The relational politics of shame: Evidence from the universal periodic review

Abstract

International human rights institutions often rely on “naming and shaming” to promote compliance with global norms. Critics charge that such institutions are too politicized; states condemn human rights violations selectively, based on their strategic interests, while protecting friends and allies. In this view, politicization undermines shaming’s credibility and thus its effectiveness. This paper offers an alternative account of such institutions and the mechanism by which they promote human rights. We argue that interstate shaming is an inherently political exercise that operates through strategic relationships, not in spite of them. While states are less likely to criticize their friends and allies, any criticism they do offer is more influential precisely because of this pre-existing partnership. We test this argument through quantitative analysis of the most elaborate human rights mechanism in the international system: the United Nations Universal Periodic Review. We find that states are more lenient towards their strategic partners in the peer-review process. Yet when they do criticize, their recommendations are accepted more often than substantially identical recommendations emanating from other states with fewer strategic ties. Insofar as shaming disseminates powerful signals regarding political relationships between states, these interactions can be meaningful and influential, even as they remain selective and politicized.

In late September 2015, international human rights groups erupted in outrage after a watchdog organization revealed that Saudi Arabia was nominated to chair an important human rights panel at the United Nations (Ingraham 2015). At the time, Saudi Arabia was facing widespread condemnation for its plans to execute a 21-year-old political dissident by beheading followed by public crucifixion. Asked about the appropriateness of Saudi Arabia leading a key human rights institution, a US State Department official replied by saying, “We would welcome it. We’re close allies” (Goldberg 2015). Meanwhile, classified documents passed through Wikileaks revealed that the United Kingdom traded votes with Saudi Arabia to assure that both countries gained a seat on the UN Human Rights Council (Bowcott 2015). “We have a relationship with Saudi Arabia,” David Cameron said when pressed on the deal (Channel 4 News 2015).

Anecdotes such as these fuel widespread skepticism of international human rights institutions, which largely depend on “naming and shaming” violators. A process that is simply about condemning adversaries while protecting friends and allies is not really about human rights at all. Others counter that while politics has a nefarious influence, such institutions can still promote human rights by revealing information about compliance or by socializing states into accepting shared norms.

In this paper, we offer an alternative account of such institutions and the mechanism by which they promote human rights. Interstate shaming is an inherently relational and political exercise, mediated by strategic ties between shamer and target. When it comes to human rights violations, states will condemn their friends and allies less harshly in order to maintain a valuable relationship. By the same token, however, strategic ties provide an important source of leverage. In a politicized environment, governments expect to be shamed by their enemies, and can easily brush off such commentary as a cynical attempt to sully their country’s reputation. That is not so easy if friends or allies offer the scrutiny. States will take shaming seriously when it emanates from a strategic partner in order to avoid damaging the relationship. In short, the efficacy of shaming is conditional on the relationship between source and target.

We illustrate this perspective through an analysis of the most elaborate multilateral human rights process in the international system: the United Nations Universal Periodic Review (UPR). In the UPR, governments voluntarily subject their human rights record to the scrutiny of their peers, who offer feedback in the form of specific recommendations. States under review must then publicly decide whether or not to accept each recommendation it receives. We examine over 40,000 recommendations from the first two cycles of UPR, testing the effects of four kinds of political relationships: geopolitical affinity, formal military alliance, arms trade, and humanitarian aid. We find strong evidence that states spare their strategic partners in the review process, giving less severe commentary on average. But when friendly states do offer criticism, their recommendations are more likely to be accepted by the state under review compared to substantively identical recommendations coming from other countries.

To be clear, our findings cannot determine the effects of the UPR on states’ domestic human rights practices. They do, however, contribute three significant advances to the study of norms and social pressure in international politics. First, we produce a more empirically refined portrait of “naming and shaming” between government actors. Prior studies typically examine shaming as a binary event (or count of events): a country either is or is not shamed (e.g., DeMeritt 2012; Franklin 2008; Hafner-Burton 2008; Krain 2012; Murdie and Davis 2012).¹ To our knowledge, this is the first quantitative analysis that disaggregates normative pressure by source, issue, and severity. The empirical leverage afforded by this approach enables us to demonstrate the importance of political relationships in the shaming process.

Second, we develop a novel theoretical account concerning the role of peer-review institutions like the UPR. Contrary to conventional models that emphasize socialization, credible expertise, or reputational effects (see below), we argue that strategic relations provide the mechanism linking rhetorical pressure to behavioral outcomes. In the UPR, recommendations are significant to the extent that they reveal information on what behavior risks undermining a mutually beneficial partnership.

Finally, we offer somewhat counterintuitive implications for the debate about politicization in human rights institutions. Most studies imply that shaming might work despite the corrupting influence of politicization. That is, politicization detracts from what the UPR is supposed to do, which is to provide objective information about human rights violations and/or put social pressure on violators by appealing to broadly accepted norms. We argue that this perspective is misleading. The UPR is political by design. As an institution, it provides governments a venue for publicly opining on the domestic behaviors of other governments in an attempt to exert influence. If the UPR “matters,” it is likely due to these political interactions, not because the process approximates an independent or impartial assessment of human rights. Indeed, the strong political ties between the United Kingdom, United States, and Saudi Arabia make it more likely that the latter country will be shielded from harsh scrutiny. But if the United Kingdom or United States did decide to offer sharply worded criticism, this rhetoric would be taken far more seriously precisely because of their pre-existing relationship. To the extent that the UPR facilitates and regularizes these exercises in social power, it may serve a meaningful role in the promotion of human rights.

