Deciphering the Dimensions of Reintegration in Post-Conflict Liberia July 2007 James Pugel1

Abstract How does the international community evaluate a Disarmament, Demobilization, and Reintegration (DDR) program? To date, the only tangible measure of success available to donors has been founded in the targeted country’s ability to maintain peace after intervention. The ambiguity inherent in the evaluation of a DDR program stems mostly from an inability to quantify suitable outcome measures – especially those concerned with reintegration. This chapter will investigate the enigma of ex-combatant reintegration with the purpose of codifying measurable outcomes, determining the extent that influencing factors play upon the dimensions of reintegration, and will quantify the interim impact of the DDR program in Liberia. While the chapter focuses on the details of the Liberia conflict, the methodologies and concepts advanced are believed to be transferable to other post-conflict scenarios. Quantifiable outcomes are essential for program evaluation and subsequent policy formulation. Fundamentally, the chapter demonstrates the importance of understanding the dimensions of ex-combatant reintegration in a way that can be translated into quantifiable outcome measures for practitioners and decision makers.

1

The chapter is drawn from a data collection effort sponsored by the UNDP Liberia while the author was doing thesis research at the Joint Military Intelligence College. The author is a research associate with the Centre for the Study of Civil War (CSCW), International Peace Research Institute of Oslo (PRIO). Many thanks go to Dr. Scott Gates and Kendra Dupuy for their helpful comments concerning this chapter and to Jamie Fuller for her keen editorial eye. The views expressed in this chapter are those of the author and do not reflect the official policy or position of the Department of Defense or the US Government. Author takes responsibility for all errors and inconsistencies.

I Introduction

How does the international community evaluate a Disarmament, Demobilization, and Reintegration (DDR) program? To date, the only tangible measure of success available to donors has been founded in the targeted country’s ability to maintain peace after intervention. The ambiguity inherent in the evaluation of a DDR program can be linked to an inability to quantify suitable outcome measures, especially those concerned with reintegration. This chapter will investigate the enigma of ex-combatant reintegration with the purpose of codifying measurable outcomes, determining the effect that influencing factors have on the dimensions of reintegration, and quantifying the interim impact of the DDR program in Liberia. Quantifiable outcomes are essential for program evaluation and subsequent policy formulation.2 Fundamentally, the chapter demonstrates the importance of understanding the dimensions of ex-combatant reintegration in a way that can be translated into quantifiable outcome measures for practitioners and decision makers. Utilizing a data set of 590 ex-combatants from Liberia that was collected randomly from a nationwide sample, this chapter will address the research question through an examination of the literature surrounding reintegration and the conjecture on unique dimensions. Facets of reintegration will be defined and prepared for analysis by operationalizing outcomes determined to yield successful measures of positive impact on the lives of former fighters. A variety of factors— individual demographics, wartime experiences, and community factors—will then be modeled as a regression to evaluate influence in a given dimension of reintegration and determine the impact of the national DDRR program upon the lives of those who registered for benefits.

2

The Handbook on Monitoring and Evaluating for Result (UNDP, 2002) articulates the need for quantifiable outcomes as a cornerstone of the organization’s approved management strategy – Results Based Management (RBM).

2

The chapter proceeds with an overview of the Liberian civil war that situates the postconflict challenges that the international intervention program faces. Section 3 examines the discussion of reintegration dimensions in the literature and defines three components—societal, economic, and political—for further analysis. Section 4 identifies a number of influencing factors that impact the reintegration process and hypothesizes the relationship between them and outcomes of reintegration success. The details of the data collection effort in Liberia are then explained and followed with a section that seeks to operationalize reintegration success outcomes for each of the three dimensions selected for analysis. Sections 7 through 9 constitute the analysis portion of the chapter and present reintegration-related findings associated with the specified dimensions, influencing factors, and ultimately the impact of the DDRR program. A concluding section summarizes the key findings presented in this chapter and reinforces the necessity for rigorous evaluations in all future international intervention program analyses.

II The Years the Locusts Have Eaten3 On 2 March 1807, through an Act of Congress, the importation of slaves into America had been strictly prohibited.4 Unfortunately, slaves that were smuggled in fell into a loophole that was not enforceable by the statute. It was as a direct result of attempting to mitigate the effects of this loophole that the U.S. House of Representatives passed a bill empowering the president of the United States “to make such regulations and arrangements as he may deem expedient for the safekeeping, support, and removal beyond the limits of the United States, of all such negros . . . and

3

This heading is a reference to the title of a historical narrative of Liberia and her people. From the inception of the Liberian state by the American Colonization Society (ACS) in 1822 until the exile of Charles Taylor in August of 2003, the author argues that “Liberian leaders never made a pungent attempt to establish and maintain democratic institutions. The nation remained mired in a labyrinth of self-inflicted wounds, brought on by authoritarian rule, rampant corruption, ethnic hatred and intolerance, and anarchism." Joseph K. Tellewoyan, The Years the Locusts Have Eaten: Liberia 1816-2004 (USA: Xlibris, 2005). 4 J.H.T. McPherson, History of Liberia (USA: Kessinger Publishing, 2004), 11.

3

appoint a proper person or persons residing upon the coast of Africa as agent or agents for receiving the negros.”5 With this authorization, President Monroe on 25 April 1822 empowered the American Colonization Society (ACS) and a contingent to establish a foothold on the African continent, and former slaves now began to establish a new home. In 1847 the settlers claimed independence and began the formation of an oligarchic state that would ultimately keep the True Whig Party in power until the bloody coup and assassination of President William Tolbert in 1980. The systematic oppression of the indigenous peoples of Liberia coupled with fifteen major deadly conflicts can be easily identified as the precipitator of over 158 years of tension.6 As former master sergeant and coup conspirator Samuel Doe took office as the new leader of the Liberian Republic, military oppression constricted the society, and the ethnic Khran, Doe and his “emerging ethno-autocracy did not purport to be democratic but acquiesced in the politics of violence and domination.”7 By 1989, rebellion was ripe again, and Charles Taylor led a multi-ethnic coalition to remove Doe from power. In 1990, rebel leader Prince Johnson of the Independent National Patriotic Front of Liberia (INPFL) captured Samuel Doe and killed him, which created a leadership void for almost seven years. During the period between 1990 and the elections of 1997 (Taylor elected president), fourteen peace accords were attempted by the international community in order to bring peace to Liberia.8

5

McPherson, 11. Liberia is a country composed of 16 major ethnic tribes. Despite the efforts of many of these tribes to strive for integration with the newly arrived settlers, a “competitive nation” was established instead. Traditional settler or “Congo” rivals included the Bassa, Dei, Grebo, Gola, Kru, and Vai, who fought the conflicts of the nineteenth and twentieth centuries. It was the Krahn and Mandigo, however, that brought the indigenous-led coup in 1980 that turned Liberian society upside-down. Jeremy I. Levitt, The Evolution of Deadly Conflict in Liberia (Durham: Carolina Academic Press, 2005), 5-7 and 194. 7 Levitt, 202. 8 Charles Taylor won a landslide victory against the now president of Liberia, Ellen Johnson-Sirleaf, who at the time only captured 9% of the vote. Levitt, 210. 6

4

Prophetically, the election slogan “You kill my ma, you kill my pa, I will vote for you!”9 swept Taylor and a war-weary, cynical nation to electoral victory. Predictably, however, Charles Taylor’s problems began almost immediately upon his election. Distanced from the International Monetary Fund (IMF) for shady business practices and involved in constant disputes with ECOMOG (Economic Community of West African States Monitoring Group) over the rebuilding of the national army, Taylor was placed in difficult international standing. Domestic issues followed and threw Liberia back into civil war. The many years of conflict have destroyed Liberia inside and out. Infrastructure and basic services such as running water and electricity have been destroyed and are nonexistent. Liberia’s children, its most precious resource, have been savagely scarred and have suffered irreparable generational damage. With the signing of the Comprehensive Peace Agreement (CPA) in August 2003, all former warring factions, including the government troops, began the post-conflict process of Disarmament, Demobilization, Reintegration, and Rehabilitation (DDRR). On 19 September 2003 the United Nations Mission in Liberia (UNMIL) was established by Security Council resolution 1509 with a mandate to support the implementation of the cease-fire agreement. The National Transitional Government of Liberia (NGTL) was organized, with Monrovian businessman Charles Gyude Bryant assuming leadership duties as chairman. Multi-party elections were peacefully charted for and carried out in October 2005, and a constitutionally mandated two-candidate runoff occurred in November 2005. The inauguration of the first female African head of state took place in Monrovia on 16 January 2006, and the republic has since begun the process of reconciling and rebuilding. The UNMIL was able to disarm over 100,000 former combatants and secure approximately 28,000 9

"'You kill my ma, you kill my pa, I will vote for you', sung during the 1997 campaign season that ushered into power the NPP (Taylor's National Patriotic Party) by masses of the electorate in Liberia was a vociferous protest to the quasi-politicians who had nothing to offer our people but a plethora of half-baked reasons why they should not vote for the NPP, primary among was that they were killers. The song was a sincere and genuine

