Promoting compliance with human rights : the performance of the UN Universal Periodic Review and treaty bodies

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website. • The final author version and the galley proof are versions of the publication after peer review. • The final published version features the final layout of the paper including the volume, issue and page numbers.


Introduction
Monitoring states' compliance with international human rights obligations is a highly complex task. First, although international organizations strongly rely on states' selfreporting, in several cases of human rights breaches the executive branches of states were the prime violators of the norm; why, then, would they provide information to the international organizations, accepting the ensuing conclusions? Second, human rights violations do not normally have an impact on other states, in contrast to other policy areas such as pollution or trade. For these reasons, the rational incentives for states to establish forceful human rights mechanisms or to pressure each other into compliance are lower than in other policy domains (Underdal 1998;Risse, Ropp, and Sikkink 1999;Dai 2002;Hathaway 2002;Hawkins 2004;Simmons, 2009Simmons, , 2010. Finally, it is doubtful whether a common normative framework allowing for a universal interpretation of human rights across regions and cultures exists at the global level (Samhat 1999;Donnelly 2007;Heywood 2011;Jun 2015).
Relatedly, there is a wide gap between states' inclination to ratify international human rights treaties, and their actual compliance with those treaties (Avdeyeva 2007). Scholars thus wonder under what conditions states conform to international human rights standards, and how they can be led to modify their policies in case of noncompliance. As further elaborated in this article, the three main theoretical schools on compliance-enforcement, managerial, and constructivist-provide different answers to these questions.
A multitude of instruments to monitor state adherence to international human rights obligations exists. A prominent place is occupied by the United Nations' (UN) treaty bodies and Universal Periodic Review (UPR), two global mecha-Valentina Carraro is a Postdoctoral Researcher in the Faculty of Arts and Social Sciences at Maastricht University, The Netherlands, and a Rubicon Research Fellow at the Ludwig Boltzmann Institute of Human Rights in Vienna.
Author's note: Research for this article received funding by the Dutch Research Council (NWO -grant numbers 452-11-016 and 446-17-006). nisms for assessing states' human rights performance. Treaty bodies are committees of independent experts monitoring states' implementation of the ten major UN human rights treaties. One of their major tasks is to undertake the state reporting procedure, during which state parties are evaluated on the implementation of their treaty obligations and receive recommendations for improvement. In contrast, the UPR is a peer review where the assessment of states' human rights performance is carried out by other states. As in the treaty bodies, reviewed states receive a list of recommendations for improvement.
What remains unclear, however, is the extent to which these mechanisms are effective in improving state compliance with human rights, and why. While debates on reasons for state compliance with international obligations abound in the literature, assessing the performance of a public organization is challenging, as it requires isolating the role played by the organization in stimulating compliance from a variety of intervening factors, such as other bodies promoting the same standards. This article follows the steps laid down by Gutner and Thompson (2010) by acknowledging that assessing the performance of an international organization by only looking at its ability to reach its ultimate goal (e.g., improving human rights on the ground) might prove misleading. An organization might be successful in producing all necessary conditions to stimulate compliance, and yet states might still be breaching international obligations due to a variety of other reasons. Hence, the performance of organizations is best assessed by studying what Gutner and Thompson (2010) term "process-based performance," namely, the ability of an organization to reach smaller-scale objectives, which might be helpful towards the achievement of the overall goals. This article takes this framework further, by assessing the extent to which these smaller-scale objectives are seen as instrumental in improving states' human rights compliance. It thus proposes and applies a model to assess the extent to which human rights governance mechanisms-in this case the UPR and the state reporting procedure of the treaty

The Treaty Bodies and the Universal Periodic Review
The UN treaty bodies were established to monitor and supervise states' implementation of the ten core UN treaties on human rights. They are composed of committees of independent experts, elected by the state parties to the treaty. Treaty bodies may perform a variety of tasks: receiving periodic reports by states on the implementation of the treaty; deciding on individual complaints for purported treaty violations by a state party; conducting country inquiries if they suspect that the state party has severely violated the treaty; and providing their interpretation of treaty provisions in documents called General Comments. This article focuses exclusively on the state reporting procedure, arguably the main function carried out by committees (Melish 2009;O'Flaherty and Tsai 2011;Flinterman 2015) and, as discussed below, whose procedural functioning is highly similar to the UPR. 1 The UPR was created in 2007, as part of a reform of the UN human rights machinery that culminated in the establishment of the Human Rights Council, of which the UPR is a part. The UPR is a peer review-hence, states assess each other's human rights performances, alternating in their roles as reviewers and reviewees. States are reviewed every four years, on the basis of all the human rights obligations they have undertaken. 2 The UPR and the state reporting procedure of the treaty bodies are highly comparable in their functioning. Both are based on information provided by states in a self-assessment report, complemented by the UN secretariat with additional information from UN and non-UN sources. On this basis, reviewers issue a set of recommendations for improvement. In the treaty bodies this document is called Concluding Observations, while in the UPR, recommendations are included in the outcome report. A key difference between these two mechanisms concerns the nature of the reviewing body: in the UPR, reviewers are states, whereas in the treaty bodies, they are independent experts. In addition, while treaty body recommendations are adopted by consensus, UPR recommendations are exclusively attributed to the country issuing them.
