A Community Driven Governance Programme in Somalia/Somaliland.', Project Report

Durham Research Online

Deposited in DRO: 03 November 2020 Version of attached le: Published Version Peer-review status of attached le: Not peer-reviewed Citation for published item: Bakonyi, Jutta and Cohen, Gidon and Bedard, Pierre-Olivier (2015) 'An evaluation of Hogaan iyo Nabad : a community driven governance programme in Somalia/Somaliland.', Project Report. Durham University , Durham. Further information on publisher's website: https://www.dur.ac.uk/dgsi/cdgsomalia/

Publisher's copyright statement:

Additional information:

Use policy

The full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-prot purposes provided that: • a full bibliographic reference is made to the original source • a link is made to the metadata record in DRO • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders.

Please consult the full DRO policy for further details.

Durham University Library, Stockton Road, Durham DH1 3LY, United Kingdom Tel : +44 (0)191 334 3042 | Fax : +44 (0)191 334 2971 https://dro.dur.ac.uk An Evaluation of Hogaan iyo Nabad: A Community Driven Governance Programme in Somalia/Somaliland

Dr Jutta Bakonyi Dr Gidon Cohen Dr Pierre-Olivier Bedard´ Durham University, UK This report is the ﬁnal evaluation of the Hogaan iyo Nabad Community Driven Recon- struction programme in Somalia/Somaliland. The programme, including the evaluation was funded by DFID. Hogaan iyo Nabad was implemented by a consortium of international organisations (IRC, CARE and DRC). The evaluation was conducted by researchers from the School of Government and International Aﬀairs, Durham University in cooperation with OCVP. https://www.dur.ac.uk/dgsi/

First release: September 2015 Contents

List of Tables iii

List of Figures iv

Abbreviations viii

Acknowledgements ix

Executive Summaryx

1 Context of the Hogaan Programme and Evaluation1 1.1 Community Driven Development/Reconstruction...... 1 1.2 Evaluations of Community Driven Programmes...... 2 1.3 Hogaan iyo Nabad (Governance and Peace)...... 3 1.3.1 Problem Statement...... 4 1.3.2 Theory of Change...... 5

2 Evaluation Design and Implementation8 2.1 Research Design...... 8 2.2 Sampling Techniques...... 15 2.2.1 Qualitative Sampling...... 15 2.2.2 Survey Sampling...... 15 2.3 The Surveys...... 17 2.3.1 The Household Survey...... 17 2.3.2 Leadership Survey...... 18 2.4 Evaluation Implementation...... 18 2.5 Sample: Description and Results of Experiments...... 19 2.5.1 Satellite Image versus EPI sampling...... 19 2.5.2 Experiment Result...... 19 2.5.3 Within Household Randomisation...... 21 2.6 Statistical Methods...... 23

3 Evaluation Findings 24 3.1 Participation (Outcome 1)...... 27 3.1.1 Citizens’ Participation in Decision-Making...... 27 3.1.2 Participation and Inclusiveness...... 31 Conclusion: Hogaan’s impact on participation and inclusion...... 47 3.2 Improving Governance Capacity (Outcome 2)...... 48 3.2.1 Coordination and Administration...... 48 3.2.2 Service Delivery...... 53

i 3.2.3 Conflict Management...... 55 3.2.4 Trust in Village Governance...... 57 Conclusion: Hogaan’s Impact on Local Governance...... 58 3.3 The Survey Experiment...... 59 3.3.1 Findings...... 61 3.3.2 Comparison with Non-implementation Villages...... 62 3.3.3 Impact of Financial Information on Project Preferences...... 62 Conclusion: Summary of survey experiment results...... 63 3.4 Training Effects...... 64 3.4.1 Spillover...... 70 Conclusion: Training or Selection Effects?...... 71

4 Summary and Discussion of Findings 73 4.1 Hogaan’s Impact on Participation...... 73 4.2 Hogaan’s Impact on Village Level Governance...... 74 4.3 Training and Selection Eﬀects...... 75 4.4 Systematic Overall Summary of Findings...... 76

5 Recommendations Relating to the Theory of Change 78

Bibliography 88

A Appendix: Data Analysis Plan Results 91 A.1 Programme Mechanisms Analysis As Identiﬁed in the Data Analysis Plan. 91 A.1.1 Citizen Participation in decision-making...... 91 A.1.2 Mechanism 2: Strengthening Institutions...... 92 A.1.3 Data Challenges...... 92 A.1.4 Overview of Data Analysis Plan Table 3 Results...... 92 A.2 Detailed Results...... 93

B Appendix: Further Results 102 B.1 Further Model Results (Hogaan and Non-Implementation Villages).... 102 B.2 Further Model Results (Gender Interaction)...... 108 B.3 Further Model Results (Clan Interaction)...... 119 B.4 Further Model Results (Age Interaction – Ref: 35+)...... 127 B.5 Further Model Results (Age Interaction – Ref: <25)...... 137 B.6 Institutional Support for Services and Rights...... 147

C Appendix: Diagramatic Representation of Original and Revised Theory of Change 154

ii List of Tables

2.1 Comparison of Implementation and Non-implementation Villages..... 11 2.2 Villages in (Matched) District Burtinle...... 12 2.3 Villages in (Matched) District Galkayo...... 13 2.4 Villages in (Matched) District Erigabo...... 14

3.1 Clan Distribution of Burtinle Survey Respondents...... 33 3.2 Clan Distribution of Galkayo Survey Respondents...... 34 3.3 Clan Distribution of Erigabo Survey Respondents...... 34

4.1 Overall Summary of Quantitative Evaluation of Hogaan Project...... 77

5.1 Village Size Estimates...... 81

A.1 Overall Data Analysis Plan Results...... 93 A.2 Regression Results for H1...... 93 A.3 Model Comparison for H1...... 94 A.4 Regression Results for H2...... 94 A.5 Model Comparison for H2...... 95 A.6 Regression Results for H3...... 95 A.7 Model Comparison for H3...... 96 A.8 Regression Results for H4...... 96 A.9 Model Comparison for H4...... 97 A.10 Regression Results for H5...... 97 A.11 Model Comparison for H5...... 98 A.12 Regression Results for H8...... 98 A.13 Model Comparison for H8...... 99 A.14 Regression Results for H9...... 99 A.15 Model Comparison for H9...... 100 A.16 Regression Results for H10...... 100 A.17 Model Comparison for H10...... 101

iii List of Figures

1 Location of the Hogaan Villages in Somaliland and Puntland...... 3

2 Example Village and Cluster Sampling Maps...... 17 3 Mapping vs EPI Sampling Eﬀect...... 20 4 Household Randomisation Eﬀect...... 22

5 Institutions Providing Social Services: Comparison Village Type...... 25 6 Participation: Comparison Youth Groups...... 26 7 Participation: Comparison Village Type (1/5)...... 28 8 Participation Customary Authority(5/5): Comparison Village Types.... 30 9 Participation: Comparison Village Type (2/5)...... 30 10 Participation: Comparison Village Type (3/5)...... 31 11 Participation: Comparison Village Type (4/5)...... 31 12 Participation: Comparison Village Type (1/4)...... 36 13 Participation: Comparison Gender (2/4)...... 37 14 Participation: Comparison Gender (3/4)...... 37 15 Participation: Comparison Youth (2/3)...... 39 16 Participation: Comparison Youth (3/3)...... 40 17 Participation: Comparison Clan (1/3)...... 41 18 Participation: Comparison Clan (2/3)...... 42 19 Participation: Comparison Clan (3/3)...... 42 20 Participation: Comparison Clan...... 43 21 Women’s Perception and Factual Involvement in Planning...... 44 22 Youth Perception and Factual Involvement in Planning...... 44 23 Minority Clan Perception and Factual Involvement in Planning...... 44 24 Young Men/Women Perception and Factual Involvement in Planning... 44 25 Minority Clan Men/Women Perception and Factual Involvement in Planning 45 26 Women’s Rights...... 45 27 Youth Rights...... 45 28 Conflict Mechanisms: Comparison Village Type...... 46 29 Leader Perception Women’s Rights and Inclusion: Comparison Village Type 47 30 Institutional Responsibility for Social Services: Comparison Village Type. 51 31 Institutions Responsible for Security: Comparison Village Type...... 51 32 Relations Village and Government: Comparison Village Type...... 53 33 Service Delivery: Comparison Village Type...... 54 34 Service Delivery: Comparison Youth Adults...... 55 35 Conflict Resolution: Comparison Village Type...... 56 36 Conflict Issues: Comparison Village Type...... 56 37 Inclusiveness of Conflict Resolution: Comparison Village Types...... 57

iv 38 Trust: Comparison Village Type...... 58 39 Citizens and Leaders Preferences in Hogaan Villages from Survey Experiment 61 40 Citizens Preferences from Survey Experiment, Comparison of Hogaan and Non-implementation Villages...... 62 41 Citizens Preferences with and without Financial Information from Survey Experiment...... 63 42 Training Effects on Participation (1/2)...... 65 43 Training Effects on Participation (2/2)...... 66 44 Training Effects on Service Delivery...... 67 45 Training Effects on Women’s Rights...... 68 46 Training Effects on Institutional Inclusion (1/2)...... 69 47 Training Effects on Institutional Inclusion (2/2)...... 70

48 Group Membership and Participation in Hogaan Programme...... 76

49 Comparison of Clan Groups on Participation in Hogaan Programme.... 86

50 Perceptions of Institutional Inclusion...... 103 51 Comparison of Non-Implementation with Levels of Hogaan Activity on Community Perceptions of Inclusion...... 104 52 Youth Perception of their own and Youth General Involement in the Plan- ning Process...... 104 53 Minority Clan Perception of their own and Minority Clan General Invole- ment in the Planning Process...... 105 54 Alternate Participation Measures: Comparison Village Type...... 105 55 Leader Perception of Youth’s Rights and Inclusion in Hogaan and Com- parison Villages...... 105 56 Percentage Having Made Contributions of Different Kinds – Further Models106 57 Included in Top Three Rights Priorities in Project and Comparison Villages – Further Models...... 107 58 Conflict Mechanisms in Project and Comparison Villages – Further Models 107 59 Comparison of Gender Groups on Participation Measures (4/4)...... 109 60 Women’s Rights in Comparison to Men’s...... 110 61 Women’s Rights in Comparison to Men’s...... 110 62 Comparison of Gender Groups on Trust Measures...... 110 63 Comparison of Gender Groups on Service Delivery Measures – Further Models111 64 Comparison of Gender Groups on Community Perceptions of Inclusion – Further Models...... 112 65 Percentage Having Made Contributions of Different Kinds – Further Models113 66 Comparison of Gender Groups on Top Three Rights Priorities – Further Models...... 114 67 Youth Rights in Comparison to Older People – Further Models...... 115 68 Minority and Outcast Clan in Comparison to Majority Clan – Further Models115 69 Comparison of Gender Groups on Conflict Mechanisms – Further Models. 116 70 Comparison of Gender Groups on Issue Resolution – Further Models... 117 71 Perceptions of Institutional Inclusion – Further Models...... 118 72 Perceptions of Institutional Inclusion, by Clan Groups...... 119 73 Comparison of clan Groups on Trust Measures...... 120 74 Participation: Comparison Clan – Further Models...... 120

v 75 Comparison of clan Groups on Service Delivery Measures – Further Models 121 76 Comparison of clan Groups on Community Perceptions of Inclusion – Fur- ther Models...... 122 77 Percentage Having Made Contributions of Different Kinds – Further Models122 78 Comparison of clan Groups on Top Three Rights Priorities – Further Models123 79 Minority and Outcast Clan in Comparison to Majority Clan – Further Models124 80 Comparison of clan Groups on Issue Resolution...... 125 81 Perceptions of Institutional Inclusion – Further Models...... 126 82 Comparison of Youth Groups on Participation Measures (1/3)...... 128 83 Comparison of Youth Groups on Trust Measures...... 128 84 Comparison of Youth Groups on Community Perceptions of Inclusion – Further Models...... 129 85 Percentage Having Made Contributions of Different Kinds – Further Models130 86 Comparison of Youth Groups on Top Three Rights Priorities – Further Models...... 131 87 Women’s Rights in Comparison to Men’s – Further Models...... 132 88 Youth Rights in Comparison to Older People – Further Models...... 132 89 Minority and Outcast Clan in Comparison to Majority Clan – Further Models133 90 Comparison of Youth Groups on Conflict Mechanisms – Further Models.. 134 91 Comparison of Youth Groups on Issue Resolution – Further Models.... 135 92 Perceptions of Institutional Inclusion – Further Models...... 136 93 Participation: Comparison Youth – Further Models...... 138 94 Comparison of Youth Groups on Service Delivery Measures – Further Models139 95 Comparison of Youth Groups on Community Perceptions of Inclusion – Further Models...... 140 96 Women’s Rights in Comparison to Men’s – Further Models...... 140 97 Youth Rights in Comparison to Older People – Further Models...... 141 98 Minority and Outcast Clan in Comparison to Majority Clan – Further Models141 99 Perceptions of Institutional Inclusion – Further Models...... 142 100 Comparison of Youth Groups on Trust Measures – Further Models..... 143 101 Percentage Having Made Contributions of Different Kinds – Further Models143 102 Comparison of Youth Groups on Top Three Rights Priorities – Further Models...... 144 103 Comparison of Youth Groups on Conflict Mechanisms – Further Models.. 145 104 Comparison of Youth Groups on Issue Resolution – Further Models.... 146 105 Comparison of Hogaan and Non-Implementation Villages on Institutions Responsible for Roads...... 148 106 Comparison of Hogaan and Non-Implementation Villages on Institutions Responsible for Resource Management...... 148 107 Comparison of Hogaan and Non-Implementation Villages on Institutions Responsible for Providing Clean Water...... 149 108 Rights to Justice: Comparison Village Type...... 149 109 Institutions and Equality Rights: Comparison Village Type...... 150 110 Institutions and Education: Comparison Village Type...... 150 111 Institutions and Health: Comparison Village Type...... 151 112 Institutions and Consultation: Comparison Village Types...... 151 113 Institutions and Free Speech: Comparison Village Type...... 152 114 Institutions and Right to Life: Comparison Village Type...... 152

vi 115 Government Service Perceptions and Social Services and International Con- tributions...... 153

116 Original Hogaan Theory of Change...... 155 117 Replacement Hogaan Theory of Change...... 156

vii Abbreviations

CAP Community Action Planning CARE Care International CBO Community Based Organisation CDD Community Driven Development CDR Community Driven Reconstruction CDD/R Community Driven Development/Reconstruction DC District Commission DFID Department for International Development DRC Danish Refugee Council DRC Democratic Republic of Congo EPI Expanded Program of Immunization of the World Health Organization FGD Focus Group Discussant GIP Governance Improvement Plans GPC Governance and Peace Building Consortium IDP Internally Displaced People IRC International Rescue Committee KII Key Informant Interviews NGO Non-Governmental Organisation OCVP Observatory of Conﬂict and Violence Prevention ODK Open Data Kit QGIS Quantum Geographic Information System RCT Randomized Control Trial ToC Theory of Change VC Village Council SWOT Strengths, Weaknesses, Opportunities and Threats

viii Acknowledgements

The production of this report would not have been possible without the continuous support of IRC and the larger GPC team, among them particularly Marie-France Guimond, Pasteur Ruberintwari, Brenda Engola, Abnet Kassa Debay, Mira Miriano and, during the pre-inception phase, Sheree Bennet. We would also like to thank Gareth Rannamets from DFID who handled implementation difficulties calmly and with the humour necessary for working in the Somali context. The Steering Committee provided valuable comments on the design and implementation of the evaluation. Our thanks go Markus Hoehne who led the research team which conducted the qualitative research in Erigabo and to the people who have implemented the survey, among them the Observatory of Conflict Resolution and Violence Prevention (OCVP) under the leadership of Abdullahi Odowa and, in Puntland, Ismael Elmi. Special thanks and admiration for the brilliant job they have done go to Said Dahir, Roble Ahmed Dahir, Ahmed Dualeh and Said Haji Nuur, who conducted the qualitative research in Puntland and Erigabo. Said Haji Nuur, Ahmed Dualeh and Aisha Ahmed Haji-Hussein have supervised the survey implementation and have equally done a brilliant job. They were supported by a dedicated team of enumerators who delivered high quality work under very difficult condi- tions. Their names in alphabetical order: Abdiqadir Abdilkahi, Abdiwahid Abshir, Abdi- rizak, Ahmed Abdirahmaan, Ahmed Ibrahim, Dalmar, Mariam Abdi Mahamed, Maryan Jaamac Xassan, Warfaa, Rahma, Hamda, Omar, Bilan, Geesood, Halima. In Durham the evaluation was greatly supported by Lorraine Holmes, Cynthia Garibay- Lopez, Raphaela Kormoll, Antoine Baudon, Fabian Stroetges, Lindsey Young and Victoria Keating. Our sincere thanks go to all of them.

ix Executive Summary

The report presents the results of an evaluation of Hogaan iyo Nabad, a Community Driven Reconstruction (CDR) programme in Somaliland and Somalia.The programme was funded by DFID and implemented by CARE, DRC and IRC, and aimed at strengthening local governance in rural and rather remote districts in Somaliland (Erigabo) and Puntland (Galkayo and Burtinle). To work towards making local government more responsive, accountable and eﬀective, the programme aimed to achieve two outcomes:

Outcome 1: Citizens’ participation in decision-making and conﬂict management is enhanced.

Outcome 2: Village-level institutions have improved their ability to plan, manage and advocate for community priorities.

Based on a mixed methods design, the evaluation assesses whether these two outcomes have been achieved and examines the mechanisms by which the programme worked. Qual- itative research was conducted to find out if basic assumptions and concepts of the programme corresponded with local meanings and practices. Findings were used to operationalise and, if necessary, refine these concepts for the evaluation and to support the interpretation of evaluation results. Quantitative tools (household and leadership survey) were used to estimate programme effects. The originally planned baseline/endline comparison was dropped in favour of a design that compares villages that have received the Hogaan programme (implementation villages) with similar villages that have not (non- implementation villages).

The main evaluation ﬁndings were:

Participation Positive Results

• An increase in participation, driven by the relationship of citizens (including women and youth) with the formal structures of governance at the village (Village Council) and district (District Council) levels.

• An increase in citizens’ perception of inclusiveness in decision-making.

• (Unintended) A decrease in the diﬀerence in participation rates between majority and minority clans (not occupational caste groups)

We did not ﬁnd evidence of any overall change in:

x • Overall responsiveness of local governance institutions • Inclusion rates with respect to gender and youth. • Empowerment of youth or women to participate in planning and decision-making.

Governance • Signiﬁcant change in citizens’ views of governance responsibilities at village and district levels. – Signiﬁcant increase in the opinion that the Village Council (VC) should protect rights and provide services. – Parallel decrease in the opinion that they should be provided by the District (DC).

Positive Results • An improvement in the relationship between village authorities and the District Council. • An increase in trust in Village and District Councils. • An improvement in the resolution of leadership conflicts. We did not find evidence of any overall change in: • Overall citizens’ satisfaction with local governance service delivery. • The balance of responsibilities between formal (VC and DC) and customary (elders and religious leaders) authorities. • Overall Perceptions of conflict resolution (despite leadership conflict effects). • Overall satisfaction with locally governing institutions (despite increased trust in VC and DC).

Mechanisms • Strong relationship between programme participation and positive view of village level governance. • Large attitudinal differences between programme participants and non-participants in Hogaan villages. • No major attitudinal differences between programme non-participants in Hogaan villages and citizens in non-implementation villages. • Membership in community groups provides a major mechanism for selection of participants in programme. • No evidence of standard spill-over effects. • Evidence of selection effects.

xi Chapter 1 Context of the Hogaan Programme and Evaluation

The chapter provides background information on Community Driven Development and Reconstruction (CDD/R) programming and evaluation, and outlines objectives and implementation of the Hogaan programme.

1.1 Community Driven Development/Reconstruction

The support for participatory processes and Community Driven Development (CDD) has rapidly gained popularity among development organisations and donors.1 CDD approaches focus on the local and aim at strengthening the ability of social groups to manage their own development and to gather their own resources for it without depending too much on formal government structures, which are often not functioning adequately or have been destroyed during violent conflicts. Since the new millennium, the focus of peace-building has gradually shifted towards state-building, and international resources were increasingly spent to strengthen legitimate forms of governance (Chandler 2010). Although labelled as state or institutional reconstruction, many of these programmes entail a significant institutional overhaul including provisions to support transparency, efficiency and accountability (Suhrke 2007: 1792). CDD approaches were adapted to this general shift towards institution and state-building and are, under the label of Community Driven Reconstruction (CDR), increasingly used as instruments to strengthen, re-establish, overhaul or build up local level governance in post-conflict settings. One of these CDR programmes was, between 2012 and 2015, implemented in Somalia, in particular in Somaliland and Puntland. The programme was evaluated by Dr Jutta Bakonyi (lead consultant), Dr Gidon Cohen (lead quantitative analysis) and Dr Pierre- Olivier Bédardfrom Durham University (UK). Dr Markus Hoehne from Leipzig University conducted parts of the qualitative research and analysis in Erigabo. Field research and especially implementation of the quantitative surveys was organised by the Observatory of Conflict and Violence Prevention (OCVP) in Hargeysa and Garowe.

1 According to Wong (2012: IV) the World Bank supports 400 CDD projects in 94 countries with more than $30 billion. Mansuri and Rao (2013: IX) claim that the World Bank has, in the last decade, spent about $85 Billion in support of local participatory development.

1 1.2 Evaluations of Community Driven Programmes

Evaluations of CDD/R approaches have not shown conclusive results. Large variations of evaluation questions, evaluation designs and methodologies have rendered a synthesis of findings difficult. A first review of CDD projects concluded that evidence of CDD results lags behind the rate and speed at which new projects are implemented or scaled- up (Mansuri and Rao 2004: 3). In the last decade, some organisations, among them DFID, IRC and the World Bank, have invested in rigorous impact evaluations in Afghanistan, the Democratic Republic of Congo (DRC), Aceh in Indonesia, Liberia and Sierra Leone. Additionally the World Bank has undertaken two reviews of evaluations of projects that were based on induced participation (Mansuri and Rao 2013)2 and of CDD programmes more generally (Wong). The findings of the evaluations and reviews are mixed, ranging from overall zero impact (DRC, Humphreys et al. 2014) to scarce or modest impact on social cohesion (Liberia, Fearon et al. 2008), weak (Aceh, Barron et al. 2009) to mixed impact on governance (Afghanistan, Beath et al. 2013; Sierra Leone, Casey et al. 2013) and either mixed or positive impacts on social welfare and on poverty reduction (Aceh, Sierra Leone) and access to basic services (Wong 2012). These results are rather disappointing, given that CDD/R programmes are ambitious and aim at initiating broad and long term attitudinal and behavioural change of citizens and local leaders, which will improve social cohesion, provide more inclusive forms of governance, and increase welfare provision (King 2013: 3f.). The evaluation findings have given rise to a number of general questions on the validity of theoretical and conceptual underpinnings of CDD/R, the change hypothesis that guides CDD/R programming, and even if CDD/R is at all able to address problems of governance and social cohesion, and if so to what extent. The evaluation of the CDR programme in Somalia will contribute to some of these questions. In addition to the disappointing results of rigorous evaluations, theoretical and conceptual concerns were raised that challenge general assumptions as well as the presumed logic of change expressed in CDD/R projects.3 Accordingly, CDD/R approaches provide the allure of optimism and purpose as they are embedded in a seductive mix of development frames such as local ownership, empowerment or participation, which evoke moral authority and display normative power (Cornwall and Brock 2005: 1045, 1043), but are conceptually too vague to guide implementations. Instead, the concepts are interpreted, (re-)defined and contested by people designing and implementing CDD/R programmes, and by those who are receiving development projects. This process often leads to new alliances between international and particular groups of local actors, and to elite capture (Chopra and Hohe 2004; Chesterman 2007). The evaluation has thus assessed the meanings and practices of community, participation and governance in Somali villages, in order to assess if the proposed logic of change holds in the Somali context and, if necessary and possible, to provide alternatives.

2 Induced participation is introduced and facilitated by a state or an external actor. Mansuri and Rao (2013) pose it in opposition to organic participation which is mobilised or initialised by the social groups independent of a government or any other external stimulus. 3 See among others Mohan and Stokke 2000; Cooke and Kothari 2001; Hickey and Mohan 2005; Reich 2006; Cornwall and Brock 2005, Cornwall 2008.

2 1.3 Hogaan iyo Nabad (Governance and Peace)

Hogaan iyo Nabad (Hogaan) is funded by DFID (£4,519,723) and implemented by a consortium of three international organisations: Care International (Care), Danish Refugee Council (DRC) and the International Rescue Committee (IRC). Programme activities in the Somali villages were carried out from September 2013 until April 2015.4 Its main aim is to support institutions of governance and governance capacity of local authorities and citizens in 60 villages in Somaliland and Puntland. Local governance is defined broadly as decision-making at village level, while the two main goals of governance are identified as service delivery and conflict management (ToC Document, no date).

50 0 50 100 150 200 km

Figure 1: Location of the Hogaan Villages in Somaliland and Puntland

4 The evaluation ﬁeldwork was conducted in November/December 2014 (qualitative); February/March 2015 (qualitative) and March/April (quantitative), thus while activities are still ongoing (cf. Caveat 1: Timing of Evaluation).

3 1.3.1 Problem Statement Hogaan is concerned with local level governance. In line with recent academic literature on hybridity and, in particular, on hybrid forms of governance (Boege et al. 2008)5, the Hogaan programme recognises the diversity of village institutions in Somaliland and Puntland, and identifies co-operation between the formal (Village Council, Village Head- man, Mayor) and traditional or customary institutions6 (elders, religious authorities) as a critical precondition to both sustainable service delivery, and peace and security in the village and beyond (ToC Document no date). The programme, however, identified two main problems in the way local level governance is structured and practiced, and stated that governance is 1) ineffective and 2) not inclusive. For the first problem (inefficiency) the ToC identifies causes, while the second is merely described.

1. Local level governance is ineﬀective due to

1.1. Informal and non-bureaucratic practices of administration 1.1. Village institutions lack legal basis 1.2. Role confusion, role duplication and unclear responsibilities between Vil- lage Council, clan elders and committee members 1.3. Limited links between village institutions, the District Council and the Mayor, which again impacts on the effectiveness of local government. 1.2. Lack of capacity of local authorities 1.1. to effectively govern and manage conflicts, which in turn hampers decision- making and service delivery 1.2. to lead development by assessing and managing communities’ needs 1.3. Lack of capacity of citizens 1.1. to demand services from the village authorities 1.2. to participate in decision-making 1.4. Lack of Resources 1.1. impedes service delivery; and 1.2. causes continued dependency on NGOs

2. Local level governance is not inclusive Minority clans, women and youth

2.1. are not represented in village institutions 2.2. do not participate in decision-making 2.3. do not have their rights protected

Textbox 1 : Problem Statement (extracted from Theory of Change)

5 Since Boege et al (2008) introduced the concept of hybridity in their analysis of fragile or weak states, it has gained popularity and is used to describe a broad spectrum of political orders. 6 The report prefers the term “customary” as it indicates that these rules and institutions are based on everyday interactions and change with them, and often (but not necessarily) lack formal, bureaucratic structures and legal anchorage.

4 1.3.2 Theory of Change Based on this problem statement, the programme has developed a Theory of Change (ToC) which is centred around the core hypothesis that increased and more inclusive citizen participation in decision-making and conflict management, and enhanced service delivery by locally governing bodies, will lead to more effective, responsive and accountable governance. The ToC outlines expectation of how a sequence of activities and outputs will lead to two intended outcomes. It addresses questions of what, how and who, and thus clarifies the main objectives (what), mechanisms (how) and addressees (who) of the programme.

Main Objectives of Change (What): The ToC outlines two main outcomes of the programme:

Outcome 1: Citizens’ participation in decision-making and conﬂict management is enhanced.

Outcome 2: Village-level institutions have improved their ability to plan, manage and advocate for community priorities.

Under Outcome 1 a particular focus is given to inclusiveness and thus to the participation of groups that are traditionally excluded from politics (see Text Box 1: Problem 2), such as minorities, women and youth. Increased participation is also reﬂected in the accessibility of the governing bodies by the citizens (albeit that this goal is speciﬁed in the ToC under Outcome 2).

Under Outcome 2 Hogaan speciﬁes that locally governing institutions will improve their capacity in:

• Co-ordination and administration;

• Service delivery;

• Conﬂict management.

Local institutions have further increased their:

• Accessibility for citizens (responsiveness)7

The programme additionally intended to strengthen the links between village and higher level governance at the district and regional level. While this goal was not separated in the ToC, it was explicated in the problem statement.8

Main Mechanisms of Change (How): The ToC additionally outlines the mechanisms used to address the identiﬁed problems. Mechanisms thus deﬁne the main assumptions of how the intended change will be brought about. The programme followed a fairly

7 Albeit under Outcome 2 in the ToC, the evaluation ﬁndings on governance accessibility are outlined in section 4.1 8 Although this goal was not included in the ToC, the evaluation has evaluated local authorities’ and citizens’ views on the relation between village and district.

