15 - Identifying Custom in Universal Periodic Review Recommendations

Summary

Since 2008 the Universal Periodic Review (UPR) mechanism of the UN Human Rights Council has become a fixture of the international human rights regime, reviewing the human rights record of every single state once every few years. Over the course of the UPR reviews to date thousands of recommendations have been issued about the protection of human rights with a growing body of evidence showing that states modify their behaviour in response to the outcome of reviews. This chapter analyses how UPR recommendations can be used to identify customary international law. Using the discussion on the identification of custom in General Assembly resolutions as a starting point the chapter examines how scholarship on the identification of custom has argued that there are a different set of rules for the identification of customary international human rights law. Identifying custom in UPR recommendations can help resolve some of the ambiguities and contradictions in relation to the identification of customary international human rights law. This chapter concludes by outlining a three-part model for identifying custom in UPR recommendations which would complement existing rules for identifying customary international law.

1 Introduction

In January 2019 the Dominican Republic went before the United Nations (UN) Human Rights Council (HRC) for its review during the third Universal Periodic Review (UPR) cycle. In the report it submitted to the council in advance of their review it was at pains to demonstrate the various legislative measures it had taken in respect of domestic violence.¹ The country had been criticised by NGOs for systemically failing the survivors of domestic violence both in legislative terms – prior to 1997 it was not a crime in the country – and in relation to the training of law enforcement officials.² As such the Dominican Republic was keen to demonstrate that, following on from its review in the second UPR cycle in February 2014, it had made changes to the law to reflect the recommendations made to them on the issue of domestic violence by states conducting the review at the HRC.³ When considering the Dominican Republic’s report in 2019 the Australian representative to the HRC expressed concern that some of the domestic legislative reforms did not go far enough, leading it to offer a series of further recommendations on domestic violence and the strengthening of police accountability.⁴ Throughout 2019 recommendations were offered to states undergoing the UPR on domestic violence, some going further than the standards set out by the Committee on the Convention of the Elimination of all forms of Discrimination Against Women (CEDAW Committee) others more general in nature, and some building on recommendations offered to states over the previous two review cycles. Universal Periodic Review recommendations are numerous and wide ranging but, when aggregated, can demonstrate certain trends in relation to human rights protection. Although some UPR recommendations are relatively trivial in nature and others concern matters relating to commitments under human rights treaties, there is a large class of recommendations framed in legal language, recommending a specific practice in relation to the protection of human rights to the state under review that is novel or relates to a particular interpretation of a widely acknowledged right. An analysis of these recommendations can show the emergence of customary international human rights law.

The HRC was created by UN General Assembly (GA) in 2006 and the UPR process was one of the most significant features of the new body, which replaced the UN Commission on Human Rights.⁵ There has been a wide-ranging debate about the role of GA resolutions in the formation of custom. At the 1945 San Francisco Conference which founded the UN a proposal by the Philippines to give the GA the power to enact rules of international law which would become effective and binding upon members was defeated 26–1. Yet, from the beginning UNGA resolutions were often shaped in a way that suggested that they aimed to have some form of legal effect on states. In 1951 in the Advisory Opinion on Genocide Reservations Judge Alverez observed that GA resolutions had ‘not yet acquired a binding character’ but noted that if resolutions had the support of ‘public opinion’ they might be recognised as having some form of force over a state.⁶ During the 1960s it became clear that certain declarations contained in GA resolutions were treated as quasi-legal statements of authority – in particular Resolution 1514 which called for the end of western colonialism which was often recited in subsequent resolutions.⁷ Rosalyn Higgins concluded in a 1965 paper that the repeated practice of UN political organs was of ‘probative value as customary law’.⁸ There are a number of issues that arise with any analysis of the legal status of GA resolutions, such as the status of opposition to resolutions, which makes their customary status contentious.⁹ The importance of the debate over custom and UNGA resolutions is that it provides a useful comparison point for understanding how custom can be observed in UPR recommendations.

Understanding the status of UPR recommendations is important in the context of understanding customary international human rights law. Section 2 of this chapter shows that human rights law poses a number of problems for the traditional assumption that custom requires both state practice and opinio juris.¹⁰ Due to the way that UPR recommendations shape state behaviour and because of the importance of the review recommendations, the remainder of this chapter argues that UPR recommendations can be a useful means for observing the formation of customary human rights rules. The International Court of Justice (ICJ) has held that certain features of GA resolutions such as the context of their emergence, the language of their substantive provision and the reaction to states can be evidence of a consensus supporting the emergence of a new customary rule.¹¹ It is possible to trace similar features within a series of UPR recommendations on a particular norm and a framework for analysing UPR recommendations and identifying custom is set out in the final part of this chapter. But as the conclusion goes on to outline this raises wider questions about the rules for identifying custom.

