To facilitate flexible commitments, states are permitted when they join treaties to enter reservations – statements that claim to modify or release them from being legally bound by specific treaty provisions (Vienna Convention on the Law of Treaties [VCLT] 1969: Article 2[1][d]). States use reservations to modulate their treaty obligations, particularly in the area of human rights (Neumayer Reference Neumayer2007) and especially when obligations are demanding (Zvobgo, Sandholtz, and Mulesky Reference Zvobgo, Sandholtz and Mulesky2020). Reservations reduce treaties' effect on state behavior by limiting executives' legal exposure under a treaty (McKibben and Western Reference McKibben and Western2020), keeping domestic standards low and undercutting judiciaries that would uphold international standards (Hill Reference Hill2016). Islamic law countries (McKibben and Western Reference McKibben and Western2020; Simmons Reference Simmons2009) and common law countries (McKibben and Western Reference McKibben and Western2020; Simmons Reference Simmons2009; Zvobgo, Sandholtz, and Mulesky Reference Zvobgo, Sandholtz and Mulesky2020) are also more likely to enter reservations than civil law countries.Footnote 1
Still, the practical impact of reservations is not entirely clear, and reservations create complex legal dynamics. Fellow treaty members are allowed to object to other members' reservations, which poses the problem of how to resolve disputes.Footnote 2 Most of the time, the disputing countries agree not to let their disagreement prevent them from being mutually bound by the treaty (Neumayer Reference Neumayer2007). An objecting country can also ‘sever’ (refuse to accept) a reservation and, from the objecting country's perspective, both it and the reserving country remain mutually bound by the full treaty. Another possibility is that the reserving country withdraws its reservation. Although scholars identify some of the factors that lead states to enter reservations, as well as lodge objections to reservations (Eldredge and Shannon Reference Eldredge and Shannon2022), we know relatively little about what motivates states to withdraw them.Footnote 3 Yet the withdrawal of reservations may reveal interesting aspects of state behavior. States do not need to rescind their reservations, especially because severance and withdrawal usually have the same legal effect – states are mutually bound by the treaty (McCall-Smith Reference McCall-Smith2014). Still, many states go to the trouble of publicly withdrawing reservations. Why?
We propose that states withdraw reservations in response to international social pressure. We identify two pathways for social pressure leading states to withdraw treaty reservations: peer pressure by fellow states and institutional pressure by human rights treaty bodies. First, a state withdraws a reservation because it seeks to appease states that object to the reservation and maintain social relationships with them. Second, a state withdraws a reservation because it seeks to comply with a recommendation made by a human rights treaty body in a periodic review.
We evaluate these propositions with the first broad empirical analysis of reservation withdrawals across the nine core treaties of the international human rights regime. We focus on human rights treaties because reservations and objections are most extensively used in this regime (Koremenos Reference Koremenos2016). Other regimes such as labor, environment, and arms control tend to rely on other mechanisms to give their members flexibility in treaty commitments (Helfer Reference Helfer2006). While it is possible that the dynamics we study also apply to other regimes, they are likely most pronounced in the human rights regime.
The analysis reveals that peer and institutional pressure often leads states to withdraw reservations to human rights treaties. While previous work emphasizes domestic factors that prompt reservation withdrawals, our findings show that the international community influences states' decisions to lift treaty restrictions. This indicates that the international community is a force in persuading states to improve their commitments to the human rights regime and international law.
Entering and Withdrawing Treaty Reservations
In a 1951 advisory opinion, the International Court of Justice (ICJ) concluded that countries can make reservations to treaties (ICJ REP 15).Footnote 4 A reservation is a tool that countries can use during the treaty ratification process to formally declare that certain treaty provisions will not bind them, or at least not to the full extent envisioned by treaty designers. The primary benefit of reservations is providing countries flexibility in committing to treaties and embedding them in an agreement in which they might not otherwise participate (Edry Reference Edry2020). The primary drawback of reservations is the contention they can elicit.
States are not unconditionally permitted to enter reservations; there are limits to their use. Building on the ICJ's 1951 advisory opinion, the VCLT sets conditions on reservations, which are allowed ‘unless (a) the reservation is prohibited by the treaty; (b) the treaty provides that only specified reservations, which do not include the reservation in question, may be made; or (c) in cases not falling under sub-paragraphs (a) and (b), the reservation is incompatible with the object and purpose of the treaty’ (1969: Article 19). Reservations that violate these stipulations are invalid, or carry no legal effect. Interestingly, there does not exist an automatic review process or a dedicated international body that reviews reservations for their validity. Rather, validity is tested in part by fellow treaty members' reactions to reservations. One way for treaty members to proclaim that reservations are invalid is through objections. Another is through the involvement of treaty bodies that evaluate a state's commitment to and compliance with a particular treaty, though they only have authority over the state after ratification.