5

individual weapons. Gun violence has been almost nonexistent10 since the completion of the Disarmament and Demobilization phase. The dynamics of the former warring factions provide insight into the challenges that the reconciliation and peace-building process face as Liberia and the international community attempt to mend the socio-economic fabric of the country.11 During the contemporary stage of the conflict 1999-2003, there were three main factions that ultimately decided to disarm and disband – the Government of Liberia forces/militias, the Liberians United for Democracy and Reconciliation (LURD), and the Movement for Democracy in Liberia (MODEL). Each faction had unique social and political origins, organizational recruitment techniques and incentives, interaction with noncombatants, and interstate activities that characterized each warring faction. The core of Charles Taylor’s forces while he served as president was the group of fighters that served the leader when he began his insurgency in 1989 – the National Patriotic Front of Liberia (NPFL). Taylor’s military force was resident not only within the capital but also throughout the hinterland. Family protection and loyal support to Taylor’s political ideals were central reasons for joining. Taylor’s troops, both NPFL and the contemporary militia forces, typically utilized non-combatant forced labor and were quite brutal when the civilians did not submit to supporting their war efforts. Pecuniary incentives were used as the primary means of recruitment by the sub-factions under Taylor’s control. The two rebel groups LURD and MODEL had distinct geographic origins, Guinea and Lofa County, Liberia for the former and Cote d’Ivoire and Grand Gedeh County, Liberia for the latter. During operations the LURD maintained the lowest percentage of combat soldiers, when compared with the other competing factions, within their faction (71%). Dedicated camp workers made up

expression of the conviction of the people that they knew the NPP was the right choice and that they wanted the world to know it." URL:< http://www.theperspective.org/lina.html> accessed 3 February 2006. 10 LNP (Liberia National Police). 2006. Annual Crime Report and Statistics for 2005. October. 11 This section describing the dynamics of the warring factions in Liberia is derived from Pugel (2007b: 4243).

6

over 15% of the fighting force. Violent introduction was the predominant method employed by the LURD in securing new recruits, and coercion, through either abduction or scaring, was the primary reason for joining the group. Pecuniary incentives – job, food, money – were often promised as rewards promised for joining. The LURD also had a significant percentage of foreign fighters within their ranks. Alternatively, the MODEL had limited membership from Monrovia and had an attraction that induced a fair number of its members to actively seek them out and join. Fighters of this faction primarily joined to protect their families and non-pecuniary incentives topped the rewards scheme for those joining the MODEL. The faction was found to have the lowest reported rate of civilian impressments during the conflict and was the most likely to release those noncombatants who preferred not to support the faction’s efforts with their labor.

III Identifying the Dimensions of Reintegration

The literature abounds with qualitative postulations that speak to the dimensions of ex-combatant reintegration during the post-conflict recovery period. Ironically, authors usually address only a portion of the reintegration spectrum and never attempt to describe the total environment. In May 2005, the United Nations (UN) adopted the following definition for reintegration: 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.

Using the UN doctrinal definition as a starting point for the discussion of identifying dimensions of reintegration, one can clearly identify two components—“social” and “economic.” Many authors agree with and cite these two facets in their attempts to outline the reintegration construct, but are these the only two components of ex-combatant reintegration that the international community need worry about during the post-conflict intervention? Do the categories “social” and “economic”

7

accurately and adequately provide an accurate basis for a description of reintegration? With dimensions of reintegration defined, does the international community have operative outcomes to assist in measuring the efficacy of an intervention targeted to improve the situation in any given country? This section will explore the various questions posed above in order to establish an empirical framework for testing hypotheses-related measures of reintegration success. Myriad dimensions of reintegration are referred to in the literature on post-conflict intervention scenarios. A January 2000 UNDP internal evaluation recommends that the organization “should concentrate its support to political, social and economic reintegration of waraffected populations on restoring social and human capital while contributing to political and economic stability.”12 Kees Kingma agrees with the three components of reintegration highlighted by the UNDP evaluation team but goes further by noting that there are distinct psychological aspects as well.13 Mats Berdal acknowledges the importance of both the economic and social dimensions of reintegration but also argues that within a post-conflict environment these two areas are inexorably linked to political and security considerations.14 While our cursory examination of the literature tends to corroborate the necessity of categorizing reintegration along social and economic dimensions, political/security-related concerns as well as identification of a particular psychological component appear to require further investigation. The remainder of this section will examine the identified dimensions of reintegration in detail and identify operative outcomes that may be used to assist in measuring the efficacy of intervention programs. While the psychological dimension of reintegration has been tendered as a facet to be examined within the context of a complete

12

UNDP (UN Development Program). 2000. Sharing New Ground in Post-Conflict Situations: The role of UNDP in Support of Reintegration Programmes, January, 55. Accessed at . 13 Kees Kingma, “Demobilization, Reintegration, and Peacebuilding in Africa,” International Peacekeeping (Frank Cass) 9, issue 2 (Summer 2002), 183. 14 Mats R. Berdal, Disarmament and Demobilization after Civil Wars: Arms, Soldiers, and the Termination of Conflict, Adelphi Paper No. 303 (Oxford: Oxford University Press), 14.

8

reintegration construct, the author, unfortunately, does not have sufficient empirical data to investigate any further in this chapter.

SOCIAL VERSUS SOCIETAL The purpose of using a single word to describe a facet of reintegration is to focus practitioners, academics, and policy makers on the overarching purpose of a particular field of intervention—the specific term should enable an intuitive set of relatable outcomes to be derived. While the term “social’ appears at first glance to fulfill this role, further examination leads us to an alternate descriptive. After an in-depth study of ex-combatant reintegration, Anders Nilsson (2005: 27) concludes that reintegration as a field is “theoryless” and proposes an alternate definition:15 that it should be seen as a “societal process”. Historically, the literature on DDR utilizes the term “social” to characterize aspects of reintegration relating to the constructs of acceptance, community integration, and relationship capacities. Lichem (2006) contends that the international community is mislabeling this particular dimension of reintegration and engaging in post-conflict reconstruction from an imperfect datum. In a keynote address for a conference on post-conflict peace-building in Africa in Accra, Ghana, Lichem suggests that “societal” should be the term of choice as a baseline for informing DDR interventions: “Societal”…refers to the relational capacities of a human being or of a community, the capacities for being able to live with others in community, the capacities for religious, ethnic, political plurality, the capacity for peace, partnership, friendship, solidarity etc. It also includes the capacity to forgive and to include as well as the capacity to become integrated and included. The “societal” dimension is basic to any post-war rehabilitation and reintegration process. (Walther Lichem, 2006)

The contrast between the two terms lies in how Lichem defines “social”—the “productive capacities of a human being such as health, education, age, gender, hunger, poverty, etc.” The inadvertent misleading that results from an improperly defined dimension has the potential for introducing an

15

Nilsson’s proposed definition: “Reintegration is seen as a societal process aiming at the economic, political, and social assimilation of ex-combatants and their families into civil society.” (p. 27)

9

order of ambiguity that only makes the deciphering of a particular component of reintegration unnecessarily complex. In the case of this dimension, if the focus of programs is to be on the communally related aspects of reintegration vice individual demographics of productive capacity, then “societal” is the right word. In the case of Liberia, the designers of the DDRR program were most certainly focused on the communally related aspects of reintegration when conceiving their program. Community reconciliation and participation were understood as essential to post-conflict recovery. The Strategy and Implementation Framework for the Liberian DDRR program was drafted on 31 October 2003 —a little more than two months after the CPA was signed. The framework (2003) outlined five program results that would be tied to the overall objective to “contribute to the consolidation of peace, national security, reconciliation and development through the disarmament, demobilization, and reintegration of ex-combatants into civil society.” Result 4 focused on the “societal” dimension of reintegration: “Ex-combatants are received into and contribute to the development of their communities.” Based upon the indicative activities and the indicators envisioned by the original Liberian program designers, the following societal outcomes will be examined in this study. Excombatants are more likely to integrate into civilian communities if: Outcome 1: They gain the acceptance of the community. Outcome 2: They actively participate in community social and traditional events. Outcome 3: They extend their social networks beyond their ex-military circles.

ECONOMICS OF REINTEGRATION Besides a “social” aspect of reintegration, the UN doctrine distinguishes the “economic” component as equally essential to formulating an approach to a DDR intervention. The socio-economic landscape of a post-conflict setting is characterized by the devastation of a country’s infrastructure,

10

economy, and societal fabric. Essential to integrating former fighters into society is providing a mechanism for them to engage productively in a livelihood-producing activity. Initially, the labor market in a conflict-torn country will have a difficult time absorbing not only the flood of new workers looking for employment—ex-combatants, refugees, and Internally Displaced Persons (IDPs)—but also determining what markets will be successful. Key to the economic dimension and the outcomes we aim to evaluate is an understanding of how ex-combatants engage in livelihoodproducing activities and their ability to maintain a standard of living above the poverty line. Collier et al. (2003: 121) emphasize the macro importance of grappling with poverty and war-torn economies, citing these as the “core of the future global risk of civil war”. Others address poverty at the micro-level with regard to the linkages to reintegration outcomes. Nelson-Richards (2006) states that “poverty spawns many offspring,” including illiteracy, lack of income, insecurity, vulnerability, and armed conflict. Fearon and Laitin (2003: 80), looking to explain which countries have been at risk for civil war, point out that insurgency recruitment is easier when the “economic alternatives are worse”. While there are complexities and ambiguities that inherently confound the understanding of the societal dimension due to cultural biases and relationships as well as the “unobserved” community cleavages that smolder under the surface, the economic measures and outcomes are more readily applicable universally. In the Liberia program, the designers of the DDRR program were most certainly focused on an approach that accounted for the economic aspect of reintegration. Specifically, result 5 focused on the “economic” dimension of reintegration: “Ex-combatants are able to engage in economic activities.” Based upon the indicative activities and the indicators envisioned by the original Liberian program designers, the following economic outcomes will be examined in this study. Ex-combatants are more likely to integrate into civilian communities if: Outcome 4: They are engaged in livelihood-producing activities.