1 The universal periodic review: Promise and perils

The Universal Periodic Review (UPR) is a process conducted by the UN Human Rights Council to periodically review the human rights practices of all UN member states.² It is the first international human rights mechanism to achieve 100% voluntary participation, addressing all 193 countries in the UN.³ Reviews are facilitated by the UPR Working Group, consisting of the 47 member states of the Human Rights Council, though any UN member can participate. The UPR working group meets three times per year, reviewing 14–16 states per session (the order of reviews is determined by lot). The first cycle ran from 2008 to 2011, while the second cycle started in 2012 and will conclude in 2016.

Reviews take place through an interactive dialogue between the state under review (SuR) and other UN members. First, the SuR presents a self-assessment in the form of a national report. Non-governmental organizations and other stakeholders may also submit information, but only the SuR presents directly to the Working Group. Following the state’s presentation, 140 min of interactive dialogue takes place, during which any other UN member states (and permanent observers, i.e., Holy See and Palestine) can make recommendations towards the improvement of the SuR’s human rights record.⁴ The SuR must then respond, declaring which recommendations it does and does not support. Once the review is complete, an outcome report is compiled by a group of three states known as a “troikas,” which are selected through a random drawing.⁵ States have 4.5 years to act on the recommendations it supports before undergoing another review.

The UPR arose in 2006 from the institutional ashes of the UN Human Rights Commission, which was heavily criticized for being too politicized. In a major 2005 UN reform report, Secretary-General Kofi Annan concluded that:

the Commission’s capacity to perform its tasks has been increasingly undermined by its declining credibility and professionalism. In particular, States have sought membership of the Commission not to strengthen human rights but to protect themselves against criticism or to criticize others. As a result, a credibility deficit has developed, which casts a shadow on the reputation of the United Nations system as a whole (United Nations Secretary-General 2005a, 182).

The UPR’s peer review system, Annan argued, would “help avoid, to the extent possible, the politicization and selectivity that are hallmarks of the Commission’s existing system” (United Nations Secretary-General 2005b, para. 8). While voting in the Human Rights Council remains as politicized as its predecessor (Hug 2016), many observers maintain that the UPR is a significant advancement in the fight for human dignity (Gujadhur and Limon 2016).

Underwriting this optimism is a robust theoretical literature on international norms, suggesting a number of mechanisms by which institutions like the UPR can promote human rights. One possibility is that the UPR works as a socializing influence. Constructivists view states as social creatures that care about prestige, status, and self-image, in addition to material rewards and punishments (Finnemore 1996; Finnemore and Sikkink 2001; Goertz and Diehl 1992; Katzenstein 1996; Kelley and Simmons 2015; Ropp, Sikkink, and Risse 1999; Towns 2012). Norms such as human rights are said to diffuse through “peer pressure”: conformity brings praise, increased social worth and esteem, while violation is met with shame, disapproval and isolation. Over time, states may internalize communal norms as part of their identities, i.e., members of the “community of civilized nations,” giving such norms a “taken for granted” quality (Finnemore and Sikkink 2001, 903–5).

For constructivists, institutions like the UPR are powerful because they embody social environments where states learn about shared expectations of appropriate behavior, and face social consequences for their ability or failure to adhere to those expectations (Goodman and Jinks 2013; Greenhill 2010; Johnston 2001). A number of empirical findings support this view. Using cross-national data, Greenhill (2010) demonstrates that state interactions in IGOs can promote human rights compliance, even in IGOs without an obvious human rights mandate. He attributes this effect to the socializing influence of IGOs. Other empirical studies have shown that “naming and shaming” – in and outside the context of IGOs – can improve compliance by harnessing the power of social inducements (DeMeritt 2012; Cole 2012b; Krain 2012; Hafner-Burton and Tsutsui 2005; Hendrix and Wong 2013; Murdie and Davis 2012). In sum, the UPR may be helpful by facilitating peer-evaluation, fueling the socializing process.

For scholars working in the liberal institutionalist tradition, states do not necessarily care about social approval per se, and yet institutions like the UPR may still influence human rights practices through indirect mechanisms leveraging information and reputation. For instance, evaluation in the UPR may inform an international audience on the degree to which a state has abided by international agreements. This could in turn shape its overall reputation for compliance and limit beneficial forms of cooperation such as trade agreements or foreign aid (Guzman 2007). Information mediated in such environments may also trigger domestic mobilization against a government (Simmons 2009).

Constructivists and liberal institutionalists converge on the belief that normative pressure can be effective to the extent that it accompanies bad behavior. Both perspectives view international norms as precisely that – international, i.e., emanating from an accepted standard that all (or most) states strive to embody. Even if governments do not fully internalize these standards, they may still be susceptible to shaming insofar as the failure to live up to communal norms damages their reputation or standing within the international community. Thus the UPR holds promise to the extent that it provides accurate information on states’ human rights performance, teaches states about communal standards, and/or harnesses the power of social inducements.