Much of the political fight over reservations concerns the ‘object and purpose’ of the treaty, as our data will show. Thirty percent of state objections to reservations by peer states allege violations of the object and purpose of the treaty under consideration.Footnote 5 There are several outcomes that can occur as a result of state objections. First, the most common outcome is for the objecting country to say that the reservation will not prevent the treaty from entering into force between the two countries (McCall-Smith Reference McCall-Smith2014; Neumayer Reference Neumayer2007). In such situations the reserving country is bound by the treaty except for the provision against which it has reserved (Goodman Reference Goodman2017). Second, the objecting country may avoid clarifying the effect of its objection, leaving the legal status of the treaty unclear. A third outcome is the objecting country declares the reserving country fully bound by the treaty. In other words, the objecting country refuses to accept, or ‘severs,’ the reservation (McCall-Smith Reference McCall-Smith2014).Footnote 6 Finally, in the most extreme cases, the objecting country may reject the reservation and dissolve its legal relationship under the treaty with the reserving state.Footnote 7
To avoid these contentious outcomes, a reserving state can withdraw its reservation (Edry Reference Edry2020), a process outlined in Article 22 of the VCLT. Withdrawal is the approach human rights committees emphasize in their periodic reviews of countries (Edry Reference Edry2020). McCall-Smith (Reference McCall-Smith2014) describes withdrawal as a ‘state-sensitive’ approach: it allows states to save face when confronted by objections, as reserving states can withdraw their reservations rather than have them severed by other states. Pakistan, for example, withdrew its reservation in 2011 to Article 40 of the International Covenant on Civil and Political Rights (ICCPR), which declared Pakistan would not recognize the authority and jurisdiction of the Human Rights Committee, the treaty's formal monitoring and enforcement arm.
Despite the advantages, it is puzzling that states choose to withdraw their reservations. If states make reservations to preserve their sovereignty, and if reservations signal an intention not to be bound by certain parts of a treaty, it is not clear why states withdraw them. And though controversial, severance is widely accepted in practice, meaning states can join treaties with reservations even when their peers object.Footnote 8
What then compels states to withdraw reservations? Relatively little research addresses this question, which is surprising: withdrawal can be an important signal of a state's intention to improve its treaty commitments and foreign relations. When states withdraw reservations, they remove constraints they had previously placed on a treaty, rendering themselves subject to a greater number of binding obligations. Withdrawing reservations further embeds states into the regime. For this reason, non-governmental organizations (NGOs) like the Citizens' Initiatives on CEDAW and the Women's Learning Partnership use targeted campaigns to get states to withdraw reservations to human rights treaties (Edry Reference Edry2020).
This research extends a line of work that argues behaviors besides ratification can illuminate important aspects of treaty participation and international politics (Comstock Reference Comstock2021; Simmons Reference Simmons2009). Investigating what compels states to withdraw reservations can help scholars and legal practitioners understand the domestic and international processes that lead states to strengthen their commitments under international law. Perhaps the most direct investigation of reservation withdrawals is Edry (Reference Edry2020), who focuses on how domestic pressure toward greater human rights treaty participation leads countries to withdraw reservations to the Convention on the Elimination of All Forms of Discrimination Against Women (CEDAW). We build on Edry's study in two ways to develop a deeper understanding of reservation withdrawals. First, we explore a broader sample of human rights treaties, allowing us to generate inferences about the withdrawal process across human rights issue areas.Footnote 9 Second, we explore how the reactions of treaty members and treaty bodies can persuade countries to withdraw their reservations, while accounting for the domestic politics that influence withdrawal, as Edry does. In the next section, we discuss how peer and institutional pressure compels countries to withdraw reservations to human rights treaties.
In summary, states can make reservations when ratifying treaties. Fellow treaty members can accept the reservations or can lodge objections to the reservations. Treaty bodies can also review reservations for validity after ratification, though their recommendations on reservations are not binding. A number of outcomes result from reservations and objections, some of which are more contentious than others. One possible and less contentious outcome is for a country to withdraw its reservation. We now discuss some of the international dynamics that motivate states to withdraw reservations.
International Social Pressure and Treaty Reservation Withdrawal
We argue that international social pressure induces states to withdraw treaty reservations. We suggest two pathways: peer pressure by fellow states and institutional pressure by human rights treaty bodies. In the case of the former, a state withdraws a reservation to appease objecting states and maintain its social relationships. In the case of the latter, a state withdraws a reservation to comply with a recommendation (or an expected recommendation) by a human rights treaty body in a periodic review. We elaborate on these in turn.
Peer Pressure via State Objections
Human rights treaties reflect a desire by states to mutually bind themselves to specific legal commitments. Since treaties benefit from broad participation, treaty members with stronger commitments have an incentive to persuade fellow members to strengthen their participation. In fact, previous research demonstrates how external forces influence states' participation in treaties. Social networks among states, for instance, encourage particular behaviors, strategies, and actions (Hafner-Burton and Montgomery Reference Hafner-Burton and Montgomery2006). If state-to-state pressure, particularly from human rights ‘stewards,’ encourages target states to improve their human rights practices (Hafner-Burton Reference Hafner-Burton2013), it is likely that states increase their treaty commitments as a result of this peer pressure. As Simmons notes, ‘No doubt does brushing up against international society have some influence on governments’ decision to ratify human rights treaties…’ (2009, 64). We can expect that international society similarly encourages states to lift treaty reservations and improve their treaty participation.
While states possess numerous tools of persuasion, objection is the most immediate, legal, and visible way to signal dissatisfaction and influence a reserving state. An objection occurs when a treaty member publicly announces that they do not accept another state's reservation. According to the VCLT, states have one year from the filing of a reservation to object. Ginsburg (Reference Ginsburg2018, 235) reasons, ‘In many cases, the objector … hopes that the reserving state will modify or withdraw the reservation.’ For example, Australia objected to Pakistan's reservations to the ICCPR, reasoning that the reservations violated the purpose of the treaty, and expressed in its objection ‘the hope that the Islamic Republic of Pakistan will withdraw its reservations.’