Considering its recent establishment, a notable amount of literature on the UPR is currently available. The large majority of existing scholarship focuses on its functioning, often by studying state behavior within the mechanism, and on its potential to add value to the global human rights landscape (Abraham 2007;Gaer 2007;Domínguez-Redondo 2008;Lilliebjerg 2008;Rathgeber 2008;Abebe 2009;Freedman 2011;McMahon and Ascherio 2012;Cowan and Billaud 2015). Some authors discuss the establishment of the UPR in a highly detailed manner (Alston 2006;Abraham 2007;Gaer 2007;Freedman 2011Freedman , 2013McMahon and Ascherio 2012), while others focus on its functioning, for example by discussing the participation of certain (groups of) states, such as African states (Abebe 2009;Smith 2014) or China (Smith 2011), or by highlighting the existence of namingand-shaming (Terman and Voeten 2018). While scholars are divided about the potential of the UPR to generate compliance (for an overview of this debate, see Etone 2019), empirical work assessing the extent to which UPR recommendations are implemented is scarce. Most of existing analyses have been carried out by the Geneva-based nongovernmental organization (NGO) UPR Info (UPR Info 2014 or consist of single case studies by practitioners (e.g., Roesdahl 2017). Although UPR Info is positive concerning states' implementation of recommendations, its studies do not take into account possible intervening factors, as also pointed out by Elizalde (2019).
Academic accounts on the treaty bodies abound (Nowak 1993;McGoldrick 1994;Alston and Crawford 2000;Bayefsky 2001;Nowak and McArthur 2008;Bassiouni and Schabas 2011;Keller and Ulfstein 2012b;Alston and Goodman 2013;Creamer and Simmons 2015). Works assessing their impact tend to focus on a selected number of cases, and generally show that recommendations are implemented only in some cases, and with high variation across countries (see Heyns and Viljoen 2001;McQuigg 2011;Krommendijk 2014aKrommendijk , 2014bKrommendijk , 2015. When scholars find a link between treaty bodies and policy change, they generally argue that this impact derives from a favorable domestic context: for example, treaty bodies are seen as instrumental in providing opportunities for domestic actors to push for their preferred policy outcomes (Krommendijk 2015;Creamer and Simmons 2019).
Few comparisons of the UPR and treaty bodies have been undertaken, so far. Exceptions are work by Gaer (2007) and Rodley (2012), who compare the two mechanisms and discuss their potential for complementarity or duplication, and by Carraro (2017), who studies the effects of politicization on their ability to advance human rights. This article aims to broaden the debate on the UPR and treaty bodies by explaining, in a comparative manner, their perceived ability to stimulate change.

Compliance with International Provisions
The reasons why states comply with international obligations have long been debated (Franck 1990;Chayes and Chayes 1995;Koh 1997;Downs 1998;Simmons 1998Simmons , 2010Underdal 1998;Checkel 2001Checkel , 2005Checkel and Moravcsik 2001;Johnston 2001;Dai 2002Dai , 2005Dai , 2006Dai , 2013. Some scholars argue that enforcement is essential for compliance. Downs (1998) defines enforcement as, "The overall strategy that a State or a multilateral adopts to establish expectations in the minds of state leaders and bureaucrats about the nature of the negative consequences that will follow noncompliance" (Downs 1998, 321-2).
Deterrents for noncompliance could, for example, be political pressure, withdrawal of positive incentives, or economic sanctions. Compliance is seen as the result of cost-benefit calculations, influenced by incentives for good performance and tools to punish noncompliance. In a typical prisoner's dilemma situation, benefits for a state are conceptualized as what that state can obtain by withdrawing its cooperation with, for example, a multilateral agreement, while other states continue to comply with it. If these expected gains are high, the threatened punishment must be sufficiently severe to act as a deterrent for the state to defect from the agreement (Downs 1998). Actors are therefore assumed to act strategically, with the goal of maximizing their gains and minimizing their losses (Downs 1998;Checkel 2001;Dai 2005), following what March and Olsen (1998) call the logic of anticipated consequences and prior preferences.
Existing critiques of enforcement models center around two main concerns. First, some scholars claim that actors cannot foresee all the consequences of their actions (March and Olsen 1998). Second, it is generally claimed that enforcement models tend to ignore the processes leading to interest formation: preferences are assumed to be exogenously formed, and there is no attention to the way such preferences evolve based on the context in which actors operate (March and Olsen 1998;Checkel 2001). Stemming from these considerations, constructivist scholars claim that to explain compliance one has to focus on the processes that lead to the formation of preferences and the construction of identities. These processes are endogenous to the organizations in which actors operate and take the form of social learning, norm diffusion, and socialization (Franck 1990;Checkel 2001Checkel , 2005Checkel and Moravcsik 2001;Johnston 2001). When explaining compliance from a constructivist standpoint, Checkel (2001) mentions that two mechanisms are at play. On the one hand, some argue that the roots of compliance with international obligations are to be found in social mobilization: according to this view, compliance is not exempt from cost-benefit calculations, yet these calculations are shaped by a process of norm diffusion that drives societal forces to put pressure on decision makers to conform to these rules. On the other hand, some scholars argue that compliance is the result of social learning and norm in-ternalization (Checkel 2001). This follows what March and Olsen (1998) term the logic of appropriateness. The model is centered on the crucial role that institutions and norms play in the construction of identities, which in turn drive human behavior. Actors are believed to make their choices not by following cost-benefit calculations, but on the basis of what they perceive to be appropriate behavior. Rules are therefore followed only to the extent that they are perceived to be legitimate and conforming to actors' identities (March and Olsen 1998). Persuasion and social learning are thus conducive to compliance by leading to a convergence of states' preferences and behavior (Franck 1990;Checkel 2001Checkel , 2005Checkel and Moravcsik 2001;Johnston 2001).
A third strand of compliance literature is that of the management school (Chayes and Chayes 1995;Koh 1997). In this view, states are generally committed to complying with international treaties because they dedicate time and resources in negotiating them. Chayes and Chayes (1995, 4) argue that compliance is the "normal organizational presumption." The fact that states signed an agreement means that they believed it was in line with their interests, and that they acknowledge the legal obligation to comply with it. Starting from these assumptions, noncompliance will have to be explained by factors such as treaty ambiguity or limitations in country capabilities. Successful strategies to ensure compliance will therefore include disseminating information on regime requirements, providing practical support to states, and, ultimately, persuading noncomplying actors to alter their course of action (Chayes and Chayes 1995).