Other scholars remain skeptical, both of the UPR as well as of the power of IGOs to promote human rights more generally. For skeptics, governments are motivated primarily by security and economic interests, not social inducements or diffuse reputational concerns. Because human rights compliance is rarely enforced in any material sense, espousals to such ideals are often “window-dressing,” superficial and misleading when it comes to actual policy and behavior (Hafner-Burton and Tsutsui 2005). Likewise, states will shame one another in order to promote their own interests, not the universality of human rights. Indeed, there is considerable evidence that “naming and shaming” is driven by factors that are extraneous to actual violations (Boockmann and Dreher 2010; Edwards et al. 2008; Hafner-Burton and Ron 2013; Hill, Moore, and Mukherjee 2013; Hug and Lukács 2013; Lebovic and Voeten 2006; Murdie and Urpelainen 2015; Ramos, Ron, and Thoms 2007; Ron, Ramos, and Rodgers 2005). Importantly, states tend to shame their geopolitical adversaries in order to cast them in a bad light, while going easy on friends, even if they, too, violate human rights norms.

Indeed, the UPR’s critics lament that it has once again fallen victim to the politicization and selectivity that discredited its predecessor (e.g., Schaefer and Groves 2016). Rights-respecting states such as Canada tend to be attacked for their liberal economic policies, while oppressive states like Cuba and Iran enjoy praise from their ideological sympathizers (UN Watch 2009). As one editorial argues, this kind of selectivity damages the UPR’s credibility, transforming it into “a forum in which abusive governments can both trivialize their own crimes and belittle the enjoyment of fundamental civil and political rights by citizens in liberal democracies” (Mchangama and Rhodes 2013).

In many ways, this kind of politicization is institutionalized into the UPR’s procedures. As one report put it, “once states are judge and jury, foreign policy is never far from their thoughts when they take the floor” (FIACAT 2009, 6). While nongovernmental organizations can participate, the UPR is dominated by states, who are in the sole position to publicly criticize the records of other states. Thus the process is explicitly relational and political. As the late High Commissioner for Human Rights Sergio Viera de Mello diagnosed the problem: “Let’s be frank. Most of the people in this room work for governments. That is politics. For some people in this room to accuse others of being political is a bit like fish criticizing one another for being wet.” (Vieira de Mello 2003).

The selective and political nature of the UPR is there by design. And yet governments appear to care a great deal about their evaluations (McMahon 2012b). Moreover, research suggests that, under some conditions, IGOs and other socialization efforts can indeed affect state behavior or impose material costs (Ausderan 2014; Barry, Clay, and Flynn 2012; Cole 2012b; DeMeritt 2012; Hafner-Burton 2008; Krain 2012; Lebovic and Voeten 2009; Meernik et al. 2012; Murdie and Davis 2012; Hendrix and Wong 2013). We argue that if the UPR “matters,” it is likely because of its political nature, not in spire of it.

2 The informative value of politicized peer-review

2.1 Theoretical argument

We propose that governments are sensitive to symbolic political assessments, even when those assessments are neither independent nor impartial. In doing so, we build off a growing literature concerning the politicized roots of international norm diffusion. Instead of approaching norms such as human rights as representative of some monolithic “international community,” these works cast normative pressure as an exercise in social power, promoted and contested between specific states in relational terms. For instance, Kelley and Simmons (2015) argue that monitoring and indexing state behavior on human trafficking can be influential to the extent that the creator of these indicators wields status and credibility in the international system. Likewise, Lebovic and Voeten (2009) argue that condemnations in the now defunct UN Human Rights Commission exposed information both about a state’s poor rights record as well as its inability to muster a sufficiently large coalition to shield itself from multilateral scrutiny. The combination of these qualities allowed multilateral lending agencies to use Commission votes as signals reflecting state power. Finally, Goodliffe and Hawkins (2009) show that governments supported (or failed to support) a strong International Criminal Court because they were influenced by the positions of other states on whom they depend for valued goods.

Assuming this relational approach, our explanation for shaming in the UPR emphasizes the political relationship between source and target.⁶ Because states behave primarily to further their material interests, they strive to avoid alienating other states on which those interests depend. Goodliffe and Hawkins call this set of partners a “dependence network”: those with whom a state “regularly engages in exchanges of valued goods, where those exchanges would be costly to break” (Goodliffe and Hawkins 2009, 978). These goods can be economic, security, or political in nature, including humanitarian aid, foreign policy support, and military alliance. States exchange such goods in the context of a strategic partnership, where each party cooperates as a means to further their own interests. Insofar as states value these relationships, they will strive to maintain them by anticipating potential reactions from their partners. As Goodliffe and Hawkins note, these reactions need not involve explicit rewards or punishments contingent on the behavior in question. Rather, states merely anticipate the potential costs (or benefits) to a valued relationship, and factor these into their decision calculus (Goodliffe and Hawkins 2009, 978). All else equal, states would rather avoid provoking a negative judgment from their strategic partners.