5 standardized CDD/R process relying on the three core mechanisms commonly applied in community-driven programming: 1. community entry and Community Action Planning (CAP); 2. the delivery of block grants; and 3. capacity building and training. Community entry and Community Action Planning (CAP) include a series of initial village meetings (community entry) where the problems are explained and discussed, and ways to solve them are described. Villagers then select a committee (around 30 people) whose members co-operate with the Village Council to identify and to prioritise main needs. However, in villages in which the VC was either dormant or not functional at all, Hogaan encouraged the (re)building of formal institutional structures and advo- cated for the inclusion of women in the council. The results of the needs prioritisation is then again presented to a village meeting, to be discussed, challenged and revised until a (majority) decision is reached. The lack of resources (Problem 1.4 in Textbox 1) is addressed by the provision of block grants.9 The international organisation provides two consecutive block grants (each $17,000) for the realisation of development projects chosen by the villagers. A minimum of 20% of each grant is provided by the villagers themselves in order to ensure seriousness of needs identification and communal ownership of the project. The lack of capacity of both villagers and local leaders (Problem 1.2 and 1.3 in Textbox 1) is addressed by capacity building and training. This provides the core mechanism for stimulating social change and accompanies every step of the CDR processes on every impact level. In order to stimulate citizens’ participation in decision- making, the programme for example provides training and capacity building, organises dialogue forums and provides further support for citizens to play a more active role in decision-making, to participate in collaborative discussions and to co-ordinate priorities with government stakeholders. In particular, selected citizens10 and local government officials receive training in civic education, advocacy, conflict resolution and general support for community dialogue and peace-building.11 In order to facilitate planning, management and community orientation at the village level, training and support is provided for members of the Village Council and the development committee in the provision of transparent, accountable and accessible services to communities.12 However, besides training in specific topics and for selected groups of citizens or members of locally governing bodies, capacity building included regular consultations with and advice from members of the international organisation. Another crucial part of CDD/Rs capacity-building was based on a learning-by-doing approach, for example during community action planning and needs prioritisation, through the management of the block grant and the implementation of the projects. The assumption was that community members who have gained planning, management and implementation practice

9 Block grants are also important part of organisation development and ’learning by doing’, as outlined in the next section. 10 The programme documents do not outline the criteria for the selection of training participants. 11 The Theory of Change outlines precisely: IF Community members are provided with training and supported to play a more active role in public decision-making, co-ordinate priorities with government stakeholders and participate in collaborative discussions about local development, AND selected community members and local government officials in GPC target areas are trained in civic education, advocacy or conflict resolution in support of community dialogue and peace-building efforts, THEN citizens participate in decision-making and conflict mitigation. 12 Again the ToC states IF village and district council members in GPC target areas are trained and supported to provide transparent, accountable and accessible services to communities THEN village level institutions are able to plan, manage and advocate for community priorities.

6 will continue to manage projects in a similar manner, but without being dependent on either the central government or an international organisation. Main addressees (Who): The ToC refers to community members and local government officials, the latter specified as selected village and district council members (ToC Document, no date). The general claim that CDR programmes challenge existing structures of authority by putting power and resources in the hands of community members (Fearon et al. 2008: 1) does thus not apply to the Hogaan programme. Hogaan was implemented in districts where the control of the state over local resources was either weak or completely absent, and where governance was radically localised (organised by citizens), thus resulting in the complex mix of self-governing institutions that today characterise most parts of Somalia and at least remoter places in Somaliland.13 Hogaan did thus not so much challenge local structures as work through them and with them, in an attempt to formalise and if necessary build-up these structures, and to link them more closely to the district government and hence the state. Its second focus was on the intersection of government and society. Like other CDR programmes, Hogaan aimed at empowering citizens in general and disadvantaged groups in particular, by providing avenues to increase interaction between citizens and locally governing institutions and officials, by including villagers in decision-making processes and by empowering them to articulate demands.

13 Literature on the organisation of governance in Somalia/Somaliland includes Menkhaus, Bakonyi.

7 Chapter 2 Evaluation Design and Implementation

This chapter sets out the main evaluation design, and outlines the evaluation’s main objectives and methods used to achieve its objectives (section 2.1). It also explains why and how the evaluation diverted from an originally proposed design and analysis plan, and discusses challenges of this diversion and how they were mitigated. The sampling methods and details about the two surveys conducted are outlined in (section 2.2), followed by information on the implementation of both the qualitative and quantitative research component (section 2.3). Given that little solid information is available on the social composition and structure of Somali villages, a proper randomisation of the household sample proved challenging. The evaluation therefore conducted two sampling experiments, aimed at shedding light on which techniques allow for a proper randomisation of the sample, especially in contexts where little general information is available. The chapter ends with the presentation and discussion of the results of these experiments (section 2.4).

2.1 Research Design

The evaluation uses a mixed-method design to contribute to three core objectives (O).

O1: To unpack some of the concepts and assumptions underlying community-based approaches and change hypotheses.

O2: To measure if the programme has led to the intended outcomes that is if it has increased citizens’ participation in decision-making and conﬂict management, and strengthened the ability of village-level institutions to plan, manage and advocate for community priorities.

Based on the ﬁndings of O1 and O2:

O3: To provide conjectures about causal mechanisms that may lead to social change and to develop an alternative theory of change.

Under O1, as a point of departure for the evaluation, and with the aim to operationalize and if necessary reﬁne the concepts that underlie the programme’s Theory of Change, rapid ethnographic assessments based on semi-standardised interviews and focus group discussions were conducted, in three implementation villages and one comparison (non- implementation) village per district. The qualitative research component focused on social relations, practices and meanings that structure concepts of community, participation and

8 community contribution; it additionally examined membership and relations within and between formal and customary governing institutions at the village level. In order to address O2, two surveys were conducted. They were guided by the question of whether the programme caused the changes anticipated. Villages in which Hogaan was implemented (implementation villages) were compared to villages where the programme was not implemented (non-implementation villages).This allows the evaluation to assess what would have happened if the programme had not taken place and thus the change induced by the Hogaan programme.1 The non-implementation villages were not randomly selected before the programme was implemented (as would be the case with a Randomized Controlled Trial). Instead they were selected in the course of the evaluation after the programme was already implemented. Regression methods were used to control for systematic diﬀerences between the implementation (Hogaan) and non-implementation villages.

Caveat 1: Timing of the Evaluation Ideally an endline survey should be implemented a while after a programme is finalised. The final evaluation of the Hogaan programme was, however, conducted while programme activities were ongoing. This has serious implications, as the findings do not determine sustainability and long term change, but rather measure effects that are directly produced by the programme. It is thus not possible to assess if the measured perception and governance effects are routinized and therefore likely to last.

Caveat 2: Change in the Initial Evaluation Strategy The research design outlined above represents a change from the original design that was set out in the evaluation proposal and subsequently outlined in the data analysis plan. Following the evaluation Terms of Reference (TOR), a pre-post evaluation design was designed to achieve O2. Using a baseline survey the programme had conducted, the original design aimed to look at changes that had taken place in the course of delivering the programme. The evaluation, however, recognised that pre-post evaluations provide a less robust measure than alternatives, particularly Randomized Control Trials (RCTs). However, a feasibility study ruled out the possibility of an RCT for the Hogaan programme (Grant et al. 2013), hence a pre-post evaluation design was requested whilst recommending that results were interpreted with caution. The pre-post research design made comparability with the baseline survey an overrid- ing necessity, and considerable eﬀort was put into making the endline data as comparable as possible with the baseline data (see for example Section 2.2 below on sampling). For this reason the main body of the endline survey followed the baseline survey as closely as possible. Additional modules were added to test some of the ﬁndings of the qualitative research, and to assess the impact of programme activities. A leadership survey was designed to enable the testing of the aspects of the theory of change relating to leadership. A small-scale comparison with six non-implementation villages was included in the initial design as a ‘sanity check’. Despite this, the evaluation presented here depends almost entirely on comparison with non-implementation villages. A comparison with the baseline data had to be dropped when the baseline data was shown to be at least partly unreliable. Large data entry errors were discovered only relatively late, after the endline data was already collected and

1 Non-implementation villages are thus used as the basis for a model of a counterfactual.

9 an initial comparison between baseline and endline data identified implausible patterns. After further investigation, implausible data was also found in the paper copies of the baseline survey. These problems came on the back of a number of additional difficulties with the interpretation of the baseline. Amongst these were difficulties caused by the change of evaluator teams from the baseline and endline evaluation. Although the baseline evaluation set out criteria for an endline assessment, the assumptions or hypotheses that underlined some of the baseline survey questions were not completely clear to the endline evaluation team. Additionally, parts of the baseline were conducted while the programme, and in particular community action planning (CAP), was already ongoing. Instead of providing baseline data, that is information on the situation in the villages before the programme was implemented, the baseline already measured implementation effects. At the same time as the baseline data proved particularly problematic, the identification of non-implementation villages proved successful. In line with the growing realisation of baseline data errors, the decision was taken to extend the survey to further non-implementation villages and to make the comparison between implementation and non-implementation villages the basis of the evaluation.

Non-implementation Villages One reason for the rejection of an RCT for the evaluation of the Hogaan programme was the difficulty of identifying control villages (Grant et al. 2013). Very little systematic information on Somali villages is available, including even basic information on the number of villages in the different districts or the exact district boundaries, let alone other characteristics such as approximate village size, security situation and accessibility (with respect to conflict), social structure and composition of villages etc. Compared to Eri- gabo in Somaliland, the districts in Puntland are quite small, and district borders were redesigned in both Somaliland and Puntland while the Hogaan programme was implemented. For some villages it was thus not fully clear to which district they belonged; others changed their status to district capitals and thus experienced fundamental changes in the course of the project period, with the central appointment of the District Commis- sioner and the reshuffling of the Village to a District Council. The selection of comparison villages (non-implementation villages) thus faces some challenges, in particular unclear district boundaries and the possibility of there being no untreated village of the minimum size available as comparison. The turbulent political developments and internationally sponsored attempts to re-build state structures in both Somaliland and Puntland makes estimates of treatment effects rather noisy. The evaluation findings are therefore not as robust as a large scale RCT would be. Nevertheless, the findings of the evaluations do provide evidence of impact. In the absence of any systematic village information, advice from local experts and residents on village composition and selection was taken, and the survey extended to eight further non-implementation villages in Galkayo (North) and Erigabo. Due to heavy rains, Burtinle could not be included in the extension, as most villages were not accessible. However remaining villages in Burtinle did anyway not match the selection criteria, especially the criteria of minimum size. There are development programmes on-going in most villages in Puntland and Somaliland (including non-Hogaan programmes in many of the Hogaan villages).2 The following Table 2.1 compares both village types with respect to

2 For example, in Erigabo there were CARE programmes on-going in ﬁve of the non-implementation villages. The GPC team suggested there was particularly pronounced possibility of programme spillover

10 years of residence of survey respondents, livestock ownership, clan diversity, average HH size and poverty levels, to determine comparability. Urban areas were separated because they differ in many significant measures. The comparison shows that the only significant difference between implementation (Hoogaan) and non-implementation villages is the clan composition. Implementation villages were more clan homogeneous than villages participating in the Hogaan programme.

hogaan hogaan non- sig. diﬀ. villages urban village implementation (hogaan v. non-imp) av. residence (yrs) 16.44 13.76 13.01 none hh with livestock % 0.29 0.61 0.62 none clan diversity 1.32 0.97 0.60 ∗∗∗ av. hh size (people) 7.69 6.11 6.27 none extreme poor % 0.47 0.48 0.49 none permanent roof % 0.90 0.68 0.69 none

Table 2.1: Comparison of Implementation and Non-implementation Villages

The three following tables 2.2 to 2.4 provide further details on the villages in each district. The villages in which surveys were conducted are listed alphabetically, but ﬁrst the implementation (Hogaan) villages and then non-implementation (non-imp) villages (Type). The table also indicates the number (N) of Households (HH) and leadership (Lead) ques- tionnaires collected in each village and gives an estimate of household (HH) size of each village.3

on the question of VC meeting attendance in Erigabo as other CARE programmes had taken to working through VC rather than programme speciﬁc committees as a result of the perceived sucess of Hogaan programme. However, the data does not suggest that the CARE processes in Erigabo led to a substantial increase in participation in VC meetings. On the contrary higher VC participation was found in the non-implementation villages in Galkayo: Erigabo: Hogaan 79%/Non-impl 79% Galkayo: Hogaan 78%/Non-impl 83% 3 Household size in the table is based on estimates by the supervisors of the survey teams. Table 5.1 in Chapter5 provides another table in which household estimates based on programme documents, on the satellite images and on supervisors are compared. We use here and in the following the supervisors’ estimates, not because we think they are more accurate, but because they provide estimates for all villages.

11 Village Type N (HH) N (lead) HH Size Bahaley Hogaan 22 5 40 Ballidacar Hogaan 27 5 55 Birrecaad Hogaan 11 5 30 Birta Dheer Hogaan 24 4 50 B. Town (Hawlwadag) Hogaan 29 5 530 B. Town (Horumar) Hogaan 29 5 550 B. Town (Israac) Hogaan 28 5 530 B. Town (Wadajir) Hogaan 27 5 580 Faratooyo Hogaan 24 5 60 Farjano Hogaan 25 5 33 Godobyar Hogaan 25 5 62 Hayanle Hogaan 15 5 45 Jalam Hogaan 29 5 5000 Kalabayr Hogaan 21 6 900 Koryal Hogaan 17 5 40 Lacle Hogaan 10 5 25 Maga’ley Hogaan 29 5 70 Megag Hogaan 29 5 54 Meygagle Hogaan 28 5 60 Xaarxaar Hogaan 20 5 50 Awrculus non-imp 26 5 40 Meeraysane non-imp 5 10

Table 2.2: Villages in (Matched) District Burtinle

12 Village Type N (HH) N (lead) HH Size Adaygebagebo Hogaan 2 3 3 Agaran Hogaan 29 5 200 Badweyn Hogaan 25 5 5000 Balanbal Hogaan 4 75 Balisbule Hogaan 27 5 500 Bayra Hogaan 29 5 300 Beer dhagaxtuur Hogaan 23 4 50 Bilcil Hogaan 19 5 20 Bursaalax Hogaan 26 5 1500 Darusalam Hogaan 21 5 25 Dhagaxyo Cado Hogaan 26 4 30 G. City (Hormar4) Hogaan 4 1000 G. City (Israc) Hogaan 26 5 800 G. City (X/Garsoor) Hogaan 31 8 1000 Harfo Hogaan 32 5 3000 Malaasle Hogaan 26 5 60 Qalanqal Hogaan 27 6 45 Roox Hogaan 26 5 50 Sallah Hogaan 30 5 320 Dhobocantug non-imp 19 5 100 Elbardale non-imp 12 4 20 Gosol non-imp 24 5 45 Labilamane non-imp 18 5 120 Mayle non-imp 18 5 200 Shakaal non-imp 17 5 50

Table 2.3: Villages in (Matched) District Galkayo

13 Village Type N (HH) N (lead) HH Size Ardaa Hogaan 25 4 35 Biyoguduud Hogaan 17 5 30 Buq Hogaan 25 5 30 Carmala Hogaan 27 5 450 Damal Hagare Hogaan 30 5 200 Daryle Hogaan 8 3 50 Dibqarax Hogaan 7 2 24 E. Town (Daya’an) Hogaan 30 5 150 E. Town (Hafad Somale) Hogaan 33 5 150 E. Town (Shacab) Hogaan 27 4 200 Godcaanood Hogaan 26 5 120 Godmobias Hogaan 30 3 400 Goof Hogaan 24 4 50 Jidali Hogaan 23 5 40 Mait Hogaan 30 4 400 Rugey Hogaan 20 5 50 Shimbirale Hogaan 31 5 100 Sibaayo Hogaan 10 5 25 Yube Hogaan 23 5 500 Yuﬂe Hogaan 36 4 150 Darasalam non-imp 16 5 200 Fadhigaab non-imp 18 4 250 Gar-Adag non-imp 18 5 2000 Midhisho non-imp 22 4 25 Xiingalool non-imp 18 5 3000 Xiis non-imp 2 4 10

Table 2.4: Villages in (Matched) District Erigabo

14 Impact of Design Change and Mitigation The necessary change of the evaluation approach has had a detrimental effect, although every effort has been made to minimize it. Due to the late change in the evaluation character, the data analysis plan had to be discarded. Additionally, the evaluation team would have chosen a larger size of the non-implementation village sample had this research strategy been designed from the outset. This is particularly a problem for the study of anything other than very large interaction effects. At the same time, the change in evaluation design had only limited effects on the overall quality of the evaluation. The dropping of the data analysis plan carries the risk that results are cherry-picked. In order to address this issue we use a standard model for all our analyses, and systematically report all our findings relating to a standard set of dependent variables. The overall pattern of results is therefore fully transparent in the report. With respect to the robustness of the conclusions, both pre-post and non-random non-implementation villages are less robust than an RCT for making causal inferences. Non-implementation village comparisons however may even provide a better basis for a comparison in situations of rapidly changing contexts. If the selection of reasonable non- implementation villages had been deemed possible at the outset of the evaluation, this is very likely the approach which would have been chosen. Therefore, the unavoidable late change in the analytical strategy may well have had overall positive consequences for the evaluation.

2.2 Sampling Techniques

2.2.1 Qualitative Sampling As the qualitative study followed the aim to capture a broad range of citizens’ views on general concepts of community, participation and governance, selection of villages was based on maximizing variation of the following criteria: clan composition (homogene- ity/heterogeneity; minority clan representation); size of the village (fewer/larger number of households); accessibility (distance from urban centres and type of road) and project composition (type and contribution). In Erigabo, the criterion of clan affiliation gained importance to ensure that the views of the four major clans in the district (Isaaq/Habar Yonis and Isaaq/Habar Jeclo, Darood/Dulbahante and Darood/Warsangeli) are captured. The selection of villages was constrained by security restrictions for international staff. While all villages were accessible in Burtinle district, the requirement to take a hotel in Garowe and to travel forth and back every day, made the selection of remote villages impossible. Villages in Galkayco were not accessible at all for international staff and, due to ongoing conflicts, two villages were also off limits for national staff. Two national researchers were trained in Burtinle and collected the data in Galkayo. In Erigabo, only 13 villages (including the 3 urban quarters in Ergibao) could be selected due to Care International’s security protocol for international staff.

2.2.2 Survey Sampling The evaluation aimed to survey a representative sample of the village population, which would also be directly comparable to the results from the baseline survey.

15 Insufficient details on the precise conduct of the baseline led to the decision to develop two survey experiments that would enable the evaluation to generate comparable data, but simultaneously to assess the extent to which the survey results were influenced by sampling methods. The first experiment is related to the selection of households, the second to the within household selection of the interviewee.

Household Selection Experiment In the design phase, a variation on EPI sampling methods were proposed. EPI sampling4 is a standard method used to find random samples when there is no sampling frame available. EPI and other random-walk methods work by selecting a random initial household, and then randomly selecting further households from that starting point until a fixed number of households have been interviewed. The specific proposal was that an initial household was selected by taking a starting point near to the village centre then selecting a random direction by spinning a pen to indicate a direction and finding the closest household to the point in that direction half way to the village periphery. Subsequent households would then be selected by choosing the closest house in a random direction from the current location (again using pen spinning to determine direction). Reasonable concerns were raised in the design phase about the proposed EPI sampling techniques (almost certainly used in the baseline) as it is well known that such methods under-represent the periphery of villages, where it was believed that minority and poorer population may more frequently live. In order to address this, the evaluation randomly selected half of the villages and undertook sampling in those places using satellite images. In order to ensure comparability with the baseline, 25 implementation villages (50% of the sample of implementation villages) were randomly selected for the satellite image approach while the remaining 25 used EPI-like methods. In the villages randomly selected for satellite image sampling, a map of the village was created using QGIS and satellite images primarily from Bing Areal. Conditional on size, the villages were divided into geographical clusters of approximately equal size and about twenty households (using the tcluster package in R). Within clusters, households were randomly sampled for the interviews. In all, 550 A3 maps were created for the enumerators. These maps, as illustrated by the examples in Figure2, show both the location of the cluster within the village (on one side of the A3 map), and the precise location of the randomly selected households (on the other side of the map). Each enumerator is allocated a cluster within the village, and they were told to visit the first six locations in the cluster (according to the numbers of the map). A protocol for attempting to locate the inhabitants of empty houses was given, but in case that this did not yield an interview, or that the building was not an inhabited location, enumerators were told to replace with other locations in the cluster in strict numerical order. The number of interviews conducted in each village is shown in tables 2.2-2.4.

4 EPI Sampling was ﬁrst developed by the World Health Organisation for its Expanded Programme on Immunization, thus EPI.

16 Yube cluster A Yube cluster A

13 19 7

15 9

14 18 10 5

8 17 1 4

20 12 11 21

3 16

2 22

0 0.1 0.2 km 0 0.02 0.04 km

(a) Location of Cluster in Village (b) Location of HH in Cluster Figure 2: Example Village and Cluster Sampling Maps

Within Household Selection Experiment There were concerns about the practice of interviewing the person who answered the door in each household, but these were mitigated by the development of a household roster which allowed the randomisation of the interviewee. To ensure comparability with the baseline whilst still addressing concerns about representativeness, the door openers were sampled in some households and individuals randomly selected from the household roster in other households. The results of both sampling experiments, that is mapping versus EPI sampling, and randomised selection from household roster versus questioning the door opener, are presented in Section 2.5

2.3 The Surveys

As the addressees of the programme were divided into citizens and local leaders, and training and capacity-building was provided to both groups, the evaluation designed two surveys: an endline household survey and a local leadership survey. This enables the evaluation to differentiate programme effects on citizens from effects on local leaders.

2.3.1 The Household Survey The endline household survey was designed to replicate the questions posed in the baseline survey. However, two modules were added. The ToC identified training and capacity building as core mechanism for initiating change. One module was added to evaluate participation of respondents in the implementation villages asking how they perceived the quality of training and capacity-building activities and, more generally, the implementation of the programme. This module is mainly used to draw inferences of direct programme effects, as people who have, for example, participated in training are supposed to answer some questions differently to people who have not received the same training. As a result of the findings from the qualitative research, a module on community contributions was added to the survey. Moral obligations to contributions, and thus

17 mutual dependencies of villagers, were emphasized as a constitutive element of communities and as a factor that facilitates social cohesion. The module on contributions aims at a further understanding of contribution practices, and is used to evaluate the impact of (different types of) contributions on village level planning and governance. Further questions on clan affiliation, and membership in local organisations (such as youth groups, women’s groups, NGOs) were added to the survey to capture if and how clan affiliation and membership in organisations influences perceptions and practices of communal relations, village level governance and community-based development, and if certain affiliations provide or block opportunities for participation. In order to further examine the impact of the programme, a discrete choice survey experiment was added. The GPC programme seeks to raise awareness on inclusive participatory approaches towards governance, to foster inclusive decision-making and to clarify roles of the different governing bodies at village level. However, international projects also provide substantial financial assistance. One long-standing concern is that the entire structure of responses is driven by anticipated future support. Villagers may well give answers that they believe international actors want to hear, answers that hence reflect the preference for the continued international assistance. Villagers may also hide what they believe are internationally undesirable responses. The endline household survey was composed of ten modules: background, personal and household details; Hogaan engagement; survey experiment; community contributions; rights and rights prioritisation; institutional roles and responsibilities; participation and interaction with leaders; conflict resolution; governance, service provision and identities. The implementation of the survey took an average of 30 minutes.

2.3.2 Leadership Survey The leadership survey consisted of nine modules: background and personal details; leadership roles; conﬂict resolutions; right and rights prioritisation; Hogaan training; institutional roles and rights; governance; survey experiment and identities. The questions were designed to enable a comparison between implementation and non-implementation villages, to understand the impacts of speciﬁc forms of participation in the Hogaan programme on attitudes towards leadership and to compare leadership and citizens attitudes on governance. The implementation of the survey took approximately 15 minutes.

2.4 Evaluation Implementation

The qualitative research was conducted from 16 November to 6 December 2014 in Puntland and from 18 February to 2 March 2015 in Somaliland. The second team build on lessons learnt from the ﬁrst sequence. In total, 79 Key Informant Interviews (KII) and 47 Focus Group Discussions (FGD) were conducted.5 Findings from this research stream fed into Objective 2. They provided a starting point for descriptive inference and the examination of how the GPC programme has aﬀected community relations and power dynamics at the village level, and if change initiated by the programme followed the pathway outlined in the programme’s Theory of Change (cf. Alternative ToC, Chapter 5).

5 Divided among the districts, 29 KII and 15 FGD were conducted in N/Galkayo; 21 KII and 17 FGD in Burtinle; and 29 KII and 15 FGD in Erigabo.

18 The quantitative research started on 16 March 2015 with training (3 days) for enumerators in Hargeysa. Three teams, each comprising five enumerators and headed by one supervisor, implemented the survey in the three districts. Data collection took between 28 (Erigabo) and 30 days (Burtinle and N/Galkayo). Mobile data collection tools on tablets and based on a freeware Open Data Kit (ODK) platform were used. Data was transmitted from the field back to the research team via the internet using a secure ODK Aggregate server by the supervisor, whenever the field team had access to the internet. The household and the leadership survey were implemented in 57 of 60 programme villages and in 13 non-implementation villages. Three implementation villages were not accessible for the team, one of them because local authorities insisted that interviews are conducted in their presence and additionally refused the utilisation of tablets6, the two others because of internal conflicts. In all, the evaluators collected 1604 household and 347 leadership interviews in 70 villages.

2.5 Sample: Description and Results of Experiments

2.5.1 Satellite Image versus EPI sampling As outlined in Section 2.2.2 the evaluation experimented with the utilisation of satellite images and aimed at testing if they provide a feasible alternative to standard EPI sampling techniques. 50% of implementation villages were randomly selected for the satellite image approach. However, as the mapping technique used was relative time consuming, the evaluation decided to exclude mapping of urban areas. Additionally the evaluation was unable to identify the exact location and thus to map one of the randomly selected villages. In total, as part of experiment the evaluation mapped 24 villages and used standard EPI sampling technique in 26 (excluding non-implementation villages and urban areas).

2.5.2 Experiment Result The evaluation did not find evidence that the utilisation of satellite images changed the representation of the village periphery and thus of minority households or poorer popula- tions. There are no significant differences with respect to mapped and unmapped villages on any of the following measures:

1. Household location with respect to village periphery (as classiﬁed by the enumerator)

2. Indicators that are likely to be connected to location such as

• size of house • roof type of house • wealth

However, the evaluation found provisional evidence that sampling by maps led to an increase in sampling of members with minority clan (note: not caste!) background. In

6 While the team carried a paper version of the survey and could have diverted to it, the rejection of anonymity was not acceptable.

19 mapped villages, a significantly larger number of householders with minority clan background were sampled (** with clustered standard errors).7 The result, however, should be considered provisional because (a) the main mechanism by which we expected this result, that is by including the periphery, does not appear to have been a factor; (b) the experiment was conducted on a fairly small scale (50 villages); and (c) alternative mechanisms or the robustness of the result against alternative specifications of the minority/majority clan specifications should be checked before final conclusions are drawn.

periphery minority.clan year.here 15 0.3 0.4

0.3 0.2 10

0.2 0.1 5 0.1

0.0 0.0 0 livestock num.people.hh extreme.poor 0.5 0.6 6 0.4

0.4 4 0.3 mapping (intended)

value 0.2 random walk (intended) 0.2 2 0.1

0.0 0 0.0 hh.permanent clan.diversity 1.00 0.6 0.75

0.4 0.50

0.2 0.25

0.0 0.00

Figure 3: Mapping vs EPI Sampling Eﬀect

The evaluation team would point to two auxiliary advantages and one disadvantage of producing and using maps:

Advantage

7 This result also holds under an intent to treat analysis, in which the evaluation compares those villages that were originally intended to be mapped, including those which were not actually mapped, with those where there was no intention to map.

20 • Mapping and maps helped to understand the size, structure and potential household composition (size and wealth of households, shops) of the villages. It also raised concerns about the size of some villages, which were signiﬁcantly under the minimum size for projects identiﬁed by Hogaan (70 households).

• Maps served as an important tool in the training of enumerators. First, enumerators started to think about village structure themselves. While conducting the training in the outskirts of Hargeysa, for example, enumerators did exclude IDP huts with the argument that IDPs are not part of the community. The maps helped to explain that everybody living in the village, independent of status, wealth, duration of residence etc. qualiﬁes as interview partner.

Disadvantage

• The production of maps is time consuming and expensive. The latter is especially relevant if contemporary maps need to be produced. Publicly available images are often a number of years out of date whilst access to real-time or contemporary images is relatively expensive.

2.5.3 Within Household Randomisation Figure4 shows the results of the within household randomisation procedure. Each graph relates to a diﬀerent demographic characteristic. The ﬁrst bar shows the proportion with that characteristic amongst door-openers; the second bar shows the proportion when using within household randomization. The third bar refers to the proportion of the characteristic within the whole population (on basis of the household roster). For a representative sample of the population the samples should approximate to the population proportions (thus should approximate the third graph).

21 female child.over.18 no.education

60%

40%

20%

0% dooropener fiftyfive.plus householder spouse random value household roster

60%

40%

20%

Figure 4: Household Randomisation Eﬀect

The results show that sampling door-openers will lead to an oversampling of:

• women (∗∗∗)

• older ages groups (∗∗∗)

• those with no formal education (∗) and an undersampling of:

• grown-up children of the head of household (∗∗∗)

Figure4 shows that within household randomization obtains results which are generally much closer to the patterns found in the household roster. Nevertheless, even the random selection of household members does not fully represent the household roster itself. This is not because our random sample was ‘unlucky’, but a result of the multi-stage sampling procedures that are almost always used in survey research. The key stage here is the selection of households followed by the selection of one individual within that household. In cases where characteristics are correlated with household size relative to the population, this approach produces ‘unbalanced’ results, as can most obviously be demonstrated in the case of a high number of adult children. If we consider household size simply in terms of the number of adults eligible for interviews, very large households tend to be those with large number of children living there. Hence, grown-up children have a systematically lower

22 probability of being selected for interviews even if their household has been selected. In Puntland and Somaliland a large number of demographic characteristics appear to be correlated with household size, and hence even within household randomization will produce samples which are not representative of the population as a whole. These results show that the bias that results from interviewing only door-openers is likely to be substantial and needs to be systematically addressed. The use of within household randomization is able to correct many but not all of these biases. The data collected from the experiment will enable us to examine the extent to which model-based corrections can be used to rebalance the data. However this analysis has not yet been conducted.