2 Identifying Customary International Human Rights Law

The traditional concept of customary international law (CIL), which the International Law Commission’s (ILC) Draft Conclusions adopts, is justified on the basis that the two-element concept is necessary in order to maintain the ‘unity and coherence of international law’.¹² The idea of different branches of international law having different rules in relation to law formation was rejected by the ILC even though, as Hugh Thirlway notes, there has been a widespread literature arguing that certain fields, such as international human rights law and international criminal law, ought to be treated differently with only the requirement for opinio juris to be present.¹³ As Jean d’Aspremont also notes, exceptionalist thinking about the role of international human rights law, drawing on the non-reciprocity of rights and the importance of human rights for a constitutional framework of international law, led to a call for the ‘argumentative structures of general international law … not [to] apply’ to international human rights law or for them to be in some way ‘loosened’.¹⁴ The seeming inflexibility of the rules surrounding custom has led other scholars to argue that the requirements of CIL are actively detrimental for the protection of human rights.¹⁵ There is not really space here to re-examine the different schools of thought in this debate, but there are three important practical issues which arise with any attempt to identify customary international human rights law.

Firstly, the requirement to demonstrate consistency of state practice is hobbled by the basic reality that there are widespread human rights abuses perpetrated by states and human rights practice is often wildly inconsistent.¹⁶ Realist critics, such as Goldsmith and Posner, have explained this by arguing that divergence of practice is simply a reflection of state interest as the ‘behavioural regularities’ require external incentives, such as coercion from more powerful states.¹⁷ Other realist critics have taken this further arguing that state interest is in practice reducible to what is required for state survival and as a consequence custom is an inappropriate vehicle for human rights, when compared to treaty law which clearly binds states and defines their obligations.¹⁸ The counter argument to this is that the practice requirement matters less in relation to international human rights law, and, as the ICJ has clarified, the search should be for consistency of practice rather than rigorous conformity.¹⁹ A lot of this argument depends on what is recognised as state practice, as the implementation of rights is different from commitment to the protection of rights.²⁰ Even though there are potentially good reasons for acknowledging that acceptance of human rights norms through instruments such as GA resolutions is important in altering the normative consensus that leads to the protection of human rights, this leads to uncertainty about what precisely constitutes state practice and where the intention of a state behind a practice can be distinguished from the intention to be bound by that practice.²¹

Secondly, there is genuine debate about the nature of the prohibited practices involved in the protection of human rights, sometimes referred to as the secondary rules problem.²² D’Amato frames the problem thus: ‘[w]hat are the parameters of torture? … Is the battering of wives “torture”? … what constitutes “inhuman treatment or punishment”?’.²³ The existence of a strong general consensus over particular norms – such as the prohibition of torture – does not necessarily mean clarity over the practical implications of what they entail, even if the customary status of certain rights are citied and recited by courts.²⁴ For example, Ghana is a party to the International Covenant on Civil and Political Rights (ICCPR) and the Human Rights Committee has since 1994 consistently interpreted the right to privacy under Article 17 as being incompatible with the criminalisation of same sex relations; yet Ghana not only ignores this but has also rejected UPR recommendations to decriminalise sexual orientation.²⁵ This speaks to a third related problem in relation to customary international human rights law – the idea that some states do not regard human rights law or human rights practice as creating a legally binding obligation upon them. Even though almost every state in the world is party to at least one legally binding international human rights instrument, mechanisms such as reservations have allowed states to manage the scope of their obligations, and states have used a variety of arguments to maintain they are not bound by treaty obligations in relation to specific rights, relating to practices considered culturally sensitive, or in areas where they have security concerns.²⁶ The persistent objector doctrine can also perform a similar function in that it allows states to define the limits of what should and what should not be considered human rights.²⁷

