INDIRECT HORIZONTALITY: FORMS AND LIMITS

In the previous chapter, while considering the state action doctrine, I argued that the central purpose of the ‘functional equivalence approach' is to prevent the state from evading its obligations by removing itself from the sphere of action.

There are two archetypal solutions that claim to address this failure. The first is the ‘third-party effect', which has its origins in post-war German constitutional doctrine. The third-party effect understands a constitution as an ‘objective order of values' that pervades the private legal realm. The second solution flows from the scholarship of the US legal realists. It holds that the state itself constitutes the public/private divide, and it therefore remains present in the private legal realm. While both these approaches go through the state to get to the private relationship at issue (and are therefore variants of ‘indirect' horizontality), there are nonetheless important conceptual differences between the two.^[203] I consider each in turn.

A. The Third-Party Effect

The third-party effect (mittelbare Drittwirkung) - which is the dominant form of indirect horizontality - originated with the German Federal Constitutional Court.

It is not my task here to add to this substantial body of literature by presenting a fresh analysis of the third-party effect. My limited objective is to examine indirect horizontality as an archetypal departure from the default vertical approach, in light of the conceptual assumptions that I set out in chapter one. Consequently, I shall focus on the form of the legal reasoning employed by the Federal Constitutional Court in the Luth judgment,^[208] where the third-party effect was first articulated, and which remains the point of departure for any conceptual exploration of the doctrine.

Erich Luth was a German writer and film director. In the aftermath of World War II, Luth protested against the effective rehabilitation of Veit Harlan, also a film director as well as a former Nazi propagandist, by calling for a public boycott of Harlan’s new film. Harlan sued Luth. Relying upon German civil law, a lower court injuncted Luth from making any further calls for a boycott. The Federal Constitutional Court set aside the injunction, on the basis that the lower court had not taken into account Luth’s constitutional free speech rights vis-a-vis Harlan in its application of the relevant private law provisions. But what was the relevance of the Constitution to a dispute between two private parties, which turned upon the existing provisions of the civil code? According to the Court:

The Constitution erects an objective system of values in its section on basic rights, and thus expresses and reinforces the validity of the basic rights.

At the first instance, this analysis - where the rules of private law were to be infused by constitutional values - was to be performed by the lower courts that were seized of the private dispute. If, however, the lower courts failed to do so, or did so in a patently incorrect fashion, then the Federal Constitutional Court could review the decision and set it aside.

The holding of Luth, therefore, was that as a ‘system of values’, constitu­tional principles ‘radiate’ into ‘existing and older [private law] rules’,^[210] which ‘receive from it a definite constitutional content which thereafter determines their construction’.^[211] At the same time, this does not mean that existing private law rules are subsumed into constitutional law and thereby effaced in their own right. The Constitutional Court was careful to note that ‘it is private law which is interpreted and applied even if its interpreters must follow the public law of the constitution’.^[212] This interpretive technique relies primarily upon pre-existing ‘general clauses’ in the civil code - for example, the duty of good faith in contract law - and aims to achieve the ‘deliberate interpretation [of these provisions]... in ways that render them substantially compliant with the norms of [the Constitution]’.^[213] As an illustration: in the Burghschaft case, the German Constitutional Court invoked the good faith and good morals clauses (that were part of the civil code); these clauses were then interpreted in light of the Constitution’s ‘objective order of values’, in order to justify refusing the enforcement of an onerous personal guarantee that had been entered into between unequal parties (a bank and a 21-year-old woman, the daughter of a businessman who had defaulted on his debts).^[214]

Thus, as Matthias Kumm writes, the teaching of Luth (and the third-party effect) is that ‘constitutional principles radiate to affect the rights and duties of all actors within the jurisdiction’.

[I]n the course of private litigation... [an individual] C could insist that the court, within its jurisdiction, is required to do what is in its power to ensure that freedom of contract, as guaranteed as an instantiation of the general constitutional right to liberty, is adequately protected. In that sense the basic value commitments underlying constitutional rights ‘radiate’ throughout the legal order to also establish require­ments for the interpretation of private law by civil courts. this means that private law is to be interpreted so as to reflect an adequate balance between the respective constitutional interests at stake.^[215]

The underlying theory, as Kumm notes, is that of absolute constitutional sover­eignty, or the ‘total Constitution’.^[216] The constitution is deemed to exercise complete normative sovereignty over the entire legal domain, so that there is no transaction or relationship that remains untouched by the ‘objective order of values’.