2.6 Statistical Methods

The main body of the evaluation compares response patterns in implementation villages with responses in non-implementation villages, in order to establish eﬀects of the Hogaan programme on participation practices, inclusiveness of decision-making and governance performance. The Hogaan programme has a multi-level structure. The programme aims to have an impact on individuals who live in villages that form districts. When we interview citizens across the 60 implementation villages, we gather information from more than 1,000 independent individuals, but also from 60 clusters of individuals living in the same village, which are again grouped together into three more or less closely related groups (districts) and these again attached to two (sub-)states. The statistical analysis of this kind of data requires either clustered standard errors or multi-level modelling. The evaluation chose to use multi-level modelling because it provides additional estimates of substantive parameters, here in particular the modelling of village and district eﬀects. In order to keep the analysis transparent, a common set of independent variables is used in all the models. These independent variables are:

Individual level: age group [18-24, 25-34, 35-54, 55+], gender [male, female], clan [majority clan, minority clan, occupational caste]

Village level: type [city/town, village], clan diversity (Shannon diversity measure of sub-clans in the village), comparison status [implementation, non-implementation village]

The analysis on which the evaluation rests uses on multi-levels models. To ease interpretation of the results, the findings are presented in a simplified form. Rather than presenting regression coefficients, the results of the multi-level models are presented as effect sizes that is the expected change in the dependent variable due to a change in the category of one particular independent variable (usually the difference between implementation and non-implementation villages). The effect size errors are estimated across 1,000 draws of the model parameters. The effect size estimates appear in the subsequent chapters as simple percentages.

23 Chapter 3 Evaluation Findings

The presentation of findings in the report is structured by the question of whether the Hogaan programme has reached its main objectives from the viewpoint of the citizens. The answers of respondents in Hogaan villages are compared with answers of respondents from non-implementation villages in order to determine if Hogaan villages are more effectively governed and if governance is more inclusive than in villages which have not received the programme. Those findings of the qualitative evaluation component that provide additional background information against which the survey results can be interpreted, will be introduced in the findings. Others will be presented in the final chapter as they were used to develop recommendations for further programming (see chapter5) in conjunction with survey results.1 The chapters are structured in accordance with the main outcomes identified in the ToC. The first section (3.1) addresses participation and assesses whether Hogaan has enhanced villagers’ participation in decision-making, and if the governing bodies have indeed become more accessible for villagers and increased their responsiveness towards citizens’ needs. The findings on Hogaan’s impact on the capacity of village-level institutions are presented in the second section (3.2). In particular, it addresses the question of whether capacities to govern have improved with respect to Co-ordination and Adminis- tration (3.2.1); Service Delivery (3.2.2); and Conflict Management (3.2.3). It also examines whether Hogaan has more generally impacted on citizens’ trust in local governance (3.2.4). The results of a survey experiment that aimed to determine villagers’ governance preferences with respect to development (3.3) and the effects of trainings (3.4) are presented in the final section. While each section ends with a brief conclusion, a summary of all findings will be presented in the next chapter (4).

Presentation of Results: How to read the graphs The results of the survey are shown in a series of parallel graphs. The graphs also show if there are statistically significant differences in the answers to questions, for example if there are differences between implementation and non-implementation villages. The graphs show effect sizes of regressions, but can be read as showing percentages.

1 Qualitative ﬁndings were partly used to operationalise survey questions. However, the survey design was in this respect quite restricted by the requirement to conduct a pre-post comparison. It could thus only expand the survey but not signiﬁcantly alter existing modules.

24 A very quick overview Each set of graphs answers a question. Usually the question is something like: Did the Hogaan Programme Make a Diﬀerence? The answers are colour-coded in the right-hand graph: • Green: Hogaan led to a measurement increase • Grey: No evidence Hogaan made a diﬀerence • Red: Hogaan led to a measurement decrease Just looking at the colours means that the headline message from the graph can be seen very quickly.

An only slightly less quick overview The numbers on the graphs can also be read. When there are two graphs, they are normally comparing Hogaan and non-implementation responses to a question like ‘should the VC provide social services?’ The numbers on the right hand graph indicate how big the differences were (for example 14% more people think so in Hogaan villages), and the numbers on the left hand graph give the baseline percentages (for example 44% in Hogaan against 30% in non-implmentation villages). When there are three graphs this address a question like has the gap between youth and older people in participation changed with the programme. The right hand graph shows difference in the gap between youth and older people in Hogaan and non-implementation villages (eg the the gap between youth and older people participation is 13% more in non- implementation than Hogaan villages). The middle graph shows the gaps in the different villages types (eg participation is the same in Hogaan villages, and 13% lower for youth in non-implementation villages) and the left-hand graph gives the baseline percentages (eg 16% participation for both youth and older people in Hogaan villages and in non- implementation villages 6% for youth and 19% for older people).

In a bit more detail

Elders: Social Services Elders: Social Services

Hogaan 20 (14,● 26)

2 (−5,● 8)

Non−imp 18 (11,● 26)

VC: Social Services VC: Social Services

Hogaan 44 (33,● 54)

14 (6, ●22)***

Non−imp 30 (20,● 40) significance a● negative DC: Social Services DC: Social Services a● none a● positive Hogaan 22 (19,● 24)

−14 (−21,● −6)***

Non−imp 35 (28,● 42)

RL: Social Services RL: Social Services

Hogaan 4 (2,● 8)

−2 (−7,● 1)

Non−imp 6 (2,● 13)

0% 20% 40% −20% −10% 0% 10% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 5: Institutions Providing Social Services: Comparison Village Type

25 In the two parallel graphs, the left-hand graph shows percentages (with a 90% confidence interval) and the right-hand graph shows the difference in percentages (with a 90% confidence interval). For example the four graphs in Figure5 relate to the question of which institutions should provide social services. The top graph, entitled ’Elders: Social Services’ represents those who answered that elders should provide social services. The top line in the left hand graph shows the most likely estimate for the percentage of respondents in implementation (Hogaan) villages who think that elders should provide social services. The percentage is 20%, but has a 90% confidence interval spanning from 14-26%. The bottom line provides the same information for the non-implementation villages, in this case estimated at 18% (with a 90% confidence interval spanning the range 11-25%). The right hand graph addresses the question of whether there is a significant difference between the figures in the left hand graph for the implementation (Hogaan) villages and the non-implementation villages. Of course on average the difference between the implementation estimate of 20 and the non-implementation estimate of 18 is 2. However, the confidence interval of the difference spans the range from -5 to 8. Because this confidence interval includes zero, the difference between the implementation and non-implementation villages is not statistically significant. The second figure provides the same information for the question of whether the Village Council should provide social services. In this case 44% (33-53%) of implementation respondents compared to 30% (21-40%) of non- implementation respondents think so. The difference of 14% has a confidence interval from 6-21% which does not include zero, hence the difference is statistically significant. To assist with the recognition of overall patterns we have colour-coded the significance of responses. Positive significant responses (where implementation respondents are higher than non-implementation respondents) are coloured green, non-significant responses are coloured grey and negative significant responses (where implementation respondents are lower than non-implementation respondents) are coloured red. The number of stars in the left hand figure indicate the level of statistical significance, ∗ indicates significance at the 90% level (p < .1), ∗∗ indicates significance at the 95% level (p < .05), and ∗∗∗ indicates significance at the 99% level (p < .01). Cases where the differences we find are not statistically significant (p > .01) are described as null results. A null result means that we cannot reject the null hypothesis of no difference between the cases. Note that not being able to reject the null hypothesis is not the same as demonstrating that there actually is no difference between the cases. Other figures include three parallel running graphs, such as for example:

Respondent made request of local government Respondent made request of local government Respondent made request of local government

b. Hogaan 16 (13,● 19) Youth b. Hogaan 0 (−4, 4) ●

b. Hogaan 16 (13,● 19) difference 35+ 13 (4, 20)*** significance ● a● positive a. Non−imp 6 (3,● 11) Youth a. Non−imp −13 (−19, −6)*** ● difference a. Non−imp 19 (12,● 25) 35+

0% 10% 20% −20% −15% −10% −5% 0% 5% 0% 5% 10% 15% 20% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 6: Participation: Comparison Youth Groups

In the above presented three parallel graphs, the left-hand graph shows percentages, the middle graph shows the diﬀerence of percentage and the right-hand graph shows the difference in the diﬀerences. These comparisons are largely used to examine questions about inclusion. The Hogaan programme might have had an impact on inclusion without bring- ing about full equality. Figure6 addresses youth inclusion on the question of whether the

26 individual has made a request of local government. The left hand figure give the percentages of youth (16% with CI 13-19%) and adults (16% with CI 13-19%) in implementation villages in the top two lines, and the percentages of youths (6%, 90% CI 3-11%) and adults (19% CI 13-26%) in non-implementation villages in the bottom two lines. The middle graph addresses the question of whether there is a difference between youths and adults in implementation villages and non-implementation villages. The top line relates to the implementation villages, where there is no significant difference, indeed the point estimate of the difference is 0% (-4,+4%). The bottom line relates to non-implementation villages where youths are about 13% (-20 to -6%) less likely to have made a request, which is a statistically significant negative difference at the 99% level. The third parallel graph addresses the key question from the point of view of the evaluation of whether there is a significant difference in these differences. The point estimate of the difference in the differences is 13%, with a confidence interval from 5% to 21% which does not include 0. Hence there is a statistically significant difference in youth inclusion on having made a request of local government.

3.1 Participation (Outcome 1)

Given the increasing importance of participatory approaches to development since the 1980s, there is an enormous body of literature available, including empirically driven studies on the challenges of participation more generally and in development in particular.2 Most of these studies emphasise that participation comprises a broad range different activities ranging from mobilizing people, to attending meetings, discussing ideas, sharing information, consulting, organising self-help groups or making joint decision. In CDD/R, participation is frequently connected to inclusiveness and thus implies that those most affected by the proposed intervention are included in decision-making. However, the form of this inclusion can again range from nominal and passive participation (that is mere membership in a group or being informed of decisions), to consultative participation (where one is at least consulted and requested to raise an opinion although without any guar- antee that the opinion is considered in the decision-making), to interactive participation (which refers to having voice and power to influence decisions) (Agraval 2001: 1624f.). Participation can furthermore be organised by a group of people themselves, it can be induced or requested by local leaders or by a government, or it can be induced exter- nally to the country or region by an international organisation. The qualitative research included an assessment of how citizens in Somali villages understand participation and how they participate in decision-making processes. The findings provide the basis for understanding and interpreting results of the survey, which assessed if Hogaan has increased villagers’ participation in decision-making, and enhanced accessibility, responsiveness and inclusiveness of local governing bodies (Outcome 1).

3.1.1 Citizens’ Participation in Decision-Making In the interviews and focus group discussions, interviewees emphasized the inclusive and participatory nature of decision-making in their villages. They pointed to regular discus-

2 To name but a few in alphabetical order: Agarwal 2001; Cohen and Uphoﬀ 1980, Cornwall 2000, Mohan and Stokke 2000, Cooke and Kothari 2001, Chopra and Hohe 2004, Cornwall and Brock 2005, Hickey and Mohan 2005, Reich 2006, Cornwall 2008, Jha et al. 2010, Hoenke and Hoenke 2012, Kyamusugulwa 2013, Mansuri and Rao 2013.

27 sions and consultations about important village matters. Several interviewees outlined that villagers come together to discuss and decide about village matters and projects: ’whenever, there is a big issue they [villagers, JB] come together and decide together’. Although some members of the development committees attributed these meetings to the Hogaan programme, respondents in non-implementation villages referred to similar processes. The participatory nature of decision-making in Somali villages was conﬁrmed in the survey. Asked if they participate in decision-making and governance processes, the majority of respondents in both implementation villages (60%) and non-implementation villages (55%) emphasized their participation in community meetings and in VC meetings (84/82%) (Figure7). 3

Currently Attends Community Meetings Currently Attends Community Meetings

Hogaan 60 (49,● 70)

6 (−3,● 14)

Non−imp 55 (42,● 65)

Should Attend Community Meetings Should Attend Community Meetings

Hogaan 85 (63,● 94) significance 3 (−3,● 10) a● none Non−imp 82 (59,● 93)

Attend VC Meeting This Year Attend VC Meeting This Year

Hogaan 84 (73,● 90)

1 (−6,● 9)

Non−imp 82 (71,● 90)

0% 25% 50% 75% 100% −5% 0% 5% 10% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 7: Participation: Comparison Village Type (1/5)

While respondents thus conﬁrmed the participatory nature of decision-making, the evaluation found no evidence that the Hogaan intervention increased the level of participation or that it sensitized villagers for the importance of these meetings, as 85% of respondents in implementation villages compared to 82% of respondents in non-implementation villages are convinced that they should attend these meetings (Figure7). However, during interviews and discussions it became increasingly clear that meetings of the whole village do not happen regularly, but mainly when larger problems occur, for example ‘when natural disasters happen, like war, drought or widespread disease.’ Village meetings are also conducted when larger development projects are envisaged, such as ‘the digging of boreholes, mobilizing the community for a construction of a police station and even when the construction of this oﬃce4 started, DRC and the community came together.’ With few exceptions, development-oriented village meetings were related to initiatives started by international organisations. Villagers also regularly mentioned religiously driven community meetings and initiatives, such as the building of mosques, or the organisation of religious festivities. Interviewees again described a fairly democratic process for such major decision- making, which is based on the selection of representatives and sometimes on public debate. One villager, for example, described how a female representative for the VC was selected:

First, we [referring to the whole village, J.B] meet, then we divide ourselves in the way we think is the best. At times we give our decision to people we

3 Here and in the following the ﬁrst number refers to the percentage of respondents from implementation villages, the second to the percentage of respondents from non-implementation villages. 4 The interviewee refers here to the Hogaan community hall in which the interviews were conducted

28 select and we trust. Once we give our conﬁdence to those we select, they bring what they decide back to us and we endorse it because they are the once we assigned. This is how we reach community decisions.

Another quite common practice of participation is through clan-based representation. Diﬀerent interviewees in several villages described these representation mechanisms. For example one of the descriptions was as follows:

when we are making decisions here, we have a style to follow. We use the four-clan system. [...] For example, when there is an organization that visits us, we share things in a good way. We share things between the four clans; among them, every clan sub-divides things on their own. Nothing comes back [goes wrong] [...]. Clans have one elder as their representative in every task. This is how we handle it and nobody complains about it. We share everything.

Participation in decision-making was frequently equated with information sharing. Sev- eral interviewees also understood participation as public aﬃrmation of decisions already made by relevant authorities. Indeed, after discussions, many interviewees conﬁrmed that decisions are usually made by local authorities, most notably elders and the Village Chair- man or a committee. If local leaders think that they require more information for a decision, they invite relevant stakeholders for consultation (for example representatives of a women’s or youth group or of a particular clan segment), but authorities may well decide that they already know enough about the case.

When elders sit together they inform and call others in the community including youth, elders and women, all parts. Then they announce the decisions that came out of their meeting.

Decisions are then reached among the local authorities after (sometimes lengthy) discussions and are based on consensus. Once a decision is reached, people are informed about it either through community meetings or more often through more informal channels. People are supposed to, and usually do, accept these decisions:

When there is a problem, a particular group of trusted community members are assigned to solve it. They are the ones who make decisions regarding the problems. So, girls agree with their decisions, older women also accept them, everybody in the community, once elders decide, tend to accept.

The survey results conﬁrm that the overwhelming majority of villagers appreciated these forms of decision-making. Villagers in both implementation and non-implementation villages are especially supportive of decisions made by customary authorities (Figure8). The overwhelming majority of respondents rate customary authorities as most responsive and accessible. 88/89% of respondents think that elders consider their viewpoint, and 93/92% feel that religious leaders do. They also consider customary authorities as most accessible (the latter exceeding 90%). The evaluation did not ﬁnd evidence that the Hogaan programme impacted in any way on the (very high) appreciation of customary forms of governance or the interaction of villagers with customary authorities, which exceeds 50% in both village types.

29 Elders Consider my Viewpoint Elders Consider my Viewpoint

Hogaan 88 (76,● 93)

−1 (−6,● 5)

Non−imp 89 (78,● 94)

Interact Elders Interact Elders

Hogaan 59 (48,● 68)

7 (−1,● 14)

Non−imp 52 (41,● 62)

Elders Accessible Elders Accessible

Hogaan 94 (86,● 97)

0 (−4,● 5)

Non−imp 94 (84,● 98)

significance Religious Leaders Consider my Viewpoint Religious Leaders Consider my Viewpoint a● none

Hogaan 93 (89,● 96)

2 (−2,● 7)

Non−imp 92 (85,● 95)

Interact with Relgious Leader Interact with Relgious Leader

Hogaan 63 (50,● 73)

5 (−2,● 12)

Non−imp 58 (44,● 69)

Religious Leaders Accessible Religious Leaders Accessible

Hogaan 97 (95,● 98)

2 (−1,● 6)

Non−imp 95 (89,● 98)

0% 25% 50% 75% 100% −5% 0% 5% 10% 15% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 8: Participation Customary Authority(5/5): Comparison Village Types

While customary governance was not aﬀected by the programme, the evaluation found signiﬁcant evidence that Hogaan increased the interaction between villagers and members of the formal institutions of governance (VC and DC). Compared to non-implementation villages, 18% more respondents in implementation villages interacted with the VC. Addi- tionally 5% more respondents in implementation villages rated the VC as accessible and 6% more people are convinced that the VC considers their viewpoint (Figure9). From the viewpoint of the villagers, responsiveness and accessibility of the VC thus increased.

VC Considers my Viewpoint VC Considers my Viewpoint

Hogaan 85 (69,● 91)

5 (0,● 12)*

Non−imp 79 (60,● 87)

Interacted with VC Member Interacted with VC Member

Hogaan 60 (49,● 70) significance 18 (9, ●26)*** a● positive Non−imp 43 (30,● 54)

VC Members Accessible VC Members Accessible

Hogaan 90 (76,● 96)

5 (0,● 12)*

Non−imp 85 (69,● 94)

0% 25% 50% 75% 100% 0% 10% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 9: Participation: Comparison Village Type (2/5)

Additionally, the interaction between villagers and the DC increased signiﬁcantly, and is 13% higher in implementation villages (32/45%). In spite of the higher interaction rates, however, villagers do not rate the DC as more accessible (85/82%) or responsive to their views (Figure 10).

30 DC Considers my Viewpoint DC Considers my Viewpoint

Hogaan 83 (66,● 91)

5 (−1,● 11)

Non−imp 78 (61,● 88)

Interacted with DC Member Interacted with DC Member

Hogaan 46 (34,● 57) significance 13 (5, 22)** ● ● a none ● Non−imp 33 (20,● 44) a positive

DC Members Accessible DC Members Accessible

Hogaan 85 (67,● 92)

2 (−3,● 8)

Non−imp 82 (65,● 91)

0% 25% 50% 75% 0% 5% 10% 15% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 10: Participation: Comparison Village Type (3/5)

Although villagers interact more with formal authorities, the evaluation found no evidence that the Hogaan intervention has improved villagers’ ability or willingness to articulate demands (16/13%) or that people in implementation villages are signiﬁcantly more involved in the village planning (65/57% – Figure 11).

Respondent made request of local government Respondent made request of local government

Hogaan 16 (13,● 18)

3 (−2,● 7)

Non−imp 13 (9,● 18)

significance Respondent Involved in Planning Process Respondent Involved in Planning Process a● none

Hogaan 65 (49,● 76)

7 (0,● 14)

Non−imp 58 (41,● 70)

0% 20% 40% 60% 80% 0% 5% 10% 15% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 11: Participation: Comparison Village Type (4/5)

To summarize, the evaluation found mixed evidence on Hogaan’s impact on participation and decision-making. The survey results indicate that Hogaan managed to build on available mechanisms of participation, that is consultation and information sharing, and expanded them to formal governance as it signiﬁcantly increased interaction between citizens and the VC and DC. It especially improved villagers’ views of the VC with respect to accessibility and responsiveness. There is, however, no evidence that Hogaan increased the (direct) involvement of citizens in the village planning process or that it enhanced their ability and/or willingness to raise demands. While governance in Somalia is already participatory to a certain extent, this does not imply that villagers are equally represented or that they have an equal voice. The next section will therefore assess if Hogaan managed to increase inclusiveness of governance, in particular the inclusion of women, youth and clan minorities.

3.1.2 Participation and Inclusiveness The programme aimed to include women, youth and minority clans. The definition of both ‘youth’ and minority clans’ is, however, controversial. In the absence of further clarification by the programme about whom they understand to be youth or minority, the following section outlines and explains the definitions for youth and majority/minority clan affiliation that were used by the evaluation team to further disaggregate the data.

31 Operationalizing Inclusion: On majorities, minorities and occupational castes The Hogaan programme explicitly states the intention to enhance participation of minority clans, without further specifying the minority groups they wish to include or de- veloping particular measures around how to improve their participation.5 The evaluation diﬀerentiated between two forms of belonging to a minority group:

a) as a member of a minority clan that is better described as occupational caste (Luling 1984) and

b) as a member of a politically and numerically smaller clan group in the speciﬁc location of the project.

Although not explicitly stated, the programme targeted group a, that is occupational castes6 and aimed at increasing their participation. With respect to this group, the evaluation decided to use Luling’s terminology and labelled this group as occupational caste. More than minority clan, the caste label indicates the social and political status of members of these groups. People with caste status are forming endogamous groups that are not, or not fully, integrated in the segmentary lineage or clan system.7 Due to their occupation as blacksmiths, leather workers, barbers, hunters or ‘ritual specialists’ member of caste groups are considered inferior, and are discriminated against in many aspects of the social and political life. Indeed most castes members are extremely poor and live in deprivation (Luling 1984; Cassanelli 1995; Hill 2010). The evaluation differentiates this group from people from politically and numerically smaller clan groups. Please note that the Hogaan programme did not include any reference to clan affiliation. Although the segmentary lineage system provides a social matrix through which social and political life is organised and interpreted in Somalia, the Hogaan programme (like other international interventions) avoided any reference to clan. How- ever, given the importance of clan affiliation for structuring political and social live in Somalia, the evaluation decided to integrate clan as one variable. We acknowledge the academic debate that accompanies the utilisation of the lineage system and clan affiliation to explain politics in Somalia. While some authors recognise the clan system as a prime feature of Somali politics and to some extent interpret the clan-based character of politics as a cause of violence (Lewis and Luling for example), others heavily criticise this view as it ignores other social features, such as class or gender, and overlooks social hierarchies that were introduced with commercialisation and (colonial and post-colonial) state-building (Samatar 1992; Besteman 1996; Cassanelli and Besteman 1996; Kapteijns 2011). Clan is considered directly in the analysis of inclusion; however, clan is also a control variable (along with gender and age) in the analysis of the overall effect of the Hogaan project (in other words, we aim to estimate the effect of the programme if Hogaan and non-implementation villages had the same clan, gender and age composition). This decision may require further justification: the political systems in Somalia are based to a large extent on clan representation. Although Somaliland moved away from clan-based

5 The evaluation team acknowledge the difficulties the programme had in identifying minorities and their ‘fear’ that the programme could stir up conflicts. Both problems were explained to the evaluation team. 6 Note again that they were not labelled as castes but as minority clans 7 Lewis (1961) first described the segmentary lineage system in Somalia. Based on patri-linear descent, Somali society, or rather ethnic Somalis, are divided into groups that represent different levels of segmentation as clan-families, clans, lineages and diya-paying groups. Other authors use sub-clans instead of lineages or have introduced further levels (sub-clans, sub-sub-clans etc.).

32 representation with the introduction of a multi-party democracy, it nonetheless has an unelected Upper House that is based on clan representation the Guurti (chamber of elders). Somaliland undoubtedly made remarkable steps towards democratisation. However clan still plays a signiﬁcant role in determining political alliances as well as voter decisions (for example RVI 2013). Puntland’s politics is, in contrast, explicitly clan based. Already its territorial claim is grounded in the demarcation of Darood/Harti (i.e. Majerteen, Dhulbahante, Warsangeli) clan territory, and the parliament is constituted through representation of the Harti clans. The Parliament is thus dominated by Majerteen, which are considered the numerically largest Darood/Harti clan.8 The three numerically largest Majerteen sub-clans, Cumar Maxamud, Cisse Maxamud and Cismaan Maxamud, largely dominate the government, and to date the Presidency has rotated between these three sub-clans. Each of the three sub-clans again dominates one of the three Puntland regions, that is Mudug, Nugal and Bari (ICG 2013). Members of other Majerteen clan thus have a minority status, without necessarily being discriminated against. Many of the villages in Burtinle and North Galkayo district where the Hogaan programme was implemented were inhabited by people from the Majerteen/Cumar Maxamud clan, which left village inhabitants of other (sub-)clans numerically in a minority status. Village and district level governance have a similar representational mechanism, and while Cumar Maxamud may dominate a district, they nonetheless do not necessarily dominate every single village. One clan can hence be a minority in a district, but still form a majority in a particular village or, the other way round, be a majority in a village but a minority in the district. The latter is, for example, the case in villages mainly inhabited by Darood/Leelkaase or the Darood/Awrtable. While descending from the Darood clan family, they are not part of the Majerteen clan and thus form a clan minority in the districts of Puntland. The following tables (3.1-3.3) provide an overview of the clan-aﬃliation of the interviewees in the three districts. Clan N Percentage Darood/Majerteen/Cumar Maxamud 263 57 Darood/Majerteen/Other 64 14 Darood/Majerteen/Ciise Maxamud 61 13 Darood/Awrtable 30 6 Darood/Other (not Majerteen) 22 5 Other Clan Families 11 2 Darood/Lelkasse 9 2 Darood/Majerteen/Cismaan Maxamud 3 1

Table 3.1: Clan Distribution of Burtinle Survey Respondents

Beyond the national level, village level governance is, in both Somaliland and Punt- land, structured by clan affiliation, although of course not exclusively so. As outlined above, clan affiliation plays an important role in how participation and representation are organised. It was also identified in interviews and focus groups as crucial for structuring village level governance as it constitutes the main principle for selecting Village Council representatives. Members of village administrations are usually selected by clan elders,

8 Please note that numerical diﬀerences between clan segments, and independence of the level of segmentation, are not based on statistically reliable information but rather on powered negotiations of clan sections within the lineage system.

33 Clan N Percentage Darood/Majerteen/Cumar Maxamud 291 60 Darood/Majerteen/Other 87 18 Darood/Lelkasse 42 9 Darood/Other (not Majerteen) 29 6 Other Clan Families 18 4 Darood/Majerteen/Ciise Maxamud 9 2 Darood/Awrtable 7 1 Darood/Majerteen/Cismaan Maxamud 3 1

Table 3.2: Clan Distribution of Galkayo Survey Respondents

Clan N Percentage Isaaq/Habar Yonis 159 30 Darood/Warsangeli 142 27 Isaaq/Habar Jeclo 141 27 Darood/Dulbahante 57 11 Darood/Other 21 4 Isaaq/Other 5 1 Other Clan Families 5 1

Table 3.3: Clan Distribution of Erigabo Survey Respondents often in co-operation with other, locally powerful actors such as wealthy business people or religious authorities. The composition of the council is intended to match the (nego- tiated not calculated!) numerical strength of clan, sub-clans or lineages in the village. A local leader in Erigabo district explained the procedure during a Focus Group Discussion (FGD):

We are Somalis and you know we have diﬀerences. We are from diﬀerent clans. We wanted to be cautious and we wanted to go to every clan and get their representation so that people are balanced. Then every clan could appear in the committee. So every clan living here has a member in the committee; this was the case till the 1940s and it continues like that now.

Clan-based representation was also stressed by respondents from villages in Puntland. Here the example of a young businessman who explained composition of the administration in one village in North Galkayo district:

We have a chairman and a vice chairman who are from the two main tribes in [village name]. The chairman is always from Reer Xirsi and the vice chairman is usually from Reer Mahad. Interviewer: Why always these two tribes? What do you think is the reason? They are the majority in terms of numbers, they have the majority of the business here, and they are the ones who contribute the largest share for the contributions that are collected, but the other tribes are still members of the administration committee.

34 Given the political importance of clan affiliation for representational mechanisms, the evaluation recorded the clan family, clan, subclan and lineage of respondents.9 The evaluation then calculated the strength of each clan at all levels in all villages and identified the highest level at which a significant division between groups in the village was detected. The threshold was set at 60%, which means that if 60% or more respondents were from one clan segment, this segment was classified as the majority clan of the village, while others gained the minority label. The evaluation team acknowledges that this method does not address many of the complexities of clan affiliations. Nevertheless, in the absence of previous attempts to quantify the relational dynamics of clan affiliations in villages in Somaliland and Puntland, this approach is considered as a starting point. Certainly, the resulting variable is a highly significant predictor of many attitudes, whilst more straightforward measures, which are not relative to the village majority (eg just the clan), are not significant.

Youth in Somalia/Somaliland In the Theory of Change (ToC), young people are described as being marginalised in decision-making processes. The Hogaan intervention thus aimed at increasing participation of young people in village level governance. However, no definition of youth was given. While age obviously serves as the main criterion for the definition of youth, social practices in Somalia do not provide clear age delineations or other criteria that allow for a precise definition of youth. Instead a combination of personal attributes, such as age, marriage status, children, sex, occupation (going to school/university) or income determine if a person is considered as youth or adult. Often enough people can be considered youth by some people and on one occasion, but as adults by other people or on other occasions, or may self-define their status differently at different occasions.10 In the absence of clearly formulated alternatives, for simplicity, the evaluation team decided to classify youth exclusively by age. However, to account for the flexibility of age, youth was in the first comparison defined narrowly as people aged 18-25 years (youth 1), in the second comparison more broadly as people aged 18-35 years (youth 2). Youth 2 was selected as a major reference. However, in case the analysis provided different results on programme effects of these age groups, they will be presented in the findings. If no specific reference is made, the results refer to youth 2.

