Kelsenian Courts in an Ethnic Key: A Nightmare or a Noble Dream?

1 Introduction

A safe sign that an institution – in our case, a constitutional court performing constitutional review – has had a global triumph is when it is perceived as ‘the Holy Grail of constitutionalism’.^[1] Nevertheless, despite the fact that constitutional review today is not seen as a ‘deviant institution’,^[2] it still raises questions as to its proper place in deviant contexts, notably those where ‘[e]thnocultural diversity translates into political fragmentation’.^[3] Considering the vital role constitutional courts potentially play in such contexts – notably in arbitrating inter-community conflicts, but also in enabling the hydraulics of informal constitutional change against the background of legislative status quo due to chronic gridlocks^[4] – such courts are not exempt, but may in fact be a focus of constitutional engineering.^[5] That should not come as a surprise considering that constitutionalism in divided societies is often strained, and that ‘without appropriate tailoring’ it may be ‘worse than useless’.^[6]

That the particular design of the court, as well as its very role, might have to be moulded to the peculiarities of its constitutional, political and societal context has been recognized since the inception of the institution.^[7] Indeed, in the same sense in which Dieter Grimm has suggested that the decision on the recognition of constitutional adjudication ‘is not one of principle but of pragmatics’, it could be held that, by the extension, the same applies to the shape of the court when it exists. For Grimm, the choice on the matter is ultimately one ‘between different types of democracy’.^[8] Taking that idea as a starting point, I will consider a model of the court that I call an ‘ethnic’ court, one robustly reflecting a particular type of salient cleavage, namely ethnicity, characterizing many consociational democracies. Such a model is instantiated at different levels of intensity in three constitutional courts existing in Bosnia and Herzegovina (BiH), at the level of the state and its two federal units, and they will be the focus of our inquiry.

What I wish to examine here are only the ways in which the court model under examination swerves from the established constitutionalist grammar, and, furthermore, to explore the possible limits, if any, to such constitutional engineering. I will examine these issues starting with a brief overview of what might be considered a standard presupposition of ‘courtness’ in constitutional courts. I will then analyze the model of an ethnic constitutional court and its theoretical underpinnings, which challenge the standard presupposition. I will conclude with the critical analysis of that model.

2 The ‘Courtness’ of Constitutional Courts

One of the widely accepted assumptions concerning constitutional courts, or apex courts wielding the power of constitutional review, is that they indeed are courts.^[9] It is also assumed that this is not (just) a result of strategic labelling, a cheap way of assuring legitimacy for its rulings, but that it signals something about its nature. This designation is complicated, though not denied, by the fact that they are also strong political institutions, or at least that their adjudicatory function may have a strong political character.^[10] Nevertheless, even though such courts may be said to ‘occupy their own “constitutional space”’,^[11] for the reasons just stated they are still (as courts), in Wechsler’s words, ‘bound to function otherwise than as a naked power organ’.^[12] If they are part of the broader genus of ‘courts’, what is it that they have in common?

It seems that the basic conceptual analysis of ‘courts’ points to an institution composed of permanent officials ‘who settle disputes by applying pre-existing standards in doing so’.^[13] This condition is qualified in context of constitutional courts, due to the very nature of the constitutional text they are dealing with, but it is still assumed that judges on such courts heed the ‘ideal of adjudication according to law’,^[14] and do not merely engage in ideological and completely discretionary law-making. Among further distinguishing features of ‘courtness’ (to use Martin Shapiro’s term)^[15] are usually listed the facts that they are supposed to be independent and impartial from other branches of government, that they normally perform their functions ‘in the context of procedures initiated by others’,^[16] and that their decisions are binding even if mistaken.^[17] In particular, clear absence of independence and impartiality in an institution designated as a ‘court’, or designated otherwise but expected to act as one, would undermine the first necessary element, and the idea that the officials are not resolving disputes under complete discretion.^[18] These features are supposed to distinguish courts not only from political institutions, such as legislatures, but also from other conflict-resolving institutions, such as mediations or arbitrations.^[19] Even though Kelsen considered conceptual analysis of adjudication in this context to be unnecessary, he still insisted that constitutional courts must resemble ordinary courts from an ‘organizational point of view’, notably for ensuring their independence and impartiality.^[20] These latter features are generally accepted as indispensable for the very justification of constitutional courts,^[21] and even the realist would accept that it presents the way the judges themselves comprehend their roles (even if falsely), or at least that is the way they present themselves to others.^[22] What differentiates a constitutional court, particularly in Kelsenian typology, from other courts is its particular function, namely that it is an ‘independent organ of the state whose central purpose is to defend the normative superiority of the constitutional law within the juridical order’.^[23]

