Policy Research Working Paper 9916 Local Peace Agreements and the Return of IDPs with Perceived ISIL Affiliation in Iraq Jacqueline Parry Olga Aymerich Social Sustainability and Inclusion Global Practice January 2022 Policy Research Working Paper 9916 Abstract Between 2015 and 2020, communities across Iraq signed same time, indirect representation of IDPs (through tribal local peace agreements (LPAs) aimed at setting the foun- leadership) enabled negotiations to occur that would oth- dations for peaceful relations between internally displaced erwise be impossible; that the inclusion of women, youth persons (IDPs) accused of siding with ISIL and their ‘home’ and civil society contributed to the sustainability of IDP communities. This paper maps the characteristics of eight return but did not appear to influence safe return; that LPAs such LPAs and compares each with the scholarly expecta- were an effective tool for addressing community concerns tions regarding what makes a ‘good’ peace agreement. It around the return of IDPs but typically prioritized com- also considers how the characteristics of each LPA affected munity concerns over IDP rights, potentially creating new conditions for safe and sustainable IDP return. Overall, grievances; and finally, that mechanisms to implement and the study found that a failure to include IDPs directly in monitor each LPA were important for building community LPA negotiations did not appear to compromise safe return acceptance but empowered new actors and were vulnerable but did potentially undermine sustainability—but at the to co-optation. This paper is a product of the Social Sustainability and Inclusion Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at parry.jacqueline@gmail.com. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Local Peace Agreements and the Return of IDPs with Perceived ISIL Affiliation in Iraq Jacqueline Parry and Olga Aymerich Key words: Peacebuilding, Displacement, Iraq, Stabilization JEL classifications: D74: Conflict; Conflict Resolution D81: Criteria for Decision-Making under Risk and Uncertainty This paper was commissioned by the World Bank Social Sustainability and Inclusion Global Practice as part of the activity “Preventing Social Conflict and Promoting Social Cohesion in Forced Displacement Contexts.” The activity is task managed by Audrey Sacks and Susan Wong with assistance from Stephen Winkler. This work is part of the program “Building the Evidence on Protracted Forced Displacement: A Multi-Stakeholder Partnership''. The program is funded by UK aid from the United Kingdom's Foreign, Commonwealth and Development Office (FCDO), it is managed by the World Bank Group (WBG) and was established in partnership with the United Nations High Commissioner for Refugees (UNHCR). The scope of the program is to expand the global knowledge on forced displacement by funding quality research and disseminating results for the use of practitioners and policy makers. This work does not necessarily reflect the views of FCDO, the WBG or UNHCR. Section 1: Introduction Modern intrastate conflicts have become known for their intensity and complexity, generating not only mass displacement, loss of life, and destroyed infrastructure, but also severely damaging social relationships and undermining trust in state institutions. In this context, resolving displacement is a complex task: in order for people to return ‘home,’ they need to feel that the issues driving the conflict have been sufficiently addressed and will not recur in future, to the detriment of both stability and prospects for sustainable development.1 In order to generate this sense of confidence and understand the concerns of displaced persons, it has become increasingly common to include refugees and internally displaced persons (IDPs) in peace processes that occur at or near the end of war. 2 Peace processes entail a mix of ‘politics, diplomacy, changing relationships, negotiation, mediation, and dialogue’ and aim to solve conflict by peaceful means, 3 typically concluding with a formal peace agreement that articulates and addresses the substantive issues underlying the conflict. 4 Iraq offers a valuable case study in this respect. Following a period of conflict associated with the Islamic State (ISIL) from 2014 until 2017, numerous local peace agreements (LPAs) were negotiated between IDPs accused of siding with ISIL and their ‘home’ communities, members of which including those who rejected, fought against, or were victimized by ISIL. These LPAs aimed to lay the foundation for reconciliation between IDPs and their home communities and negotiate the return of IDPs in a safe and sustainable manner. This study examines eight LPAs established in Iraq between 2015 and 2020. Each LPA applies to a single town, sub-district or governorate and was drafted and signed by prominent tribal and community leaders. Alternatively known as ‘return agreements’ or ‘tribal peace accords,’ the LPAs grapple with one of the most pressing issues for durable peace in Iraq’s post-war context: how to resolve the protracted displacement of families with a perceived affiliation to ISIL. Each agreement attempted to establish or recognize principles and a mechanism to govern the return of IDPs, as well as a commitment to the rule of law, with the expectation that this would mitigate revenge attacks and collective punishment while simultaneously addressing security concerns. While these LPAs have received substantial attention by the international community, 5 there is little analysis regarding the 1 O’Neill, W. (2009) Internal displacement and peacebuilding: Institutional responses. Refugee Survey Quarterly, 28(1), pp.151-180. 2 For an overview, see: Brookings Institution, Addressing Internal Displacement in Peace Processes, Peace Agreements and Peace-Building, September 2007. Available at: https://www.brookings.edu/wp- content/uploads/2016/06/2007_peaceprocesses.pdf 3 Saunders, H. (2001) Turbulent Peace, Washington DC: US Institute for Peace. 4 Although peace agreements still only make up just 18.4 per cent of total terminations. Kreutz, J. (2010). How and when armed conflicts end: Introducing the UCDP Conflict Termination dataset. Journal of Peace Research, 47(2), pp.243-250. 5 The LPAs in this study have been designed, implemented and/or supported in other ways by the United States Institute of Peace (USIP), Sanad for Peacebuilding (Sanad), Peace Paradigms Organization (PPO), the German Corporation for International Cooperation (GIZ), The Government of Canada-PSOPs Program, and the United Nations Development Program (UNDP). 2 affect they have had on community relationship and safe and sustainable IDP return. This study aims to address this gap in knowledge by examining the terms, negotiation process, and implementation of eight LPAs, comparing these findings with the scholarly expectations regarding what makes a ‘good’ peace agreement, 6 and then considering how the LPA characteristics affected conditions for safe and sustainable IDP return. This paper complements the work of other scholars in this series who describe the ways that community cohesion and social conflict are affected by displacement, and how fractured social cohesion associated with IDP return can be addressed. The paper authored by Tellez and Balcells highlights the willingness of conflict-displaced IDPs to contribute to peace-building efforts, even when they expect insecurity to continue or when the context is unstable, 7 a perspective that reinforces the finding in our paper that IDPs should be involved in local peace agreements even while community relations are tense, utilizing representatives if direct negotiations are not feasible. Analysis by Ruiz and Vargas-Silva regarding returned refugees to Burundi offers further insight, pointing to the negative impact IDP return can have on certain aspects of social cohesion: if there is a negative attitude in the community towards the return of conflict-affected populations, it may affect the willingness of community members to help each other and may interfere with the perception of the community as a peaceful space. 8 Our study emphasizes the role that LPAs can play in situations where there is a negative perception towards returning IDPs, since an LPA can offer a way to navigate the return of IDPs rejected by their communities and build community trust in the return process. Finally, the paper by Bove, Salvatore and Elia indicates how the international community can enable conditions for voluntary, safe and sustainable return of displaced people: their study of UN peacekeeping operations in South Sudan found that displaced people are more likely to return home if peacekeepers are deployed in their province of destination, and that the presence of peacekeepers contributes to better relations between local populations and returnees. 9 Our paper aligns with these findings by demonstrating that local peace agreements in Iraq, most of which were brokered or supported by international actors in cooperation with local peacebuilding actors, also contributed to safe and sustainable IDP return and better community relations post-return. 6 Scholars have started to identify the type of peace agreements most likely to contribute to durable peace, and the type of engagement with IDPs during peace processes most likely to support durable solutions. This issue is discussed throughout the present study, but for an overview, see: Andersen-Rodgers, D.R. (2015). Back home again: Assessing the impact of provisions for internally displaced persons in comprehensive peace accords. Refugee Survey Quarterly, 34(3), pp.24-45. 7 Juan F. Tellez and Laia Balcells (2021), “The social, economic, and political effects of displacement. Evidence from rural Colombia”, Unpublished Working Paper. Commissioned as part of the “Preventing Social Conflict and Promoting Social Cohesion in Forced Displacement Contexts” Series. Washington, DC: World Bank Group. 8 Isabel Ruiz and Carlos Vargas-Silva (2021), “Refugee Return and Social Cohesion”, Unpublished Working Paper. Commissioned as part of the “Preventing Social Conflict and Promoting Social Cohesion in Forced Displacement Contexts” Series. Washington, DC: World Bank Group. 9 Vincenzo Bove, Jessica De Salvatore and Leandor Elia (2021), “How does UN peacekeeping shape attitudes toward displacement and return for IDPs and host communities?”, Unpublished Working Paper. Commissioned as part of the “Preventing Social Conflict and Promoting Social Cohesion in Forced Displacement Contexts” Series. Washington, DC: World Bank Group. 3 Section 2: Context From 2014 to 2017, Iraq experienced a displacement crisis of unprecedented national severity, with more than 5 million people (roughly 15% of the Iraqi population) displaced by ISIL’s violent insurgency and the responding military campaign. 10 Ethno-religious minorities were targeted with brutal violence that caused the displacement of whole communities, particularly in Sinjar and the Nineveh Plains (where Yezidi and Christian communities historically live). While the atrocities committed against ethno-sectarian minorities were among the most grievous of this period, the majority of people forcibly displaced during 2014-2017 were Sunni Arabs originating from the Sunni-majority western and northern governorates of Iraq (including Anbar, Kirkuk, Ninewa, and Salah al-Din), 11 since these areas were targeted and subsequently occupied by ISIL and later became the focus of the military campaign. Across these areas, some Sunni Arab residents fled immediately upon ISIL’s arrival, while others remained throughout the occupation, for diverse reasons: economic resources, social networks, family structures, information, threat perceptions, ideology, and – particularly in the early days – a perception that the quality of governance under ISIL would improve relative to the Iraqi state. 12 Residents who remained in ISIL-held territory typically displaced with the start of the military campaign to retake each location. 13 This meant that whole communities ultimately became displaced across Anbar, Kirkuk, Ninewa, and Salah al-Din, but the timing and circumstances of their displacement differed, and this difference created deep intra-community divisions and shattered community cohesion. 14 This had serious implications for the return of IDPs, which we discuss further below. Since the end of the military campaign in December 2017, more than 3.5 million IDPs have returned to their areas of origin. 15 Yet still, in 2021, some 1.2 million IDPs remain displaced. 16 Many IDPs face material barriers to return such as damage to housing, lack of livelihoods, and lack of basic services, 17 but what appears even more difficult to overcome are the social and security-related 10 IOM Iraq (2018) Iraq Displacement Crisis: 2014-2017. 11 For a full overview of the ISIL displacement crisis see: IOM Iraq (2018). ISIL Displacement Crisis: 2014-2017. 12 Revkin, M. (2021). Competitive Governance and Displacement Decisions Under Rebel Rule: Evidence from the Islamic State in Iraq. Journal of Conflict Resolution, 65(1), 46-80. 