In an attempt to refocus the debate surrounding the nature of the subjective element of custom, Brian Lepard has argued that a rule or principle ought to be considered customary law if it can be shown that ‘states generally believe that it is desirable now or in the near future’ to make a ‘rule or principle’ legally authoritative ‘for all members of the global community’.²⁸ Lepard’s argument is that ‘state practice is importance of evidence of the belief that a norm should be universally binding’ but that that it is not an ‘essential independent requirement’; a position criticised by some scholars as fusing the two elements together.²⁹ Yet, as Lepard goes on to argue, his reformulation clarifies the role of opinio juris, because it looks at what should or ought to be binding, rather than what is or may be perceived as binding.³⁰ A case study of this approach to custom is the GA resolution containing the Declaration on the Granting of Independence to Colonial Countries and Peoples.³¹ This recognised emerging practice from Western states to grant colonial independence and was recognised by many states (particularly those who were newly independent) as having a legal quality to it because it supported the desirable goal of independence, whereas others (European states with colonial territories) were far more reluctant to concede its legal status.³² Although this is not a formula for replacing the two-element rule – it instead refines how the individual elements are identified and examined – it has come in for criticism as being a form of ‘norm entrepreneurship’ which dilutes the meaning of custom.³³ Given the competing difficulties that the identification of customary human rights law poses, there is a need not so much for new rules but for a broader consideration of the materials used in the identification of the two elements of custom. An analysis of UPR recommendations, as the remainder of this chapter sets out, provides two clarifying functions for the identification of customary international human rights law. Firstly, it allows for a transparent and more democratic way of measuring the existence of a common consensus on a particular human rights norm. Secondly, the way the UPR process works allows for the contours of any norm to be defined, which is important when that norm is a secondary or interpretative norm about the scope of a particular practice which emerges outside of an agreed codification of a particular right in a treaty or one that is recognised elsewhere as jus cogens.

3 UPR Recommendations and Their Effect on State Behaviour

Under UPR rules every UN member has their human rights record reviewed around once every four years – known as UPR cycles.³⁴ The review is conducted based on documentary evidence from the state under review, reports from treaty bodies to which the state under review is a party, stakeholder reports from civil society groups in the state under review, international NGOs and other consultative bodies. Every state has participated in the process since its inception in 2006 and the treatment of states as equal peers has been a significant attraction of the process, differentiating it from its predecessor the UN Commission on Human Rights. During the review there is first a documentary review, examining the state under review’s performance at treaty bodies, reports from stakeholders, and its own report. Then there is an interactive dialogue between the state under review and other members of the review panel. After that there is the opportunity for all states to issue recommendations to the state under review about changes in domestic law in order to improve their human rights practices. The UPR is a political process and was not intended to be law making but it involves scrutiny and discussion of a state’s human rights obligations which leads to it sometimes overlapping with other international legal processes.³⁵ It was meant to complement and not duplicate the work of treaty bodies and in relation to some treaties the UPR has played a role in reinforcing obligations, by recommendations being cited by treaty bodies as evidence of state practice in relation to a particular norm.³⁶

At the time of writing in October 2021 there have been nearly 79,000 recommendations issued to states in the 11 years that the review process has been in operation. Assessing recommendations is difficult because they have no real set form and are constructed by states acting individually rather than collectively or drafting them in concert with others, as is the case with UNGA resolutions. Edward McMahon has devised a system for categorising recommendations based on the nature and quality of action required of a state, ranking them from one to five.³⁷ Category two recommendations concern general comments about what the state under review is currently doing and a request to continue an ongoing course of action. For example a recommendation to Brazil from Senegal during the second review cycle to ‘continue fighting violence against women’ simply asked the state under review (Brazil) to do nothing beyond what they were currently doing.³⁸ There has been some criticism of these sorts of recommendation being little more than offerings of praise from the states conducting the review. Sometimes when the state under review was their political ally, states would praise-bargain hoping that favourable recommendations would lead to the state under review affording them similar treatment when it was their turn for review.³⁹ In 2011, at the end of the first review cycle, a lot of emphasis was placed on the process of following up the implementation of recommendations accepted by states in subsequent review cycles.⁴⁰ Follow up can occur in the portion of the review dedicated to interactive dialogue. Sometimes recommendations specifically cross-reference previous commitments; for example, the recommendation issued to India by Botswana during its third cycle review to ‘Ratify the Convention against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment, as previously recommended’ was designed to reinforce earlier commitments.⁴¹

Recommendations in action categories four and five are framed using language which requires positive action on the part of a state, and directly reflects the legal language of a treaty or declaration. Recommendations classed as category five are those using legal verbs such as ‘abolish’, ‘accede’, ‘adopt’, ‘amend’, ‘implement’, ‘enforce’, or ‘ratify’.⁴² To illustrate this with another example from Brazil in its second cycle review; Spain issued a recommendation to ‘adopt Bill No. 2442 in order to guarantee the independence and autonomy of the members of the National Preventive Mechanism, in conformity with Brazil’s obligations under OPCAT’ which is both precise in its intent and framed with clear instructions to the state party on the course of action to take.⁴³ After recommendations have been issued it is up to states to either ‘Support’ (UPR terminology for accept) or classify a reservation as ‘Noted’ (UPR terminology for reject). There has been some criticism of McMahon’s framework as being too narrow, and concern that its focus on the type of action required to implement it obscures the utility of the recommendation for the state under review.⁴⁴ Yet, the linguistic framing of recommendations is vital for identifying and distinguishing what could appear to be the language of obligation and commitment from general descriptive language about human rights. Category five recommendations are the recommendations that are most likely to be rejected by the state under review – in the first UPR cycle which ran from 2008 to 2011, 60 per cent of category five recommendations were rejected. In the second cycle, which ran from 2012 to 2016, 55 per cent were rejected.⁴⁵ Significantly, in spite of this rejection rate the number of category five recommendations has remained steady over successive review cycles – 35 per cent of all recommendations in the first cycle were category five, in the second cycle it was 37 per cent of recommendations, and of the data available so far for the third cycle, 38 per cent have been category five recommendations. It is, however, important to remember that the categories are analytical tools for understanding the framing of recommendations and not a formal part of the UPR process. Whilst shaping the wording of recommendations, as they author them, states are not necessarily conscious that they are making a recommendation of one particular category or the other. They are best understood as an instrument of measure; ascertaining both the quantity of recommendations and the relative severity of linguistic framing. Moreover, while they may serve as a useful proxy for state intention, they are not always definitive proof of it.