While international cooperation in all fields of international law presents its challenges, human rights are a particularly complex case, as discussed above. Despite the increasing prevalence of international human rights treaties posing legal obligations on states, human rights enforcement mechanisms are still lacking at the global level. Whereas regional human rights judicial mechanisms exist-most notably, the European Court of Human Rights, the Inter-American Court of Human Rights, and the African Court on Human and Peoples' Rights-no such parallel court can be found on a universal scale. The only global court dealing with human rights violations is the International Criminal Court (ICC), but this only deals with gross human rights violations such as genocide and war crimes.
The international system-including, although not limited to, the case of human rights-thus heavily relies on voluntary compliance. Its fundamental premise is that states will deal with the enforcement of international human rights provisions at the domestic level (Donoho 2006). Even though international institutions are often able to induce states to comply-for example by exerting pressure-they lack enforcement abilities, as they are unable to "compel direct consequences … under the threat of meaningful sanction" (Donoho 2006). As a consequence, no global legally binding enforcement mechanism exists that can hold states accountable for violating their international human rights obligations (Forsythe 2009).
It is in this context that the existence of the UPR and treaty bodies should be understood. Whereas most states are reluctant to commit to coercive international oversight systems, their legally nonbinding nature 3 is arguably the reason why they exist in the first place. Yet, lack of coercion does not necessarily mean that these mechanisms are irrelevant. While their nonbinding nature implies that hard enforcement and coercion do not apply, if we follow the enforcement model, compliance might still be triggered by outside pressure and naming-and-shaming, or by fear of losing material advantages such as donor aid (Downs 1998). In a more constructivist vein, the reviews might also lead states to progressively internalize those norms, adopting them as the "right thing to do" (Checkel 2001;Finnemore and Sikkink 2001). Countries, finally, might be willing to comply, yet lack the resources to do so (Chayes and Chayes 1995).
The next subsection introduces the concept of processbased performance to provide a framework to assess the ability of the UPR and treaty bodies to achieve a set of goals, linked to the different compliance schools. This is done with the aim of testing the extent to which reaching those goals is instrumental in promoting states' implementation of the human rights recommendations received. Gutner and Thompson (2010, 236) stress the importance of distinguishing between an organization's process performance and its outcome performance. Whereas the outcome performance of an organization is identified by its ability to solve a specific problem-in our case, to improve states' human rights record-its process performance relates to its ability to reach what scholars call micro outcomes, namely, its ability to carry out specific tasks.

Process-Based Performance
Treaty bodies and the UPR work towards the achievement of their ultimate goal by delivering recommendations, whose implementation by states would arguably lead to an improvement of their human rights performance (outcome performance). To reach the ultimate goal of policy change, other more concrete and measurable results might be achieved by these mechanisms (process-based performance). Yet, what would these micro results be, concretely? And to what extent does the achievement of these results contribute to the outcome performance of these mechanisms? A list of these possibly meaningful outcomes was developed both inductively and by employing secondary literature on soft governance mechanisms. Each of these goals is linked here to the theoretical debate on compliance.
The first result considered in this study is the exertion of pressure (Pagani 2002;Pagani and Wellen 2008;Carraro and Jongen 2018;Jongen 2018). Following the enforcement model, a possible meaningful result of nonbinding mechanisms is the generation of pressure on states (Downs 1998), which might aid compliance for fear of material or reputational losses. Pressure can be exerted in several ways, ranging from pressure on states to submit accurate and timely reports, to pressure for the follow-up of recommendations. In addition, pressure can be exerted by peers (particularly in the case of peer reviews), as well as by the broader public.
Second, the article turns to these mechanisms' ability to trigger learning (Pagani 2002;Lehtonen 2005;Tanaka 2008;Smith 2011;Kälin 2012;Cowan and Billaud 2015;Carraro and Jongen 2018;Jongen 2018), in a more constructivist vein. Indeed, while reviews can be seen as sanctioning fora aimed at exposing countries to criticism (as in the enforcement model), they can also be viewed as nonconfrontational mechanisms where the focus is on stimulating learning, socializing states to the "right" approach in dealing with human rights norms.
Third, this study focuses on the potential ability of reviews to provide an accurate overview of states' human rights per-  (Pagani 2002;Lehtonen 2005;Rathgeber 2008;Carraro and Jongen 2018;Jongen 2018). To achieve this goal, the review output should adequately reflect the internal human rights situation in the reviewed state, highlighting all relevant human rights issues and suggesting areas for improvement. From both a managerial and constructivist perspective, it is essential for reviewed states to be aware of what areas need attention, in order to devise the appropriate strategy to improve. The fourth potential result relates to the reviews' ability to deliver practically feasible recommendations (Ikhsan 2008;Carraro and Jongen 2018;Jongen 2018). What is of interest here is an assessment of whether recommendations delivered in these mechanisms are perceived to be feasible, and an understanding of what characteristics a recommendation should have in order to be considered as such. From a managerial perspective, rule ambiguity is one of the most likely culprits of poor rule implementation. If these mechanisms are unable to provide clear guidelines for implementation, how can states take the correct actions? Table 1 summarizes the possible outcomes identified above, and the respective school of compliance with which they more closely align.
As also stressed, among others, by Versluis and Tarr (2013), although compliance schools depart from distinct theoretical viewpoints, the dynamics they highlight might very well be observed to work in combination, empirically. In this article, these categories are therefore employed as heuristic devices to study the extent to which each of these logics is at play, but they are not intended as mutually exclusive categories.
Following the "Data Collection and Methodology" section, this article first assesses the extent to which actors involved in the UPR and treaty bodies perceive that the reviews are able to achieve these outcomes. Second, it turns to their perceived capacity to reach the ultimate goal of improving human rights on the ground, and asks to what extent this can be considered a result of their (in)ability to achieve the outcomes outlined above.