11

Outcome 5: They are earning a wage that keeps them above the poverty line. Outcome 6: They are exhibiting spending patterns indicative of excess earning capacity.

THE MISSING DIMENSION: POLITICAL The doctrine for international intervention and execution of a DDR operation has only recently begun to surface from the offices of the UN. While DDR has been conducted in many areas throughout the world for some time now, there had been only a disconnected collection of lessons learned and academic analysis until the publishing of the Integrated Disarmament, Demobilization, and Reintegration Standards by the UN in late 2006. While doctrine is not “law” or “regulation.” it does serve to codify the approach and methodology that is to be employed operationally. Organizational doctrine serves to guide the practitioners of operations, not to limit or to dictate. Ultimately, doctrine is meant to serve as datum for all specific projects and actions to relate to for overall program coherency. Unfortunately, while much has been published on the importance of a“political” dimension, the newly produced “standards” guide is devoid of any mention on how to incorporate mitigating actions during a DDR operation. Without an understanding of this dimension in doctrine, programs will be developed without a fundamental grounding of importance, and initial designs will not incorporate the indicators required to adequately monitor program progress and ensure a successful mission. What is meant by “political” as applied to this facet of reintegration? As mentioned above, Berdal (1996) and Kingma (2002) discuss the dimension in a qualitative manner along with numerous other scholars. In his theory on political violence, Gurr (1972) contends that deprivationinduced discontents derived from imbalances in the political environment are highly likely to precipitate rebellion. Key among Gurr’s variables is overt cultural and subcultural aggression, the extent to which political violence is successful, and the legitimacy of the political system (Gurr, 1972,

12

13). Humphreys and Weinstein (2007) place the political dimension center stage in their analysis of DDR in Sierra Leone. The authors examined the strength of ex-combatant factional ties in a postconflict environment and noted some quantifiable linkages to sources of political instability in a nation that had just emerged from civil war. Ironically, the UNHCR (United Nations High Commissioner for Refugees) identifies the need to understand the political dimension of reintegration in its repatriation programs of refugees to the same war-torn countries. The Handbook for Repatriation and Reintegration Activities (2004: One-4) states that reintegration is directly related to an environment that assures “political, economic, [legal] and social conditions needed to maintain life, livelihood and dignity.” This political environment includes the reestablishment of local judiciaries and enforcement means as well as improved security. In Liberia, the designers of the DDRR program put little emphasis on the political dimension of reintegration and did not provide any strategic results in this area aside from disarming the combatants. Critically, while political instability was recognized by the framework authors, there were only two references to the political/human security element of reintegration that were brought out for “action” in the document. The first was the observation that social tensions between former fighters and the civilians would be reduced through the promotion of the country’s traditional mechanisms for dispute resolution. The second was the identification of the “socio-political risk” that required monitoring—the “chances of organized political demonstrations due to general unhappiness, unmet expectations or perception of unsatisfactory peace-dividends”16. Based upon the indicative activities and the indicators envisioned by the original Liberian program designers and supplemented by the literature, following political outcomes will be examined in this study. Excombatants are more likely to integrate into civilian communities if: Outcome 7: They have confidence in the communal mechanisms for dispute resolution. 16

Draft Interim Secretariat (2003: 78)

13

Outcome 8: They sever all social and economic ties with former faction members. Outcome 9: They are empowered by nonviolent political expression in changing government.

IV Factors That Influence

Given a categorization of the facets of reintegration, this chapter now explores a few of the factors that are believed to influence resultant outcomes. The literature is peppered with anecdotal evidence that postulates which factors are influential in the reintegration and reconciliation process. In an effort to identify a few correlates of reintegration success, four categories of independent variables are offered for discussion. The examination includes programmatic as well as nonprogrammatic related variables. The programmatic variables identify the former combatants of Liberia and segregate them into two distinct groupings—those who chose to register with the NCDDRR and those who did not and opted to reintegrate on their own. These categories will provide the basis for evaluating the impact of the international intervention program in Liberia. A handful of nonprogrammatic factors will also be evaluated against the outcomes posited to be oriented toward (positive) reintegration. Besides the programmatic variables, this chapter will examine three unique categories of nonprogrammatic factors—individual demographics, personal wartime experiences, and community factors.

PROGRAMMATIC VARIABLES Assessing the impact of DDR programs has historically presented numerous difficulties. While the macro-level outcome (observed at the country level) of a DDR program is generally how well the

14

country keeps the peace after conflict,17 this does little to inform the international community on the efficacy of the program itself. A plethora of factors swirl and torture a country emerging from a significant period of conflict. Societal cleavages that were catalyzed need mending, infrastructure that was decimated requires repair, and economic capacity destroyed requires reestablishment. While all of these factors play a part in the rebuilding of communities in post-conflict, there is a need for program planners of reintegration to understand how to discern the impact of their interventions amid the chaos of other societal “rebirthing” activities. Very few studies have attempted to quantify the impact of a DDR program through an empirical examination of the micro-level determinants considered to be reflective of the more readily apparent macro-level country observables. Significantly, Humphreys and Weinstein (2007) failed to observe any evidence of micro-level impact of the national DDR program in Sierra Leone. Since the DDRR program in Liberia commenced in December 2003 and by March 2006 only 21% of the registered ex-combatants (over 18) had completed their assigned course of reintegration training, a need to segregate the population in accordance with their stage in the intervention process was deemed essential. The disaggregation of the intervention independent variable will allow not only a testing of those receiving reintegration assistance but also an investigation of any nuances associated with the track of progression in the DDRR program from one stage to the next. Accordingly, the grouping that includes the former combatants who registered with the national DDRR program is further disaggregated into three categories—those who completed the course of reintegration training, those who were currently enrolled in reintegration training, and those who disarmed and demobilized but had yet to enter a reintegration training program. With an international focus on funding, training, and equipping the former combatants

17

UNDP (2006: 23) idenifies rigorous monitoring and evaluation of DDR programs as impediments to understanding the efficacy of any intervention effort.

15

with the purpose of reintegrating into Liberian society, the following will be true of the DDRR program in Liberia:

H1 = Former fighters who register and complete a course of reintegration training will reintegrate better than those former fighters who decided not to register and reintegrate on their own

INDIVIDUAL DEMOGRAPHICS The literature contains a great deal of research that encompasses individual demographics and their correlations with reintegration success in a post-conflict environment. Notable among these are those that focus on age and gender. These two demographics have been important as they are used to disaggregate samples of the beneficiary population thought to be vulnerable in the post-conflict setting and in need of targeted assistance. For age, the study excluded respondents under the age of 18, so we evaluate influence as a function of age given a limited sample of the population. That said, however, approximately 18% of the respondents surveyed would have been under the age of 18 at the time of the signing of the CPA. The war ran for the better part of 14 years in Liberia, which meant that for the majority of the population—which is quite young, with a mean age of 26-1/2 — conflict is all that most ex-combatants know. UN Security Council Resolution 1325 on women, peace, and security, as well as a subsequent Secretary-General 2002 Report, calls for stronger measures to be implemented toward adopting a gender perspective both in the negotiating process and during the rebuilding of the peace and reconciliation. Despite this resolution, DDR programs all too often ignore women or fail to account for gender differences in mobilization and demobilization. Carlson and Mazurana (2004) echo this in their study of female combatants in Sierra Leone by warning of the marginalization of

16

women and faulty design of DDR interventions. McKay and Mazurana (2004) continue to illustrate the severe hardships that former women fighters face during reintegration as a result of the sexism, misogyny, and sexual exploitation that occurred for many during the conflict. One demographic that has had little attention paid to it in the literature with regard to reintegration success is marital status. In Africa, Khapoya (1998) indicates, the significance of marriage is not necessarily based on love or common interests, as in the West, but is considered fundamental to the propagation of life and is part of a greater responsibility to the community. Those who are married have a significant stake in ensuring that their village can live in peace. Education and property ownership are other factors that have been linked to successful reintegration in the post-conflict environment. Nelson-Richards (2006) believes that education plays a major underlying role in diminishing the onset or reoccurrence of armed conflict. The researcher hypothesizes that education enables employment, which in turn makes joining an armed insurrection less appealing. While there are counterpoints that argue that when educational opportunities in developing countries exceed the capacity of a labor force to absorb them at an appropriate employment level, the result may be disillusionment, ‘alienation and political disaffection,’18 Gurr challenges this theory through presentation of findings from empirical tests. Gurr (1972: 99) states that education ‘seems to serve less to raise expectations to an unsatisfiably high level than to provide ambitious men with the sense that they have better means for attaining their expectations.’ With regard to property and ownership, Doyle and Sambanis (2006) postulate that interventions must support and enable one’s right to a claim on the soil. Ruminating over an adage often quoted by economists—“no one washes a rented car”—the authors stress the importance of the will to protect ownership. Those who have are less likely to abandon under pressure.

18

Gurr (1970:96) quotes from a study conducted by Archibald Callaway, “Unemployment Among School Leavers,” Journal of Modern African Studies 1 (September 1963), 351-371.