Differential Participation Effects: Gender The Hogaan programme put considerable effort into increasing the participation of women in decision-making processes. However, the evaluation found no evidence that the programme has differentially increased women’s participation in formal governance.11 VC meetings are, for example, fairly gender balanced in both implementation and non-implementation villages (Figure 12). This can be contrasted with community meetings. There is no significant difference between the proportions of women and men who think they

9 The baseline also tried to identify clan affiliation of respondents, but skipped the question due to worries about enumerators’ safety. The endline survey included training of how to ask about clan affiliation of respondents and successfully gathered clan data of 98% of respondents (1571 out of the total of 1604 respondents). No safety issues occurred. 10 The same applies to the term elder. While a lot of elders are indeed older men, there is no age threshold that restricts becoming an elder. Sometimes men in their thirties are respected as elders; sometimes old men have no voice as elders in committees. 11 Remember that significance of the difference is now presented in the third parallel running graph.

35 should attend community meetings (over 80% think this). However, fewer women than men actually do attend these meetings, (-10/-16%). The difference between the 10% fewer women in implementation and and -16% fewer women in non-implementation villages, is not statistically significant and therefore we cannot rule out the hypothesis that there is no difference between Hogaan and non-implementation villages.

Currently Attends Community Meetings Currently Attends Community Meetings Currently Attends Community Meetings

b. Hogaan 68 (55,● 77) Male b. Hogaan −10 (−16, −5)*** ● difference b. Hogaan 58 (44,● 68) Female 6 (−6, 17) ●

a. Non−imp 66 (49,● 77) Male a. Non−imp −16 (−27, −6)** ● difference a. Non−imp 49 (36,● 62) Female Should Attend Community Meetings Should Attend Community Meetings Should Attend Community Meetings

b. Hogaan 86 (65,● 95) Male b. Hogaan −2 (−6, 2) ● difference b. Hogaan 84 (63,● 94) Female −1 (−11, 8) significance ● a● none a. Non−imp 81 (58,● 93) Male a. Non−imp 0 (−10, 9) ● difference a. Non−imp 81 (57,● 93) Female Attend VC Meeting This Year Attend VC Meeting This Year Attend VC Meeting This Year

b. Hogaan 84 (74,● 90) Male b. Hogaan −1 (−6, 5) ● difference b. Hogaan 84 (74,● 89) Female −6 (−18, 7) ●

a. Non−imp 79 (62,● 89) Male a. Non−imp 5 (−6, 17) ● difference a. Non−imp 84 (71,● 92) Female

0% 25% 50% 75% 100% −20% −10% 0% 10% −10% 0% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 12: Participation: Comparison Village Type (1/4)

There is also no evidence that the Hogaan intervention differentially increased the accessibility or responsiveness of the VC for women (Figure 13). Women in both village types are significantly less likely to interact with the VC (-17%/-25%). Although rates of engagement of women in non-implementation villages are much lower than those of women in Hogaan villages, the rates of engagement of men are also much lower and the difference in the difference is not significant. In spite of their significantly lower interaction with the VC, the majority (80%) of female respondents in both implementation and non-implementation villages rate the VC as accessible and are convinced that the VC considers their views. There is thus no very clear correlation between higher rates of engagement and the assessment of the responsiveness of governance institutions.

36 VC Meetings Consider my Household Viewpoint VC Meetings Consider my Household Viewpoint VC Meetings Consider my Household Viewpoint

b. Hogaan 93 (87,● 97) Male b. Hogaan 0 (−3, 5) ● difference b. Hogaan 93 (88,● 97) Female −67 (−97, 15) ●

a. Non−imp3 (0, ●100) Male a. Non−imp 67 (−14, 96) ● difference a. Non−imp 93 (78,● 98) Female VC Considers my Viewpoint VC Considers my Viewpoint VC Considers my Viewpoint

b. Hogaan 84 (68,● 91) Male b. Hogaan 1 (−4, 5) ● difference b. Hogaan 85 (68,● 91) Female −6 (−17, 3) ●

a. Non−imp 74 (55,● 85) Male a. Non−imp 6 (−2, 17) ● difference a. Non−imp 81 (65,● 90) Female significance Interacted with VC Member Interacted with VC Member Interacted with VC Member a● none

b. Hogaan 72 (59,● 81) Male b. Hogaan−17 (−22, −12)*** ● difference b. Hogaan 55 (41,● 66) Female 8 (−3, 20) ●

a. Non−imp 59 (43,● 71) Male a. Non−imp−25 (−35, −15)*** ● difference a. Non−imp 34 (21,● 46) Female VC Members Accessible VC Members Accessible VC Members Accessible

b. Hogaan 89 (74,● 96) Male b. Hogaan 3 (−1, 7) ● difference b. Hogaan 92 (78,● 97) Female −1 (−10, 7) ●

a. Non−imp 82 (65,● 92) Male a. Non−imp 4 (−3, 12) ● difference a. Non−imp 87 (70,● 95) Female

0% 25% 50% 75% 100% 0% 50% 100% −100% −75% −50% −25% 0% 25% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 13: Participation: Comparison Gender (2/4)

A similar pattern characterises the interaction of women and men with the DC (Figure 14). Women are significantly less likely to interact with the DC than men in implementation and non-implementation villages (-13/-23%). However, comparing interaction differences between men and women, the difference between Hogaan and non-implementation villages is statistically not significant. Lower interaction does again not correspond with a lack of accessibility or responsiveness. More than 70% men and over 80% women in implementation as well as non-implementation villages are convinced that the DC considers their views and over 78% of respondents of both sexes rate the DC as accessible.

DC Considers my Viewpoint DC Considers my Viewpoint DC Considers my Viewpoint

b. Hogaan 80 (63,● 89) Male b. Hogaan 4 (0, 9) ● difference b. Hogaan 84 (68,● 91) Female −3 (−15, 7) ●

a. Non−imp 72 (53,● 85) Male a. Non−imp 8 (0, 18) ● difference a. Non−imp 80 (63,● 89) Female Interacted with DC Member Interacted with DC Member Interacted with DC Member

b. Hogaan 55 (40,● 66) Male b. Hogaan −13 (−18, −8)*** ● difference b. Hogaan 42 (28,● 54) Female 10 (−1, 23) significance ● a● none a. Non−imp 48 (31,● 62) Male a. Non−imp −23 (−35, −13)*** ● difference a. Non−imp 24 (12,● 36) Female DC Members Accessible DC Members Accessible DC Members Accessible

b. Hogaan 82 (65,● 91) Male b. Hogaan 4 (−1, 8) ● difference b. Hogaan 86 (69,● 93) Female 1 (−9, 10) ●

a. Non−imp 80 (62,● 90) Male a. Non−imp 3 (−5, 11) ● difference a. Non−imp 83 (66,● 92) Female

0% 25% 50% 75% −30% −20% −10% 0% 10% 20% −10% 0% 10% 20% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 14: Participation: Comparison Gender (3/4)

37 In contrast to formal governance, the evaluation did find significant effects on the interaction of women with customary authorities (See Figure 59 in Appendix). Women generally interact much less with elders and religious leaders than men, but the difference between men and women is significantly lower in implementation villages where 20% fewer women interact with elders compared to 39% fewer women in non-implementation villages, amounting to a 19% difference. A similar pattern is found with respect to religious authorities. In both village types women are much less likely to interact with religious leaders, but the difference between men and women is significantly lower in implementation villages. In both implementation and non-implementation villages, women rank elders and religious leaders as most accessible (>93%) and responsive (elders >86%; religious leaders >91%). The evaluation did not find evidence that the Hogaan programme impacted on any other participation measures with respect to gender (see Figures in Appendix: Further Results). To summarize, the Hogaan programme did differentially increase interaction of women especially with customary authorities. Recall that in implementation villages, women do interact significantly more with the VC and the DC than in non-implementation villages; however, there is no evidence of a particular gender effect as men also interact more. The evaluation additionally found no evidence that higher rates of engagement with local institutions impacted on how respondents assessed either the accessibility or responsiveness of these institutions. For example, women interact much less with formal and customary institutions than men, but nonetheless share the views of men on accessibility and responsiveness of these institutions. And although interactions of women with customary institutions increased in implementation villages, this increase did not translate into significantly higher ratings of responsiveness or accessibility of elders or religious authorities.

Differential Participation Effects: Youth The evaluation did not find evidence that the programme had a significant differential effect on youth participation in community or VC meetings, the accessibility of the VC and DC for young people or the participation of youth in planning processes (Figures 15, 16 and 82 (Appendix)). Young people rate the accessibility of institutions higher than adults, if anything. Irrespective of the village type, more than 80% of youth respondents rate formal institutions as accessible, and over 90% rate customary institutions as accessible. Young people, however, interact significantly less with formal and with customary institutions: they interact less with the VC (-10/-17%), the DC (-11/-10%), elders (-13/-17%), and religious leaders (-13/-20%) (Figures 15). We do not find a significant difference in the difference between youth and elder interactions with local authorities in implementation and non-implementation villages. However, in implementation villages young people rate the responsiveness of formal and customary institutions higher, and youth here are more convinced than adults that the DC considers their viewpoint (85% youth, 81% adults, difference: 4%). In non-implementation villages young people are -11% less likely to rate the DC as responsive, and are also more critical than adults in the same village type (72% youth, 79% adults, difference -7%). Additionally, significantly more young people in implementation villages rate elders and religious leaders as responsive. If youth group 1 (18-25years) is taken as reference, these differences are slightly larger (see Appendix: Further Results).

38 VC Meetings Consider my Household Viewpoint VC Meetings Consider my Household Viewpoint VC Meetings Consider my Household Viewpoint

b. Hogaan 95 (90,● 98) Youth b. Hogaan 2 (−1, 6) ● difference b. Hogaan 93 (88,● 97) 35+ 6 (−2, 17) ●

a. Non−imp 93 (75,● 98) Youth a. Non−imp −3 (−15, 3) ● difference a. Non−imp 97 (85,● 100) 35+

VC Considers my Viewpoint VC Considers my Viewpoint VC Considers my Viewpoint

b. Hogaan 87 (72,● 93) Youth b. Hogaan 5 (1, 10)** ● difference b. Hogaan 82 (66,● 90) 35+ 4 (−5, 14) ●

a. Non−imp 79 (61,● 89) Youth a. Non−imp 1 (−8, 9) ● difference a. Non−imp 79 (61,● 88) 35+

Interacted with VC Member Interacted with VC Member Interacted with VC Member

b. Hogaan 55 (42,● 65) Youth b. Hogaan −10 (−14, −5)*** ● difference b. Hogaan 64 (52,● 74) 35+ 8 (−3, 18) ●

a. Non−imp 33 (20,● 46) Youth a. Non−imp −17 (−27, −8)*** ● difference a. Non−imp 50 (37,● 63) 35+

VC Members Accessible VC Members Accessible VC Members Accessible

b. Hogaan 94 (79,● 98) Youth b. Hogaan 4 (1, 9)** ● b. Hogaan difference 89 (71,● 96) significance 35+ 0 (−8, 7) ● none ● a ● positive a. Non−imp 88 (71,● 96) a Youth a. Non−imp 5 (−1, 12) ● difference a. Non−imp 83 (67,● 93) 35+

DC Considers my Viewpoint DC Considers my Viewpoint DC Considers my Viewpoint

b. Hogaan 85 (69,● 92) Youth b. Hogaan 5 (1, 9)* ● difference b. Hogaan 80 (64,● 89) 35+ 11 (2, 21)** ●

a. Non−imp 74 (55,● 86) Youth a. Non−imp −6 (−15, 1) ● difference a. Non−imp 81 (63,● 90) 35+

Interacted with DC Member Interacted with DC Member Interacted with DC Member

b. Hogaan 40 (28,● 52) Youth b. Hogaan −10 (−15, −6)*** ● difference b. Hogaan 51 (37,● 62) 35+ 0 (−11, 11) ●

a. Non−imp 27 (15,● 40) Youth a. Non−imp −10 (−20, 0)* ● difference a. Non−imp 38 (24,● 50) 35+

DC Members Accessible DC Members Accessible DC Members Accessible

b. Hogaan 87 (68,● 93) Youth b. Hogaan 3 (−1, 7) ● difference b. Hogaan 84 (66,● 92) 35+ 2 (−7, 11) ●

a. Non−imp 83 (64,● 92) Youth a. Non−imp 1 (−6, 8) ● difference a. Non−imp 82 (63,● 91) 35+

0% 25% 50% 75% 100% −20% −10% 0% 10% −10% 0% 10% 20% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 15: Participation: Comparison Youth (2/3)

Although we did not find evidence of a general effect on the ability of citizens to articulate demands, we nevertheless do find a significant effect of the Hogaan programme on young people. Whilst youth in implementation villages are as likely as adults to articulate requests, they are significantly less likely to do so (-13% difference in the difference) in non-implementation villages (Figures 16).

39 Elders Consider my Viewpoint Elders Consider my Viewpoint Elders Consider my Viewpoint

b. Hogaan 89 (77,● 93) Youth b. Hogaan 2 (−2, 6) ● difference b. Hogaan 87 (74,● 92) 35+ 11 (3, 21)** ●

a. Non−imp 84 (68,● 91) Youth a. Non−imp −9 (−18, −2)** ● difference a. Non−imp 92 (82,● 96) 35+

Interact Elders Interact Elders Interact Elders

b. Hogaan 52 (40,● 62) Youth b. Hogaan −13 (−18, −8)*** ● difference b. Hogaan 65 (53,● 74) 35+ 5 (−6, 16) ●

a. Non−imp 43 (30,● 55) Youth a. Non−imp −18 (−27, −8)*** ● difference a. Non−imp 61 (48,● 71) 35+

Elders Accessible Elders Accessible Elders Accessible

b. Hogaan 95 (87,● 98) Youth b. Hogaan 2 (0, 5) ● difference b. Hogaan 93 (84,● 97) 35+ 4 (−3, 13) ●

a. Non−imp 92 (78,● 97) Youth a. Non−imp −2 (−10, 4) ● difference a. Non−imp 95 (84,● 98) 35+

Religious Leaders Consider my Viewpoint Religious Leaders Consider my Viewpoint Religious Leaders Consider my Viewpoint

b. Hogaan 94 (90,● 96) Youth b. Hogaan 1 (−2, 4) ● b. Hogaan difference 93 (88,● 96) significance 35+ 11 (5, 20)*** ● none ● a ● positive a. Non−imp 86 (75,● 92) a Youth a. Non−imp −10 (−19, −4)*** ● difference a. Non−imp 97 (91,● 99) 35+

Interact with Relgious Leader Interact with Relgious Leader Interact with Relgious Leader

b. Hogaan 56 (42,● 67) Youth b. Hogaan −13 (−18, −9)*** ● difference b. Hogaan 70 (57,● 79) 35+ 7 (−4, 19) ●

a. Non−imp 46 (32,● 59) Youth a. Non−imp −20 (−30, −10)*** ● difference a. Non−imp 67 (52,● 77) 35+

Religious Leaders Accessible Religious Leaders Accessible Religious Leaders Accessible

b. Hogaan 96 (94,● 98) Youth b. Hogaan −1 (−3, 1) ● difference b. Hogaan 97 (95,● 98) 35+ 0 (−6, 8) ●

a. Non−imp 94 (85,● 98) Youth a. Non−imp −1 (−8, 5) ● difference a. Non−imp 96 (88,● 99) 35+

Respondent made request of local government Respondent made request of local government Respondent made request of local government

b. Hogaan 16 (13,● 19) Youth b. Hogaan 0 (−4, 4) ● difference b. Hogaan 16 (13,● 19) 35+ 13 (5, 21)** ●

a. Non−imp6 (3,● 11) Youth a. Non−imp −13 (−20, −6)*** ● difference a. Non−imp 19 (12,● 26) 35+

0% 25% 50% 75% 100% −30% −20% −10% 0% 0% 10% 20% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 16: Participation: Comparison Youth (3/3)

To summarize, evidence of the impact of Hogaan on youth’s inclusion and participation in decision-making is mixed. Measured relative to older people, the programme did not signiﬁcantly enhance the rates of participation of young people in meetings, nor did it increase interaction of young people with the VC or informal authorities. Nonetheless, the programme had a positive impact on how young people rate responsiveness of formal as well as customary institutions. Relative to older people, they are signiﬁcantly more likely in Hogaan villages to believe that their viewpoint is considered by the DC, elders and religious authorities. We also found evidence that Hogaan improved the relative ability of young people to raise demands.

40 Differential Participation Effects: Clan Clan affiliation in general and membership of a majority or minority clan in particular have a significant impact on most participation measures. For example, 8/14% fewer people from minority clans attend community meetings and 2/16% fewer attend VC meetings. Although the programme did not aim at influencing the clan-based character of politics in Somalia, the evaluation found evidence that the programme had a differential positive impact on some participation measures (Figures ??-19). Relative to majority clan members, people from minority clans in implementation villages are 11% more involved in village planning. Evidence also suggests that Hogaan improved the relative interaction of minorities with the DC (18%) and with elders (9%), and increased the relative accessibility of elders and religious leaders (7%) for members of minority clans.12

Currently Attends Community Meetings Currently Attends Community Meetings Currently Attends Community Meetings

b. Hogaan 56 (43,● 66) Minority b. Hogaan −8 (−14, −3)*** ● difference b. Hogaan 64 (53,● 73) Majority 5 (−7, 18) ●

a. Non−imp 47 (32,● 60) Minority a. Non−imp −14 (−25, −2)** ● difference a. Non−imp 61 (47,● 72) Majority Should Attend Community Meetings Should Attend Community Meetings Should Attend Community Meetings

b. Hogaan 81 (59,● 93) Minority b. Hogaan −4 (−10, −1)** ● difference b. Hogaan 86 (66,● 95) significance Majority −5 (−15, 4) ● none ● a ● positive a. Non−imp 82 (58,● 94) a Minority a. Non−imp 1 (−8, 10) ● difference a. Non−imp 81 (59,● 94) Majority Respondent Involved in Planning Process Respondent Involved in Planning Process Respondent Involved in Planning Process

b. Hogaan 64 (47,● 76) Minority b. Hogaan −2 (−7, 3) ● difference b. Hogaan 66 (49,● 78) Majority 11 (0, 22)* ●

a. Non−imp 48 (30,● 63) Minority a. Non−imp −13 (−23, −4)** ● difference a. Non−imp 61 (44,● 74) Majority

12 Please note here that non-implementation villages were more clan homogeneous than implementation villages. Although we control for this, comparability with respect to clan dynamics may be somewhat limited and caution should be used in the interpretation of the programme eﬀects we discover.

41 Attend VC Meeting This Year Attend VC Meeting This Year Attend VC Meeting This Year

b. Hogaan 83 (71,● 89) Minority b. Hogaan −2 (−8, 3) ● difference b. Hogaan 85 (74,● 90) Majority 14 (0, 29)* ●

a. Non−imp 70 (51,● 83) Minority a. Non−imp −16 (−31, −4)** ● difference a. Non−imp 87 (75,● 93) Majority Interacted with VC Member Interacted with VC Member Interacted with VC Member

b. Hogaan 58 (45,● 68) Minority b. Hogaan −4 (−10, 0) ● difference b. Hogaan 63 (50,● 72) significance Majority 11 (0, 22) ● none ● a ● positive a. Non−imp 33 (20,● 46) a Minority a. Non−imp −15 (−26, −5)** ● difference a. Non−imp 48 (35,● 60) Majority Interacted with DC Member Interacted with DC Member Interacted with DC Member

b. Hogaan 45 (31,● 56) Minority b. Hogaan −2 (−8, 3) ● difference b. Hogaan 47 (33,● 59) Majority 18 (6, 30)** ●

a. Non−imp 17 (7,● 29) Minority a. Non−imp −21 (−32, −10)*** ● difference a. Non−imp 39 (25,● 50) Majority

0% 25% 50% 75% −30% −20% −10% 0% 0% 10% 20% 30% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 18: Participation: Comparison Clan (2/3)

Elders Accessible Elders Accessible Elders Accessible

b. Hogaan 95 (89,● 98) Minority b. Hogaan 0 (−3, 3) ● difference b. Hogaan 95 (89,● 97) Majority 9 (1, 19)* ●

a. Non−imp 88 (73,● 95) Minority a. Non−imp −8 (−19, −1)** ● difference a. Non−imp 97 (88,● 99) Majority significance Religious Leaders Accessible Religious Leaders Accessible Religious Leaders Accessible a● positive

b. Hogaan 96 (92,● 98) Minority b. Hogaan −1 (−4, 1) ● difference b. Hogaan 97 (95,● 98) Majority 8 (0, 18)* ●

a. Non−imp 88 (75,● 95) Minority a. Non−imp −9 (−19, −2)** ● difference a. Non−imp 98 (91,● 99) Majority

0% 25% 50% 75% 100% −20% −15% −10% −5% 0% 0% 5% 10% 15% 20% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 19: Participation: Comparison Clan (3/3)

With respect to occupational castes, the programme explicitly aimed at increasing their participation but largely failed to identify caste members in villages. No specific measures were thus developed to increase their participation or at least to limit discrimination.13 While numbers of caste members are too small to allow for a meaningful comparison between implementation and non-implementation villages, and thus implementation effects can not be determined, the evaluation analysed participation and inclusion of caste members within the implementation villages to provide a clearer analysis of their marginalised status. The analysis shows very clearly that members of occupational castes differ significantly in most measures from people with majority clan background. 26% fewer members of occupational castes attend community meetings. Caste members have 15% lower interaction rates with VC members and are 12% less convinced that their viewpoint is considered by the VC. They also find the VC less accessible and are significantly less likely to participate in village planning processes (-16% – Figure 20).

13 Many hurdles in the identification of caste members were identified by the programme staff. Generally speaking, however, the identification of caste members is neither very difficult nor particularly conflict- sensitive. Implementation of measures to improve the socio-economic or even political status of outcast groups, on the other hand, is very likely to cause tensions.

42 Interacted with a VC Member Interacted with a VC Member 43 (26, 57) Outcast ● −17 (−27, −6)** Outcast ● 54 (41, 65) Minority ● −6 (−11, −1)** 60 (48, 70) Minority ● Majority ●

VC Members Accessible VC Members Accessible

78 (59,● 91) Outcast −12 (−23,● −5)*** Outcast significance 88 (71, 95) ● negative Minority ● a −4 (−8, −1)* a● none 92 (76, 97) Minority ● Majority ●

Respondent Involved in Planning Process Respondent Involved in Planning Process 48 (29, 64) Outcast ● −17 (−28, −6)*** Outcast ● 60 (44, 73) Minority ● −4 (−9, 0) 65 (49, 76) Minority ● Majority ●

0% 25% 50% 75% 100% −20% −10% 0% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 20: Participation: Comparison Clan

The evaluation cannot assess whether there has been a change in the position of occupational castes, but does clearly show that their participation in planning and decision- making is significantly lower that participation of people within the segmentary lineage system. However, to summarize the effect of Hogaan on participation of minority clans remains difficult, as the operationalisation of clan dynamics for quantitative purposes is not yet fully developed. Hogaan, however, seems to have had a positive effect on participation of people from minority clans.

Inclusiveness of Planning and Decision-Making The evaluation found evidence that the Hogaan programme had a positive effect on the householders’ perception of inclusiveness in decision-making. Overall, a large majority of villagers even in non-implementation villages believe that village level decision-making is inclusive, but nevertheless this view was significantly more prevalent in implementation villages. The conclusion that women, youth and minority clans are significantly more involved in decision-making remains intact even if the analysis is restricted to women, youth and minority clans. Women in implementation villages feel that women are significantly more involved in the planning process (13%); similarly youth feel that youth are significantly more involved (15%). However, the evaluation did not find significantly more women or young people who answered that they themselves have actually participated in village planning. On the other hand, the evaluation did find that minority clan members were more likely to believe themselves to be be involved in planning in implementation than in non-implementation villages (17%). There is therefore evidence that the programme has had an impact on both the perception of group inclusion, and on the individual level feelings of inclusion of marginalised groups. However, there is not a simple one-to-one relationship between these. Group levels feelings of inclusion in particular are not necessarily reflected in individual level patterns of inclusion (Figures 21-23).

43 Respondent Involved in Planning Process[Female Respondents] Respondent Involved in Planning Process[Female Respondents]

Hogaan 59 (41,● 73)

5 (−4,● 14)

Non−imp 53 (36,● 68) significance Women Involved in Planning Process[Female Respondents] Women Involved in Planning Process[Female Respondents] a● none a● positive Hogaan 83 (70,● 90)

13 (5, ●23)***

Non−imp 68 (54,● 79)

0% 25% 50% 75% −5% 0% 5% 10% 15% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 21: Women’s Perception and Factual Involvement in Planning

Respondent Involved in Planning Process[Youth Respondents] Respondent Involved in Planning Process[Youth Respondents]

Hogaan 61 (43,● 74)

6 (−5,● 17)

Non−imp 55 (37,● 70) significance Youth Involved in Planning Process[Youth Respondents] Youth Involved in Planning Process[Youth Respondents] a● none a● positive Hogaan 83 (71,● 89)

15 (5,● 26)**

Non−imp 68 (51,● 79)

0% 25% 50% 75% 0% 10% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 22: Youth Perception and Factual Involvement in Planning

Respondent Involved in Planning Process[Minority Clan Respondents] Respondent Involved in Planning Process[Minority Clan Respondents]

Hogaan 62 (47,● 74)

17 (6,● 29)**

Non−imp 45 (28,● 61) significance Minority Clans Involved in Planning Process[Minority Clan Respondents] Minority Clans Involved in Planning Process[Minority Clan Respondents] a● none a● positive Hogaan 79 (68,● 87)

9 (−2,● 23)

Non−imp 69 (52,● 81)

0% 25% 50% 75% 0% 10% 20% 30% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 23: Minority Clan Perception and Factual Involvement in Planning

If answers of marginalised people are further disaggregated by gender, profound imbal- ances become visible. Young women in implementation villages are with -17% signiﬁcantly less involved than young men in planning and -14% fewer female than male members of minority groups participate in planning. However independent of their gender and age, men and women are nearly equally convinced that youth (80%) and minority groups (77%) are indeed integrated in the planning processes.

Respondent Involved in Planning Process[Youth Respondents] Respondent Involved in Planning Process[Youth Respondents]

57 (39,● 71) Female −13 (−21, −6)*** Female ● 70 (52, 82) Male ● significance

● Youth Involved in Planning Process[Youth Respondents] Youth Involved in Planning Process[Youth Respondents] a negative a● none 80 (67,● 87) Female 0 (−7, 6) Female ● 80 (68, 88) Male ●

0% 25% 50% 75% −20% −10% 0% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 24: Young Men/Women Perception and Factual Involvement in Planning

44 Respondent Involved in Planning Process[Minority Clan Respondents] Respondent Involved in Planning Process[Minority Clan Respondents]

56 (36,● 72) Female −14 (−23, −7)*** Female ● 71 (51, 84) Male ● significance

● Minority Clans Involved in Planning Process[Minority Clan Respondents] Minority Clans Involved in Planning Process[Minority Clan Respondents]a negative a● none 77 (64,● 86) Female 1 (−7, 8) Female ● 77 (62, 86) Male ●

0% 25% 50% 75% −20% −10% 0% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 25: Minority Clan Men/Women Perception and Factual Involvement in Planning

Inclusiveness and Equality In order to determine if the project had an impact on citizens’ views on equality norms, respondents were asked if they believe that marginalised groups (women, youths, clans, occupational castes) should be protected by the law, if they should have equal rights and if they currently have equal rights. While an overwhelming majority of respondents believe that women (98/98%) and youth (97/96%) have the right to be protected by the law, 7% less respondents in implementation villages believe that rights should be equal for women (69/77%) and signiﬁcantly less (-9%) that it should be equal for youth (77/86%).

Do Women Have a Right to Protection Under the Law Do Women Have a Right to Protection Under the Law

Hogaan 98 (96,● 99)

0 (−2,● 4)

Non−imp 98 (93,● 100)

Should Women Have Same Rights as Men Should Women Have Same Rights as Men

Hogaan 69 (65,● 73) significance −7 (−14,● 0) a● none Non−imp 77 (69,● 82)

Do Women Have the Same Rights as Men Do Women Have the Same Rights as Men

Hogaan 50 (42,● 57)

−2 (−10,● 6)

Non−imp 53 (43,● 61)

0% 25% 50% 75% 100% −10% −5% 0% 5% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 26: Women’s Rights

Do Youth Have a Right to Protection Under the Law Do Youth Have a Right to Protection Under the Law

Hogaan 97 (92,● 98)

1 (−2,● 6)

Non−imp 96 (89,● 98)

Should Youth Have Same Rights as Elderly People Should Youth Have Same Rights as Elderly People

Hogaan 77 (74,● 80) significance −9 (−14, −3)** ● ● a negative ● Non−imp 86 (80,● 90) a none

Do Youth Have the Same Rights as The Elderly Do Youth Have the Same Rights as The Elderly

Hogaan 57 (51,● 62)

1 (−6,● 9)

Non−imp 55 (47,● 63)

0% 25% 50% 75% 100% −15% −10% −5% 0% 5% 10% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 27: Youth Rights

An overwhelming majority of respondents agree with the view that people from minority

45 clans should have equal rights, and that they are entitled to legal protection. Substantially fewer respondents believe that occupational caste members either should have (79/82%) or indeed do have (64/62%) equal rights. The evaluation finds no significant difference between implementation and non-implementation villages on these measures. In general, gender, age and clan affiliation of respondents had no significant effect on the perception of equality norms. In non-implementation villages, 11% more women support gender equality. However, we find no evidence that the lower difference in implementation villages is attributed to the project. Hogaan did not sensitize villagers on equality questions, as people in implementation as well as non-implementation villages are relatively aware of the fact that women and youth do not have the same rights. Only half of the respondents in both village types are convinced that women (51/52%) and youth (57/55%) indeed have the same rights and there is no difference between men and women in how they rate the equality status (Figures 60 and 61 in Appendix).

