Preferences or blocs? Voting in the United Nations Human Rights Council

“Apologise or we’ll cut your funding, US envoy tells the UN” The Times, June 9, 2006 (see also Rajagopal 2007).

Brett D. Schaefer “Elections for U.N. Human Rights Council Underscore the Need for Reform” Backgrounder. No. 2417 June 2, 2010. Published by the Heritage Foundation and “Le Conseil des Droits de l’Homme sera revu” Le Temps Septembre 6, 2010 http://​emploi.​letemps.​ch/​Facet/​print/​Uuid/​9018569c-b9bf-11df-9ea7-6791aedb1a60/​Le_​Conseil_​des_​Droits_​de_​lHomme_​sera_​revu.

Partial exceptions are the studies by Cox (2010), Seligman (2011) and McMahon (2012), focusing on particular aspects of the new body.

Other studies focus on international bodies like the European parliament (e.g., Attina 1990; Brzinski 1995; Hix et al. 2006), the assembly of the International Labor Organization (e.g., Boockmann 2006, 2003) or the United Nations Security Council (UNSC) (e.g., O’Neill 1996; Dreher et al. 2009).

Editorial “The Shame of the UN” The New York Times, February 26, 2006. http://​www.​nytimes.​com/​2006/​02/​26/​opinion/​26sun2.​html.

UN Special Report 2005. In Larger Freedom Addendum http://​www.​un.​org/​largerfreedom/​add1.​htm.

UN Special Report 2005. In Larger Freedom http://​daccess-dds-ny.​un.​org/​doc/​UNDOC/​GEN/​N05/​270/​78/​PDF/​N0527078.​pdf?​OpenElement, p. 45.

UN Special Report 2005. In Larger Freedom Addendum, 5 http://​www.​un.​org/​largerfreedom/​add1.​htm.

“Sudan’s U.N. post provokes anger” Seattle Times, May 5, 2004 http://​seattletimes.​nwsource.​com/​html/​nationworld/​2001920167_​sudan05.​html. See also the position of Human Rights Watch on this issue in the Press Release Democracy Coalition Project 2004. http://www.democracycaucus.net/pdf/ undc_press_release_may04.pdf.

UN Special Report 2005. In Larger Freedom Addendum http://​www.​un.​org/​largerfreedom/​add1.​htm, p. 45.

See also Brett D. Schaefer “Elections for U.N. Human Rights Council Underscore the Need for Reform” Backgrounder. No. 2417 June 2, 2010. Published by the Heritage Foundation.

Resolution UN/A/RES/60/251 http://​www2.​ohchr.​org/​english/​bodies/​hrcouncil/​docs/​A.​RES.​60.​251_​En.​pdf.

http://​www2.​ohchr.​org/​english/​bodies/​hrcouncil.

United Nations 2006. A/Res/60/251 http://​www2.​ohchr.​org/​english/​bodies/​hrcouncil/​docs/​A.​RES.​60.​251_​En.​pdf.

In order to have systematic rotation within the Council, certain countries were elected only for one or two year terms in 2006. This has had the effect that by April 2010 64 UN member states had been elected to the Council. Countries are elected depending on geographical distribution, the so-called UN regional groups. These groups are the following (numbers of members in the Council in parentheses): Group of African States (13); Group of Asian States (13); Group of Eastern European States (6); Group of Latin American and Caribbean States (8); Group of Western European and other States (7).

Strictly speaking, these selection mechanisms are likely to influence the working of the UNHRC, and thus also its voting process. We refrain from addressing this issue in this paper directly for three reasons. First, as we focus in our analysis on how characteristics of UNHRC members influence their voting records in controversial resolutions, and the latter can be introduced even by nonmembers, we believe that taking into account the membership selection would not affect our analyses. Second, our analyses do not pretend to offer more general insights about decisions on human rights in international bodies. We only wish to show what influences the voting decisions of members of the UNHRC. Third, as our analyses indirectly allow for whether the politicization in the UNHRC is lower than in the UNCHR, the fact that membership rules have only slightly changed allows us to keep these constant. This even more so as the election of members proposed by regional groups is almost systematically endorsed by the UNGA.

http://​www.​ohchr.​org/​EN/​HRBodies/​HRC/​Pages/​AboutCouncil.​aspx.

One could maybe argue that in the 1950s and 1960s when referred to the Arabs, the Muslims were also included.

See also the study by Marin-Bosch (1987).

We consider this problem to be akin to the issue in research on parliaments whether members of the same party vote together because they have the same political preferences or because belonging to a particular party influences their voting record (see most notably Krehbiel 1993).

Relatedly McMahon (2012) offers a systematic analysis of the “universal periodic review.”

The OIC serves as “collective voice of the Muslim world and ensuring to safeguard and protect the interests of the Muslim world in the spirit of promoting international peace and harmony among various people of the world” (Source: http://​www.​oic-oci.​org/​page_​detail.​asp?​p_​id=​52).