Data Collection and Methodology
This study focuses on the views of actors who are directly involved in the two mechanisms under study as either representatives of states under review, or as reviewers. In the UPR, these are all the involved diplomats-acting as reviewees and reviewers in turn-whereas in the treaty bodies these are the diplomats as reviewees, and the independent experts composing the reviewing committee.
As also acknowledged by Gutner and Thompson, evaluating performance might wield different results depending on who is conducting the analysis. Performance assessments suffer from the "eye of the beholder problem" (Gutner and Thompson 2010, 233), as different stakeholders (e.g., member states, the wider public) are likely to hold different views. In this light, this article argues that focusing on the perceptions of directly involved actors is preferable from both a methodological and a theoretical point of view. Methodologically, involved actors are the most suitable to provide an accurate assessment of these mechanisms' ability to achieve certain outcomes. Indeed, most of the outcomes measured in this article cannot be subject to an external empirical measurement: for example, it would be impossible to objectively establish, by means of some external yardstick, whether the UPR is able to trigger learning-much more logical is to ask directly involved actors whether they experience that the UPR generates learning. Similarly, objectively assessing the ability of these instruments to generate compliance would be extremely challenging, as it would require isolating the role played by these instruments from a variety of other intervening factors (Heyns and Viljoen 2001;Krommendijk 2015). For example, other international or national actors may have been advocating the same goals, making it difficult to estimate the extent to which each player caused the observed outcome. In addition, as observed in the earlier section "The Treaty Bodies and the Universal Periodic Review", the few existing scholarly works studying the implementation of UPR and treaty body recommendations do so in a highly selected number of cases. Such an assessment would become even more difficult in a comparative study on the UPR and treaty bodies-two mechanisms with largely overlapping mandates and goalsas is the present one. Finally, and from a more theoretical perspective, this article takes the position that these mechanisms can only be effective if participating actors believe in their relevance. Since recommendations cannot be legally enforced, for policy change to occur, participants must take the process seriously, and must believe that these mechanisms can make a significant contribution (Donoho 2006;Creamer and Simmons 2015). Conversely, if involved actors were to find these mechanisms meaningless, they would be unlikely to invest the necessary time and resources required for participating in the reviews, and for implementing the recommendations. Thus, it is essential that both reviewees and reviewers regard these instruments as able to achieve concrete results (also see Jongen 2018). In a similar line of reasoning, scholars studying the effectiveness of international organizations have argued for the importance of factors such as socialization, reputation, and ideas as key drivers of policy change (see for example Schmidt and Radaelli 2004;Greenhill 2010; Hafner-Burton and Schneider 2019), moving beyond an exclusive focus on domestic implementation to measure the effectiveness of an international organization (Schmidt and Radaelli 2004). While this article does not claim that perceptions necessarily coincide with actual levels of compliance, nor does it argue against the importance of studies tracing the implementation of recommendations by international bodies, it does argue, for the reasons listed above, that measuring perceptions is a more methodologically viable alternative in the present case (also see Tompkins and Amundsen 2008), and one that provides important insights on factors facilitating compliance with recommendations by human rights bodies.
Such an approach focusing on the views of directly involved actors comes with limitations. A larger-scale study focusing on the views of additional actors-for instance, governmental officials or NGOs-would provide a more encompassing overview of perceptions on these instruments. Non-directly involved actors could either be more skeptical or, conversely, they might more strongly believe in their appropriateness, as they are not directly exposed to their weaknesses. However, due to time and scope constraints, this article exclusively focuses on directly involved officials, for two reasons. First, these officials know the mechanisms more accurately, and are therefore best suited to provide an assessment of their performance. Second, as the goal of the article is to compare the UPR and treaty bodies, possible biases towards the Geneva-based mechanisms should not influence the respondents' ability to assess their strengths and weaknesses comparatively.
The survey and interviews first assessed the extent to which respondents believe that the two mechanisms are able to achieve the goals identified earlier. Whereas survey respondents were provided with a list of possible outcomes (see Table 2), interviewees were asked to elaborate on the results that they would see as most valuable. All interviewees' answers fit in the four pre-identified categories. Methodological considerations in terms of the survey and interviews can be found in Annexes 1 and 2 respectively in the supplementary data archive.
In a second step, the extent to which these mechanisms stimulate compliance was measured exclusively by means of interviews, due to the complex nature of compliance. For the reasons discussed earlier in this section, it is very difficult-for researchers and respondents alike-to assess the extent to which each of the two mechanisms triggers policy change, as this depends on a variety of factors unrelated to the reviews. Consequently, interviews were chosen as the only data collection method in this regard, as they allowed officials to explain and qualify their answers, and to link these results to the goals discussed above. Interviewees generally provided nuanced answers, explaining the extent to which, in different situations, they believe that the mechanisms are able to achieve such a goal.

Assessing the Performance of the UPR and of the State Reporting Procedure of the Treaty Bodies
The Ability to Exert Pressure Survey results show a marked difference between the two mechanisms in relation to their ability to generate pressure. As Figure 1 illustrates, the majority of respondents believe that the UPR exerts peer pressure to a large extent or completely. Concerning the ability of the UPR to generate public pressure, results are similar. A majority of respondents-albeit slightly fewer than in the case of peer pressure-believe that the UPR is able to generate public pressure to a large extent or completely. In contrast, as Figure 2 illustrates, a majority of respondents believe that the treaty bodies are only to some extent able to generate public pressure, or even not at all, while over a third believe that this is to a large extent or completely the case.