13 For detailed information on displacement patterns during the war with ISIL in Iraq, see: IOM Iraq (2018). Iraq Displacement Crisis 2014-2017. Available at: http://iraqdtm.iom.int/files/DurableSolutions/20203224827300_IOM- Iraq_Displacement_Crisis_2014-2017.pdf 14 For detailed information on the need to reconciliation and a correlation of reconciliation indicators to residential destruction, illegal occupation of private residences, concerns over revenge, the multiplicity of security actors see: IOM Iraq, Social Inquiry, RWG (2019) The growing role of reconciliation in return movements. Return Index Thematic Briefing Series. Available at: http://iraqdtm.iom.int/images/ReturnIndex/20203145217451_iom_returnindex_thematic_seriesNov2019.pdf 15 For detailed information on IDP return in Iraq see: IOM Iraq (2021). An Overview of Return Movements in Iraq. 16 A breakdown of the ethno-sectarian identity of IDPs is not available, since this data is not typically collected by humanitarian actors due to the potential stigmatization or discrimination it may cause. Overall, as of 30 April, 2021, according to IOM’s Displacement Tracking Mechanism (DTM), there are 1,198,940 IDPs displaced across Iraq, and 4,867,050 returnees. IOM Iraq (2021) Displacement Tracking Matrix (DTM). Available at: http://iraqdtm.iom.int/ 17 For example, an IOM study reported that IDPs perceived the greatest barriers to return as damage to housing (affecting 71% of IDPs), lack of livelihoods (53%), and lack of basic services (62%.) IOM Iraq (2021). Protracted Displacement in Iraq: Revisiting Categories of Return Barriers. 4 barriers in areas of return. 18 One group to emerge as particularly vulnerable to protracted displacement is IDPs with perceived affiliation to ISIL, an issue that predominantly affects the Sunni Arab (and Sunni Turkmen) population. The precise number of such families is difficult to quantify, in part because humanitarian agencies do not record this perceived status, 19 and in part because the concept itself is defined differently in each community and has changed over time. The perception that a family is affiliated to ISIL has little to do with legal definitions but draws upon community understandings of who is considered guilty and deserving of punishment. ISIL’s offensive and occupation divided communities and tribes into two broad groups, linked to the circumstances under which families displaced and their experience during the war: on one ‘side’ is those who fled ISIL’s arrival and occupation – and sometimes participated in the military campaign to retake territory from ISIL via area-based tribal mobilization forces 20 – and on the other, those who remained under ISIL occupation and displaced when the military campaign arrived. For those who remained and lived under ISIL occupation, communities differ in the perception of what ‘level’ or ‘type’ of collaboration warrants an accusation or perception of affiliation to ISIL. Many communities recognize the complex reality of living under ISIL occupation: although some residents in ISIL- occupied territory undoubtably collaborated with the group, this collaboration was extremely diverse and ranged from simply paying taxes (per regulations enforced by ISIL) to committing horrific acts of violence. 21 Immediately after the war ended, it was more common for communities to judge all those who lived under ISIL rule as affiliated in an indiscriminate and sweeping fashion; however, as time has progressed, the perception of affiliation has typically narrowed to those who held certain roles under ISIL (such as senior leadership) or committed serious crimes on behalf of the group. 22 Typically, the accusation of affiliation goes beyond the individual to the kin level and attaches to 18 IOM Iraq, Social Inquiry, Returns Working Group (2021). Home Again? Categorizing Obstacles to Returnee Reintegration in Iraq. 19 To avoid stigmatization and discrimination, the humanitarian community does not collect data on the number of people considered to be affiliated with extremist groups such as ISIL. In mid-2018, some estimates placed the number at several hundred thousand. See: Interaction, ‘Moving Forward Together, Leaving No One Behind: From Stigmatization to Social Cohesion in Post-Conflict Iraq,’ August 2018, available at: https://protection.interaction.org/wp- content/uploads/2018/10/2018-10-31-InterAction-Iraq-report_Moving-Forward-Together.pdf. 20 Iraq’s Popular Mobilization Forces (PMF) is a conglomeration of 60 to 70 armed groups, formally recognized by the Government of Iraq, which rose to prominence in 2014 for halting the expansion of ISIL and retaking Iraqi territory from its control. It includes Sunni units known as the Tribal Mobilization Units, composed of fighters from the Sunni tribes from the areas occupied by ISIL. van Veen, E., Grinstead, N., & El Kamouni-Janssen, F. (2017). A house divided; EASO (2021). Popular Mobilisation Units and Tribal Mobilisation Militias. Country Guidance Iraq 2021. Available at https://easo.europa.eu/country-guidance-iraq-2021/12-popular-mobilisation-units-and-tribal-mobilisation-militias.) 21 Kao, K., & Revkin, M. R. (2018). To Punish or to Pardon?: Reintegrating Rebel Collaborators After Conflict in Iraq. Reintegrating Rebel Collaborators After Conflict in Iraq (November 1, 2018). The Program on Governance and Local Development Working Paper Series at the University of Gothenburg, (17). 22 For example, in the governorate of Anbar, some communities consider as ISIL affiliates those who cohabited with the group, whereas other communities have a more nuanced understanding of affiliation and consider as affiliated families those with one first-degree relative allegedly was a full-fledged member of the group, meaning the person was an ISIL combatant or was appointed by ISIL to a core civil servant post such as mukhtar. Certain actions are also considered as redlines for returning to the community, which tends to apply to those families whose first-degree relatives killed other community members or burn their properties as part of ISIL. For more information see: Aymerich, O. (2020). Managing Return in Anbar: Community Responses to the Return of IDPs with Perceived Affiliation. IOM Iraq. 5 first- or second- degree relatives, although in some communities, it even reaches up to fourth-degree relatives. 23 The perception that a family supported ISIL has serious repercussions for return prospects. Families with perceived affiliation to ISIL often find their return blocked by security actors,24 experience community rejection and stigmatization, 25 and are at high risk of revenge attacks and violence. 26 Female returnees to ISIS-affected communities are also at a higher risk of SGBV, particularly female-headed houses and those in rural areas 27. At the same time, it is common for people living in the area of return to fear that the return of families they believe supported or continue to support ISIL will destabilize their communities and create new risks for security and social relations. In this context, the buy-in of community members has been vital to avoid revenge attacks or the start of new conflict when families with perceived ISIL affiliation return, 28 and communities across Iraq have adapted a range of strategies to navigate this issue, often relying on tribal structures and customs. The reliance on tribal structures to address community conflict is not surprising: an estimated 75% of Iraq’s population identify themselves with a tribe, 29 and tribal lineage is a key determinate of identity, status, and social relations in many communities (particularly in Sunni Arab areas). 30 Tribes often play a very practical role in daily life, with many Iraqis relying on their tribe for physical protection, economic resources, and dispute mediation on a day-to-day basis, 31 not only in rural areas (where adherence to tribal customs, traditions and hierarchies is typically strongest) but also in urban areas, particularly where rural-to-urban migration is common. 32 Tribal justice and formal justice often work in a complementary manner in Iraq, and coordination occurs regularly between tribal leaders and security forces to address criminal behaviour; 33 in some cases, law enforcement officials and courts choose to refer cases to the tribal system for settlement.34 Conversely, in controversial or 23 Genat, M. 2020. Tribal Justice Mechanisms and Durable Solutions for Families with a Perceived Affiliation to ISIS. IOM Iraq. 24 Blocked returns are commonly enforced by security actors on the ground, usually on the grounds of the IDPs families not possessing the required documentation (mostly a security clearance) to leave the area if displacement or re-enter the area of origin (IOM Iraq, 2021). 25 Genat, M. (2020) and IOM Iraq (2018). 26 See for example Human Rights Watch (2017). Iraq: Looting, Destruction by Forces Fighting ISIS. Available at: https://www.refworld.org/docid/58a5b4344.html 27 Oxfam (2021). Community perceptions of Sexual and Gender Based Violence. 28 Aymerich, O., 2020. Community Resilience and the Return of Iraqi IDPs with Perceived Affiliation to the Islamic State. Refugee Survey Quarterly, 39(4), pp.552-563. 29 Hussein H. (2007) Iraq: Tribal structure, social, and political activities. Otterman, S. (2005). Iraq: The Role of Tribes. Available at: https://www.cfr.org/backgrounder/iraq-role-tribes 30 Gospodinov, I. D. (2015). The Sunni Tribes of Iraq. Leiden: Leiden University. 31 IOM Iraq (2020). Perceptions of Security and Police in Iraq. Baseline survey findings. 32Ibid. 33 Bobseine, H. (2019). Tribal Justice in a Fragile Iraq. The Century Foundation, 1-24. 34 Carroll, K. B. (2011). Tribal Law and reconciliation in the new Iraq. The Middle East Journal, 65(1), 11-29. 6 sensitive cases such as the trial and punishment of captured ISIL members, tribal leaders usually prefer to rely on state security or judicial actors. 35 In the aftermath of the ISIL conflict, tribal leaders have played an active role in facilitating the return of members of their tribe. Wary of the risk of violence and disruption of social peace that threatens their communities and fearing that grievances created by the conflict causes community members to take matters into their own hands through revenge and retaliation, 36 tribal leaders have adapted customary processes to facilitate the return of perceived ISIL-affiliated families and address the security concerns and grievances of community members and victims of ISIL violence. 37 Prominent examples include the practice of sponsorship, by which a tribal sheikh sponsors the IDP head of household to obtain a security clearance (issued by the state) to authorize return, and disavowal practices, by which relatives of ISIL members pledge in front of a judge or tribal committee to cut ties to their accused relative. 38 Although these practices continue to be used by Iraqi communities to resolve protracted displacement and prevent conflict, they are not typically supported by the international community due to the perception that they undermine the rule of law and do not comply with human rights standards. However, another strategy that has received consistent support by international actors is the establishment of LPAs. Beginning in 2015, tribal leaders and other key stakeholders (including government officials and security actors) in communities across Ninewa, Anbar, and Salah al-Din negotiated and signed LPAs aimed at facilitating the return of families with perceived ISIL affiliation while simultaneously addressing community security concerns. The remainder of the report discusses these LPAs in detail. Section 3: Theoretical Motivation This study situates the analysis of Iraq’s LPAs within the scholarship concerning IDP involvement in peace agreements. Peace processes have increasingly engaged IDPs in recognition that resolving displacement is critical for sustainable peace after conflict. 39 The watershed Dayton Accords first articulated the right of return for all refugees and displaced persons in 1995, 40 and since then, the right to return and redress for IDPs has been echoed in almost every subsequent peace agreement. 41 35 Revkin, M. (2018). After the Islamic State: Balancing Accountability and Reconciliation in Iraq. United Nations University Centre for Policy Research. 36 IOM Iraq, Social Inquiry, RWG (2019) The growing role of reconciliation in return movements. Return Index Thematic Briefing Series. Available at: http://iraqdtm.iom.int/images/ReturnIndex/20203145217451_iom_returnindex_thematic_seriesNov2019.pdf 37 Aymerich, O. (2020). Managing Return in Anbar: Community Responses to the Return of IDPs with Perceived Affiliation. IOM Iraq. 38 For more information in these practices see Genat, M. (2020). 39 Andersen-Rodgers, D.R. (2015). 40 General Framework Agreement For Peace In Bosnia and Herzegovina, 35 I.L.M. 117 (14 December 1995) annex 7, art 1. 41 This includes peace agreements in Burundi, Bosnia and Herzegovina, Cambodia, Ethiopia, Eritrea, Guatemala, Kosovo, Mozambique, and Sierra Leone. Phuong, C. (2005). Forcible Displacement in Peace Agreements, International Council on Human Rights Policy; Bradley, M. (2006) Reparations, Reconciliation and Forced Migration (Research Guide, Forced 7 International legal standards emphasize that return is a fundamental right, 42 as well as one of the pathways to achieving a durable solution for IDPs, 43 and establishes that states may not restrict the right to return except on the relatively narrow grounds of ‘national security, public order, public health or morals, or the rights and freedoms of others,’ and only when such restrictions are appropriate, proportionate, and established by law. 44 Generating the confidence of IDPs requires more than a rights-based peace agreement; IDPs need to feel convinced that that their security, rights, and freedoms will be guaranteed following return, otherwise the mere existence of a peace agreement is unlikely to lead to a durable solution. 