Many objections to the proportionality test claim that it was not clear what is weighed.1 Tsakyrakis, for example, criticizes the view that both rights and the public interest are considered to be interests that can be weighed against each other on the same level.2 In his opinion, rights should rather be treated as ‘trumps, or as side constraints’.3 Nevertheless, he doesn’t give any further advice how to understand this idea. A more precise understanding, however, is necessary in order to assess the merits of the argument. We will develop this understanding here by taking into account the general debate on the understanding of rights and on the relationship between rights and rights’ limitations. This is the debate on ‘rights as interests’ and ‘rights as trumps’.
Most often scholars distinguish between an ‘interest model’ and a ‘trump model’.4 The terminology differs. Harbo, for example, distinguishes between a ‘Weak Rights Regime’ and a ‘Strong Rights Regime’.5 Kumm calls the interest model a ‘weak conception of rights’.6 Apart from these differences, there is agreement in substance. The ‘interest model’ notions ‘rights’ as important interests7 which can be outweighed by competing interests, such as the rights of others or public interests.8
Aleinikoff has put this point as follows:
The metaphor of balancing refers to theories of constitutional interpretation that are based on the identification, valuation, and comparison of competing interests … We usually think of balancing cases as pitting the interests of the individual against the interests of the government.9
The fact that a person has a right thus does not imply that he holds a position that gives him any kind of priority over competing considerations.10 The trump model, in contrast, notions rights as ‘trumps’,11 ‘side constraints’12 or as a ‘firewall’.13 This model is based on a basic liberal intuition that rights enjoy some kind of special priority, which gives them lexical priority over other considerations,14 in particular over any public interest.
This rather rough distinction between an ‘interest model’ and a ‘trump model’, however, cannot fully cover the implications of the understanding of rights as interests or rights as trumps. It is rather heuristic than legal. Therefore, the ideas of rights as interests and rights as trumps have to be analysed and restructured in a more detailed approach. We will distinguish between the ‘interest model’ (section I) and three ‘trump models’ (sections II-IV) here.
The interest model presumes that rights are relative rights, which means rights with limitations. Whether a right’s restriction is justified depends, inter alia, on the proportionality test. In the interest model, the proportionality test is applied as follows: the legitimate aim is any lawful aim. Therefore, any interest can legitimately be pursued by a state’s measure which infringes a fundamental right. In particular, the state can pursue a wide range of public interests. The broad understanding of the legitimate aim in turn means that a wide range of aims, in particular a wide range of public interests, are allowed to play out as competing interests on the balancing stage. Rights and all the public interests would compete on the same level.15 The right can thus be, at least theoretically, outweighed by any lawful, even if minor, interest.
Opponents of the interest model argue that this model deprives rights of their normative power: since the legitimate aim test contained no requirements, Tsakyrakis argues, ‘everything… is … up for grabs’.16 Even minor competing interests could outweigh rights. We will assume that this criticism is correct. Since any interest is allowed to play out on the balancing stage if it is considered to be a legitimate aim, even minor interests will find their way into the balancing stage and can, at least theoretically, outweigh rights. Constitutional rights which have constitutional status could thus be outweighed by minor interests without constitutional status. Such an approach in fact doesn’t protect fundamental rights properly. It guarantees no effective protection for the fundamental rights and deprives them of their normative power. Therefore, the interest model is not convincing.
Trump models, in contrast, are based on a basic liberal intuition that rights enjoy some kind of special priority, which gives them lexical priority over other considerations,17 in particular over the public interest. Tsakyrakis clearly prefers this understanding of rights by saying that ‘by definition, any treaty for the protection of human rights gives priority to rights’. He continues that:
there are, at least, two controversial assumptions…: first, that public interests, as a matter of principle, can always be weighed against human rights; and, second, that measures aimed at promoting a public interest may prevail unless they impose an excessive restriction compared to the benefit they secure.18
However, Tsakyrakis doesn’t explain what exactly his favourite model of rights is. His idea of rights as trumps is rather unclear. However, the idea of rights as trumps can be reconstructed in three different approaches. We will deal with each of them in turn.
On the one hand, there is the ‘strong trump model’. It assumes that rights are absolute rights, construed by applying the so-called internal theory,19 and thus not amendable to limitations. There is only one legal item: the fundamental right with certain content.20 Therefore, the extent of a right is not a matter of its content and its limitations combined, but of its content alone. The protection is determined by the definition of the right’s scope alone.21 As a consequence, any interference with the right is tantamount to the violation of the right.22 This model has, in fact, the best possible normative power:23 since the right is absolute, all competing interests will always be overruled by the right. The right to freedom of speech, for example, would be a rule that prohibits any interference with speech.