17

With regard to the nonprogrammatic characteristics to be explored in this chapter, the following individual demographics hypotheses are offered with regard to their capacity as indicators of reintegration success:

H2 = An increase in reintegration success will be observed as a function of age H3 = Males will reintegrate better than females H4 = Married ex-combatants will be better off than those without a partner in the postconflict environment H5 = Former fighters with higher levels of education will reintegrate better than those with limited educational exposure H6 = Former fighters who are established home owners will show increased levels of reintegration success

WAR-TIME EXPERIENCES The effects on the reintegration process of individual experiences during conflict are also an area worth examining. Conflict by definition is socially disruptive and those who are party to it will be inevitably altered by the experience. The nature of participation is felt not only by those who volunteer but also by those who were coerced into the conflagration. In order to account for this extreme in participation, two variables will be examined for their potential as an influential factor— volunteer officers and abducted. During the conflict in Liberia, officers were looked to lead, organize, train, recruit, and serve in staff roles that supported field commanders. Officers were typically given greater responsibilities than enlisted soldiers and were looked to for direction and orders. Officers are typically the representatives who negotiate with village chiefs and elders when on operations and by whom

18

information is to be garnered, resources acquired, and protection afforded. The decisions made by officers in the field can have effects that linger long after the conflict and may still exist in areas where the officer’s unit maneuvered. Those in command are usually well known and chosen based upon their ability to complete a mission—no matter what the cost, and sometimes no matter how it is done. In order to obtain the category of individuals that we seek to examine in detail, we must exclude officers who reported that they were abducted and include only those who voluntarily joined. Alternatively, coercion plays an equally disruptive role in the lives of ordinary Liberians attempting to stay out of the fighting. Abduction as a means of recruitment was used heavily throughout the 14 years of conflict and more recently by all three major warring factions— Government forces, LURD (Liberians United for Reconciliation and Democracy), and MODEL (Movement for Democracy in Liberia). Pugel (2007b) reveals that nearly 40% of all combatants in the Liberian civil war were coerced in some way to join a fighting faction with approximately a quarter of all forces reporting that they were abducted. The traumatic nature of abduction in conflict has also been shown to negatively affect the reintegration of this subpopulation of ex-combatants. In many other studies of DDR reintegration, officers and those who were forced into military service were cited as subsamples of the population of former fighters who had encountered difficulties in assimilation.19 Given the nature and involvement of individuals who served in leadership roles and those experiencing highly coercive actions during their “recruitment” process, the following hypotheses are offered with regard to their capacity as an indicator of reintegration success:

19

Readers can refer to Humphreys & Weinstein (2007), UNDP (2006), and Bouta (2005).

19

H7 = Those who were abducted during the conflict will be less likely to reintegrate than other ex-combatants H8 = Former volunteer officers will be less likely to reintegrate than other ex-combatants

COMMUNITY FACTORS The final set of factors that will be considered concerns the influence of the environment in which the former fighter is embedded. While individual demographics and wartime experiences will inform us on the influential capacity of subsample characteristics that affect reintegration, an examination of a few environmental factors will allow a connection with more macro-level measures. Locality of resettlement must be considered when examining reintegration. Kingma (2002) discusses the challenges that face ex-combatants who converge upon the nation’s capital after a civil war believing that jobs and opportunity are available only in the urban centers. Most are disappointed with the failure to rebuild the infrastructure at a noticeable pace and even more frustrated with the lack of economic capacity that comes with saturation of self-employed entrepreneurship. Resettling in one’s prewar community is the recognized preference of the UNHCR when repatriating the refugees of the conflict.20 The literature indicates a bit more ambiguity for ex-combatants, though, as a population in general. Under some circumstances, the former fighters might be revered as great defenders of the community that they hailed from. Other may have chosen to “take advantage of the war,” as Kalyvas (2003: 476) points out, in order to settle old disputes and are thus seen as monsters in the eyes of the communities that they once called home. Personal security is also required if reintegration success is to take hold. An environment of fear and isolation has been shown to be incompatible with viable intervention approaches. Doyle

20

UNHCR (UN High Commissioner for Refugees). 2004. Repatriation and Reintegration of Liberian Refugees: January-February 2004.

20

and Sambanis (2006) emphatically state that the first step in building peace is securing the environment for an unobstructed reconstruction of society. This implication is echoed by Kingma (2002), who states that if the resettlement environment does not afford protection for reintegrating ex-combatants and their families, the former fighters may be forced into picking up the gun again out of necessity. Given the diversity and nature of environmental factors, the following hypotheses are offered with regard to their capacity as an indicator of reintegration success:

H9 = Former fighters who have chosen to resettle their original pre-conflict communities will reintegrate better than those who chose to settle in a community away from home H10 = Former fighters who have chosen to settle in the Liberian capital region will encounter a more difficult time reintegrating than those who chose to settle in the hinterland H11 = It will be more difficult for former fighters to reintegrate in communities with a higher rate of crime

V Empirical Approach to a Qualitative Field of Study

The data collection effort took place during February and March 2006, slightly more than a year after the formal end of the Disarmament and Demobilization period in Liberia. The study was sponsored by Mr. Charles Achodo, the senior program advisor for DDRR in the UNDP Liberia. In partnership with a Monrovian-based nongovernmental organization (NGO), the African Network for the Prevention & Protection Against Child Abuse & Neglect (ANPPCAN) Liberia, a large-scale randomly sampled survey of ex-combatants was undertaken. Elenora Koeb led the project staff, Raj Karikolrag facilitated the data entry, and Shecku Silla, head of the Monitoring and Evaluation Unit, was instrumental in assisting with the cleaning of the data set.

21

The survey was heavily influenced by a similar ex-combatant study carried out by Humphreys and Weinstein in Sierra Leone in 2003.21 Implementation of the design was conducted in a blended approach that incorporated a closed-ended quantitative interview intake questionnaire with a qualitative participatory method, a focus group discussion. A criterion for entry into the study was that a prospective former combatant must have self-reported that he or she served with a warring faction at some point during the Liberian civil war after 1989. Participation in the formal DDRR program was not a requirement. The study was inclusive of both genders and all former warring faction participants. The study collected interviews only from respondents who were over the age of 18. The study was designed to collect 600 interviews from 30 geographic clusters of 20 respondents each. All 66 administrative districts of Liberia were considered in a random selection process that was weighted against the resettlement preferences indicated by former combatants on their demobilization in-take forms. In the end, 590 usable surveys were collected (255 from Greater Monrovia and 355 from outside the capital region). The field collection spanned 10 of Liberia’s 15 counties and included 23 unique district locations. Variation across Greater Monrovia was gained through access to 12 metropolitan zones. Community locations were also randomly selected within each administrative district through incorporation of unique identifiers (P-codes)22 in the random number generation technique.

21

A large-N study of ex-combatants in Sierra Leone was conducted by Humphreys and Weinstein during June-August 2003. The UNDP-funded study in Liberia largely replicated the Sierra Leone study. The Sierra Leone report, “What the Fighters Say: A Survey of Ex-combatants in Sierra Leone,” and the survey instrument used can be found at URL:. 22 The P-codes were provided by the Humanitarian Information Center from a mapping project that was completed in 2005. The data lists all villages with a unique “P-code” (Place Code) that can subsequently be linked to many other sources of information. The database currently associates the following information as well: county, district , number of houses, percentage of displacement, percentage that have returned, presence and functionality of schools and health facilities, accessibility by car in wet and dry seasons. The database can be accessed at URL:< http://www.humanitarianinfo.org/liberia/ mapcentre/ reference/village_ws_maps/index.asp>.

22

At the community level, field collection teams introduced themselves to the appropriate chiefs or elders and solicited their assistance in obtaining a pool of prospective candidates for the study. Consent was gained from each candidate and a final random selection of the respondents available in the targeted location was conducted. A report on the status of the population of ex-combatants registered with the NCDDRR was requested to be specially run for the purpose of this study. The status report was generated by the Monitoring and Evaluation Unit of the Joint Implementation Unit (JIU). The purpose of the report was to establish the basis for the sampling frame. Table 1 presents the breakdown of the respondents in accordance with their participation status in the formal DDRR intervention program. (Table 1) (Table 2)

VI Measurable Outcomes

Measuring reintegration outcomes in an empirical approach is essential to any evaluation of a DDR program. In order to characterize the dimensions of reintegration that have been identified, nine measurable outcomes have been operationalized with three defined for each component.

SOCIETAL The three outcomes modeled within the societal dimension are accepted, participates, and social

network. Accepted (Outcome 1) was derived from two questions on acceptance—one that acknowledged the level of acceptance gained through self-reporting and the second that recorded a perceived level of acceptance from the community that the former combatant resided in. The

accepted variable took a measure of 1 only if the combatant reported no problems with acceptance AND perceived that the community reciprocated. A 0 was recorded for any respondent self-

23

reporting problems OR perceiving that the community viewed the population of ex-combatants with fear or distrust.

Participates (Outcome 2) was modeled on a single, four-item Likert scale question that queried how often the ex-combatant participated in community events led by citizens who did not take up arms during the conflict. A measure of 1 was achieved for those respondents who participated “often” or “sometimes,” while a 0 was recorded for those who participated “seldom” or “never.” The outcome of social network (Outcome 3) was derived from a single question that asked respondents to indicate whom they spent their free time with. A measure of 1 was recorded for responses that included “family” or “friends” met either before or after the war. A 0 was recorded for those former fighters who responded with “friends” from the war or that they spent their time alone.