Objections provide an initial and visible form of pressure to encourage states to withdraw reservations. If they are not immediately effective in encouraging withdrawal, they set the stage for additional pressure from states, treaty bodies, and NGOs. As Ginsburg argues, objections ‘can be seen as “fire alarm” claims that seek to mobilize other states to follow the particular interpretation that has been advanced’ (2018: 235). In other words, an objection is a diplomatic tool a state uses to signal its discontent, and it can drive states and other actors to similarly signal their dissatisfaction. Returning to the Pakistan example, not only did numerous states object to the country's reservations to the ICCPR, groups like Amnesty International and the International Commission of Jurists publicly urged Pakistan to withdraw its reservations in June 2011 (Amnesty International 2011). Pakistan did so later that year.
While Simmons (Reference Simmons2009) argues that ‘few states take on the policing role’ of objection, Ginsburg (Reference Ginsburg2018) shows objections are more common than previously believed. From the time the treaties in the sample entered into force through 2014, there were 597 incidents of objections.Footnote 10 Objections are a costly action by treaty members, and not every treaty member is willing to bear such costs (Ginsburg Reference Ginsburg2018).Footnote 11 Yet objections are not as costly as other forms of coercion like economic and military sanctions, making objections a valuable mechanism for treaty members to pressure reserving states. And social pressure from the international community is arguably more effective than material inducements or coercion (Goodman and Jinks Reference Goodman and Jinks2013).Footnote 12 International actors seem to be more effective through quiet diplomacy, explaining why it is a frequently-used tool in human rights advocacy (Forsythe Reference Forsythe2017). Objections, while public, are a quieter tool of diplomacy.
When states object to treaty reservations, they publicly provide reasons for their actions. We review all objections to the treaties in the sample and code the reason for objection using a scheme provided by Ginsburg (Reference Ginsburg2018). The reasons for objection are presented in Table 1. The most common reason given for objecting is that the reservation violates the spirit or purpose of the treaty. Other common reasons include the lack of clarity about a reservation and doubts about a state's commitment, the invocation of domestic or religious law contrary to the agreement, and the reservation setting a precedent that can undermine international law. Overall, many objections ‘call out’ reserving states in various ways for a lack of commitment to a treaty. The reasons behind objections illuminate the process by which flexibility in treaty commitments is tested. States make reservations to test the waters and ultimately render themselves less legally obligated to a treaty's terms. But as discussed earlier, there is no specified review process or dedicated legal body that reviews reservations for their validity. Instead, objections are important for determining the validity of reservations. Objections by fellow members draw public attention to what they consider weak or invalid commitments to a treaty.Footnote 13 States may attempt to lessen their treaty commitments, but objections from fellow treaty members seek to bring reserving states back into the fold.Footnote 14
Reasons for objections to reservations across major human rights treaties through 2014
a Countries can give multiple reasons when they lodge objections, so the total number of reasons (1,393) is more than the total incidents of objection (597).
As Ginsburg notes, it is difficult to pinpoint how objections encourage reservation withdrawals, but we speculate on two mechanisms. First, diplomatic pressure could accompany objections. Fellow treaty members may use behind-the-scenes pressure in addition to public objections to persuade reserving states to change their behavior. International law is predicated on states entering agreements in good faith. Objections can cast aspersions on a reserving state, indicating it has not entered an agreement – and does not enter agreements – in good faith. Concerns about being seen as a bad and unreliable partner may lead a state to withdraw its reservations. Withdrawing would help the state maintain a good reputation and preserve its relationship with the objecting state(s).
Second, domestic pressure could accompany objections. States join human rights treaties to signal not only to the international community but also to their domestic populations that they are committed to particular standards of behavior. If fellow states object to a country's reservations, the country's human rights commitments are visibly questioned, if not undermined entirely. Even states that are not committed to human rights want to appear committed (Hafner-Burton and Tsutsui Reference Hafner-Burton and Tsutsui2005). Concerns about negative reactions from domestic audiences may thus compel a state to withdraw its reservations. Lifting reservations would renew the state's good standing at home.
Objections are not the only tool states can use to influence their peers, but they are a form of persuasion that compels reserving states to improve their human rights commitments. The pressure is additive (Ginsburg Reference Ginsburg2018) so that the more states that object, the more visible the reservation becomes, and the more pressure grows on the reserving state to withdraw. This logic motivates the first hypothesis:
H1: States are more likely to withdraw a reservation when they receive a higher number of objections than when they receive a lower number of objections.