For the sake of completeness, it is important to note that the doctrines of Luth have been advanced in two different (although unopposed) directions. Luth has been invoked in evolving a framework of positive obligations that are enforceable against the state.^[217] As indicated above, however, this book does not engage with the important and much-debated topic of positive obligations. The second direction - with which I am concerned here - is that of indirect horizon- tality and the third-party effect, summed up by Kumm in the quote above.

In order to set up the framework of the critique that will follow later in this section, the following observation is important: conceptually, the third-party effect retains the idea of the public/private divide as it is. That is, the distinc­tion between the public and the private spheres is first taken as pre-existing and given, and it is within this framework that public values ‘radiate’ into the private sphere.^[218] This is marked by how - as pointed out above - the primary vehicles of this form of indirect horizontality are meant to be the general clauses of the (private) civil codes - a fact that has been remarked upon by numerous scholars^[219] and has been influential in other jurisdictions (such as South Africa^[220]).

B. The Ubiquity of the State

As discussed above, the third-party effect leaves the public/private divide untouched. A second manifestation of the ‘failure of constitutional scrutiny’ approach, however, comes at the problem from the opposite side. According to this view, it is the state itself that constructs the public/private divide.^[222] Any a priori distinc­tion between public and private fails because every private act or private ordering necessarily exists against the backdrop of state action. As Larry Alexander puts the point, ‘private power is the product of public laws’.^[223]

This argument takes two further, overlapping, albeit not identical, forms. According to one, every private act derives its legal validity from the background legal landscape - which, in turn, is the creation of the state. This fact was judi­cially recognised by Mahomed DP in his concurring opinion in Du Plessis v De Klerk, a judgment of the South African Constitutional Court that dealt with the horizontal application of rights in the absence of specific textual authority for the same. Mahomed DP observed that:

It is, I believe, erroneous to conclude that the law operates for the first time only when [a] sanction is invoked. The truth is that it precedes it and is indeed the ulti­mate source for the legitimation of any conduct... Inherently there can be no ‘right’ governing relations between individuals inter se or between individuals and the state the protection of which is not legally enforceable and if it is legally enforceable it must be part of law.^[224]

Mahomed DP’s articulation had a long intellectual lineage. It was most force­fully articulated by the US legal realist thinkers in the mid-twentieth century.^[225] Taking up the language of sovereignty and delegation (which was discussed in the previous chapter), these thinkers argued that ‘much of the private power over others is in fact delegated and that all of it is “sanctioned” in the sense of being permitted’.^[226] For example, contract law, acting ‘through judges, sheriffs, or marshals puts the sovereign power of the state at the disposal of one party to be exercised over the other party.

A slightly different version of the argument points out that the existing legal order is as much the product of active state choice as is any ‘departure from’ or ‘intervention into’ it. As Robert Hale argued, ‘when a state passes certain social legislation, there is a necessity of making a choice between the preservation of one kind of property right and liberty and another’.^[229] At all times, the very existence of the legal order reflects the state’s choices to privilege certain rights at the cost of others. Any ‘private act’ within that framework, therefore, can only be understood as a product of that choice, and inseparable from it.^[230]

Both these arguments - ie that the state is implicated by virtue of having created the private legal landscape and that both action and inaction count equally as an ‘intervention’ into the private law status quo - come together in Louis Henkin’s reinterpretation of the famous Shelley v Kraemer judgment. In Shelley, the US Supreme Court held that a racially restrictive covenant was unen­forceable, on the ground that the judiciary was a wing of the state. Therefore, its involvement in enforcing a discriminatory contract would effectively amount to state action.^[231] Articulating a different justification of the outcome, Henkin argues that Shelley ought to be understood as a case where there was a clash between the Fourteenth Amendment right to equal treatment of the Shelleys (the Black buyers of the property) and the freedom-of-contract rights of the objecting property holders in the neighbourhood. Note that, first, the prop­erty regime was a creation of the state; and secondly, the Court was required to prioritise one right over the other, where either option would have required an intervention, whether it took the side of the Shelleys in setting aside the covenant or the side of the property owners in allowing for its enforcement. The Court confronted the choice before it, and prioritised equal treatment in view of the greater constitutional sanctity enjoyed by it at the time.^[232] Perhaps unsurprisingly, this argument involves the same kind of ‘balancing' test that is the hallmark of the third-party effect. This suggests, in turn, that when it comes to application, the two approaches will require something similar from the courts.