Interviews confirmed that most respondents believe the UPR is able to generate peer and public pressure on states (interviews 1-3, 10-13, 22, 24, 26, and 40). Concerning public pressure, one of the major strengths of the UPR is identified as the fact that NGOs are an essential part of the process: even though they are not formally part of the review, they play an active role both in the information-collection and the implementation phases, exerting pressure on governments to live up to their commitments. Even though treaty bodies, on paper, are equally inclusive of civil society, respondents believe that it is easier for NGOs to exert pressure in the UPR, due to the fact that all UPR-related information is easily accessible, including what specific recommendations were accepted by states. This in turn makes Similarly, the UPR is considered very successful in generating peer pressure. Interviewees mentioned that the bilateral nature of the UPR-where recommendations are delivered by one governmental representative to another-increases pressure on countries to follow up on recommendations (interviews 1, 2, 10-13, 22, 24, 26, and 40). Finally, in the case of both peer and public pressure, interviewees mentioned that the periodicity of the UPR adds a further layer of pressure, as states must report back every four years. Reportedly, government officials feel the need to show their domestic and international audiences, as well as their colleague diplomats in the room, that they have already acted upon many of the recommendations received in the previous review (interviews 1, 2, 4, 10-13, and 40).
Interviews confirmed that treaty bodies are able to generate some pressure on states for compliance, although to a limited extent (interviews 12, 25, 28, and 34-36). Reportedly, it does happen that Concluding Observations are taken up by the media and the general public, but this is perceived to occur to a much more limited degree than in the UPR. One interviewee for instance mentioned that a successful example of exertion of public pressure relates to the "Chicago police case," where a former Chicago police chief was being accused of systematically employing torture as an interrogation technique. The matter was discussed during the examination of the United States' report by the Committee Against Torture in 2006, and, according to the interviewee, "because that case was mentioned in the Concluding Ob-servations…the pressure was so strong that he actually was investigated and then convicted" (interview 25).

The Ability to Trigger Learning
Survey results show that the UPR is only partially able to trigger learning. As Figure 3 illustrates, a majority of respondents believe that the UPR successfully triggers (mutual) learning only to some extent or not at all, although over 40 percent of respondents find this to occur to a large extent or completely. Only slightly more positive are the results for the treaty bodies, as opinions are strongly divided: the majority of respondents believe that learning occurs to a large extent or completely, although a substantial proportion of respondents believes that it only occurs to some extent or not at all.
Interestingly, no interviewee mentioned learning as one of the successful outcomes of the UPR. Even though interviewees refer to the UPR as a constructive exercise (interviews 1, 2, 3, 8, and 12), learning and the exchange of best practices were not mentioned as playing any substantial role. Finally, interviewees mentioned that the UPR is equally unable to trigger external assistance in the implementation of recommendations (interviews 10 and 11), which could, in turn, lead to indirect learning.
In a somewhat more positive light, interviewees reported that treaty bodies are relatively successful in triggering learn-ing, albeit often in an indirect way (interviews 12, 27, 34, and 39). Specifically, it was argued that treaty bodies can bring a valuable contribution when it comes to stimulating domestic dialogue and changing the way audiences talk about certain issues (interviews 27 and 34): People underestimate the role that the treaty bodies have in changing … the semantics of discussions, simple things like talking about persons with a disability as exactly that, persons with a disability rather than what was almost universally accepted terminology just a decade ago … where the reference would be to a disabled or handicapped person. (Interview 27)

The Ability to Provide an Accurate Overview of Reviewed States' Performance
Survey results show mixed views on the capacity of the UPR to provide an accurate overview of states' performance. Figure 4 shows that half of respondents believe that the UPR achieves this result to a large extent or completely; however, the other half believe this to be the case only to some extent, or not at all. Treaty bodies are remarkably more successful: a substantial majority of respondents believe that they achieve this goal to a large extent or even completely, about a quarter believe that this occurs only to some extent, and no one finds this is not the case. Interviewees expressed criticism concerning the ability of the UPR to provide an accurate overview of states' performance. While most recommendations are judged to be objective and well informed, others are reportedly politically motivated or stemming from unfamiliarity with reviewed countries (interviews 1-4, 6, 8, and 12). As a consequence, by looking at the list of recommendations, it is not always possible to identify the level of human rights violations in a country, nor what the serious areas of concern are. As an interviewee put it, "you will have the same country praised and criticized" (interview 8). Interestingly, participants' perceptions are partially in contradiction with findings by Rathgeber (2008), who-on the basis of his own observations-argued that the output of the mechanism successfully depicts states' internal human rights situations.
The situation appears markedly more positive in the treaty bodies, as recommendations are considered to be much more objective and of a higher quality than those in the UPR (interviews 4, 6, 16, 25, and 30). An interviewee provided an interesting example on the subject: You take [the UPR outcome reports of] Finland and China. You remove the names…and you give it to people and you say … "tell me which country do you think is more respectful of human rights than the other," and then you see the answer [will not correspond to expectations]. But if you do the same exercise with Concluding Observations, there you can see the one [which is most respectful of human rights]. (Interview 25) These results highlight the main perceived weakness of the UPR when compared with the treaty bodies. As previous studies have argued (Freedman 2011;Carraro 2017;Terman and Voeten 2018), political relations play a crucial role in determining the content and harshness of the recommendations received by states. This, in turn, reportedly damages the credibility of the UPR, as it shows that some countries receive a much more lenient treatment than their human rights record would warrant. Whereas these dynamics are also partially present in the treaty bodies, in the UPR they take place to a much larger extent (Carraro 2017).