Once accepted, a recommendation does not create an obligation upon the state under review to implement its substance. Recommendations do however affect state behaviour; firstly, the process of follow up, or at least the expectation of follow up, and the deliberative nature of the review process does encourage states to make incremental change to their behaviour in respect to their laws and policies respecting human rights.⁴⁶ Jane Cowan and Julie Billard’s observation that states under review can treat their review like an ‘exam’, with state delegations to the UPR concerned about proving they have met minimum standards or demonstrating the implementation of legal reforms, shows that the UPR process can change state behaviour, even if it does not meet some of the loftier objectives about the promotion of dialogue on human rights intended by the UPR’s creators.⁴⁷ Secondly, in terms of implementation and delivery there is a correlation between states accepting recommendations and implementing changes to their law and policy surrounding human rights in response to those recommendations.⁴⁸ Sometimes this relates to a course of action already decided upon by a state party and recommendations help reinforce this course of action.⁴⁹ Yet on other occasions there are signs that recommendations act as drivers of reform independently – for example in relation to protection from human trafficking and maternal health there has been some research showing recommendations correlate with the adoption of higher standards on these issues in countries which have accepted recommendations.⁵⁰ There was also an upsurge in the number of states signing up to and ratifying treaties in the wake of the first UPR cycle, again seemingly in response to a wave of recommendations in the first cycle relating to treaty provisions.⁵¹

Regular patterns of state behaviour in accordance with the terms of some GA resolutions has been cited by some scholars as proof of their customary status, even though such behaviour might not be in conformity with all of the terms of a resolution.⁵² Regardless of whether states consider the UPR to be a form of ritualised audit or approach the review as an opportunity to advance strategic interests, states’ behaviour towards the process indicates that issuing and accepting recommendations carries a degree of importance.⁵³ A crucial distinction between UPR recommendations and GA resolutions however is that certain resolutions, such as those on outer space or environmental issues, were claimed to create instant custom, without state practice.⁵⁴ Some of the most compelling arguments in this debate related to recommendations containing declarations which made normative pronouncements about what the law ought to be in a particular area, and which was then put in a codified text of a resolution and voted on.⁵⁵ In 1977 this argument was made by the Group of 77 (a UN grouping principally consisting of newly independent states in Africa and Asia) who contended that a UN resolution on the seabed should be regarded as binding because it represented a true international consensus on what the law ought to be, had been expertly drafted, and the GA resolution containing it passed without any votes in opposition.⁵⁶ Universal Periodic Review recommendations differ from GA resolutions as they are drafted by individual states and reflect their own interest and priorities. As Gujadhur and Limon note, one of the problems the UPR process faces is the sheer weight of recommendations, many of them of variable quality and relatively poorly constructed.⁵⁷ Acceptance and rejection of individual recommendations can mean relatively little and can vary considerably from state to state. It is therefore more important to look at recommendations in aggregate on a particular issue to see the reflection of a normative consensus on any one human rights issue.

4 A Lens through Which to See Custom in UPR Recommendations

Custom’s formation is often described as being observed rather than generated. As Anthea Roberts notes, traditionally custom was ‘inductive’ in that it was derived from an observation of state practice, whereas modern custom is ‘deductive’ in that is deduced from international instruments, such as declarations and reservation.⁵⁸ As Stefan Talmon notes, however, it is incorrect to think of this as a choice between the two methods and at the ICJ there have been situations where it was simply not possible to use an inductive method to identify custom.⁵⁹ What is presented here therefore is a deductive framework to use for identifying the emergence of customary human rights norms in the UPR. This framework breaks into three parts and arguably provides greater clarity in the context of the UPR than the two-element approach.