Now, which states lodge objections to reservations, and does this facilitate reservation withdrawal? Numerous works in the human rights literature find that the actors most effective in pressuring states to change their treaty behavior and human rights practices are those with influence over the state in question (Goodliffe and Hawkins Reference Goodliffe and Hawkins2009; Greenhill Reference Greenhill2010). We can therefore surmise that the states most likely to lodge objections and encourage their peers to withdraw reservations are those with influence over and relational ties to reserving states. Eldredge and Shannon (Reference Eldredge and Shannon2022) show that the states that most often lodge objections have greater social power in relation to reserving states, measured by shared inter-governmental organization (IGO) memberships between the objecting and reserving state. Moreover, when we look at the states that most often lodge objections across human rights treaties, we see numerous ‘defenders of the faith’ (Terman and Voeten Reference Terman and Voeten2018), with strong human rights records and a history of human rights advocacy. The Netherlands, Sweden, Germany, Finland, and Norway are the top-five most frequent objectors to reservations across the human rights treaties in the sample (see Table 2 for a full list of objecting countries across treaties). Further evidence reveals that countries with stronger domestic human rights bureaucracies are more likely to object to reservations (Eldredge and Shannon Reference Eldredge and Shannon2022; Ginsburg Reference Ginsburg2018). While exploring patterns of state objections to human rights treaty reservations deserves attention in future research, the evidence thus far reveals that states that object are active stewards of human rights or have stronger social relationships with reserving states.
Objections lodged by countries to reservations across human rights treaties through 2014
Does pressure by states with more influence and clout effectively change more reluctant states' behavior within the human rights treaty regime? This is a broad empirical question that has not yet been addressed in the context of treaty reservation withdrawal. Yet some evidence supports the idea that states with more influence are better at applying pressure for change. Eldredge and Shannon (Reference Eldredge and Shannon2022) find for the CAT that as more states object to treaty reservations, more reservations are withdrawn. These findings are bolstered by Terman and Voeten (Reference Terman and Voeten2018), who show that states are more likely to accept recommendations from the UN Human Rights Council's Universal Periodic Review (UPR) when the states making recommendations have a strong history of promoting human rights, or when recommenders have strategic partnerships with the state under review.Footnote 15 It seems that states with clout in the human rights regime are influential, as are those with close relationship ties to reserving states. It is likely that objections from these states are most effective in persuading states to withdraw reservations and improve their treaty participation.
Institutional Pressure via Periodic Review
International human rights treaty bodies or committees are another source of pressure for states to withdraw treaty reservations. Each major human rights treaty has an associated monitoring body that periodically scrutinizes states' self-reports on their human rights practices (Reiners Reference Reiners2021). States report on their policies and indicate whether and how they adhere to the treaty's provisions. After reviewing state reports, human rights treaty bodies make observations and recommendations for how states can further reform their policies and improve their performance. While human rights committees' recommendations are not binding, they constitute a ‘constructive dialogue’ between institutions and states on treaty adherence (O'Flaherty Reference O'Flaherty2006). In their observations, human rights treaty bodies frequently recommend that states reconsider and remove their reservations (Edry Reference Edry2020; O'Flaherty Reference O'Flaherty2006). For example, the CEDAW committee has adopted official recommendations that states withdraw reservations to Articles 2 and 16 in particular (De Pauw Reference De Pauw2013). The Committee on the Rights of the Child also entreats states to withdraw their reservations (Cowell Reference Cowell2021). Moreover, though not associated with a particular treaty, the UPR has issued almost 400 recommendations on reservations to various human rights treaties, most of which encourage states to withdraw reservations (Cowell Reference Cowell2021).
Observations and recommendations from human rights committees increase pressure on states to withdraw reservations. Though the tone and intention of committee reports are not necessarily to condemn states, the high-profile reports illuminate problems in states' human rights practice, and this frequently involves targeting states' reservations. Committee reports are not likely to universally or single-handedly persuade states to withdraw reservations, but they constitute an iterative institutional practice that works to change state behavior through social interaction and pressure from other treaty members (Cowell Reference Cowell2021). For example, after its first review by the CEDAW committee, Mauritania accepted the committee's recommendation to amend or withdraw its recommendation. By its second review, Mauritania was able to report compliance with the recommendation (Cowell Reference Cowell2021, 280).
By drawing attention to a state's reservations, a human rights treaty body puts fellow members of the treaty regime on notice. Treaty members and other actors can then increase pressure on reserving states to withdraw. As with state objections, treaty body criticisms can unravel the credible commitments states promised their peers and their publics. While institutional pressure from treaty bodies may occur at any time, and may develop over months and years, the pressure likely peaks around periodic reviews. For example, during the CRC committee's review of Somalia in 2016, Germany recommended that Somalia withdraw its reservations to the Convention, bolstering its previous objection to Somalia's reservation (Cowell Reference Cowell2021, 280).
Periodic review can also mobilize non-governmental and domestic actors to pressure governments to deepen their human rights commitments (Creamer and Simmons Reference Creamer and Simmons2018; Edry Reference Edry2020). This makes it important to account for domestic forces, as we do in the empirical analyses to follow, but does not negate the fact that international actors can increase pressure on reserving states in light of a periodic review.
While periodic review is an iterative process between a human rights committee and treaty member, again, it is likely that pressure on a state to withdraw its reservations peaks around periodic reviews. Reserving states may withdraw their reservations just before periodic review, if they anticipate being called out by the committee, or soon after, because of pressure from the review and fellow treaty members. This logic motivates the second hypothesis.
H2: States are more likely to withdraw a reservation when undergoing periodic review than when not undergoing periodic review.
Analysis of Treaty Reservation Withdrawal
We evaluate the hypotheses using recent data from Zvobgo, Sandholtz, and Mulesky (Reference Zvobgo, Sandholtz and Mulesky2020) that captures states' use of reservations at the provision level for nine major human rights treaties.Footnote 16 A provision is a treaty-article-paragraph. An example is Article 2, paragraph 1 of the CAT, which says, ‘Each state party shall take effective legislative, administrative, judicial or other measures to prevent acts of torture in any territory under its jurisdiction.’ Paragraphs 2 and 3 of the same article are each considered separate provisions from paragraph 1.