Although there are obvious ways in which constitutional courts depart from this impressionistic list of features of courts in general, most debates on constitutional courts presuppose something like this understanding, at least as a baseline. In the following we will consider a degree to which the court model under consideration swerves from these presuppositions, and the effects that might have on its plausibility.

3 A Noble Dream: Power-Sharing Courts

The proponents of the ethnic court model take a serious look at the divided societies as they actually are, and attempt to shape the court to those circumstances. One characteristic of such societies is an apparent absence of a unitary constituent power or rather the presence of multiple demoi in a single polity.^[24] One way to understand the case behind this model is to notice that it starts from an idea that in divided societies traditional constitutional courts, or other apex courts exercising constitutional review, do not offer an obvious answer to what could be called ‘the first legitimation question’, namely in whose name they are doing all that they are tasked with.^[25] Ordinary run-of-the-mill debates about the legitimacy of constitutional courts, such as those concerning the counter-majoritarian difficulty, presuppose that at least the first legitimation question has already been adequately answered. Thus, it might be argued that in the absence of substantial societal homogeneity, further representational requirements must be imposed upon such courts.

Indeed, one might wonder if the standard Kelsenian model is itself based on a presupposition of a non-divided society, thus necessitating conceptual engineering where such is not the case. If one takes the view that Kelsen’s pure theory of law, including his view of (constitutional) courts, formed part of his larger project in political philosophy,^[26] one would have a reason to think that he would see a constitutional court in deeply divided places more as a poisoned chalice, than a Holy Grail. The reason is that the primary role of such a court – at least under some interpretations of Kelsen – would be to ‘protect the integrity of the democratic political process’,^[27] but taking place within a well-functioning democracy where every minority ‘is a majority in waiting’.^[28] Vinx, a leading interpreter of this view of Kelsen, identifies places of deep fragmentation, notably those in Northern Ireland and Bosnia and Herzegovina, as not satisfying such a condition.^[29] In those contexts, he claims, minorities would not have a reason to accept the outcomes of majority decision-making, presumably even under the guardianship of a constitutional court, but would focus their efforts on ‘acquire[ing] veto rights or protective privileges disabling the powers of a potentially hostile democratic majority’.^[30]

That practice may not follow theory is seen in Issacharoff’s claim that new democracies with ‘fractured societies’ have rejected consociational power-sharing settlements for an ‘assertive form of constitutional democracy’ with constitutional or other apex courts exercising constitutional review.^[31] As an empirical claim, this view itself may be questioned considering that half a century after its theoretical elaboration, consociationalism remains ‘a theory that matters’,^[32] promising a form of government suitable for segmented or multinational societies, often politically deeply divided. It is thus still on the top menu of constitutional solutions for such societies, for better or worse. However, precisely because consociational solutions are often seen as another ‘Holy Grail’ in constitutional engineering, at least for some sort of societies, one may expect their coincidence with constitutional courts in the same constitutional package. But would such an arrangement be coherent? Beyond Kelsenian worries surveyed previously, a reason to believe otherwise would hinge on the fact that constitutional courts – as traditionally conceived – importantly diverge and even seem to defeat the point of consensual decision-making settlement promised by consociational solutions.^[33]