C. Critique

(i) Individual Responsibility and Abstract Freedom

The critique of indirect horizontality has been dominated by concerns about its ‘vagueness and unpredictability',^[233] and the vast amounts of discretion it offers up to judges.^[234] As Matthias Kumm concedes in his defence of the approach, indirect horizontality only provides us with the form of the analysis: it does not tell us how the balance is to be struck.^[235] Scholars draw a distinction between the (relatively) more clear-cut question of whether state infringement of consti­tutional rights is justifiable and the ‘open-ended and unwieldy process of “balancing” competing fundamental rights... in the absence of detailed and specific legislative guidance'.^[236] Any balancing analysis, therefore, runs the risk of being labelled as ideologically driven.

While I am not focusing upon the mechanics of indirect horizontality here, the critique reveals a deeper issue with this archetypal approach towards depart­ing from default verticality. Indirect horizontality, in practice, has often come to be identified with the enterprise of ‘balancing' clashing individual rights because it effectively ‘transplants'^[237] the rights framework from the vertical/ public domain to the horizontal/private domain, without considering the differ­ent types of relationships that exist between private parties and the state on the one hand and inter se between private parties on the other.^[238]

This transplant is based on two presumptions. The first presumption is that, as far as the rights framework is concerned, it can be applied in the same manner to a private relationship as it is to a state-private party relationship. Secondly, as Jean Thomas observes, ‘in moving from the “vertical” to the “horizontal” [indirect horizontality approaches] effectively make the legal presumption that non-state actors are all the same, so far as rights are concerned'.^[239] We may change this to ‘rights and duties': a clash of rights arises because duties are imposed upon private parties to respect the rights of other private parties. In other words, in line with the third underlying conceptual assumption of the default vertical approach that was discussed in chapter one (‘individual respon­sibility'), these two assumptions continue to operate on the premise that parties in the private domain are located in the same ‘horizontal' plane with respect to each other, where they remain formally equal.

The balancing requirement then comes to dominate indirect horizontality because one of the parties to the relationship (the state), against which rights were unilaterally enforceable, has been simply and directly substituted by a private party, with the only relevant difference being that this private party now has rights of its own that must be taken into account.^[240] It is in this sense that the passing observation made above - that indirect horizontality ‘retains the idea of the public/private divide as it is' - assumes significance. Instead of asking what it is about certain presumptively ‘private' relationships that justifies the application of a bill of rights, these approaches take private relationships as they are already characterised (ie between two formally equal, rights-bearing private parties) and then engage in a ‘balancing test' between the rights puta­tively possessed by both parties.^[241]

This leads us to a further insight: as the ‘balancing' is conducted within the overarching framework of private law, it encounters resistance in taking into account the social and economic background that structures private relation­ships (it thus reproduces the second underlying assumption discussed in chapter one - ‘abstract freedom'). This is evident when we look at the illustrations provided by the defenders of indirect horizontality. In a much-discussed article, for example, Aharon Barak notes that:

My intuition tells me that the restaurant owner has an obligation to give service, that is, to make a contract, without discriminating on the basis of gender, race or religion. By contrast, the same intuition tells me that the private party renting out a room in his apartment is entitled to choose a lodger as he sees fit. This intuition is based primarily on the proper balance between the freedom to make a contract of the restaurant owner and apartment owner and the right of the persons seeking the service (food or dwelling) not to be discriminated against. I accept that the restaurant owner and apartment owner have the constitutional freedom to decide with whom to contract. Similarly, I accept that the person wishing to dine in the restaurant or rent the room is entitled not to be discriminated against (whether by the state or by private parties) and that if he is refused on the basis of gender, religion or race, this constitutes discrimination. (emphasis added)^[242]