Inclusiveness of Governance Institutions The evaluation also assessed how inclusive community needs prioritisation is and which of the locally governing bodies is considered the most inclusive. The evaluation found no evidence that the Hogaan programme had an impact either on deﬁning or on clarifying responsibilities for inclusive governance. In both implementation and non-implementation villages, religious leaders are considered the most inclusive, but signiﬁcantly lower in implementation villages with respect to considering priorities of women (94/99%) and occupational castes (82/90%). The VC, DC and elders rank fairly equally in both village types, with prioritisation of women or youth issues ranking between 80-90%, and caste group between 70 and 75%. Villagers also seem to agree that priorities of occupational castes are least considered by all governing institutions (Figure 50 in Appendix).

Inclusiveness of Conflict Resolution The evaluation did find a significant difference between implementation and non-implementation villages with respect to inclusiveness of conflict resolution. Generally the majority of people in both implementation and non-implementation villages (81/82%) state that conflict resolution is effective. However, 9% more respondents in Hogaan villages (77/68%) believe that woman are addressed in conflict resolution and 9% more (77/68%) believe that youth issues are. This result corresponds with the general higher perception of inclusiveness in implementation villages (Figure 28).

Women's Issues in Conflict Resolution Women's Issues in Conflict Resolution

Hogaan 79 (73,● 83)

9 (2, ●15)**

Non−imp 70 (61,● 77)

Youth Issues in Conflict Resolution Youth Issues in Conflict Resolution

Hogaan 77 (68,● 82) significance 9 (3, 16)** ● ● a none ● Non−imp 67 (57,● 76) a positive

Minority Clan Issues in Conflict Resolution Minority Clan Issues in Conflict Resolution

Hogaan 76 (67,● 82)

3 (−2,● 10)

Non−imp 73 (61,● 81)

0% 20% 40% 60% 80% 0% 5% 10% 15% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 28: Conﬂict Mechanisms: Comparison Village Type

46 Local Leaders on Equality Norms The leadership survey asked local authorities for their views on equality and rights protection. The majority of local leaders in both implementation (77%) and non-implementation villages (76%) think that women and men should have equal rights. A similar percentage of leaders in both village types (71/68%) are convinced that equal rights are already realised. However, authorities in Hogaan villages are more sensitized towards the difficulties of including women in decision-making: 21% more authorities in Hogaan villages (58/36%) acknowledge that women often play a limited role in decision-making and 22% more authorities admit the difficulty to including women (51/28%). Hogaan thus had a positive effect on sensitizing local leaders on gendered access to participation and decision-making.

Should Women Have Same Rights as Men Should Women Have Same Rights as Men

Hogaan 76 (63,● 85)

−3 (−14,● 9)

Non−imp 79 (62,● 90)

Do Women Have Same Rights as Men Do Women Have Same Rights as Men

Hogaan 71 (56,● 82)

3 (−8,● 15)

Non−imp 68 (49,● 81) significance In Some Villages Women Play a Limited Role in Village Decisions In Some Villages Women Play a Limited Role in Village Decisions a● none a● positive Hogaan 57 (39,● 70)

21 (9, ●33)***

Non−imp 36 (18,● 53)

It is Difficult to Include Women in Decision Making Process It is Difficult to Include Women in Decision Making Process

Hogaan 51 (31,● 66)

22 (11,● 34)***

Non−imp 27 (11,● 46)

0% 25% 50% 75% −10% 0% 10% 20% 30% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 29: Leader Perception Women’s Rights and Inclusion: Comparison Village Type

There are no signiﬁcant diﬀerences on the local leaders’ views on rights of youth, minority clans or caste groups. The majority of authorities in both Hogaan and non-implementation villages is convinced that youth and minority clans should have and actually do have the same rights as older people and people from majority clans (See Figure 55 in Appendix).

Conclusion: Hogaan’s impact on participation/inclusion

The evaluation finds mixed evidence for programme effects on citizens’ participation in local level governance. The programme improved interaction of villagers with formal governing bodies and interaction of women with customary authorities, and increased responsiveness of the Village Council and District Council (from the viewpoint of villagers). The evaluation however found no evidence that the intervention had an overall effect on other participation measures, such as participation of citizens in village planning or other decision-making processes. Additionally, the capacity or willingness of citizens to raise demands was not significantly affected by the Hogaan intervention. Given the priority and effort Hogaan puts into sensitizing villagers to define needs and demand adequate services and, especially during village entry and the CAP process, to participate in planning, this result is rather disappointing. Evidence for the programme’s objective to particularly increase participation of marginalised groups - women, youth and outcasts - is rather weak. Participation of member of

47 these groups is much lower, and there is no evidence that Hogaan increased either women’s or youth’s participation in decision-making. However, it enhanced interaction of women with customary authorities and managed to support young people to articulate demands. The evaluation was unable to determine an intervention eﬀect on occupational castes, and caste members participate much less in decision-making processes than members of other clan groups. There is evidence that the participation of minority clans has increased. The intervention had a positive eﬀect on the responsiveness of local institutions. Villagers in implementation villages are generally convinced that their viewpoints are considered, especially by formal authorities. People in Hogaan villages are also inclined to believe that local institutions are inclusive, although the evaluation found limited individual level evidence to support this. The programme did not have a positive impact on the promotion of equality norms, and may have negatively impacted on norms surrounding gender and youth equality. However, it sensitized local leaders to the priorities of marginalised groups although, like ordinary citizens, most leaders are convinced that inclusion is already realised.

3.2 Improving Governance Capacity (Outcome 2)

In its problem statement, the ToC elaborated that village level governance is challenged by its unclear legal status, acknowledged hybridity14 and suggested among others that lack of role clarification between VC, elders and sub-committees hampers the efficiency of local institutions.The programme thus aimed at strengthening planning and management capacities of local authorities, while at the same time enhancing service delivery and conflict management. The following section presents the results of the evaluation of governance capacity. While the qualitative interviews assessed how village governance actually works, focusing on the role of elders and the VC, the survey results show if the programme has indeed improved local level governance, from the viewpoint of the citizens, with respect to overall service delivery and conflict management, and has thus strengthened leaders’ capacity to steer and manage development.

3.2.1 Coordination and Administration The qualitative research confirms that village-level institutions lack clear definitions of roles and responsibilities. Neither the local authorities themselves nor ordinary citizens were clear about the tasks or division of labour between the institutions. When asked what and who exactly constitutes the village government or village administration (ma- muul), villagers were quite divided in their views. In Somaliland (Erigabo), some villagers referred exclusively to formal institutions, such as the VC and the mayor, and interpreted administration as a representation of the Somaliland government. Reflecting this view, a woman in a village in Erigabo-district even insisted that her village does not know any forms of administration as ‘the ordinary people administer themselves’. Most villagers in both Somaliland and Puntland, however, were of the opinion that the village administra-

14 The term hybridity is not used in the ToC but introduced by the consultants referring to those sections in the document that elaborate on the mix of formal and customary mechanisms to conﬂict resolution. In the ToC document the relevance of both formal (i.e. recognised by law) and informal (’traditional’) governance structures in Somalia and Somaliland and the importance both play for delivering a broad range of governance services, among them conﬂict resolution, is emphasized.

48 tion consists of both the Village Council and non-state actors, such as elders or religious leaders. As outlined above, clan-affiliation constitutes the main principle for selecting Village Council representatives. There is a clear overlap between elders and the VC, as the VC in many villages, both implementation and non-implementation, includes elders.15 Most interviewees, including local authorities and citizens, outlined that elders and VC members co-operate in the running of everyday affairs. Some respondents saw the VC mainly as institution in support of the elders, while others referred to complementary and co-operative roles between them. Accordingly, elders deal with conflicts, work for social cohesion and represent the interest of the clans, while the village administration provides a link to the district and to international organisations. However, other interviewees insisted that elders, and not the village administration, provide the more important link to higher level governance. Beyond representing clan-members in distress and in situations of conflict, elders also represent the interest of their clan group more widely. One high ranking elder (nabadoon) summarized this view:

The committee plays its part. But conﬂicts are always solved by the Odayaasha [elders] and Nabadoonada [clan leaders]. However, the committee works in facilitating negotiations and they work with the Nabadoonada and Odayaasha for ﬁnding sustainable solutions. They also take part in awareness raising, and they bring people together and they are good role examples.

All (!) interviewees acknowledged the importance of elders in regulating the daily affairs of the village and in particular in dealing with conflicts.16 One elder was quite explicit in a focus group discussion about the role and significance of elders compared to the VC:

Odayaashu (elders) are more important to the community. The nominated administration is ineﬀective, but Odayaal (elders) are the ones who have ﬁlled the gap. You cannot compare someone who is working with someone who has just a title but is doing nothing. For us our priority is Odayaal (elders).

While this view may be extreme, many interviewees, including those who stressed the complementary and co-operative relations between the village administration and elders, identiﬁed elders as the ﬁnal instance of governance and decision-making at the village level. This was often related to the collapse of the state and especially the state’s provision of security, which left people in need of organising security themselves. As one member of a development committee outlined:

Since the administration has no police, it is Odayaasha (Elders) who help in conflict mediations and they are the law enforcement. At this moment the chairman is in [village name] with a group of elders to mediate a conflict about water. 15 Just a reminder that the term ’elder’ (oday) is used rather flexibly and usually means respected man. It has no age threshold and even a young man in the 30s can be classified as oday. However, the particular status of an elder depends on different factors, among them his genealogical position, individual skills and personal reputation, and not least his individual and family wealth. These things contribute to whether he is viewed as important and if he is respected enough to negotiate for his clan constituency, for example in conflicts, or whether he has a voice in a committee. 16 This view was only not shared by two interviewees, one woman on one member of the VC, who identified the VC as most important institution for conflict management.

49 Interviewer: Which one is more important, Odayaasha or the administration? Odayaashu are more important.

Although interviewees in both implementation and non-implementation villages emphasized the importance of elders, most interviewees saw the main responsibility of the VC in the development of the village and in co-operation with international organisations. This was confirmed in the quantitative analysis, which found clear differences across implementation and non-implementation villages between governance responsibilities of different institutions. Classical development issues, such as road construction (Figure 105), provision of water (Figure 107) and education (Figure 110) were assigned to formal institutions by the majority of respondents. There was some ambiguity as to the precise goals of the role clarification aspects of the programme. It was not clear if the aim was a general strengthening of the formal institutional mechanisms relative to the traditional mechanisms, or if simply there was to be agreement across leaders and citizens that certain tasks were carried out by certain bodies. In the quantitative assessment, we looked for any of two patterns. First, the evaluation looked for a general preference for formal institutions in carrying out tasks relative to traditional institutions in implementation villages (relative to non-implementation villages). Second, we looked to see if responsibilities for specific tasks differed between implementation and non-implementation villages. Then we compared citizens’ views with leaders’ views on this specific task, looking for similar patterns of change in both. The quantitative assessment did find very clear and systematic differences between implementation and non-implementation villages. However, what the evaluation found did not conform to either of the patterns outlined above. In general, there was no significant difference between implementation and non-implementation villages in the proportion of citizens assigning tasks to formal relative to traditional institutions, nor was there a common change in views across citizens and leaders. Instead, the evaluation found evidence that Hogaan had a profound impact on how villagers define roles between formal village level and district level institutions. Significantly more respondents in implementation villages assign major institutional roles and responsibilities to the VC, and significantly fewer to the DC on every task. Accordingly, more people in implementation villages think that the VC is responsible for the delivery of social services (44/30%), security (34/16%), development (road: 27/19%, water: 38/25%), resource management (37/26%), representation of the community (32/21%) and even conflict resolution (21/9%) (See Figures in Appendix B.6). The two figures below on social services (Figure 30) and security provision (Figure 31) are provided as examples to show that the intervention caused a clear preference shift from the DC to the VC, while the number of people in implementation and non-implementation villages who assign these responsibility to customary institutions are fairly similar.

50 Elders: Social Services Elders: Social Services

Hogaan 20 (14,● 26)

2 (−5,● 8)

Non−imp 18 (11,● 26)

VC: Social Services VC: Social Services

Hogaan 44 (33,● 54)

14 (6, ●22)***

Non−imp 30 (20,● 40) significance a● negative DC: Social Services DC: Social Services a● none a● positive Hogaan 22 (19,● 24)

−14 (−21,● −6)***

Non−imp 35 (28,● 42)

RL: Social Services RL: Social Services

Hogaan 4 (2,● 8)

−2 (−7,● 1)

Non−imp 6 (2,● 13)

0% 20% 40% −20% −10% 0% 10% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 30: Institutional Responsibility for Social Services: Comparison Village Type

Elders: Provide Security Elders: Provide Security

Hogaan 22 (19,● 24)

−2 (−9,● 4)

Non−imp 24 (18,● 30)

VC: Provide Security VC: Provide Security

Hogaan 34 (22,● 46)

18 (11,● 24)***

Non−imp 16 (9,● 26) significance a● negative DC: Provide Security DC: Provide Security a● none a● positive Hogaan 31 (28,● 35)

−21 (−29,● −14)***

Non−imp 53 (44,● 60)

RL: Provide Security RL: Provide Security

Hogaan 3 (1,● 7)

0 (−4,● 2)

Non−imp 5 (1,● 9)

0% 20% 40% 60% −30% −20% −10% 0% 10% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 31: Institutions Responsible for Security: Comparison Village Type

This pattern is repeated with respect to other institutional responsibilities. Overall, respondents in implementation village are more inclined to attribute development and governance responsibilities to the VC. If data is further disaggregated, the programme had a particular impact on adults (>34 years) with respect to institutional responsibilities. Compared with youth (18-35), the difference between adult preferences on institutional responsibilities is significantly greater. In non-implementation villages, adults are more inclined to assign governance and development responsibilities to the DC, including the provision of social services, security, conflict resolution and water: while adults in implementation villages tend to assign these roles to the VC. The Hogaan intervention thus strengthened citizens’ confidence in formal village institutions (VC), and particularly enhanced the confidence of adults. In contrast to age, gender and affiliation to a majority or minority clan had no consistent effect on how people assign responsibilities (see Figures in Appendix: Further Re- sults). Overall, more men than women in non-implementation villages tend to attribute

51 responsibilities to the DC. The relation is the opposite with respect to the VC, in which fewer men seem to have confidence. However, evidence for a significant effect of the programme on gender can only be found with respect to security provision, where 16% more men than women in non-implementation villages attribute responsibility to the DC compared to only 4% more men in implementation villages. The answer pattern is even more inconclusive with respect to majority/minority clan affiliation of respondents. Overall there is no significant difference in how people with majority and majority clan background assign governance responsibilities, with the exception of provision of clean water. Here 22% more people with majority clan background were inclined to rely on district authorities (61% majority background compared to 38% from minority background), while in implementation villages there were no major differences between people from these clan groups with around 35/38% respondents supporting the DC. Minority clans in the same time decreased their support for elders by 13% and instead opt more, and in a fairly similar percentage (36%) than people from majority clans (39%) for the VC.

Institutional support for rights The shift from the DC to the VC in implementation villages can also be detected in the view of people on which institutions should support speciﬁc rights. With respect to the right to justice (Figure 108), the right to equality (Figure 109), education (Figure 109), health care (Figure 111), consultation (Figure112) and free speech (Figure 113), signiﬁ- cantly more villagers in implementation villages assigned supportive roles to the VC while the number of people attributing support to the DC shrinks. Customary authorities in contrast were in all cases rated fairly similarly in implementation and non-implementation villages (Figures in Appendix: Further Results). The only exception to this pattern is found in citizens’ views on the right to life (Figure 114). While 8% more respondents in implementation villages think that the VC should support the right to life, this increase appears to impact on the elders rather than on the DC. Although the majority of respondents in implementation and non-implementation villages (45/55%) retain the view that elders should support the right to life, the number in implementation villages was 10% lower. However, this single case should not be given too much emphasis, especially against the background of the very consistent shift in attitudes seen on all other measures of citizens’ views.

Local Leaders views on Roles and Responsibilities Despite the substantial change in citizens’ perceptions, the evaluation did not find evidence that the Hogaan programme substantially changed the view of roles and responsibilities of local leaders. In particular the public tendency in implementation villages to assign more responsibilities to the VC is not mirrored by the leaders, which confirms the conclusion that the impact of the programme was to a lesser extent on clarifying institutional roles and responsibilities than on changing citizens’ view on the institutional importance and service delivery capacities of the VC. However, from the viewpoint of local leaders, the intervention has very significantly improved their relation to the district government: 36% more local leaders in implementation than in non-implementation villages state that their relations with district authorities have improved and 33% more leaders in implementation villages are convinced that their relation to the government is improved.

52 Positive change in relations with DC over the past three years Positive change in relations with DC over the past three years

Hogaan 87 (77,● 92)

37 (23,● 52)***

Non−imp 50 (31,● 64)

significance Positive change in relations with Somaliland or Puntland governement over the pastPositive three change years in relations with Somaliland or Puntland governement over the past threea● positive years

Hogaan 79 (72,● 84)

33 (16,● 50)***

Non−imp 45 (27,● 61)

0% 25% 50% 75% 0% 10% 20% 30% 40% 50% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 32: Relations Village and Government: Comparison Village Type

To summarize, the evaluation did not find evidence that supports the conclusion that the programme has contributed to the clarification of role and responsibilities of village level institutions. We did not find evidence of a significant change in the division of responsibilities between formal and traditional institutions, nor did we find significant shifts in the attribution of roles and responsibilities consistent across citizens and leaders. Nev- ertheless, despite not changing the level of support for formal institutions as a whole, the programme did substantially shift citizens’ support for the VC relative to the DC. In implementation villages, citizens (but not leaders) assigned significantly more responsibility to the VC and correspondingly less to the DC, leaving the roles and responsibilities assigned to customary authorities largely unchanged.

3.2.2 Service Delivery The Hogaan programme aimed to build confidence in local governance institutions by involving those institutions, particularly the VC in delivering services. It is noteworthy that the satisfaction rate of villagers with service-delivering capacities of the formal local institutions is quite high. More than 70% of respondents in both implementation and non-implementation villages agree that the formal governing bodies provide services, can manage projects, and are transparent and accountable. The evaluation did not find evidence that Hogaan influenced villagers’ assessment of the service-delivering capacities of local institutions, with the exception of management capacity of the VC. 8% more people in implementation villages are convinced that the VC can manage projects. Despite the central focus of the Hogaan programme on enabling the VC and development committee to provide services, it had neither a significant effect on how villagers rate the service delivery capacity of locally governing institutions nor on how they rate the management capacities of both VC and DC.

53 Does VC Provide Services Does VC Provide Services

Hogaan 71 (55,● 81)

−3 (−9,● 3)

Non−imp 74 (58,● 83)

Can VC Manage Projects Can VC Manage Projects

Hogaan 79 (67,● 86)

8 (1, ●14)**

Non−imp 71 (59,● 80)

Does VC Manage Money Transparently and Accountably Does VC Manage Money Transparently and Accountably

Hogaan 75 (63,● 83)

5 (−2,● 12)

Non−imp 70 (58,● 79) significance Does DC Provide Services Does DC Provide Services a● none a● positive Hogaan 75 (64,● 83)

4 (−3,● 12)

Non−imp 71 (59,● 81)

Can DC Manage Projects Can DC Manage Projects

Hogaan 75 (64,● 82)

3 (−4,● 10)

Non−imp 72 (60,● 80)

Does DC Manage Money Transparently and Accountably Does DC Manage Money Transparently and Accountably

Hogaan 73 (64,● 79)

−2 (−8,● 5)

Non−imp 75 (64,● 81)

0% 25% 50% 75% −10% −5% 0% 5% 10% 15% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 33: Service Delivery: Comparison Village Type

How inclusive is the high satisfaction rate? With respect to age, young people under 34 in implementation villages are significantly more satisfied with services delivered by the VC and DC, and again significantly more youngsters are convinced that the formal institutions are able to manage projects, and that they do so in a transparent and accountable way. In non-implementation villages, the differences between young people and adults are not significant on most measures. However, the differences between implementation and non-implementation villages were not sufficiently large for us to be confident that there is a programme effect.

54 Does VC Provide Services Does VC Provide Services Does VC Provide Services

b. Hogaan 74 (59,● 83) Youth b. Hogaan 6 (2, 11)** ● difference b. Hogaan 68 (53,● 79) 35+ −3 (−14, 7) ●

a. Non−imp 79 (62,● 88) Youth a. Non−imp 9 (0, 18)* ● difference a. Non−imp 70 (54,● 80) 35+ Can VC Manage Projects Can VC Manage Projects Can VC Manage Projects

b. Hogaan 83 (71,● 89) Youth b. Hogaan 7 (3, 12)*** ● difference b. Hogaan 76 (63,● 84) 35+ 1 (−10, 11) ●

a. Non−imp 75 (60,● 84) Youth a. Non−imp 7 (−3, 16) ● difference a. Non−imp 68 (53,● 79) 35+ Does VC Manage Money Transparently and Accountably Does VC Manage Money Transparently and Accountably Does VC Manage Money Transparently and Accountably

b. Hogaan 80 (69,● 87) Youth b. Hogaan 8 (4, 13)*** ● difference b. Hogaan 72 (59,● 80) 35+ 7 (−4, 17) ●

a. Non−imp 72 (57,● 82) Youth a. Non−imp 2 (−7, 11) ● difference a. Non−imp 70 (55,● 80) 35+ significance Does DC Provide Services Does DC Provide Services Does DC Provide Services a● none

b. Hogaan 80 (68,● 87) Youth b. Hogaan 9 (5, 14)*** ● difference b. Hogaan 71 (58,● 80) 35+ 7 (−3, 18) ●

a. Non−imp 73 (58,● 82) Youth a. Non−imp 2 (−8, 11) ● difference a. Non−imp 71 (56,● 81) 35+ Can DC Manage Projects Can DC Manage Projects Can DC Manage Projects

b. Hogaan 80 (70,● 86) Youth b. Hogaan 9 (5, 14)*** ● difference b. Hogaan 71 (59,● 79) 35+ 9 (−2, 20) ●

a. Non−imp 73 (58,● 83) Youth a. Non−imp 1 (−10, 10) ● difference a. Non−imp 73 (58,● 81) 35+ Does DC Manage Money Transparently and Accountably Does DC Manage Money Transparently and Accountably Does DC Manage Money Transparently and Accountably

b. Hogaan 78 (70,● 84) Youth b. Hogaan 10 (5, 15)*** ● difference b. Hogaan 68 (59,● 75) 35+ 10 (−1, 20) ●

a. Non−imp 75 (63,● 83) Youth a. Non−imp 0 (−9, 10) ● difference a. Non−imp 75 (63,● 82) 35+

0% 25% 50% 75% −10% 0% 10% −10% 0% 10% 20% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 34: Service Delivery: Comparison Youth Adults

Gender and clan affiliation had no significant effect on the assessment of service and management capacities of local governments, and the programme found no evidence that the intervention had an impact on gendered or clan-based views of respondents. Men and women rate the capacities of local governing bodies relatively equally. Members of minority clans in both implementation and non-implementation villages are in some cases, especially relating to the DC, significantly less satisfied with the service and management capacities of the formal institutions, but we do not find a significant difference in these differences between implementation and non-implementation villages (cf. Appendix: Further Results).

3.2.3 Conflict Management A large number of villagers in both implementation (81%) and non-implementation villages (82%) are convinced that conflict resolution mechanisms are effective. However, 5% more people in implementation villages report that conflicts are resolved peacefully (94/89%).

55 Conflict Mechansims Effective Overall Conflict Mechansims Effective Overall

Hogaan 81 (69,● 89)

−1 (−6,● 4)

Non−imp 82 (70,● 90) significance Conflicts Resolved Peacefully Conflicts Resolved Peacefully a● none a● positive Hogaan 94 (92,● 96)

5 (0,● 11)*

Non−imp 89 (81,● 94)

0% 25% 50% 75% 100% −5% 0% 5% 10% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 35: Conﬂict Resolution: Comparison Village Type

The evaluation found no evidence for a significant difference in the resolution of most single issue conflicts (clan, water, grazing conflicts etc.), with the exception of leadership conflicts. 10% more respondents in implementation villages report that leadership conflicts are solved.

Resolution: Clan Issues Resolution: Clan Issues

Hogaan 90 (84,● 93)

2 (−4,● 8)

Non−imp 89 (80,● 93)

Resolution: Water Issues Resolution: Water Issues

Hogaan 91 (85,● 93)

3 (−4,● 11)

Non−imp 88 (78,● 93)

Resolution: Grazing Issues Resolution: Grazing Issues

Hogaan 88 (80,● 92)

−1 (−7,● 4)

Non−imp 90 (81,● 94) significance Resolution: Land Issues Resolution: Land Issues a● none a● positive Hogaan 86 (78,● 91)

−2 (−8,● 4)

Non−imp 89 (79,● 94)

Resolution: Charcoal Issues Resolution: Charcoal Issues

Hogaan 83 (74,● 88)

−2 (−9,● 6)

Non−imp 85 (75,● 91)

Resolution: Leadership Issues Resolution: Leadership Issues

Hogaan 89 (82,● 93)

10 (3,● 19)**

Non−imp 78 (67,● 86)

0% 25% 50% 75% −10% 0% 10% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 36: Conﬂict Issues: Comparison Village Type

This view corresponds with the self-assessment of local authorities. In general, on most conflict types, leaders in implementation and non-implementation villages are equally likely to report the successful resolution of conflicts. As in the citizen survey, leadership conflicts are the only conflict type where a significant difference is found. 16% more local leaders in implementation villages are convinced that leadership issues are successfully resolved, while there is no significant difference between them with respect to other types and issues of conflict. There is no evidence that age or gender had an overall effect on the assessment of conflict resolution and conflict management capacities (cf.Appendix: Further Results). The programme however had a positive effect on the inclusiveness of conflict resolution.

56 In accordance with the general perception of increased inclusiveness (Section 3.1.2 of this report), 9% more people in implementation villages are convinced that women’s issues (78/69%) and youth issues (77/68%) are considered in conﬂict resolution.

Women's Issues in Conflict Resolution Women's Issues in Conflict Resolution

Hogaan 79 (73,● 83)

9 (3, ●15)**

Non−imp 70 (60,● 76)

Youth Issues in Conflict Resolution Youth Issues in Conflict Resolution

Hogaan 77 (69,● 82) significance 9 (3, 16)*** ● ● a none ● Non−imp 68 (57,● 76) a positive

Minority Clan Issues in Conflict Resolution Minority Clan Issues in Conflict Resolution

Hogaan 76 (67,● 82)

4 (−2,● 10)

Non−imp 72 (61,● 80)

0% 20% 40% 60% 80% 0% 5% 10% 15% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 37: Inclusiveness of Conﬂict Resolution: Comparison Village Types

Clan did not affect villagers assessment of conflict resolution, with the exception of the resolution of clan issues. Here the programme had a significant impact on the satisfaction of minority clans. In non-implementation villages, 11% fewer people from minority clans (81%) think that clan-issues are resolved, while the percentage of minority clans who believe that clan issues are resolved rose to 90% in implementation villages and is thus nearly equal to people with majority clan affiliation (91%).

3.2.4 Trust in Village Governance The evaluation also addressed the level of trust of citizens in diverse institutions of governance. In general, levels of trust in customary institutions, elders and religious authorities are very high. Trust rates exceed 90% in both implementation and non-implementation villages. The Hogaan programme did not aﬀect these high levels of trust in customary institutions. Trust in formal institutions, that is Village and District Councils, was considerably lower than trust in customary institutions. However, trust in formal institutions was signiﬁcantly higher in implementation villages than in non-implementation villages with respect to the VC (9%) and to the DC (8%). Overall 88% of respondents stated that they trust both the VC and DC, compared to 79% in non-implementation villages. This provides evidence that Hogaan increased citizens’ trust in formal governance institutions at village and district level.

57 Trust VC Trust VC

Hogaan 89 (79,● 93)

9 (3, 15)***●

Non−imp 79 (67,● 87)

Trust DC Trust DC

Hogaan 88 (77,● 93)

8 (3, ●15)**

Non−imp 79 (66,● 88) significance Trust Elders Trust Elders a● none a● positive Hogaan 94 (92,● 95)

2 (−3,● 8)

Non−imp 92 (84,● 96)

Trust Religious Leaders Trust Religious Leaders

Hogaan 96 (94,● 97)

2 (−1,● 7)

Non−imp 94 (88,● 97)

0% 25% 50% 75% 100% 0% 5% 10% 15% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 38: Trust: Comparison Village Type

Inclusiveness of trust measures: gender, age, clan Women are generally more likely than men to trust formal authorities. In both implementation and non-implementation villages 5% more women than men express trust in the VC and DC. The general increase in trust in implementation villages is relatively equally shared by men and women, and Hogaan did thus not have a gendered effect on trust measures. There are no significant gender differences in trust in customary institutions, with men and women in both implementation and non-implementation villages expressing extremely high levels of trust in elders and religious authorities, exceeding 90% (Figure 62). Although youth has the general reputation of being most critical and potentially re- bellious, the evaluation found that adults over 34 years in Somaliland and Puntland are more sceptical towards formal institutions while customary institutions are widely trusted by both youth and adults. In implementation villages over 91% young people under 35 years express their trust in formal institutions compared to 86% of adults trusting the VC and 85% of adults trusting the VC. Although the discrepancy between youth and adults in non-implementation villages is smaller and not significant, the evaluation found no evidence that Hogaan had a particular effect on young people with respect to trust. Customary institutions are widely (>90%) and equally trusted by both youth and adults (Figure 83). Clan affiliation of respondents had no significant effect on governance trust measures. A nearly equal percentage of people from minority and majority clans express their trust towards formal as well as customary institutions (Figure 73).