How does this insight pertain to interstate interactions in the UPR? We propose that peer-review is mediated by strategic relations between reviewer and target, affecting the behavior of each party. This involves two distinct yet interrelated mechanisms. First, we expect governments to vary their evaluations depending on their political ties to the SuR. Although delegations are mandated to be objective in their reviews, the reality is that recommendations vary widely in tone, ranging from disparaging to congratulatory. On average, states will condemn their strategic partners less harshly so as to signal their commitment to a valued relationship. In general, states do not take well to accusations of human rights abuse. When evaluating a partner, delegations on average mute their criticisms so as to avoid this negative reaction.

At the same time, some states genuinely do care about human rights. This concern may originate in domestic societal pressure (e.g., an outraged public demanding condemnation of an abuse abroad) or the normative beliefs of government officials. In some cases, these preferences are strong enough to warrant a tough stance and risk alienating an otherwise beneficial partnership.⁷ When a state publicly shames its strategic partner, it serves as a credible signal reflecting the shamer’s preferences vis-à-vis human rights, and the target will take such criticism seriously in order to maintain the relationship.

We examine this second mechanism by investigating when the SuR “accepts” UPR recommendations. In institutional terms, accepting a recommendation forces the SuR to follow up on that item during its next review. In theoretical terms, accepting a recommendation involves a kind of public commitment that may entrap governments in their own rhetoric, leaving them vulnerable to normative pressure from transnational and domestic advocacy groups (Ropp, Sikkink, and Risse 1999; Simmons 2009). A number of empirical studies support this view, demonstrating that public commitments (such as treaty ratifications) improve a state’s human rights performance, even where there is a lack of coercive mechanism to enforce compliance (Cole 2012b; Cole 2012a; Simmons 2009). Indeed, observers of the UPR note that states do, in fact, implement at least some of the recommendations they accept. One study calculated that 48% of accepted recommendations were either fully or partially implemented by mid-term (2.5 years after initial review) versus just 19% of the non-accepted recommendations (UPR Info 2014, 5). For these reasons, we view the acceptance of UPR recommendations as a meaningful, if not determinative, sign of compliance.

When deciding whether or not to accept a recommendation, states must factor in political context, because recommendations reveal very different signals depending on the source. In a politicized environment, governments interpret criticism by their enemies as a cynical attempt to sully their country’s reputation. Not only are there few incentives to comply in such cases, doing so may confer costs on the part of the SuR if they are seen as “kowtowing” to the enemy. On the other hand, recommendations mean something very different when coming from a state that shares strong political, economic, or security ties with the target. Since there are few strategic incentives to criticize friends, shaming in this case serves as a credible signal reflecting the critic’s preferences on a particular norm. In this context, the SuR is more likely to accept the recommendation in order to avoid damaging a valuable partnership.

2.2 Hypotheses

To evaluate the validity of these theoretical arguments, we focus on four types of strategic relationships that exchange, respectively, foreign policy support, formal military alliance, weapons/arms, and development aid. We chose these ties because they offer a broad sample of strategic relationships that are likely to influence interstate criticism in the UPR. First, we capture foreign policy support by examining the degree of geopolitical affinity. In domestic realm, citizens are more likely to accept messages from elites that share their partisanship (e.g., Bartels 2002; Rahn 1993). A similar mechanism operates at the level of international governance. States who have embraced the neoliberal order advanced by the U.S. and other Western states may easily brush off critiques from those who rebel against it (and vice versa). Governments can exploit this by highlighting the source of a recommendation rather than its content, implying it was motivated by political hostility rather than normative commitments. Moreover, publicly dismissing a human rights concern from a geopolitical adversary is less likely to endanger other cooperative endeavors, especially with states that share its ideological position. In contrast, harsh criticisms from governments with similar foreign policy dispositions cannot be dismissed as easily.

This leads to two testable hypotheses. First, all else equal, we expect that states will go easier on those that share their geopolitical agenda, and reserve their most severe criticism for states that challenge it. At the same time, recommendations coming from like-minded states are more likely to be accepted, controlling for the severity and content of the recommendation.

• Hypothesis 1A : All else equal, states are more lenient on states that share their geopolitical ideology.

• Hypothesis 1B : All else equal, states are more likely to accept recommendations coming from states that share their geopolitical ideology.

Second, we look at the effects of formal military alliances on state interactions in the UPR. Formal military alliances hinge on the belief that states will come to each other’s defense. Public discord may undermine that belief and sabotage a mutually-benefitial partnership. Thus states are more likely to be constrained when addressing their military allies. If alliance partners do criticize each other, there may also be a stronger incentive to accept the recommendation in order to squash concerns about a conflictual relationship.

• Hypothesis 2A : All else equal, states are more lenient on states with which they share a formal military alliance.

• Hypothesis 2B : All else equal, states are more likely to accept recommendations coming from states with which they share a formal military alliance.

In the hypotheses above, we make the simplifying assumption that foreign policy support and military alliance reflect symmetric relationships. In reality, however, inter-state relations often involve some sort of dependency. To explore this dynamic, we examine two forms of relational ties that explicitly involve some kind of directional dependence: weapons/arms exports and humanitarian aid. We posit that arms importers and aid recipients may fear criticizing their supplier too harshly, lest they sabotage this valued relationship. As one report documented, an African diplomat conveyed that “he would think twice about producing a criticism of western states who are donors, such as the U.S. and the U.K.” (McMahon 2012a, 16).