Reservations are often made to particular provisions, and the provision as a unit of analysis allows us to control for provision-level characteristics as well as country-level characteristics.Footnote 17 The data include all instances of states attaching reservations to treaty provisions that obligate states parties, and withdrawing them if such an act occurs, from the point each treaty opens for signature (the earliest year is 1966) through 2014, when the analysis period concludes. We restrict the analysis to treaty provisions that establish obligations for states parties, to focus on the most consequential reserving and withdrawing behaviors (see Zvobgo, Sandholtz, and Mulesky Reference Zvobgo, Sandholtz and Mulesky2020, 786–7). This approach presents a more difficult test of the hypotheses due to the high-stakes nature of the provisions under scrutiny.Footnote 18
As we discuss in more detail later in this section, we model the process in two stages: one where a country decides to enter a reservation and one where it decides to withdraw a reservation. The data-generating process and the first stage begin when all treaty ratifiers are paired with all treaty provisions. The result is a cross-sectional sample where each observation is a country-provision. When a country ratifies, it can make a reservation to a particular provision. If a country does not make a reservation, the first-stage variable Reservation is coded as a 0 and the country-provision does not enter the second stage of analysis.Footnote 19
If a country makes a reservation to a particular provision, the first-stage variable Reservation is coded as a 1 and the country-provision enters the second stage.Footnote 20 International pressure can develop during this stage from other treaty members lodging an objection to the reservation, or by the reserving state undergoing periodic review by human rights treaty bodies. Either of these occur after the reservation is submitted. The second stage in the process captures the reserving country's choice to maintain or withdraw the reservation. A reserving country can choose to withdraw reservations whether or not it receives pressure from objections or periodic review. If the reserving country chooses to withdraw the reservation, it can do so at any time by notifying the UN Secretary-General. The second-stage variable Withdrawal is then coded as a 1 if the reserving state withdraws the reservation, as recorded on each treaty's website.
Our interest in whether states withdraw reservations raises two unique challenges for empirical analysis. First, though human rights treaties have near-universal membership, not all states have ratified the treaties in the sample and are thus not candidates for entering or withdrawing reservations. However, we are confident that the risk of bias arising from this form of selection is minimal. Multiple analysts studying the determinants of reservations fail to uncover bias induced by membership selection (Hill Reference Hill2016; Zvobgo, Sandholtz, and Mulesky Reference Zvobgo, Sandholtz and Mulesky2020). Moreover, by the end of the analysis period, the sample of treaties enjoyed membership from the vast majority of states. Given the small numbers of states that have not ratified or acceded, any bias resulting from membership selection would have to be severe before undermining the analysis.
The second empirical challenge arises from the fact that not all states enter reservations when joining the agreements in the sample. These states are not candidates for withdrawing reservations. We must therefore account for states that select into the subgroup of those eligible for withdrawing a reservation. Hence, our two-stage censored probit model. As explained earlier, the first stage explores the factors that drive states to file reservations when joining a human rights treaty, and the second models the process of withdrawing those reservations. This approach allows us to draw upon the parameter estimates from the first stage to account for the possibility of selection-induced bias in the second (Heckman Reference Heckman1979; Von Stein Reference Von Stein2005).
This empirical strategy requires a suitable instrument for inclusion in the first-stage selection equation. The instrument must affect the dependent variable in the second stage only through its influence on the first stage. If the instrument affects the second-stage dependent variable in ways other than the first-stage outcome, then coefficient estimates of variables of interest in the second stage may be biased. We use a country's legal system as an instrument and, by doing so, we assume that a country's legal system influences the withdrawal of reservations only through the path of making reservations. This assumption is plausible, as we are not aware of any research that asserts that legal systems influence reservation withdrawal independently. Edry (Reference Edry2020) convincingly argues that legal systems influence the withdrawal of reservations through the path of making reservations, and consequently also uses the domestic legal system as an instrument. Simmons (Reference Simmons2009) and McKibben and Western (Reference McKibben and Western2020) also argue that countries with common law systems and Islamic law systems are more likely to make reservations because of the occasional incompatibility of domestic law with treaty law. It is likely that the legal system influences reservation withdrawal in a number of ways, but all of them hinge on the process by which a country makes a reservation. For instance, NGOs often raise concerns about reservations, and this may be more likely to happen in countries where the legal system favors reservations. The process of periodic review also calls attention to reservations, as do objections, contingent on the country making a reservation in the first place. We cannot conceive of another pathway by which legal systems lead to reservation withdrawals other than the pathway that leads to the reservation, and are therefore confident that the legal system as an instrument allows for unbiased coefficient estimates in the second-stage model.
Independent Variables in the Outcome Model: International Pressure to Withdraw Reservations
Our primary interest is to determine whether pressure by fellow treaty members compels reserving countries to withdraw reservations, so we focus on parameter estimates from the second-stage model. The first expectation is that objections by fellow treaty members are strong signals of opposition that can lead states to lift their reservations and improve their treaty commitments. To test this, we gather data on the countries that object to reservations across the agreements in our sample. We include the variable State Objections to reflect the sum of states parties that formally object to a state's reservation, expecting that as this number increases, a reserving state will be more likely to withdraw the relevant reservation.Footnote 21 In many cases, temporal variation exists in the number of states objecting to a particular reservation. The measure captures the highest number of objections targeting any individual reservation during the life of the reservation. The average number of objections for provisions with reservations is 1.8. We expect a positive relationship between State Objections and Withdrawal.