ECONOMIC The variable engaged (Outcome 4) represents an outcome that is broader (and more relevant) than a simply stated dydad of employment and unemployment. The measure utilizes a construct for determining employment status though a series of five questions.23 Ultimately, engaged is derived from the resultant categories of the construct and indicates the level of engagement in a productive livelihood activity. A measure of 1 is designated if the respondent was economically employed—

23

The construct for employment was formulated in coordination with the Bureau of Statistics, Government of Liberia. A series of six question based upon a 1999 Government of Liberia demographic survey were designed to assess the respondents into a particular employment status. Ten possible categories of employment are used: employer, employee, self-employed, unemployed, home duties, student, retired, not looking for work, don’t know/other, and sick/disabled. A response of “unemployed” distinguishes the respondents from those who are actively engaged in a livelihood-producing activity. With unemployment estimated at 80 percent in the formal sector, a construct that provides more variation on employment status is critical to analysis.

24

“employer,” “employee,” or “self-employed” or engaged in “home duties” or as a “student.” A measure of 0 was given for those who were unemployed, “not looking,” or sick/disabled. A question soliciting a respondent’s daily income in Liberian Dollars (LD) represented the basis for the measure to gauge resistance to poverty as an outcome. For above poverty (Outcome 5), a measure of 1 was given for each respondent indicating a daily wage above 55 LD (approximately $1US at the time of the survey). In an effort to identify another relevant interval variable that could be considered a quantifiable indicator of economic freedom, the variable spending (Outcome 6) represented a dollar amount that the former fighters spent on clothes for the family during the Christmas/Ramadan/New Year’s holiday period that Liberians observe as a traditional time of spending. The variable was kept in its true interval form for subsequent analysis.

POLITICAL The outcome measure impartial (Outcome 7) was generated from a single four-item Likert scale question that asked the respondents to state their opinion as to the degree they believed that the judges of community conflict resolution were fair and impartial. A measure of 1 was given only for those respondents who indicated that their mechanism for resolving conflict was “almost always fair.” The measure received a 0 for responses of “mostly fair,” ”mostly unfair,” “almost always unfair.” In order to measure an outcome of factional disassociation, broken ties (Outcome 8) was derived from a single, four-item Likert scale question that asked the ex-combatants how often they were in contact with their former faction leaders for financial or employment assistance. A measure of 1 was given only to those respondents who reported that they “never” had any contact. A score of 0 was given when the ex-combatant reported any kind of contact with former faction leaders.

25

The final outcome measured is vote counts (Outcome 9). This measure is derived from a single question that asks the former combatants the degree to which they believed they could effect change through a democratic process by expressing opinions during elections. A measure of 1 was given for responses stating that expressing opinions during elections would “help a lot.” A measure of 0 was recorded for those responding that their opinions would help “a little” OR “not at all.”

VII Deciphering the Dimensions of Reintegration

Examination of the outcomes represented under unique dimensions of reintegration has yielded a number of interesting findings. This section of the chapter examines dimensional correlation and provides a probit analysis of each dependent reintegration outcome variable (with the exception of one interval variable where a standard ordinary least squares regression is utilized). Marginal effects for each measure are reported in the following tables. Each regression is based upon fixed effects for factions and clusters errors around the administrative districts where the interviews were taken. The exploratory alignment of the tested outcomes under qualitative dimensional headings proved to be informative. As can be observed in the correlation matrix at Table 1, the majority of the dependent variables appear to maintain positive correlations within their designated groupings. While the societal and economic variables appear to exhibit strong inter-dimensional relationships, the Political dimension appears to be more enigmatic. In fact, one of the political dimension variables— vote counts—appears to be negatively correlated with the perception of impartiality in regard to the communal mechanism for conflict resolution as well as four other positive outcomes of reintegration success. While the general consistency of the correlations maintained by the societal and economic dimensions illustrates the beginnings of a quantitative understanding of these two facets of reintegration, the radical sign changes in inter-dimensional correlations between the political and

26

other two dimensions indicates that the complexity of the third exploratory dimension has not yet been captured.

(Table 3)

SOCIETAL Of the three dimensions examined, the societal outcomes provided the least empirical insight. While the acceptance variable appeared to inform quite well when controlling for the programmatic factors, participation, and social network did little to surface any societal impact evidence. Additionally, none of the factors that were controlled for in the categories of individual demographics, wartime experiences, and community factors indicated any empirical relevance across the three dimensional outcomes selected. Of the ten factors, only three yielded a statistically significant finding within the societal dimension—education level (higher), marital status (married), and gender (female). All three factors appear consistent with reintegration success within the societal dimension.

(Table 4)

ECONOMIC The findings related to economic reintegration success were fairly informative. All three outcomes that were evaluated surfaced a number of statistically significant results. Across the dimension, six of the ten controlling factors indicated an influence on reintegration success with four of the statistically significant variables impacting at least two of the three postulated dimensional outcomes. Importantly, the magnitudes of each factor move in the same direction, which helps to eliminate

27

ambiguity from interpreting the findings. Each regression model investigated in the economic dimension also exhibits a fair bit of strength as demonstrated by the fact that each model is estimated to explain from 10 to 15% of the variance in each outcome. Interestingly, married respondents proved highly motivated in the area of economic reintegration success and consistently influenced the three economic outcomes. Positive influences were also observed by those with higher levels of education and those who were older. Alternatively, those who were abducted during the war as well as those who chose to resettle in the greater Monrovia area were shown as disadvantaged when it came to economic reintegration success. Most surprising was the finding that those who chose to resettle in their prewar communities were much more disadvantaged than those former fighters who opted to settle away from home. The programmatic control variables yielded limited empirical evidence with regard to economic reintegration, only superficially highlighting potentially influential indicators.

(Table 5)

POLITICAL The findings related to understanding the political dimension of reintegration success are informing with regard to the relational associations between factors and outcomes but tend to be problematic when one attempts to characterize them as a unique dimension. While eight of the ten nonprogrammatic factors provide insight into political reintegration success and four of those eight factors cut across two outcomes, many of the factors introduce ambiguity in the interpretation by reacting in opposite directions. Of the four factors that influence two of the political outcomes, only one—those former fighters who were married—exhibited a consistent magnitude of direction which was shown as negatively associated with this chapter’s definition of political reintegration

28

success. The three other factors—education, crime, and gender (female)—informed on the same two outcomes, impartiality and vote counts, and yielded contradictory effects against our measures of reintegration success within the dimension. Volunteer officers were shown to be disadvantaged with regard to political reintegration success, and those abducted during the war demonstrated a propensity for positive reintegration success within this dimension. Again, only limited programmatic impact was assessed from the models that were formulated for the political dimension. (Table 6)

While this chapter has endeavored to decipher the dimensions of reintegration under an empirical light, the investigation has revealed the complex nature of this undertaking and has hopefully surfaced many more avenues for further research.

VIII Influential Factor Determination: Nonprogrammatic Factors

A number of indicators were introduced into the regression models in an effort to identify any patterns that might be useful to understanding the three dimensions of reintegration that this chapter examines. Ten factors were considered in the analysis drawn from three tiered topical areas —individual demographics, wartime experiences, and community factors. Much of the literature has referred to the factors discussed as influential. The findings below quantify the influences empirically as a function of their impact on reintegration success. While all categories exhibited some level of influence upon the reintegration outcomes that were tested, the set of individual demographic factors proved the most consistent at informing on success.

29

INDIVIDUAL DEMOGRAPHICS While all five of the individual demographics indicated a level of influence on the reintegration process, only two of the factors could inform across all three dimensions—education level and marital status. Age and home ownership provided only limited value in understanding reintegration success, with both demonstrating an effect on only one outcome of the nine that were investigated. Surprisingly, the examination of gender challenges did not yield conclusive evidence of the literature’s generalization that females encounter extreme difficulties. In fact, it appears that females are actually outpacing males in some areas of reintegration. Females are more likely to gain acceptance within their communities and are also more committed to the democratic process as their preferred method for political change. Politically, however, females remain skeptical of the mechanisms for conflict resolution within their communities. Explanation for the increased female societal resilience in post-conflict Liberia may be a result of the national DDRR program concentrating on gender issues, which was by far one of the most inclusive intervention programs to date.24 As for understanding the positive findings on democratic empowerment, the inauguration of the first female African head of state only one month before the survey was conducted undoubtedly played a role in influencing the responses of the former combatant women. The most promising factors for understanding reintegration success were related to the former fighter’s education level and marital status. These two individual demographic measures were consistently correlated with the success outcomes presented under the societal and economic headings—both of which had stated programmatic objectives as defined by the designers of the Liberian DDRR in the original framework. While the societal dimension of reintegration remains largely enigmatic (empirically) even after this study, education and marital status appear as two of the few factors that may provide clues to understanding how to tailor a reintegration program.