45 As such, the process by which an agreement is negotiated, designed and implemented will determine whether – and which – IDPs are willing or able to return home, and under what conditions. 46 That is to say, the ‘who, what, and how’ of each agreement is vital to its success and must be tailored to each context. 47 Scholars stress the importance of engaging displaced persons in all levels of a multi-track peace process: in ‘track-one,’ which involves official government diplomacy; 48 ‘track-two,’ which involves conflict resolution specialists, private citizens, non-governmental organizations or businesses; and ‘track-three,’ which refers to unofficial interventions at the grassroots. 49 The scholarship also recognizes there may be times when direct participation of IDPs in one or all of these tracks may be undesirable or impossible, due to insurmountable barriers to participation, if participation entails risks for the displaced, or when the absence of democratic structures to elect IDP leaders means that IDP interests are compromised. In situations where it is not possible to include Migration Online; Leckie, S. d. (2003), Returning Home: Housing and Property Restitution Rights of Refugees and Displaced Persons, Transnational Publishers, pp. 15-16. 42 The right to return is founded enshrined by the right to return is Article 12 of the International Covenant on Civil and Political Rights, which protects liberty of movement and the freedom to choose one’s own residence. See: W Kälin, Guiding Principles on Internal Displacement: Annotations (rev edn, American Society of International Law and Brookings–Bern Project on Internal Displacement 2008) at 126, 129–30. 43 Principle 28 of the Guiding Principles on Internal Displacement states that IDPs ‘have the right to choose freely between return, local integration or resettlement.’ See: Francis Deng, Compilation and Analysis of Legal Norms, Part II: Legal Aspects Relating to the Protection Against Arbitrary Displacement, U.N. Doc. E/CN.4/1998/53/Add.2 (11 February 1998). 44 International Covenant on Civil and Political Rights (adopted 16 December 1966, entered into force 23 March 1976) 999 UNTS 171 at Article 12(3). See also: UN Human Rights Committee, ‘General Comment No 27: Article 12 (Freedom of Movement)’ (2 November 1999) UN Doc CCPR/C/21/Rev.1/Add9. 45 Purkey A. (2016) Justice, reconciliation, and ending displacement: Legal empowerment and refugee engagement in transitional processes, Refugee Survey Quarterly, 35, 1-25. 46 Fagen, P.W. (2009). Peace processes and IDP solutions. Refugee Survey Quarterly, 28(1), pp.31-58. 47 Ruta, C., Ruaudel, H. and Bongard, P. (2018). The Guiding Principles and armed non-State actors. Forced Migration Review, (59), pp.37-39. 48 Examples of direct IDP participation in track-one peace processes are rare, since they are typically difficult to access and displaced persons are further disadvantaged by lack of resources, education, political skills or influence, and the lack of a formal interlocutor to support their rights. See: Brookings Institution, Addressing Internal Displacement in Peace Processes, Peace Agreements and Peace-Building, September 2007. Available at: https://www.brookings.edu/wp- content/uploads/2016/06/2007_peaceprocesses.pdf 49 IDPs have participated in track-two initiatives in a number of settings, joining a broader coalition that can offer them greater leverage, and also in track-three process that aim to resolve conflict and encourage reconciliation at the community level, and thereby support IDP return. See: Cohen, R. and Deng, F. (1998) Masses in Flight, Washington DC: Brookings Institution Press, p.259 8 IDPs directly in a multi-track peace process, it is important to develop complementary strategies to represent their interests. 50 The scholarship articulates some expectations regarding what makes a peace agreement more likely to contribute to the safe and sustainable return of IDPs. We use these expectations to structure our analysis of the LPAs in Iraq and focus on the following key factors identified in the academic literature: • Of greatest importance is the ability of IDPs to participate in the negotiation and implementation of a peace agreement, via ‘a process that carries weight,’ 51 meaning that IDP participation has an identifiable impact. 52 However, scholars also acknowledge that even lesser forms of participation (such as outreach or community consultations) can also make an important contribution. 53 • The more comprehensive the involvement in a peace agreement, the more likely it is to achieve a durable peace. 54 That is to say, the inclusion of multiple parties’ interests, beyond those of the armed actors, improves an accord’s chances for overall success. 55 Beyond the inclusion of IDPs, scholars highlight that peace negotiations with greater inclusivity of civil society actors are more likely to contribute to durable peace, 56 as are those that include women in the negotiation process. 57 • In addition to the right to return, it is vital for peace agreements to address the specific concerns held by IDPs, which may include issues such as property restitution, discrimination, and reconciliation. 58 Intra-community relations are highlighted as a key issue to address, given that in many situations, IDPs are not welcomed by local communities and encounter discrimination or even acts of violence upon return. 59 50 This may include international mediators prompting political leaders to incorporate displacement issues in peace negotiations in consultation with IDPs, or focusing on the legal rights of IDPs through international, regional and national mechanisms. See: Brookings Institution, Addressing Internal Displacement in Peace Processes, Peace Agreements and Peace-Building, September 2007. Available at: https://www.brookings.edu/wp- content/uploads/2016/06/2007_peaceprocesses.pdf 51 Magarrell, L. (2007). Outreach to and engagement of victims on reparations–Lessons learned from truth and reconciliation processes. In Speech at the Conference Reparations for Victims of Genocide, Crimes against Humanity and War Crimes: Systems in Place and Systems in the Making. 52 Magarrell, L. (2007). 53 Taylor, D. (2014). Victim participation in transitional justice mechanisms: Real power or empty ritual. Impunity Watch Research. 54 Badran, R. (2014). Intrastate peace agreements and the durability of peace. Conflict Management and Peace Science, 31(2), pp.193-217. 55 Andersen-Rodgers, D.R. (2015). 56 Nilsson, D. (2012). Anchoring the peace: Civil society actors in peace accords and durable peace. International Interactions, 38(2), pp.243-266.; Paffenholz, T., (2014). Civil society and peace negotiations: beyond the inclusion– exclusion dichotomy. Negotiation Journal, 30(1), pp.69-91. 57 Mary, C., Nielsen, R. and Hudson, V.M., (2017). Women and Post-Conflict Settings. In Peace and Conflict 2010 (pp. 91-102). Routledge. 58 Fagen, P.W. (2009). Institute; O’Neill, W. (2009) 59 For example, see: E/CN.4/2006/71/Add.4 (Bosnia-Herzegovina), para. 36; E/CN.4/2006/71/Add.5 (Serbia-Montenegro including Kosovo), para. 56. 9 • Strong commitments written into peace agreements are insufficient in themselves to bring about durable solutions, and the existence of a peace agreement alone cannot ensure return in safety and dignity. Rather, this depends upon the degree of compliance with the full range of provisions in the accord and requires sufficient follow up post-signature. 60 • Finally, it is essential to understand how actors participating in a peace agreement are chosen, whose views they represent, and whether they have their own unstated interests. 61 Peace agreements do not necessarily empower IDPs and generate improved outcomes; in reality, they may in fact empower external actors, reproduce or reinforce power imbalances, or entrench marginalized actors. 62 The existing scholarship thus expects that LPAs will be more likely to facilitate safe and sustainable IDP return if they include IDPs in the negotiation process; include a broader range of perspectives, including civil society actors and women; address complex issues that are of highest concern to IDPs and communities, beyond the issue of return; and establish a mechanism to implement and/or monitor implementation post-signature. Our study aims to contribute to this body of literature by mapping the characteristics of eight LPAs, in order to understand how each was negotiated, its terms and signatories, and how it was implemented and monitored. We then compare these characteristics with the scholarly expectations regarding what makes a ‘good’ peace agreement, according to the four key criteria listed above. Specifically, we ask: How did each LPA align with or diverge from the factors identified in the scholarship? Which factors or practices appeared to support safe and sustainable IDP return, and which practices potentially undermined durable solutions? Section 4: Methodology and Research Design At the outset of this study, the authors identified and mapped all LPAs signed in Iraq between 2015 and 2021, reaching 11 LPAs in total. 63 Of the 11 LPAs originally identified, the authors could not find sufficient information about the negotiation, terms, and implementation of three LPAs, and therefore these agreements were excluded from the study. 64 The remaining eight LPAs are included in the study, and span the governorates of Anbar, Ninewa, Kirkuk, and Salah al-Din. 60 Fagen, P.W. (2009). 61Taylor, D. (2014). Victim participation in transitional justice mechanisms: Real power or empty ritual. Impunity Watch Research.; Cooke, B and Kothary, U. (2001). ‘The Case for Participation as Tyranny, in Cooke, B and Kothary, U. (eds.), Participation: The New Tyranny?, London, Zed Books, 13. 62 Hickey S. and Mohan G. (eds.) (2004) Participation: From Tyranny to Transformation? London, Zed Books. Cooke, B and Kothary, U. (2001). ‘The Case for Participation as Tyranny, in Cooke, B and Kothary, U. (eds.), Participation: The New Tyranny?, London, Zed Books. 63 To the knowledge of the authors at the time of finalizing this study, these are the only LPAs in existence in Iraq from this period. 64 The excluded agreements are: the Ninewa Peaceful Coexistence Pact signed in March 2016; the Heet Pact signed in November 2016; and the al-Sabaawi Agreement of Qayyarah (Mosul) signed in May 2017. 10 The primary data informing this study consists of 32 semi-structured interviews. Interview data was used to understand why each LPA was initiated; how each LPA was negotiated and by who; why specific terms were valued (and by who); how terms were interpreted in practice; how each LPA was implemented after signature; and how the key stakeholders involved in drafting, signing, and implementing each LPA perceived the LPA affected durable solutions and community relations. All interviews were conducted in Iraq in May 2021 by a female Iraqi researcher. Due to government restrictions on intra-governorate travel and meetings (in response to the COVID-19 pandemic) the interviews were all conducted by phone, in Arabic. Interviewees were drawn equally from the eight sub-districts associated with the LPAs, and were chosen based on their direct involvement in, or direct knowledge of, the process of negotiating, drafting, and implementing each LPA. As a result, most interviewees were tribal leaders (13 respondents) or representatives of the local authorities (12), with a smaller number of interviewees drawn from civil society, NGOs or academia (5) or the federal government (2). All interviewees were Iraqi and resided in the area where the LPA applied. Respondents were identified using snowball sampling, with the ‘start’ of each snowball typically a signatory of each LPA or a prominent tribal leader in the sub-district. After the data collection was completed, the authors coded the (translated) interviews utilizing a line-by-line reading of the transcripts in order to identify common themes, and then instigated a second round of focused coding that refined the themes and identified sub-themes for analysis. 65 The interview data exhibits several limitations. The first is that community members and IDPs affected by each LPA were not interviewed, and all interviewees occupied a relative position of power (whether as a tribal leader, government official, or as an active member of civil society). The primary reason for this narrow approach was that data collection took place during the COVID pandemic, when government restrictions on movement and gatherings were enforced across Iraq. The topic of the return of families with perceived ISIL affiliation remains sensitive, and while in- person interviews could have overcome these sensitivities, it was not feasible to conduct phone interviews with ‘ordinary’ community members or IDPs without the risk of creating suspicion upon them; moreover, ordinary residents would be unlikely to discuss the issue on the phone (due to the sensitivity). Powerbrokers, on the other hand, have a wider latitude to talk about these topics since broaching sensitive issues is typically perceived as part of their social responsibilities. 