At the same time, arms exporters and aid donors face similar incentives towards their recipients. The U.S. exports arms to Saudi Arabia not just for financial gain but also because it values a strategic partnership. More generally, both foreign aid (e.g., Alesina and Dollar 2000) and arms exports (Blanton 2005) are at least partially driven by shared interests between donor and recipient. Thus even powerful states benefit from the maintenance of such relationships, and avoid provoking negative judgments from their clients (Esarey and DeMeritt 2016). Moreover, powerful states may wish to mitigate the perception that they aid and abett a human rights violator.

Consider, for example, when the Obama administration approved $1.3 billion in military aid to Egypt following Sisi’s crackdown on dissenters. The move attracted widespread scrutiny, with journalist Glenn Greenwald bemoaning that the administration “lavished the regime with aid, money and weapons” while ignoring its repression (Greenwald 2015). In fact, Secretary of State John Kerry did address Egypt’s human rights issues while announcing the renewed military relationship, but only obliquely, and in a series of platitudes as he sat next to his Egyptian counterpart in a joint press conference. As the New York Times put it, Kerry was signaling that American officials “would not let their concerns with human rights stand in the way of increased security cooperation with Egypt” (Gordon and Kirkpatrick 2015). In sum, both aid/weapons suppliers as well as clients are incentivized to be more lenient in their human rights recommendations towards their partner.

• Hypothesis 3A : All else equal, states involved in an arms trade are more lenient towards each other.

• Hypothesis 4A : All else equal, states involved in an aid relationship are more lenient towards each other.

On the other hand, these relationships are asymmetric, involving a power disparity between supplier and client. Although both states benefit from a strategic relationship, patron states wield greater influence over their clients than vice versa. Specifically, donors and arms exporters have the opportunity to make the continuation of an exchange relationship conditional on human rights performance in a way that recipients and importers do not. Indeed, there is some evidence that human rights concerns play a role in the allocation of foreign aid and arms, albeit selectively (Blanton 2005; Nielsen 2013). If a donor or arms exporter offers criticism, this may be interpreted by the target state as a signal that the relationship is at risk. In this sense, recipients of aid and arms are more sensitive to the judgments of their patrons than the other way around.

In the UPR, this asymmetry may be irrelevant for the first mechanism; the costs of giving a muted review are so low that both patrons and clients will do it in order to avoid sabotaging a mutually beneficial relationship. However, the costs of accepting a recommendation are significantly higher, as it involves a potential constraint on domestic rule. In other words, there is more at stake in responding to criticism than offering it. Thus we expect that only the powerful partner in an asymmetric relationship can wield influence in this regard.

• Hypothesis 3B : All else equal, arms importers are more likely to accept recommendations coming from their exporters.

• Hypothesis 4B : All else equal, aid recipients are more likely to accept recommendations coming from their donors.

Of course, these four hypotheses do not exhaust the range of political dynamics in the UPR. A number of other relational facets may exert influence, but we do not consider them in this essay. Notably, previous studies suggest that regional and cultural affinity mediate norm socialization (Goodliffe and Hawkins 2009; Simmons, Dobbin, and Garrett 2007).⁸ Without dismissing this insight, our argument specifically focuses on the influence of strategic ties that facilitate the exchange of material goods. Likewise, we limit our analysis to dyadic relations, excluding potential audience effects that may result from the fact that state behavior is observed by third-party states, civil society, and domestic parties. While recognizing the potential influence of these mechanisms, we maintain that the analysis of geopolitical and military/economic dependence offers valuable insight into the relational politics of the UPR.

3 Research design

We use data collected by the non-profit organization UPR Info on all recommendations made during the first 20 sessions of the UPR working group (n = 41,066). The data are mined from outcome reports from all 192 countries reviewed during the first cycle, and 112 countries in the second cycle.⁹ For each recommendation, data was collected on the state offering the recommendation (the Reviewer), the state receiving the recommendation (the Target, coterminous with the SuR), and the SuR’s response to the recommendation (the Response). UPR Info researchers also hand-labeled each recommendation according to the kind of action demanded on the part of the SuR, and the specific human rights issue involved, explained below.¹⁰

3.1 Dependent variables

Our hypotheses concern two main dependent variables: the Severity of recommendation, and the Response to the recommendation. Response records, simply, whether or not the SuR supports a recommendation. Severity measures the level of leniency/severity in the recommendation’s content based on the kinds of actions demanded of the SuR. UPR Info researchers coded each recommendation according to a 5-point categorical variable based on the first verb and the overall action contained in the recommendation. We recoded this measure as a 3-point ordinal measure, which we call Severity.¹¹ Recommendations coded as 1 on this scale would not be considered “shaming” by any typical definition; they either praise the SuR or request minimal change (e.g., “share best practices,” “request technical assistance”).¹² Unsurprisingly, recommendations of this type are very likely (96%) to be accepted by the SuR. Recommendations coded as 2 are the most common, and contain a general behavioral element (e.g., “encourage,” “engage with”). Of these, 84% are accepted. In contrast, only 56% of level 3 recommendations, requesting that the SuR change their behavior, are accepted.¹³ Thus higher values on the Severity scale denote more severe or demanding recommendations. In some of the subsequent analyses, we remove recommendations with Severity levels 1 and 2 in order to make sure our hypotheses hold with only the most critical interactions.