30

Economically, the two factors do much to inform on reintegration success. In both instances, former combatants with higher levels of education and those who were married were consistently found to be living above the poverty line and able to indulge in more generous gift spending over the winter holiday period.25 Where the average holiday spending was found to be approximately $45 US (2475 LD), married and educated respondents reported levels that were on average almost $9 US (495 LD) more than those for other former combatants. Last, married respondents (besides those who had completed a course of reintegration training) were found to be more engaged in a livelihood-developing activity than the rest of the population of ex-combatants. With both education and marital status, however, interpretation of reintegration success as a comprehensive, multi-faceted measure tends to get muddled with the introduction of the political component. Interestingly, just as the female respondents lacked confidence in the impartiality of the local judges who preside over conflict resolution, those with higher levels of education and married former fighters displayed similar frustrations. While this finding may be indicative of a system really in its infancy, attempting to return to normalcy after years of brutal war and infrastructure devastation and not conclusive evidence for informing on reintegration success, it is disturbing to note that while the married ex-combatants appear to assimilate quite well both societally and economically, they are also disposed to maintaining ties to their former faction commanders for employment and economic assistance.

24

Carames et al. (2006: 17) note in their study of DDR programs worldwide, that women are often not included. Significantly, however, they point to Liberia as an exception to this finding. 25 In Liberia, there are traditionally two times per year that the people indulge in spending on gifts for friends and family. One is during the Independence Day celebrations (26 July), and the other is over the Christmas/New Year’s period (December).

31

WARTIME EXPERIENCES Uniquely, only two characteristics were disaggregated for the analysis of wartime experiences, those who had self-reported being abducted from the war and those former combatants who left the war as officers and did not report that they were coerced into joining. The two factors did not demonstrate any influence (positive or negative) on the societal dimension outcomes and very little on the economic outcomes, where only those who were abducted demonstrated a disadvantaged engagement in livelihood-producing activities. The two factors were, however, able to inform on the reintegration outcomes characterized within the political dimension. Volunteer officers, as expected, were shown dependent on their former faction for employment assistance and economic aid. This finding underlines the need to further understand the bridge between warring faction mobilization and the demobilization that occurs when conflict has ceased and hard-core fighters are now expected to become contributing members of civil society. Understanding how to break the ties of dependency is critical in the design of a post-conflict intervention program. Alternatively, those former combatants who were coerced into the fighting surprisingly demonstrated positive motivations for the newly installed democratic process in Liberia and are actually more apt to believe that their vote can force political change than those former combatants who were not abducted.

COMMUNITY FACTORS The final set of factors that will be considered in the analysis revolves around the influence of the environment within which the respondents chose to reintegrate. Like the factors that looked to inform on the influence of individual wartime experiences, the selected community factors provided little empirical evidence within the context of societal reintegration outcomes. The puzzling lack of evidence within the societal dimension again illustrates the complexity of the dimension and that there must be many other confounding factors that the literature has yet to upturn. The factors did,

32

however, help promote understanding of reintegration success in both the political and economic dimensions. The factor on crime was the least informing. A higher rate of perceived crime and lawlessness yielded contradictory effects within the political dimension only, with indications that higher crime is associated with disillusionment in political expression, but this was an unexpected finding that demonstrates a belief in the impartiality of the system for conflict resolution. While settlement in the Liberian capital was indeed shown as generally disadvantageous for former fighters as predicted—less engagement in a livelihood-producing capacity and a propensity for living below the poverty line—a surprisingly positive indicator did in fact surface. Former fighters within Monrovia were more apt to have confidence in the impartiality of the conflict resolution system than those who chose to settle in the hinterland and locations away from the capital. This finding may be explained as a result of the concentration of forces in Monrovia and the buildup of the National Police Force, which took great measures to secure the nation’s capital and allow “judging” to occur without fear of retribution and without bias that usually accompanies areas which do not allow their citizens a confident sense of personal security. Disturbingly, when examining the factor of hometown resettlement, the magnitudes of all six coefficients within the models formulated for the political and economic dimensions were negative, implying that ex-combatants might be disadvantaged if they chose to resettle their prewar communities of origin. Two of the variables were statistically significant—livelihood engagement and poverty. Those respondents who chose to return home were found not as likely to be engaged in a livelihood-producing function as those respondents who chose to settle elsewhere. The capacity of those same former fighters to earn a wage that kept them above the poverty line was also shown as significantly less than the capacity of those who chose to establish a new life elsewhere.

33

IX Reintegration and DDR: What Is the Impact?

Disaggregation by dimensional component of reintegration provides a limited understanding of the impact of the national DDRR program in Liberia. The study segregated two groupings of excombatants—those who registered with the DDRR program for benefits and those who opted to reintegrate on their own. The grouping that includes the former combatants who registered with the national DDRR program was further disaggregated into three categories—those who completed the course of reintegration training, those who were currently enrolled in reintegration training, and those who disarmed and demobilized but had yet to enter a reintegration training program. The regressions presented in tables 2 through 4 illustrate the effects of the categories of DDRR participants on the reintegration outcome variables for all three dimensions under examination. In each instance of the regression modeling, the three categories of DDRR-registered combatants are tested against the category of combatants that did not formally disarm and register for program benefits. As demonstrated in the analysis, results of assessing the impact of the DDRR program are mixed. The strongest finding, within the societal dimension, indicates that when examined as two distinct groups—participants and nonparticipants—those ex-combatants who formally disarmed and registered with the NCDDRR gained acceptance within their chosen resettlement areas at a greater rate than did the nonparticipants. A second major finding focuses on testing the chapter’s first hypothesis that former fighters who register and complete a course of reintegration training will reintegrate better than those former fighters who decided not to register and reintegrate on their own. When only these two categories are compared against the reintegration success outcomes, the ex-combatants who completed a course of reintegration training tended to fare better in two of the three dimensions. Only two of the outcomes were evaluated as statistically significant, and both only weakly so. While acceptance has already been identified as a discriminator of the DDRR program, the other significant outcome was

34

the economic variable of engaged , which shows that the DDRR program completers were more apt than nonparticipants to be pursuing some form of livelihood-producing status. A third major finding concerned with addressing the question of program impact notes the variance in outcomes exhibited by the category of registered former fighters who had yet to enroll in a course of reintegration training. While this subsample of ex-combatants was found to have the highest acceptance rates of any other category and was also shown as a category of former fighters that appears to have been successful in breaking their ties with former faction leaders, the grouping was also found, unfortunately, to be encumbered politically and economically. As table 6 shows, this category of the population appears to be disillusioned with the democratic process and doesn’t believe that their vote will make a difference. As a result of an ordinary least squares regression, table 5 shows that the former fighters who had yet to enroll in a course of reintegration training have a reduced excess spending capacity when unique cash outlay periods (over winter holidays) are examined. While disaggregation by category of ex-combatant has highlighted the differences between groups, more research is needed before the DDR program’s impact can be properly assessed. Of the nine outcomes examined, only one proved to be consistent as an indicator of reintegration success when evaluating the two groups, participants and nonparticipants. The question of evaluation timing must also be considered when attempting to interpret the findings presented here. Berdal (1996) cites the ILO in stating that reintegration assessments shouldn’t be undertaken until a significant amount of time has passed. Last, the finding that demonstrates the nonlinearity in the benefits returns process as characterized by the irregularity of reintegration outcomes success by the category of former fighters who formally disarmed but had yet to enroll in a course of reintegration training provides evidence that examining only those who participated in a program and those who did not could introduce confounding effects into the analysis.

35

X Conclusion

The purpose of this chapter is to acknowledge the challenges in evaluating a DDR program and provide an argument for more empirically based research and evaluations for the formulation of programs and policies. While substantive empirical research is beginning to emerge on DDR evaluation, a number of areas are in desperate need of attention if academics and practitioners are to stay ahead of the programming curve. With an increasing emphasis on quantifying the efficacy of international DDR intervention programs, a rigorous approach to evaluation has been shown to illuminate issues in all three reintegration dimensions examined in this chapter as well as to spotlight some influencing factors and their ability to impact reintegration success. The approach taken is not without shortcomings. Foremost, the examination was not an inclusive look at the whole of ex-combatant reintegration, as pointed out by the need to examine the psychological dimension in more detail. While not discussed in this chapter, preliminary research into self-esteem constructs by the author appears to hold promise for informing on reintegration success, but results at this point are certainly not ready for scrutiny. Additionally, the study was quasi-experimental and recorded measurements from respondents only after the DDR “treatment” had been applied. The inferences in this chapter are based on findings believed to be accurate, but they would most certainly carry more weight under controlled experimental conditions. Finally, truth telling during self-reporting is always a concern when interviewing former combatants. While Liberia was a quite conciliatory environment during the time of field collection and the overwhelming majority of the respondents were more than willing to participate, the Liberian Truth and Reconciliation Commission was inaugurated (21 February 2006) at almost the same time that n field collections began. It is hard to be certain what effect this had on the respondents’ answers, but is the effect assumed to have been quite minimal.