66 A second limitation was that the authors struggled to identify any women who were involved in, or even had direct knowledge of any of the LPAs – a point that speaks to the broader exclusion of women from 65 Emerson, R. M., Fretz, R. I., & Shaw, L. L. (2011). Writing ethnographic fieldnotes. University of Chicago Press; Pokorny, J. J., Norman, A., Zanesco, A. P., Bauer-Wu, S., Sahdra, B. K., & Saron, C. D. (2018). Network analysis for the visualization and analysis of qualitative data. Psychological methods, 23(1), 169. 66 Based on multiple experienced with data collection on the same topic by the authors, as well as advice by the Iraqi researcher. 11 the LPA process, and which we discuss later in this report. As a result, all interviewees were male, which limited not only the diversity of perspectives but also the quality of gendered analysis. To complement the personal perspectives of key-stakeholders, the authors utilized secondary data taken from public databases and published sources. The purpose of the secondary data was to offer a data-based perspective on the context in each location – including security, social relations, and extent of IDP return – before and after each LPA was signed. To do this in a systematic way, the authors selected a set of indicators associated with safe and sustainable IDP return. The choice of indicators was based on two factors: first, we identified indicators which reflected the rights and standards set out in the IASC Framework on Durable Solutions, and second, we excluded indicators that did not have data available throughout the study period (2015-2021). The final set of indicators focused on the number of returnees, security conditions, and social cohesion, as set out in Table 2 below. All indicators were measured at sub-district level, and at regular intervals: as close as possible to LPA signature, and quarterly or bi-annually after signature.67 The data source for the IDP return indicators was IOM DTM’s Master Lists, 68 while the security and social cohesion indicators were drawn from two sources: IOM DTM’s Integrated Location Assessments (ILA), 69 as well as IOM DTM’s Return Index. 70 67 Specifically, IDP return indicators were measured at the closest point in time before each LPA was signed, as well as approximately 3, 6, and 12 months after each LPA was signed, while security and social cohesion indicators were measured at the closest point in time before each LPA was signed, as well as approximately 6 months after each LPA was signed (due to the lower frequency of data collection for these indicators). 68 The study relied on IOM DTM’s Master Lists from Round 21 (May 2015) to Round 121 (May 2021), available at: http://iraqdtm.iom.int/. Note that the DTM database defines ‘returnee’ as an Iraq who became displaced after January 2014, and subsequently returned to their location of origin. 69 The study relied on ILA Round 1 (July – October 2016), Round 3 (March – May 2018) and Round 4 (May – June 2019), as well as the Return Index Round 7 (November – December 2019) to Round 12 (March – April 2021), all available at: http://iraqdtm.iom.int/ Note that the indicators included in the ILA dataset vary with each yearly assessment. For this reason, ILA 2 was excluded from this study because the indicators included in ILA 2 were not comparable to ILA 1 or ILA 3, and no suitable proxy indicators could be identified in the dataset. Note also that the dates of each round refer to the period in which data was collected. 70 The Master Lists contain data on IDP and returnee movements (from the onset of the ISIL crisis in 2014) and are updated every two months. The Integrated Location Assessment and the Return Index each provide location-level data on demographic composition, state of infrastructure, services, security, social cohesion, utilizing different but complementary indicators. While these datasets offer valuable insights into population movements and return conditions, it is important to keep in mind that DTM data aims to provide general trends on population figures, rather than precise figures. This methodology introduces some limitations to the dataset, and therefore to this study: DTM data is collected through interviews with key informants such as mukhtars, local authorities or local security actors, and not via household surveys or registration, and the number of individuals is calculated by multiplying the number of displaced / returnee households by six (being the average size of a family in Iraq.) Iraq’s difficult environment when it comes to limited access and security issues also affects the data collection process. Yet, while recognizing these limitations, it is important to note that no alternative datasets exist with the same geographic scope and breadth of time as DTM; and as such, the authors chose to rely on these datasets as one tool of analysis. For more information on DTM methodology and data collection procedures see: http://iraqdtm.iom.int/MasterList#Methodology 12 Table 1: Secondary Data Indicators IDP Return Social Cohesion Security Number of returnees Whether returnees feel welcome Existence of blocked IDP returns Threats between tribal groups Revenge attacks Mistrust between returnees and host Fear of revenge attacks community Perceptions of ‘daily public life’ 71 The purpose in presenting these indicators is to understand changes in the context rather than imply that the LPA was directly responsible for changes. This is an important limitation to recognize: due to the myriad of factors affecting IDP return, security, and social relations in each community between 2015 and 2021, it is impossible to isolate the impact of each LPA, and we do not intend to imply there is a causal link between the signing of an LPA and changes in the context. Rather, the secondary data offers a way to understand the context that existed in each community before and after each LPA was signed and therefore make sense of the way each LPA was perceived. There were several limitations in comparing indicators across the eight LPAs. Each LPA was signed at a different time and there is no consistent (single or consolidated) dataset that can cover all eight agreements. This means that it is impossible to compare identical indicators across all LPAs; or even to compare the same indicator across all LPAs – to give one example, the returnee population prior to LPA signature was available days before signature for some LPAs, months before signature for others, and not available at all for yet others (since in some locations, international actors were not present when the LPA was signed). 72 The ILA and Return Index also utilize different methodologies, varied timing for data collection, different indicators for similar issues, and different questions for similar indicators. However, despite these limitations, the combined data sources still offer significant contextual information that enables us to identify key markers associated with IDP return, security, and social cohesion in the eight areas before and after each agreement. Another limitation in comparing LPAs was the absence of data related to the return of IDPs with perceived ISIL-affiliation, since this information is not available in public datasets and is not recorded by humanitarian actors. In addition, data by ethnoreligious background is not systematically collected, so it is not possible to disaggregate population movements (or other indicators) by ethnoreligious group. Instead, the number of returned IDPs refers to the overall returns to the sub- district, 73 while the reported return of ISIL-affiliated families pursuant to each agreement is captured via interview data and secondary sources. 71 This indicator measures how comfortable people feel in engaging in public spaces. 72 We rely on the ILA datasets for agreements signed prior to mid-2018 (since no other datasets are available), while indicators for more recent agreements are drawn from the Return Index. 73 Genat, M. (2020). 13 Finally, the authors undertook a thorough review of existing policy and advocacy reports, press releases, and grey literature in order to triangulate and complement the data drawn from interviews and datasets. In doing so, and despite the methodological challenges set out above, we hope to offer a more systematic analysis of LPAs in Iraq than has been attempted previously, and which may enrich the understanding of how local peace processes can resolve protracted displacement after war. Moreover, the methodological limitations highlighted by this study may also encourage actors involved in negotiating local peace agreements to consider how the impact of an agreement may be captured more effectively in future, through timely and targeted monitoring of contextual changes. Section 5: Results This section presents the key findings of the study, in two parts. The first part summarizes each LPA and how it compares with the expectations regarding what makes a ‘good’ peace agreement according to the four key criteria set out in the scholarship: (1) Did the LPA include IDPs in the negotiation process? (2) Did the LPA include diverse perspectives, including civil society actors and women? (3) Did it address complex issues of highest concern to IDPs and communities? and (4) Was a mechanism established to implement and/or monitor the LPA? The second part of this section presents the consolidated findings from the eight LPAs. It considers the factors that were associated with safe and sustainable IDP return, and whether these findings align with, or depart from, the existing scholarship. It also identifies practices that supported safe and sustainable IDP return, and practices that potentially undermined durable solutions for IDPs with perceived ISIL affiliation. Section 5.1: Overview of LPAs in Iraq The table below sets out the key characteristics of each LPA reviewed in the study. Table 2: Summary of LPAs in Iraq Tikrit Peace Accords (2015) The Tikrit Peace Accords took a particularly hierarchical approach to the local peace process, with senior tribal figures and high-level security actors playing a stronger role compared with other LPAs. IDPs accused of ISIL affiliation were not included in negotiations directly and were not even represented by their own tribal leaders, but by intermediary tribal leaders free of accusation of ISIL affiliation. Civil society and women were not included in the peace process. In addition to IDP return, the Accords addressed issues of justice and accountability and approved payments to victim representatives. No mechanism was put in place to implement or monitor the Accords; instead, this was left to the hierarchical structures associated with tribal and security actors involved in 14 negotiations. The context after the Tikrit Peace Accords improved and become more conducive of safe return, and returns took place at a very high rate following the LPA. 74 Covenant of Anbar (2016 and 2018) The Anbar Covenant is the only LPA whose terms have been revised. Neither the original 2016 Covenant nor the 2018 Covenant included IDPs accused of ISIL affiliation in negotiations directly; however, they were represented by their tribal leaders. Civil society and women were not included in either process. The 2016 Covenant focused on punishing accused persons via tribal mechanisms, while the revised 2018 Covenant committed to refer tribesmen accused of ISIL affiliation to the state legal system for prosecution. The 2016 Covenant established (and the 2018 Covenant re- mandated) a Committee headed by tribal leaders and government officials to monitor implementation and take the lead in reviewing decisions made by government actors to permit (or deny) the return of ISIL-affiliated families. The context after the LPA appears to have improved in Ramadi city and seems to be more conducive of safe return. 75 Declaration for Southwest Kirkuk (2017) The Declaration for Southwest Kirkuk was significantly more securitized than other LPAs. IDPs accused of ISIL affiliation were represented by their tribal leaders, and civil society and women were included by the end of negotiations. The Declaration dealt primarily with issues of security and justice and committed to using state institutions not tribal mechanisms. A central Committee was established (comprised of tribal leaders, security actors, officials) and worked with regional (localized) Committees to pursue security-related objectives. The central Committee was responsible to identify ISIL suspects and refer them to the state judicial system, and the regional committees were tasked to prepare evidence of ISIL crimes and cooperate with local security actors. There is no reliable context data available regarding context changes as the agreement was signed while ISIL was still in power; however, after the area was retaken from ISIL, the agreement and context appeared to support safe IDP return to some degree. 76 Peaceful Coexistence and Stabilization Agreement for Yathrib (2018) IDPs accused of ISIL affiliation were represented by their tribal leaders and did not engage directly in negotiations. Civil society was present at the signing and was given a role in the Committee post- signature. 