For three reasons, however, this model is not convincing. First, it doesn’t reflect the constitutional reality in modern human rights catalogues. Most modern human rights catalogues are aware of the fact that rights must be limited in certain cases. Therefore, they provide either written or unwritten limitation clauses. They either contain a written general limitation clause that applies to all rights, or written specific limitation clauses for single rights.24 In Canada, for example, the Canadian Charter of Rights and Freedoms contains a limitation clause that reads: ‘The Canadian Charter of Rights and Freedoms guarantees the rights and freedoms set out in it subject only to such reasonable limits prescribed by law as can be demonstrably justified in a free and democratic society.’ This means that all the rights guaranteed in the Charter are expressly subject to justified limitations. Absolute rights therefore do not exist in the Canadian Charter.25 Similar general limitation clauses can be found in Article 52(1) of the Charter of Fundamental Rights of the European Union, in the New Zealand Bill of Rights 1990, s 1, and in the Constitution of the Republic of South Africa 1996, s 5. As for specific limitations clauses, we may refer to the ECHR in which the rights conferred are ‘not absolute’.26 A typical specific limitation clause, for example, is Article 10(2) ECHR:
(1) Everyone has the right to freedom of expression…
(2) The exercise of these freedoms… may be subject to… restrictions…
as prescribed by law and are necessary in a democratic society…
But even if there is no written limitation clause for a certain right, the need for limitations is widely accepted. Several rights in the German Basic Law, for example, such as Articles 4(1) and (2), 5(3) and 8(1), are guaranteed without limitation clauses. The German Constitutional Court nevertheless ruled that:
conflicting constitutional rights of third parties and other legal values of constitutional status … are capable, in exceptional circumstances, of limiting unqualified constitutional rights.27
This is, in Alexy’s words, an unwritten limiting clause enacting constitutionally immediate limits.28 With regard to the ECHR, Rivers stated:
Thus at least as a rule of thumb it would seem that no Convention right is unqualified.29
Most human rights are therefore relative rights. Absolute rights are rare.30 An approach of absolute rights can at best be applied to rights without any limitation clauses. It is an America-centred approach,31 since many of the most important American rights do not contain express limitations.32 Therefore, US constitutional law tends to give rights priority over other considerations.33 It is thus much more rule-based than, for example, European constitutional law.34 But considering rights with limitations as absolute rights does not seem realistic. The overwhelming body of rights guaranteed in human rights catalogues is relative. This is neglected by assuming that all rights were absolute.
Secondly, the strong trump model is not practicable. In practice, there may occur a need for the state, in some cases such as hate speech, to interfere with the right.35 This would have to be taken into account by defining the right if rights are considered to be absolute, since limitations do not exist. A good example for that is the ECtHR’s jurisprudence on Article 3 of the ECHR: on the one hand, it insisted on the absolute character of Article 3 which is guaranteed without express limitations;36 on the other hand, it applied a balancing approach.37
An absolute right would thus almost always trump other considerations, except for extreme situations in which, for example, it is necessary to ‘prevent a clear and serious danger’.38 The definition of the right to freedom of speech could run like this: ‘The right to freedom of speech protects all speech except for hate speech’.39 This approach is typical for the jurisdiction of the First Amendment in the United States for example, as Stone Sweet points out:
The Court has constructed some rights (eg, political speech) as quasi-absolute—shields against (or ‘trumps’ with respect to) state acts.40
Thus, the definition of an absolute right would have to take into account the reasons that may justify an infringement of the right,41 or alternatively would have to exclude behaviour that will never prevail over other considerations.
Such narrow definitions are, thirdly, deficient.42 In fact, they rely on balancing, since the definition is based on reasons for and reasons against the protection.43 Rivers has clearly demonstrated this, both generally:
Rights are only absolute … once that process of balancing has been completed44
and also with respect to the European Court of Human Rights which:
engages in balancing even in the context of unqualified, rule-based, Convention rights.45
He pointed out that:
relatively precise rules … are to be seen as the outcome of an underlying balancing approach.46
In this point, Rivers is sided by Stone Sweet, who remarked that:
all of these outcomes, to the extent that they are stable, can be traced back to seminal acts of balancing.47
Stone Sweet illuminatively summarized this point as follows:
Pushed out the front door, balancing comes in through the back, where it is used to create ever more nuanced rules and exceptions.48
Assuming that the right to freedom of speech is absolute and, by definition, does not protect hate speech, then this definition is the outcome of a balancing test that takes both interests into account, namely the right of the freedom of speech and the interests of the addressed person. It then over-weighs the rights of the insulted person over the freedom of speech. Balancing is not only unavoidable; it also takes place in a hidden way and is thus no longer traceable. This prevents proper judicial review of the balancing. The definitions are, furthermore, structurally deficient, since the rights’ content and the considerations for the rights’ restrictions are mixed up. The argumentation is unstructured and obscure. Moreover, the definitions cause legal uncertainty, since the rights’ content depends on balancing and thus varies from case to case, dependent on the competing considerations taken into account. Furthermore, protection can be denied without openly giving reasons for it. The strong trump model reduces the state’s duty to justify rights’ restrictions and promotes judicial arbitrariness. Therefore, the strong trump model is deficient. Rights are to be constructed as relative rights.49