One way of cutting this Gordian knot is by offering a hybrid solution, a sort of consociationalised constitutional court. Graziadei has offered the most elaborate explication of the model by identifying a number of such courts in comparative constitutional law, but also normatively defending it. His contention is that the legitimacy of apex courts in consociational contexts can only be secured if they themselves become ‘power-sharing courts’, strongly reflecting the salient cleavages dividing the society, and incorporating power-sharing elements in its structure and decision-making rules.^[34] In societies where such cleavages were ethnic ones, and they often are, the constitutional court would be an ethnic court. In terms of the first legitimation question: each ethnic judge would speak in the name of her own ethnic community. For Graziadei an ideal type of such an institution would be one where the court is composed of the members of dominant (ethnic) groups in society on equal basis, and where ‘no decision is possible against the judges who are expression of one constitutive nation acting collectively’.^[35] The power-sharing elements are thus included in the composition, decision-making procedures, and the allocation of positions of power within these institutions. It is made clear that these arrangements ‘go far beyond international standards of representativeness’,^[36] or what might go under the term of ‘reflective judiciary’. The apex courts in Belgium and BiH, as well as the Cypriot one pursuant to the Constitution of 1960, are put forward as examples of such a model.^[37] That this model swerves away from ‘courtness’, despite the retention of the court-talk and designation, is made obvious with Graziadei’s frequent references to arbitration, as a distinct form of dispute resolution. The reference is appropriately made not to the International Court of Justice, but to the Permanent Court of International Arbitration. The connection lays both in what is perceived as the most important function of these institutions in the contexts under consideration, namely the ‘arbitration of salient inter-community conflicts’, but also for justifying their paritarian composition, as each party of the ethnic cleavage would want to appoint its own ethnic arbitrators.^[38] Otherwise, we are back to Shapiro’s ‘brute fact of being outnumbered’,^[39] which no party to arbitration would have a reason to accept.

To make further sense of a power-sharing court, and the reasons supporting it, it might be helpful to contrast it with other power-sharing institutions in general. Take for instance the Bosnian antitrust agency. The relevant legislation provides that the body has six members, at least three of which must come from the three dominant ethnic groups. The decisions of the institution are adopted by a majority, which also must include at least one vote from each of those three ethnic groups.^[40] What is the point of these provisions for an administrative agency famous for its highly technical decisions filled with dry economic analysis? Perhaps its decisions could not be legitimate and thus acceptable without such power-sharing rules in a divided society? But it is not clear why that would be true for this subject matter not touching on any politically salient issues, but on questions of consumer welfare and protection of market competition. Provisions were certainly not motivated by ‘mere’ interest in diversity, seen in the fact the rules promote the interests only of those groups that already hold dominant position in the society, and not of its marginalized and excluded members. The major difficulty appears to be that the institution is governed by the principles of the rule of law, implicit in the applicable administrative procedure, so that its (ethnic) members are primarily expected to be legal and economic experts independent of any extraneous influences in the impartial application of the law. Now, whatever else the latter requirements could mean it is generally agreed that their purpose is to insulate decision-makers from ruling based on ‘irrelevant or inappropriate considerations’.^[41] Ethnicity appears to be such an irrelevant consideration in this particular context,^[42] and insofar as the institutionalized rules may suggest that it could play a decisive role in the adoption of decisions, it makes them seriously suspect.

However, even if the Bosnian antitrust agency might be a mere consociational reflex,^[43] we should still consider if constitutional courts are in an importantly different position, so that the above considerations would not count against its ethnicization. An affirmative answer would aim to show strong disanalogies with the antitrust agency. Firstly, it would point out that the decisions of the court are often highly politically salient. Indeed, the matters to be decided may fall under what Chaudry calls ‘constitutive constitutional politics’, namely issues concerning ‘existential questions that go to the very identity of the political community as a multinational political entity.’^[44] Secondly, unlike the administrative agency which applies the law, the constitutional courts are interpreting a document notorious for its open-textured language, and where the law frequently runes out and the court is forced to act – in realist terms – as ‘a super-legislature’.^[45] This is quintessentially a political function and in a divided society such judicial law-making may only be legitimate under special conditions since in the area of politically highly salient issues, where the court is forced to legislate, any result reached could only be perceived as ‘the brute fact of being outnumbered’. In this context, ethnicity, it is claimed, would be neither an irrelevant nor inappropriate consideration in decision-making but rather a necessary ingredient for the wider acceptability of any decision reached. These arrangements, it is argued, would in fact guarantee the impartiality of the body,^[46] and would thus present the only model capable of having requisite legitimacy. As noted previously, Kelsen was certainly adamant that such guarantees were indispensable if the constitutional court were to successfully adjudicate disputes of federal competence, which by their nature are ‘political questions of life and death’.^[47] These questions, of course, only become more dramatic for consociational federations.