36

In examining the dimensional reintegration outcomes, the economic dimension proved most empirically informing. All three outcomes that were evaluated surfaced a number of statistically significant results. Across the dimension, six of the ten controlling factors indicated an influence on reintegration success, with four of the statistically significant variables impacting at least two of the three postulated dimensional outcomes. Importantly, all statistically significant indicators reacted in the same direction across the three outcomes designated as economic, providing a fair bit of confidence in the relationship between the factors under study and the dimension as a conceptual grouping. While many interesting findings surfaced in the societal and political dimensional areas, much more research is required in the areas of reintegration. Of the three groupings of influencing factors—individual demographics, wartime experiences, and community factors—the individual demographics were found to be the most telling on interpreting reintegration success. While all five factors that were examined demonstrated influence on reintegration success, the variables marital status (married) and education level (higher) proved the most influential. Significantly, it was found that those respondents with a higher education level are reintegrating better socially and economically than those with lesser levels of education. Most important, married respondents appear to represent a group that can be examined the most accurately within the framework of this study. Again, this group demonstrates a significant advantage in achieving reintegration success in both the economic and social dimensions over excombatants without partners. Ironically, though, married former fighters struggle in the political dimension; they were shown to be skeptical of the impartiality of the judges of conflict resolution and were also found to maintain ties with former faction leaders at a higher rate than their comrades without partners. An important finding identified in this chapter is that Liberian ex-combatants who registered with the NCDDRR were found to find acceptance within their chosen resettlement communities at

37

a higher rate than the former fighters who chose not to formally disarm. When program completers were compared with those fighters who did not register for benefits, the completers, in addition to being accepted, were more likely to be engaged in a livelihood-producing activity—employment, school, home duties. Both findings were only weakly significant, but they certainly indicate an impact of the DDRR program on the lives of those who have participated. Unfortunately, however, there is a major risk of leaving behind a very vulnerable grouping of NCDDRR-registered excombatants—those who have disarmed and demobilized but have yet to receive training. This category of former fighters is the least educated, most agriculturally oriented, and poorest of the four classes under investigation. While these fighters were found to have gained acceptance within their chosen resettlement communities and also demonstrated a commitment to breaking ties with their former factional leaders, they were also found to be more economically disadvantaged than those former fighters who chose to reintegrate on their own. The group was also shown to have less faith in the democratic system and elections for political change than the category of ex-combatants that opted to self-reintegrate. This chapter has endeavored to investigate the enigma of ex-combatant reintegration with the purpose of codifying measurable outcomes, determining the effect that influencing factors have on the dimensions of reintegration, and quantifying the interim impact of the DDR program in Liberia. The ambiguity inherent in the evaluation of a DDR program stems mostly from an inability to quantify suitable outcome measures—especially those concerned with reintegration. Quantifiable outcomes are essential for program evaluation and policy framing. Fundamentally, the chapter demonstrates the importance of understanding the dimensions of ex-combatant reintegration in a way that can be translated into quantifiable outcome measures for practitioners and decision makers.

38

References Amara, Joe, Freida M’Cormack, James Pugel and Allan Quee. 2006. A Situation Analysis of Excombatant Reintegration in Liberia. Monrovia: UNMIL RRR. December. Anderson, Mary B. and Lara Olson. 2001. Confronting War: Critical Lessons for Peace Practitioners. Cambridge, MA: CDA, Inc. Annan, Koffi. 2005. ‘Secretary-General's statement to the Sudan Donors' Conference.’ Presented in Oslo, Norway. 11 April. Azam, Jean-Paul. 2004. “On Thugs and Heroes: Why Warlords Victimize Their Own Civilians.” Working Paper, Institute for Economic Analysis. Berdal, Mats R. 1996. Disarmament and Demobilisation after Civil Wars. Adelphi Paper 303. Bøås, Morten and Anne Hatløy. 2006. After the Storm: Economic activities among children and youth in return areas in post war Liberia – the case of Voinjama. Fafo report 523. Bouta, Tsjeard. 2005. Gender in Disarmament, Demobilization, and Reintegration. Clingendael Conflict Research Unit. March. Butler, Christopher, Scott Gates and Michele Leiby. 2005. Social Networks and Rebellion. Manuscript. Caramés, Albert, Vicenç Fisas, and Daniel Luz. 2006. Analysis of Disarmament, Demobilization and Reintegration (DDR) programs existing in the world during 2005. Barcelona: Escola de cultura de Pau. February. Carlson, Khristopher and Dyan Mazurana. 2004. From Combat to Community: Women and Girls of Sierra Leone. Women Waging Peace Policy Commission. Hunt Fund Alternatives. http://www.huntalternatives.org/pages/32_case_studies.cfm. Collier, Paul et. al. 2003 Breaking the Conflict Trap. Washington, D.C.: World Bank and Oxford University Press. CPA (Comprehensive Peace Agreement) between the Government of Liberia and the Liberians United for Reconciliation and Democracy (LURD) and the Movement for Democracy in Liberia (MODEL) and political parties, Accra, Ghana, 18 August 2003. Doyle, Michael and Nicholas Sambanis. 2006. Making War & Building Peace. Princeton, NJ: Princeton University Press. Draft Interim Secretariat. 2003. Liberian Disarmament, Demobilization, Rehabilitation, and Reintegration Programme: Strategy and Implementation Framework. Monrovia. October. Fearon, James, Macartan Humphreys and Jeremy Weinstein. 2007. Community-Driven Reconstruction in Lofa County Baseline Survey Preliminary Report.

39

Fearon, James and David Laitin. 2003. “Ethnicity, Insurgency, and Civil War.” American Political Science Review 97: 75-90. Gates, Scott. 2002. “Recruitment and Allegiance: The Microfoundations of Rebellion.” Journal of Conflict Resolution 46: 111–130. Gurr, Ted R. 1972. Why Men Rebel. Princeton, New Jersey: Princeton University Press. GoL (Government of Liberia). 2003. Community Based Support in Reintegration and Recovery Process. Project outline coordinated with UNDP Liberia. 15 December. Gleichmann, Colin et al. 2004. Disarmament Demobilisation and Reintegration: A Practical Field and Classroom Guide. GTZ, NODEFIC, PPC, and SNDC. HRW (Human Rights Watch). 2004. How to Fight, How to Kill: Child Soldiers in Liberia. Vol 16, No 2 (A). February. Humphreys, Macartan and Paul Richards. 2005. Prospects and Opportunities for Achieving the MDGs in Post-conflict Countries: A Case Study of Sierra Leone and Liberia. CGSD Working Paper No. 27. October. Humphreys, Macartan and Jeremy Weinstein. 2007. ‘Demobilization and Reintegration.’ Journal of Conflict Resolution (forthcoming). August. Humphreys, Macartan and Jeremy Weinstein 2006. ‘Who Rebels? The Determinants of Participation in Civil War.’ Unpublished Paper, Columbia and Stanford Universities. July. Humphreys, Macartan and Jeremy Weinstein 2005. “Handling and Manhandling Civilians in Civil War: Determinants of the Strategies of Warring Factions.” Unpublished Paper, Columbia and Stanford Universities. ——. 2004. What the Fighters Say: A Survey of Ex-Combatants in Sierra Leone. CGSD Working Paper No. 20. August. ICG (International Crisis Group). 2003. Liberia: Security Challenges. Africa Report No 71. November. IPA (International Peace Academy). 2006. Counting What Counts: Ten steps toward increasing the relevance of empirical research in the UN system. Meeting Note. New York. February. ——. 2003. Transforming War Economies: Challenges for Peacemaking and Peacebuilding. Report of the 725th Wilton Park Conference ‘Economic Agendas in Civil Wars’ Heiko Nitzschke rapporteur. October. ISS (Institute for Security Studies). 2004. ‘Trends in DDR in Peacekeeping in Africa’ in Identifying Lessons from DDR experiences in Africa. Malan, Mark and Sarah Meek ed. Monograph no. 106. October. Jennings, Kathleen. 2007. “The Struggle to Satisfy: DDR through the Eyes of Ex-combatants in Liberia.’ International Peacekeeping. Summer.

40

Kalyvas, Stathis. 2003. “The Ontology of ‘Political Violence’: Action and Identity in Civil Wars’. Perspectives on Politics 1:3, 475-494. Kingma, Kees. 2002. ‘Demobilization, Reintegration and Peacebuilding in Africa.’ International Peacekeeping Vol 9, No 2. Summer. ——. 1997. ‘Demobilization of combatants after civil wars in Africa and their reintegration into civilian life.’ Policy Sciences 30. Khapoya, Vincent B. 1998. The African Experience: An Introduction, 2nd ed. Upper Saddle River, New Jersey: Prentice Hall. Nelson-Richards, Melsome. 2006. Youth in Post-Conflict Africa: Community-Based Rehabilitation, Rebuilding and Reconstruction of Society. Paper presented at the Expert Group Meeting on Youth in Africa, Windhoek, Namibia. 16 November. Nilsson, Anders. 2005. Reintegrating Ex-combatants in Post-Conflict Societies. Paper commissioned by Sida to the Department of Peace and Conflict Research, Uppsala University. May. Landry, Guillaume. 2006. Child soldiers and Disarmament, Demobilization, Rehabilitation and Reintegration in West Africa. Coalition to Stop the Use of Child Soldiers. November. LNP (Liberia National Police). 2006. Annual Crime Report and Statistics for 2005. October. Levitt, Jeremy I. 2005. The Evolution of Deadly Conflict in Liberia. Durham: Carolina Academic Press. Lichem, Walther. 2006. DDR Processes and Societal Development. Address to the Conference on PostConflict Peacebuilding in Africa: Assessing DDR Process. Kofi Annan International Peacekeeping Training Center, Accra, Ghana. August. McKay, Susan and Dyan Mazurana. 2004. Where are the Girls? Canada: Rights & Democracy. McPherson, J.H.T. 2004. History of Liberia. USA: Kessinger Publishing. Nichols, Ryan. 2005. ‘Disarming Liberia: Progress and Pitfalls.’ In Nicolas Florquin and Eric G. Berman, eds. Armed and Aimless: Armed Groups, Guns, and Human Security in the ECOWAS Region. Geneva: Small Arms Survey. Paes, Wolf-Christian. 2005. ‘The Challenges of Disarmament, Demobilization and Reintegration in Liberia.’ International Peacekeeping Vol 12, No 2. Summer. Pugel, James. 2007a. Disaggregating the Causal Factors Unique to Child Soldiering. Manuscript. _____. 2007b. What the Fighters Say: A Survey of Ex-Combatants in Liberia. Monrovia: UNDP Liberia. Republic of Liberia. 2004. Community Resettlement and Reintegration Strategy. June.