The Ability to Deliver Practically Feasible Recommendations
As Figure 5 illustrates, survey results are relatively positive with regard to both mechanisms' ability to deliver practically feasible recommendations. A majority of respondents believe that the UPR is to a large extent or completely able to do so, although almost half of respondents believe this to be the case only to some extent. Treaty bodies score even more positively: a clear majority of respondents believe that practically feasible recommendations are provided to a large extent or completely, while about a third of respondents believe this to be the case only to some extent or not at all. 4 Interviewees were first asked to discuss what constitutes a practically feasible recommendation, in their opinion (interviews 6, 7, 10-13, 16, 23, 29, 34, and 40). Their views can be summarized into three main points. First, recommendations should be specific and measurable, outlining the steps that countries should take toward implementation. For example, a recommendation that asks the state party "to do more to reduce poverty levels" can be made implementable if it is accompanied by "the state party should within two years have commissioned and finalized a report looking at the obstacles to poverty reduction" (interview 29). Second, recommendations should take into account a clear timeline: what can realistically be achieved by the next review? Third, recommendations should show awareness of countries' internal situations, and of what progress they can be 4 Responses varied between committee members and diplomats, with committee members holding more positive views. Treaty bodies are deemed to deliver practically feasible recommendations completely by 16.3 percent of committee members and 10 percent of diplomats; to a large extent by 60.5 percent of committee members and 43.3 percent of diplomats; to some extent by 20.9 percent of committee members and 43.3 percent of diplomats; and not at all by 0 percent of committee members and 3.3 percent of diplomats. reasonably expected to make. In sum, interviewees believe that practically feasible recommendations are: (1) specific and measurable; (2) achievable within a specific timeframe; and (3) able to take into account the national context of reviewed states.
As concerns the ability of the mechanisms to deliver practically feasible recommendations, the majority of UPR recommendations are considered to be rather general. Interviewees often mentioned that this is one of the characterizing features of the UPR, where soft, broad recommendations represent the large majority as compared with specific ones (interviews 1, 4, 5, 8, 10-13, 24, and 38-40). When analyzing UPR recommendations, this indeed seems to be the case. The NGO UPR Info publishes a database containing a record of all recommendations issued in the UPR and divides them into five different action categories: (1) minimal action; (2) continuing action; (3) considering action; (4) general action; and (5) specific action. Recommendations in categories 1, 2, and 3 are extremely minor and encouraging; category 4 recommendations are very general; whereas category 5 recommendations are very specific. A total of 46,584 recommendations were issued from the first UPR session up to May 2016. 5 Of these, 40.86 percent requested minimal or general action, 25.02 percent were related to continuing or considering action, whereas 34.12 percent requested specific action. These data therefore empirically confirm that nonspecific recommendations represent the large majority in the UPR. This falls in line with findings presented in the subsection "The 5 The month when data collection was concluded. The database is accessible at: https://www.upr-info.org/database/ Ability to Provide an Accurate Overview of Reviewed States' Performance," which showed that the UPR is not deemed successful in providing an accurate overview of country performances. The tendency of UPR states to deliver general recommendations points in the same direction, as overall states seem reluctant to criticize each other too harshly, to avoid jeopardizing their diplomatic relations (interviews 1-5, 8, 10-13, 24, 25, 39, and 40).
Some interviewees, however, acknowledged that, on several occasions, UPR recommendations-although general-are at least realistic. This is because recommendations take into account the starting situation of reviewed state, considering what is feasible for the country to achieve both within the timeframe of the upcoming review, and with regard to its internal capacities (interviews 3, 6, 7, 26, and 40). Thus, while UPR recommendations do not fare well on the first point, namely specificity, they perform much better when it comes to the second and third points, as they take into account what can be realistically achieved by the state in question, in the timeframe of the upcoming review.
Interviews show the opposite picture for the treaty bodies: Concluding Observations are generally praised for their high specificity, and criticized for being unrealistic (interviews 2, 6, 10-12, 16, 25-28, 30, 38, and 39). Indeed, by looking at Concluding Observations, one can clearly observe their very specific and detailed nature. 6 Respondents consider this to be highly positive, as recommendations are reportedly successful in guiding states toward implementation (interviews 6, 16, 25, 26, 28, 38, and 39). Yet, they are often deemed by interviewees to aim at excessively ambitious goals, which are not realistic when taking into account country capabilities. Respondents criticized treaty bodies for aiming at an unachievable ideal of perfection, which creates frustration in state representatives (interviews 2, 6, 10-12, 27, and 30).

The Ability to Generate Compliance with Human Rights
The section above showed that both treaty bodies and the UPR have strengths and weaknesses when it comes to their ability to achieve certain outcomes. The major perceived strength of the UPR lies in its capacity to trigger pressure, on which the treaty bodies fare rather poorly. Conversely, treaty bodies are mostly appreciated for providing accurate overviews of reviewed states' performances and learning opportunities, on which the UPR scores lower. Finally, both appear quite successful in providing states with practically feasible recommendations. The current section qualitatively investigates the degree to which the UPR and treaty bodies are deemed able to stimulate compliance with recommendations. It further assesses the extent to which reaching the outcomes identified above is instrumental for compliance.
First, interviewees commented that, to improve human rights domestically, these mechanisms should guide states toward implementation, providing clear guidelines. This goal is to be achieved via the formulation of clear and practically feasible recommendations (interviews 6, 7, 10-13, 16, 23, 29, 34, and 40). In this sense, the results presented above show that both the UPR and treaty bodies fare relatively well, with the treaty bodies being slightly more successful.