The noble dream driving the model of an ethnic court is to maintain an institution of an effective Kelsenian constitutional court, in an otherwise deeply divided society characterized by deep distrust. The unique features of such an institution would grant it robust legitimacy that would enable it to perform its basic function of arbitration of salient inter-community conflicts. Losers would now have a reason to comply, thus solving one of the noted challenges often affecting constitutional courts in consociational contexts: non-enforcement.^[48] Precisely because such a model would guarantee robust legitimacy, the court could wield the powers of strong review, even in a context of a deeply divided society.

4 A Nightmare: Ethnic Arbitrations

There seem to be two basic challenges with the proposed ethnic court model. First, the realist presuppositions, that are motivating the model of an ethnic court, are implausible in their most extensive form. The constitutional courts deal with much more law than it is supposed to, and if that is true, the undermining of the ‘courtness’ of the institution would not only make it less probable that any decisions reached would be legally ‘correct’ but also would undermine the legitimacy of the institution in relation to its adjudicatory functions, which might be too high of a price to pay. The second problem, connected to the first, is that despite its notation as a court, the resulting body would significantly move in the continuum of the dispute-resolving institutions. This would largely be the consequence of the idea that the court is a representative institution. Insofar as ‘courtness’ of an institution has a value in itself, and it will be argued that it does, considering the spectrum of functions that it performs, this will be a significant challenge.

Firstly, there is reason to believe that the actual domain within which the model proposed would find its proper place could be much narrower than initially perceived. Depending on the scope of competences of the court, the domain of politically salient cases may be diminished. As Hirschl has pointed out in differentiating between judicialization of mega-politics touching on ‘polity’s very identity’,^[64] such as issues on vital national interests in Bosnian context, and its other manifestations, political salience is a scalar property. Taking the BCC as an example, its main function does not, in fact, concern matters of mega-politics on which the ethnic judges would ‘arbitrate’. The vast majority of its caseload, more than 99 % of it,^[65] concerns decisions on constitutional complaints by individuals alleging violations of human rights where, in the words of the former judge at the court, ‘the considerations of ethnic representation are as good as meaningless’.^[66] Indeed, in this, the BCC seems to be following a general predicament of most constitutional courts, whose work can usually be described as ‘routine, mundane, and almost random in its incidence’.^[67] Furthermore, there is reason to believe that the area within which the court would be engaging in law-making, would equally be narrow. While Graziadei is correct in rejecting the Dworkinian type of ‘right answer’ legal formalism, he also seems to embrace a suggestion that judges at the constitutional level engage in a ‘game of scorer’s discretion’.^[68] He thus starts with the premise that law in that context ‘rarely predicts judicial outcomes’, and ends with the Posnerian conclusion of the ‘fiction of impartial legal interpretation’.^[69] However, as the BCC’s output on constitutional complaints shows, even in that area where provisions are notoriously open-textured, judges are not free to decide as they wish, motivated by the protection of ethnic interests, since even in such cases their ‘discretion’ or choice is ‘neither arbitrary nor mechanical’.^[70] If this is true, the intentional ethnurgy across the board threatens to undermine the Court’s authority insofar as the ethnicity is obviously an irrelevant consideration for the vast majority of the Court’s output and would only strengthen the risks related to epistemic hazards inherent to ethnocentrism.^[71]