Conclusion: Hogaan’s Impact on Local Governance

There is evidence that Hogaan had an impact on local level governance. In particular, there was a clear increase in citizens’ trust and conﬁdence in the capacities of the Village Council. Citizens in implementation villages were more likely to attribute responsibility for all services and the protection of all rights to the VC (ranging from security provision, development and service delivery through to resource management and conﬂict resolu-

58 tion). By itself, this increase in confidence in the VC conforms to the general CDD/R aim of increasing villagers’ self-reliance, and in the case of Hogaan especially to strengthen their ability to self-govern. However, this shift in each case comes at the cost of the DC and therefore does not support villagers’ confidence in the governance capacities and responsibilities of the state. Despite this, the evaluation did not find evidence of adverse effects on vertical linkages between the villages and districts or regions. While villagers seem to attribute more responsibilities and tasks to the VC and fewer to the DC this does not translate to a lack of trust or confidence of citizens in the district authorities. At the same time, local leaders in implementation villages overwhelmingly report that they have improved their relations with the district.17 Although the programme shifted citizens’ (but not leaders’) perceptions of governance responsibilities from the district to the village, the evaluation did not consider this as clarification of roles and responsibilities of the different village institutions. We did not find evidence of either a systematic shift towards formal institutions or of co-ordinated shift of both citizens’ and leaders’ views. The evaluation also did not find evidence that Hogaan strengthened the service delivery capacity of local authorities. Respondents in implementation villages rated service delivery, planning and management capacities of local authorities fairly similarly to those in non-implementation villages. Given the programme’s support for at least two social infrastructure projects in each village (e.g. community centre, school, etc.), this result is surprising. There is also no evidence that the programme had an impact on conflicts with the exception of leadership conflicts. As violence in Somalia and Somaliland is often related to leadership conflicts, the relevance of this contribution should nonetheless be acknowledged.

3.3 The Survey Experiment

It is sometimes argued that answers to direct survey questions may be misleading because they tend to elicit responses which are thought to be desirable. Further, preferences given in answer to direct survey questions may actually be preferences for some other (implicit) outcome. Both these factors might tend to inflate responses to the evaluation survey in line with the broad outlook of international organisations. For example, if asked whether a project should be organised by the VC or by the Clan, respondents may claim that they prefer a project organised by the VC because they directly think that this is the desired answer. They might reach the same answer because their actual preference is for projects which are funded by international organisations, and they think that a project organised by the VC has a better prospect of receiving funding. Discrete choice experiments are designed to address these sources of bias. In these experiments respondents are asked to choose between options with a number of different attributes. The inclusion of a number of different attributes, which may differ in social desirability, greatly reduces the tendency towards desirable answers. At the same time bias from factors that may be implicitly driving choice (such as sources of funding) can be examined by including these factors as an explicit attribute. The evaluation overall sought to examine if the intervention had an impact on villagers preferences and, if so, if the intervention strengthened citizens’ confidence in local

17 We also acknowledge the GPC interpretation of the shift from the VC to the DC which they understand as reﬂecting the strengthening of the link between the VC and the DC as villagers now turn directly to the VC

59 institutions. In order to examine if our regular survey responses were mainly driven either directly by the tendency to give socially desirable answers or indirectly by giving preferences which were believed to be more likely to be in line with the funding priorities of international organisations, the evaluation team conducted a survey experiment. The experiment aimed at gauging practical preferences of villagers and local authorities around the organization of projects, with a number of diﬀerent attributes, one of which was the source of funding. The experiment was guided by Hypothesis 1: that the intervention has altered citizens’ preference on the actors responsible for managing and implementing development projects

Hypothesis 2: that respondents in implementation villages articulate preferences towards the VC Taking the example of building a water catchment to prevent eﬀects of recurrent droughts, respondents were asked to express their project preferences. They could select between two projects that diﬀered in the actor that organised the building and the actor responsible for maintenance. The project thus had each two attributes, built by and maintained by, with randomly varying answers for each attribute. Answers included

Build by: the water catchment will be built by • an enthusiastic group of villagers.

• clan elders

• the Village Council

• the District Commission and Government

• an International Organisation

Maintained by: the water catchment will be maintained • as need arises

• by your clan

• by the Village Council In order to assess if financial information has an additional or maybe the prime influence on project choices, interviewees were given differing information about the financing of the project. Each interviewee was given the same financial information for all projects, so finance attribute did not vary between projects. There were three different levels to the finance attribute: • Finances were not mentioned in describing the projects

• An international organisation provides the funding

• An international organisation provides funding, but citizens are required to contribute 20%.

60 3.3.1 Findings The results indicate the impact of diﬀerent answers for both attributes on project choice. Answers on built by’ question are compared with the baseline category, an enthusiastic group of villagers. The baseline category for maintenance is as needs arises’, thus that the water catchment is maintained as needs arises. Figure 39 shows the main results of the project preferences for citizens (Figure 39a) and local leaders (Figure 39b) in implementation villages. There is some consistent structure to citizens’ responses with respect to the question of who should be responsible for building the catchment. The least favoured option is by an enthusiastic group of villagers. A project built by the VC is 9% more likely to be selected than one built by the enthusiastic villagers. However, citizens had less clear preferences between other options, with projects built by international organisations, DC and clan elders are all about 14% more likely to be selected. With respect to maintenance of the catchment, respondents were least likely to choose a project which is maintained as need arises. Compared to this, the option that the clan maintains the catchment was 4% more likely to be selected. However, most people were in favour of the VC, an option that was 10% more likely to be selected.

households leadership v. enthusiasts v. enthusiasts 15(11, 18) 17(10, 24) build build clan clan v. enthusiasts v. enthusiasts

14(11, 18) DC/Gov 32(24, 39) DC/Gov build build v. enthusiasts v. enthusiasts 13(10, 17) 13(5, 20) int.org. int.org. build build label label v. enthusiasts v. enthusiasts 9(5, 12) 16(8, 24) build build VC VC maintain maintain

4(1, 6) v. need −6(−13, 0) v. need clan clan maintain maintain

10(8, 13) v. need 3(−3, 9) v. need VC VC

0% 5% 10% 15% −10% 0% 10% 20% 30% 40% est est

coef.type build maintain coef.type build maintain

(a) Hogaan Households (b) Hogaan Leaders Figure 39: Citizens and Leaders Preferences in Hogaan Villages from Survey Experiment

Local leaders (Figure 39b) also articulated clear preferences. As in the household survey, the least favoured option was that the catchment is built by a group of enthusiastic villagers. They had an approximately equal preference for the international organisation, my clan and the VC. Compared to the baseline category, these options were about 15% more likely to be selected. However, village leaders’ clear preference was for projects built by the DC and the Government, an option that leaders were 32% more likely to

61 select than projects built by an enthusiastic group of villagers. In contrast to the citizens, village leaders had no clear preference with respect to maintenance. This indicated that their selection was mainly driven by the building of the catchment while the maintenance of the water catchment did not signiﬁcantly aﬀect leaders’ project choice.

3.3.2 Comparison with Non-implementation Villages As shown in Figure 40 there is no signiﬁcant diﬀerence between the project and the comparison villages neither with respect to who should build the project nor with respect to maintenance.

hogaan villagers non−implementation villagers maintain 4(1, 6) 2(−5, 9) v. need clan maintain 10(8, 13) 18(11, 25) v. need VC v. villagers 15(11, 18) 19(10, 27) build clan v. villagers label 14(11, 18) 20(12, 29) DC/Gov build v. villagers

13(10, 17) 18(9, 27) int.org. build v. villagers 9(5, 12) 14(6, 22) build VC

0% 10% 20% 30% 0% 10% 20% 30% est

coef.type build maintain

Figure 40: Citizens Preferences from Survey Experiment, Comparison of Hogaan and Non- implementation Villages

3.3.3 Impact of Financial Information on Project Preferences The second analysis is driven by the question: did ﬁnancial information change preferences, and if so in which direction?

62 hh b:clan v. villagers 13(9, 17) with info

20(15, 26) no info b:DC/Gov v. villagers

12(8, 16) with info

20(14, 26) no info b:int.org. v. villagers

10(5, 14) with info

23(18, 29) no info b:VC v villagers label 7(3, 11) with info

14(8, 20) no info

3(0, 6) m:clan v. need with info

4(0, 8) no info

12(9, 15) m:VC v. need with info

10(6, 15) no info

0% 10% 20% 30% est

coef.type build maintain

Figure 41: Citizens Preferences with and without Financial Information from Survey Experiment

The results shown in Figure 41 provide support for the view that financial information changes citizens’ preferences. When villagers were informed about where the finances for the project were coming from, they were significantly more likely to select enthusiastic villagers for the building of the catchment, albeit that the enthusiasts remained the least favoured option. Compared to the enthusiasts, people were still more likely to choose clan elders (13% more likely than enthusiasts), the DC (12%), or an international organisation (10%). The VC remained the second least preferred choice, but was still 7% more likely to be chosen than the enthusiasts. There was a particularly large shift towards a preference for international organisations building the project when villagers were not provided with financial information. The provision of financial information did not affect the preferences of local leaders. The different specification of financial arrangements, whether all finance was provided by an international organization or that a contribution was required had no effect on the answers of either citizens or village leaders.

Conclusion: Summary of survey experiment results

The overall pattern of responses shows that there are some important differences between village leaders and citizens. In terms of project organization, village leaders more clearly prefer projects organized by formal institutions, and particularly by the DC and government. At the same time, citizens’ preferences are affected by information about project maintenance whilst village leaders’ preferences are not. We found evidence that project preferences were affected by the provision of financial information. Projects implemented by international organisations are more likely to be chosen when no information about

63 funding is given, which suggests that villagers prefer to organise projects without external support, but lack resources to do so. However, we did not find evidence that the programme significantly affected project choice, and certainly not that it enhanced confidence of villagers in locally governing institutions, with preferences for both VC building and maintenance if anything slightly higher in non-implementation villages. This therefore increases our confidence in the validity of the main evaluation.

3.4 Training Eﬀects

As stated in Section 1.3 the provision of training, here understood broadly and ranging from consultations and content specific workshops to learning by doing, are considered a core mechanism of CDD/R. In order to determine effects of training on participation and governance measures, respondents were asked if they participated in training provided by the programme, and if so in what training. For the analysis of training effects, respondents were clustered into five different types of training recipients, and their responses in the analysis were compared with response patterns in non-implementation villages. The groups are:

1. CAP: People engaged in Community Action Planning (238 respondents)

2. Train: people who engaged in speciﬁc training sessions but not in CAP (136 respondents)

3. Community Entry: people who engaged in Community Entry but not CAP or training (109 respondents)

4. Other: people who were engaged with the project in some other way, but not in CAP, training, or Community Entry (99 respondents)

5. None: Respondents in Hogaan Villages who did not receive any type of training and were not engaged with the project (794 respondents)

6. Non-implementation: people in non-implementation villages (228).

The analysis of training eﬀects shows clear patterns. People who have received training are signiﬁcantly more likely to attend community and VC meetings, and are more involved in village planning. The highest attendance rates are displayed by respondents who participated in CAP. Interestingly, people who have not participated in trainings are even 9-10% less likely to attend such meetings than people in non-implementation villages (Figure 42).

64 Currently Attends Community Meetings Currently Attends Community Meetings

71 (56,● 81) 18 (6, 27)*** Other Other ● 91 (81, 95) CAP ● 37 (27, 46)*** CAP ● 87 (76, 93) Training ● 33 (24,● 42)*** 84 (73, 91) Training Community Entry ● 30 (21, 40)*** 44 (34, 54) Community Entry ● None: Hogaan ●

53 (42, 63) −9 (−17,● 0)* Non−imp ● None: Hogaan

Should Attend Community Meetings Should Attend Community Meetings

89 (66,● 97) 8 (1, 18)* Other Other ● 97 (85, 99) CAP ● 17 (6, 31)*** CAP ● 95 (78, 99) Training ● 14 (5, ●27)*** 94 (73, 99) Training Community Entry ● 13 (4, 25)*** 76 (53, 92) Community Entry ● None: Hogaan ● −2 (−10, 4) 79 (56, 93) ● Non−imp ● None: Hogaan significance a● negative Attend VC Meeting This Year Attend VC Meeting This Year a● none 91 (78, 96) 9 (−1, 18) a● positive Other ● Other ● 99 (94, 100) CAP ● 17 (10, 26)*** CAP ● 97 (90, 99) Training ● 15 (7, 24)*** Training ● 95 (86, 98) Community Entry ● 13 (5, 22)** 72 (60, 79) Community Entry ● None: Hogaan ● −10 (−18, −1)* 81 (70, 89) ● Non−imp ● None: Hogaan

Respondent Involved in Planning Process Respondent Involved in Planning Process 79 (62, 88) 20 (10, 30)*** Other ● Other ●

89 (75,● 95) 30 (21, 40)*** CAP CAP ● 86 (72, 93) Training ● 28 (18, 37)*** Training ● 71 (53, 82) Community Entry ● 13 (3, 22)** 55 (37, 68) Community Entry ● None: Hogaan ● −3 (−10, 4) 58 (42, 71) ● Non−imp ● None: Hogaan

0% 25% 50% 75% 100% −20% 0% 20% 40% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 42: Training Eﬀects on Participation (1/2)

Training participants also interact significantly more with locally governing institutions: over 30% more with the VC; over 28% more with the DC; over 26% more with elders, and over 21% more with religious authorities. While people without training in implementation villages had approximately the same level of interaction with formal institutions (VC and DC) than people in the non-implementation villages, significantly fewer people (- 9%) without training interacted with customary authorities. They are also most critical towards the responsiveness of customary institutions, albeit they overwhelmingly rated elders and religious leaders as accessible and do not differ in this view from training participants or from people in non-implementation villages (Figure 43).

65 Currently Attends Community Meetings Currently Attends Community Meetings 71 (57, 82) ● 17 (7, 28)*** Other Other ● 91 (80, 95) CAP ● 36 (27, 46)*** CAP ● 87 (77, 93) Training ● 33 (24, 42)*** 85 (73, 92) Training ● Community Entry ● 31 (21, 40)*** 45 (34, 55) Community Entry ● None: Hogaan ● −9 (−17, −1)* 54 (42, 64) ● Non−implementation ● None: Hogaan

Should Attend Community Meetings Should Attend Community Meetings 89 (65, 97) ● 8 (0, 19)* Other Other ● 97 (85, 99) CAP ● 17 (6, 32)*** CAP ● 95 (78, 99) Training ● 14 (5, 27)*** 94 (74, 99) Training ● Community Entry ● 13 (4, 26)*** 77 (53, 91) Community Entry ● None: Hogaan ● 79 (56, 93) −3 (−10, 4) ● None: Hogaan ● Non−implementation significance a● negative Attend VC Meeting This Year Attend VC Meeting This Year a● none 91 (80, 96) ● positive ● 9 (0, 19) a Other Other ● 99 (95, 100) CAP ● 17 (9, 26)*** CAP ● 97 (90, 99) Training ● 15 (8, 24)*** 95 (87, 98) Training ● Community Entry ● 13 (5, 22)** 73 (61, 80) Community Entry ● None: Hogaan ● −9 (−18, 0)* 82 (70, 89) ● Non−implementation ● None: Hogaan

Respondent Involved in Planning Process Respondent Involved in Planning Process 78 (61, 87) ● 20 (10, 29)*** Other Other ● 89 (74, 94) CAP ● 30 (21, 40)*** CAP ● 86 (70, 93) Training ● 28 (19, 37)*** 70 (52, 82) Training ● Community Entry ● 13 (3, 22)** 54 (37, 68) Community Entry ● None: Hogaan ● −3 (−10, 5) 58 (40, 70) ● Non−implementation ● None: Hogaan

0% 25% 50% 75% 100% −20% 0% 20% 40% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 43: Training Eﬀects on Participation (2/2)

Respondents who have received training are, in general, also more satisﬁed with service delivery and management capacities of the VC and DC, again with the exception of people who merely participated in community entry. Respondents from the latter group have comparable views to people without training from implementation villages and respondents from non-implementation villages.

66 Does VC Provide Services Does VC Provide Services

83 (67,● 91) 8 (0, 16) Other Other ● 84 (70, 91) CAP ● 9 (2, 18)** CAP ● 71 (52, 80) Training ● −4 (−13,● 4) 72 (56, 83) Training Community Entry ● −2 (−12, 7) 68 (52, 78) Community Entry ● None: Hogaan ●

74 (59, 83) −6 (−13,● 0) Non−imp ● None: Hogaan

Can VC Manage Projects Can VC Manage Projects

88 (77,● 94) 17 (8, 27)*** Other Other ● 96 (89, 98) CAP ● 25 (17, 33)*** CAP ● 88 (78, 93) Training ● 17 (9, ●26)*** 79 (66, 87) Training Community Entry ● 8 (−1, 17) 73 (62, 80) Community Entry ● None: Hogaan ●

71 (58, 79) 2 (−4,● 10) Non−imp ● None: Hogaan

Does VC Manage Money Transparently and Accountably Does VC Manage Money Transparently and Accountably

85 (73,● 92) 14 (5, 23)** Other Other ● 94 (88, 97) CAP ● 23 (15, 32)*** CAP ● 88 (78, 93) Training ● 16 (8, ●24)*** 73 (60, 82) Training Community Entry ● 2 (−7, 12) 69 (59, 77) Community Entry ● None: Hogaan ● −2 (−9, 5) 71 (59, 79) ● Non−imp ● None: Hogaan significance a● negative Does DC Provide Services Does DC Provide Services a● none 86 (73, 92) 14 (5, 23)** a● positive Other ● Other ● 89 (79, 93) CAP ● 17 (9, 25)*** CAP ● 87 (76, 92) Training ● 15 (7, 23)*** Training ● 70 (56, 79) Community Entry ● −1 (−12, 8) 71 (60, 78) Community Entry ● None: Hogaan ● −1 (−8, 7) 71 (59, 80) ● Non−imp ● None: Hogaan

Can DC Manage Projects Can DC Manage Projects 86 (75, 92) 14 (4, 23)** Other ● Other ● 87 (79, 92) CAP ● 15 (8, 24)*** CAP ● 83 (74, 89) Training ● 11 (2, 19)** Training ● 75 (63, 83) Community Entry ● 3 (−6, 12) 70 (61, 77) Community Entry ● None: Hogaan ● −1 (−9, 6) 72 (61, 79) ● Non−imp ● None: Hogaan

Does DC Manage Money Transparently and Accountably Does DC Manage Money Transparently and Accountably 88 (80, 93) 14 (5, 22)*** Other ● Other ●

88 (81,● 92) 13 (6, 21)*** CAP CAP ● 79 (69, 85) Training ● 4 (−5, 13) Training ● 71 (60, 78) Community Entry ● −3 (−13, 5) 66 (60, 72) Community Entry ● None: Hogaan ● −7 (−15, −1)* 74 (66, 80) ● Non−imp ● None: Hogaan

0% 25% 50% 75% 100% −10% 0% 10% 20% 30% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 44: Training Eﬀects on Service Delivery

With respect to perceptions of inclusion of marginalised groups, people who have merely participated in community entry tend to be the most critical. While not signiﬁcantly diﬀer- ent from respondents in non-implementation villages, people who participated exclusively in Hogaan’s community entry are least convinced of all respondents in implementation villages (including people without Hogaan training) that village planning actually includes women, youth or caste members. Again members from marginalised groups who only participated in community entry are least convinced that their own group is involved in planning (Figures 51-53). Another similar pattern with respect to community entry participants can be detected with respect to views on equality norms. Respondents from this group are least convinced about equality norms. Compared to respondents from

67 non-implementation villages, significantly fewer community entry participants think that women(-21%), youth (-26%), and caste groups (-12%) should have equal rights. The patterns detected are in some respects clearly contrary to the aims of community entry and may indicate a selection effect, that is that those people who were involved in community entry but did not receive any other form of training were specifically excluded either because they hold particular views and behaviours which were identified as not productive for further engagement during community entry, or they have self-excluded due to their unfavourable views towards the programme or for other reasons. People who have participated in the CAP process, on the other hand, are more likely to think that equal rights of marginalised groups are already realised, but nonetheless not more convinced that youth, women or outcasts actually should have equal rights.

Do Women Have a Right to Protection Under the Law Do Women Have a Right to Protection Under the Law

95 (87,● 98) −3 (−9, 1) Other Other ● 100 (0, 100) CAP ● 0 (−100, 4) CAP ● 0 (0, 100) Training ● −94 (−100, 4) 0 (0, 100) Training ● Community Entry ● −90 (−100, 4) 98 (95, 99) Community Entry ● None: Hogaan ●

99 (94, 100) −1 (−3,● 3) Non−imp ● None: Hogaan Should Women Have Same Rights as Men Should Women Have Same Rights as Men 74 (63, 81) −3 (−13, 7) Other ● Other ● 74 (66, 80) CAP ● −3 (−11, 5) CAP ● 70 (61, 77) significance Training ● a● negative −6 (−15,● 2) Training ● 55 (45, 64) a none Community Entry ● −21 (−32, −11)*** a● positive 70 (64, 74) Community Entry ● None: Hogaan ● −7 (−14, 1) 77 (69, 82) ● Non−imp ● None: Hogaan Do Women Have the Same Rights as Men Do Women Have the Same Rights as Men 66 (55, 74) 14 (3, 24)** Other ● Other ●

60 (51,● 67) 8 (−1, 17) CAP CAP ● 57 (47, 66) Training ● 5 (−4, 16) Training ● 34 (25,● 44) Community Entry −17 (−27, −6)*** 47 (40, 53) Community Entry ● None: Hogaan ● −5 (−13, 3) 52 (43, 60) ● Non−imp ● None: Hogaan

0% 25% 50% 75% 100% −80% −40% 0% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 45: Training Eﬀects on Women’s Rights

A similar pattern is also detected on institutional inclusiveness. Asked which institutions consider priorities of marginalised groups, community entry participants are much more critical of all institutions, but particularly when assessing inclusiveness of elders. The same applies to trust measures. They can be contrasted with people who have participated in the CAP, and who are significantly more likely to believe that elders and formal institutions actually consider viewpoints of all marginalised groups. The CAP process thus either seems to have had a large impact on how participants rate inclusion, or those who believed that institutions were inclusive were particularly selected to participate in the CAP, or some combination of the two. The same applies with respect to trust in local institutions. Again, people who participated in trainings, with the exception of community entry participants, are far more likely to trust the formal institutions. Here, however training and participation did not affect trust in customary institutions. The general negative pattern with respect to people who participated merely in community entry is again strongly suggestive of a selection effect.

68 Elders: Priority Women Elders: Priority Women

93 (76,● 98) 4 (−2, 12) Other Other ● 96 (84, 99) CAP ● 7 (2, 17)*** CAP ● 83 (64, 94) Training ● −4 (−12,● 2) 75 (53, 88) Training Community Entry ● −13 (−23, −5)*** 83 (64, 93) Community Entry ● None: Hogaan ●

88 (71, 96) −5 (−10,● −1)* Non−imp ● None: Hogaan

Elders: Priority Youth Elders: Priority Youth

91 (71,● 97) 4 (−2, 12) Other Other ● 94 (79, 98) CAP ● 8 (2, 16)** CAP ● 90 (72, 97) Training ● 4 (−1,● 11) 67 (45, 82) Training Community Entry ● −17 (−28, −9)*** 82 (63, 92) Community Entry ● None: Hogaan ●

86 (68, 94) −3 (−9,● 1) Non−imp ● None: Hogaan

Elders: Priority Outcast Clans Elders: Priority Outcast Clans

79 (55,● 93) 7 (−2, 17) Other Other ● 85 (63, 95) CAP ● 12 (5, 22)*** CAP ● 76 (52, 90) Training ● 4 (−5,● 13) 54 (36, 73) Training Community Entry ● −16 (−26, −6)*** 73 (50, 88) Community Entry ● None: Hogaan ● 1 (−6, 10) 72 (49, 88) ● Non−imp ● None: Hogaan significance a● negative Relgious Leaders: Priority Women Relgious Leaders: Priority Women a● none 96 (84, 99) −3 (−12, 0) a● positive Other ● Other ● 98 (92, 100) CAP ● 0 (−4, 2) CAP ● 97 (86, 99) Training ● −2 (−9, 1) Training ● 96 (87, 99) Community Entry ● −2 (−9, 0) 93 (82, 97) Community Entry ● None: Hogaan ● −6 (−12, −2)*** 99 (95, 100) ● Non−imp ● None: Hogaan

Religious Leaders: Priority Youth Religious Leaders: Priority Youth 90 (77, 96) −5 (−14, 1) Other ● Other ● 96 (86, 99) CAP ● 1 (−4, 6) CAP ● 92 (80, 97) Training ● −2 (−9, 2) Training ● 95 (83, 98) Community Entry ● 0 (−6, 5) 90 (79, 95) Community Entry ● None: Hogaan ● −4 (−10, 0)* 95 (86, 98) ● Non−imp ● None: Hogaan

Religious Leaders: Priority Outcast Clans Religious Leaders: Priority Outcast Clans 83 (59, 96) −5 (−14, 2) Other ● Other ●

89 (68,● 98) 0 (−6, 7) CAP CAP ● 82 (57, 95) Training ● −6 (−16, 1) Training ● 74 (49, 91) Community Entry ● −14 (−25, −4)*** 80 (56, 94) Community Entry ● None: Hogaan ● −8 (−17, −1)* 89 (68, 98) ● Non−imp ● None: Hogaan

0% 25% 50% 75% 100% −30% −20% −10% 0% 10% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 46: Training Eﬀects on Institutional Inclusion (1/2)

69 VC: Priority Women VC: Priority Women

89 (70,● 97) 0 (−7, 7) Other Other ● 96 (81, 99) CAP ● 6 (1, 15)* CAP ● 91 (72, 98) Training ● 2 (−4,● 8) 83 (63, 94) Training Community Entry ● −5 (−12, 0) 83 (65, 93) Community Entry ● None: Hogaan ●

89 (71, 96) −5 (−11,● −2)** Non−imp ● None: Hogaan

VC: Priority Youth VC: Priority Youth

87 (67,● 95) 2 (−6, 9) Other Other ● 91 (73, 97) CAP ● 6 (0, 13)* CAP ● 93 (75, 98) Training ● 7 (1, ●15)** 78 (58, 91) Training Community Entry ● −6 (−13, 0) 81 (61, 91) Community Entry ● None: Hogaan ●

85 (66, 94) −4 (−9,● 0) Non−imp ● None: Hogaan

VC: Priority Outcast Clans VC: Priority Outcast Clans

77 (53,● 93) 4 (−3, 12) Other Other ● 86 (62, 96) CAP ● 12 (4, 21)*** CAP ● 76 (51, 92) Training ● 3 (−4,● 11) 64 (43, 85) Training Community Entry ● −7 (−15, 1) 73 (49, 90) Community Entry ● None: Hogaan ● 0 (−6, 7) 73 (49, 90) ● Non−imp ● None: Hogaan significance a● negative DC: Priority Women DC: Priority Women a● none 89 (69, 97) 2 (−5, 9) a● positive Other ● Other ● 93 (76, 98) CAP ● 6 (0, 14)* CAP ● 92 (75, 97) Training ● 4 (−1, 12) Training ● 85 (67, 95) Community Entry ● −1 (−9, 4) 82 (63, 93) Community Entry ● None: Hogaan ● −4 (−10, 1) 87 (69, 95) ● Non−imp ● None: Hogaan

DC: Priority Youth DC: Priority Youth 89 (71, 96) 4 (−3, 12) Other ● Other ● 91 (76, 97) CAP ● 7 (0, 14)* CAP ● 92 (76, 97) Training ● 7 (1, 15)* Training ● 77 (58, 89) Community Entry ● −7 (−15, 0) 80 (63, 91) Community Entry ● None: Hogaan ● −4 (−9, 1) 84 (67, 93) ● Non−imp ● None: Hogaan

DC: Priority Outcast Clans DC: Priority Outcast Clans 80 (56, 93) 6 (−3, 15) Other ● Other ●

84 (63,● 95) 10 (3, 19)** CAP CAP ● 77 (55, 91) Training ● 4 (−4, 12) Training ● 61 (42, 79) Community Entry ● −11 (−20, −3)** 71 (49, 87) Community Entry ● None: Hogaan ● −2 (−9, 5) 73 (52, 88) ● Non−imp ● None: Hogaan

0% 25% 50% 75% 100% −20% −10% 0% 10% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 47: Training Eﬀects on Institutional Inclusion (2/2)

3.4.1 Spillover The evaluation was not designed to assess spillover effects. However, the above analysis also provides evidence relevant to understanding spillover. The different groups can be thought of as either directly treated or affected through spillover by the programme. Directly Treated Groups: Including all of those who directly participated in the Hogaan programme (CAP, Training, Community Entry, Other). Within Project Spillover: Individuals in implementation villages who did not participate in the programme individually.

70 Between Village Spillover: Individuals in non-implementation villages.

Given the available data it is not possible to provide a watertight case either for or against the presence of spillover. Many complex patterns of spillover can be imagined, and spillover stands in a complex relationship with selection effects. Nevertheless, specific spillover mechanisms do create alternative expectations about the patterns of attitudes we would expect to observe between these different groups. Below we outline three basic mechanisms and resulting patterns; we then consider how these relate to our observations.

