Divergent Compliance with Art. 3 ECHR Judgements: Explaining the factors leading to compliance and non-compliance in ECHR member states

(1)

DIVERGENT COMPLIANCE WITH

ART.3 ECHR JUDGMENTS

Explaining the factors leading to compliance and non-compliance in ECHR

member states

Leiden University – Faculty for Global Governance & Affairs

Heleen van Weers – S2283360 Master Thesis – PA IEG – Capstone:

ADCRIMIN Date: May 28, 2020

(2)

2

Preface

The thesis: ‘’DIVERGENT COMPLIANCE WITH ART.3 ECHR JUDGMENTS: Explaining the factors leading to compliance and non-compliance in ECHR member states’’ is laying in front of you. It has been written as component of the Master Public Administration with the specialisation ‘International and European Governance’ at Leiden University in The Hague, the Netherlands. This thesis is part of the capstone ‘’The Administration of Criminal Injustice (ADCRIMIN)’’ (Leiden University, 2019). The research process lasted from October 2019 until May 2020.

The research process was supervised by Dr. Andrei Poama. The start of the research process was difficult, but after months of conducting research, I gained more trust in delivering this thesis. Fortunately, Dr. Andrei Poama was willing to answer my questions along the way. Furthermore, during the kick off with my co-students, I gained significant feedback from both my supervisor and co-students on how to proceed my research.

I would like to thank my thesis supervisor, Dr. Andrei Poama, for guiding me through this process. Moreover, I would like to thank Dr. Suzuki for his advice on the research methods. Additionally, as I reached out to Prof. Beth Simmons and the Center for Systemic Peace for additional information concerning methods and missing data, I wish to thank to them as well. Lastly, I would like to thank my family and friends for their continuous support.

Heleen van Weers

(3)

3

Abstract

In the last decades multilateral treaties regarding human rights have gained significance. One of those treaties is the European Convention on Human Rights (ECHR) including art.3 describing the prohibition on torture and inhuman or degrading treatment (ECHR & Council of Europe, 2010, p.7). This article holds exclusive characteristics as it is an absolute right which should be implemented fully and by all member states equally. If states fail to uphold the requirements of this article, members of the civil society who think that they are victimised by states could file a lawsuit against the state. The European Court of Human Rights (ECtHR) could then launch a judgment to which states should comply. Nonetheless, the degree of compliance differs among states. The question that arises is how these differences in compliances come about. Therefore, the main research question of this thesis is:

Which domestic factors explain compliance and non-compliance with ECtHR judgments concerning art.3 of the ECHR among the ECHR member states?

On the basis of Beth Simmons’ compliance theory (2009), this thesis studies whether two domestic characteristics, a state’s regime type and the level of respect for the rule-of-law, influence the degree of compliance to art.3 ECHR judgments. Consequently, the following hypotheses are tested:

H1: States with partial or transitional democracies are more likely to comply with art.3 ECHR judgments than stable democracies or stable autocracies.

H2: States with moderate respect for the rule-of-law are more likely to comply with art.3 ECHR judgments than states with strong or weak respect for the rule-of-law.

Furthermore, these variables are being controlled for by a state’s freedom of the press, degree of federalism, level of violent conflict and GDP/capita. By means of a multiple linear regression, the influence of the independent variables as well as the control variables is measured.

The research shows that both a state’s regime type and level of respect for the rule-of-law explain for approximately a quarter the variance in compliance with art.3 ECHR judgments among the member states. The basic model and the controlled model result in different outcomes. The basic model confirms the hypothesis that states with moderate respect for the rule-of-law would account for higher compliance rates than states with strong respect for the rule-of-law. The controlled model confirms that partial/transitional democracies are more likely to comply with art.3 ECHR judgments than stable democracies. The control variables accounted for another quarter of the variance. Especially the freedom of the press and the degree of federalism seem to affect these differences. Although this

(4)

4 research to some degree either confirms or rejects the hypotheses and the compliance theory of Beth Simmons, some questions remain unanswered, for example about the effect of non-democracies or states with weak respect for the rule-of-law on art.3 ECHR compliance rates. Therefore, further research including a larger number of cases as well as other control variables is recommended in order to provide further answers as well as to increase statistical significance.

(5)

5

List of figures & tables

Tables

Table 1: Descriptive statistics independent variables 31

Table 2: Model summary 36

Table 3: Regression statistics 38

Annex I – Table 1: Categorisation/Data from dependent variable, 55 independent variables & control variables

Table 2: Numerical values of categories ‘Regime type’ 56 Table 3: Numerical values of categories ‘Respect for the rule-of-law’ 56 Table 4: Numerical values of categories ‘Freedom of the press’ 56 Table 5: Numerical values of categories ‘Degree of federalism’ 56 Table 6: Numerical values of categories ‘Level of violent conflict’ 56

Annex II - Table 1: Dummy coding scheme 57

Table 2: Abbreviations dummy coding 57

Table 3: Dummy labels ‘Regime type’ 58

Table 4: Dummy labels ‘Respect for the rule-of-law’ 58

Table 5: Dummy labels ‘Freedom of the press’ 58

Table 6: Dummy labels ‘Degree of federalism’ 58

Table 7: Dummy labels ‘Level of violent conflict’ 58

Annex III – Table 1: Pearson Correlation 59

Table 2: Regression output of 2 multiple linear regression models 60 Table 3: Tests of Between-Subjects Effects for ‘Regime type’ 61 Table 4: Tests of Between-Subjects Effects for ‘Respect for the rule-of-law’ 61

(8)

8 Graphs

Graph 1: Number & distribution of ECHR judgments 16

Graph 2: Type of art.3 ECHR judgments 16

Graph 3: Percentage of repetitive art.3 ECHR judgments as part of the 18 total number of art.3 ECHR judgments/state

Graph 4: Histogram ‘Regime type’ 32

Graph 5: Histogram ‘Respect for the rule-of-law’ 32

Graph 6: Histogram ‘Freedom of the press’ 32

Graph 7: Histogram ‘Degree of federalism’ 32

Graph 8: Histogram ‘Level of violent conflict’ 33

Graph 9: Histogram ‘GDP/capita in US Dollars’ 33

Graph 10: Normal P-P Plot 34

Graph 11: Scatterplot 34

(9)

9

List of abbreviations

CAT = Committee Against Torture

ECHR = European Convention on Human Rights ECtHR = European Court of Human Rights EU = European Union

GDP = Gross Domestic Product GDP = Global Peace Index H0 = Null hypothesis H1 = Hypothesis 1 H2 = Hypothesis 2

HUDOC = Human Rights Documentation IEP = Institute for Economics and Peace NGO = Non-Governmental Organisation

Normal P-P Plot = Normal Probability-Probability Plot P-value = Probability value

RESID = Unstandardized residuals

SPSS = Statistical Package for Social Sciences U.S. = United States

(10)

10

Introduction

State compliance with international agreements has become a corner stone of the state’s foreign policy. In former times, bilateral agreements were the dominant method to achieve political compliance. In the contemporary, interdependent world, multilateral treaties underlining the importance of regional or global cooperation to solve complexities in multiple policy areas are significant (Chayes & Chayes, 1995). One of those policy areas is human rights and one corresponding multilateral treaty is the European Convention on Human Rights (ECHR). The ECHR was established in November 1950 in order to secure human rights of all EU citizens and was complemented by its own court, the European Court of Human Rights (ECtHR) in 1959. It is generally assumed that the ECHR and the ECtHR together could have the legitimate power to account for significant human rights developments among its member states on the basis of ECtHR judgments, however, this is not always the case in all states (Council of Europe, 2016).