Secondly, an important assumption behind this model, as reflected also in the ‘first legitimation question’ discussed above, is that the constitutional court is a representative institution, and that it suffers a legitimacy deficit insofar as it does not live up to this promise. But this is a controversial assumption.^[72] Constitutional courts are best seen, for better or worse, as counter-majoritarian institutions, ‘functionally oriented towards law’.^[73] It seems that it is this error that conflates the issues of representation and more defensible idea of the need for diversity in judiciary. Graziadei is clear that the proposed model, particularly in its strong form that is considered here, goes far beyond the normal expectations behind the idea of a ‘reflective judiciary’, normally understood as requiring ‘that diverse identity groups within the citizenry’ be included in the court.^[74] It is important to see what type of representation is at play. The representation usually presupposed in the discussions of the design of courts for the purpose of accommodation of difference is the descriptive representation, one where the ‘representative stands for or resembles someone or something else’.^[75] This is often made explicit, such as when Chaudry and Stacey claim that the aim of the institutionalization of ethnic difference in a court is ‘not to transform judges into agents to serve the narrow political agendas of the ethnic minority politicians who appointed them’, but rather to have the diversity of the bench affect the ‘interpretative agenda that is congenial to the protection of minority interests’.^[76] The structural features of the ethnic model point rather to a different understanding of representation, namely one where the ‘representative acts for someone or something else,^[77] or in words of Anne Philips, where the members are supposed to ‘refer back to and speak for their group’.^[78]

It should be noted that the accommodation of difference in constitutional courts, even in a more robust sense, should not (necessarily) have deleterious effects. This is evidenced by examples from Belgium and Canada.^[79] The difference may lie in the fact that the latter countries, particularly Belgium, are demo (i) cracies and not ethnocracies, and thus put emphasis on language-based differences and not on ethnic ones, which is of moral significance.^[80] As Susanne Baer has emphasized, ‘diversity matters’ but the ‘point is not biology, but perspective’.^[81]

All the previous considerations seem to directly undermine Graziadei’s ‘strongest case’ for power-sharing courts, namely the idea that ‘there must be no suspicion that the national background of a judge influences his decisions’,^[82] since these ‘arbitration-court’ hybrids would, like real arbitrations, necessarily be ‘circumscribed to administering “private justice.”’^[83] As noted by Fleiner in his study of Yugoslav institutions, such rules of ethnic power-sharing force the members of collegial institutions to act not as statesmen, but ‘much more like lobby representatives’.^[84] Of course, the situation is more challenging for courts, as judges cannot legitimately engage in special pleading, or see themselves as representing sectarian interests,^[85] which they would necessarily have to do in such an institutionalized ethnic arbitration.^[86] If this is the case then there can be a reason to believe that the legitimacy of the institution would suffer if it is perceived to lack the requisite ‘courtness’ usually associated with a constitutional court. Furthermore, to the degree it is true that the mere existence of the constitutional court may result in ‘negative perceptions of the independence and fairness of the court system’,^[87] a model where there is not even pretention of impartiality could have deleterious effects on the rest of judiciary.

However, even if we exclude all the above considerations, could one not argue that there may still be a very narrow field of politically highly salient cases where ethnic considerations should take centre stage through explicit institutionalization? This would cover cases where the law is highly indeterminate, or not even amenable to adjudication, where judges may be forced to engage in completely unconstrained law-making or political arbitration. It could be argued that in such a context the previous criticisms would not even matter, because once we accept that not everything the courts do is adjudication, the standard rule of law principles, including the guarantees of independence and impartiality, would not apply in full stringency or at all to such non-judicial activities.^[88] In such an area any result reached could only be perceived as ‘the brute fact of being outnumbered’.^[89] One might consider issues raising vital national interests in Bosnian context, particularly in the competence of constitutional courts on the adjudication of the ethnic veto, as falling among such cases. If there has ever been a concept that on its face does not seem to be amenable to judicially discoverable and manageable standards it is a notion of a ‘vital interest’ of a particular segment of a society. It seems to be plagued by that very same vagueness of phraseology that Kelsen deplored in the constitutional context.^[90]