Barak’s intuition about the ‘balance’ between competing rights is, of course, not shared by the many jurisdictions that consider housing discrimination (in situations involving commercial renting) to be illegal. Indeed, in chapter five, I shall offer an alternative model that argues, contrary to Barak, that, at least in some cases, housing discrimination should be considered a straightforward violation of constitutional rights. What is interesting, however, is that Barak’s distinction between the two situations is based entirely on his intuition about the ‘balance’ between the rights of two abstract parties to a private relationship. In the succeeding paragraph, Barak explains that the basis for his intuition is that a contractual right is ‘stronger’ when it relates to personal privacy and ‘weaker’ when it is ‘directed against the public at large’.^[243] The argument, therefore, simply shifts from one right (freedom of contract) to another (privacy), until the intuition feels adequately grounded. What it misses - or treats as irrelevant - however, is the ‘relational context’^[244] between the parties to the relationship.

By ‘relational context’, I mean that often the parties whose rights are to be balanced will be in different positions vis-a-vis each other. For example, social practices based on the denial of letting or selling property to members of an oppressed minority, with the express intention of forcing them into segregated ghettoes, are well known in history.^[245] A case where a landlord who acts to advance such a practice with respect to a particular tenant appears, therefore, to be distinct from a case where the denial occurs in the absence of any insti­tutionalised practice (for example, I refuse to let out my flat to a rival, out of professional or personal jealousy). Barak’s analysis - focused solely on balanc­ing the competing ‘rights’ at play - is unable to distinguish between these two situations.

Something similar is also at work in Matthias Kumm's distinction between a bigoted dinner party host deciding whom to invite home for dinner on the basis of their political views (discrimination permitted) and an employer deciding whom to employ on the same basis (discrimination not permitted).^[246] Tellingly, the discussion only arises because Kumm is anxious to allay the concerns of critics that applying the rights framework to private relationships will destroy personal freedom (the second assumption, once again).^[247] Kumm argues that the issue is essentially about the ‘delimitation of respective spheres of autonomy' (emphasis added) of the parties.^[248] He then claims that the ‘competing auton­omy interests balance out very differently'^[249] in the employment context, but says nothing further.

Barak and Kumm's examples demonstrate implicitly what Peter Benson makes explicit. Citing Barak's article, Benson argues that indirect horizontality is subsumed within the ‘normative idiom'^[250] of private law, which is to ‘specify fair and reasonable terms for voluntary and involuntary interactions between individuals'^[251] located within a relationship of ‘formal or abstract... equality'.^[252] This is, of course, a direct invocation of the third assumption discussed in chap­ter one, and with which I began this argument: that parties in the private domain are located in the same ‘horizontal' plane with respect to each other, shorn of their respective structural or institutional locations, which may make such a description inapposite.

It is important to clarify that the above issues have arisen in the course of how the indirect horizontality has been applied in practice, and in the accounts of some of its prominent defenders. Notwithstanding Benson's point about the ‘normative idiom' of private law, indirect horizontality need not be conceptually committed to ignoring the material dimensions of private relations and the rela­tional context that I have highlighted above.^[253] Indeed, in the labour law context, it could be argued that cases such as Wilson v United Kingdom have, indeed, taken the relational context into account while applying indirect horizontality.^[254]

It is therefore true that in its strongest form, indirect horizontality can accommodate the concerns outlined in this section (although, as the arguments advanced above demonstrate, there is often resistance to its doing so). However, there are two reasons why, even in this strongest form, indirect horizontality will not always be enough.

The first is that indirect horizontality still depends upon the existence of private law as a trigger for its application. There are a range of situations, however, where unequal power relations between parties - and the horizontal violation of rights - will simply not be mediated by existing law. There will thus be nothing that can be suitably interpreted or modified by deploying indirect horizontality. I will consider some of those examples in greater detail in chapter five, and in the case studies in part II of this book.

The second reason is that indirect horizontality remains bound to at least one of the assumptions of default verticality (that of normatively undivided and absolute sovereignty). This leads to indirect horizontality seeing a horizon­tal violation of rights not as a problem involving power relations between two parties, but as a problem of insufficient state action. I will consider this in more detail in the following section.