Complete spillover: expected observation on diﬀerent treatment groups: no observable diﬀerences between groups.

Partial spillover: based on features such as proximity and contact. Expected observation: gradient eﬀect, with greatest eﬀect on treated individuals, intermediate impact on untreated individuals in treated villages, lowest impact on individuals in untreated villages.

No spillover: impact of programme only on treated individuals. Expected observation: treated individuals significantly different than others, but no significant difference between untreated individuals in treated and untreated villages.

In general, the patterns in the observed data conform most closely to the expectations generated by the hypothesis of no spillover. There is a positive difference between treated individuals and untreated groups, but no positive difference between untreated individuals in Hogaan villages and those in non-implementation villages. Consider for example the 14 measures shown in Figures 42, 43, and 44. All of these conform most closely to the no spillover expectation. In every case some, or all, of the treated groups have significantly more positive responses, which rules out complete spillover. Neither are the observations consistent with partial spillover, as there is no significant positive difference between the untreated individuals in the Hogaan villages and those in non-implementation villages (there is no significant difference in nine cases, and a significant negative difference in five cases). We do not rule out the possibility of some spillover. However, we consider that there is moderately strong evidence against spillover being an explanation of the broad patterns found in the data.

Conclusion: Training or Selection Eﬀects?

The evaluation found a very clear relationship between participation in the programme and the positive opinions of respondents of village level governance. The more intensive engagement people have had with the programme, the more satisﬁed they are in general with village level governance. People who have participated in the CAP, that is citizens and local authorities who underwent the long process of identifying and prioritising needs and suggesting projects, were the most supportive of local governance institutions and procedure. It is possible that these views are caused by the programme. However, the existence of these relationships between training or engagement with the programme and high levels of satisfaction, does not by itself demonstrate causal direction. It is equally possible that holding these positive opinions of village level governance causes people to be selected for trainings (See Section 4.3). People who were involved in community entry, that is were informed by the international organisation about the programme and have

71 discussed it with them, but have not participated in any other trainings, show similar answer patterns than those who were not trained. This general pattern hints at a selection effect. Community entry people were either excluded from further engagement with Hogaan by relevant programme stakeholders or they have excluded themselves, be it because they were not convinced about the programme’s goals or methods or for any other reasons. An alternative explanation could be that Hogaan entry has sensitized people to the shortcomings of local government, and thus influenced a particular negative view on inclusion, decision-making and capacities of local government, which were not mitigated by later trainings. More important, is the finding that people who had not participated in programme activities hold attitudes and recalled behaviours that were very similar to those found in non-implementation villages, but different from those who had individually participated in the programme. On this basis it seems unlikely that this broad pattern of results is explained by any straightforward pattern of spillover effect of trainings to non-participants.

72 Chapter 4 Summary and Discussion of Findings

This chapter provides a summary of the detailed ﬁndings of the evaluation from the previous chapter with respect to participation, village governance, and training and selection. Following the summary, the chapter concludes with a systematic quantitative overview of the detailed quantitative ﬁndings, to address concerns about the cherry-picking of results.

4.1 Hogaan’s Impact on Participation

The evaluation found mixed evidence for programme effects on citizens’ participation in local level governance. The programme improved interaction of villagers with formal governing bodies and increased responsiveness of the Village Council and District Council (from the viewpoint of villagers). The evaluation however found no evidence that the intervention had an overall effect on other participation measures, such as participation of citizens in village planning or other decision-making processes. Additionally the capacity or willingness of citizens to raise demands was not improved by the Hogaan intervention. Given the priority and effort Hogaan puts into sensitizing villagers to demand services and, especially during village entry and the CAP process, to participate in planning, this result may be somewhat disappointing. Evidence that the Hogaan intervention increased participation of women, youth, or occupational caste groups is weak. Women and young people continue to have lower participation rates in meetings and interact much less with all local institutions. While the intervention does not appear to have strengthened participation of women or of young people in decision-making, there is evidence that it increased interaction between women and customary authorities, and strengthened young people’s ability to articulate demands. Due to the small number of caste respondents, the evaluation was not able to determine an intervention effect on occupational castes, but it established that caste members participate much less in meetings and are least involved in decision-making processes. One factor that might explain the limited impact of the programme on participation is the lack of clarification about which forms of participation were envisaged. With the collapse of the state, elders have taken over many governance functions in Somalia and Somaliland. Through the segmentary lineage (clan) system all Somalis are at least theoretically represented by elders, and therefore, perhaps indirectly, participate in the decision-making process. Consultation and information sharing are participatory mechanisms that are commonly practiced by elders. Hogaan seems to have built on these already widely available mechanisms and expanded them to formal institutions, especially to the Village Council. This had a positive effect on how citizens perceive formal governance at village and also at district level, but did not automatically promote citizens’ ‘voice’ and

73 ‘choice’ with respect to village planning or other forms of decision-making. The relation between the form of participation and the goal of increasing voice and choice of citizens needs additional elaboration. Increased interaction between citizens and the local institutions do not necessarily give citizens more voice or affect how they assess responsiveness of these institutions or how satisfied they are with them. Although, for example, women interact much less with the VC than men, they rate accessibility and responsiveness quite similarly. If anything, then, women are slightly more satisfied with the VC, despite being much less involved in village planning and decision-making. A similar pattern is evident with respect to youth. Their participation in meetings is lower and they interact significantly less with the VC than adults, but nonetheless have higher appreciation for the VC’s capacities. Although not intended by the programme (at least not explicitly so), the intervention had an effect on participation of minority clans in local governance. Governance in Somalia and Somaliland is largely clan based, and numerical strength of clans in the village is reflected in the number of representatives in the VC. Numerical strength is also mirrored in other forms of participation, such as contributions or benefit sharing. It is thus not surprising that clan minorities, here defined as people from numerically smaller clan groups in a particular location, have lower levels of participation and are less involved in decision- making than people from majority groups. The programme however had a positive effect and increased participation of people from minority clans in planning processes. It also increased their interaction with customary authorities. The programme had a large impact on the perception of villagers, and especially so on the perception of governance inclusiveness. While there is overall limited evidence that marginalized individuals are more likely to participate in decision-making, people in implementation villages are significantly more convinced than their counterparts in non- implementation villages that marginalised groups are included in decision-making and planning processes. This may well be a result of the advocacy and discussions initiated by Hogaan. The intervention also had an effect on the responsiveness of local institutions. Villagers in implementation villages were more convinced that their viewpoints are considered, especially by formal authorities. The programme did not have a positive impact on the promotion of equality norms, and may have negatively impacted on the norms surrounding gender and youth equality. It did, however, sensitize local leaders to priorities of marginalised groups, although, similarly to ordinary citizens, most leaders are convinced that inclusion is already realised.

4.2 Hogaan’s Impact on Village Level Governance

The Hogaan programme had a significant impact on local level governance. The evaluation found systematic evidence that the project has increased citizens’ reliance on the Village Council and also increased trust in it. Interestingly, this change is not fully mirrored in how citizens access the VC’s governance capacities. While they have a higher apprehension for the VC’s management capacity, citizens did not change their views on the VC’s ability to provide services or its accountability. With respect to conflict management, the Hogaan intervention had little overall impact but contributed to the resolution of leadership conflicts. The evaluation did not find evidence that the programme succeeded in clarifying governance roles and responsibilities, nor did it systematically change the balance between formal and traditional institutions. It did nevertheless expand citizens’ view of the remit

74 of the VC relative to the responsibilities of the district government and thus the state.1 While this shift is in line with the general CDD/R aim of increasing villagers’ confidence in self-government and abilities to self-govern, it also creates the risk of neglecting the responsibilities of the state. Given that governance is already ‘radically localized’ (Menkhaus) throughout Somalia and (although to a lesser extent) Somaliland, the evaluation team would like to hint at the risk of interventions to contribute to the further localisation of governance, and to the tension this could produce for the more general aim of statebuilding. However, it is important to emphasize that the evaluation did not find any evidence that would suggest other adverse effects on vertical links between village, district and region. While villagers seem to attribute more responsibilities and tasks to the VC and less to the DC, this does not translate to a lack of trust or confidence of citizens in the district authorities. On the contrary, respondents in Hogaan villages have higher levels of trust in the DC, although they assign less responsibilities to the district. More importantly local leaders in implementation villages overwhelmingly report that they have improved their relations with the district.

4.3 Training and Selection Eﬀects

There are very clear relationships between participation in the programme and positive opinions of village level governance. The more intensive engagement people have had with the programme the more satisﬁed they were in general with village level governance. It is possible that these relationships are caused by the programme. However, the existence of these relationships does not by itself demonstrate causal direction. It is equally possible that holding these positive opinions of village level governance causes people to be selected for training. While many people during the qualitative research emphasized for example that all segments of the village community among them men, women, youth, rich and poor are involved in village decision-making, further discussions revealed that they are often not directly involved but through representatives, mostly elders (See Section 3.1.1). Some interviewees equally emphasized that the village community participated in the Hogaan intervention. However, many villagers do not automatically feel entitled to participate in meetings or training, but wait until they are explicitly approached by a local leader, be it an elder, VC member or member of a development committee (and often enough they are invited because an international organisation demands it). Chances of people, and especially women and young people, to be invited seem to be higher if they are organised in community groups or other organisations, as elders or the VC tend to approach organisations when they mobilise people for speciﬁc activities. See the following interview sections that indicate that membership in an organisation can facilitate participation in the project. Often it is the chairpersons of community-based organisations who are invited to represent the wider group. A young man, for example, listed some of the people that participated in Hogaan initiated village planning sessions:

Those who participate included the chairman, and the secretary of the village, chairlady of the women organization, chairman of the youth organization and representatives from the business community and some of the Odayasha (elders).

1 It is noteworthy here that local leaders do not show the same tendency to shift responsibility to the VC.

75 Several interviewers indicated that membership in local organisations enables and en- courages participation. It also enhances the chances that own issues and concerns are addressed, and helps members of these groups to receive a share in the beneﬁts of the projects (participation in beneﬁts). For example, the following is an answer of an elder who was asked if youth and women participate in decision-making:

Yes, they have a youth organizations that is very well organized, they raised and collected contributions among themselves, and they took a big share in the construction of the Mother Child Health Centre. Interviewer: What about women? They have their organization as well, and they take part in all the necessary issues.

In order to assess if membership in a community based organisation (CBO) supports participation, the survey collected data on the organisational membership of respondents. The date provides very clear results: people who are organised were 30% more likely to participate in the Hogaan programme and 17% more likely to receive information about it.

Heard of Hogaan Project Heard of Hogaan Project

Not in a Group 68 (55,● 77)

−17 (−24,● −12)***

Group Member 86 (75,● 91)

significance Participated in the Hogaan Project Participated in the Hogaan Project a● negative

Not in a Group 41 (28,● 52)

−30 (−36,● −23)***

Group Member 71 (57,● 79)

0% 25% 50% 75% −30% −20% −10% 0% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 48: Group Membership and Participation in Hogaan Programme

Based on these findings, it is likely that the programme, instead of increasing inclusiveness, mainly worked through people who were already quite active and already participated to some extent in decisions that concerned the village or at least their particular group in the village. While this does not necessarily influence the project negatively, the lack of inclusion or the self-exclusion of other people needs further assessment. Self-exclusion is usually not addressed in participatory processes (Cornwall 2008: 279), but can lead to frustration and have negative impacts for the programme. This is also suggested by the finding that people who were involved in community entry but have not participated in any other trainings, are the least satisfied with local governance. While people who participated in the programme had very different views and be- haviour, people who did not participate at all in programme activities generally held attitudes and recalled behaviours which were very similar to those found in non-implementation villages. The evaluation could thus not detect a spillover effect to non-participants.

4.4 Systematic Overall Summary of Findings

In order to further address concerns about cherry-picking results we also present a summary analysis. Rather than looking at each outcome from the survey individually we group them together and examine whether there is an overall pattern across the questions

76 collectively. Table 4.1 provides a summary of the analysis which shows the overall effect and whether there was a significant change in the differential for women, minority clans and young people. The stars in the table indicate statistical significance of the model parameters.2

Overall Gender Clan Youth Participation: Formal *** - * - Participation: Customary - - ** * Participation Overall ** - ** - Perception Minorities Involved *** - - - Inclusion Overall - - - - Formal Institutions Strengthened - ** - - Service Provision - - - - Governance Trust ** - - - Overall Governance - - - - Issue Resolution - - - -

Table 4.1: Overall Summary of Quantitative Evaluation of Hogaan Project

Outcome 1: Participation We find systematic evidence for an effect on overall levels of participation, driven by greater engagement with formal institutions. We also find evidence that the project has led to significant improvement in the participation of minority clans (not outcast groups). We do not find systematic evidence that there has been an improvement in the participation of women or youth.

Outcome 1: Inclusion of women, youth and minority clans Looking at patterns of participation across all measures we do not ﬁnd systematic evidence for an eﬀect on the inclusion of women, young people or minority clans with village institutions.

Outcome 2: Co-ordination and Administration [through strengthening of formal governance institutions] Looking across all measures of co-ordination and administration we do not find systematic evidence of a programme effect on the overall population although women are significantly more supportive of formal institutions in implementation villages.

Outcome 2: Service Delivery We do not ﬁnd systematic evidence of a programme eﬀect.

Outcome 2: Conflict Management We do not find systematic evidence of a programme effect. While no overall improvement is registered, it is noteworthy that systematic evidence was found on the resolution of leadership conflicts.

2 The summary is calculated with a multi-level linear regression using the standard model on the mean number of positive responses across all questions in the category. The stars in the table indicate whether the inclusion of the relevant parameter leads to a signiﬁcant increase in the goodness of ﬁt of the model using an anova F-test.

77 Chapter 5 Recommendations Relating to the Theory of Change

To suggest an alternative ToC would imply suggesting alternative objectives (what), mechanisms (how) and/or to target or partner with an alternative group of people (who). It would address the questions of which objectives would have been more realistic, which mechanisms would have worked better and whether the programme should have chosen a diﬀerent target population. Instead of suggesting substantive alternatives to goal setting, implementation and partnerships, we suggest that it would be more fruitful to rethink how the three elements should be developed. Two broad and well established points underpin our recommendations:

1. Be context speciﬁc: Take the context into account in more detail in all three components of the ToC: the objectives (participation, inclusion, good governance), the mechanisms (participation, learning by doing and training) and the target group (communities, women, youth, minorities).

2. Enhance clarity through specificity (Cohen and Uphoff 1980:) Following Cohen and Upphoff (1980:214) we suggest that participatory approaches should specify the meaning of participation with respect to 1) the kind of participation 2) the question of who is supposed to participate and 3) how participation is supposed to occur (on which level, direct or indirect).

The following sections explain how these recommendations could be applied in the context of the Hogaan and other participatory programmes in Somalia and Somaliland.

Recommendations 1: Increase Focus without Losing Specificity Hogaan’s original ToC was reviewed and adapted with the intention of increasing programme focus and thus making impact manageable, and being able to monitor purposes (See Figures in AppendixC). The revised theory of change did reduce the number of outcomes from five to two. However, simply reducing the number of outcomes was not sufficient to focus the project in central respects. It is the case that some of the outcomes were explicitly dropped (particularly those relating to the relationship between the villages and the DC and regional policy change). However, many outputs and outcomes were now grouped together into much shorter, but as a consequence also less specific, goals. For example, in the original ToC, one outcome was the village level institutions demonstrating an ability to manage development funds in a transparent and equitable way, and another was village institution being able to advocate more effectively at the district level. In the replacement ToC one of the specified outcomes is that village institutions are able to plan, manage and co-ordinate. This is a shorter formulation, but is much more general, in fact

78 completely encompassing both of the outcomes from the original ToC and much more besides. Further, even where outcomes were dropped from the ToC the aspects connected to them remained in the programme. For example conflict management was originally considered a separate outcome. While it was dropped as an outcome in favour of a shift to accountability, transparency and responsiveness of governance, conflict resolution training still remained part of the mechanism suggested to contribute to this outcome. The ToC thus displays a high degree of vagueness with respect to both the outcomes themselves and the pathway or logic of change, as it is not clear which activities and outputs are actually meant to catalyse which type of change. Another example for this vagueness is the dropping of mechanisms for increasing responsiveness of local institutions. While the original ToC at least specified, for example, the output that citizens have access to formal mechanisms of engagement with village institutions’ (failing to clarify the kind of mechanisms or what is meant by formal), the new ToC does not provide any mechanism at all, and instead just states that citizens’ engagement and participation will increase responsiveness. To conclude, we fully agree with the attempt to increase focus (and also with the dropping of many outcomes including conflict management). However, we suggest that the programme needs to better specify the sequences of change and to fine-tune the logic in which the sequences are supposed to build on each other.

Recommendation 2: Adapt implementation practice to context The Hogaan programme used a fairly standardised CDD/R approach (see also section 1.3). It put considerable effort into working with different village groups and applied a broad range of participatory tools (Venn diagram, community mapping, seasonal calendar, income/poverty analysis, SWOT analysis etc.) to stimulate discussion about villagers’ needs and ways to address them. Despite these attempts, the outcomes of the needs analysis and priority development are nonetheless surprisingly uniform, and reveal a clear priority from villagers of the building of community centres. In all, villagers chose 28 community centres, 18 schools, 10 health centres and 9 Berkets. A second quite uniform approach was chosen for the distribution of block grants. The village was constructed as a uniform object of the programme, and an equal amount of money was released to each village independent of, for example, its size (the number of households varies considerable) or the prioritised project. Urban areas were divided into sections but despite obvious differences were otherwise treated in a very similar way to villages, and the standard implementation procedure applied with the delivery of an equal grant for each section of the village. The following table indicates the mismatch between the size of the funds and the size of the village (number of households). The fact that the numbers of households is difficult to estimate is already indicated in the table, as we have included three different estimates in the second row, one extracted from the programme documents (Prog Docs),1 the other from the satellite maps we used for the randomisation experiment, the last from estimates of the Survey Supervisors (Survey Super). We use the survey supervisors’ estimate to calculate the investment per household (app. Invest

1 Please note that for several villages we found quite different estimates in the documents. If available we chose to use the number in the CAP document. However, this points to an issue with the management of data. Crucial data is spread across a different reports, some of them very detailed and well developed, others clearly suffering from copy/paste errors. There would be considerable benefits to a more systematic approach to the storage and distribution of key administrative and management data.

79 per HH), not because we think that the supervisors’ estimates are the most accurate but because the supervisors provided estimates of every village.

Household Size Estimates Prog. Satellite Survey app. Invest District Village Docs Maps Super per HH Burtinle Jalam 1300 5000 $3 Galkayo Badweyn 3000 5000 $3 Galkayo Harfo 3000 3000 $6 Galkayo Bursaalax 3900 1696 1500 $11 Galkayo City (X/Garsoor) 1500 1000 $17 Galkayo City (Hormar4) 1500 1000 $17 Burtinle Kalabayr 380 900 $19 Galkayo City (Israc) 1500 800 $21 Burtinle Town (Wadajir) 580 $29 Burtinle Town (Horumar) 500 550 $31 Burtinle Town (Hawlwadag) 500 530 $32 Burtinle Town (Israac) 530 $32 Galkayo Balisbule 1800 453 500 $34 Erigabo Yube 360 203 500 $34 Erigabo Carmala 800 450 $38 Erigabo Mait 300 158 400 $42 Erigabo Godmobias 200 289 400 $42 Galkayo Sallah 320 $53 Galkayo Bayra 500 300 $57 Galkayo Agaran 700 200 $85 Erigabo City (Shacab) 200 $85 Erigabo Damal Hagare 320 273 200 $85 Erigabo Yuﬂe 260 150 $113 Erigabo City (Daya’an) 1000 150 $113 Erigabo City (Hafad Somale) 850 150 $113 Erigabo Godcaanood 350 60 120 $142 Erigabo Shimbirale 200 121 100 $170 Galkayo Balanbal 500 205 75 $227 Burtinle Maga’ley 350 70 $243 Burtinle Godobyar 90 62 $274 Burtinle Faratooyo 180 60 $283 Burtinle Meygagle 250 60 $283 Galkayo Malaasle 300 60 $283 Burtinle Ballidacar 80 55 $309 Burtinle Megag 70 54 $315 Burtinle Xaarxaar 200 103 50 $340 Burtinle Birta Dheer 100 94 50 $340 Galkayo Beer Dhagaxtuur 300 31 50 $340 Galkayo Roox 300 111 50 $340 Erigabo Daryle 100 50 $340 Erigabo Rugey 50 42 50 $340 Erigabo Goof 190 50 $340 Burtinle Hayanle 80 45 $378

80 Galkayo Qalanqal 150 45 $378 Burtinle Bahaley 75 74 40 $425 Burtinle Koryal 40 $425 Erigabo Jidali 250 111 40 $425 Erigabo Ardaa 100 35 $486 Burtinle Farjano small 33 $515 Burtinle Birrecaad 70 19 30 $567 Galkayo Dhagaxyo Cado 100 30 $567 Erigabo Biyoguduud 100 63 30 $567 Erigabo Buq 80 30 $567 Burtinle Lacle 17 25 $680 Galkayo Darusalam 80 25 $680 Erigabo Sibaayo 150 45 25 $680 Erigabo Dibqarax 80 45 24 $708 Galkayo Bilcil 100 42 20 $850 Galkayo Adaygebagebo 150 3 $5667

Table 5.1: Village Size Estimates

The size of some villages varies considerable due to seasonal migration. Some of the smaller villages are completely abandoned during the dry season as pastoralists move to other places in search for pasture and water. This explains, for example, the household size of Adaygebagebo in the above table (5.1). When the enumerators arrived in Adaygebagebo, they only found a few older men who were left behind to protect the property, while all other villagers were on the move. While most of them likely came back, seasonal migration impacts not only on village size, but also influences governance and politics more widely. The effects of seasonal migration on village governance, including service provision, management and maintenance of projects etc., were not discussed by the programme and are thus not considered in the change logic. The Hogaan programme should hence think through the impact of seasonal migration on village governance and the impact of mobility on governance more generally. The uniformity of approaches and outcomes raises some doubts about the community- driven nature of the programme. We acknowledge that the uniformity in the selection of projects can have many reasons: villages are likely to face similar problems; some of the prioritised projects can be supported by the implementing organisation under another programme title2; the grants were not sufficient for other projects; some projects are excluded by the organisations; a tendency of villages to copy each other; etc. However, the uniformity of outcomes points to more general or structural constraints between the need to standardize approaches (to get the work done or to scale up) and the localised nature of community-driven work. The donor related need to pre-design outcomes and approaches even if local knowledge is somehow limited creates another challenge for community-driven work. This general critique indicates the need to increase flexibility of programming and to adapt the implementation practice to the (changing) context.

Recommendation 3: Specify roles and responsibilities of local institutions The Hogaan programme conducted a participatory governance assessment and developed Government Improvement Plans (GIP) for most villages. The GIPs outline strength, weak-

2 This was given as a reason for the uniformity of projects by Care in Erigabo.

81 nesses, opportunities and threats (SWOT-analysis) of the VC and provide suggestions for improvements. The GIP did not include an analysis of all governing institutions, but concentrated on formal institutions, in particular the VC. Elders were not considered although elders are acknowledged as important political players in the ToC. Indeed, as outlined in section 3.2, elders are in many villages responsible for selecting VC members and are themselves often members. They also enjoy very high levels of trust among the population. The programme’s failure to include elders in the GIP does not therefore mean that elders were not integrated in the programme. Rather, it means the programme failed to specify elders’ role in the programme especially in relation to the VC and the village development committee, as well as in relation to the DC. The programme itself did not make clear the different roles and responsibilities of village level institutions, and thus suffered from the same lack of clarity it attributes to village governance in Somalia/Somaliland. Beyond stimulating further discussion on the roles and responsibilities of the different institutions (among authorities themselves, among citizens and between authorities and citizens), CDD/R would need to tailor training and capacity-building initiatives to the specific responsibilities and needs of the different local authorities. To conclude, the programme needs to expand its addressees beyond “local government officials” and “selected village and District Council members”, and to specifically integrate elders and religious leaders. It also needs to tailor trainings to the specific duties and responsibilities that are assigned to the particular institutions. With respect to recommendation 3, it is however necessary to keep in mind that the differentiation of institutions and offices, the clarification of roles and responsibilities, and the assignment of clear office portfolios and duties, amount to a formalisation and bureaucratisation of governance, and thus would imply significant social and political transformations. Governance in Somalia/Somaliland is performed through overlapping roles and responsibilities. To de-differentiate and thus to formalise it, if realistic at all, would need a more careful discussion and conflict sensitive approach, as it would initiate socio-political transformations that are likely to trigger conflicts. The general project assumption that current village governance lacks efficiency would itself require further investigation. It seems that the programme document was influenced by a liberal view on how governance should work (through formal procedures and processes, with policies implemented in a bureaucratic way) rather than by an assessment of governance problems in Somalia. On one side the ToC criticizes inefficiency, on the other it acknowledges hybridity, and thus the blend between ideal typically differentiated governance principles. Governance in Somaliland in Puntland is mostly process oriented and not focused on outcomes (modern administrations in contrast are outcome oriented), and may thus be inefficient from a liberal point of view. In the case of Somaliland and Puntland these forms of governance have nonetheless achieved some considerable success in peace - as well as statebuilding.3

Recommendation 4: Specify types and modes of participation Hogaan aimed at increasing participation of citizens in village-level decision-making. How- ever, the programme did not specify the modes and types of participation it would like to encourage, but instead used modes of participation that were already in place. The programme thus failed to move towards the ideal of improving citizens’ voice and choice,

3 Hybridity of course provides some substantial challenges with respect to democratisation. This became recently quite apparent in Somaliland, cf. for example RVI 2013; APD 2010.

82 but multiplied participation types that are already commonly practiced, such as representation, information sharing and consultation, and built on already available modes of participation through, for example, the clan system, CBOs or personal contacting/invitation. Being informed about a process or even being involved in it does not automatically lead to having voice and agency. To stimulate a shift from these rather passive to more interactive forms of participation, the programme needs to specify or to facilitate discussions around the types and modes of participation already practiced, and to outline the types and modes it envisages. This should include an assessment of the different types of decisions that are being or should be made, who exactly is involved in which types of decision-making, and how different stakeholders and villagers more generally could and should participate in decision-making processes (direct/indirect; passive/selective; active/interactive), if at all. One focus in these assessments should be on the people who do not participate, those who are absent or retreat in the course of an activity or project, and why. On the basis of such an assessment, specific types and modes of participation could be initiated for the different areas of decision-making and specific hypotheses on the pathway of change developed. To give an example: indirect participation through clan-based representation is quite common in Somalia and Somaliland. Another often and maybe increasingly practiced form of indirect participation is through community-based organisations (See above section 3.1.1). Programmes (sometimes knowingly, sometimes not) tend to rely on the same representational mechanisms, and thus go, for example, through women groups when they aim at integrating women into their programmes or through youth groups when approach- ing young people. The participation of a representative of this group in a meeting is then used as evidence for women’s or youth’s participation. If, for example, the chairlady of a women’s group participates in a meeting or programme, organisations seem to assume a) that the chairlady represents ‘the women’ of the village, b) that the chairlady discusses decisions to be made with ‘the women’ and reflects this discussion back in the meetings, and c) that she informs ‘the women’ about decisions made. However, exactly these assumed mechanisms of representational participation would need further specification. This would enable the programme to build-up a plausible logic of change based on differentiated forms and mechanisms of participation. It would also allow the programme to develop more differentiated pathways of change and would thus help to specify objectives. Again a note of caution is required. Participation is not simply a technique that can be stimulated but an inherently political and thus potentially conflictive process. It thus requires careful and conflict-sensitive considerations, consultation and negotiations with a broad range of village stakeholders.

Recommendation 5: Specify expected effects of participation Hogaan additionally seems to assume that participation automatically enhances satisfaction with decision-making. Again, this assumption would need further investigation. The evaluation found no evidence that higher rates of engagement with local institutions significantly changed how respondents assessed either the accessibility or the responsiveness of local institutions. For example, women interact much less with both formal and customary institutions, but nonetheless share the views of men on accessibility and responsiveness of these institutions. Similar results are found for youth: Although young people (18-25 years) interact the least frequently with all governing institutions and are least involved in planning and decision-making, people from this age group are among the most satisfied with the service delivery, management capacity and management style of

83 formal authorities. Indeed older people (>54 years) seem to be the most critical of the capacities of local authorities. With respect to the Village Council, older people are significantly less likely than young people to think that the VC can manage projects, and that the VC handles money transparently and accountably. The diﬀerences are even greater in the assessment of the DC where older people are signiﬁcantly less likely to think that the DC provides services, can manage projects, or can handle money transparently and accountably. This age group also has lowest levels of trust in the VC and DC. Here it is most likely not their lack of opportunities towards participation, but past experiences and in Somalia/Somaliland the experience of war and decay of formal institutions that may contribute to the rather negative view of current forms of governance.