However, unsurprisingly, interviewees stressed that policy change is ultimately dependent on states' willingness to improve (interviews 2, 4, 5, 7, 9-11, 16, and 39). If such willingness is not present, no mechanism will be able to produce results. In this light, the UPR is seen as a helpful tool to provide an extra incentive for states to improve their standards, due to its ability to trigger pressure. Improvement must, however, be seen as a process. In some cases, this could mean bringing a formerly closed-off state to at least publicly discuss their human rights situation: The DPRK [Democratic People's Republic of Korea] is a positive example…In the first cycle the DPRK didn't want to accept any of the recommendations, even they didn't want to cooperate … Coming this year, the second cycle from DPRK in the UPR, they have accepted [over one hundred recommendations]. This is huge progress. (Interview 9) In other instances, the UPR reportedly produces tangible results (interviews 2, 4, 5, 7, and 39). Interviewees mentioned that, when revising the implementation of previous recommendations, they observed concrete improvement in many member states (interviews 2 and 39). The UPR is generally seen as the "extra push,…the political move that makes it happen" (interview 39). Even though it is admittedly impossible to understand whether the cause of such policy developments is really the UPR or any other body, some respondents believe that it is likely the combination of all these mechanisms that triggers change (interviews 2 and 39): It is more about adding the recommendations, if you have one single recommendation on one topic you can say it is just an erroneous assessment but when everything starts to add [up], meaning a committee recommended, UPR recommended, a special rapporteur recommended it, your own civil society has identified that as a problem,…the sum of all these different elements will be able to have an impact on the ground. (Interview 2) In the absence of political pressure, the improvement of human rights domestic situations is, in the treaty bodies, even more dependent on states' will. If states are not already inclined toward improving their performance, there is nothing that committees can do (interviews 5, 7, 16, 25, 27, 29, 35, 36, and 38). When states are committed to improving, however, treaty bodies provide guidance and specific advice on the steps to be taken. Many interviewees recalled instances in which it could reasonably be assumed that policy change happened as a direct consequence of treaty body recommendations, or as a consequence of treaty body recommendations adding up to advice received in the context of other human rights mechanisms (interviews 2, 8, 16, 25, 27, 29, 34-36, and 38). One respondent mentioned the example of Angola, where a law combating violence against women was adopted shortly after a recommendation by the Human Rights Committee on the subject, and where several local stakeholders reportedly confirmed that there was a direct causal effect between the recommendation and the adoption of that law (interview 38). Treaty bodies might also be considered to have an impact on improving the human rights situation by stimulating dialogue between different actors within a country (interviews 27 and 34).
Moreover, albeit their ability to generate pressure is limited, treaty bodies are reportedly successful in strengthening certain claims made by civil society or other actors (interviews 25 and 34-36): "Sometimes what treaty bodies [recommend is accomplished]…because it then became available as a legitimate credible objective recommendation that was picked up by someone with power" (interview 36). These findings fall in line with those by Krommendijk (2014bKrommendijk ( , 2015 and Creamer andSimmons (2018, 2019), who stress that Concluding Observations are most effective when domestic actors push governments to act upon them.
Interestingly, pressure in the UPR does not seem to become a confrontational form of "naming-and-shaming": while, undoubtedly, the UPR does "name" and "shame" transgressors, this is done within the framework of a highly constructive mechanism, where states often avoid being excessively harsh with each other (interviews 1, 4, 5, 7, 8, 10-13, 23-25, 39, and 40). In this sense, treaty bodiescomposed of nongovernmental experts-may be more willing to engage in "naming-and-shaming" activities (also see Creamer and Simmons 2013; Kahn-Nisser 2018). However, their limited political power-as discussed earlier-makes their naming-and-shaming less visible and politically compelling for states.

Discussion of Results in Light of Existing Compliance Theories
What can we learn from this, in the context of the debate on state compliance with international law? From a managerial perspective, states are naturally inclined toward compliance, and defection will have to be explained by factors other than states' lack of willingness. Hence, compliance is facilitated by providing states with the tools to address the problems they are facing (Chayes and Chayes 1995). This article shows that the provision of practically feasible recommendations is considered by interviewees as an important feature, and that recommendations are considered at their most helpful when they are realistic, specific, and measurable. In this sense, whereas both treaty bodies and the UPR present different strengths and weaknesses, treaty bodies appear to be better equipped in providing guidelines to states on how to proceed with implementation. Additionally, they are more successful than the UPR in providing an accurate overview of states' performance, which is instrumental in helping states understand how to tackle implementation. From a managerial perspective, therefore, treaty bodies are better suited at providing states with the instruments they need to bring their performance up to standard.
Moreover, treaty bodies appear to be better able to stimulate learning, helping states to understand where their problems lie and how to tackle them. From a constructivist viewpoint, providing states with learning opportunities is crucial to ensure compliance: states become socialized with the best way to approach certain problems, and eventually learn the rules of the game. This, in turn, will lead states to respecting human rights provisions as the "right thing to do," rather than for fear of repercussions (Checkel 2001;Finnemore and Sikkink 2001). Thus, also from a constructivist point of view, treaty bodies are better able to guide states toward compliance. As Creamer and Simmons (2018) argue, the main impact of treaty bodies is not to be found as a direct consequence of a given reporting cycle: "It was never intended or designed to affect rights practices through a single report submission" (Creamer and Simmons 2018, 47). Rather, the authors continue, it is the constant process of engaging with the reporting procedure that brings forward change (Creamer andSimmons 2018, 2019). Similarly, Heyns and Viljoen (2001) stress that the impact of the treaties derives from shaping national understandings of what human rights are, rather than from the specific activities undertaken by treaty bodies.
Yet, this article also showed that, while treaty bodies offer states the best tools for improvement, this only works when states are sincerely willing to change. Helping states to comply by means of learning and providing clear guidelines is not sufficient when political willingness is lacking. Findings by Heyns and Viljoen (2001) support this argument, as the scholars conclude that states that are sincerely committed to participating in the treaty bodies will benefit from reporting; in contrast, the system seems to have no impact on states that do not engage meaningfully with it. When states need an extra push toward implementation, the UPR is better at applying political pressure on reluctant states, motivating them to comply even if they would otherwise not have done so. In this light, the enforcement logic seems to be better able to explain compliance in the case of states that are not inclined towards implementation in the first place.