Art.3 of the ECHR, describing the prohibition on torture and inhuman or degrading treatment, is quite unique as it is an absolute right. This implies that there are no exceptions in which these actions will be justified and that it is stricter maintained than other articles (Council of Europe, n.d.). However, this does not mean that states automatically obey this rule. Firstly, the ECtHR is still dealing with art.3 ECHR infringements by states and secondly, if the ECtHR executes a judgment it does not mean that states comply with the judgment. This research is focused on compliance with ECtHR judgments. Previous research concluded that judgments from the ECtHR concerning art.3 ECHR did neither cause a decline in the number of infringements on torture nor on degrading or inhuman treatment. The difference between torture and degrading or inhuman treatment is the severity of the pain or suffering caused as well as the intentionality behind actions (Hafner-Burton, 2008, p.701). This finding was supported by evidence from the ECtHR stating that several states are facing structural problems with repeated violations of art.3 ECHR (European Court of Human Rights, 2019, pp.36-37). However, there additionally are states who complied with art.3 ECHR judgments in domestic law (Council of Europe, 2019b, 2019c). This means that compliance with art.3 ECHR judgments differs among the ECHR member states.

National compliance with judicial rulings has not been widely studied. On the European level, the EU did execute research on the enforcement of agreements and rulings rather than on compliance by the assigned states (Haas, 1998). In the last decade this was still not the case as research into this topic has been conducted by a few scholars solely (Zhelyazkova, Kaya & Schrama, 2016). Research in the area of human rights shifted from theoretical opposing approaches towards examining instruments combining multiple theoretical approaches (Simmons, 2009, p.275). One of the examining instruments was

(11)

11 designed by Beth Simmons. Beth Simmons conducted research into factors leading to national compliance with international agreements and international judicial rulings, on the basis of Oona Hathaway’s theory (2002) on treaty ratification. The result of Hathaway’s research was that domestic mobilisation could explain divergent compliance with human rights treaties between countries. Simmons (2009) specified this further by focusing on compliance with anti-torture treaties and judicial rulings specifically. Two factors within domestic mobilisation are of great importance here. Firstly, the type of political regime and secondly, the rule-of-law.

In the case of art.3 ECHR, the violations themselves have been widely studied while compliance with ECtHR judgments resulting from those violations was left behind (Varju, 2014). Although research into compliance with anti-torture agreements was studied by scholars including Simmons, it was mainly focused on compliance with measures from the Committee Against Torture (CAT) and other globally oriented anti-torture measures. In other words, there is no evidence that her findings can be generalised to compliance with ECtHR judgments among ECHR member states. Therefore, research into this topic can be relevant in order to gain knowledge about compliance with anti-torture judgments on the European level. Secondly, 7,865 out of a total of 43,100 ECtHR judgments in 2018 are violations of art.3 ECHR which means that this is the second biggest group of ECtHR judgments nowadays, only preceded by the right to a fair trial (Council of Europe, 2019a, pp.167-171). Lastly, the degree of compliance is crucial to measure the long-term capabilities, validity and reliability of the ECHR and the ECtHR as a system to combat torture and inhuman or degrading treatment or punishment by states (Anagnostau, 2013, p.3).

Accordingly, this research aims to explain divergent degrees of compliance between the ECHR member states in ECtHR judgments focused on art.3 violations. As Simmons did not focus on compliance on the EU-level, there is no evidence that Simmons’ theory additionally applies to art.3 ECHR judgments. Therefore, this research is focused on testing if the factors derived from Simmons’ theory, the type of political regime and the degree of respect for the rule-of-law, result in differences in compliance with art.3 ECHR judgments between the ECHR member states. Consequently, the main research question of this thesis is:

Which domestic factors explain compliance and non-compliance of ECHR member states with ECtHR judgments concerning art.3 of the ECHR?

(12)

12 In order to answer the main research question, this thesis consists of the following elements. Firstly, theoretical framework describes the state-of-the-art concerning this topic including the theory of Simmons, and key concepts and indicators that will be applied to this research. Secondly, the methodology of this research is described and explained in detail. Thirdly, the collected data and research results are outlined and analysed. Finally, a conclusion is drawn thereby answering the main research question.

(13)

13

Theoretical framework

In this part of the research, relevant theories, key concepts and indicators are discussed. Furthermore, the theory applied to this specific research will be explained in detail. Thereafter, the contribution of the research to the academic world will be justified. The theoretical framework is structured on the basis of the concepts.

1. Dependent variable: Number of repetitive art.3 ECHR judgments

Art.3 of the ECHR describes the prohibition of torture stating that: ‘’No one shall be subjected to torture or to inhuman or degrading treatment or punishment’’ (ECHR & Council of Europe, 2010, p.7). Art.3 ECHR is treated as an absolute right, meaning that this right can neither be limited nor confined in any situation. Nonetheless, the image of art.3 ECHR of being an absolute right should be nuanced as this is not always the case. Indeed, the ECtHR judges states on either lawful or unlawful behaviour, however the ECHR lacks to describe details of what is considered as torture or degrading or inhuman treatment or punishment. Therefore, the absolute right is not always absolute as the article does not include details defining what falls under breaches (Mavronicola, 2015). Therefore, it could be possible that cases concerning art.3 ECHR could relatively frequently come for the ECtHR. This does not necessarily mean that compliance differs due to national-bound factors. Rather, it is the result of the ECHR and ECtHR system. However, it visualises the complexity or unclarity for states in defining and assessing actions in terms of violation of art.3 ECHR.

Gerards & Fleuren (2014) state that the ECHR is ‘’deeply embedded in national law, national policy and national case law’’ (p.3). This would mean that art.3 ECHR is embodied in the domestic law of its member states. However, Varju (2014) argues that the ‘’EU judicial power in the protection of human rights is limited by a parallel constitutional authority situated externally and respect for that authority is expressed in the interpretative practice of alignment to ECHR law’’ (p.217). These parallel constitutional authorities are national legislative bodies. It underlines the relative freedom of states to apply human rights laws and judicial rulings from the ECHR/ECtHR-level to national laws and policies. Another research complements this statement by suggesting that ‘’interpretative principles and standards result in interpretative authority’’ (Martens in Gerards & Fleuren, 2014, p.2). The interpretative character of the article could explain the divergent compliance of ECtHR judgments. However, the fact that art.3 ECHR specifically is an absolute right should leave no room for interpretation of the article itself and judgments resulting from violations of this article. Therefore, this cannot empirically explain divergent compliance with judgments between states. Nevertheless, it does mark the importance of domestic factors in compliance or non-compliance of the assigned states.