If one takes seriously this suggestion, but also the challenges identified above with respect to ethnic courts, one way out of this problem could be the creation of a separate ethnic arbitration body charged with dealing only with the issue of vital interests, thus freeing the court from this task.^[91] It is, however, hard to see how that would make sense since the arbitration body would be charged with breaking the deadlock in the legislature composed of ethnic delegates, while itself being a power-sharing institution of ethnic delegates endowed with ethnic vetoes. An alternative proposal might be the creation of a special body within a constitutional court which would be tasked with dealing with politically salient issues under its ‘ethnic hat’ only on certain matters, but would otherwise act with the same membership as a ‘normal’ court, not smeared by its association with the other body when dealing with other issues.^[92] Something like this already exists at the level of federal units in BiH, where the constitutional courts act as pure ethnic courts when adjudicating on the ethnic veto in the ‘Council for the Protection of Vital Interests’, with vestiges of ethnurgy reflected to a lesser degree in their other functions. Such cleavage in its organization, however, proved to be ill-motivated. The ‘Councils’ are constitutional courts in all but name,^[93] and the analysis of their output shows that their work on vetoes does not consist in ethnic arbitration on indeterminate matters, but in formalist application of pre-existing law and developed tests.^[94] Contrary to Kelsen’s fears, and similar to adjudication of human rights, vital national interests appear to be fully amenable to judicial management. Ultimately, those seeing the institution of an ethnic court in more nightmarish terms might wonder if it presents a needless reduplication of already existing power-sharing institutions. Insofar as the court was supposed to be some sort of solution to chronic gridlocks expected in the system saturated with veto points, making the court just one more of such institutions may appear redundant and counterproductive. To paraphrase Scott Shapiro from another context, ‘it reintroduces the disease it was supposed to cure’.^[95]

As a side note not to be pursued further at this point, it is not clear that an ethnic court would even be legal, strictly speaking. It is useful to approach the challenge by first considering it from the aspect of political morality. It takes the form of a problem that has recently been termed ‘exclusion-amid-inclusion dilemma’ to describe a long-known political phenomenon afflicting consociations, namely that ‘[t]he decision to expressly include some groups will result in the exclusion of others’, and which ‘is also likely to limit the ability of new identity groups to emerge and find representation in power-sharing structures in the future’.^[96] Appointment of judges on the basis of ascriptive criteria, such as ethnicity, would seem to enforce the problematic corporate form of consociations, which tend to freeze existing identitarian cleavages (and accompanying existing exclusions), and thus run contrary to recommendations for the adoption of more flexible solutions.^[97] But this type of exclusion is also legally problematic insofar as it would run counter to the relevant anti-discrimination norms in international human rights law. Graziadei gives a short shrift to this criticism in respect of the BCC noting that the relevant provisions of the Constitution are framed in liberal and not corporate terms.^[98] Nevertheless, if one takes seriously the claim that there is a constitutional custom – understood as an established practice of ethnic exclusion, accompanied by (social or legal) sense that such exclusion is obligatory – as it is accepted both by domestic constitutional scholarship and apparently by Graziadei himself, the problem appears more challenging.

5 Conclusions

Ethnic ‘courts’ are structurally incapable of living up to the promise of the dream they are supposed to realize: the establishment of robust legitimacy leading to acceptance of its (far-reaching) role in a deeply divided society. It is not likely that they can achieve that, while remaining courts. And if they were anything else but courts there is no a reason to believe they would fare any differently from other political institutions in a divided society, to which they were supposed to be a solution. If this is true, the emphasis on ‘technocratic legitimacy’ might have greater prospects of success, at least in bolstering the basic ‘courtness’ of the institution.^[99] In that light the actions undertaken by the BCC to de-ethnicize its procedures should be celebrated, not lamented.^[100]

Discussions on constitutional engineering considered here are largely based on the assumption that in an otherwise dysfunctional system the constitutional court is expected to take on the weight of many tasks that political institutions do not seem to be capable of performing. This presupposition – and its accompanying lack of imagination resulting from what Moyn has called the ‘juristocracy of fear’^[101] – provides no space for rethinking of the court’s role, while maintaining the ‘courtness’ of such an institution.