Repression Substitution: Shifting Human Rights Violations in Response to UN Naming and Shaming

ABSTRACT

Does the United Nations naming and shaming of specific violations of human rights decrease government repression? In this article, we argue that international shaming of specific human rights violations can weaken the target government, bringing new challenges and making the government cessation of repression less feasible. When international naming and shaming campaigns target specific repressive tactics, they increase the costs of some – but not all – means of repression. Using original data on naming and shaming by the United Nations Human Rights Council (HRC), we show that the shaming of one physical integrity violation is jointly associated with decreases in that violation and increases in other violations of human rights.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed here.

Notes

1. For tractability, we focus in this article on the effect of international naming and shaming campaigns on one set of repressive tactics: personal or physical integrity violations, which we define below.

2. See also Poe and Tate (1994, note 6).

3. We focus on UN naming and shaming because the Cingranelli et al. (2014) data that we use to measure our dependent variables draws on naming and shaming by the United States State Department and Amnesty International (AI) to generate measures of rights violations.

4. The argument that the costs of repression vary on an international dimension is consistent with Risse et al. (2013), who assume that the costs of repression vary internationally. In instances where the leaders do not directly order repression, we assume that they pass the costs on to repressive agents (Mitchell 2004; Conrad and Moore, 2010, DeMeritt 2014).

5. See Spilker and Bohmelt (2013) for an alternative explanation for this empirical relationship.

6. Franklin (2008) focuses on physical integrity violations. Although Hafner-Burton (2008) distinguishes between physical integrity violations and empowerment rights violations, she does not disaggregate to the individual repressive tactic.

7. A notable exception that does not focus on international naming shaming as an independent variable is the analysis of the co-occurance and substitution patterns of human rights practices by Fariss and Schnakenberg (2014).

8. We use the term interchangeably with physical integrity violations.

9. These rights can be contrasted with civil liberties restrictions, which involve ‘state or state-affiliated limitations, such as arrests, banning, and curfews, being placed on expression, association, assembly, and beliefs.’ We focus on personal integrity violations for two reasons. First, violations of personal integrity rights ‘offend the most widely-shared norms of appropriate government conduct’ (Walsh and Piazza 2010, 552). Second, we consider physical security to be fundamental in securing civil rights; without freedom from personal integrity violations, citizens are unlikely to mobilize in favor of political and civil rights. It is difficult to demand the right to vote before one’s right to life is secured. We have no theoretical reason to believe that our theory should not apply to other types of human rights violations, or to the substitution of repression with accommodation.

10. In the worst cases of civil and international war, governments engage in a brutal combination of these tactics (e.g. Rasler 1986, Cingranelli and Richards 1999, Wood and Gibney 2010).

11. For more on the under provision of goods because of free-riding, see Olson (1971).

12. These dynamics are similar to those described by Wood (2008) in his discussion of the effect of sanctions on government respect for human rights. Admittedly, few leaders are likely to view being shamed for government repression and facing economic sanctions as similar in terms of their domestic consequences; sanctions are argued to impose severe economic and social costs that fall directly on civilians (e.g. Cortright and Lopez 2002). Although the direct economic and social costs of shaming to civilians are less obvious, naming and shaming nonetheless increases the leader’s perception of threat by increasing the salience of dissatisfaction with the regime and creating focal points for domestic and international mobilization in favor of human rights.

13. For more on the extent to which formal or written standards create expectations and coordinate behavior, see Weingast (1997); Carey (2000).

14. In May 2017, local media in the Philippines reported on the fact that the country had been condemned by the HRC for extrajudicial killings (de la Cruz 2017); the government of the Philippines thought the allegations important enough to counter both at the UN and at home (de la Cruz 2017).

15. Murdie and Davis (2012), for example, argue that naming and shaming only decreases violations of human rights in states with a thriving domestic NGO community.

16. The idea that actors switch or substitute tactics is not new. Lichbach (1987) argues dissidents using violent tactics that are repressed by thr government will switch to nonviolence, and vice versa. Empirical evidence supports that expectation (Francisco 1996, Moore 1998). Like Lichbach (1987), we are interested in how actors choose among available policies in the pursuit of a single goal: we argue that leaders can select from a myriad of repressive techniques to quell their political opposition. Our work differs from that of Lichbach (1987) in that we seek to explain government choices among repressive tactics rather than dissident responses to repression.

