In his interesting and thoughtful defence of proportionality, Madhav Khosla concedes that “certain types of balancing [are] objectionable”. He cites with approval the Israeli Supreme Court decision in Public Committee Against Torture in Israel, which affirmed a blanket prohibition on torture, and he endorses Ronald Dworkin’s warning against the “balancing metaphor” which, if taken to its extreme, might be thought to warrant an assessment of “how much torture we should facilitate”. Nevertheless, Khosla resists the idea that, at least in some cases, proportionality should not be applied. That is because in his view the application of the proportionality test does not necessitate balancing. He claims that there will be cases in which the impugned measure fails the first subtest of proportionality that examines whether the purpose of the impugned measure is legitimate. In these cases, the proportionality inquiry stops at the first step, and hence no balancing comes into play. He uses this account to explain how the proportionality test, properly understood, would vindicate the prohibition of segregation (Brown) and the prohibition of discrimination (the Eskimo-haters hypothetical), without any recourse to balancing. In short, he holds the view that balancing may sometimes be wrong, but that does not render proportionality defective. Accordingly, his basic critique against my argument is that I underestimate the first two subtests of proportionality (legitimate purpose and suitability) and thus mistakenly confuse the entire test with its third component (proportionality in the strict sense).
Khosla offers a welcome reformulation of the proportionality test and its operation. Still, I remain unconvinced. I stand by my original view that bringing these cases under the proportionality test goes against our basic intuitions about the meaning of human rights.
Is there any way to consider proportionality without referring to balancing? Proportionality is commonly understood to be a specific method of balancing. It is premised on the same basic idea that the only rational way to assess the legitimacy of a limitation of a right (as Alexy for example says) is by balancing the purpose served by the limitation against the severity of the infringement of the right and appears to be a specific heuristic device for the achievement of that purpose. I believe that this pedigree exerts a strong influence on the application of the proportionality test. As a result, although it is no doubt true that the proportionality test consists of three subtests, it is the third one that is still the most crucial. This is so in at least two senses. First, as a matter of fact it is treated as the most crucial. Second, the architecture of the proportionality test would be destabilized if the first subtest were beefed up to accommodate the human rights philosophy that I argued for. Running the risk of repeating some of the claims of the original article, I want to say a few words about each of these two senses.
To begin with, although judges as well as the most ardent defenders of proportionality pay lip service to the first two subtests, they really don’t attribute much significance to them. This is evinced in their endorsing an approach with regard to what constitutes a right that I called definitional generosity. It is also evinced in the rather cavalier approach they adopt when specifying the value that is set against the right. To illustrate this approach, I discussed the Otto-Preminger and IA cases, where it did not really matter whether the purpose of the limitation was the protection of a right or a public interest. Why are these tendencies in the case-law and the doctrine relevant to the role of the first sub-test? They are, because an analysis of the content of the right that is more closely attuned to its moral point would yield priorities between rights and interests and dismiss certain interests as inappropriate grounds for restriction. In other words, it would do precisely the kind of work that Khosla wants to assign to the first sub-test.
There is a natural explanation for these tendencies. Many people are reluctant to engage in a morally informed analysis of human rights. They fear that it necessarily leads to indeterminacy. This fear is assuaged if any explicit recourse to moral reasoning is avoided and all the crucial issues are deferred to the balancing stage, since balancing is typically portrayed as a morally neutral exercise. But, as I tried to show, when the more demanding inquiry that I urge for is foregone, the outcome of such balancing is not only controversial but seems hopelessly arbitrary, and the Eskimohaters hypothetical aimed to illustrate this.
Khosla suggests that we can avoid such mistakes, if, instead of entirely reducing human rights adjudication to questions of relative weight in order to bypass the moral discourse on values and priorities, we emphasize the first subtests of proportionality. But I think his proposal fails because either he follows the traditional path or he masks moral discourse with the rhetoric of proportionality.
To see this, we should examine what would be the implications of his proposal, if it were put into practice. Firstly, we must notice that if the assessment of a right is possible without reference to the intensity or the severity of its restriction, there is no reason whatsoever to say that this inquiry assesses the proportionality of the restriction in the first place. But, it is more important to reflect on the substance of his warning not to treat the first test of proportionality as farcical. What kind of inquiry are we supposed to undertake? How should we establish whether the purpose of the restriction of a right is “sufficiently important,” “legitimate” or “constitutional”? In a well ordered society the limitation of a right is rarely completely irrational and its aim is rarely obviously illegitimate. Moreover, when we deal with such an obvious illegitimacy or irrationality, we hardly find it problematic to condemn it. But the more common and theoretically more interesting cases are those involving less blatantly illegitimate purposes. For instance, in Otto-Preminger the purpose invoked was public order and the rights of others, not bigotry. “Separate but equal” policies are unconstitutional not necessarily because they are actually animated by racism but because they offend equality, properly understood. Similarly, we cannot dismiss out of hand the view that torture is “suitable” or “necessary” in some situations of extreme emergency. That is, not without engaging in an exercise in political morality that relates values, sets priorities and orders them into a system. The more of that we do, the less room there is for the balancing stage.
Is it enough to save the proportionality test, if there is still some (however little) room left for the last stage? Khosla seems to think so. He cites approvingly Grimm’s hypothetical of a law permitting the police to shoot someone to prevent him from destroying property. He says that “here the purpose is appropriate as it aims to protect property; there is a rational connection since the shooting will stop the perpetrator; and the means are narrowly tailored as there is no other way to stop the destruction”. In this situation he argues only proportionality in the strict sense will resolve the conflict between life and property. In other words he thinks that it is possible to resolve this conflict as the traditional proponents of proportionality contend, by just investigating if the shooting is too intrusive a measure compared to the loss suffered in terms of the right to life.
In fact, though, the example shows the instability of his proposal. If the scrutiny at the first stage rests content with the abstract identification of the protection of property as the purpose served by the measure, then it leads us to the balancing stage without the appropriate theoretical resources to resolve the conflict between the right to life and the right to property. If, on the other hand, it embraces the more searching analysis of the competing rights from the perspective of political morality, then the balancing stage is rendered redundant. For, this analysis will have already assigned priorities to the values of life and property. Either way it is untenable to argue that there is need for a moral discourse of the sort I argue for and at the same time to engage in a proportionality assessment.
One final point. I think that Khosla’s attempt to safeguard proportionality runs into the same risk I envisaged in the Otto-Preminger and IA cases. That is, it risks obscuring the decision making process. To see this, consider Khosla’s claim that our disapproval of the outcome of these cases has more to do with the poor implementation of proportionality or the margin of appreciation doctrine of the Court and less with flaws of the principle per se. On the contrary, Khosla thinks, the F v. Switzerland case (a challenge of a law imposing a three-year prohibition on remarriage) was decided through the proper application of proportionality, that is, by inquiring whether the means of the impugned measure were appropriate and not by evaluating the proportionality of the length of the prohibition.
Now one may wonder why the Court had to use the proportionality language if the real ground of its ruling was the doctrine of the margin of appreciation. The same is true for the F. v. Switzerland or the Lusting-Prean and Beckett v. UK case. To my mind, in both cases the ruling of the Court was based on substantive moral reasoning but was marred due to the use of the language of proportionality. As a result, the final judgment obscured the Court’s decision-making process. It stands to reason that the more explicit the moral considerations of a case are made, the clearer the reasoning of the court.
[*] Associate professor of constitutional Law, University of Athens
Published in International Journal of Constitutional Law, (2010) Volume 8, Number 2, pp. 307-310.