DIRECT HORIZONTALITY: WHAT IT SOLVES (AND WHAT IT DOESN'T)
The above section has demonstrated that while indirect horizontality is useful in some respects, it nonetheless remains tied to some of the founding assumptions of the default vertical model.
It also generates additional problems, both conceptual as well as in its application. The question might then arise: if default verticality is no longer fit for purpose, and if both the state action doctrine and indirect horizontality are unsatisfactory substitutes, then why not simply apply constitutional rights directly (direct horizontality)? Some constitutions do bite this bullet. As discussed above, Article 20 of the Kenyan Constitution, for example, stipulates that ‘the Bill of Rights applies to all law and binds all state organs and all persons’[265] (emphasis added) (what I labelled ‘unbounded direct horizontality', and will refer to simply as ‘direct horizontality' in this section).In three respects, direct horizontality straightforwardly solves the concerns discussed above. First, by displacing the state entirely from the picture, it is no longer tied to sovereignty as a necessary precondition for the application of the rights framework. Secondly, by treating the relationship between the two private parties as the relevant subject of constitutional concern, direct horizontality avoids the mischaracterisation problem involved in going through the state to get at the private relationship and the private parties. And thirdly, it allows courts to adjudicate upon the rights and obligations of the private parties directly.
The third aspect indicates, however, that direct horizontality suffers from problems that indirect horizontality is explicitly set up to avoid. Traditionally, we think that it is the regimes of private law - tort and contract, to name just two - that regulate rights and obligations between private parties.
If constitutional rights are made horizontally applicable per se, widespread overlap will be inevitable, and questions of normative priority will arise.[266] For example, if the ‘right to life and personal liberty' is directly applicable between private parties, where does that leave the regime of criminal law, which is concerned with precisely the question of preventing people from depriving each other of their life or personal liberty? As we have seen, indirect horizontality seeks to resolve the question of overlap by limiting its own intervention in two ways: first, through interpretation, by using the wide scope provided by ‘general clauses' in private law codes to enforce a rights framework between private parties; and secondly, through modification, by subjecting aspects of the private law regime to constitutional scrutiny and modifying or striking them down if required (with due deference to the first-order legislative balance that has already been struck).But neither of those two options is open to a model of direct horizontality, as it requires constitutional rights to be applicable to private parties and private relationships without mediation through private law. In other words, overlap, rather than reconciliation, is a necessary consequence of direct horizontality.[267] This, in turn, raises a separation of powers issue, in a much starker fashion than in indirect horizontality. As both scholars[268] and courts[269] have pointed out, balancing private rights and obligations appears to require the kind of granular judgement that legislatures are better positioned to make. Furthermore, as we shall see in greater detail in subsequent chapters, such a balance can be struck in a range of ways, all within the spectrum of reasonableness. A complete judicial takeover of the ‘horizontal domain', then, raises serious concerns about institutional competence (again, a concern that indirect horizontality attempts to address through norms of deference, but which is not open to direct horizontality).
It is therefore evident that limiting the scope and reach of direct horizontal- ity in some form is necessary. The law of the European Union, for example, provides for direct horizontality within certain domains (such as equal pay).[270] The South African Constitution predicates it upon the ‘nature of the right' and the ‘nature of the duty'.[271] But that, of course, raises the question: what is the principled basis for restricting horizontality to certain domains, and how is one to decide the basis on which the ‘nature' of rights and duties necessitates the application of the rights framework to private conduct?
The debate about direct horizontality, therefore, cannot avoid engaging with the delineation of its limits and boundaries. This delineation has been the focus of both courts and scholars that have had occasion to engage with direct hori- zontality. The approach of prominent courts, however, has been unsatisfactory. I will briefly flag two examples where direct horizontality has been entrenched within the constitutional landscape for a while and, therefore, where courts have had some time to work out its contours. Note that, in the discussion that follows, the limited point I will consider is the existence - or lack thereof - of a fleshed out normative framework that determines both the scope and the boundaries of direct horizontality.
The first example is the Court of Justice of the European Union (CJEU), which is admittedly something of a sui generis case, given the interplay between different levels of law-making, different types of legal instruments, and tensions between the supranational court and the municipal courts. It is not my aim here to encapsulate the long-standing debate within the EU pertaining to the relationship between national and EU law, or the role of direct horizontality in shaping that relationship. Therefore, for the limited purposes of this section, the following snapshot should suffice: the CJEU began by limiting the application of direct horizontal effect to measures that are ‘precise', ‘clear' and ‘unconditional' enough, so that no further legislative intervention - either by the states or by the EU - is required to effectuate the right in question.[272] Subsequently, in later judgments, the CJEU referred to Treaty intent (preventing ‘discrimination on the labour market'[273]) and to Treaty guarantee of the freedom of movement as ‘fundamental objectives of [the European] Community'.[274] These, it held, were justifications for expanding the reach of direct horizontal effect in cases of discrimination on grounds of nationality.