3.2 Explanatory variables

To test the effects of Geopolitical Affinity (Hypothesis 1), we estimate the absolute distance between country ideal points estimated using votes in the United Nations General Assembly (Bailey, Strezhnev, and Voeten, Estimating Dynamic State Preferences from United Nations Voting Data, unpublished), and multiply this by minus one, thereby transforming it into a measure of affinity. Larger values represent smaller distances and thus higher levels of ideological convergence on global issues. To test effects of formal Alliance (Hypothesis 2), we use the Correlates of War Formal Alliance (version 4.1) data indicating whether or not a formal alliance exists between reviewer and target (Gibler 2008). To test the effects of Arms trade (Hypothesis 3), we include two binary variables indicating whether the reviewer supplies arms to the target and vice versa, using data from the Stockholm International Peace Research Institute Arms Transfers Database. Finally, to test the effects of Aid relations, we calculate the proportion of total aid receipts (donations) the reviewer receives (gives to) the target based on data from the Organization for Economic Co-operation and Development. For instance, Aid (Target to Reviewer) captures the proportion of total foreign aid donations the reviewing country gives the SuR.¹⁴ The results are substantively identical when we use dummy variables for aid donor and aid recipient status.

We also consider a number of potentially confounding variables. The most glaring possibility concerns the substance of recommendations. Countries may consider some human rights issues, such as torture or genocide, as particularly serious, demanding stronger actions. Likewise, countries are likely to accept or reject recommendations based on the issue involved. For this reason, we control for the thematic Issue involved in each recommendation. We took UPR Info’s 54 hand-labeled codes and used an automatic clustering algorithm to aggregate them into 8 more manageable categories: (1) Women, Children & Trafficking, (2) Physical Integrity Rights (including the death penalty) (3) Justice, (4) Speech & Political Participation, (5) Race, Ethnic, & Religious Discrimination, (6) Migration, (7) Socio-Economic Rights, and (8) Vulnerable Populations. ¹⁵ Similarly, because states are more likely to accept recommendations that are more lenient, we control for the Severity level of recommendations in the analyses of state response. By controlling for Issue and Severity, we are able to isolate the effects of dyadic relationships across substantively similar recommendations.

Another potential confound involves the tendency for norm-abiding states to shame norm-violating states in harsher terms. Given that some of our hypothesized explanatory variables are plausibly correlated with human rights records, we include a measure of Physical Integrity Rights Protections (Fariss 2014), taking the difference between reviewer and target level in order to capture relational dynamics. Given that all our models use country fixed effects over a short time period, we conclude that the difference measure is the appropriate choice in this case.

We also coded for whether the reviewer country was undergoing a review in the same year as the target (Reviewer UPR). Reviewer states who themselves undergo a UPR in the same year may wish to be seen as participating, but might shy away from politically-sensitive commentary due to expectations of reciprocity. Finally, many observers note that co-regionals face more pressure to deal tactfully with one another (FIACAT 2009; McMahon 2012b; UN Watch 2009). Shared region is strongly correlated with UN voting patterns and may thus confound relationships between our variables of interest. Thus we code for whether the target and reviewer countries come from the same region (Region), using classifications from the Correlates of War project.

3.3 Method

Our analysis proceeds in two stages.¹⁶ First, we examine the behavior of reviewer states. Given a review, does a delegation participate and, if so, how severe are their recommendations? These are clearly related decisions. Indeed, the UPR’s process offers incentives for geopolitical allies to participate in each other’s reviews, but not necessarily to give honest feedback.¹⁷ In order to account for the possibility of selection bias, we estimate a two-step Heckman selection model. The sample consists of all directed dyads between states undergoing a review in a given year and members of the United Nations (potential reviewer states), yielding 58,224 observations.¹⁸ The selection variable is Participation, a binary variable indicating whether a potential reviewer offered a recommendation to the SuR (true in 25% of cases), and the outcome variable is the mean level of Severity in the recommendation(s) offered by a reviewer state. The exclusion criterion is whether the reviewer state is an HRC member at the time of a review. The reasoning is that HRC members are much more likely to participate in the UPR, but there is no reason to expect HRC members to request more or less severe recommendations, all else equal.¹⁹

Diagnostic tests confirm the appropriateness of the Heckman model: both the likelihood ratio test (first stage model) and the Wald test (whole model) are significant at high levels, which allows for a rejection of the null-hypothesis that all coefficients jointly equal zero. Further, rho is negative, suggesting that the Heckman approach is an appropriate way to model this data. We estimate two models, with and without Geopolitical Affinity, given that shared foreign policy interests are likely collinear with our other relational variables.

The second analysis examines the actions of the SuR. Here, the unit of observation is an individual recommendation (n = 41,066) and the dependent variable is a dichotomous Response variable indicating whether or not the SuR accepted the recommendation. We report results from logit models (although all results are robust to a straightforward OLS estimation) that include each of our hypothesized explanatory variables as well as a cumulative model. In all tables, we present estimates of marginal effects holding other variables at their means in order to facilitate substantive interpretations.