41

Search for Common Ground. 2005. Key Findings from 2005 Risk and Conflict Assessment: Liberia before Elections. < http://www.sfcg.org/programmes/liberia/programmes_liberia.html> Swarbrick, Peter. 2007. Avoiding Disarmament Failure: The Critical Link in DDR—An Operational Manual for Donors, Managers, and Practitioners. Working Paper 5. Geneva: Small Arms Survey. Toki, Hinako. 2004. Peace-building and the Process of Disarmament, Demobilization, and Reintegration: the Experiences of Mozambique and Sierra Leone. Japan International Cooperation Agency. March. UN (United Nations). 2006. Disarmament, demobilization, and reintegration. Report of the Secretary General A/60/705. March. UN (United Nations). 2006. Operational Guide to the Integrated Disarmament, Demobilization and Reintegrations Standards. UN (United Nations). 2005. ‘Seminar on the Challenge of Reintegration of Ex-Combatants in DDR Programs in West Africa.’ Final Communique presented in Dakar, Senegal. 8 April. UNDP (UN Development Program). 2006. Practice Note: Disarmament, Demobilization and Reintegration of Ex-combatants. UNDP (UN Development Program). 2002. Handbook on Monitoring and Evaluating for Results. June. New York, NY: UNDP. UNDP (UN Development Program). 2000. Sharing New Ground in Post-Conflict Situations: The role of UNDP in Support of Reintegration Programmes, January. . UNHCR (UN High Commissioner for Refugees). 2004. Repatriation and Reintegration of Liberian Refugees: January-February 2004.

42

Table 1: Sampling Framework and Distribution of Ex-Combatants over 18 Years Old

NCDDRR Totals Frequency Percent Disarmed/Demobilized Participants Reintegration Participants Completed DDRR Reintegration Training DDRR Nonparticipants Total

Survey Sample Frequency Percent

46,152

43.7%

265

44.9%

26,536

25.1%

170

28.8%

19,244

18.2%

86

14.6%

13,73726

13.0%

69

11.7%

105,66927

100%

590

100%

Table 2: Sampling Framework and Distribution of Female Ex-Combatants over 18 Years Old

NCDDRR Totals Frequency Percent Disarmed/Demobilized Participants Reintegration Participants Completed DDRR Reintegration Training DDRR Nonparticipants Total

Survey Sample Frequency Percent

13,081

12.4%

63

10.7%

5,465

5.2%

34

5.8%

3,896

3.7%

11

1.9%

996

0.9%

11

1.9%

23,438

22.2%

119

20.2%

Table 3: Correlation Matrix of Dimensional Outcomes Accepted? Accepted?

Participates?

Social Network?

Impartial?

Broken Ties?

Vote Counts?

Above Poverty?

Economically Engaged?

Spending?

0.12

0.21 0.05

-0.05

0.16

-0.04

0.04

0.06

-0.10

0.05

-0.01

-0.03

0.02

0.14

0.02

Participates? Social Network?

0.12 0.21

0.05

Impartial?

-0.05

0.05

Broken Ties?

0.16

-0.01

0.10

0.08

Vote Counts? Above Poverty?

-0.04

-0.03

-0.04

-0.03

-0.06 -0.06

0.10

-0.04

0.03

0.04

-0.01

0.08

-0.03

-0.01

-0.15

0.08

0.03

0.01

0.10

0.05

0.00

0.03

-0.04

0.14

0.25

0.03

0.04

0.02

0.03

-0.01

0.01

0.00

Economically Engaged?

0.06

0.14

0.04

-0.15

0.10

0.03

0.14

Spending?

-0.10

0.02

-0.01

0.08

0.05

-0.04

0.25

26

27

Not a number reported by the NCDDRR; represents an estimated rate of 13%. Exact NCDDRR database total was 81,368.

43

0.14 0.14

Table 4: Societal Outcomes

(1) Accepted?

(2) (3) Participates? SocialNetwork?

Class: Completed

0.155 0.069 -0.063 [1.82]* [1.54] [1.31] Class: Enrolled 0.223 0.087 0.027 [3.41]*** [1.92]* [0.95] Class: D1/D2 only 0.273 -0.010 0.037 [4.04]*** [0.26] [1.08] ------------------------------------------------------------------------------------------------Education Level 0.033 -0.010 0.017 [0.68] [0.27] [1.73]* Home Owner 0.058 0.049 -0.010 [1.06] [1.48] [0.38] Married 0.039 0.083 0.005 [0.74] [2.17]** [0.16] Female 0.099 0.039 -0.013 [1.98]** [0.80] [0.54] Age 0.002 0.063 -0.001 [0.66] [1.17] [0.30] Volunteer Officer -0.017 0.084 -0.031 [0.25] [1.44] [0.98] Abducted -0.048 -0.005 -0.008 [0.83] [0.11] [0.24] Resettled Home 0.069 0.031 0.023 [1.28] [0.76] [0.84] Greater Monrovia -0.043 -0.099 -0.071 [0.31] [1.18] [1.40] High Crime 0.009 0.048 -0.026 [0.17] [0.79] [0.69] Observations Pseudo R2

571 0.05

562 0.06

571 0.08

NB: “Class” is compared against the class: nonparticipants (DDRR program). * Significant at 10%; ** significant at 5%; *** significant at 1%. Robust z statistics are in brackets. Marginal coefficient estimates from probit analyses reported. All models allow errors to be clustered geographically at the district (subset of county) level.

44

Table 5: Economic Outcomes

(4) Above Poverty?

(5) Engaged?

(6) Spending?

Class: Completed

0.086 0.141 84.857 [1.41] [1.67]* [0.23] Class: Enrolled 0.013 0.112 -129.668 [0.32] [1.23] [0.44] Class: D1/D2 only -0.039 -0.023 -467.005 [1.04] [0.26] [1.78]* ----------------------------------------------------------------------------------------------------------------------Education Level 0.054 0.040 405.454 [2.21]** [1.40] [3.29]*** Home Owner 0.022 -0.101 -240.210 [0.66] [1.09] [1.25] Married 0.093 0.206 578.443 [2.43]** [6.17]*** [2.80]*** Female 0.018 -0.029 -162.856 [0.57] [0.72] [1.07] Age 0.007 -0.002 46.008 [1.25] [0.91] [2.75]** Volunteer Officer -0.041 -0.002 142.298 [0.86] [0.04] [0.76] Abducted -0.016 -0.122 -121.287 [0.51] [1.68]* [0.87] Resettled Home -0.038 -0.142 -98.659 [1.67]* [2.09]** [0.69] Greater Monrovia -0.093 -0.251 68.992 [1.89]* [2.10]** [0.22] High Crime 0.055 -0.056 257.205 [1.18] [1.08] [1.49] Observations Pseudo R2 R-squared

571 0.10

487 0.11

563 0.15

NB: “Class” is compared against the class: nonparticipants (DDRR program). * Significant at 10%; ** significant at 5%; *** significant at 1%. Robust z statistics are in brackets. Marginal coefficient estimates from probit analyses reported. All models allow errors to be clustered geographically at the district (subset of county) level.

45

Table 6: Political Outcomes

(7) Impartial?

(8) BrokenTies?

(9) VoteCounts?

Class: Completed

-0.039 0.009 -0.043 [0.50] [0.18] [1.06] Class: Enrolled 0.013 0.045 -0.052 [0.17] [1.02] [1.39] Class: D1/D2 only -0.042 0.116 -0.047 [0.49] [2.38]** [1.83]* ---------------------------------------------------------------------------------------------------Education Level -0.093 -0.009 0.028 [3.02]*** [0.38] [2.59]*** Home Owner 0.178 0.010 0.002 [2.11]** [0.36] [0.08] Married -0.115 -0.065 0.002 [3.09]*** [1.79]* [0.07] Female -0.094 0.002 0.042 [1.82]* [0.04] [2.17]** Age -0.002 0.004 -0.000 [0.45] [1.53] [0.11] Volunteer Officer 0.025 -0.072 -0.017 [0.32] [1.68]* [0.53] Abducted 0.068 0.030 0.032 [1.13] [0.85] [1.97]** Resettled Home -0.038 -0.048 -0.002 [0.69] [1.13] [0.11] Greater Monrovia 0.223 -0.076 -0.019 [2.79]*** [1.41] [0.89] High Crime 0.243 0.057 -0.067 [2.39]** [1.41] [2.38]** Observations Pseudo R2

566 0.14

571 0.05

569 0.07

NB: “Class” is compared against the class: nonparticipants (DDRR program). * Significant at 10%; ** significant at 5%; *** significant at 1%. Robust z statistics are in brackets. Marginal coefficient estimates from probit analyses reported. All models allow errors to be clustered geographically at the district (subset of county) level.

46