17. In 2006, the UNCHR handed human rights advocacy responsibilities to the newly-formed UNHRC. Because of the transition, the organization(s) issued an unusually low number of shaming resolutions during that year.

18. Governments may substitute toward non-physical integrity violations when they are shamed. To make our empirical model tractable, we focus here on physical integrity violations.

19. Naming and shaming is a non-random treatment. If the CHR and HRC only shame states with a certain type of human rights record, then all countries in the UN do not face an equal chance of facing criticism. On average, states are more likely to be criticized by the CHR and the HRC as their human rights worsen (Lebovic and Voeten 2006). Because states with worse human rights records are also the least likely to change their behavior, the selection effect biases our coefficients toward null findings.

20. One might suspect that states are not often shamed for one violation in isolation from others. While shaming for one form of abuse in our data often comes alongside shaming for others, this is not always the case. Seventeen percent of country-years that include shaming for at least one physical integrity abuse include shaming for only one form of abuse.

21. Although we make an argument about actual government behaviors in response to shaming, we are cognizant of the fact that the data we use to assess this argument are designed to capture ‘reporting’ of human rights abuse and not actual ‘levels’ of abuse (e.g. Conrad et al., 2013; Fariss 2014). We are also attentive to the possibility of changing standards of accountability and corresponding changes to the reporting of repression over time (Fariss 2014). We expect that our time variables absorb at least some of the measurement error introduced by these changes. We prefer these controls to the Fariss (2014) latent repression measure interacted with time because that control is based, in part, on the same CIRI data that inform our dependent variables.

22. These two cases are also different in interesting ways to which we return in the conclusion.

23. Results using a binary construction are included in our supplemental appendix and discussed in the conclusion below.

24. We are sensitive to the argument by Bellemar, Masaki, and Pepinsky (Bellemare et al. 2017) that lagging explanatory variables merely replaces a ‘selection on observables’ assumption with an equally untestable ‘no dynamics among unobservables’ assumption. We include this lag primarily for theoretical reasons. Nonetheless, we hesitate to argue that these results are causal; instead, our results provide interesting evidence of a correlation between naming and shaming and the substitution of rights violations.

25. Much of this work disaggregates democracy (e.g. Voice, Veto, etc.). Since the focus of this paper is naming and shaming, and we engage democracy only as a control, we instead focus on democracy as a general concept.

26. This combination of controls produces the best model fit, but we remain sensitive to the fact that state wealth tends to relate to state coercion (Rodwan and Cingranelli 2006; Poe and Tate 1994, Poe et al. 1999, Ziegenhagen 1986). In the supplemental appendix, we show that our results are robust to controlling for the natural log of gross domestic product (GDP) using data from the World Bank. We also show their robustness to controlling for the level of each form of physical integrity abuse.

27. For descriptive statistics, please refer to our supplemental appendix.

28. Our control variables are in line with theoretical expectations.

29. These percentages sum to 131% because states substitute by reducing one abusive tactic and introducing up to three new forms.

30. By using probit in our empirical models, we assume that substitution vis-à-vis each tactic is independent of substitution vis-à-vis all the others. But perhaps the process that produces reduced killing and increased torture are related to the process that produces reduced disappearances and increased imprisonment. There may be benefits in modelling these processes as (both theoretically and empirically) conditional rather than fully independent.

Additional information

Notes on contributors

Jacqueline H.R. DeMeritt

Jacqueline H.R. DeMeritt is Associate Professor in the Department of Political Science at the University of North Texas. Her research and teaching interests center on reasons and remedies for repression, and especially on opportunities for international actors to protect vulnerable individuals. Other interests include the applications of quantitative methodology and formal theory to social scientific processes.

Courtenay R. Conrad

Courtenay R. Conrad is Associate Professor in the Department of Political Science at the University of California, Merced. Her research and teaching interests focus primarily on political violence and human rights, particularly on how repressive agents make decisions in the face of domestic and international institutional constraints.