11% of signatories were women. The Yathrib Agreement addressed issues related to security and justice, and focused on preventing retaliation via tribal justice (including collective punishments. The Agreement established a Committee comprised of religious leaders, community members, government actors and NGOs mandated to meet regularly to resolve conflict peacefully. However, this Committee does not appear to play a role in the return of IDPs or engagement with security or judicial actors in determining ISIL affiliation. The context after the Yathrib LPA seemed to improve and become more conducive of safe return. 77 74 According to IOM’s DTM Master List, 1,500 IDPs returned within 2 months of the Tikrit LPA signature, and 63,360 IDPs returned within 6 months. No incidents of discrimination or conflict between IDPs and host communities was registered by the IOM Integrated Location Assessment 1 during this period. 75 By the time the Anbar Covenant was signed, IDP return had taken place at a high level: according to IOM’s DTM Master List, 405,696 IDPs had returned to the provincial capital Ramadi by roughly the same time as the LPA was signed, and this increased incrementally, with 478,596 IDPs returning within 12 months of signature. Context indicators show notable change, with threats, blocked returns and ‘returnees feeling unwelcome’ all reducing to zero in Ramadi within 6 months of signature, and no concerns over revenge attacks or tribal tension reported, according to IOM’s Integrated Location Assessment 4. 76 According to IOM’s DTM Master List, within 6 months of the Southwest Kirkuk LPA signature, IDP returns increased from 3,000 (at 3 months post-signature) to 22,248 (at 6 months post-signature). No context indicators were available due to lack of systematic data collection during that period. 77 According to IOM’s DTM Master List, 21,360 IDPs had returned by the time the Yathrib LPA was signed, and this increased incrementally over 12 months, with 28,368 IDPs returning by 12 months post-signature. Context indicators showed marked improvement, with threats between tribal groups, blocked IDP return, and returnees ‘feeling unwelcome’ 15 Ayadhiya Pact of Honor(2018) The Ayadhiya Pact of Honor was more inclusive than other LPAs, since the peace process included civil society, youth activists and women activists as signatories and also gave them a role during the post-signature monitoring process. IDPs accused of ISIL affiliation were represented by their tribal leaders. The Pact affirmed both tribal and judicial structures as legitimate for dealing with ISIL suspects. The Pact established criteria for IDP return and mandated a Tribal Council (Committee) to assess IDPs against that criteria and give (or deny) approval to return, meaning that the Ayadhiya Committee was more involved in IDP return than other LPAs. There was a slight improvement in the security and social cohesion context after the LPA was signed, although conduciveness of safe return remained low, with threats towards IDPs, physical attacks, mistrust and concern over tribal tension reported in the post-agreement period. 78 Covenant for Peaceful Coexistence in Zummar (2020) The Zummar Covenant is less operational and more principle-based than other LPAs and also has a comparatively wider mandate. IDPs accused of ISIL affiliation were represented by their tribal leaders, and civil society and women were not included in negotiations. The dialogue sessions preceding signature identified land disputes as a key source of instability in Zummar, and the Covenant therefore specified land ownership disputes as within the mandate of the Council to solve. Yet the Covenant offered no concrete information on the type of disputes or authority of the Council. The Agreement established a Committee titled the ‘Community Dialogue Council for Conflict Resolution,’ to address local conflict related to property disputes, IDP return, tribal conflicts, and conflicts over resources and services, affording it a wider mandate than other Committees. The indicator concerning daily public life slightly deteriorated after the LPA, possibly due to the COVID-19 pandemic. 79 Tal Afar Covenant (2020) The Tal Afar Covenant had a relatively strong focus on security. IDPs accused of ISIL affiliation were represented by their tribal leaders, and while civil society actors did not participate in negotiations, female activists did. There was a strong focus on cooperation with security authorities to identify ISIL suspects, and on the role of security actors to vet potential returnees and permit or refuse their return. The pre-existing Local Peace Committee in Tal Afar was mandated to monitor and implement the Covenant. Overall, the context appeared to improve after signature, including improvements to blocked returns, social cohesion, and movement restrictions. 80 reported before the LPA was signed and absent 3 months post-signature, and no concerns over revenge attacks or tribal tension reported, according to IOM’s Integrated Location Assessment 4. 78 According to IOM’s DTM Master List, 18,984 IDPs had returned by roughly the time the Ayadhiya LPA was signed, and this increased to 25,086 by 12 months post-signature. Context indicators indicated poor return conditions: the rate of physical violence, threats, mistrust and returnees ‘feeling unwelcome’ decreased within 12 months, but still remained, according to IOM’s Integrated Location Assessment 4. 79 According to IOM’s DTM Master List, the rate of IDP return increased only marginally following the signature of the LPA in Zummar (likely influenced by the timing of the LPA and the fact that many IDPs had returned already.) Prior to signature, 95,124 IDPs had returned, and this increased to 97,410 after 12 months. Context indicators did not show significant change, with IOM’s Return Index showing similar rates of concern over revenge attacks and ethnoreligious tension pre- and post-signature. 80 According to IOM’s DTM Master List, the rate of IDP return increased only marginally following the signature of the LPA in Tal Aafar (likely influenced by the timing of the LPA and the fact that many IDPs had returned already.) Prior to signature, 169,686 IDPs had returned, which increased to 172,098 after signature. Context indicators such as blocked return, social cohesion and movement restriction improved according to IOM’s Return Index 11. 16 Mahalabiya Covenant of Honor (2020) The Mahalabiya Covenant was brokered to facilitate the return of a comparatively small number of IDP families and had comparatively high rates of community engagement. IDPs accused of ISIL affiliation were represented by their tribal leaders, although family members of returnees were included in the peace process. The Covenant engaged community members as signatories and incorporated civil society actors in activities post-signature to counter radical and extremist views. The Covenant addressed accountability for ISIL atrocities, preventing false accusations, and countering extremist views. The pre-existing Local Peace Committee was mandated to monitor the Covenant, although did not appear to play a role in vetting IDPs prior to return. The context before and after the LPA appears similar, but there is a slight increase in blocked returns and tensions in public life (although these may be partially attributed to military operations in the surrounding areas). 81 81 According to IOM’s DTM Master List, the rate of IDP return increased only marginally following the signature of the LPA in Mahalabiya (likely influenced by the timing of the LPA and the fact that many IDPs had returned already.) Prior to signature, 28,572 IDPs had returned, which increased to 29,994 after signature. Context indicators such as blocked return, social cohesion and movement restriction remained similar, according to IOM’s Return Index 12. 17 Section 5.2: LPAs as a method of Facilitating Safe and Sustainable IDP Return While the eight LPAs in the study shared a common objective – the return of IDPs with perceived ISIL affiliation without revenge attacks – they differed in their terms, inclusivity, and method of implementation. These differences, in turn, had varied implications for safe and sustainable IDP return. The table below sets out the key findings as compared to the expectations set out in the academic scholarship. The remainder of this section then explains these findings in detail. Table 3: Consolidated Findings # Expectation Summary Findings: Iraq LPAs Tribal representation of IDPs enabled negotiations that would not have been possible with direct IDP involvement, and did not appear to undermine safe return in the short-term – in fact, the locations that relied upon tribal hierarchies to represent IDPs (namely, Tikrit LPAs should include IDPs and Yathrib) saw the highest rates of return and the most significant 1 in the negotiation process changes in context indicators. However, excluding IDPs from negotiations potentially undermined the sustainability of IDP return, particularly if community trust in tribal authorities wanes in future. The absence of IDP engagement also contributed to prioritizing community concerns over the IDP right to return. LPAs that engaged civil society and women in the peace process were not associated with safer conditions for return, which appears at least partly due to the limited influence that civilian actors have on high profile security issues in Iraq’s post-conflict environment, LPAs should include a and because LPAs were signed in areas where tribal leaders were 2 broader range of expected to take on the role as interlocutors. However, engaging perspectives civil society and women in the peace process appeared to strengthen the sustainability of IDP return due to the role these community actors played in increasing community awareness and trust in the LPA, monitoring the implementation of the LPAs, and holding actors accountable for breaches. All LPAs aimed to facilitate safe IDP return and prevent revenge attacks. This was pursued by shifting the attribution of guilt from a collective to an individual model by establishing a definition concerning who was ‘innocent’ and therefore permitted to return, and substituting tribal justice mechanisms (which relied on LPAs should address collective punishments) with the state justice system. While this complex issues of highest approach appeared to generate community trust in the return 3 concern to IDPs and process and mitigate revenge attacks, thereby enabling IDPs communities previously under suspicion of ISIL affiliation to return without harm, it also prioritized community concerns (related to security) over the individual right of IDPs to return. In addition, some LPAs created (or supported existing) vetting mechanisms that lacked transparency and created insurmountable barriers to return for some IDPs. LPAs should establish a Mechanisms that had a decision-making role in the IDP vetting 4 mechanism to implement process took on a securitized function and interpreted “safe” IDP monitor post-signature return by emphasizing security criteria and community concerns, 18 and sometimes contradicted the terms of the LPA and the rule of law. Some mechanisms with an implementation function were co- opted by security and tribal actors who were empowered by the LPA and their role in the LPA mechanism, undermining the sustainability of IDP return. Section 5.2.1: IDPs included in the negotiation process Across all locations IDPs had very little direct involvement in the LPA peace process. The equivalent of track-one negotiations exclusively engaged tribal leaders, local authorities, and security actors, and while track-two occasionally engaged civil society and female activists, their role typically focused on awareness-raising and monitoring of the LPA (further discussed in the next section) and did not directly involve IDPs. Track-three appeared to be the only track that engaged IDPs with perceived affiliation, but stakeholders in all locations noted that the purpose of this engagement was to let IDPs know that the LPA was progressing and may enable them to return safely, but it did not involve hearing or trying to incorporate their personal concerns with the peace process. This applied equally to other families in the community, such as victim families, or those who had already returned; across all locations, key stakeholders involved in LPA negotiations noted that there was no involvement at the ‘family’ level. The only way that IDPs could be involved in the peace process was via their tribal leaders, who were expected to represent the views of their constituents, and, equally, to encourage their constituents to accept and abide by the terms of each LPA. While this exclusion of IDPs contradicts the expectations in the scholarship of what makes a ‘good’ peace agreement, making tribal leaders the primary interlocutors enabled negotiations to move forward that would otherwise have been impossible given the high community tensions, since the authority of tribal leaders allowed them to initiate negotiations on behalf of their constituents. LPAs were conceived as a matter of negotiation between tribes – in broad terms, between tribes whose members opposed or were victimized by ISIL, and tribes that had members who joined ISIL 82 – and not between ‘IDPs’ and ‘host communities.’ Describing a sentiment that was repeated across locations, a senior tribal leader in Anbar explained: “There were no host community members or IDPs involved in the negotiation…it wasn’t at this level, it wasn’t about IDPs and the host community” 83 while a political representative in Zummar noted that: “[The LPA involved] tribal leaders and the government, we didn’t include specific families, we [the tribal leaders and government] represented the families.” 84 Most interviewees involved in the negotiation process insisted that representation at the tribal level meant that the process was inclusive, in the sense that individual families or competing ‘sides’ were not excluded; as one tribal leader in Anbar noted: “The 82 As noted earlier, many tribes were themselves split, with some members joining and some members opposing ISIL. 83 Interview with senior tribal leader, Anbar, 25 April 2021. 84 Interview with political representative, Zummar, 17 May 2021. 19 negotiations took place at the tribal level, so there were no specific family that was prevented from participating.” 85 In six LPAs (Anbar, Hawija, Zummar, Ayadhiya, Tal Afar, Mahalabiya) IDPs were represented by their own tribal leaders, while in two LPAs (Tikrit and Yathrib) the the participation of IDPs in the peace process was further diluted as their own tribal leaders delegated their authority to more senior leaders in the tribal hierarchy. According to stakeholders involved in the Tikrit and Yathrib LPAs, this delegation was necessary in order for negotiations to take place, since tribal leaders whose constituents were victims of ISIL were not willing to meet directly with tribal leaders whose members were implicated with ISIL atrocities. At the time of signing the LPA – sometimes only months after the war had ended – IDPs with perceived ISIL affiliation had very poor community relations and many families staunchly opposed their return, and delegating authorities to accepted figures (such as senior tribal leaders) was a way of avoiding confrontation. 86 One tribal leader (and mukhtar) in Tikrit explained a practice that was reported by multiple key informants in Tikrit and Yathrib: “The [head] Sheikh did [initiated and led] the agreement negotiations and we authorized him to do so (…) I wasn’t involved directly, I just gave my vote of agreement to the [head] Sheikh.” 87 While this method of engaging in negotiations diluted IDP representation, the LPAs in both Tikrit and Yathrib witnessed significant returns with relatively low rates of community-led revenge attacks, which was a notable achievement given that Tikrit 88 and Yathrib 89 witnessed some of the highest tensions preceding return and were amongst the most diverse communities (with Sunni and Shia populations living in close proximity.) Tribal leaders also offered a way to ‘guarantee’ the terms of the LPA – that is, to compel compliance on the part of their constituents, whether victimized by or accused of affiliation with ISIL – reflecting the influential role that tribal hierarchy plays in daily life in Iraq, particularly in rural areas. 90 A tribal leader in Tikrit expressed a view that was repeated across locations, noting that: 85 Interview with senior tribal leader, Anbar, 25 April 2021. 86 All interviewees in Yathrib expressed similar sentiment. 87 Interview with Mukhtar #1, Tikrit, 1 May 2021. 88 Tikrit witnessed a surge in returns following the LPA signature: 1,500 IDPs returned within 2 months of the Tikrit LPA signature and 63,360 IDPs returned within 6 months, according to IOM’s DTM Master List. No incidents of discrimination or conflict between IDPs and host communities was registered by the IOM Integrated Location Assessment 1 during this period, which is striking given that although no quantitative data is available pre-signature, other sources point to high levels of tensions and threats of violence (for example, see: Human Rights Watch (2015). Ruinous Aftermath. Militia Abuses Following Iraq’s Recapture of Tikrit and Gaston, J. and Maas, F (2017). ‘Iraq after ISIL: Tikrit and Surrounding Areas’. GPPI. 30 August Available at: https://www.gppi.net/2017/08/29/iraq-after-isil-tikrit-and-surrounding-areas). 89 In Yathrib, 21,360 IDPs had returned by the time the Yathrib LPA was signed, and this increased incrementally over 12 months, with 28,368 IDPs returning by 12 months post-signature, according to IOM’s DTM Master List. Context indicators showed marked improvement in Yathrib, with threats between tribal groups, blocked IDP return, and returnees ‘feeling unwelcome’ reported before the LPA was signed and absent 3 months post-signature, and no concerns over revenge attacks or tribal tension reported post-signature, according to IOM’s Integrated Location Assessment 4. 90 IOM Iraq (2020). Perceptions of Security and Police in Iraq. Baseline survey findings. 20 It took time and negotiations to convince all the people that this agreement is for the best, and to convince the community to follow this agreement … this was the role of tribal leaders and mukhtars (many of whom are also tribal leaders) after the agreement was signed, to follow up to ensure that safe return was happening. 91 However, while reliance on tribal structures facilitated difficult negotiations and contributed to community compliance, it often meant that community concerns were amplified and prioritized over the rights of IDPs. Most LPAs established – or confirmed – criteria for IDP return, typically a security clearance, and some mechanisms established to monitor or implement the LPA were actively involved in IDP vetting. While security vetting was critical for community trust in the process and acceptance of IDP return, the security vetting process was often unclear and inconsistent, with little if any opportunity to appeal a negative decision. 92 The absence of IDP voices in the negotiation phase as well as a lack of formal recognition for IDPs in the monitoring process meant that these concerns were not addressed in any of the LPAs. As discussed further below, this is particularly concerning given the marginalized position of IDPs with perceived affiliation, as well as the tendency of LPA mechanism to become co-opted by actors whose interests contradicted those of Sunni Arab IDPs. Section 5.2.2: Include a broader range of perspectives Three LPAs (Tikrit, Anbar, and Zummar) did not engage actors such as civil society, women, or youth in the LPA negotiations or signature process. Interviewees in these three locations insisted that tribal leaders represented all members of the society and therefore it was not necessary to involve more diverse actors. One tribal leader (and mukhtar) in Tikrit expressed a view that was repeated across all these locations: “Nobody was excluded because all the tribes authorized the [head] Sheikh to manage the negotiations.” 93 Five LPAs did involve women, youth or civil society actors in the peace process (Yathrib, Hawija, Ayadhiya, Tal Afar, and Mahalabiya). However, even in these locations where diverse actors were included in the peace process, many interviewees (who were mainly tribal leaders and local authorities) appeared to share the same view, with one interviewee from Hawija explaining: “We didn’t need to include other layers, all the people authorized their tribal leaders to talk and vote for them.” 94 As such, in these areas the participation of civil society, youth, and women often appeared symbolic, at least during the negotiation and signature phase. Women, youth, or civil society actor were invited to attend the signing ceremony – and in the case of Yathrib and Mahalabiya, join as signatories – but did not contribute to the substantive terms of each LPA, since the negotiation process and decision-making was managed exclusively by tribal 91 Interview with tribal leader in Tikrit, 29 April 2021. 92 Genat, M. (2020). 93 Interview with mukhtar #1, Tikrit, 1 May 2021. 94 Interview with civil society representative, Hawija, 8 May 2021. 21 leaders and government authorities (and to a lesser extent, actors such as mukhtars.) Tribal leaders typically understood that it was their role to meet with constituents (i.e. people who belonged to their tribe) who were also civil society actors, academics, teachers, or activists, in order to discuss the proposed terms – similar to a track-two process – and the responsible tribal leader would then present community views during the negotiations. As one tribal leader in Yathrib explained: The head of the tribal leaders, me and some local [people] from Yathrib like researchers, doctors, teachers… initiated the agreement negotiations. … The [head] Sheikh played a big role by discussing the issue, and explaining more to the other tribes, encouraging us that we should vote for this agreement and the aim is to create a safe environment after return. 95 Sometimes the role of engaging civil society actors in a track-two process was taken on by a government representative (typically local authorities such as the Mayor) although even in that case, tribal leaders remained important constituents and the role of civil society actors was downplayed. A government official in Mahalabiya noted that: My role (in the process) was assembling a group of tribal leaders, activists, and representatives from different areas, and we agreed on which items to add to the agreement… the most influential person was [government representative] and also the support of the tribal leaders … voting [took place] amongst the representatives of the community layers … it was inclusive because we included all the tribes and the governmental side too. 96 The absence of diverse actors from the LPA peace process did not appear to affect immediate conditions for safe IDP return, with contextual indicators giving mixed results. Of the three LPAs that did not engage diverse actors, both Tikrit and Anbar witnessed high rates of IDP return 97 and both locations showed marked contextual changes: in Tikrit, there were no reports of discrimination or conflict between IDPs and host communities post-signature, a notable finding given the high rates of tension and threat of violence reported pre-signature, 98 and in Anbar, threats, blocked returns and ‘returnees feeling unwelcome’ all reduced to zero within 6 months of signature in the provincial capital of Ramadi, with no concerns over revenge attacks or tribal tension were reported. Of the five LPAs that did involve women, youth or civil society actors in the peace process, conduciveness of safe return remained low in Ayadhiya, where threats towards IDPs, physical attacks, mistrust and concern over tribal tension remained relatively high in the post-agreement period despite the LPA 95 Interview with Tribal Leader #1, Yathrib, 26 April 2021. 96 Interview with political representative, Mahalabiya, 12 May 2021. 97 In Tikrit, 1,500 IDPs returned within 2 months of the LPA signature, and 63,360 IDPs returned within 6 months, according to IOM’s DTM Master List. For Anbar, 405,696 IDPs had returned to Ramadi by roughly the same time as the LPA was signed, and this increased incrementally, with 478,596 IDPs returning within 12 months of signature, according to IOM’s DTM Master List. 98 See footnote 88 above. 22 process showing the highest engagement with civil society, women, and youth; 99 context indicators remained relatively poor (without change) in Mahalabiya,100 and context indicators showed improvements in Yathrib 101 and Tal Afar. 102 While there are a myriad of reason why people may not report instances of violence or threats (including fear, or perceived pressure to give socially-desirable answers), and it is not possible to isolate the LPA as a source of contextual change, the available data does not suggest that the inclusion of diverse actors led to higher rates of, or safer returns. An important contextual factor that may explain the limited influence of civil actors over the LPA process, as well as low value placed on their participation by government and tribal actors involved in the LPA negotiations, was that most LPAs were signed in an environment characterized by insecurity, a fractured security sector, and instances of revenge attacks and ongoing military operations in some areas. In this context, opportunity for community members to influence issues related to security was highly constrained, particularly high-profile issues like the return of persons with perceived ISIL affiliation. Rather, these issues are typically dealt with by a collection of security actors who often compete for dominance, alongside government officials and tribal leaders. 103 However, LPAs that engaged civil society and women appeared more likely to support the sustainability of IDP return, and the long-term success of the LPA, in three ways. First, the return of families suspected of ISIL affiliation took place gradually; some families with ‘lower’ levels of perceived affiliation who were less fearful of revenge attacks returned even while negotiations were ongoing, encouraged by the fact that the peace process had been initiated. 104 This gradual approach helped to build trust in the return process, both at the community level (that the return of families would not destabilize their community) and with IDPs with perceived ISIL affiliation (that they were safe from revenge attacks). Community members such as civil society and women played a vital role as the peace process advanced, as they shared information amongst their networks about the existence of the LPA and how it was being implemented. 