Finally, there are likely unobserved characteristics of reviewer and target states that affect their propensity to receive and send recommendations. In both analyses, we include fixed reviewer and target country effects, which control for un-modeled and stable state characteristics. We also include fixed effects for the year in which the UPR review took place in order to control for possible learning effects or unobserved contextual factors that shape the review process at particular times. Notably, more recommendations were made in the second round of the review process than in the first. Dyadic fixed effects are impossible in our context, given that individual dyads occur once or twice in the data.

4 Findings

4.1 The determinants of UPR participation and recommendations

Our first analysis focuses on the actions of reviewers as they evaluate other states in the UPR. The findings, summarized in Table 1, confirm that reviewer states participate more in UPR reviews of their strategic partners but are less severe in their commentary. First, we find strong support for Hypotheses 1A. While Geopolitical Affinity makes UPR participation more likely, a one-point (one standard deviation) increase in Geopolitical Affinity reduces the mean level of Severity by .13 (on a 3-point scale).

Table 1 Heckman selection models on UPR participation and severity

Formal military allies (Hypotheses 2A) are more likely to participate and slightly less likely to offer strong recommendations, although the magnitude of this effect is not as strong as Geopolitical Affinity (.04 on the three point scale). We also find that recipients of Arms imports (Hypothesis 3A) make slightly less stringent recommendations of their suppliers (.09 on the 3-point scale), especially once we remove Geopolitical Affinity from the model. On the other hand, Arms exporter status does not have a perceived effect on Participation or Severity.

We find that aid donors (Hypothesis 4A) are more likely to participate but also more likely to offer weaker recommendations (0.64) when the SuR is an important recipient state. This confirms the reasoning that aid relationships are strategic. For example, the U.S. gives large proportions of its aid to Israel and Egypt; it also has reasons to participate cautiously in human rights reviews of those countries. Similarly, countries often give large proportions of aid to their former colonies. They may well have an interest in participating in these countries’ rights reviews, but avoiding strong criticism. There is also some evidence that aid recipients exhibit the same pattern when they evaluate donors, although the effect doesn’t reliably reach conventional levels of statistical significance across specifications. Region has a similar influence. Indeed, in most reviews, the regional grouping most represented among participants is the group to which the SuR belongs (FIACAT 2009, 26). At the same time, co-regionals tend to take a more lenient tone with one another.

A good illustration of these dynamics is Tunisia’s 2008 review. Of the 64 states that spoke during the interactive dialogue, over 50 used their time to congratulate Tunisia on its human rights performance. Japan, a donor, called it a model of democracy. France, former colonial power and current donor, delicately avoided the sensitive issues of torture and freedom of expression. Nearly half of the participants belonged to the Organization of the Islamic Conference, whose members tend to act in concert in UN institutions (FIACAT 2009, 26).

There is also some evidence of institutional effects. Countries sitting on the HRC are more likely to participate, as well as countries that are undergoing a review in the same year. Given reviewer fixed effects and the fact that the timing of UPR reviews is random, the UPR effect could be interpreted as causal. Yet there is no evidence that countries undergoing a UPR review are more lenient in their recommendations. We also find that states are more likely to offer more, and more severe, comments when their records of physical integrity rights are superior to the SuR.

In all, geopolitical friends, military allies, aid donors, and co-regionals are less likely to issue recommendations demanding strong remedial actions. Still, governments do rebuke their strategic partners somewhat regularly. For example, we count 1454 instances in which a military ally issued a recommendation at the highest level of severity in our coding, and 339 times when such a recommendation emanated from an arms exporter. What happens when states shame their strategic allies in such harsh terms? The next section examines this question.

4.2 Responses to shaming

The second analysis explores the behavior of the SuR, asking what drives states to respond the way they do to UPR recommendations. Table 2 reports estimates of marginal effects from our logit model, holding other variables at their means in order to facilitate substantive interpretations. Figure 1 plots the marginal effects from the first model to get a better sense of substantive effects. Unsurprisingly, states are more likely to accept those recommendations that involve vague or congratulatory language over those involving specific demands. Similarly, states appear to be more open to some issues than others. Namely, recommendations relating to “Women, Children & Trafficking” are 10 percentage points more likely to be accepted than the default. Those involving “Socio-economic Rights” also appear to be amenable to states. In contrast, recommendations involving “Physical Integrity Rights” and especially “Migration” (i.e., citizenship and refugee issues) are more contentious.

Table 2 Logit models on SuR response to UPR recommendations

Fig. 1

Substantive effects on probability of recommendation acceptance. Note : These estimates were drawn from the same model reported in Table 2 , Model 1

Even after controlling for the substantive content of recommendations, we find that relational variables correlate strongly with state responses. For instance, states are more likely to accept recommendations when the reviewer has a better human rights record, speaking to the potential credibility of “defenders of the faith” such as Canada or Norway. At the same time, strategic ties exert an influence that is at least as strong as normative credibility. As predicted by Hypothesis 1B, states are more likely to accept recommendations coming from geopolitical sympathizers. Countries under review are six percentage points more likely to accept a recommendation as the reviewer state moves one standard deviation in its UN voting pattern, even after controlling for substantive characteristics of the recommendation. States are also more likely to accept recommendations from their Arms suppliers (Hypothesis 3B). On the other hand, we find mixed support for the effects of Alliance (Hypothesis 2B), which is only significant once Geopolitical Affinity is excluded from the model. One plausible explanation is that geopolitical interests shape alliance formation, resulting in a correlation between these two variables and suppressing the effect of Alliance. Similarly, Aid is only significant in some models.