The second independent variable captures the proximity of a periodic review by the relevant human rights committee. Pressure from periodic review can work in an anticipatory manner, where states withdraw their reservations because they expect to be called out in an upcoming periodic review. Or, periodic review can formally entreat states to lift their reservations while galvanizing treaty members to apply additional diplomatic pressure behind the scenes. As discussed, periodic review is an iterative process, and it is possible that the pressure of periodic review operates even in years when there is no formal review. However, we believe that the pressure is most intense and focused when the review begins, making the year of review a suitable proxy for the persuasion of the review process. We include the variable Ongoing Periodic Review, which is coded as a 1 if the country is undergoing periodic review by a treaty committee in the year it withdraws a reservation, or zero otherwise. We expect a positive relationship between Ongoing Periodic Review and Withdrawal.
Additional Independent Variables in the Outcome Model
We control for a range of factors that may make states more susceptible to international pressure to help isolate and determine the effect of objections and periodic review on the decision to withdraw. First, democracies may be more susceptible to international pressure. Leaders of democratic polities are more accountable to domestic actors for their decisions, and may therefore be more likely to respond when their reputations or international ties are threatened. Domestic actors are also better able to mobilize in democracies and supplement existing international pressure to compel states to withdraw (Edry Reference Edry2020). We account for this possible relationship by controlling for a country's political regime type, as measured by its Polity Democracy Score from the Polity IV project. A state's Democracy score can range from −10 to 10, with higher scores reflecting higher levels of democracy. We expect a positive relationship between Democracy and Withdrawal.
Another factor that enhances international pressure is the presence of national human rights institutions (NHRIs). Countries with NHRIs have a more developed human rights bureaucracy to provide information to advocates. This monitoring function may be particularly important for understanding states' reservation and withdrawal behavior, because such actions may not otherwise be visible to the general public. Countries with strong NHRIs are more susceptible to international pressure because domestic and international advocates can more easily observe and target such countries for their reserving behavior. Drawing on Zvobgo, Sandholtz, and Mulesky (Reference Zvobgo, Sandholtz and Mulesky2020, 793) and sourced from Conrad et al. (Reference Conrad2013), we include an index variable, Strong NHRI, that captures the strength of NHRIs within a state. It ranges from 0 to 15, where higher values indicate stronger NHRIs. We expect a positive relationship between Strong NHRI and Withdrawal.
It is possible that treaty members and influential states will target and encourage withdrawal of more important reservations. To the extent that other treaty members want states to be legally obligated by a treaty, it is logical that they would put pressure on states to be bound by more obligatory provisions. We include a variable, Demanding Treaty Provision, that takes a value of 1, where some provision establishes an obligation that is simultaneously strong and precise and requires domestic action for a state, or zero otherwise. Provisions that take a value of 1 contain the highest level of ‘demandingness’ for states parties; these provisions entail the highest compliance costs. An example of a demanding treaty provision is Article 3(1) of the CAT, which prohibits refoulement, enjoining, ‘No State Party shall expel, return (‘refouler’) or extradite a person to another State where there are substantial grounds for believing that he would be in danger of being subjected to torture.’ This provision is strong, precise, and stipulates domestic action. Roughly 40 percent of the provisions in the data are demanding (Zvobgo, Sandholtz, and Mulesky Reference Zvobgo, Sandholtz and Mulesky2020). We expect there will be greater international pressure to withdraw reservations to demanding treaty provisions, and so we account for them to isolate the effects of objections and periodic review.
For another measure capturing a treaty provision's importance, we include a variable, Non-derogable Treaty Provision, that takes a value of 1 if the provision is subject to a non-derogation clause, or zero otherwise. These provisions are not necessarily demanding (that is, strong, precise, and requiring domestic action), but states cannot derogate from them, even in a national emergency, and fellow treaty members expect compliance. An example is Article 6(5) of the ICCPR, which stipulates, ‘Sentence of death shall not be imposed for crimes committed by persons below eighteen years of age and shall not be carried out on pregnant women.’Footnote 22 Again, we expect reservations to non-derogable treaty provisions will receive more international pressure to withdraw.