105 This helped to generate momentum and encourage additional returns; as a political representative in Zummar noted: 99 According to IOM’s Integrated Location Assessment 4, the rate of physical violence, threats, mistrust and returnees ‘feeling unwelcome’ decreased within 12 months, but still remained. 100 Context indicators such as blocked return, social cohesion and movement restriction remained similar, according to IOM’s Return Index 12. 101 Context indicators in Yathrib showed marked improvement, with threats between tribal groups, blocked IDP return, and returnees ‘feeling unwelcome’ reported before the LPA was signed and absent 3 months post-signature, and no concerns over revenge attacks or tribal tension reported, according to IOM’s Integrated Location Assessment 4. 102 Context indicators in Tal Afar such as blocked return, social cohesion and movement restriction improved according to IOM’s Return Index 11. 103 For governorate level mappings of actors involved in the return process, as well as specific challenges see: RWG Return Process Mapping series available at: http://iraqrecovery.org/RWG. For example, Anbar governorate mapping available at: http://iraqrecovery.org/Files/2051/2191.pdf shows the multiplicity of competing security actors, governmental actors and tribal leaders involved in the process of return. 104 Interview with thematic expert, Erbil, 20 May 2021. 105 Nonviolent Peaceforce (2020). 23 “When you see your friends, relatives, return safely you will return too and this is what happened, families started to encourage each other to return.” 106 Second, community members – and particularly civil society actors – played a vital role in monitoring the implementation of LPAs. This was demonstrated most clearly in the case of Anbar, where advocacy by civil society actors captured the attention and support of the international community, and their joint efforts led to revisions of the Anbar Covenant to address the application of collective punishments and inconsistency with the rule of law and human rights standards. More recently, advocacy by civil society actors in Tal Afar and Ayadhiya have led to calls for the revision of those LPAs also. In Ayadhiya, 80% of female respondents interviewed in a study on the LPA (all of whom were returnees in Ayadhiya) stated that they saw a role for themselves in implementing the Pact through awareness raising, community engagement and monitoring breaches. 107 Here, it appears that involving community members during the negotiation phase enabled them to play a stronger role in monitoring, since they became familiar with key interlocutors as well as with the intention of the terms of the LPA and could then identify shortcomings or irregularities of each LPA more effectively. Third, community involvement helped to maintain political attention on the peace process and momentum in the post-war community recovery phase. Reflecting on the shortcomings of the Anbar Covenant over time due to the lack of strong engagement with broader society, a security actor in Anbar noted: Community committees, made of doctors, tribal leaders, security employees and myself were following up, but we didn’t hit the goal because there was neglect by the government itself, then the community became busy with new subjects: the political conflict, the elections… which made the activation of the Agreement decrease over time.108 Section 5.2.3: Address complex issues of highest concern to IDPs and communities The primary objective of all LPAs was the safe return of IDPs, and specifically IDPs who were not criminally liable under Iraq’s anti-terrorism legislation but were nonetheless perceived as affiliated to ISIL. 109 Interviewees across all locations agreed that an LPA was ‘successful’ if IDP return took place in the absence of revenge attacks, and consistently pointed to the safe return of IDPs as the biggest achievement of each LPA. A second and closely related priority, again expressed across all locations, was to distinguish ‘degrees’ of ISIL affiliation by establishing agreed criteria and a mechanism to make that distinction. The aim in doing so was to dispense with the sweeping accusations of guilt that dominated the immediate post-war context and encourage communities to 106 Interview with political representative, Zummar, 17 May 2021. 107 Nonviolent Peaceforce (2020). 108 Interview with security actor, Anbar, 24 April 2021. 109 As noted earlier, there are many reasons for perceived affiliation, including the actions of an extended family member or a deceased (or missing) husband, or even the length of time spent living under ISIL occupation. 24 accept a much more narrow and nuanced interpretation of guilt; and also to stop the application of collective punishments by replacing tribal justice methods with a commitment to the rule of law. To do this, each LPA put in place a system that prevented the return of IDPs whose family member had a direct and more ‘serious’ affiliation to ISIL but enabled those with a ‘lesser’ or more distant affiliation to ISIL to return in safety. However, across all locations it was clear that the two goals – namely, the safe return of IDPs and distinguishing guilt at the individual level – reinforced each other, since safe return was only possible if community members trusted that families with direct and genuine ties to ISIL were not amongst those returning. The criteria distinguishing ‘innocent’ from ‘guilty’ IDPs was defined and implemented differently in each LPA, discussed further below, and the commitment to the rule of law and state justice was also expressed differently in each place. In order to distinguish those who were ‘cleared’ of ISIL affiliation and therefore could return from those who could not, all LPAs relied (as a minimum) on the pre-existing government security vetting process. 110 Most LPAs explicitly stated that this security vetting superseded community perceptions about guilt; the Tikrit Accords, for example, stated that regardless whether IDPs “fled ISIL’s arrival or remained during ISIL occupation,” they were entitled to return, “provided they passed the security vetting process, clearing them of ISIL affiliation.” While the LPAs helped communities move away from sweeping perceptions of collective guilt, these security clearances still applied to the family unit, meaning that if the head of household failed the vetting process, his wife and children would similarly be without a clearance. Equally, women whose husbands were missing or imprisoned often struggled to obtain a security clearance.111 For individuals who wished to separate themselves from their accused family member in order to obtain a security clearance, several LPAs recognized an existing practice – the disavowal process – to do so. 112 The LPAs in Ayadhiya and Anbar (2018) stated that an IDP family with a ‘first degree relative’ who joined ISIL could return ‘if they formally renounce the family member’ through the disavowal process, while the Mahalabiya LPA stated that IDP families that ‘rejected the crimes committed by relatives who were associated with ISIL’ may also return. The requirement in the Tal Afar Agreement was slightly different, noting that ‘individuals found guilty during the security clearance process of being an ISIL supporter or member cannot return; however, their families can 110 The LPAs in Tikrit and Yathrib did not specific any additional criteria for IDP return beyond the standard security clearance. The LPA for Southwest Kirkuk simply stated that “anyone suspected of ISIL affiliation must be referred to the state’s legal system for adjudication.” 111 Genat, M. (2020). 112 The disavowal process has become common across Iraq in the post-war period as a means of enabling the return of families that had a first-degree relative with close ties to ISIL. It is practiced in two forms: one, known as tabriya, occurs in front of a tribal committee and does not have a basis in Iraqi law, while the second, known as ikhbar, is a form of ‘notification’ that takes place in front of a judge and has been recognized in Iraqi law for years preceding ISIL’s arrival. The core idea of both processes is similar: namely, that the person formally pledges that their relationship with the accused relative is severed. For more information in these practices see Genat, M. (2020). 25 return provided they publicly condemn extremism and pledge to support the legal ruling against their family member.’ This approach of moving from collective ideas of guilt to a more nuanced approach had a positive effect on the scale of IDP return, since it allowed large groups of IDPs previously under collective suspicion of ISIL affiliation to return. IDP return also appeared to take place with fewer instances of revenge attacks. 113 The criteria and system recognized by each LPA were vital to this, since they generated community trust and acceptance of the process of return and thus prevented revenge attacks. 114 A political representative from Ayadhiya who participated in negotiations explained that: Some Yazidi families (who live close to this area) faced issues with some Arab tribes who joined ISIL, and the Yazidi families refused to participate in the negotiation for this Peace Agreement, but we promised them all that anyone with a dirty hand (meaning anyone involved with ISIL) will never return. 115 However, by recognizing specific criteria and mechanisms to facilitate IDP return, the LPAs also introduced new barriers to return for some IDPs. If an IDP family was rejected by the mechanism associated with the LPA, it reinforced their inability to return. As one senior tribal leader in Anbar noted: There were families who were proven guilty after the vetting process and proven to be one hundred percent affiliated to ISIL, and the agreement couldn't help them because of their direct support to ISIL. 116 The problem with this perception that the vetting process determined ‘one hundred per cent affiliation to ISIL’ is that the security vetting process has been widely criticized for its opacity and lack of due process; moreover, it is not clear which security body is entitled to screen IDPs and which criteria they should apply. 117 A Human Rights Watch report dated July 2019, for instance, listed five different actors involved in the security screening of 3,500 IDP families in Hamam al-‘Alil camp: military intelligence, national security, Special Weapons and Tactics (Ministry of Interior), local police, and, most controversially given their role in abuse of civilians and destruction of civilian property as well as interference in the process of IDP return, the Popular Mobilization Forces. 118 These different 113 There were some exceptions to safe returns. In Ayadhiya, interviewees reported that a small number of revenge attacks took place in the period after the LPAs signature, while secondary data also points at revenge attacks occurring in Anbar post-signature of the LPA, and that fear of revenge attacks continued to prevent IDP returns in Yathrib despite the LPA. For more information on each location please refer to the relevant Location Profile, at Annnex 1. 114 Other studies re-iterate the importance of community acceptance to avoid revenge attacks or the start of new conflict when families with perceived ISIL affiliation return – for example, see: Aymerich, O., 2020. Community Resilience and the Return of Iraqi IDPs with Perceived Affiliation to the Islamic State. Refugee Survey Quarterly, 39(4), pp.552-563. 115 Interview with political representative, Ayadhiya. 10 May 2021. 116 Interview with senior tribal leader, Anbar, 25 April 2021. 117 Genat, M. (2020). 118 HRW: Iraq: Military Enter Camp, Occupy School for ‘Screening’. Displaced People Fear Arrest over Actions of their Relatives. July 2019 ,18. Available from https:// www.hrw.org/news/18/07/2019/iraq-military-enter-camp-occupy-school- screening 26 groups and agencies typically hold their own lists of alleged ISIL members, of unclear source and which are often contradictory. 119 LPAs did not acknowledge these shortcomings in the vetting process, nor did they provide any way to challenge a negative vetting outcome. A final but separate issue raised by interviewees concerned the relevance of each LPA over time. Several interviewees involved in earlier LPAs noted that they focused on immediate issues – namely, the safe return of IDPs and the absence of revenge attacks – but did not address broader issues related to IDP reintegration, such as housing or livelihood support. In the view of some interviewees, this undermined the ongoing relevance of the peace process and its ability to contribute to sustainable IDP return, since it did not extend to long-term reconciliation or reintegration issues, but limited its focus to the immediate post-war phase. One tribal leader in Yathrib noted: I can’t say if the agreement is positive or negative, because it helped people return but it didn’t fix their status after return, like rebuilding their houses or supporting them financially.120 An academic in Anbar also noted: I think the agreement, in the beginning, it addressed the most important issues related to return, as it enabled safe returns and it really helped, but it needs more items, and edits now [to address the issues that are important today]. 121 Section 5.2.4: Establish a mechanism to implement monitor post-signature With the exception of Tikrit, which relied on tribal hierarchy to implement and monitor its LPA, all LPAs established a mechanism (typically a Committee) to oversee implementation and monitor the agreement. 