4.1 Acceptance of Recommendations and Practice

Although state practice was historically conceived as the physical acts of states, for example by controlling which ships were allowed into a particular area, there is now a general recognition that verbal acts can in certain circumstances constitute state practice.⁶⁰ UPR recommendations are issued through an official process, created by a GA resolution that has a broad-based international acceptance and is treated by human rights treaty bodies as being authoritative evidence of state practice in relation to a particular norm. Accepted recommendations go beyond mere statements of practice.⁶¹ International decisions – such as a GA resolutions – are exceptional as their institutional provenance means that they are considered indicative of either how states are acting, ought to act or ought not be acting.⁶² Conclusion 4 of the ILC’s Draft Conclusions on the Identification of Customary International Law makes specific reference to international organisations in the context of state practice, and Draft Conclusion 7 notes that practice includes a ‘wide range of forms’ including ‘resolutions adopted by an international organization’.⁶³ Strictly speaking recommendations are not resolutions but they are advanced as part of an organisational process and therefore would be analogous to the processes outlined in both the ILC’s Draft Proposals and the International Law Association’s (ILA) final report.⁶⁴ In fact as the ILA’s report goes on to note the practice of international tribunals ‘is replete with examples of verbal acts being treated as examples of practice’ so the concept of practice is viewed in relatively expansive terms.⁶⁵

Yet, this raises the issue of what precise moment in the UPR process – acceptance of a recommendation or implementation of the substance of the recommendation – constitutes state practice. Implementation of a recommendation would demonstrate the existence of a concrete human rights protection within a state and therefore be the physical manifestation of a principle. But, as the ILA notes, ‘statements in international organizations and the resolutions these bodies adopt’ are more common than ‘physical acts, such as arresting people or seizing property’ leading to the conclusion that if a claim is publicly communicated it would constitute an act for the purpose of custom.⁶⁶ Following UNHRC Resolution 16/21, a state is required to ‘clearly communicate to the Council … its positions on all received recommendations’ entailing that there is a requirement on states to take a public position in relation to the recommendations they have been offered.⁶⁷ Acceptance of a recommendation is therefore made in public, recorded in an official UN document, with an expectation that its terms will be put into practice by a state; thus when a recommendation is accepted, it is state practice. As Malcom Shaw puts it, if practice is considered as simply ‘what States actually do’, then publicly making a commitment to implement a specified human rights reform and consenting to be examined on progress towards that reform in four years’ time, is what states ‘do’.⁶⁸ Yet, an individual recommendation and acceptance of it by a state under review would not really be sufficient to establish that there was state practice as a recommendation applies to a particular state.⁶⁹ Practice, according to the ICJ, needs to be widespread as well as ‘sufficiently extensive and convincing’ in order for it to be considered the basis of custom.⁷⁰ Therefore, multiple accepted recommendations of the category four or five type, which by their nature require a specific course of action on a human rights norm by states, would need to be shown in order to demonstrate a practice.

Even if a chain of accepted recommendations on the same subject can be identified, there are likely to be some rejected recommendations on the same subject. The rejection rate of recommendations in action category five supports the idea that states act with the belief that because of their framing, such commitments are in some way consequential.⁷¹ Yet Elvira Domínguez Redondo notes that this pattern of behaviour can be interpreted narrowly, as simply the state under review ‘asserting its reluctance to be monitored by the UPR on the implementation of such a recommendation during its next review’.⁷² Therefore, a state may not actually object to the substance of the recommendation but wish to avoid, for a variety of reasons, accepting a UPR recommendation on the subject. The multiplicity of motivations behind rejected recommendations means that they are difficult to read as a conclusive manifestation of the persistent objector doctrine.⁷³ As Joel Trachtman notes, the doctrine only applies ‘when the customary rule is in the process of emerging’ but that this is a somewhat problematic part of the principle as custom ‘is always in a zen-like process of becoming and un-becoming’.⁷⁴ As UPR recommendations (as shown below) on the same subject might be accepted in some cases but rejected in others, this means that it is difficult to pinpoint the moment of becoming for a norm. Furthermore, in relation to customary human rights law, the persistent objector rule could mean that a right that ought to be universal is essentially opted out of by a state, although human rights tribunals have rejected this argument where the right is considered jus cogens.⁷⁵

Rejected recommendations also do not give much of an insight into the substantive objection to a recommendation. As Lynn Loschin identifies in her four-part model for analysing the persistent objector doctrine in international human rights law, ‘the quality and quantity of the State’s objection’ would be important for validating whether the objection reflects a genuine preference of a state.⁷⁶ It is, for example, entirely possible that a recommendation is rejected based on part of its text and not as a reflection of the whole recommendation. Even if a state’s rejection of a recommendation is relatively consistent over review cycles, that also may not be grounds for saying that a customary norm should not be universal. As Lepard argues, customary human rights law should be about what rights ought to be protected as a matter of international law, not an assessment of the often-inconsistent nature of state practice.⁷⁷ Accepted and rejected recommendations therefore need to be considered in tandem in order to ascertain the nature of a norm which emerges from recommendations, but there would need to be a high number of accepted recommendations which leads onto the next issue – quantification.