Simmons (2009, 83f) finds that Muslim countries do not strongly differ from those of other cultures when it comes to ratifying human rights treaties, except those relating to women’s rights.

See for a related, more theoretically grounded argument Hillman and Potrafke (2011).

We do not go beyond this partial fix of the problem, as to our knowledge, there does not exist a selection model that deals with the empirical approach we need to adopt here (see below).

We refrain from comparing explicitly voting in the UNHRC with voting in its predecessor (the UNCRH), as despite some similarities (see above) they still differ in terms of membership and procedure (for an attempt in such a comparison, see Hug (2013)). For all documents concerning the voting see the UNHRC website.

UNHRC 2010. http://​www2.​ohchr.​org/​english/​bodies/​hrcouncil/​.

The remaining 73 resolutions were submitted by several countries or in the name of a group, and thus we omit these from Table 1. Glancing at these resolutions suggests that including them would not alter the picture presented in Table 1.

See the “Political Terror Scale” (Wood and Gibney 2010) for a similar assessment.

We refrain from calculating averages per membership session as resolutions can also be introduced by non members and the president.

The EU countries are the following (with the number of resolutions in parentheses); Finland (1), Germany (2), Poland (1) Russia (1) Slovenia (1) Portugal (1), and Spain (1). The remaining countries are the following (with the number of resolutions in parenthesis); Algeria (4), Burkina Faso (1), Canada (1), Japan (1), Nicaragua (1) and South Africa (2).

We determined the EU’s position as being the modal response among “yea” and “nay” votes. Only if all EU member countries did abstain did we consider the vote as characterized by EU’s abstention.

We exclude from our analysis six resolutions adopted with no opposing votes. The six resolutions are the following ones: “Elimination of intolerance and discrimination based on religion or belief,” “Role of good governance in the promotion and protection of Human Rights,” “Mandate of special rapporteur on the promotion and protection of the rights of freedom and expression,” “Torture and cruel treatment: the role of medical personnel,” “Situation of human rights in the Democratic Republic of Congo,” and “Right to development.” See the online appendix on this journal’s webpage for the full list of votes.

For South Korea the 2008 coding is missing in this data. We coded this country as democracy. In a robustness check we also relied on the Polity indicator (Marshall et al. 2002).

She covers the following treaties: 1987 Convention Against Torture and Other Cruel, Inhuman or Degrading Treatment (CAT), 1976 International Covenant on Civil and Political Rights (ICCPR), 1981 Convention on the Elimination of All Forms of Discrimination Against Women (CEDAW), Optional Protocol to the ICCPR, and Articles 21 and 22 to the CAT.

Anonymous reviewers suggested these control variables, and we gratefully acknowledge these useful suggestions.

See, however, Heckman and Snyder’s (1997) proposal how a simple linear probability model might be underlying a factor analytic estimation procedure (though see Poole 2005).

Carroll et al. (2009) offer an empirical evaluation of these different estimators.

These various techniques are discussed in detail in Poole (2005).

See for a similar approach to address a different problem Lauderdale’s (2010) hierarchical item-response theory model.

Proceeding in a more traditional way by estimating in a panel framework a model trying to explain individual voting decisions would not allow us to take into account both preferences and bloc membership in addition to information on who proposed the resolution voted upon.

Estimating these effects together with effects of blocs and preferences on voting requires an IRT-model as used here. Using a more traditional type of regression model would require us to manually recode all votes to ensure that our independent variables have the same directional effect in all votes. Our IRT-model directly estimates the relevant “direction,” allows for estimating all the relevant parameters in one single model and in addition imputes the missing data for the dependent variable (i.e., the votes).

In an earlier version of this paper we also relied for the simpler models on Malecki’s (2008) implementation of hierarchical IRT-models in MCMCpack (Martin and Quinn 2004).

See the online appendix on this journal’s webpage for the full list of votes considered.

In the online appendix we depict some convergence diagnostics for all parameters estimated in the model. Convergence for this model is more problematic even after 50000 burn-in iterations. Nevertheless we report estimates that characterize the 5000 iterations thinned by 5.

We refrain from depicting the results graphically as some of the credible intervals are very large and would distort a figure. In the online appendix we depict again some convergence diagnostics for all parameters estimated in the model. Most parameter distributions seem to have converged but some, especially those based on few data points, hardly converged after the 1000000 burn-in iterations. The reported estimates characterize 5000 iterations thinned by 5 after the burn-in iterations.

This was suggested by reviewers, and we gratefully acknowledge this useful suggestion.

For these two variables we have two cases with missing data. For the “Political Terror Scale” based on US State Department reports the value for the US is missing. As the score based on Amnesty International reports equals 2, we used this value. Similarly, according to the Polity website the Polity2 value for Bosnia Herzegovina is equal to 5, so we used this value.

Hug (2013) can show that in the last ten sessions of the UNCHR Cuba was the second most frequent proposer of resolutions, following the United States (the latter country not being a member of the UNHRC at its beginning).