Keeping out of harm’s way? Constitutional due process and state repression

Abstract

The question concerning a state leader's decision to employ repression against domestic audiences extends beyond a binary choice. Leaders act strategically in choosing among a variety of repressive methods through a cost-benefit analysis largely predicated on whether the state will be held responsible for its actions. The divergent costs of repression, particularly physical integrity repression, vary across two dimensions: whether a particular form of repression is custodial or non-custodial, and whether legal means exist to hold leaders accountable. Non-custodial repression, which occurs in contexts outside direct state detention, is harder to link to the state. Furthermore, a state can employ non-custodial repression while maintaining the façade of abiding by constitutional expectations, particularly promises of pre-trial, due process. I use propensity score matching to identify four separate samples in which states vary on the presence or absence of four types of constitutional provisions: pre-trial release, habeas corpus actions, speedy trials, and the right to counsel. The empirical results support the theoretical expectations that the existence of pre-trial, due process provisions in a state's constitution will increase the likelihood of non-custodial repression (disappearances and extrajudicial killings) and decrease the probability of a state employing custodial repression (political imprisonments).

Notes

Acknowledgments

I would like to thank Jessica Maves Braithwaite, Faten Ghosn, and Chad Westerland for their guidance and feedback, which improved this article at every stage. I also wish to thank Alex Braithwaite, the editors, and the anonymous reviewers for their helpful recommendations.

Notes

1 The following factors have also been identified as correlates of state repression: economic conditions, population levels, conflict, dissent, leadership insecurities, international institutions, civil society, and nongovernmental organizations (Poe and Tate 1994; Fein 1995; Poe et al. 1999; Moore 2000; Davenport and Armstrong 2004; Neumayer 2005; Carey 2006; Davenport 2007; Powell and Staton 2009; Simmons 2009; Murdie and Bhasin 2011; Hollyer and Rosendorff 2011; Mitchell and Powell 2011; Bell, Clay, and Murdie 2012; Meernik, Aloisi, Sowell, and Nichols 2012; Hafner-Burton 2012, 2014; DeMeritt and Young 2013; Nordas and Davenport 2013; Danneman and Ritter 2014; Hill and Jones 2014; Lupu 2015; Richards, Webb, and Clay 2015).

2 The costs linked to state repression can take a variety of forms. In the domestic sphere, repression can lead to further dissent (Moore 2000; Carey 2006; Ritter 2014) or revolution (Young 2013), either of which can lead to the ouster of the state’s leadership. Constitutional violations may incur judicial rebukes, damaging the credibility of the regime (Mitchell et al. 2013), or set the precedent for future human rights prosecutions (Kim and Sikkink 2010; Sikkink 2011). Although the adoption of constitutional norms may increase a state’s legitimacy, the violation of constitutional expectations creates reputation costs, thereby potentially incurring the imposition of sanctions, the removal of foreign direct investment, or the cessation of foreign aid (Murdie and Bhasin 2011; Meernik et al. 2012; Murdie and Peksen 2013; Albertus and Menaldo 2014).

3 Generally, the literature includes within the ambit of physical integrity repression the act of torture (Cingranelli and Richards 1999, 2010). Although torture would likely constitute a custodial form of repression, the ability of states to use torture without leaving physical marks makes it more difficult to link state action with torture (Conrad and Moore 2010; Conrad 2014; Payne and Abouharb 2016). Given the variation in the forms of torture states may leverage if inclined toward repressive strategies, it becomes difficult to isolate a specific cost associated with “torture” writ large, in order to analyze how constitutional due process would impact a state’s willingness to use torture relative to other repressive tactics. Given these differences, I focus only on disappearances, extrajudicial killings, and political imprisonments; however, Table 17 in the Online Appendix presents results from a series of models analyzing the effects of pretrial, due process provisions on torture, which reflect a negative relationship, and is accompanied by a brief discussion of an explanation for the results.

4 Clear, observable periods of imprisonment may not be the only method of using state power to incarcerate individuals with the intent of reducing or limiting antiregime behavior; rather, states may engage in clandestine detentions. Because clandestine detentions potentially obscure the link between the state and this form of repression, clandestine detentions are likely to be a noncustodial form of repression. Following Cingranelli et al. (2014), I treat clandestine detentions as disappearances; thus, clandestine detentions do not disturb the underlying assumption that political imprisonments constitute a custodial form of repression.

5 International and domestic actors may perceive disappearances and extrajudicial killings to be more heinous acts relative to political imprisonments, which would suggest that the costs of the former outweigh those of the latter. However, the imposition of costs on a repressive actor is preceded by an evidentiary chain that can link the actor to the acts of repression. If the state cannot be easily linked to repressive activity, it becomes difficult for international and domestic forces to impose costs on the state. Therefore, the custodial nature of political imprisonments can increase the costs associated with this type of repression relative to disappearances and extrajudicial killings, which can be carried out through clandestine means.

6 The “psmatch2” package in Stata is used to generate the matched samples (Leuven and Sianesi 2003). The Online Appendix includes descriptive data for the treatment variables and control variables for the prematched and postmatched samples.

7 In order to test the robustness of the results in Tables 1, 2, and 3, I ran a series of ordered logit models using the preprocessed samples, and the results are generally consistent with the empirics in the primary analyses. The results from the ordered logit analyses can be found in Tables 18 through 20 in the Online Appendix.

Additional information

Notes on contributors

Joseph M. Cox

Joseph M. Cox is a PhD candidate in the School of Government and Public Policy at the University of Arizona. His research interests include the effects of legal organizations and institutions on conflict, human rights, and state behavior.