4.2 Quantification: Accepted Recommendations Making a Rule?

The ICJ’s Advisory Opinion on Nuclear Weapons noted that a ‘series of [GA] resolutions may show the gradual evolution of the opinio juris required for the establishment of a new rule’.⁷⁸ Empirical studies on the formation and identification of custom in court briefs and submissions before international tribunals have been relatively inconclusive in establishing trends of how custom is identified, yet this does not mean that thresholds of practice cannot be established.⁷⁹ As Christopher Joyner noted in relation to GA resolutions, even though the law making competence of the assembly was qualified, when ‘delegates representing almost all the world’s national governments cast votes on a resolution, they are in effect providing a common confirmation (or rejection) of the presence and acceptance of that issue in international law’.⁸⁰ In the case of UPR recommendations it would mean establishing a common linguistic framing, a common subject matter and a pattern of acceptance from states with a reasonably wide geographic spread – all of which is possible using a database such as UPR Info to track the emergence or existence of such a trend.⁸¹ A constant series of recommendations all aimed at a particular practice, which are accepted and over the course of multiple cycles are adopted by states, could amount to what the ICJ describes as a ‘general recognition’ that a law or legal obligation is involved.⁸² Quantifying recommendations helps to establish to what extent a consensus surrounding a particular norm actually exists, which is important for establishing the existence of practice in the formation of custom.⁸³ Even adopting a theory of customary international human rights law, of the sort outlined by Lepard, there would need to be some acknowledgement of the scale of moral consensus surrounding a particular norm, to give weight to the claim that it ought to be universal.⁸⁴ Because of the nature of recommendations, as outlined in the second section above, it is necessary to trace a particular norm through recommendations and quantify the use of certain words in a series of recommendations on that subject.

A good case study of how this process might work is the prohibition on corporal punishment. According to the Global Initiative to End Corporal Punishment of Children, at the time of writing around 140 states prohibit corporal punishment in the criminal justice system and 132 prohibit it in the education system.⁸⁵ There are, however, far fewer states that have prohibited corporal punishment in the home or care system and in total only fifty-six states have a total prohibition on corporal punishment as a matter of law. The European Court of Human Rights has been clear that state sanctioned corporal punishment constitutes inhuman and degrading treatment.⁸⁶ An advisory opinion of the Inter-American Court of Human Rights stated that the American Convention on Human Rights required state parties to take ‘positive measures … to ensure protection of children against mistreatment’ especially in ‘relations among individuals or with non-governmental entities’ but stopped short of formally requiring the prohibition of corporal punishment.⁸⁷ Article 19 of the Convention on the Rights of the Child (CRC) requires state parties to take ‘appropriate legislative, administrative, social and educational measures to protect the child from all forms of physical or mental violence’ which does not explicitly prohibit corporal punishment.⁸⁸ In General Comment 1, the Committee on the Rights of the Child noted that in the context of protecting the right to education it had previously made clear in its concluding observations that the use of corporal punishment did ‘not respect the inherent dignity of the child nor the strict limits on school discipline’ protected in the convention.⁸⁹ In General Comment 8 the committee went further, arguing that corporal punishment was incompatible with the requirement to protect children from harm and that laws permitting corporal punishment in education needed to be repealed.⁹⁰ But crucially it stopped short of recommending the prohibition on corporal punishment in the home.

There have been a large number of UPR recommendations submitted concerning corporal punishment and overall, 58 per cent of them have been accepted. The vast majority of accepted and rejected recommendations are in action category 4 and 5, containing the words ‘prohibit’, ‘end’, ‘ban’ or ‘eliminate’. For example, France in its second cycle review accepted a recommendation from Uruguay ‘to explicitly ban the corporal punishment of children in all settings, including the family, schools and institutions’.⁹¹ Other recommendations can be more explicit in cross referencing the CRC and existing obligations in international law. For example, Uruguay recommended to Algeria in its second cycle review, after commending it during the interactive dialogue for introducing a prohibition on corporal punishment in schools, that it extend the prohibition to ‘home care institutions, penitentiary centres and any other settings, in conformity with Article 19 of CRC’.⁹² Other states have, however, been wary of recommendations which include a prohibition that would entail them prohibiting corporal punishment in the home, potentially entailing the introducing of laws which might criminalise parents; Switzerland in the third cycle accepted one recommendation on the prohibition of corporal punishment but rejected another which specifically referenced prohibition ‘in all settings, including in the home’.⁹³ Out of all rejected recommendations referring to corporal punishment, 25 per cent of them are category four or five recommendations referring to the ‘home’, ‘family’ or other term referring to prohibition on the domestic sphere.