(14)

14 Before exploring these domestic factors further, it is important to analyse the different parts of art.3 ECHR by dividing the terms ‘torture’ and ‘inhuman or degrading punishment’.

1.1 Torture & degrading or inhuman treatment or punishment

The Declaration on the Protection of All Persons from Being Subjected to Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment from December 1975 laid down the basis for the definition of torture. Additionally, it was the first treaty to prohibit torture without exceptions (Simmons, 2009, p.261). This basic definition was further defined by the Committee Against Torture (CAT), stating that:

"Torture means any act by which severe pain or suffering, whether physical or mental, is intentionally inflicted on a person for such purposes as obtaining from him or a third person information or a confession, punishing him for an act he or a third person has committed or is suspected of having committed, or intimidating or coercing him or a third person, or for any reason based on discrimination of any kind, when such pain or suffering is inflicted by or at the instigation of or with the consent or acquiescence of a public official or other person acting in an official capacity. It does not include pain or suffering arising only from, inherent in or incidental to lawful sanctions.’’ (OHCHR, 1984, p.1).

The EU takes this definition into account in the ECHR and the ECtHR (Simmons, 2009, p.261). The exact definition of torture is described as: "deliberate inhuman treatment causing very serious and cruel suffering" (Council of Europe, n.d.). This definition is short and ambiguous, however, in reports and judgments the ECtHR does make the definition more explicit by adding situations, persons and specific activities considered as torture. Besides the role of defining torture, the ECtHR has the right to form or change state acts in order to decline violations (Moravcsik, 2000). Accordingly, the ECHR and the ECtHR seek to minimise the interpretative character leading to repetitive cases due to the lack of an explicit definition.

Besides torture, art.3 ECHR additionally describes the prohibition of degrading or inhuman treatment or punishment. The ECHR divides torture from degrading and inhuman treatment or punishment in the intensity of the suffering and the deliberation of the act. Inhuman treatment is described as ‘’causing either actual bodily harm or intense mental suffering" (Council of Europe, n.d.). There must be a certain degree of severity, but it does not have to be deliberate nor executed for a certain purpose. The most occurring case is people suffering from abuse in prisons. Contrary to suffering, degrading treatment includes humiliation without necessarily deliberation. The most occurring case from this

(15)

15 part of the article is poor conditions in penitentiaries. The severity is here less than in cases of inhuman treatment or punishment. Nonetheless, if for example these prison conditions are too severe, it would be categorised as inhuman instead of degrading (Council of Europe, n.d.).

1.2 ECtHR decisions on art.3 violations

Torture and degrading or inhuman treatment or punishment are very complex topics for the ECtHR to examine in an impartial and orderly way. Firstly, states are not interested in exhibiting their shortcomings. Consequently, figures and numbers about torture and degrading or inhuman punishment or treatment are usually gathered by non-governmental organisations (NGOs) instead of states. Secondly, if the government decides to exhibit these practises, they are interested in minimising the gravity of the situation. Lastly, examining torture is complicated as the responsible actor -the government- could not oversee all acts as it is for example distributed over numerous police headquarters and penitentiaries (Simmons, 2009, p.259-260). Therefore, art.3 ECHR judgments can be implemented by the government into domestic law, but might not be executed by other government agents, for example police officers or local policy makers. Accordingly, one would suggest that repetitive cases and judgments could easily occur relatively easy.

The ECtHR delivered 844 judgments concerning art.3 ECHR violations (HUDOC, 2020b). In Graph 1, the distribution of these judgments is visualised. How do these judgments come about? The ECHR is actually a tool to implement changes in national laws and demands that all states alter national judicial systems if it is necessary to meet the requirements of the ECHR. If a state fails to uphold these requirements, a single citizen, NGOs or other state actors can litigate a state at the ECtHR. The state thus empowers the ECtHR to decide if the state violated the ECHR and if a state should change their activities. Before the ECtHR launches its judgment, the state and prosecutor could opt for a friendly settlement. This implies that (in most cases) the ECtHR proposes a certain amount of repayment, as written down in art.41 ECHR (Roper & Barria, n.d.). Friendly settlements are no judicial rulings as they involve an admissibility decision which additionally means that the execution of a friendly settlement is not checked by the Committee of Ministers (Keller, Engi & Forowicz, 2010, p.123). Therefore, friendly settlements will not be included in this research. Although the ECtHR plays a significant role in examining violations, decisions from the ECtHR cannot overrule national law. This means that states themselves decide the degree of compliance and what to implement in order to avoid future violations and corresponding prosecutions. The law can be changed as a result of the judgment, but the ECtHR is unable to stipulate actions to be taken (Hawkins & Jacoby, 2008). Consequently, this could lead to a situation where there is actually non-compliance with ECtHR judgments. Thereafter, the state could

(16)

16 face new litigation due to repeated violations. Nevertheless, the Committee of Ministers checks if and how a member state complies to ECtHR judgments (Von Staden, 2018, p.36).

Graph 1: Number & distribution of ECHR judgments. Source: HUDOC (2020b).

Out of the 844 art.3 ECHR judgments, 458 judgments are the result of repetitive cases (see Graph 2) (HUDOC, 2020b). As more than 50% of all the judgments are repetitive, it is important to understand what repetitive cases actually mean. Repetitive judgments are defined by the ECtHR and indicate that the ECtHR have dealt with these cases previously (European Court of Human Rights & Council of Europe, n.d., p.6). Moreover, these cases identify that there are continuing and fundamental complexities on the domestic level impeding compliance (McKenzie et al., 2005, p.38). Lastly, repetitive cases are the indication for non-compliance with ECtHR judgments by member states (Lambert Abdelgawad, 2018, p.8).

Graph 2: Type of art.3 ECHR judgments. Source: HUDOC (2020b).

The fact that there is a high number of repetitive judgments is quite problematic. Therefore, it is important to understand why these cases are repetitive, i.e. why states do not comply after the first judgment. Non-compliance with ECtHR judgments is a problem recognised by the ECHR itself (Council of Europe, 2016). Non-compliance with art.3 ECHR judgments increases the chance of overloading the

0 20 40 60 80 100 120 140 A lb an ia A n d o rr a A rme n ia A u str ia A ze rb aijan Bel gi u m Bo sn ia &… Bul gari a Cr o ati a Cyp ru s Cz ec h R ep u b lic De n mark Es to n ia Fi n lan d Fr an ce G e o rg ia G e rman y G re ec e H u n gary Ic el an d Ire lan d Italy La tv ia Li ec h te n st ei n Li th u an ia Lu xe mb o u rg Mal ta R ep u b lic o f… Mo n ac o Mo n te n eg ro N e th e rl an d s N o rt h -… N o rwa y Po la n d Po rt u gal R o man ia R u ss ian … San Mari n o Se rb ia Sl o vak R ep u b lic Sl o ve n ia Sp ain Sw ed en Sw itz er lan d Tu rk e y Uk rai n e Un ite d Kin gd o m

Nr. of art.3 EHCR judgments by member state

(total: 844)

(17)

17 court system with repetitive cases (Von Staden, 2018). This is disadvantageous for the ECtHR in terms of an expanded workload as well as for member states in terms of extra costs and inconvenience (McKenzie et al., p.66). Moreover, the UN underlines the importance of compliance with judicial rulings i.e. the decrease in repetitive cases for ECHR member states as non-compliance questions the authority of the ECtHR and the ECHR system in total (OHCHR, 2005, p.49). Lastly, compliance with ECtHR decisions, whether national or international, is a characteristic of a democratic society which is one of the corner stones of modern-day Europe (Council of Europe, 2016).

In order to study the rate of compliance or non-compliance of states with art.3 ECHR judgments, this research takes into account the number of repetitive art.3 ECHR judgments as, as mentioned previously, repetitive judgments indicate non-compliance with previous ECtHR judgments. However, the absolute number of repetitive art.3 ECHR judgments could not result in reliable results. When a state has a high number of art.3 ECHR judgments, chances are higher that such a state received more repetitive judgments as well. Therefore, it is important to include the relative number of repetitive art.3 ECHR judgments instead. In Graph 3, one can find the number of repetitive cases as percentage of the total amount of art.3 ECHR judgments. The average percentage of repetitive art.3 ECHR judgments relative to the total of art.3 ECHR judgments for a member state is 27.3%. However, as one can see in Graph 3, the relative number of repetitive art.3 ECHR judgments is distributed with great differences among the member states. Is the unequal distribution of repetitive art.3 ECHR judgments a result of the interpretative character of the article described by Mavronicola (2015) or the relative freedom for states to interpret the judgments explained by Varju (2014)? In order to find possible factors of compliance and non-compliance with ECHR states with art.3 judgments, one should first outline compliance theories.

Graph 3: Percentage of repetitive art.3 ECHR judgments as part of the total number of art.3 ECHR

(18)

18 2. Compliance theory

Theories of compliance are nothing new, however, the area of human rights is unique as it does not possess the characteristics that forces states into compliance as other areas such as threats of boycotting or restrictions in trade areas. Consequently, there are fewer reasons for states to check if the state itself or other states comply with international human rights agreements (Henkin in Hathaway, 2002, p.1938). Nevertheless, in the last decades human rights treaties were under a magnifying glass. Scholars generally recognised the necessity of a certain degree of compliance from nations in order for an international law to have effect (Hathaway, 2002, p.1943). Before turning to the traditional theories, it is important to define the term ‘compliance’. Compliance is a concept that universally described as ‘’conformity of an actor’s observed behaviour with the behavioural requirements of a normative pre- or proscription applicable to that actor’’ (Von Staden, 2018, p.30). In the case of the ECHR, the behavioural requirements are laid down in the ECHR or in ECtHR judgments and elucidated by the ECtHR. This rather descriptive definition leaves a gap for interpretation on what factors generates compliance and non-compliance. This is further defined by judicial rulings of the ECtHR.

Emily Hafner-Burton and Kiyoteru Tsusui (2005) were the first researchers to underline the significance of compliance to international human rights agreements and judicial rulings to result in better human rights practices. They created a theory grounded in institutional isomorphism with the claim that treaty or measure acceptance is nothing more than a hollow commitment to work in favour of other states and, therefore, might not have the intended effects. Simmons contrasts on this aspect as she concludes that international human rights treaties as well as judicial rulings do have those effects in all states, however, the effects are different among states (Simmons, 2009, p.16). The contribution of these studies is the attention to the potential of quantitative data to test the functioning of international human rights agreements and rulings. Nevertheless, previous studies into human rights treaties were based on comprehensive case studies in individual states. Therefore, these studies lack in generalisability (Simmons, 2009, pp.10-11). Besides that, there has been hardly any interest in factors leading to compliance. Furthermore, the study in variation between states was left behind (Hathaway, 2002, p.1943). The first research to take into account generalisability and causes of variation between states was conducted by Oona Hathaway. In that study, Hathaway (2002) tested state activities compared to treaty requirements (p.1964). In other words, she studied whether or not countries that signed international human rights treaties are violating conditions agreed upon in the treaty. The conceptualisation and operationalisation of her research is relevant for this research as well. Hathaway combined the traditional theoretical approaches into two categories: the rational approach and the

(19)

19 normative approach (p.1944). Her theory of treaty ratification combines several aspects of the two approaches.

2.1 Rational approach

The rational approach is focused on the realist view that individual, self-interested motivations are the explanation for compliance. These motivations could involve for example philosophies, (refraining from) conflicts, power and reputation, but differ among the sub-streams of the rational approach. The realist and instrumentalist strand argues that in terms of human rights, states make a cost-benefit analysis of compliance and non-compliance only sometimes with a specific focus on the state’s reputation and direct actions that could abuse their reputation (Snidal, 1985). Hathaway rejects this idea as human rights violations are hardly handled with those direct (financial and military) actions. The liberal rational actor model grounded by Immanuel Kant is focussed on domestic characteristics of states that decide on compliance of non-compliance with international agreements (Doyle, 1983, pp.343-344). The significance of domestic characters is additionally underlined by Von Staden who claims that human rights treaties and measures have weak extrinsic enforcement and thus a strong intrinsic enforcement (2018). The most important domestic characteristic that Hathaway has found to influence compliance or non-compliance with international human rights treaties is a states’ regime type.

2.2 Normative approach

Contrary to rational model approach, the normative approach is focused on the influential power of legal agreements. The way states act regarding compliance with human rights treaties is not merely a result of a realist cost-benefit analysis, it is rather focused on ‘’the influence and importance of ideas’’ (Hathaway, 2002, p.1955). The type and degree of influence differ among normative theories. The persuasive strand argues that states comply to human rights treaties due to their symbolic adaption, which means that states comply to promote a political position (Ovink et al., 2016). Consequently, globally widespread norms concerning human rights are easy to comply to as all countries want to show their rejection of human rights violations. This theory is countered by Hathaway as it lacks to explain why states choose to comply not with human rights treaties. The managerial strand states that compliance to international human rights agreements or rulings is the result of the ‘norm of compliance’. States comply to human rights agreements as they involve normative responsibilities which makes them obliged to comply. This is one of the few theories that additionally explains why states do not comply. According to the managerial theory, non-compliance is due to a state’s inability to manage the norm of compliance to an agreement originating from capacity shortages. These

(20)

20 capacity shortages can mainly be found in the domestic structure of a state (Chayes & Chayes, 1995). Hathaway as well as Simmons elaborated on this theory and found one factor that could be a sign of domestic capacity shortages, that is the level of respect for the rule-of-law.

One of the main institutions of the rule-of-law is a state’s court system. For a court to be an actual tool to push for compliance, it needs judicial autonomy, capability and integrity. It is rare for states with these judicial characteristics to be listed as one of the worst tortures as their judicial system does not allow for it (Hathaway, 2002, p.1971). Moreover, it would mean that states with high and medium respect for the rule-of-law are expected to make improvements after judgments as they do have the characteristics, which would result in higher compliance rates (Kaufmann et al., 2004, p.255). Simmons’ research underlined this theory partly in the area of anti-torture rulings after finding improvements in torture practises among states with only medium respect for the rule-of-law. This difference could be explained by the focus of the different studies (Simmons, 2009, pp.282-283). For Hathaway and Kaufmann this was on human rights in general, whereas Simmons specifically focused on anti-torture measurements.

(21)

21 3. Independent variables

The theory of Simmons (2009), which will be used in this research, combines two aspects of each theoretical approach of Hathaway’s theory of compliance (2002). According to Simmons (2009), the focus from the rational actor approach on regime type and the focus from the normative approach on the level of respect for the rule-of-law are said to be the main factors of compliance or non-compliance by states with international agreements and rulings.

3.1 Regime type

The regime type of a state highly influences domestic politics and the mobilisation effects of international human rights agreements, including anti-torture measures and rulings (Simmons, 2009, p.16). Regime types are divided into three categories, that is ‘Stable democracy’, ‘Partial/unstable democracy’ and ‘Stable autocracies’. Before these terms are described, the term democracy needs to specified. Democracy could be defined as "general openness of political institutions" (Center for Systemic Peace in Hathaway, 2002, p.2030). A stable democracy is defined as: ‘’All countries that have been stable democracies (never below the 8 on the polity scale) since World War II’’ (Simmons, 2009, p.391). The structure of a liberal democratic state allows for civil society mobilisation and, therefore, these states are pushed towards compliance as their character reflects that once an agreement or judgments is made, the state should abide to it (Hathaway, 2002, p.1954). Therefore, one would expect full and stable democracies to have the highest compliance rates. Nevertheless, Simmons’ theory found that this is not the case. Especially due to this character, Simmons claims that stable democracies are not likely to alter domestic politics in the area of human rights after signing an international human rights treaty or after the launch of CAT measurements as, for example, the prohibition of torture and the prohibition of inhuman or degrading punishment of treatment are already ‘’domestically guaranteed protections’’ (Keck & Sikkink in Simmons, 2009, p.16). This means that the state could not feel the necessity to alter domestic practises in the area of human rights after receiving ECtHR judgments, which in turn could lead to non-compliance.

Partly or unstable democracies are defined as by Simmons (2009) as: ‘’All other countries other than ‘stable democracies’ and ‘stable autocracies’. This includes countries that are transitioning to democracy, moving toward authoritarianism, and hovering between 5 and 8 on the polity scale’’ (p.391). According to Hathaway’s and Simmons’ theories, this regime type has relatively the highest chances of compliance. Firstly, the development towards a state’s democratic character creates room for pro-rights alliances to secure human rights and keep the government responsible for upholding human rights. Secondly, the courts of these states could lack experience with democratic court systems

(22)

22 or could be somewhat incapable (yet) to establish human rights prosecutions. On the other side, they are expected to be less loyal to the government than before the democratic transition, thereby leading to a more independent court system (Simmons, 2009, p.153). The role of the (in)dependent court system is explained in detail in section 3.2.

In the case of the ECHR specifically, Moravcsik’s theory (2000) states that relatively new and probably unsteady democracies are more likely to endorse internationally established human rights treaties or measures than stable democracies and stable autocracies. The reason for this is that these states seek to commit to democratic procedures by means of applying those human rights treaties to the domestic political system (p.228). Due to their wish for commitment, the external pressure coming from human rights treaties is higher on partly or unstable democracies, especially in the regional context of Europe. Even if states do not have the capabilities to fully comply with a treaty or measure, they attempt to comply due to the pressure these states feel from controlling bodies (Hathaway, 2002, p.2019). Moreover, Simmons took this theoretical perspective into account in her own research and found that, indeed, in partly or unstable democracies have better compliance rates with international anti-torture measures. As stated by Simmons (2009), ‘’torture is about political and social control’’ (pp.265-266). Consequently, her research found that anti-torture measures imposed on countries had the most effects in countries where this control is challenged. This challenged control can be traced in political regimes that are shifting, for example partial or transitional democracies. Therefore, it would be interesting to test whether or not compliance rates with art.3 ECHR judgments in particular are higher in partly or unstable democracies. One complication of this theory is that stable democracies might have less absolute art.3 ECHR judgments than the other categories, thereby leading to unreliable results when compared to partial/transitional democracies or non-democracies. However, this research does not focus on the absolute number of repetitive art.3 ECHR judgments. Rather, the dependent variable is the relative number of repetitive art.3 ECHR judgments, which eliminates unreliable results due to an unequal distribution between the regime type categories.

A stable autocracy is defined as: ‘’The country never scored above 5 on the polity scale in the twentieth century’’ (Simmons, 2009, p.391). Contrary to stable democracies, stable autocracies lack the previously mentioned capacities that are necessary to facilitate the demanded changes (Simmons, 2009, p.16). One thing that stable democracies and stable autocracies have in common is that the control is not challenged, hence the term ‘stable’. According to Simmons theory, control has to be challenged in order for an human rights agreement to have the highest chance of compliance. Lastly, stable autocracies have most likely never been democratic, which makes it less likely for any human

(23)

23 rights agreement or measure to reach compliance or to have even any impact at all (Simmons, 2009, pp.266-276).

For this thesis, to study the influence of regime type on compliance with the ECtHR judgments regarding violations of art.3 ECHR the following hypothesis will be tested:

H1: States with partial or transitional democracies are more likely to comply with art.3 ECHR judgments than stable democracies or stable autocracies.

3.2 Respect for the rule-of-law

The World Bank’s definition of the rule-of-law is: ‘’Perceptions of the extent to which agents have confidence in and abide by the rules of society, and in particular the quality of contract enforcement, property rights, the police, and the courts, as well as the likelihood of crime and violence’’ (World Bank, n.d.). This definition is leading for other international organisations such as the EU and is therefore a comprehensive definition. According to the UN Deputy Secretary General (2014) , the rule-of-law is intertwined and jointly effective with human rights. The states’ rule-of-law thus involves different aspects, however for human rights there are a few aspects that are more important. International human rights treaties such as the ECHR allow citizens to appeal for their rights, especially in the area of torture or inhuman or degrading treatment or punishment as this could cause activism (Jenner, 1999, p.2). The power of the civil society (‘agents’) within a state decides on the impact of their mobilisation. Furthermore, litigation provide for citizen’s freedom to defend their rights and interests against the state. Next to the opportunity to litigate, the citizen’s knowledge of the treaty, their rights and possible procedures when there is suspected violation is additionally important. This is called legal literacy and it differs among states. Finally, the court’s degree of (in)dependency from the government is an aspect that determines the respect for the rule-of-law in a country.

States with strong respect for the rule-of-law are states that ‘’score on average above +1 on the World Bank’s Rule-of-law measure’’ (Simmons, 2009, p.392). It is said that international human rights treaties can impose prosecutions by a court, however, it lacks the ability to impose compliance deriving from the court judgments. Nonetheless, states where there is a relatively strong institution regarding the rule-of-law, this rule-of-law can press states into compliance with judgments (Simmons, 2009, p.14). Secondly, in those states the opportunity of prosecution is a significant method to force compliance with binding international human rights treaties such as the ECHR as well as with human rights rulings from, for example, the ECtHR (p.135). Nevertheless, Simmons found that these states do not improve their practices in the area of torture after international agreements or rulings (Simmons, 2009, p.280).

(24)

24 Strong respect for the rule-of-law additionally involves a court that is independent from pollical control (Cross, 1999). Lastly, citizens have a high level of legal literacy, however, there is generally no need for the society to mobilise and fight for their rights as these rights are broadly recognised by the government.

States with moderate respect for the rule-of-law are states that ‘’score between -1 and +1 on the World Bank’s rule-of-law measure’’ (Simmons, 2009, p.392). Practically, it means that the rule-of-law is fairly matured and the courts are quite autonomous from government control. Moreover, citizens have some legal literacy, which is included in the World Bank’s measure for the rule-of-law. Legal literacy enables citizens to start litigating. Litigating is a significant tool for these regimes as there is still space left for improvement. Therefore, there is an incentive for the civil society to mobilise and litigate (Hathaway, 2002, p.133). These states are thus more likely to bring cases concerning art.3 ECHR for the ECtHR, which could result in more ECtHR judgments and thus lower compliance rates. A state with weak respect for the rule-of-law is a state that is ‘’scoring below -1 on the World Bank’s rule-of-law measure’’ (Simmons, 2009, p.392). In countries with a relatively weak respect for the rule-of-law, one would expect compliance force to be less successful. Simmons’ research emphasises this as states with a relatively weak legal system do not perform better after an international human rights agreement is made or after a ruling is launched, which could be due to domestic inabilities as stated by managerial theorists. Finally, courts within these states are quite dependent on and loyal to the government. Consequently, they might be unable to prosecute the state in the case of violation (Simmons, 2009, p.280). Therefore, one would assume that an ECtHR judgment would make no noticeable difference as well. Lastly, the legal literacy rate is so low that it is less likely for the civil society to start litigating. Therefore, cases could not come before the ECtHR, indicated by less repetitive judgments and higher compliance rates.

In the ECHR area, the level of the rule-of-law differs from state to state. Some of the member states are facing structural threats to the rule-of-law. One specific area of concern within the rule-of-law is the ‘’lack of independence of courts and judges’’ (European Political Strategy Centre, 2019, p.3). Therefore, it would be interesting for this thesis to study whether the role of the rule-of-law has similar results for art.3 ECHR judgments. Therefore, the second hypothesis that will be tested is:

H2: States with moderate respect for the rule-of-law are more likely to comply with art.3 ECHR judgments than states with strong or weak respect for the rule-of-law.

(25)

25 4. Control variables

In addition to the independent variables, a number of control variables are included in this research in order to reduce the possibility of unjustly identifying the relationship between regime type or the level respect for the rule-of-law with the relative number of repetitive art.3 ECHR judgments. The first series of controls are other factors in the area of domestic accountability, that is the ‘Freedom of the press’ and the ‘Degree of federalism’. The prevalence of domestic accountability is key in achieving compliance (Simmons, 2009, p.270). The first variable in a way controls a state’s domestic accountability while the latter variable stimulates accountability through some form of competition, which makes the two variables of great significance in this case (Hathaway, 2002). The press is an institution that is able to start resistance when a governments carries out acts of torture or degrading or inhuman punishment or treatment. Moreover, a free press can reveal and criticise a government without restrictions and concerns of retaliation or persecution. Besides this, the state´s accountability could affected by a free press. Lastly, the press can make the society aware of citizens’ rights in case of acts of torture or degrading or inhuman punishment or treatment (Simmons, 2009, p.270). The second control variable regarding domestic accountability is the degree of federalism. For states with a more centralised or unitary governmental organisation, it might be less complex to check whether or not compliance with ECtHR judgments is achieved than in states with a more federalised governmental organisation (Simmons, 2009, p.270).

The third control variable, the ‘Level of violent conflicts’ is related to a states’ security situation. In the majority of research regarding human rights in times of war, it was concluded that wars increase the chances of physical and social circumstances connected with hostile behaviour (Sironi & Branche, 2002). Moreover, it is said that even well-developed states with a high level of human rights can legitimise acts of torture or degrading or inhuman punishment or treatment in times of war and deviate from compliance with rulings or agreements (Simmons, 2009, p.271). The last controlling variable is ‘GDP per capita’, which indicates the state’s developmental level (Simmons, 2009, p.273). The increase of a state’s GDP/capita is said to affect compliance with anti-torture measures and thus repetitive cases of torture or degrading or inhuman punishment or treatment as lower economic development linked with worse human rights practices. Moreover, states with a higher GDP/capita are more likely to comply with anti-torture measures than states with a lower GDP/capita (Beger & Hill, 2019, p.639).

(26)

26 5. Justification for Hathaway’s & Simmons’ theories

Hathaway’s theory of treaty ratification compliance has been the baseline for Simmons theory and research. Hathaway’s study was focused on the broader human rights treaties, where torture is only one small aspect from. As Simmons focused on reports published and measures launched by the Committee Against Torture (CAT), more attention was paid to the prohibition of torture and inhuman or degrading punishment. Moreover, the CAT is similar to the ECtHR in that it is highly influential over national governments to undertake action than the treaties and agreements applied in Hathaway’s theory (Simmons, 2009, p.15). Besides this, Simmons’ theory is more relevant for this research as it focused on measuring compliance with measures and based on litigation. Accordingly, art.3 ECHR judgments could be seen as an individual and specific anti-torture measurements. Moreover, Hathaway examined state activities compared to measure requirements, which could be scaled down to state activities (compliance or non-compliance) compared to the requirements laid down in the ECtHR judgment. As this research seeks to explain compliance with such litigation, i.e. art.3 ECHR judgments, it would be more suitable to apply Simmons’ theory, which is built on Hathaway’s theory, rather than solely on Hathaway’s theory. Moreover, Simmons’ research includes a distinct set of control variables, which makes the research more reliable in terms of the correlation between dependent and independent variables. Finally, it would be interesting to test whether factors leading to compliance with international agreements, measures and rulings found on the global level would apply to compliance with judicial rulings from the ECHR and the ECtHR as well. Accordingly, the theories by Hathaway and Simmons seem suitable to apply to the examination of compliance with ECtHR judgments, specifically concerning art.3 ECHR judgments as it is the article that is focused on anti-torture and its judgments are actually anti-torture measures.

(27)

27

Methodology

In this section, the methodology for the data collection and analysis are explained, thereby clarifying how the research is conducted in order to answer the research question. This research follows a quantitative approach. The data was gathered through the databases from the ECtHR, World Bank, the Systemic Peace Organisation and the Institute for Economics and Peace. As data on acts of torture or degrading or inhuman treatment or punishment, regime types, and the level of respect for the rule-of-law are already produced and published as public records through several databases, this data was applied to this research. Before discussing the data collection and data analysis, the number and specification of cases are listed below. The cases are 37 out of the 47 members of the ECHR. The reasoning for the elimination of 10 ECHR states can be found in the justification of the methodology (p.29). 1. Albania 2. Armenia 3. Austria 4. Azerbaijan 5. Belgium 6. Bosnia & Herzegovina 7. Bulgaria 8. Croatia 9. Cyprus 10. Czech Republic 11. Estonia 12. Finland 13. France 14. Georgia 15. Germany 16. Greece 17. Hungary 18. Ireland 19. Italy 20. Latvia 21. Lithuania 22. Republic of Moldova 23. Montenegro 24. Netherlands 25. North-Macedonia 26. Poland 27. Romania 28. Russian Federation 29. Serbia 30. Slovak Republic 31. Slovenia 32. Spain 33. Sweden 34. Switzerland 35. Turkey 36. Ukraine 37. United Kingdom 1. Data collection 1.1 Independent variables

The data for this research was collected through structured data collection instruments. For the number of art.3 ECHR violations, the information was gathered through official EU websites including those from the European Commission, the Council of Europe and the ECtHR. Moreover, the database from the ECtHR was an important source for other statistics regarding art.3 ECHR violations from the ECHR member states. The database consists of all recorded judgments of all violations and all articles from the ECHR, and allows for filtering on article number, ‘violation’ or ‘non-violation’, and the state concerned and organisation concerned, in this case the ‘European Court of Human Rights’ (HUDOC, 2020a). The database contains judgments described in English and the official language of the state concerned. Data for this research was gathered from English documents solely as it is beyond the limits of this research to include judgments described in other languages. Thereafter, the HUDOC-EXEC

(28)

28 database was used to study the state of the judgments. The database allows for filtering on states, the number of cases, the number of repetitive cases, and the number of closed or pending cases (HUDOC, 2020b). In order to test (non-)compliance of states with art.3 ECHR judgments, the relative number of repetitive judgments were applied as these are indicators of non-compliance. Pending judgments were excluded from this research as these judgments require measures that take a lot of time. This results in pending cases slowly moving towards compliance rather than simply non-compliance. The relative number of repetitive judgments visualises the ratio of the number of non-complied judgments against the total number of ECHR judgments. This ratio is necessary as some states have more art.3 ECHR judgments than other states. Consequently, this leads to a higher number of repetitive cases than other states, but is not necessarily an indication of more or less compliance.

Secondly, data on regime types was gathered through the Polity IV database which is a project from the Center for Systemic Peace. The first reason to apply data from this organisation was that Simmons as well as other researchers applied this to their research (Gretchen & Claudiu, 2003). The second reason was that it is a comprehensive database in terms of states, topics and time series. The Polity IV database contains information from all ECHR member states included in this research (Center for Systemic Peace, 2020). As mentioned previously, the states were then divided into ‘stable democracies’, ‘partial or transitional democracies’, and ‘stable autocracies’. The term democracy was be scored from the scale 0-10, where 0 is considered low in general openness to political institutions and 10 is high in general openness of political institutions (Center for Systemic Peace in Hathaway, 2002, p.2030). Thirdly, the level of respect for the rule-of-law in the ECHR member states was collected through the World Bank’s measurement of the rule-of-law which can be found in the World Bank Data Catalog (2019). Information on the rule-of-law is a combined score, consisting of ‘’the extent to which agents have confidence in and abide by the rules of society, and in particular the quality of contract enforcement, property rights, the police, and the courts, as well as the likelihood of crime and violence’’, and are known as Worldwide Governance Indicators (World Bank, 2019b). The data was then categorised in either weak, moderately weak, moderately strong or strong respect for the rule-of-law, by scoring them between -2.5 and 2.5. The moderate category was a broad category (all scores between 1 and -1), therefore this category was divided into two sub-categories. These sub-categories visualise whether the states are leaning more towards the states with strong respect for the rule-of-law (0 to 1) or towards states with weak respect for the rule-of-rule-of-law (-1 to 0) (Simmons, 2009, p.281).

1.2 Control variables

The controlling variables in the theory of Simmons that are relevant for this research were ‘Freedom of the press’, the ‘Degree of federalism’, the ‘Level of violent conflict’ and ‘GDP/capita’. The control

(29)

29 variable ‘Freedom of the press’ indicated the degree to which the press inside states is free from government intrusion (Simmons, 2009, p.391). Data was collected from the Freedom House’s database and was categorised as ‘Not free’ (0), ‘Partly free’ (1) or ‘Free’ (2). Secondly, data to measure the ‘Degree of federalism’ was collected through the World Bank’s Database of Political Institutions. The factors deciding on a states’ degree of federalism were the occurrence of locally elected municipalities, the power of local authorities to tax and the local ‘’consistencies of senators’’ (Simmons, 2009, p.384). The scale ranged from 0 (Most centralised) until 6 (Most federal). The third control variable, violent conflicts, measured if a state is involved with either a civil war or an interstate war. This data was collected through the Global Peace Index (GPI) rather than the database of the Peace Research Institute Oslo as done by Simmons’ research because the latest data originates from 2009. Therefore, this information was out of date and not relevant in this research. Instead, the GPI developed by the Institute for Economics & Peace (IEP) was applied in this research. The GPI is comprehensive and the dominant measure for international peacefulness (IEP, 2019, p.2). The GPI ranks states in terms of violent conflicts by measuring inter alia the occurrence of internal and external conflicts (IEP, 2019, p.86). The states were ranked by means of a scale ranging from 0 (Very low level of violent conflict) to 4 (Very high level of violent conflict). The last control variable, ‘GDP/capita’, was measured according to the World Bank Development indicators in current U.S. dollars (Simmons, 2009, p.385). The control variables were all entered in the regression at once as there was no theoretical order or control variables with more/less significance. The scales for the control variables as well as the scales for the independent variables were valid as they measure constructs aimed for this specific research. Additionally, the scales were reliable as they are used in previous research or have been established by scientifically valid institutes.

(30)

30 2. Data analysis

After the data collection, the data was analysed by means of a multiple linear regression thereby explaining how the dependent variable (the relative number of repetitive art.3 ECHR judgments) changes as the independent variable (regime type or respect for the rule-of-law) changes. As both of the independent variables and 4 out of the 5 control variables are categorical variables, the data was analysed on the basis of dummy variables. Dummy coding the variables enabled them to be included in the regression model from SPSS, leading to an improved analysis. The reference groups, dummy variables and coding designs are presented in Annex II (pp.59-60).

The advantage of a multiple linear regression is that it contributes to an improved predictive capacity in comparison to simple linear regression. Moreover, this method specifies to some degree the significance of each variable in the relationship. The inclusion of control variables rules out threats such as confounding variables (Suzuki & Toshkov, 2017). The disadvantages of a multiple linear regression are mainly traced back to the data being used, including missing data and the influence of outliers. The software that was used for this multiple linear regression is SPSS. SPSS automatically checks for missing data and thereby tackled the above-mentioned disadvantage. Moreover, potential outliers are detected and, if necessary, investigated properly. This way any variability in the analysis was recognised and dealt with correctly which in turn prevented for mistakes in the statistical analysis. By carefully examining the research design and instruments and excluding for example variability, the internal validity of this research was protected. The generalisation of this research was assured through the relatively large number of cases (37) with different state characteristics. Moreover, this research tested the generalisability of Simmons’ theory in the context of art.3 ECHR judgments and the ECHR member states. The reliability of this research, including the consistency and stability of the results, was secured through consistently carrying out all steps of the data collection and analysis method for each state and variable. Lastly, the conditions of this research were standardised as the measures and scales from other organisations, that is the ECtHR, the Center for Systemic Peace and the World Bank, were used for all states equally.

(31)

31 3. Justification

For this research, a quantitative approach was appropriate as it enabled to detect the relationship between ECtHR judgments and a state’s domestic characteristics. With qualitative methods or case-studies it would have been nearly impossible to make any general assumptions about this relationship (Hathaway, 2002, p.1939). By studying individual states, there was no opportunity to examine what factors would generally lead to compliance or non-compliance. Therefore, an inclusive, broader statistical analysis set apart certain factors that have effect on compliance or non-compliance with art.3 ECHR judgments. As mentioned previously, several ECHR states were excluded from this research. The first group of states were so-called microstates: Andorra, Liechtenstein, Malta, Monaco & San Marino. These microstates (countries with less than 500,000 total population in the most recent year) were not included in the data series as these states were considered too small to affect international relations in the broader, collective sense (Polity IV, personal communication, 2020). It was justified to eliminate these data from the research as the majority of these states did not possess any art.3 ECHR judgments. Only Malta accounted for 4 repetitive art.3 ECHR judgments, which is barely 0.5% of all repetitive art.3 ECHR judgments. Therefore, data from these states could be missed in this research without affecting the result (Neuman, 2014, p.229). Furthermore, Denmark, Iceland, Luxembourg, Norway and Portugal and were excluded from this research as well because no art.3 ECHR judgments for these states was found in the HUDOC database.

4. Ethical considerations

In this research, no research participants were included. Nevertheless, there were some ethical considerations that had to be clarified. Firstly, the literature and data derived from other authors and researchers used for this research were acknowledged by the means of APA referencing. Moreover, this research was conducted with objectivity in both the research process itself and the accompanied discussion and analysis.

(32)

Results

In this section the research results will be presented, structured by each research step. This section additionally includes the description of some problems faced while collecting the data.

1. Descriptive statistics

The descriptive research results help to create conditions to conduct the statistical analysis. In Table 1, one can find the mean, standard deviation, minimum/maximum scores, and the standard deviation of all independent variables (control variables included). Here, one can see that the variable ‘Regime type’ had the closest data to the mean. The mean of ‘Regime type’ (scored between 0=Democratic since World War II and 2=Never democratic), was the closest to 1 (Partial/transitional democracy). The standard error of the mean was the smallest for this variable, indicating that the mean of the model was likely to be close to the actual population mean. For the second independent variable ‘Respect for the of-law’ (scored between 0=Strong respect for the of-law and 3=Weak respect for the rule-of-law), the mean was the closest to 1 (Moderately strong respect for the rule-of-law).

Concerning the control variables, the mean of ‘Freedom of the press’ (scored between 0=Not free to 2=Free) came closest to 1 (Partly free). The variables ‘Degree of federalism’ and ‘Level of violent conflict’ were further away from the mean, which means that the data was more spread out. The mean of the ‘Degree of federalism’ (ranging from 0=Most centralised to 6=Most federalised) was near 3 (Equally centralised & federalised). For the ‘Level of violent conflict’ (scoring from 0=Very low to 4=Very high), the mean came closest to 1 (Low level of violent conflict). The frequency distribution of the values from the 37 cases among the categorical variables is presented in Graphs 4-9 below. The last control variable, ‘GDP/capita’, is a discrete variable instead of a categorical variable and had the largest standard deviation. Additionally, there was a great difference between the minimum and maximum score, i.e. there is a wide range. Furthermore, the standard error of the mean was fairly high. Nevertheless, a wide range is quite common for discrete variables comparing GDP/capita among states.

Descriptive Statistics N Minimum Maximum Mean

Std. Deviation

Std. Error of mean

Regime type 37 0 2 0,76 0,548 0,09

Respect for the rule-of-law 37 0 2 0,95 0,815 0,134

Freedom of the press 37 0 2 1,32 0,747 0,123

Federalism 37 0 6 2,76 1,847 0,304

Level of Violent conflict 37 0 4 1,24 1,011 0,166

GDP per capita 37 $3,095 $82,797 $25,542.17 $21,881.911 5,012

Valid N (listwise) 37

(33)

33 Graph 4: Histogram ‘Regime type’. Source: SPSS.

Graph 5: Histogram ‘Respect for the Rule-of-Law’. Source: SPSS.

Graph 6: Histogram ‘Freedom of the press’. Source: SPSS. Graph 7: Histogram ‘Degree of federalism’. Source: SPSS.

Graph 8: Histogram ‘Level of violent conflict’. Source SPSS. Graph 9: Histogram ‘GDP/capita in US Dollars’. Source: SPSS.

11 24 2 0 5 10 15 20 25 30 N u m b er o f c a se s Regime Type

Regime Type

6 4 6 8 6 4 3 0 1 2 3 4 5 6 7 8 9 N u m b er o f c a se s Degree of federalism

Federalism

18 14 5 0 2 4 6 8 10 12 14 16 18 20

Free Press Partly Free Press

Not Free Press

N u m b er o f c a se s

Freedom of the press

Freedom of the press

5 25 3 1 3 0 5 10 15 20 25 30

Very low Low Medium High Very High N u m b er o f c a se s

Level of violent conflict

Level of violent conflict

13 13 11 0 0 2 4 6 8 10 12 14 N u m b er o f c a se s

Level of respect for the rule-of-law

Respect for the rule-of-law

(34)

34 2. Diagnostics

Before one can run the actual multiple regression, several basic assumptions had to be examined. By doing this, the possibility of type I and II errors in further multiple regression stages can be decreased (Cohen & Cohen, 1975). The unstandardized and standardized residuals (RESID & ZRESID) were applied to check for normality and linearity. Graph 10 visualises the Normal Predicted Probability Plot (Normal P-P Plot). Here, one can see that the residuals were normally distributed. Although there were some deviations, they followed the linear line quite accurately. Secondly, by checking the homogeneity, the distribution of residuals was tested. As one can see in Graph 11 below, the data did not follow a specific line. Furthermore, the data was spread over both the X- and Y-axis. This scatterplot gave no indication that the assumptions for this multiple regression model are not met. Accordingly, it indicated a linear relationship implicating that a linear regression model was suitable for the data included in this research (Holst, 2015). Thirdly, potential outliers by applying Cook’s distance had to be checked. Examining the Cook’s distance values, the maximum value was 0.155, which could be a potential outlier. However, this was not the case as:

a) it did not influence the regression line to a problematic extent, and

b) the potential outlier was not greatly deviated from other values (see Graph 12).

Therefore, one could state that there are no outliers detected in this multiple regression model.

Graph 10: Normal P-P Plot. Source: SPSS. Graph 11: Scatterplot. Source: SPSS.

Divergent Compliance with Art. 3 ECHR Judgements: Explaining the factors leading to compliance and non-compliance in ECHR member states