122 In all locations, these committees conducted outreach to community members in order to community understanding – and acceptance – of the process of IDP return. In most locations, after the LPA was signed, Committee members held meetings with concerned residents to explain that the LPA would only allow those who were ‘clean’ of ISIL affiliation to return, and that they could rely on the government security vetting process. 123 This message – which commenced during negotiations – required engagement over time by trusted figures within the community and extending months after the LPA was signed. As one tribal leader in Hawija noted: Some families [in Hawija] gave martyrs [i.e. their sons or brothers were killed] fighting against ISIL, such as the Jabouri tribe, and as a result they were against this [Local Peace Agreement] in the beginning because they lost their people during the ISIL war. But then, 119 Nonviolent Peaceforce: A Study Amongst Displaced and Returned Communities of al-Ayadiah, Iraq. Spring 2020. 120 Interview with tribal leader, Yathrib, 26 April, 2021. 121 Interview with civil society representative, Anbar, 25 April 2021. 122 Interviewees from all seven locations noted that the Committees were active (to varying extents) in their functions. 123 Gharizi, O. and Al-Ibrahimi, H. (2021) at 4. 27 they started to see that this Agreement will create a safe environment for the returnees and protect people from ISIL, which made them support the agreement. 124 Besides this shared function, there was one important distinction between the mechanisms established or utilized by each LPA. Four mechanisms (associated with the LPAs Yathrib, Zummar, Tal Afar, Mahalabiya) had an implementation and monitoring function but, in theory, were not involved in the process of IDP return (such as vetting or identifying ISIL suspects). Three mechanisms (Anbar, Hawija, Ayadhiya) had a decision-making role in the process of IDP return (given by tasking them to participate in the security vetting process) and were also mandated to identify ISIL suspects. In this second group, some Committees were explicitly mandated to work with security actors in the IDP vetting process (and which was a deciding factor in who could return.) In Hawija, for example, a Committee member (who was also a mukhtar) explained: [After the LPA was signed] I was receiving calls from the security checkpoints when they had questions about [returning IDP] families, so I could support them in the vetting process. … The Committee [established by the LPA] was made of people from the same neighborhood and they were coordinating with the security forces to vet the [returning IDP] families and thereby complete their security checks. 125 Some Committees, such as in Ayadhiya, also had the power to set the criteria for IDP return and then decide whether an IDP family passed that criteria, based on their subjective knowledge of that family. There was no formal way to appeal such a decision. This gave tribal leaders significantly more power than they had before the LPA was signed, since prior to the LPA this role was left to security actors. 126 The power given to Committee members enabled them, in some instances, to co-opt the process of IDP return and enforce their own criteria, which sometimes contradicted the terms of the LPA and the commitment to the rule of law. For example, according to the Ayadhiya LPA, disavowal is only relevant for people with first degree affiliation. However, security and tribal actors are increasingly refusing people with first degree ISIS affiliation to return or to disavow. This causes tensions within some communities as some residents complained that the Committee shouldn’t have the authority to investigate cases, but only to advise people to follow the rule of law while investigations were the work of state authorities such as the police. 127 Even those Committees that were not given a formal mandate to engage in IDP vetting often participated in the vetting process in practice. In Zummar and Yathrib, while the Committees were only mandated to mitigate potential conflict and community tension associated with IDP returns, they 124 Interview with tribal Leader, Hawija, 8 May 2021. 125 Interview with Mukhtar #4 Hawija, 10 May 2021. 126 Nonviolent Peaceforce (2020). 127 NVP report on Ayadhiya LPA. 28 often functioned as de facto vetting mechanisms, with Committee members actively advising which families were allowed to return. As one tribal leader in Yathrib noted: The [PMF] and local committees are cooperating to support in implementing the Agreement in the right way by filtering families through the check points and giving fast security permissions. 128 As a result of the function given to the Committees in IDP vetting and identification of ISIL suspects, some Committee members received greater power and authority, particularly those involved in deciding which IDPs could return. This change in dynamics is important to acknowledge and analyse because, as noted earlier, peace agreements may empower external actors in a way that reproduces or reinforces power imbalances or entrenches marginalized actors. 129 As the literature notes, it is vital to establish a mechanism to monitor the implementation of a peace agreement – and, as the Iraq case demonstrates, it is vital that such a mechanism is neutral and ideally engages only in monitoring, not in implementation of the peace agreement terms, in order to avoid co-optation and enable objective monitoring of the peace process implementation. A failure to do this is likely to undermine safe and sustainable IDPs return and create new grievances amongst conflict-affected populations. Section 6: Policy and Program Implications The recommendations outlined below relate to the return of IDPs with perceived links to ISIL but have wider application, particularly with displaced communities divided due to wartime experiences. Local peace agreements offer a valuable way to negotiate the return of IDPs formerly rejected by their communities and build community trust in the return process. While data limitations did not allow us to isolate the characteristics of more ‘effective’ LPAs, in all locations families with perceived ISIL affiliation did return after the relevant LPA was signed and typically without revenge attacks. Situating the LPA at provincial, district, or even community level required greater resources than negotiating a single, national LPA, but enabled each LPA to contextualize its terms, address community concerns in a practical way, and commence in a timely manner for that community. As such, this localized approach is worth replicating, particularly in situations where war experiences differ across the country or where conflict ended at different points nationally. LPAs can influence practices related to IDP return and post-conflict community justice so they align with human rights standards and the rule of law. In Iraq, LPAs often committed signatories and the wider community to uphold the rule of law, and this led to a move away from collective punishments and the referral of cases to the state justice system. LPAs should include terms that moderate practices inconsistent with international human rights principles (such as collective 128Interview with tribal leader, Yathrib, 26 April 2021. 129Hickey S. and Mohan G. (eds.) (2004) Participation: From Tyranny to Transformation? London, Zed Books. Cooke, B and Kothary, U. (2001). ‘The Case for Participation as Tyranny, in Cooke, B and Kothary, U. (eds.), Participation: The New Tyranny?, London, Zed Books. 29 punishments) while also building on the positive aspects of traditional practices, such as the legitimacy they enjoy amongst communities, channels for dialogue and mediation, and trusted leadership. Local Peace Agreement negotiations should utilize a broader range of methodologies to engage IDPs in the negotiation and implementation of an LPA. IDPs who stand accused as perpetrators and face community rejection may not be able to participate directly in a peace process due to the risk of conflict or breakdown in negotiations, and may struggle to have their voices heard or their concerns addressed. The Iraq experience suggests that indirect representation (such as via tribal leadership) can enable negotiations to move forward that would otherwise be impossible; however, it also indicated that this method of engagement may contribute to prioritizing community fears or concerns over IDP rights or experiences. To give IDPs a stronger platform to voice their concerns, actors involved in LPAs should consider the use of track-two or track-three interventions, which could involve mandating an NGO or civil society actor to represent marginalized IDPs during negotiations and advocate their concerns, and then also help communicate their experiences during the implementation of the LPA. The passage of time may enable stigmatized groups of IDPs to be involved more directly in the LPA implementation, and this possibility should be considered regularly by those mandated to implement the LPA so as to acknowledge and address concerns held by IDPs regarding the terms and implementation of the LPA. Local Peace Agreement negotiations should also utilize a broader range of methodologies to create a more inclusive peace process and capitalize on the contributions of civil society, women, and youth. The Iraq experience suggests that civil society actors and women play a valuable role in building awareness of the LPA and trust in its terms, as well as monitoring implementation of the agreement. These actors will be able to fulfil these roles more effectively if they are familiar with the key stakeholders, terms of the LPA, and dynamics during negotiations, and this requires intentional engagement by parties leading the peace process. These parties should utilize broader methodologies to engage civil society, women, and youth in a meaningful way throughout each step of the LPA process, utilizing track two and track three options to mobilize, inform, and enable meaningful participation. To increase the impact of track-two and track-three activities on track-one negotiations, it is also important to shift the attitudes of decision-makers who do not value broader perspectives. This may be done by providing concrete examples of the important role that civil society and women can and have played in LPAs. The LPA negotiation process should pre-empt, mitigate, and address co-optation by actors that undermines the aim or sustainability of an LPA. The Iraq experience suggests that LPAs are at high risk of co-optation in areas where non-state actors are dominant, when the LPA empowers non- state actors to determine which IDPs can return, or when terms of the LPA do not put in place 30 sufficient safeguards. Steps to mitigate or address co-optation include recognizing and addressing asymmetrical power dynamics; identifying and tracking when new actors are empowered by the LPA; ensuring that the perspectives of marginalized groups are heard throughout the peace process; including clear terms in the LPA; and ensuring that the LPA is subject to regular monitoring and periodic review. LPAs require regular monitoring and should be periodically reviewed (and potentially modified). It is vital to monitor how the LPA is affecting all IDPs, not just those who successfully manage to return, in order to recognize new barriers to return that an LPA may be creating. The international community along with national civil society organizations and key government institutions can help establish a monitoring mechanism or adapt an existing one, with clear criteria to monitor the impact of the LPA and agreed steps to take in the event of breaches. For monitoring to be effective, it is vital that diverse actors are involved, including IDPs themselves, and that these actors are well-informed about the terms of the LPA and the type of actions that constitute a violation. Monitoring should also be used to identify instances where the LPA is creating new grievances or barriers to return, and ensure that the terms are consistent with a human-rights approach. Implementation and monitoring functions should be separated, to reduce the risk of mis-use or co-optation. The Iraq experience highlights the potential for mis-use and co-optation of an LPA when it is not accompanied by an independent monitoring mechanism, or when the same mechanism is tasked with both implementation and monitoring of implementation. To mitigate this possibility, the monitoring function should be given to a mechanism that operates independently from implementation. If LPAs are to contribute to durable solutions in the longer-term, they should be accompanied by a significant investment in reintegration assistance. An LPA alone is not sufficient to change the conditions of return, and if IDPs return to areas where resources and basic services are insufficient it will undermine the outcomes negotiated by the LPA, undermine durable solutions and potentially generate fresh conflict or grievances. Actors involved in negotiating the LPA should coordinate with humanitarian and development actors so that an increase in IDP return can be complemented by reintegration assistance such as rehabilitation of physical infrastructure, housing, and services. 31