As discussed earlier, overall, states are reluctant to agree to "hard" enforcement mechanisms at the global level, particularly in the case of human rights; hence, global human rights instruments are generally not endowed with coercive powers. Still, this article reveals that lack of coercion does not necessarily lead to ineffectiveness: it rather shows that different mechanisms for compliance, under certain conditions and to different extents, are relevant to explain states' performance. While it appears unrealistic to assume that all states are inclined towards compliance, it is certainly true that, when that is the case, countries need support to improve their record. To this aim, treaty bodies appear to be the most suitable instrument. By providing willing states with specific recommendations and an accurate overview of their own performance, they offer them the tools to increase their abidance to human rights provisions. In addition, previous studies suggest that the mere fact of engaging in reporting procedures may lead to domestic mobilization and socialization, revealing that the impact of these procedures goes beyond that of the specific reporting cycle (Heyns and Viljoen 2001;Creamer and Simmons 2018;). Yet, this article also showed that states are often uninterested, or unwilling, to improve without further incentives. In this regard, the UPR is seen as more successful in providing extra incentives for countries to comply with recommendations, mainly due to its perceived ability to generate pressureunlike the treaty bodies. Indeed, when states are committed to improving their human rights record, they will take both mechanisms very seriously (interviews 3, 22, 24, 27, and 30). Nonetheless, when such willingness is limited, the ability of the UPR to pressure states is a crucial factor in pushing them toward implementation, in line with arguments by the enforcement school (Downs 1998).
The UPR's higher performance from an enforcement school point of view can be further explained by the fact that recommendations are not endorsed by all UN states participating in the review, but are exclusively attributed to the country issuing them. Hence, they take the form of bilateral recommendations, which has strong political implications: accepted UPR recommendations become political commitments between countries. Although many recommendations are reportedly issued for political reasons, rather than out of human rights considerations, the result is that commitments are more likely to be adhered to than would be the case in a less political mechanism (interviews 1-4, 8, 22, 24-28, 30, 34, 38, and 39). This is, logically, even more the case when recommendations are delivered by a country with whom the reviewed state aims to maintain positive diplomatic relations (also see Carraro 2017; Terman and Voeten 2018). As put by a respondent: "It is so difficult to reject a recommendation because after each recommendation in the parenthesis you have the name of the country that made the recommendation" (interview 22).
Such dynamics are perceived to be lacking in the treaty bodies, where much less political pressure is put on states to comply: "It might be easier to reject a recommendation made by an expert than by a country because it is a country behind that recommendation, it is Germany, it is the USA" (interview 22). Similarly, Krommendijk (2015) finds that in the treaty bodies, international pressure alone is certainly helpful, yet not sufficient to ensure states will comply with the recommendations received.
Thus, as neither of the two mechanisms' recommendations can be legally enforced, it is up to individual states to decide whether they will implement them. The political nature of the UPR raises the stakes for noncompliance and, while this does not ensure implementation, it is perceived to at least make it more likely. To sum this up in an interviewee's words: If you are a very … committed human rights government, you will take human rights treaty bodies very seriously, probably more than the UPR, but if you are not a very human rights-oriented government, which is the majority in the world, [you will take] the treaty body … as an exercise which is sort of academic and you will forget about it the moment you leave the room, whereas with the UPR you will not, because it becomes political. (Interview 24)

Conclusions
This article proposed and applied a model to assess the performance of two UN human rights mechanisms: the UPR and the state reporting procedure of the treaty bodies. It did so by studying their process performance, namely, their ability to achieve a set of goals that might eventually lead to states' implementation of human rights recommendations. Subsequently, it evaluated whether these reviews' ability to trigger compliance is seen to derive from their capacity to achieve such goals. Data were collected by means of forty semi-structured interviews and an online survey, targeting individuals directly involved in the two procedures.
The article identified four outcomes that these mechanisms might reach: generating pressure; triggering learning; providing an accurate overview of states' performance; and delivering practically feasible recommendations. The two mechanisms show differing scores when it comes to their ability to achieve these goals. The UPR's perceived strength lies in generating peer and public pressure. The main explanation for the UPR's ability in generating public pressure is to be found in the active role that NGOs play in the process, holding states accountable for the commitments they made in the review. Additionally, the bilateral nature of UPR recommendations creates a much higher pressure on states to live up to their commitments than in the case of recommendations by nongovernmental experts. Conversely, treaty bodies fare better in providing accurate overviews of states' internal situations and learning opportunities. This is largely due to the expert nature of their recommendations, which are seen as more objective and of a higher quality than in the UPR. Finally, both reviews are successful in delivering feasible recommendations, although with a notable difference: while UPR recommendations are appreciated for being realistic (albeit often vague), Concluding Observations are praised for being very detailed, yet criticized for aiming at unattainable standards.
This article contributes to debates on state compliance with international provisions by showing that, under certain conditions, the three main theoretical schools on compliance-enforcement, management, and constructivist-offer credible explanations for states' performance in implementing human rights recommendations, although the enforcement school fares better than the other two. Managerial scholars are correct in claiming that states inclined toward compliance might need tools to correctly implement the recommendations received. In this light, when states are provided with an accurate overview of their internal challenges and receive instructions on how to tackle them, compliance is likely to improve. Similarly, in a more constructivist vein, states might still need to familiarize themselves with the rules of appropriate behavior and be exposed to a learning-stimulating environment. In both cases, treaty bodies appear to be a highly appropriate instrument to achieve those goals. However, it also occurs that states are not willing to implement the recommendations received without an extra push. This push can be reputational, or due to concern for more material consequences such as missing out on trade deals. In this case, following an enforcement logic, the UPR is a more adequate instrument to provide states with incentives for compliance due to its highly political nature, where recommendations become political commitments between countries.
From a policy perspective, these findings suggest that the UPR and treaty bodies could reinforce each other's strengths by working more closely together. Reviewing states in the UPR could more systematically consult recommendations delivered by the treaty bodies before formulating their own recommendations. 7 Even though this already occurs in several instances, it would be beneficial to make it a structural part of the process. If UPR recommendations were more strongly based on Concluding Observations, they would preserve their political force while, at the same time, providing better guidelines for states. Future research could highlight the extent to which the output of the UPR and treaty bodies is currently aligned, and the degree to which recommendations by one body inform those by the other.