When analysing recommendations, a basic three-part approach to quantification of recommendations would help identify the emergence of custom. Firstly (as detailed in Figure 15.1) there would need to be a quantification of both the practice, or the noun (i.e., ‘corporal punishment’) and the verb in connection to the noun (i.e., ‘prohibit’) because the commitment to the ‘doing’ or ‘enacting’ of a human rights norm is what differentiates a mere verbal statement from something which can be considered state practice. Given the scale of friendly recommendations and praise bargaining, there needs to be a standard to distinguish accepted recommendations which might be evidence of custom from accepted recommendations which are of largely political significance – hence recommendations looked for as evidence of custom considered would need to be in action categories four and five, as they involve an active commitment from a state party.⁹⁴ Although it is conceivable that a state might take action over recommendations in categories one to three, their vague and open-ended wording, which often lacks any clear description of subject matter or action to be undertaken by the state, means that it would be difficult to treat these recommendations as evidence of customary law. Secondly, the framing of rejected recommendations needs to be analysed to see where in the rejected recommendations there appears to be limitations of rights. For example, in the case of corporal punishment the only consistent trend in rejected recommendations on corporal punishment appears to be scepticism about extension of the prohibition to the domestic sphere. This should be read in tandem with other sources on the practice outside the UPR process – such as court decisions – to see if this would constitute a ground for rejection that indicated a substantive objection to a specific right and hence accepting the limited nature of any customary norm that could be identified in recommendations.⁹⁵ Thirdly, the recommendations analysed would need to be sufficiently numerous – in figure 1 above the number of accepted recommendations numbers over 200 – and across more than one cycle and geographic region, to demonstrate evidence of the widespread consensus on a particular issue.

Figure 15.1 Recommendations on banning corporal punishment across UPR cycles to date.

4.3 UPR Recommendations: The Sense of Obligation

The participatory nature of the UPR process arguably makes any consensus identified in accepted recommendations more justifiable on democratic grounds as the basis of a shared belief that a particular principle ought to be binding, in accordance with the interpretation of opinio juris set out in the first section of this chapter. All states have participated in at least one UPR review and all states are treated equally before it, in that they all get to be reviewed and can contribute to other states’ reviews. Unlike human rights treaty bodies, which subject states to review by panel of experts, the UPR process is genuinely participatory. Nicole Rouhgan’s work on the democratic formation of custom attempts to reconcile the way that custom’s formation ‘falls short of contemporary ideals of democracy’ and is characterised by an absence of a mechanism ‘to protect formal equality in the development of customary rules’.⁹⁶ Emmanuel Voyiakis echoes this criticism, noting how international systems are riven with inequalities, reflecting the interests of powerful states in the formation of custom to the extent that CIL as a concept lacked a firm ‘justification for generating rules with normative force’ over other states.⁹⁷ Most customs, as Anthea Roberts notes, are based on the practice of fewer than a dozen states, meaning that formation of custom skews towards states with power and knowledge of legal formation creating a situation which by default privileges powerful states.⁹⁸ In an attempt to re-found an understanding of CIL’s formation that is more democratic Roughan argues that it should be ‘understood at its core to be a matter of social participation’.⁹⁹ A series of accepted recommendations would be representative both of a broad commitment on a particular human rights norm from individual states accepting recommendations on that norm, but also would represent a positive statements from the states offering those recommendations on what they believed the law ought to be. But, whilst this would affect the validity of the consensus behind a particular norm, it would not give an insight into the subjective belief that the norm is or ought to be binding.

The process of taking part in the review and being scrutinised on the implementation of accepted recommendations is a form of ongoing interaction, which can build a sense of obligation. Jutta Brunnée and Stephen Toope argue that processes of institutional interaction on the part of a state can build a sense of fidelity to the institution encouraging them to reshape their behaviour so as to create a sense of legality.¹⁰⁰ Their thesis has received criticism from different directions, including claims that it is too reductive about the nature of obligations and fails to really interrogate the nature of international society within which states’ values are supposedly shaped.¹⁰¹ Yet, interaction has instrumental value in showing how the understanding of a norm as obligatory can emerge. As Brunnée and Toope note, within all systems of law (national or international) ‘law is constructed through rhetorical activity producing increasingly influential mutual expectations or shared understandings of actors’.¹⁰² Research on the politicised nature of UPR recommendations actually underscores the conclusion that states view the acceptance of recommendations as a process that involves accepting responsibilities.¹⁰³ This is because recommendations, when offered in a partisan manner, still appear consequential to the state under review and the overlap between the UPR and other legal processes means that the UPR process itself is seen as important. A reaction to recommendations would need to be actually observed in order to establish that changes were arising in part as a result of interaction with the UPR process, in order to meet the commonly accepted requirements of the subjective element of custom.