Independent Variables in the First-Stage Model
In specifying the first-stage selection model that explains states' use of reservations, we largely follow the guidance of Zvobgo, Sandholtz, and Mulesky (Reference Zvobgo, Sandholtz and Mulesky2020). We first include a number of measures of the costs of treaty compliance, as states are more likely to make reservations when the costs of complying with a provision are higher. We include Demanding Treaty Provision and Non-derogable Treaty Provision, expecting that states are driven by higher costs of compliance to make reservations to these provisions. Similarly, states with higher levels of judicial independence may face higher compliance costs because officials within these countries are more likely to be held accountable by domestic courts for failing to comply with some part of the treaty (Hill Reference Hill2016; Powell and Staton Reference Powell and Staton2009). We include a variable, Judicial Independence, from the Varieties of Democracy (V-Dem) Project (Coppedge et al. Reference Coppedge2018). This variable ranges from 0 to 4, with 0 reflecting low levels of judicial independence (that is, the courts always base their decisions on government wishes), and 4 reflecting the highest level of judicial independence (that is, courts never base their decisions on government wishes). We also include a variable indicating countries where treaty law has equal or superior status to domestic law. In these countries, treaty law can be invoked more quickly to punish leaders and other officials who do not comply, and governments have less domestic legal recourse in the face of violations. The variable Treaties Equal or Superior takes a value of 1 if the state's constitution specifies that treaties have equal or superior status to domestic law (Elkins, Ginsburg, and Melton Reference Elkins, Ginsburg and Melton2009), or zero otherwise. We also include GDP per capita from the World Bank's World Development Indicators to measure wealth and state capacity, as well as a measure of the reserving state's Population size. Poorer countries and those with larger populations may have more difficulty with compliance, and may therefore be more likely to make reservations.
We also control for a country's basic human rights practices, expecting that countries with better human rights practices are more likely to fully commit to treaties and are therefore less likely to make reservations. We use Fariss's (Reference Fariss2014) latent measure of respect for basic human rights; Basic Rights Respected takes on higher values where states demonstrate greater levels of protection for basic human rights, and lower values where the opposite is true.
Finally, we include the measure of a state's legal system at time of ratification as both a determinant of a state's decision to file a reservation and the instrument (Simmons Reference Simmons2009). To capture a country's legal system, we draw on the four-part classification proposed by Powell and Mitchell (Reference Powell and Mitchell2007). A state is coded as 1 if it maintained a Common Law system at the time of ratification, 2 for Civil Law, 3 for Islamic Law, or 4 for a Mixed System.Footnote 23
Results
We present coefficient estimates from both the selection and outcome models in Table 3 across a number of model specifications. As we are most interested in making inferences about what determines reservation withdrawal, we limit our discussion to the results from the outcome models. We first present bivariate outcome models with just the independent variables that measure international pressure to ensure that the results of these variables are not induced by suppressor effects from additional variables (Lenz and Sahn Reference Lenz and Sahn2021). The results support the argument that international social pressure increases the likelihood of states withdrawing reservations and improving their treaty commitments. The coefficient estimates for State Objections and Ongoing Periodic Review are both statistically significant and signed positively across all outcome models of Table 3. Figure 1 presents the marginal effect of unit increases in the number of objections on the probability of withdrawal, based on the estimates of Model 1 in Table 3. Figure 2 presents the marginal effects of objections on reservation withdrawal based on estimates from Model 4 in Table 3; this model accounts for periodic review and other potential confounders to the treatment. Both marginal effects plots reveal a positive and statistically significant relationship between objections and the likelihood of observing a reservation withdrawal. The plots also reveal a cumulative effect: as the number of objections increases, so too does the likelihood of a state withdrawing its reservation. The estimated marginal effects in Fig. 2 are not as steep as those depicted in Fig. 1, which is not surprising, given that the model on which Fig. 2 is based is more fully specified.
Two-Stage Analysis of Reservation Withdrawals in Nine Human Rights Treaties, 1966–2014
** = p < 0.05; * = p < 0.10.
Marginal effect of objections on the probability of withdrawing a treaty reservation (based on Model 1 of Table 3).
Marginal effect of objections on the probability of withdrawing a treaty reservation (based on Model 4 of Table 3).
We discuss in more detail Model 4 of Table 3, as it includes the measures of international social pressure (State Objections and Ongoing Periodic Review) in addition to other independent variables that are associated with international pressure (Democracy, Strong NHRI, Demanding Treaty Provision, and Non-derogable Treaty Provision). This model gives us the most confidence that we have isolated the effects of objections and periodic review on the withdrawal of reservations. Even after accounting for factors that may confound the treatment, the estimates indicate that as the number of countries objecting to a reservation increases, states are more likely to withdraw a reservation. As Fig. 2 indicates, each additional objection beyond one is associated with a roughly one-percentage point increase in the likelihood of reservation withdrawal, suggesting a substantively meaningful relationship. Equally notable, the effect size of a one-objection increase on reservation withdrawal eventually begins to plateau, indicating that pressure from only a handful of states in the form of an objection is often sufficient to compel a state to withdraw a reservation. The process of formal periodic review is likewise associated with an increased likelihood of reservation withdrawal. States undergoing periodic review exhibit a roughly 12-percentage point increase in the likelihood of withdrawing reservations in the same year. This substantively meaningful relationship corroborates the argument that scrutiny from formal review motivates a state to abandon its reservations and improve its treaty commitments. Overall, we find strong support for the expectation that increased peer pressure from fellow treaty members and increased institutional pressure from expert committees encourage states to withdraw treaty reservations.
We present a series of additional robustness checks in the appendix. Our results are robust to a range of modeling strategies, including different estimators and alternate approaches to calculating uncertainty. First, we estimate a series of additional two-stage censored probit models that adopt alternate strategies for estimating the error surrounding the coefficient estimates. This includes computing robust standard errors, and calculating errors clustered on country and treaty, respectively. The coefficient estimates for the primary independent variables remain positive and statistically significant across each specification. We also report a series of logit models that forgo a two-stage approach and explain Reservation Withdrawal on a subset of observations that include only provisions that have received reservations. This approach essentially explains withdrawal as a condition of having advanced a reservation previously. We also incorporate treaty fixed effects to account for the possibility of unit heterogeneity at that level, and calculate robust standard errors clustered on country and treaty, respectively. An additional group of tests replicates this approach by estimating a series of linear probability models with errors adjusted for heteroskedasticity. The results from the supplemental analyses corroborate the core findings that increased pressure from treaty members and increased scrutiny in the form of institutional review encourage states to withdraw their treaty reservations.