There are two remaining options to construe rights as relative rights: the ‘medium trump model’ and the ‘weak trump model’. The ‘medium trump model’ translates the basic liberal intuition that rights enjoy some kind of special priority into a model of relative rights. It is aware of the fact that rights need to be limited in certain cases. But since rights are considered to be trumps, the limitations are defined rather narrowly. Certain reasons must be excluded from limiting a right. Only ‘reasons of a special strength’50 are able to override the position protected by the right. Rights can only be overruled where ‘it is immediately apparent that countervailing concerns have significantly greater importance than the protected interest’.51 A proportionality test, however, is not applied in this model. In particular, the medium trump model rejects balancing. Instead, it is supposed to raise ‘the bar for justifying infringements when compared to the requirements of proportionality’.52 This model is similar to the US Supreme Court’s ‘strict scrutiny’ test, requiring inter alia that a statute must serve a ‘compelling state interest’.53
For five reasons, this approach is not convincing. First of all, it is not clear what interests will be allowed to play out. It is impossible to define the interests that are of ‘significantly greater importance’ or have a ‘special strength’ abstractly. Whether a certain interest is more important than a right depends on the right in question and on the circumstances of each single case. Therefore, secondly, the question whether an interest has ‘significantly greater importance than the protected interest’ presumes a comparison between the importance of the interest pursued and of the interest protected. This is exactly what the balancing approach provides for. It correlates the infringement of a right with the importance of the interest pursued. Thus, the medium trump model, contrary to its own inclination, is based on balancing considerations.54 But thirdly, it nevertheless doesn’t balance openly. The reasons given to establish that certain behaviour is protected or not protected are not revisable. Thus, fourthly, it is structurally deficient. The argumentation is not structured by a rational method like the law of balancing. It relies on a hidden sort of balancing. This, fifthly, promotes judicial arbitrariness, since the reasons are not presented openly and cannot be revised. We therefore hold that the medium trump model cannot realize the idea of rights as trumps properly.
The ‘weak trump model’, in contrast, incorporates the idea of rights as trumps into the proportionality test. In this model, rights as constitutional values can only be overruled by other constitutional values.55 The weak trump model thus avoids the weaknesses of the interest model. Constitutional rights always trump every consideration except for considerations which enjoy constitutional status likewise. We can refer to this rule as the first law of trumping. Within the structure of the proportionality test this means that only sufficiently important, ie constitutional, values are considered as legitimate aims. Interests that do not enjoy constitutional status are excluded from the class of legitimate aims in the sense of the first prong of the proportionality test. It follows that only interests of constitutionalvalue are allowed to play out on the balancing stage. The legitimate aims can be expressed in a written general limitation clause,56 in a written specific limitation clause,57 or in an unwritten limitation clause.58
To establish that the weak trump model is indeed convincing, we will address four challenges. The first (section IV.1) argues that a weak trump model offered no effective protection. A similar challenge (section IV.2) argues that balancing would not allow for any trumping. Closely related to these two challenges is a third one (section IV.3) claiming that it was not possible to uphold absolute rights in the balancing mode. Finally, (section IV.4), it is claimed that balancing tends to outweigh the interests of individuals and minorities.
1. Effective protection
Tsakyrakis argues that the ‘view that constitutional rights are nothing but private interests whose protection depends, on each occasion, on being balanced with competing public interests, in fact, renders the constitution futile’.59 He claims that the weak trump model, by not giving rights priority over other constitutional considerations, deprives rights of their normative power and does not protect rights effectively. Rights, in Justice Black’s words, are ‘balanced away’.60 And even Beatty, a supporter of the proportionality test, says that in proportionality, rights ‘have no special force as trumps’, but are ‘just rhetorical flourish’.61
In particular, Tsakyrakis criticizes that human rights and public interests would ‘compete on a par’ rather than giving priority to the former.62 A similar claim has been raised by Webber, who argued that this would amount to ‘undoing the constitution’ and ‘doing violence to the idea of a constitution’.63 Referring to the ECHR, Greer has argued that ‘the teleological principle suggests the principle of “priority to rights”’ and that ‘rights should be “prioritized” over collective goods in different ways according to the terms of given Convention provisions’.64 Greer concludes that Alexy’s theory could not be applicable to the Convention since the Convention formally assigned priority to Convention rights, whereas Alexy’s theory treated ‘constitutional rights and collective goods as competing principles capable of being balanced’.65
We do not agree with this argument. The weak trump model contains three ‘firewalls’ which ensure that the protection guaranteed by the constitution is effective. First, only legitimate aims of constitutional status are considered to be able to compete with a right on the balancing stage. Unlike the interest model, this guarantees that only constitutional values can outweigh a right. Minor interests would already fail to pass the legitimate aim test. Therefore, all public interests that can overrule a right must be of constitutional status. This means that rights are given strict priority over every other consideration except for considerations of constitutional status. Constitutional principles always trump all considerations that do not enjoy constitutional status. But, admittedly, rights are not given priority over, for example, public interests of constitutional status. Rather, they both compete on the same level. This has been correctly stated by Borowski:
Balancing presupposes that rights and goods to be balanced are at the same level in the hierarchy of the legal system.66
However, this doesn’t mean that the protection guaranteed by the weak trump model is inappropriate, because secondly, the law of balancing ensures that the importance of a right is taken into account in every single case. Rights can and will be given the weight they deserve to prevail over the public interest if necessary. Referring to the relation between individual rights and collective goods, Kumm has stressed this point convincingly: