When Do Disarmament, Demobilisation and Reintegration Programmes Succeed and When Do They Fail? A Qualitative Comparative Analysis.

Transcription

1 When Do Disarmament, Demobilisation and Reintegration Programmes Succeed and When Do They Fail? A Qualitative Comparative Analysis Dave van Zoonen Master s Thesis Faculty of Social Sciences, Department of Political Science June Dave van Zoonen ( )

2 Thesis submitted in partial fulfilment of the requirements for the degree Master of Arts in Political Science at the Free University of Amsterdam. Faculty of Political Science Faculty of Social Sciences VU University Amsterdam Amsterdam, June 2015 Author Dave van Zoonen ( ) Sumatraplantsoen JE, Amsterdam Supervisor Dr. E.B. van Apeldoorn Second reader Prof. Dr. A.C. Hemerijck Wordcount:

3

4 Abstract In an effort to decisively resolve civil war and breaking the conflict trap in lessdeveloped-countries, programmes of Disarmament, Demobilisation and Reintegration (DDR) are aimed at assisting ex-combatants in their transition from war to peace. While there have been many (single) case studies on the success and failure of these programmes, no-one to this date has really sought to investigate whether the conditions that lead to the success of one DDR programme also apply to others. Consequently, while we know a lot about why individual programmes succeeded or failed, we know far less about the conditions that foster or hamper the success of DDR in general. By employing an original Qualitative Comparative Analysis (QCA) approach, this study systematically compares 38 cases of DDR implemented between 1989 and The results show that theories and findings presented in the literature on peace-building also apply to DDR. In particular, DDR most often succeeds when it is accompanied by fear-reducing mechanisms such as opportunities for political participation and third-party involvement in the form of peacekeeping missions. The presence of high opportunity costs for war and the absence of ongoing conflict also lead to the success of DDR. The results of this analysis further indicate that some DDR programmes succeeded in challenging environments lacking most of the conditions conducive to success. Future research could be focused on explaining the success of these cases. iv

5 Wars can be prevented just as surely as they can be provoked, and we who fail to prevent them, must share the guilt for the dead. ~ Gen. Omar N. Bradley ~ v

6 Contents Abstract... iv Tables and figures... viii Abbreviations...viiii 1. Introduction... 1 Transitioning from war-to-peace... 1 Disarmament, Demobilization and Reintegration (DDR) programmes State of the Art... 6 Literature on conflict resolution... 6 The case for DDR... 9 DDR studies and their main findings Conceptualizing success Methodology: ISM vs QCA Research Design Case selection Operationalisation and hypotheses Findings and discussion When does DDR succeed? When does DDR fail? Conclusion Appendix A Appendix B References vi

7 Tables and figures Figure 1: Root causes and symptoms leading to failure of DDR Figure 2: Venn diagram depicting necessary and sufficient conditions Table 1: DDR programmes from Table 2: Causal conditions, indicator and applied set-logic Table 3: Intermediate solution for successful DDR Table 4: Parsimonious solution for failure of DDR.. 45 Table A1: Raw data for fuzzy-sets in DDR success model...50 Table A2: Membership score card for DDR success model 51 Table A3: Sourcebook of Qualitative Sources Consulted for Coding Table B1: Analysis of necessary conditions for DDR outcome.. 58 Table B2: Truth table for presence of outcome (success) 59 Table B3: Truth table for absence of outcome (~success) 60 Table B4: Complex solution for the analysis of the presence of outcome (success).. 61 Table B5: Intermediate solution for the analysis of the presence of outcome (success) 62 Table B6: Parsimonious solution for the analysis of the presence of outcome (success).. 63 Table B7: Complex solution for the analysis of the absence of outcome (~success). 64 Table B8: Intermediate solution for the analysis of the absence of outcome (~success) vii

8 Abbreviations CAR csqca DDR DIIS DRC fsqca GDP HDI ICTJ IDDRS IHDI INTERFRET INUS ISMs LDC LRA MDSD MILF MNLF MSSD mvqca OLS ONUCA QCA RENAMO SSR UCDP UNDP UNDPKO UNMOT UNTAC WWI WWII Central African Republic Crisp-set Qualitative Comparative Analysis Disarmament, Demobilisation and Reintegration Danish Institute for International Studies Dominican Republic of Congo Fuzzy-set Qualitative Comparative Analysis Gross Domestic Product Human Development Index International Centre for Traditional Justice Integrated Disarmament, Demobilization and Reintegration Standards Inequality-adjusted Human Development Index International Force in East Timor Insufficient Necessary Unnecessary Sufficient Inferential Statistical Methods Less Developed Country Lord s Resistance Army Most Different Systems Design Islamic Front for National Liberation Moro National Liberation Front Most Similar Systems Design Multi-Value Qualitative Comparative Analysis Ordinary Least Squares United Nations Observer Group in Central America (Spanish acronym) Qualitative Comparative Analysis Resistencia Nacional de Mozambique Security Sector Reform Uppsala Conflict Data Program United Nations Development Programme United Nations Department of Peacekeeping Operations United Nations Mission of Observers in Tajikistan United Nations Transitional Authority Cambodia World War One World War Two viii

9 1. Introduction Transitioning from war-to-peace The tendency for civil wars to recur has been extensively documented and analysed (e.g. Collier, 2003; Doyle & Sambanis, 2000; Hartzell, Hoddie & Rothchild, 2001) leading some scholars to proclaim the existence of a conflict trap (Collier & Sambanis, 2002). Quantitative analysis has shown that countries which have experienced one civil war are significantly more likely to experience another (Collier, 2003). Collier finds that countries which have experienced civil war before face around 44 percent risk of returning to conflict within five years (Collier, 2003: p.83). Most scholars of conflict resolution have focussed on the characteristics of the previous conflict to explain the onset of a new war (Gurr, 2000; Rothchild and Groth, 1995; Licklider, 1995). However, Collier and Hoeffler (1998) argue that past grievances or political goals are not the main driving forces behind recurring conflict. Rather, they identify a selfenforcing mechanism of conflict and poverty which leads less-developed countries into a vicious circle where each conflict serves to repress economic development, lowering the opportunity costs for the next rebellion. The opportunity costs are the costs associated with going to war. Generally, citizens of more developed nations have more possessions, greater access to education and more economic opportunities than their counterparts from less-developed countries. Hence, they tend to risk more by going to war than citizens of poorer nations who have trouble sustaining their livelihood in the first place, making the problem of civil-war recurrence especially pertinent for less-developed countries (LDCs). Moreover, civil wars are detrimental to a state s economy. Collier (2003) finds that, on average, a nation s GDP per capita is 15 percent lower at the end of a civil war than it would have been in the absence of conflict. This is relevant to our understanding of civil war recurrence because, as Walter (2004) rightfully points out, a fundamental difference between civil war compared to inter-state war is that civil wars are conditioned by the ability of rebel leaders to voluntarily recruit sufficient combatants for their armies. Unlike inter-state wars, 1 Introduction

10 they cannot rely on standing armies and so civil wars will have little chance to get off the ground unless individual farmers, shopkeepers, and workers voluntarily choose to enlist in the armies that are necessary to pursue war (Walter, 2004: p. 372). Although one should not dismiss the numerous instances of forced recruitment of child soldiers in particular during civil wars in Africa 1, I concur that this disparity between interstate and intrastate warfare is important to our knowledge and understanding of why civil wars often recur especially in the developing world. Since the end of the Cold War, there emerged an increasing awareness of the potential detrimental effects and threats to country s long-term peace and stability posed by large numbers of unemployed, often armed ex-combatants. The risk of them becoming spoilers to the peace process is highest when opportunity costs for rebellion are low and when it is hard for former fighters to sustain their livelihoods without returning to war. Raising such opportunity costs then, became an increasingly valued objective of post-conflict stabilisation efforts and peace-building operations. As every conflict tends to result in a large number of surplus fighting force, it became apparent there was a growing need for technical and financial support from the international community directed at programmes which aimed to reintegrate demobilized combatants into civil society (World Bank, 1993). Amongst practitioners and policy-makers, there emerged a widespread belief that reducing the number of combatants and weapons in a country through formal Disarmament, Demobilsation and Reintegration (DDR) programmes should be an integral component of peace building efforts. In 2006, the former UN Secretary-General Kofi Annan stated that the disarmament, demobilization and reintegration of former combatants and those associated with armed groups is a prerequisite for postconflict stability and recovery (United Nations, 2006: foreword). 1 The Lord s Resistance Army (LRA) in Uganda, for example, is thought to have forcedly recruited over children during its war with the government (Singer, 2006; Briggs, 2005). Introduction 2

11 Disarmament, Demobilization and Reintegration (DDR) programmes According to the United Nations Disarmament, Demobilization and Reintegration Resource Centre 2 (2015), the DDR process consists of three phases. The first phase, disarmament constitutes the collection, documentation, control and disposal of small arms, ammunition, explosives and light and heavy weapons of combatants and often also of the civilian population. Disarmament also includes the development of responsible arms management programmes. The following stage, demobilisation means the formal and controlled discharge of active combatants from armed forces or other armed groups. The first stage of demobilization may extend from the processing of individual combatants in temporary centres to the massing of troops in camps designated for this purpose (cantonment sites, encampments, assembly areas or barracks). The second stage of demobilization encompasses the support package provided to the demobilized, which is called reinsertion. The final stage, reintegration, is the process by which ex-combatants acquire civilian status and gain sustainable employment and income. Reintegration is essentially a social and economic process with an open time-frame, primarily taking place in communities at the local level. It is part of the general development of a country and a national responsibility, and often necessitates long-term external assistance (UNDDR Resource Center, 2015). The three programmatic phases generally follow consecutively. Disarmament is the first step aimed at reducing insecurity and instilling trust between formerly warring parties by reducing the number of available weapons. Demobilization starts when combatants are cantoned, often in camps, to allow for registration and a thorough needs-assessment. This phase is often concluded by repatriating ex-combatants to their home communities or another area of their choice. Reintegration can entail offering education- or cash-for-work programmes, distributing farmland, and psychological assistance for ex-combatants as to allow for the successful absorption of former fighters in communities. The sequencing of these phases, however, sometimes varies. For example, during the DDR campaign in Tajikistan, wide perceived as a success story, reintegration of former rebels started prior to their disarmament 2 Website: 3 Introduction

12 (Banholzer, 2014). It is believed that this has greatly benefitted the success of the programme as it allowed for the cultivation of trust between the government and rebels and increased their willingness to participate in the programme (Matveeva, 2012). The increasing importance of DDR programmes as part of peace building efforts is clearly expressed in numbers. Since 1989, DDR programmes have been implemented from El Salvador to Cambodia and from the Balkans to Mozambique. Schulhofer-Wohl and Sambanis (2010) report that DDR programmes were implemented after 51 civil wars, 38 of these DDR programmes followed civil wars ending after Similarly, Banholzer (2014) notes that all of the past seven peace-keeping operations mandated by the Security Council included a DDR programme. DDR is recognized as such if there is an explicit, formal programme set up by the government or international community committed to reducing the number of combatants in a country. DDR is therefore most commonly implemented in response to, and in accordance with, the desires of the national government. However, it can also, in rare instances, be imposed on a country by the international community or neighbouring countries by force (e.g. Haiti, ). Finally, the Human Security Report Project (2010) reports that while the number of international peace missions has only tripled, the number of DDR operations has seen a ninefold increase since the end of the Cold War (Human Security Report Project, 2010). Consequently, a growing body of literature focuses on the success and failure of DDR programmes. Yet, despite increased academic attention to the effects of these programmes on post-conflict societies, most (indeed nearly all) studies of DDR constitute single or small-n case studies resulting in insightful information on the conditions leading to the programme s success (e.g. Colleta, Kostner and Wiederhofer, 1996; Thruelsen, 2006; Dercon and Ayale, 1998; King, 2000; Matveeva, 2012; Paes, 2005). However, unfortunately, very few scholars have sought to systematically investigate whether the conditions that lead to the success of one DDR programme also apply to others. The lack of systematic analysis of DDR means that, while we know a lot about why certain individual programmes succeeded or failed, we know little to nothing about the conditions that foster or hamper the success of DDR in general. As Humprhey s and Weinstein (2005: p. 2) point out, there have been few systematic efforts to Introduction 4

13 evaluate the determinants of successful reintegration by ex-combatants after conflict. Likewise, Schulhofer-Wohl and Sambanis after reviewing the literature on DDR conclude: we found little evidence of systematic assessments (Schulhofer-Wohl and Sambanis, 2010: p. 4). Banholzer (2014) observes a dearth of systematic empirical evidence that particular regime types actually foster or hamper DDR efforts (Banholzer, 2014: p. 20). In fact, it seems that Banholzer (2013) is the sole study that has sought and was successful in isolating the positive effects of DDR on the durability of peace in the first place. Clearly, further research is needed in order to foster a robust and more comprehensive answer to the question which causal conditions lead to the success or failure of DDR? This study aims to advance our knowledge of DDR in three ways. Firstly, it provides a synthesis of existing knowledge on DDR and peace building in general. Secondly, it uncovers some of the conceptual problems one encounters when systematically comparing DDR programmes. Thirdly, this study employs an original medium N-size qualitative comparative approach to a research problem previously addressed exclusively through small-n research designs. This paper proceeds as follows: the next section delineates the existing literature on peacekeeping missions in general and DDR in particular. Section III presents an argument for Qualitative Comparative Analysis (QCA) over Inferential Statistical Methods (ISMs). Section IV presents the case selection and operationalization of the causal conditions. Section V presents the main findings of this study and the final chapter is reserved for concluding remarks. 5 Introduction

14 2. State of the Art DDR programmes and peacekeeping operations share the same main objective: the prevention of recurring conflict. Therefore, this study draws upon literature from both peacekeeping and conflict resolution in general, as well as DDR programmes in particular. The first section of this review delineates the main findings from the literature on conflict resolution, while section 2.2 does the same for DDR. However, as the previous section already expounded, the overall aims of DDR programmes tend to be more comprehensive than merely intending to keep the peace. Section 2.3 elaborates on this by revealing an important conceptual problem related to evaluating the success of DDR programmes and offers a solution. This chapter will conclude with the formulation of the research question. Literature on conflict resolution Broadly, scholars of conflict resolution have tended to focus on three main questions in order to explain civil war recurrence. The first two questions relate to the conflict environment while the latter is concerned with the outcome of war, and particularly in the absence of a decisive victory, the design of peace-agreements (Hartzell, Hoddie and Rothchild, 2001). Thus, three important questions regarding the recurrence of civil war are: 1) What was the initial war fought over? 2) How was it fought? 3) How did it end? Regarding the first question, Gurr (2000) finds that conflicts fought over issues of ethnicity are more persistent and more likely to recur than wars fought over territory or over competing ideologies. Gurr argues that cultural identities are stronger and rooted at a more deep-seated level than other civic and associational identities. As a result, conflicts over such issues are more difficult to resolve as identity issues become increasingly intractable. Rothchild and Groth (1995) point out that conflict between different identity groups often solidify feelings of hatred and suspicion, impeding conflict solutions that require cooperation and coexistence. 6 State of the Art

15 The second question relates to characteristics which materialise during the war. Firstly, the duration of the initial war seems to have an impact on post-conflict stability. Protracted conflicts are more likely to result in lasting peace than short wars (Hartzell, Hoddie and Rothchild, 2001; Walter, 1997). Mason, Weingarten and Fett (1999) argue that longer conflicts allow opponents to gather more information on each other s capacity and willingness to fight. As war drags on, conflicting parties become more aware of their opponent s capacity to fight and realise that they are increasingly fighting towards a stalemate. Hartzell, Hoddie and Rothchild (2001) argue that adversaries in a war juxtapose potential costs (and gains) of further conflict with the payoff of a settlement. Consequently, the prospect of a negotiated settlement becomes more attractive as war protracts, increasing the political will for a peaceful solution. Secondly, the intensity of the conflict can act as a predictor variable for the duration of peace (Hartzell and Hoddie, 2003). Hartzell and Hoddie find that one unit increase in conflict intensity (measured in monthly-battle-related-deaths) increases the probability of settlement failure, and thus war recurrence, by 124% (Hartzell & Hoddie, 2003: p. 328). The theory underpinning this is that conflicts characterised by a higher intensity in the fighting lead to more severe grievances which are harder to settle once the conflict ends. Thus, short and intense wars have a higher probability to recur than less intense and protracted conflicts (Smith & Stam, 2002). Regarding the termination of conflict, Licklider (1995) employs statistical analysis on civil wars ending between 1945 and 1991 to test Wagner s hypothesis that wars ending in decisive victory lead to longer spells of peace than those that end in a negotiated settlement (Wagner, 1993). While wars that end in a decisive victory tend to annihilate the opponent s organisation and thus the ability to retaliate, wars that end in settlements leave open the possibility of returning to war as both parties organisation is left intact. As the number of both civil wars and peace settlements increased during the 1990s, a second group of scholars focussed on the design of peace-agreements to see which were more or less likely to hold. Here, two robust findings in the literature can be discerned: the involvement of third-parties as security-guarantors and the inclusion of certain power-sharing State of the Art 7

16 arrangements. Barbara Walter s credible commitment theory (1997, 2002) posits that conflict resolution in civil wars operates under a decidedly different dynamic from interstate wars. In civil wars, adversaries cannot credibly promise to abide by the arrangements struck in negotiations. For rebels, signing a peace accord often entails relinquishing the only means those rebels have to enforce that very same contract, leaving them vulnerable to acts of deceit and opportunism by the government. As Walter explains; at a time when no legitimate government and no legal institutions exist to enforce a contract, [the rebels] are asked to demobilize, disarm, and disengage their military forces and prepare for peace (Walter, 1997: p ). She identifies third-party guarantees as the critical barrier for the successful implementation of peace agreements because only third-party involvement in the peaceprocess can overcome this prisoner s dilemma by overseeing the implementation of the peace agreement and by threatening to punish potential violators or spoilers 3. Extending this theory to the success or failure of DDR programmes, we can expect that accompanying DDR with the deployment of peacekeepers is an important, perhaps even necessary, condition for success. However, to date, no systemic analysis of DDR programmes has been carried out to test whether this hypothesis is valid. Hartzell and Hoddie (2003) emphasize the importance of power-sharing provisions included in post-conflict arrangements 4. Power-sharing provisions are designed to give the opposition a stake in the future of the country rendering peaceful solutions to disagreements more appealing. Such measures can be implemented along political, territorial, military or economic dimensions (Hartzell and Hoddie, 2003: p.320). The political dimension can involve the allocation of a certain number of parliamentary seats to the formerly armed opposition (e.g. Lebanon). In terms of territory, a government can make concessions by granting varying levels of autonomy and increasing the basis of federalism thus limiting the states capacity to exercise influence over areas controlled by former rebels (e.g. Kurdistan in Iraq). A military solution can be found in the creation of a new national armed forces by integrating former rebel factions (e.g. RENAMO rebels in Mozambique in 1992). Finally, along the economic 3 See Barbara Walter s (2002) Committing to Peace: The Successful Settlement of Civil Wars for a book-length substantiation of her theory. 4 See also Glassmyer and Sambanis (2008). 8 State of the Art

17 dimension, the central government can commit to a revised distribution of economic resources controlled by the state (e.g. Kurdistan in Iraq). What these measures have in common is that they, on the one hand, reduce fear among rebel factions by diminishing the power of the central government and, on the other hand, increase the participation of rebels in the future of the country, giving potential spoilers a stake in the peace process. The implementation of DDR programmes is sometimes accompanied by one or more elements of power-sharing, in particular the inclusion of rebel groups in the formation of a new national armed force. Building on Hartzell and Hoddie s findings, we can expect that the inclusion of such measures is conducive to the success of DDR efforts. The case for DDR The concept of DDR is not entirely new. Demobilisation of large numbers of soldiers took place, for example, after the First and Second World War. However, after WWI the failure to collect large numbers of weapons and transition soldiers back into civil society is believed to have contributed to partisan struggles and hampered the establishment of a democratic system of governance in Germany (Banholzer, 2013). This encouraged the US to improve the process after WWII. Soldiers were gradually rather than instantaneously demobilised, and allowed to finish their college degrees before formal discharge in order to facilitate a better absorption into the labour market (Banholzer, 2013). However, the formal and more comprehensive form of DDR as it is implemented today only emerged in the late 1980 s and 1990 s, mostly as a response by the international community to the increasing prevalence of intrastate wars in relation to the number of interstate wars. During the 1990 s especially, as the international community increased its involvement in peace-building and post-conflict stabilisation efforts, the number, size and scope of DDR programmes grew exponentially (Banholzer, 2014). Although a number of DDR programmes have been initiated Asia, Europe and the Americas, by far the largest number of DDR operations have been implemented in Africa. The United Nations Department of Peacekeeping Operations (UNDPKO) and the United Nations Development Programme (UNDP) as well as the World Bank play key roles in supporting national governments with the planning, funding and implementation of DDR. Continued improvement State of the Art 9

18 of the process led the UN to publish its Integrated Disarmament, Demobilization and Reintegration Standards (IDDRS) in DDR is best understood as a tool used to facilitate war-to-peace transitions by helping individuals return to civilian life, and decisively make a break with the past (Casas-Casas and Guzman-Gomez, 2013). By assisting former combatants, DDR programmes aim to raise the opportunity costs of fighting and make sure that they need not return to war in order to sustain their livelihoods. DDR also contributes to state-building as it aims to both re-emphasize the importance of institutional laws defined by the central government and to break down the formal and informal command structures of militant groups (Banholzer, 2014). Arguably, the needs of ex-combatants in conflict-torn societies do not differ greatly from other vulnerable target groups within the country, such as the poorest communities, internally displaced people and refugees (Porto, Alden and Parsons, 2007). In LDC countries, both ex-combatants and the society s poorest tend to be illiterate, have little to no access to health care, few or no non-military skills, limited access to land and a small social network due to the loss of community- and family members (Colleta et al., 1996a). Therefore, it is important that any DDR efforts do not create tension over disparities between civic communities and excombatants. The UNDPKO is increasingly aware of this hazard and argues that one of the key challenges in designing and implementing DDR programmes is how to fulfil the specific and essential needs of ex-combatants without turning them into a real or perceived privileged group within the community (UNDPKO, 2015: section 4.3: p. 3). Yet, there remain important reasons for targeted assistance to ex-combatants in particular (Knight and Ozerdem, 2004). What distinguishes former combatants from other vulnerable groups is their access to (heavy) weaponry and the fact that they are part of an organised command structure. Hence, their ability to disrupt or impair the peace process is far greater. As Porto et al. explain in their book From Soldiers to Citizens, the risk of the former combatant becoming a potential spoiler by resorting to violence, criminal activity and predatory behaviour justifies that while other war-affected groups such as refugees and IDPs may far 5 Available online at: (accessed ). 10 State of the Art

19 outnumber them, ex-combatants will usually need focused, sustainable support if they are to succeed in making the transition from military to civilian life (Porto et al., 2007: p. 27). Similarly, Douglas et al. (2004) warn that the exclusion of ex-combatants from society may encourage former fighters to resort to violence and criminal activity as a means of sustaining their livelihood. The validity of this threat is highlighted by post-conflict developments in countries such as El Salvador, Cambodia, Nicaragua, the Balkans and Angola (Porto et al., 2007). It should be noted however that, in a case-study on Mozambique, research has shown that it is not the majority of former fighters who turn to a life of crime but mainly the medium to high ranking officers who engaged in criminal activity such as drug- and arms trafficking (Lala, 2005). Further justification for the implementation of DDR programmes is the much needed psychological support they provide former fighters. As Banholzer points out, the experience of combat and living in extreme deprivation and constant fear and stress does not leave former combatants unharmed (Banholzer, 2014: p.6). Comparative studies conducted between former fighters and non-combatants found evidence of severely disruptive psychological trauma among ex-combatants. The symptoms include higher suicide rates, unemployment due to incapability of performing steady income-generating tasks, higher divorce rates, spousal and general violence, homelessness, criminal behaviour and an increased substance abuse/addiction (Winkler, 2010; World Bank, 2006). The need for a successful reintegration programme to address psychological issues among ex-combatants has only recently been included by the UNDPKO in the IDDRS: The widespread presence of psychological problems among ex-combatants and those associated with armed forces and groups has only recently been recognized as a serious obstacle to successful reintegration... If these ex-combatants do not receive adequate psychosocial care, they face an extraordinarily high risk of failing in their reintegration... Reintegration programmes should seek to prioritize psychological and physical health rehabilitation as a key measure to successful reintegration (UNDPKO, 2015: section 4.3: p. 46). The risk of former fighters becoming spoilers may also stem from a feeling of dissatisfaction due to a lack of what Schafer (1998) termed compensatory justice. Some State of the Art 11

20 fighters may feel that, as they ve served and defended their state or interest group, they are entitled to some form of compensation (Rolston, 2007). Especially since combatants returning to their home communities often have to rely on informal support structures provided by the family and community, it is important that they are not perceived as returning empty-handed. Acceptance by the community depends largely on whether returning fighters are perceived as a drain on scarce resources or as a productive asset to the community at large (Knight and Ozerdem, 2004). Useful elements of a reintegration package can thus include payment of educational fees for children of ex-combatants, health care support for both ex-combatants and their dependents, and economic assistance (either in kind or cash) aimed at providing food and shelter (Knight and Ozerdem, 2004). DDR studies and their main findings A large number of evaluating reports present lessons learned from past experiences. While these evaluations are important and sometimes contain noteworthy policyrecommendations for the future, they also frequently conclude with unclear, generalist and rather obvious points. For example, on a case study of Congo-Brazzaville, one report concluded The MDRP found that successful DDR required commitment from the parties to the conflict, effective demobilization, combating reintegration difficulties, focusing assistance, creating longterm employment, facilitating social reintegration for extended families, preventing violence and the stigmatization of ex-combatants (Carames, 2009: p. 61, emphasis added). Additionally, the Praxis Group (2000) concludes that the calibre of UN staff is important to the programme s success. The World Bank (1993: p.12) emphasises the importance of a commitment by all concerned parties, and both Kingma (1997) and Colletta (1999) identify political will and a sound needs assessment of former combatants among the main determinants for success. It is clear that, without political will, an accurate needs assessment and adequately trained staff, any assistance programme is likely to fail. Unfortunately, many of these reports and evaluations forego the more important question of why such conditions were absent. 12 State of the Art

21 Figure 1: Root causes and symptoms leading to the failure of DDR. Nevertheless, despite of the occasional open-door conclusions presented in evaluating reports and case-studies, a number of important causal conditions can be discerned from the literature on DDR. Opportunity costs In accordance with Collier & Hoeffler s theory on the opportunity costs of rebellion, a frequently cited reason for success or failure of DDR are the economic conditions under which they are implemented (Banholzer, 2013; Kingma, 1997; Kingma & Grebrewold, 1998; Rolston, 2007). DDR programmes are specifically meant to raise the opportunity costs of rebellion to such an extent that leading a civilian life becomes more appealing than resorting to violence, at least from an economic perspective. However, in war-torn societies, shattered and dysfunctional economies may jeopardize the successful reintegration of combatants when the labour market is unable absorb a large influx of new job-seekers (Kingma, 1997). When former combatants are reintegrated into poverty they may resort to criminal activities or re-join armed forces in an attempt to sustain their livelihoods (Knight and Ozerdem, 2004: 516). Banholzer, Schneider and Odenwald (2013) find that, in Somalia, combatants were more willing to disarm if they could return to an economically stable region in the country. In the DRC, State of the Art 13

22 reintegrated combatants trained as carpenters declared they were deliberating rejoining rebel groups as their businesses suffered from a perpetual lack of customers (Banholzer, 2013). Not surprisingly, good economic conditions thus seem to favour the success of DDR. Indeed, a systematic analysis of DDR programmes could reveal a particular level of economic development as a necessary condition for successful DDR. However, it is important to note that DDR programmes are distinct from (long-term) development programmes and are not able nor designed to be a tide that lifts all ships by improving a country s overall economic position. Rather, such a finding would lend empirical support to Kingma s conclusion that DDR should be implemented as part of a more multi-dimensional peace-building effort which includes long-term development goals and targets wider communities rather than individuals (Kingma, 1997b: p. 163). Third-party involvement In line with the literature on peacekeeping, the importance of third-party involvement with DDR is subject to debate. Alden (2002) argues that external actors can encourage participation and strengthen commitment to the process. Additionally, Knight and Ozerdem (2004) argue that involvement by the international community can signal a decisive end of hostilities. In the case of Liberia, Jennings (2007) argues that presence and supervision by the UN during the DDR process was a decisive condition for the programme s success. Banholzer (2014: p. 21) reports that, in Mozambique, RENAMO soldiers said they would not demobilise until the blue helmets had arrived. It seems then, that the deployment of an international peacekeeping or stabilisation force is beneficial to DDR goals. Although third-party actors are almost always involved in DDR programmes to some degree, it seems unclear whether it is best for a programme to be under national ownership of the host-country, or led by the international community. Proponents of national ownership argue that it gives former warring parties a sense of control and ownership in the programme, which helps prevent them from turning into spoilers of the peace process (Banholzer, 2014). In a case-study on Ireland, Rolston (2007) asserts that national ownership was one the prime determinants for the programme s success. In principle, the UN encourages national ownership 14 State of the Art

23 as a means of capacity building. Sceptics however, argue that local governments and institutions can lack the necessary strengths and governing abilities to successfully implement DDR (Thruelsen, 2006). For example, problems surrounding corruption and a lack of capacity regarding local partners have been reported in DDR processes in DRC, Mozambique and Afghanistan (Handson, 2007; Striuli, 2012; Giustozzi, 2012). Political participation Similar to third-party security guarantees, the need to reduce fear among demobilised combatants is what underlines the hypothesis that opportunities for political participation are important to the success of DDR programmes. It is likely that combatants are more willing to disarm if they feel that they can contribute to a political solution (Banholzer, 2014). Kingma and Grebrewold (1998) argue that democracies equipped with a functioning legal system are important for the success of DDR initiatives as they can help prevent the escalation of minor disputes at an early stage. However, to date, no systematic study has been carried out to investigate the potential link between regime type and DDR (Banholzer, 2014: p. 19). Timing of DDR Finally, the timing of the programme is an influential factor to its success. DDR can be implemented whilst hostilities are still ongoing with the aim of exerting a pacifying effect on the conflict or after war has ended in order to stabilise society. In a rare cross-sectional comparative study on DDR, Banholzer (2013) finds that programmes which are initiated while the conflict is still ongoing have a significantly higher chance of failing than programmes implemented after hostilities have already ceased. She concludes that DDR is better used as a stabilizing tool in post-conflict societies than as a means to end an ongoing war. Banholzer s study constitutes, to my knowledge, the only comparative study focussing explicitly on DDR programmes. Although thorough and informative, it also reveals some conceptual problems related to the systematic assessment of DDR programmes. In particular, one can debate what constitutes success and failure in DDR, as the following section will elaborate. State of the Art 15

24 Conceptualizing success There is no clear-cut definition of what constitutes success for a DDR programme (Williams, 2014). Banholzer (2013) argues that the most straightforward way to determine success is to see whether the primary objective of DDR the establishment of lasting peace was achieved. However, this approach is problematic for two reasons. Firstly, as the case of Cambodia illustrates, this definition of success is too narrow. In Cambodia, according to the Uppsala Conflict Data Programme (UCDP), UNTAC s DDR programme was followed by lasting peace 6. A closer look at this case, however, reveals that DDR in Cambodia was an unambiguous failure. In fact, while it was intended that some combatants were to be demobilised, the programme was actually suspended and then cancelled, after only combatants were cantoned. Consequently, a case study performed by the International Center of Transitional Justice (ICTJ) reports that much of the countryside experienced high levels of insecurity and warlike conditions throughout UNTAC s tenure (Bartu and Wilford, 2009: p.12). The endurance of peace therefore, cannot be automatically equated with successful DDR. The other side of the same problem is that it is perfectly plausible for a country to experience war or conflict despite a successful DDR programme. As section 2.1 has shown, there is a plethora of causal factors leading to a return to conflict. It is therefore inaccurate to automatically attribute the presence or absence of peace to the success or failure of DDR. As Schulhofer-Wohl and Sambanis (2010: p. 3) rightfully point out, given the presence of many confounding factors the DDR program is not necessarily connected to the observed outcome. Even if one defines success more broadly, and includes measures of residual violence and crime which fall short of war, problems persist. One issue with such an approach is that the effect of DDR on the observed outcome is likely to differ between cases. More precisely, the chances of failure of DDR would increase with a programme s scope. Consider, for example, a DDR programme in country A with a limited scope targeting 3000 combatants, while DDR in country B targets combatants. DDR in country A fails to reach any combatants while programme B demobilises and reintegrates fighters. It is likely then, 6 UCDP defines conflict as 25 battle-related-deaths while war constitutes 1000 battle-related-deaths. The case of Cambodia is coded as followed by lasting peace and thus deemed successful in Banholzer (2014). 16 State of the Art

25 ceteris paribus, that the remaining combatants in country B have a greater destabilizing effect in terms of crime rates and violence than the combatants in country A. From a programming-perspective however, one would argue that DDR in country B was more successful. In order to bypass these complications, I propose an alternative conceptualization of success. I agree with Williams (2014: p.2) that one of the ways to evaluate the success or failure of any programme is by looking at its aim and objectives in the first place. Hence, a quantitative measure of the number of combatants demobilised and reintegrated as a proportion of the number of combatants initially targeted provides a primary indication of the programme s success. In other words, to what extent has the programme succeeded in meeting its stated goals and objectives? Secondly, and complementary to the first step, it remains important to assess whether lasting peace was achieved. The final and critical step in determining the success or failure of DDR is then to link the observed outcome with the performance of the programme. If a programme succeeds in largely meeting its objectives, but the outcome is still war recurrence, then qualitative investigation of the case must be carried out to establish whether or not the resumption of violence can be traced to failures in the implementation of DDR. Ideally, one would want to measure success in terms of success stories it has produced from the perspective of individual ex-combatants, communities and visible transformation (Williams, 2014). However, while such an approach works best for case-studies with a limited sample size, it lacks pragmatism for studies as this one which encompass a larger crosssectional scale. The alternative measure for success proposed here succeeds in avoiding many of the aforementioned difficulties while still allowing for its application in a systematic and comparative study. State of the Art 17

26 3. Methodology: ISM vs QCA This study adopts a comparative approach to systematically assess when DDR programmes succeed and when they fail. The main aim of this study is to investigate whether the causal conditions presented in case studies as leading to the success or failure of DDR are robust at a larger cross-sectional scale. Hence, this research objective precludes the use of small-n comparative methodologies such as Most Similar Systems Design (MSSD) or Most Different Systems Design (MDSD) 7. It is unlikely that such an approach to this research problem would significantly improve or extend on present knowledge as the majority of research on DDR already consists of (comparative) case studies. Two comparative tools available to the researcher and suitable for larger-n studies are Inferential Statistical Methods (ISMs) and Qualitative Comparative Analysis (QCA). However, the two methodologies come with their own set of strengths and limitations. Moreover, as will be discussed below, QCA and ISM are fundamentally different in their underlying epistemology and ontology. This section evaluates the differences between QCA and traditional ISMs such as Ordinary Least Squares (OLS) regression in ontological, epistemological and practical terms and provides an argument for employing a fuzzy-set QCA (fsqca) methodology. In terms of epistemology, QCA differs from regression analysis in the unit that is central to analysis and the way in which knowledge on social phenomena is derived from it. In regression analysis, inferences are made based on correlational connections between variables. A variable-orientated approach could, for example, conclude that poverty is correlated to crime. Or that level of development is correlated to democracy. A case-orientated approach however, would conclude that criminals tend to be poor, and that economically advanced nations tend to be democratic. The difference perhaps seems subtle, but it has profound implications for the way in which social phenomena are explained. By focusing on the connections between variables, regression analysis tends to obscure cases behind indicator values. Cases then become carriers of information on the relations between dependent and 7 For a comprehensive discussion on the logic behind Mill s methods see Pennings, Keman and Kleinnijenhuis (2006: Ch. 3) 18 Methodology

27 independent variables. The researcher acquires knowledge on cases by looking at the relations between variables after all selected cases are stacked and pooled together. In QCA, the cases remain separated and thus intact as the main unit of interest. Knowledge is then acquired by analysing the relation between the outcome and different combinations of conditions present in each individual case. This difference touches upon the contrasts between QCA and regression analysis in terms of their underlying ontology of causality. There are four important differences between QCA and standards ISMs relevant to this study: 1. Set-theoretic thinking versus correlational connections 2. Calibration versus measurement 3. Configurations of conditions versus independent variables 4. Equifinality versus exclusive causality Firstly, it is important to note that employing correlational methods to establish causal connections is to maintain that causal effects are symmetrical. Contrastingly, thinking of causality in set-theoretical terms is generally asymmetrical. To use the previous example, a theory based on correlational connections that posits economic development is correlated with democracy is undermined by cases of less developed yet democratic countries. Such cases undermine the correlation coefficient and therefore weaken the central claim of the theory. The set-theoretic claim of this theory on the other hand, posits that economically advanced countries are a subset of the set of democratic countries. The claim that the causal condition is a subset of the outcome is asymmetrical and, as such, remains untouched by cases of poor and democratic countries. In the case of this study, if we hypothesize that third-party involvement is correlated with successful DDR, cases in which DDR succeeds despite a lack of third-party involvement would undermine the correlational connection. However, this theory is essentially set-theoretic in the sense that it claims that cases with third-party involvement are a subset of cases with successful DDR (i.e. a sufficient condition). The notion that this condition constitutes a subset of the outcome (successful DDR) also touches upon the assumption of equifinality, to which I will return to later. Methodology 19

28 One could also hypothesise, based on Walter s theory, that third-party involvement is a necessary condition for the success of DDR. In this case, the set-theoretic claim is that the condition (third-party involvement) constitutes a superset of the outcome. Hence, while cases of failed DDR with third-party involvement would undermine the correlational connection, it does not impede upon the set-theoretic claim of the theory. Figure 1 is a graphic depiction of set-theoretic necessary and sufficient conditions. Figure 2: Venn diagram depicting necessary (left) and sufficient conditions (right) In QCA, cases are understood as members of sets. However, the way in which the researcher assigns membership to cases varies. There are three variants of QCA research: Crispset (csqca), Fuzzy-Set (fsqca) and Multi-Value (mvqca). In crisp-set QCA, cases are coded in binary terms either as full members of a set or as full non-members (Ragin & Pennings, 2005). For example, a state is either a member of developed states (1) or not (0). Hence, crisp-set QCA does not allow for graded concepts or partial membership. The advantage of this is that it allows for clear formulations of sufficient conditions using Boolean algebra. For this research, employing csqca would mean coding cases as either democratic (1) or not (0) and as economically developed or not. Indeed, the problem with this approach is that many concepts in political science, also in this study, are best understood as graded concepts (Ragin & Pennings, 2005). fsqca provides a solution as it allows for more imprecise boundaries of membership and non-membership often encountered in social science (Klir and Yuan, 1995). But how does one determine these membership scores? 20 Methodology