In situations where private parties were in a position to impose conditions that ‘adversely affect the exercise of fundamental freedoms guaranteed under the Treaty',[275] they would be subjected to the horizontal rights framework. Still more recently, further expansion of direct horizontal effect has been justified on the grounds of ensuring ‘full effectiveness'[276] of general principles (such as non-discrimination). Sometimes these tests are used in a complementary fashion. For example, in June 2021, the CJEU handed down a preliminary ruling in a case concerning an application from the Watford Employment Tribunal, with respect to the question of equal pay for female and male workers at Tesco Stores.[277] The CJEU held that Article 157 of the Treaty of the Functioning of the European Union had direct effect both because the obligation was clear and precise, and because full effectiveness had to be given to the commitment - set out in the Treaty of the European Union and in the Charter of Fundamental Rights - to promote equality between women and men in all spheres.[278]As is evident from the above, and as many scholars have pointed out, in the cases under discussion, the ‘foundation of the horizontal direct effect is barely explained'.[279] In effect, while the CJEU's jurisprudence informs us of when rights may be applied horizontally (including the well-known distinction between horizontally applicable treaty obligations and inapplicable directives[280]), the underlying normative basis - or, at least, a normative basis internal to the question of horizontality - remains unexplored.[281] The upshot of this is that while there exist ‘technical rules' determining the application of horizontal rights on a case-by-case basis, they lack any deeper ‘constitutional coherence'[282] (of course,
Direct Horizontality: What It Solves (and What It Doesn’t) 67 it is important to flag that the constitutional status of the Treaties is itself a contested issue).
As a second example, the Irish doctrine of constitutional tort - a unique approach to horizontality - also suffers from unclear scope and boundaries. The Irish Constitution does not specifically provide for horizontal application. However, in Byrne v Ireland9 and Educational Company of Ireland Ltd v Fitzpatrick (No 2),9 the Supreme Court held that constitutional rights cast correlative duties of non-interference on citizens. Consequently,
a right guaranteed by the Constitution or granted by the Constitution can be protected by action or enforced by action even though such action may not fit into any of the ordinary forms of action in either common law or equity and that the constitutional right carries within it its own right to a remedy or for the enforcement of it.[283] [284] [285] There is a substantial body of scholarly work around the constitutional tort doctrine.[286] Most scholars agree, however, that it is not grounded within a principled, constitutional theory. In one of the more recent studies, for example, Colm O’Cinneide points out that the doctrine of constitutional tort has been uneven in the application, with ‘a serious lack of conceptual clarity as to when and how horizontal effect should be given to fundamental rights’.[287] This is because, inter alia, the case-by-case basis on which the Supreme Court has adjudicated constitutional tort claims has meant that no overarching test has been established to determine whether a constitutional right will be given horizontal effect, and if so, to what degree and extent.[288] For instance, many of the initial constitutional tort cases involved imposing obligations on trade unions (vis-a-vis individuals who wish to opt out of collective bargaining schemes, or those who are affected by industrial action). However, they contained very little analysis of how constitutional rights such as the freedom of association are to be applied when transplanted from the public to the private context, specifically when it comes to historically specific formations such as trade unions.[289] In some of its judgments, the Irish Supreme Court has hinted that constitutional rights will apply horizontally (via constitutional tort) in cases only where existing private law is ‘plainly inadequate’ to deal with the issue.[290] This, ostensibly, solves the separation of powers issue that I have outlined above. Furthermore, at a more conceptual level, the argument that horizontal rights will apply where private law does not exist is insufficient. The absence of legislation is as much a function of legislative choice as is its existence. Consequently, it can easily be argued that the absence of a private law regime represents legislative intention not to regulate a certain domain, and judicial intervention via horizontal application is unjustified.[292] In the Irish context, this is exemplified by the Supreme Court’s ‘abortion cases’, where, in the absence of common law proscriptions, the Court injuncted private health clinics, student unions and other bodies from providing advice on obtaining abortions in the UK, on the basis that the right to life of the unborn child took priority over the freedom of expression.[293] O’Cinneide and Forde both point out that the abortion cases do not engage in any analysis of issues around the scope of application and balancing of constitutional rights and obligations between the private parties.[294] It is therefore clear that the ‘inadequacy’ formulation does not obviate the need for a theory that tells us when horizontal rights are applicable to private relationships. The CJEU and Irish examples highlight the need for a normative framework that can be applied to adjudicating questions of direct horizontality. In the concluding chapter of this book, I will also demonstrate how the case law of more recent constitutional courts - such as those of Kenya and South Africa - reveals a similar need. Contemporary scholarly models, too, recognise and respond to this need. In the next chapter, I shall examine two recent models, by Johan van der Walt and Jean Thomas. While a full consideration of these models is a subject for the next chapter, it is sufficient to note here that both van der Walt and Thomas emphasise the requirement for a principled, domain-based limitation on the operation of direct horizontality, for many of the reasons discussed above. The crucial questions that then arise are: what is the underlying principle, and what does it mean in terms of a concrete model for conceptualising and applying a framework for direct horizontal rights? V.