journalists in prison in breach of Article 18 echr in conjunction with Article 5 echr.
There is, however, another possibility for disagreement. Figure 4 replicates the previous story with one difference that increases the realism of the model: what if the state does not have exact knowledge of the W(sq)? Figure 4 replicates Figure 3 but has a lighter grey shaded area, indicating the state’s uncertainty over the ECtHR’s W(sq). Uncertainty stems from the fact that the state may not be certain about the compatibility with the echr of, for instance, a particular rule, policy, or judgment by its national authorities, and are also likely to not know the preferences of the Court. As a result, these zones of uncertainty may be very wide indeed. In this scenario, the state may make a decision in the zone of uncertainty that it intends to be approved by the ECtHR, but it is instead struck down.
I can use Figures 3 and 4 to study how the strategy/choices of the ECtHR can be shaped by factors such as how likely a given judgment is to be implemented, rather than by other factors, such as the judgment fully reflecting the Court’s preferences. The ECtHR may know the levels of tolerance of an individual state and its society, and not be willing to challenge it (e.g., Ireland regarding abortions),12 or it may not know the tolerance and make a decision in the lightly shaded area of Figure 4, expecting that it will be respected and it is not (e.g., prisoners’ voting rights in cases against the UK).13 The larger the lightly shaded area (uncertainty), the more the uncertain actor (the Court or the echr states parties) will approximate the preferences of the other if they want to avoid manifest conflict. As a result, both actors would like to increase the level of uncertainty of the other side. In fact, both the Court and states would like to make the area of their anticipated decisions as precise as possible, and the dark-shaded area in Figure 4 as small as possible in order to bring the other actor as close to their own preferences as possible. Could they achieve this goal by announcements? Could, for instance, the political authorities of a state announce that they have certain positions with respect to particular issues and expect the Court to respect those positions? Or could members of the Court (e.g., its Judges writing extra-judicially) make such statements in the opposite direction? Such statements would not be credible, because, when the time of the decision comes, each one of the actors has to face the consequences of their decision, not their previous statements. Consequently, the only basis of decisions is the reputation of the other party. How is this reputation built?
In the previous rounds of the same game, there have been several possibilities: first, it could be that the state was able to anticipate the preferences of the ECtHR and develop conduct acceptable to it; second, it is possible that there is a convergence between the ECtHR’s preferences and state conduct; third, it may be that the ECtHR is afraid to contradict the state. In the first case, the ECtHR ‘prevails’, whilst in the third it is the state, and in the second there is an identity of preferences. As a result, in the case of approval by the ECtHR of the state conduct we cannot make any inference.
However, we can make inferences from the existence of manifest disagreements. These inferences are not going to be unanimous. From the disagreement between a respondent state and the Court, some echr states parties will be in agreement with the respondent state and others with the Court. For instance, some may consider one of the two actors completely unreasonable or both of them partially right. These disagreements build the reputation of each one of the actors, and this reputation is the basis for the interaction of each one of the actors with the others in the future. For example, the ECtHR’s attitude will vary according to the traditions, as well as the size and influence of a member state. With respect to traditions, there will be significantly less disagreements between the ECtHR and countries of Western Europe. This can be attributed to common values and preferences. With respect to countries where there are disagreements (e.g., the ex-communist ones, illiberal democracies), however, the Court’s judgments may be different versus strong and determined countries (such as the UK) and small and weak countries (such as Azerbaijan). Actually, the ECtHR may use the latter (in the framework, for instance, of a referral by the Committee of Ministers on the basis of Article 46(4) echr) in order to build its reputation, because it knows that it will be more difficult for a small country to defy its decisions. Another method of building reputation for the ECtHR is to be persistent in its decision-making. For instance, the Grand Chamber’s U-turn on the issue of crucifixes in classrooms in secular polities14 is a good example of what a court shall avoid doing. To build its reputation and to avoid such U-turns, the Court may be interested in how accurately it represents certain important principles/standards, how many country populations it has on its side in any particular judgment (which might explain why the Court uses European consensus analysis), how likely it is to receive support by other international institutions, and how strong, persistent, and intransigent any particular state is. Similarly, the defying country will consider, for instance, the effect that its decision has in building a reputation of reasonable or unreasonable party, or how likely it is that it will resist international pressures if such pressures build up. Building a reputation of being reasonable and steadfast on the basis of which you can have wide coalitions of supporters is the best predictor of success in the long run.
The author would like to thank the editors of the echr Law Review for many useful comments and suggestions.
G Tsebelis, Veto Players: How Political Institutions Work (Princeton University Press 2002).
See, G Tsebelis, ‘Judicial Independence, Discretion, and Preferences’ (2022, under review for publication: <http://sites.lsa.umich.edu/tsebelis/wp-content/uploads/sites/246/2022/06/Judicial-DiscretionFinal.docx>).
Decisions are, therefore, made by unanimity, since each veto player’s agreement is required.
R Cooter and T Ginsburg, ‘Comparative Judicial Discretion: An Empirical Test of Economic Models’ (1996) 16(3) International Review of Law and Economics 295; G Tsebelis ‘Decision Making in Political Systems: Veto Players in Presidentialism, Parliamentarism, Multicameralism, and Multipartyism’ (1995) 25(3) British Journal of Political Science 289; Tsebelis (n 1).
Assume that the horizontal axis represents the position of the three parties in the left-right continuum and the vertical axis their position on a particular human rights issue, such as lgbtq+ rights. The figure represents three parties, with party 2 being the most left wing, party 3 being placed at the right, and party 1 in the middle, while party 2 is the most pro-lgbtq+ rights.
It is more difficult to amend constitutional rules than ordinary legislation. Indeed, constitutional amendments require extraordinary majorities, additional bodies, etc.. For example, the US requires three quarters of the states to approve a constitutional amendment (US Constitution, Article V). There are a few exceptional countries, such as the UK, Israel, New Zealand, and India, which either do not have a constitution or where a constitutional amendment is as easy as ordinary legislation. See, G Tsebelis ‘Constitutional Rigidity Matters: A Veto Players Approach’ (2022) 52(1) British Journal of Political Science 280.
See, for instance, Article 6 and Article 8 of Protocol 14 bis to the echr.
See, for instance, Article 6 of Protocol 1 to the echr or Article 7 of Protocol 4 to the echr.
Jackson allegedly defied the Supreme Court over Worcester v. Georgia [31 US 515 (1832)], announcing: ‘John Marshall has made his decision now let him enforce it.’ See, M Warshauer, ‘Andrew Jackson and the Constitution’ (The Gilder Lehrman Institute of American History): <https://ap.gilderlehrman.org/essay/andrew-jackson-and-constitution>.
Bush v Gore, 531 US 98 (2000).
KT McGuire, G Vanberg, CE Smith, and G A Caldeira, ‘Measuring Policy Content on the U.S. Supreme Court’ (2009) 71(4) Journal of Politics 1305.
A, B and C v Ireland [gc] 25579/05 (ECtHR, 16 December 2010).
Hirst v the UK (No 2) [gc] 74025/01 (ECtHR, 6 October 2005).
Lautsi and Others v Italy [gc] 30814/06 (ECtHR, 18 March 2011).