We conduct two additional analyses to test the robustness of these findings. Table 3 re-estimates the models on a subset of data containing only the most severe recommendations; that is, those that demand a substantive change in behavior from the SuR. Again, we find that recommendations concerning “Women, Children & Trafficking” were among the most amenable to states (16 percentage points more likely to be accepted) while those involving “Migration” were among the least. Indeed, among the most severe recommendations, those involving migration issues are twelve to fourteen percentage points less likely to be accepted.

Table 3 Logit models on state response to UPR recommendations (Severity Level 3)

As for our hypotheses, we find the effect of Geopolitical Affinity (Hypothesis 1B) is even stronger in this subset (eight percentage points, from six percentage points in the full sample). Alliance (Hypothesis 2B) and Arms dependence (Hypothesis 3B) were also slightly stronger in magnitude. On the other hand, Aid did not reach conventional levels of significance among the subset of strong recommendations. We repeated the analysis with dummy variables for aid relationships and received similar findings: aid dependence is significant in some model specifications but not in others.

Finally, we estimated a Heckman selection model on the dyadic sample used in the first analysis on state participation.²⁰ The dependent variable in the outcome equation is the proportion of recommendations accepted from a specific reviewer state. The exclusion criterion is HRC membership of the reviewer state. Again, we find that Geopolitical Affinity and Arms dependence are robust correlates of recommendation acceptance.

In sum, our findings present strong support for Hypotheses 1B and 3B concerning, respectively, geopolitical affinity and arms dependence. Formal military alliance (Hypothesis 2B) receives some support as well. These relational ties appear to be important considerations for state delegations as they decide whether or not to support the recommendations they receive in the UPR. On the other hand, we are less certain about the significance of aid dependence (Hypothesis 4B). Overall, however, these results confirm the importance of political relationships in states’ receptivity to shaming. Two recommendations that address identical human rights violations, while making similar demands, can land with very different reactions depending on the source.

5 Implications and conclusions

This study makes three contributions to the literature on norms and social pressure. First, we demonstrate the importance of relational ties for both the causes and consequences of naming and shaming. As a dialectic process, shaming is not determined by target country characteristics alone. States condemn norm violations selectively, based on their strategic relationship with the violator. Likewise, a state’s sensitivity to normative pressure depends on its relationship to the source of that pressure. In other words, the critic matters just as much as the criticism. This insight has important implications for future empirical studies. Instead of asking whether shaming “works” – the conventional approach in the literature thus far – scholars should investigate when, where, and by whom the tactic is helpful in the promotion of human rights.

Second, we advance the theoretical literature by delineating the mechanism by which IGOs influence state behavior on human rights. We propose that the UPR is significant because it greatly enriches the informational environment, but not in the way constructivists or liberals would have us believe. The UPR is not a socializing domain as constructivists claim because condemnation is the result of specific political relationships, not a reflection of discordance with the international community writ large. Nor is the UPR a credible source of expertise, as recommendations are filtered through the mouths of state delegations, and rarely expose new information about actual human rights abuses.²¹ Rather, the UPR is significant because it reveals what behavior can continue without risking specific partnerships (and vice versa). This information is meaningful in its own right, both for governments, who must calculate the foreign policy costs of their behavior, and potentially non-state actors, who must bet on expected support for mobilization. In certain circumstances, such information could plausibly drive compliance.

Finally, these findings have important implications for the debate surrounding politicization in international human rights institutions. In the conventional view, the introduction of strategic interests or affinities is an inherently corrupting influence on the ideal of an impartial, independent assessment. And yet, in the UPR, these strategic ties constitute the very mechanism by which social pressure drives behavioral outcomes. Indeed, with the exception of a few pariah states that have alienated themselves from nearly all states, shame is typically the result of specific political relationships and not necessarily a reflection of discordance with the international community writ large. To the extent that institutions such as the UPR promote human rights compliance, it is likely because of these political dynamics, not in spite of them. Moreover, the basic mechanism described here may present itself in other contexts, including other IGOs. Although the UPR involves a number of distinct features, institutions that rely on peer review (such as the U.N. Convention against Corruption) or public votes may exhibit similar patterns. However, future research is needed to determine precisely when and where states influence one another in this fashion.

Finally, the dynamics surrounding politicization is normatively ambiguous. Saudi Arabia, for instance, may indeed improve human rights behavior following pressure from allies, but it may also be emboldened to continue abusing human rights on account of its allies’ acquiescence. We do not suggest that politicization is normatively desirable; simply that it is empirically integral to the shaming process. Future studies should examine the extent to which politicized shaming at the UPR affects government behavior, such as the ratification of treaties, adoption of domestic legislation, or human rights observance as measured by CIRI scores or other indicators. Moreover, we need to better understand what it takes for a country like the U.S. or the U.K. to publicly criticize a strategic partner like Saudi Arabia. Our theory suggests that the very reasons why strategic allies are shielded from scrutiny are also why shaming could actually make a difference when it does occur.