Limitations
While we find robust support for the hypothesized positive relationships between objections, periodic review, and reservation withdrawal, there are limitations to the analysis and the inferences we make. The first is that we do not directly measure some mechanisms by which peer and institutional pressure may operate. Scholars argue that international organizations and periodic review can induce pressure on states by increasing national media attention on recalcitrant states (Chaudoin Reference Chaudoin2023; Creamer and Simmons Reference Creamer and Simmons2018). Additionally, international pressure may operate when domestic legislative actors draw attention to states' treaty behavior (Creamer and Simmons Reference Creamer and Simmons2018). International pressure may work via increased media coverage and legislative mobilization to compel countries to withdraw reservations. It is difficult to systematically measure media attention and legislative mobilization across countries and in response to various treaties in a large-N analysis such as ours.Footnote 24 However, the variables capturing periodic review, NHRIs, and democracy are good proxies for some of these country-particular mechanisms.
In terms of testing the theoretical mechanisms we propose, we use the number of objections as a measure of social pressure by fellow treaty members. We believe the number of objections is a good proxy for social pressure, but acknowledge that objections do not directly reflect diplomatic communication in the form of phone calls, emails, or meetings between states. Yet, it may be impossible to capture many of these forms of communication that go unreported and are intentionally made behind the scenes. Similarly, states' bureaucracies may have idiosyncratic features that influence their decision to withdraw reservations. We have no reason to believe that these unobserved factors are correlated with the measures of international pressure in such a way as to bias the inferences we make, but we cannot eliminate the possibility.
Another limitation of this research is that it does not explore how changes over time influence countries' decisions to withdraw. The analyses presented here are a first cut in exploring broadly the determinants of reservation withdrawals, and the analyses are valuable for providing inferences across a large sample of countries and treaties. Future work may do more by way of micro-analysis to explore how changes within countries over time influence their decisions to withdraw. For instance, the process of periodic review is iterative and dynamic, and modeling more of the nuances in periodic review may provide additional insight into what leads countries to withdraw reservations. Research might also more specifically measure if the periodic review recommends withdrawal. Qualitative case studies may be especially useful in uncovering the micro-determinants of reservation withdrawals.
Conclusion
From a legal perspective, states do not need to withdraw their reservations to treaties: treaty members are often able to agree that a treaty will enter into effect despite the reservation. And yet we often observe states withdrawing their treaty reservations. What motivates reserving states to change their commitments? In an analysis of reservation withdrawals across nine major human rights treaties, we show that pressure from other states and from human rights institutions compel reserving states to withdraw reservations and improve their commitments. The findings are heartening from a policy perspective, as they indicate that state-to-state and institutional diplomacy is powerful in strengthening states' commitments to human rights law. As a number of NGOs seek to persuade states to withdraw treaty reservations, this research suggests that an effective strategy may be for NGOs to partner with committed treaty member states to put pressure on reserving states. Treaty bodies like the Human Rights Committee, which frequently recommend states withdraw reservations (O'Flaherty Reference O'Flaherty2006), can work with state partners to achieve their goals. NGOs and states may want to do more to monitor the politics of treaty ratification and file objections when reservations call into question a state's treaty commitment.
As a direction for future research, scholars should consider whether and how state-to-state and institutional pressure encourages reluctant states to improve their legal and treaty commitments, both within and beyond the human rights regime. Scholars might compare different issue regimes (Pelc Reference Pelc2018) to understand if diplomatic pressure encourages members to rescind their efforts at flexibility and strengthen their treaty participation. Research can also explore whether and how periodic review influences states to change their behavior in other legal regimes. For instance, the Paris Agreement asks states to submit reports on their progress in meeting nationally-determined contributions to mitigating climate change (Edry Reference Edry2020). Scholars might investigate if the process of review galvanizes domestic, non-governmental, and peer state actors to pressure states to improve their environmental treaty commitments.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0007123423000339.
Data availability statement
Replication Data for this article can be found in Harvard Dataverse at: https://doi.org/10.7910/DVN/V4PO0T.
Acknowledgements
Our thanks to workshop participants at Arizona State University, Centro de Investigación y Docencia Económicas, Concordia University, Harvard University, the University of Chicago, and the Southern Political Science Association for helpful feedback on earlier drafts, with special thanks to Felix Bethke, Stephen Chaudoin, Christoph Mikulaschek, Krzysztof Pelc, Ryan Welch, and the editors and anonymous reviewers at the British Journal of Political Science. We are also grateful to Wayne Sandholtz for previous research collaboration. Christopher Jackson, a graduate student at the University of Colorado Boulder, provided excellent research assistance.
Author contributions
Authors are listed alphabetically by last name.
Financial support
This work was supported by a fellowship from the University of Southern California (Provost Fellowship in the Social Sciences). In addition, this material is based upon work supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. DGE-1418060. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.
Competing interests
None.
Open access