One example of such a reaction is child marriage; there have been a few hundred recommendations issued to states in relation to the issue of early forced marriage, 68 per cent of which have been accepted by states across all three cycles. Early forced marriage is prohibited in CEDAW and in the CRC but there is a tension about both the scope of the prohibition and the age of marriage – CEDAW specifies no minimum age of marriage but the CRC implies eighteen.¹⁰⁴ The HRC has interpreted the provisions in the ICCPR on the right to a family in a way which allows for individual states to reach their own conclusion about marriage laws.¹⁰⁵ Recommendations made to Indonesia in their second review cycle to eliminate early marriage prompted the government to investigate the enforcement of marriage laws and to draft a new law raising the age of marriage to eighteen across the country.¹⁰⁶ At their third cycle review they accepted recommendations on the outright prohibition of forced early marriage.¹⁰⁷ Benin accepted recommendations during its second cycle to abolish early marriage and in its third cycle national report detailed measures it had taken to implement new legislation protecting children’s rights.¹⁰⁸ In its third cycle review during the interactive dialogue states expressed concern about the persistent prevalence of early forced marriage in spite of changes to the law and in response Benin committed to prosecutions of the perpetrators of forced marriage.¹⁰⁹ In both cases the state undertook actions indicative of a belief they were under an obligation to fulfil the substance of the recommendation. These are just two states and many other states have accepted recommendations on this subject, but these examples serve to illustrate how opinio juris can be inferred by looking at a state’s subsequent conduct in the UPR process in relation to the recommendation.

5 Conclusion

By way of conclusion, it is worth identifying two potential lines of criticism about the framework advanced here and what it means for the identification of CIL. Firstly it is open in adopting what Noora Arajärvi critically termed the ‘paradigm shift’ toward the ‘demands of humanity’ away from the more orthodox position in some of the literature, on the evidence of state practice required for custom to be identified.¹¹⁰ Fernando Tesón almost pre-empted the argument advanced in the first part of Section 3 with his description of ‘the Ad Nauseam Fallacy technique’ whereby ‘profusely citing nonbinding resolutions’ is used to advance a ‘sense of normativity’ that is not actually present.¹¹¹ Yet, this neglects the institutional framework of the UPR described above. As both the ILC and the ILA investigations into the source of custom highlight, interaction with an institution such as the UPR is a key part of what a state ‘does’ both in terms of the internal procedure and the effect it has on states and in terms of the commitments that states make. This also relates to the sense of obligation. One key criticism of the way that the ICJ has interpreted the legal effect of GA resolutions is that the court has focused more on their binding nature than on the way that they shape legal discourse.¹¹² Individual UPR recommendations do not bind the states who accept them but collectively a series of accepted recommendations demonstrate the existence of an emerging consensus on a particular norm. This means that a series of recommendations could and, from the perspective of those seeking to defend human rights, probably should, have an authorising effect – in that they identify rights that states need to protect and highlight the legal obligation to protect those rights.¹¹³

Secondly this theory involves adopting a constructivist interpretation of both elements of the two-element theory reasoning that state interests are shaped by the social conditions that surround them (in this case the rules of the UPR process).¹¹⁴ Rationalist and realist critics would probably reject this description of state behaviour, and any analysis of UPR recommendations gravitates heavily towards an empiricist understanding of custom. However, an empiricist understanding of accepted recommendations would allow for the identification of a consensus on a particular norm, and importantly, would provide grounds for establishing a definitive explanation of which rights ought to be protected. The UPR was not meant to be a legal process but its effects have altered the way states act towards certain norms and the recommendations they accept. The mechanism outlined above helps identify where a norm contained in recommendations could have a customary status and addresses some of the criticisms surrounding the identification of customary norms.¹¹⁵ Overall this theory puts a heavy institutional gloss on the identification of CIL, and questions of the practice and the binding nature of norms are answered technically with reference to the nature of the UPR process. In the context of a process which has mass buy-in from states, with every country in the world being subject to at least one review, a justification for a heavy institutional emphasis on the question of custom can be constructed on the lines that Lepard outlines above as it is possible to discern from the UPR process a strong sense of what rights ought to exist.¹¹⁶ Whilst challenging some assumptions about what constitutes practice within the existing literature, identifying custom in UPR recommendations can help give coherence to the identification of customary international human rights norms.