(ii) Sovereignty

As I have flagged above, much like the functional equivalence approach, indi­rect horizontality operates within the same understanding of sovereignty that undergirds default verticality (although it comes at the issue from the opposite direction). Rather than attempting to discern when the state is impermissi­bly abdicating from its role as the sovereign, indirect horizontality - whether through the third-party approach or the ubiquity-of-the-state approach - holds that, conceptually, the state can never abdicate, because it is always present. This argument, however, rests upon the same idea of sovereignty as being (norma­tively) absolute and indivisible within a single territory. Only on the basis of that assumption can it be argued that no legally founded action takes place without, in some way, normatively implicating the state.^[255] This, to use a familiar term, makes constitutional protection effectively ‘gapless’.^[256]

There are two corollaries that follow. First, indirect horizontality appears to mischaracterise its very framing of the problem. Instead of focusing upon how private actors violate constitutional rights, it focuses instead on how the state is responsible for those violations, through its implied acquiescence (either through the existing legal order or otherwise) or inaction. This is what is meant by Jean Thomas’s expression that indirect horizontality needs to ‘go through’ the state in order to get at the private act(s) at issue.

The point is conceded by Kumm, who argues that the only difference between indirect horizontality and direct horizontality is a difference of form: ie whether the case is brought against the public authority (arguing that the state has failed to strike the appropriate balance between private rights, or has otherwise failed in its duties) or against the private actor.^[257] That difference in form, however, matters.^[258] To start with, it appears to exempt the private actor from norma­tive scrutiny of their conduct by asking, instead, whether the state has failed to discharge its duties (although, of course, the private party continues to face a potential curtailment of its freedoms). More importantly, it does not seem to be consistent with our intuitions about the issue. When we think about the classic controversies involving horizontality - denial of service on racial grounds and control over employees’ conduct, to take just two evergreen examples - we think about them in terms of the relationships between the private parties, and not in terms of whether the state has achieved a successful balancing of the competing interests.

The second corollary - which needs to be considered in some detail - is this: the form of the argument bleeds into the question of adjudication. By going ‘through’ the state, as pointed out above, the focus is not so much on subject­ing the impugned private act or transaction to fundamental rights scrutiny, but on whether the state (in this case, the legislator) has achieved a correct balance. What the court is doing, in other words, is not reviewing the act, but how the state has dealt with it. What follows from this - as scholars have noted - is that, at the first instance, the state is accorded a degree of judicial deference.^[259] Kumm, for example, takes the example of credit card companies charging exploitative interest rates, and argues that judicial review of the practice will be consistent with allowing the legislature a range of options for deciding where and how to draw the line in order to balance freedom of contract against personal liberty interests in not being subjected to burdensome contracts.^[260]

Now, it is important to note that this, by itself, is not a criticism. Indeed, as Kumm’s very specific example (fixing interest rates) suggests, in a range of cases, this serves as a strength of the approach.^[261] In Du Plessis, Ackermann J took the view that ‘direct application should take place at the law-making level, so that all laws which are being applied by the courts do already comply with the basic rights, obviating the need for direct horizontal application by the courts’.^[262] This is true, and the precise nature of the interaction between constitutional horizontality and legislation shall be addressed in greater detail in chapter five.^[263] Here, however, I wish to flag a different concern. As Ackermann J observed immediately before the passage cited above:

[U]ncertainty is aggravated by the fact that (in contrast to a dispute between citizen and state) in a dispute between two private individuals both sides can invoke the basic rights, calling for a difficult balancing of conflicting rights which could reasonably lead different courts to different decisions.^[264]

We therefore see a return of the ‘transplant problem' at the stage of formulat­ing remedies: the same assumption of formally equal private parties facing each other across a ‘horizontal' plane, responsible for constructing the contours of their relationship.

To reiterate, as pointed out at the end of the previous section, this is not a necessary corollary. Indeed, and furthermore, in chapter five of this book, and subsequently in part II, I will argue that in cases where the subject matter of horizontal rights is covered by an existing statutory (or another kind of legisla­tive) scheme, that scheme should be given priority. However, in litigation around that statutory scheme, including litigation that takes the form of constitutional or horizontal rights claims, judicial interpretation should be informed by the institutional approach towards horizontality, especially when the statutory scheme (and private law) is not.

IV.