Recommendation 6: Be specific about the addressees Besides enhancing participation in general, the Hogaan programme aimed at creating space to enable women, youth and marginalised people’ (ToC document) to participate in decision-making. Like most programmes, Hogaan relies on collective terms to define its main addresses, such as ‘the community’, ‘the marginalised’, ‘the women’ or ‘the youth’. Collective terms are by nature unspecific and amorphous. For example when interviewees in the villages were requested to define community and to describe what it consists of, they referred to different characteristics such as location, sharing/contributing, joint activities, or social diversity. Location was often used as basic criterion: to live together in one place is to be part of the community. However, in the same time many people mentioned the rural hinterlands as part of the village community. Villagers themselves emphasized the social diversity of communities, often listing differences in sex (men and women), age (children, youth, elderly) or occupation (business people, herders, politi- cians, elders). These differentiations included emphasis on the inclusiveness of community membership. As in other contexts, the notion of community had a normative connotation in Somalia/Somaliland. Interviewees described communities as people who share work, have a high degree of co-operation and aim for a peaceful mediation of conflicts, or indeed who mutually support each other. Already the utilisation of the term thus conceals power relations and conflicts within the target group. For the purpose of specifying addressees of a CDD/R project, the utilisation of normatively coloured, collective terms should be avoided. The project instead should specify who within the village community is supposed to participate. To exchange one collective term, such as community, with a series of others, such as the women, the youth etc. does not enhance clarity. Hogaan works through representatives, usually assuming that anybody from a particular collective (women, youth etc.) is indeed representative of it and somehow speaks for it. However, in reality of course a women or a young man may not feel or act as representative of the women or the youth, but have closer bonds to their extended families, kin or clan group, to people who have the same occupation or similar wealth etc. Moreover, people within these collectives may have radically differing needs and concerns. Women for example may face quite different challenges with respect to participation depending on their age, wealth, business success, education, the family they come from and the marriage relations they have entered (or failed to enter). Similarly problematic is reference to the collectivity of youth. The ToC states that

youth have limited say in local governance mechanisms [...], that youth are largely seen as disturbers of the peace or as a labour force, but not as rele-

84 vant contributors to community decisions. In most cases, youth are excluded from Village Councils and District Councils and have little say over village aﬀairs. They are often bored, have limited access to post-primary education, and few employment or recreational opportunities. Experience [...] shows that this, combined with their lack of voice in community governance, leads to considerable frustration, and can escalate into conﬂict or insecurity.

In Somalia the dominant view identiﬁes youth as potential disturbers, and the document accepts this. Additionally, the programme document also seems to equate youth with young men, as young women in Somalia, in comparison to their male counterparts, are quite busy. Presumably the document does not refer to young women when outlining the idleness of youth (young women may nonetheless be bored given the absence of recreational opportunities). However, the same equation of youth with young men became apparent in Somalia. When we asked to be introduced to the youth of the village, we were usually introduced to young men. After further inquiry if young women are not considered as youth, the youngsters stated that they of course generally were, but that the consultant had not particularly requested the inclusion of women. This already indicates the danger of tokenism, that is representatives from ‘marginalised’ groups are invited to satisfy the need of the donor. Even if invited, they then often lack voice and power to articulate their views, let alone to inﬂuence decision-making.

Recommendation 7: Address exclusion and discrimination of caste groups sep- arately The programme largely failed in even identifying caste members in the villages. It should thus explore alternative mechanisms that ensure not only identiﬁcation but also inclusion. It is often only after repeated inquiry that villagers point to the (often very small number of) caste members living in the village. The majority of Gabooye/Madhiban interviewed had either not heard of the programme or if they had, were not invited to participate either in decision-making or in sharing the beneﬁts.4 Some interviewees from the caste groups were quite blunt about this discrimination:

This is small village, and everything depends on tribes, certain tribes are seg- regated and even sometimes they are insulted. We as the minority people, we don’t get any support from this community.

The quantitative assessment equally confirms the much lower levels of inclusion of caste groups (Figure 49). If people from these groups are supposed to be included, then special mechanisms should be developed first to ensure that they are able to participate at all, and second identifying how they are supposed to participate if the whole culture of meetings prevents them from being heard. If numbers are sufficient, it may well be worth first supporting the self-organisation capacities of caste members before starting to facilitate their inclusion in decision-making.

4 There were exceptions to the latter. In one village in Erigabo, for example, a Madhiban household received one of the Berkeds that were built.

85 Participated in the Hogaan Programme Participated in the Hogaan Programme

30 (16,● 46) Caste −20 (−30,● −7)*** Caste significance 43 (32,● 53) Minority Clan −7 (−13, −2)** a● negative 50 (40, 59) Minority Clan ● Majority ●

0% 20% 40% 60% −30% −20% −10% 0% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in % Figure 49: Comparison of Clan Groups on Participation in Hogaan Programme

Recommendation 8: Analyse how different forms of contribution impact on social cohesion The moral obligation to support community members in distress was in the qualitative assessment emphasized by many interviewees as a dominant feature of the Somali culture and constitutive element of communal relations. The moral obligation to support community members in distress is an important feature of communal relations in Somalia. It is this moral economy that specifies community boundaries, defines rules of engagement with and towards the ‘external’, and binds people through mutual obligations and pro- longed debts.5 Contributions are one of the things that define the realm of inclusion and exclusion, and circumscribe who is considered a member of a community and who is at the margins or even excluded. Three types of contributions were differentiated in the Somali villages:

1. contributions to support poor and needy people (charity);

2. (blood) compensation payments in case on injury and conﬂict; and

3. support for projects, most commonly the building of mosques but also development oriented projects.

These contributions are organized differently, the first as need arises and by anybody who identifies needy people. Several interviewees stressed the important role of women in the organization of charity. Contributions for (blood) compensation is in contrast ne- gotiated and organized mainly by elders, while contributions for development involve the village administration as well as elders. In the case of development projects, committees are often formed to oversee the organization of contributions and the implementation of the project. Villagers described different ways on how community contributions for the CDD/R program under focus were organized. Some villages collected contribution on clan base, others mainly on basis of wealth and occupation and again others combined both criteria. Few villagers seemed to have relied mainly on ‘door to door’ collections, without considering householders’ characteristics. Usually the male member of a household was requested to contribute, but women with their own income (tea shops, kiosks and restau- rants in the villages are often run by women) were also requested to donate. Some villagers who based their contribution on clan affiliation, stressed that majority clans have due to their numerical strength in the village contributed more (usually double) than people from numerically less strong clans in the villages. Based on these findings the evaluation designed and added a module on contributions to the survey, which included questions on the different types of contributions, frequency

5 Beyond the more moral debt created by gift-exchange, lending money was also regularly mentioned as indication of communal relations.

86 and organisations of contributions. The module aimed at a further understanding of contribution practices, at analysing the potential relationship between community contributions and project participation, and at evaluating their impact on village level planning and governance. The evaluation found evidence that contributions are closely related to perceptions of local governance institutions. For example, people who contribute to social or development projects are more likely to trust the VC and DC, and acknowledge their management and planning capacities (See Figure 115). Similar patterns characterise security related and even blood contributions. We cannot determine if people contribute more because their trust is higher or if contribution increases trust, so this finding may not be so surprising. However, a similar pattern characterises not just security related and but also blood contributions. In addition, patterns of contribution are substantially affected by international programmes. While further analysis will be required to draw more than preliminary conclusions, the results suggest that the way in which people contribute to development projects needs to be taken into account in more detail. It may well make a difference if, for example, contributions are organised on clan basis or neighbourhood basis, both options that were narrated in interviews. The relatively high number of people who do not contribute also needs further analysis. If there is a causal relationship between contribution and trust, this would suggest that organisations should thinks carefully about the mechanism used to collect contributions and take measure to ensure that those who do not contribute are in some ways supported to do so. Thus our final recommendation is that more considera- tion should be given to the flow of contributions into the programme, the means by which these are organised and how the forms of contribution may influence social cohesion.

87 Bibliography

Agarwal, Bina (2001). ”Participatory Exclusions, Community Forestry, and Gender: An Analysis for South Asia and a Conceptual Framework.” World Development 29(10): 1623-1648.

APD, Academy for Peace and Development (2010). Democracy in Somaliland. Challenges and Opportunities http://www.interpeace.org/index.php/documents/publications/ somali-region/220-pillars-of-peace-somali-programme-english/file Accessed 05 July 2015.

Barron, Patrick, Humphreys, Macartan, Paler, Laura and Weinstein, Jeremy (2009). Community-based reintegration in Aceh : assessing the impacts of BRA-KDP.

Beath, Andrew, Christia, Fotini and Enikolopov, Ruben (2013). Randomized Impact Evaluation of Afghanistans National Solidarity Programme. Final Report.

Besteman, Catherine (1996). “Representing Violence and Othering Somalia.” Cultural Anthropology 11(1): 120¡U+0093¿133.

Boege, Volker, Brown, Anne , Clements, Kevin and Nolan, Anna (2008). On Hybrid Political Orders and Emerging States: State Formation in the Context of Fragility’. Berlin, Berghof Research Center for Constructive Conﬂict Engagement.

Casey, Katherine, Glennerster, Rachel and Miguel, Edward (2013). The GoBifo project evaluation report. Assessing the impacts of community driven development in Sierra Leone http://www.povertyactionlab.org/publication/gobifo-project-evaluation- report-assessing-impacts-community-driven-development-sierra-l Accessed 26 June 2015.

Cassanelli, Lee (1995). Victims and Vulnerable Groups in Southern Somalia.

Cassanelli, Lee and Besteman, Catherine, Eds. (1996). The Struggle for Land in Southern Somalia. The War behind the War. Boulder, Westview.

Chandler, David (2010). International Statebuilding: The Rise of Post-Liberal Gover- nance. Oxon, Routledge.

Chesterman, Simon (2007). “Ownership in Theory and in Practice: Transfer of Authority in UN Statebuilding Operations.” Journal of Intervention and Statebuilding 1(1): 3-26.

Chopra, Jarat and Hohe, Tanja (2004). “Participatory Intervention.” Global Governance: A Review of Multilateralism and International Organizations 10(3): 289-305.

88 Cohen, John M. and Uphoﬀ, Norman T. (1980). “Participation’s place in rural development: Seeking clarity through speciﬁcity.” World Development 8(3): 213-235.

Cooke, W. and Kothari, U. (2001). Participation: the New Tyranny? . London, Zed.

Cornwall, Andrea (2008). “Unpacking Participation: models, meanings and practices.” Community Development Journal 43(3): 269-283.

Cornwall, Andrea (2000). Beneﬁciary consumer citizen: perspectives on participation for poverty reduction, Swedish International Development Cooperation Agency.

Cornwall, Andrea and Brock, Karen (2005). “What Do Buzzwords Do for Development Policy? A Critical Look at ’Participation’, ’Empowerment’ and ’Poverty Reduc- tion’.” Third World Quarterly 26(7): 1043-1060.

Fearon, James, Humphreys, Macartan and Weinstein, Jeremy (2008). Community Driven Reconstruction in Lofa County: Impact Assessment.

Grant, Gordon, Grossman, Guy and Paler, Laura (2013). Somalia Feasibility Assessment

Hickey, Sam and Mohan, Giles (2005). “Relocating Participation within a Radical Politics of Development.” Development and Change 36(2): 237-262.

Hill, Martin (2010). No Redress: Somalias Forgotten Minorities http://reliefweb.int/ sites/reliefweb.int/files/resources/DCAF9B4C12ED9AC8C12577E4004A9973- Full_Report.pdf Accessed 1 June 2015.

Hoenke, Jana and Hoenke, Jana (2012). ”Multinationals and Security Governance in the Community: Participation, Discipline and Indirect Rule.” Journal of Intervention and Statebuilding 6(1): 57-73.

Humphreys, Macartan, de la Sierra, Raul Sanchez and van der Windt, Peter (2014). Social and Economic Impacts of Tuungane: Final Report on the Eﬀects of a Com- munity Driven Reconstruction Program in Eastern Democratic Republic of Congo.

ICG, International Crisis Group (2013). Puntland’s Punted Polls http://www.crisisgroup. org/~/media/Files/africa/horn-of-africa/somalia/b097-somalia-puntlands- punted-polls.pdf Accessed 15 June 2015.

Jha, Abhas K., Barenstein, Jennifer Duyne, Phelps, Priscilla M., Pittet, Daniel and Sena, Stephen (2010). Project Implementation: Community Organizing and Participation. Safer Homes, Stronger Communities: A Handbook for Reconstructing after Natural Disasters. Washington, World Bank: 183-196.

Kapteijns, Lidwien (2011). “I. M. Lewis and Somali Clanship: A Critique.” Northeast African Studies 11(1): 2004¡U+0093¿2010.

King, Elisabeth (2013). A Critical Review of Community-Driven Development Pro- grammes in Conﬂict-Aﬀected Contexts

Kyamusugulwa, Patrick Milabyo (2013). ”Participatory Development and Reconstruc- tion: a literature review.” Third World Quarterly 34(7): 1265-1278.

89 Lewis, I.M (1961). A Pastoral Democracy. A Study of Pastoralism and Politics among the Northern Somali of the Horn of Africa. Oxford, Oxford University Press.

Luling, Virginia (1984). The Other Somali: Minority Groups in Traditional Somali So- ciety. Proceedings of the Second International Congress of Somali Studies, Vol. IV (Studies in Humanities and Natural Sciences). Labahn, Thomas. Hamburg, Buske: 39-55.

Mansuri, Ghazala and Rao, Vijayendra (2004). “Community-Based and -Driven Devel- opment: A Critical Review.” The World Bank Research Observer 19(1): 1-39.

Mansuri, Ghazala and Rao, Vijayendra (2013). Localizing Development : Does Partici- pation Work? Washington, World Bank.

Mohan, Giles and Stokke, Kristian (2000). “Participatory Development and Empower- ment: The Dangers of Localism.” Third World Quarterly 21(2): 247-268.

Reich, Hannah (2006). Local Ownership in Conﬂict Transformation Projects: Partner- ship, Participation or Patronage?

RVI, Rift Valley Institute (2013). What future for democracy in Somaliland? Learning from the 2012 local council elections http://www.riftvalley.net/publication/ what-future-democracy-somaliland#.VZTitkYXugQ Accessed 15 June 2015.

Samatar, Abdi I. (1992). “Destruction of State and Society in Somalia: Beyond the Tribal Convention.” Journal of Modern African Studies 30(4): 625¡U+0093¿641.

Suhrke, Astri (2007). “Reconstruction as modernisation: the post-conﬂict project in Afghanistan.” Third World Quarterly 28(7): 1291-1308.

ToC Document, Project Document (no date). “Theories of Change. Governance and Peacebuilding.”

Wong, Susan (2012). What Have Been the Impacts of World Bank Community-Driven Development Programs? CDD Impact Evaluation Review and Operational and Re- search Implications.

90 Appendix A Appendix: Data Analysis Plan Results

A.1 Programme Mechanisms Analysis As Identiﬁed in the Data Analysis Plan

The main evaluation of the effects of programme activities is found in Section 4.3. However, in table 3 of the data analysis plan the evaluation team proposed testing a number of hypotheses about the effects of programme activities in very specific ways. This section provides that analysis to deliver on those proposals.

A.1.1 Citizen Participation in decision-making Sub-Mechanism 1a: Citizen Training The data analysis plan identiﬁed tests against the following four hypotheses relating to the the training of citizens:

H1 Individual Quantity: Individuals who have been trained under the Hogaan programme will be more likely to participate in community forums and decision-making.

H2 Individual Quality: Individuals with higher regard for the Hogaan training they received will be more likely to participate in community forums and decision-making.

H3 Collective Quantity: Individuals in villages with greater levels of Hogaan individual training are more likely to participate in community forums and decision-making.

H4 Collective Quality: Individuals in villages where there is generally a high regard for Hogaan training are more likely to participate in community forums and village decision-making.

Sub-Mechanism 1b Leadership Training H5 Leadership Quantity Increases in levels of leadership training increase citizen participation.

H6 Leadership Quality Villages where leaders found the Hogaan training more useful will have citizens with a higher probability of participating in community forums.

91 H7 Relationship between Sub-Mechanism of Participation Increase: Both individual and leadership training contribute independently to increasing participation.

A.1.2 Mechanism 2: Strengthening Institutions H8 Leader Training and Planning Village leadership training related to citizens perceptions of ability local government’s ability to plan and mange projects.

H9 Leader Training and Service Delivery Village leadership training related to citizens perceptions of ability local government’s ability to plan and mange projects.

H10 Leader Training and Local Government Overall Village leadership training related to citizens perceptions of local government overall planning and service delivery.

A.1.3 Data Challenges The results combine the information from the leadership and the household surveys. This combination provide useable estimates on the question of whether people have or have not undertaken various forms of activity. However, in order to gauge the perceived quality of activities the evaluation directly asked people for their assessment of the quality. There was almost no variation on these measures, with almost all respondent from in household and leadership surveys reporting that all the activities that they participated in were very useful. For example on the leadership survey the most criticised activity was the training in Civic Education which received no negative comments, three who were unsure about its utility and 225 positive comments and 1 refusal. The analysis based on these utility questions is very uninformative, but nevertheless because we included it in the data analysis plan we follow through with the analysis here.

A.1.4 Overview of Data Analysis Plan Table 3 Results The regression tables and model comparison results specified in Table 3 of the data analysis plan can be found in Appendix: Data Analysis Plan Results. The table below gives a summary of the findings which can be found there relating to the above hypotheses. It can be seen that we have null findings for six of the ten hypotheses, two of which related to programme quality. We were unable to test two further hypotheses relating to programme quality. The evaluation found support for two of the ten hypotheses. It found support for the theory that participation in programme activities is related to participation in community meetings. This could be either because people who participate in community meetings anyway were more likely to participate in the Hogaan programme or because participation in the programme causes increased participation. The evaluation also found evidence that that people who live in villages where more training has taken place are more likely to participate in community meetings. This could be because there is a spillover effect, such that participation increases when there has been more engagement in the project across the whole village. However, it could also be an extension of the finding under H1. In some villages we encounter more people who participated in the Hogaan programme than in others. If, as in H1, these people as individuals were we more likely to participate in community meetings then this could show up as a village level effect. The spillover interpretation is rendered less likely, and

92 the individual interpretation more likely, by extending the analysis to include a control for individual level participation. When this is done village level Hogaan participation is no longer a signiﬁcant predictor. The overall ﬁndings of this section are consistent with the other discussion of programme activities, that individual level participation in the Hogaan programme is positively related to participation in community meetings.

Hypothesis Affecting Sub-Mechanism Slope Significance Conclusion H1 Individual Quantity Participation Citizen Training +ve *** Evidence for Mechanism H2 Individual Quality Participation Citizen Training none No Evidence for Mechanism H3 Village Quantity Participation Citizen Training +ve *** Evidence for Mechanism H4 Village Quality Participation Citizen Training none No Evidence for Mechanism H5 Leader Quantity Participation Leaders Training none No Evidence for Mechanism Insufficient negative responses on H6 Leader Quality Participation Leaders Training No Evidence for Mechanism leadership training to test hypothesis Insufficient negative responses on H7 H1 and H6 independently significant Participation Both Citizens and Leaders No Evidence for Mechanism leadership training to test hypothesis H8 Leadership improves LG Management and Planning LG Management Leaders Training none No Evidence for Mechanism H9 Leader Training improves LG Service Delivery LG Service Leaders Training none No Evidence for Mechanism H10 Leaders Training Improve LG Overall LG Overall Leaders Training none No Evidence for Mechanism

Table A.1: Overall Data Analysis Plan Results

A.2 Detailed Results

H1 Individual Quantity: Individual who have been trained under the Hogaan programme will be more likely to participate in community forums and decision-making.

Table A.2: Regression Results for H1

Dependent variable: Participate H1 compare H1 test (1) (2) Hogaan Training 2.196∗∗∗ (0.261) Female −0.597∗∗∗ (0.130) −0.536∗∗∗ (0.134) Age −0.001 (0.003) −0.002 (0.004) Minority Clan −0.482∗∗∗ (0.126) −0.445∗∗∗ (0.130) Caste −1.077∗∗∗ (0.324) −0.940∗∗∗ (0.329) Poor 0.464∗∗∗ (0.170) 0.479∗∗∗ (0.175) Better Oﬀ 0.861∗∗∗ (0.182) 0.775∗∗∗ (0.187) Rich 0.344 (0.475) 0.257 (0.493) Constant 0.608 (0.431) 0.378 (0.401) Observations 1,592 1,592 Log Likelihood −958.924 −905.481 Akaike Inf. Crit. 1,937.847 1,832.963 Bayesian Inf. Crit. 1,991.575 1,892.063 Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

93 Table A.3: Model Comparison for H1

Statistic N Mean St. Dev. Min Max Df 2 10.500 0.707 10 11 AIC 2 1,885.405 74.165 1,832.963 1,937.847 BIC 2 1,941.819 70.365 1,892.063 1,991.575 logLik 2 −932.202 37.789 −958.924 −905.481 deviance 2 1,864.405 75.579 1,810.963 1,917.847 Chisq 1 106.885 106.885 106.885 Chi Df 1 1.000 1 1 Pr(>Chisq) 1 0.000 0 0

H2 Individual Quality: Individuals with higher regard for the Hogaan training they received will be more likley to participate in community forums and decision-making.

Table A.4: Regression Results for H2

Dependent variable: Participate H2 compare H2 test (1) (2) Hogaan Useful 0.443 (0.869) Female −0.462 (0.323) −0.457 (0.323) Age 0.001 (0.010) 0.001 (0.010) Minority Clan −0.611∗∗ (0.311) −0.605∗ (0.311) Caste −0.999 (0.767) −0.977 (0.769) Poor −0.368 (0.491) −0.370 (0.492) Better Oﬀ −0.068 (0.519) −0.078 (0.520) Rich −0.888 (1.251) −0.900 (1.249) Constant 2.975∗∗∗ (0.717) 2.554∗∗ (1.105) Observations 559 559 Log Likelihood −186.054 −185.929 Akaike Inf. Crit. 390.107 391.858 Bayesian Inf. Crit. 429.042 435.120 Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

94 Table A.5: Model Comparison for H2

Statistic N Mean St. Dev. Min Max Df 2 9.500 0.707 9 10 AIC 2 390.983 1.238 390.107 391.858 BIC 2 432.081 4.297 429.042 435.120 logLik 2 −185.991 0.088 −186.054 −185.929 deviance 2 371.983 0.176 371.858 372.107 Chisq 1 0.249 0.249 0.249 Chi Df 1 1.000 1 1 Pr(>Chisq) 1 0.618 0.618 0.618

H3 Collective Quantity: Individuals in villages with greater levels of Hogaan individual training are more likely to participate in community forums and decision-making.

Table A.6: Regression Results for H3

Dependent variable: Participate H3 compare H3 test (1) (2) Village Train % 2.094∗∗∗ (0.407) Female −0.534∗∗∗ (0.141) −0.516∗∗∗ (0.140) Age −0.0001 (0.004) −0.001 (0.004) Minority Clan −0.459∗∗∗ (0.136) −0.472∗∗∗ (0.135) Caste −1.062∗∗∗ (0.333) −1.067∗∗∗ (0.333) Poor 0.461∗∗ (0.183) 0.427∗∗ (0.183) Better Oﬀ 0.925∗∗∗ (0.195) 0.873∗∗∗ (0.195) Rich 0.453 (0.483) 0.446 (0.479) Constant 0.506 (0.448) −0.434 (0.413) Observations 1,365 1,365 Log Likelihood −816.717 −805.252 Akaike Inf. Crit. 1,653.433 1,632.503 Bayesian Inf. Crit. 1,705.622 1,689.911 Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

95 Table A.7: Model Comparison for H3

Statistic N Mean St. Dev. Min Max Df 2 10.500 0.707 10 11 AIC 2 1,642.968 14.800 1,632.503 1,653.433 BIC 2 1,697.767 11.109 1,689.911 1,705.622 logLik 2 −810.984 8.107 −816.717 −805.252 deviance 2 1,621.968 16.214 1,610.503 1,633.433 Chisq 1 22.930 22.930 22.930 Chi Df 1 1.000 1 1 Pr(>Chisq) 1 0.00000 0.00000 0.00000

H4 Collective Quality: Individuals in villages where there is generally a high regard for Hogaan training are more likely to participate in community forums and village decision-making.

Table A.8: Regression Results for H4

Dependent variable: Participate H4 compare H4 test (1) (2) Hogaan Useful (village) −1.641 (1.865) Female −0.522∗∗∗ (0.141) −0.522∗∗∗ (0.141) Age −0.001 (0.004) −0.001 (0.004) Minority Clan −0.487∗∗∗ (0.137) −0.491∗∗∗ (0.137) Caste −1.116∗∗∗ (0.343) −1.118∗∗∗ (0.343) Poor 0.464∗∗ (0.184) 0.460∗∗ (0.184) Better Oﬀ 0.926∗∗∗ (0.197) 0.931∗∗∗ (0.197) Rich 0.573 (0.502) 0.590 (0.501) Constant 0.549 (0.434) 2.160 (1.882) Observations 1,339 1,339 Log Likelihood −802.766 −802.382 Akaike Inf. Crit. 1,625.531 1,626.763 Bayesian Inf. Crit. 1,677.528 1,683.960 Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

96 Table A.9: Model Comparison for H4

Statistic N Mean St. Dev. Min Max Df 2 10.500 0.707 10 11 AIC 2 1,626.147 0.871 1,625.531 1,626.763 BIC 2 1,680.744 4.548 1,677.528 1,683.960 logLik 2 −802.574 0.272 −802.766 −802.382 deviance 2 1,605.147 0.543 1,604.763 1,605.531 Chisq 1 0.768 0.768 0.768 Chi Df 1 1.000 1 1 Pr(>Chisq) 1 0.381 0.381 0.381

H5 Leadership Quantity Increases in levels of leadership training increase citizen participation.

Table A.10: Regression Results for H5

Dependent variable: Participate H5 compare H5 test (1) (2) Leaders Trained (%) −0.050 (0.986) Female −0.558∗∗∗ (0.142) −0.558∗∗∗ (0.142) Age −0.0005 (0.004) −0.0005 (0.004) Minority Clan −0.448∗∗∗ (0.138) −0.447∗∗∗ (0.138) Caste −1.080∗∗∗ (0.334) −1.080∗∗∗ (0.334) Poor 0.425∗∗ (0.185) 0.425∗∗ (0.186) Better Oﬀ 0.880∗∗∗ (0.197) 0.880∗∗∗ (0.197) Rich 0.412 (0.484) 0.412 (0.484) Constant 0.578 (0.454) 0.621 (0.968) Observations 1,341 1,341 Log Likelihood −801.830 −801.829 Akaike Inf. Crit. 1,623.661 1,625.658 Bayesian Inf. Crit. 1,675.672 1,682.871 Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

97 Table A.11: Model Comparison for H5

Statistic N Mean St. Dev. Min Max Df 2 10.500 0.707 10 11 AIC 2 1,624.659 1.412 1,623.661 1,625.658 BIC 2 1,679.272 5.090 1,675.672 1,682.871 logLik 2 −801.830 0.001 −801.830 −801.829 deviance 2 1,603.659 0.002 1,603.658 1,603.661 Chisq 1 0.003 0.003 0.003 Chi Df 1 1.000 1 1 Pr(>Chisq) 1 0.960 0.960 0.960

H8 Leader Training and Planning Village leadership training related to citizens perceptions of ability local government’s ability to plan and mange projects.

Table A.12: Regression Results for H8

Dependent variable: Local Government Mangement and Planning H8 compare H8 test (1) (2) Leaders Trained (%) 0.108 (0.207) Female 0.094∗∗∗ (0.035) 0.094∗∗∗ (0.035) Age −0.004∗∗∗ (0.001) −0.004∗∗∗ (0.001) Minority Clan −0.246∗∗∗ (0.035) −0.247∗∗∗ (0.035) Caste −0.156∗ (0.085) −0.157∗ (0.085) Poor −0.037 (0.049) −0.037 (0.049) Better Oﬀ 0.024 (0.050) 0.024 (0.050) Rich 0.100 (0.122) 0.100 (0.122) Constant 1.229∗∗∗ (0.119) 1.137∗∗∗ (0.214) Observations 1,420 1,420 Log Likelihood −2,771.230 −2,771.096 Akaike Inf. Crit. 5,562.460 5,564.192 Bayesian Inf. Crit. 5,615.045 5,622.034 Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

98 Table A.13: Model Comparison for H8

Statistic N Mean St. Dev. Min Max Df 2 10.500 0.707 10 11 AIC 2 5,563.326 1.224 5,562.460 5,564.192 BIC 2 5,618.539 4.942 5,615.045 5,622.034 logLik 2 −2,771.163 0.095 −2,771.230 −2,771.096 deviance 2 5,542.326 0.190 5,542.192 5,542.460 Chisq 1 0.269 0.269 0.269 Chi Df 1 1.000 1 1 Pr(>Chisq) 1 0.604 0.604 0.604

H9 Leader Training and Service Delivery Village leadership training related to citizens perceptions of ability local government’s ability to plan and mange projects.

Table A.14: Regression Results for H9

Dependent variable: Local Government Service Delivery H9 compare H9 test (1) (2) Leaders Trained (%) 0.357 (0.248) Female 0.094∗ (0.049) 0.095∗ (0.049) Age −0.003∗∗ (0.001) −0.003∗∗ (0.001) Minority Clan −0.219∗∗∗ (0.049) −0.223∗∗∗ (0.049) Caste −0.121 (0.120) −0.125 (0.120) Poor 0.073 (0.071) 0.073 (0.071) Better Oﬀ 0.099 (0.074) 0.100 (0.074) Rich 0.079 (0.184) 0.082 (0.184) Constant 0.369∗∗ (0.148) 0.061 (0.257) Observations 1,420 1,420 Log Likelihood −1,889.356 −1,888.325 Akaike Inf. Crit. 3,798.712 3,798.650 Bayesian Inf. Crit. 3,851.296 3,856.492 Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

99 Table A.15: Model Comparison for H9

Statistic N Mean St. Dev. Min Max Df 2 10.500 0.707 10 11 AIC 2 3,798.681 0.044 3,798.650 3,798.712 BIC 2 3,853.894 3.674 3,851.296 3,856.492 logLik 2 −1,888.840 0.729 −1,889.356 −1,888.325 deviance 2 3,777.681 1.458 3,776.650 3,778.712 Chisq 1 2.062 2.062 2.062 Chi Df 1 1.000 1 1 Pr(>Chisq) 1 0.151 0.151 0.151

H10 Leader Training and Local Government Overall Village leadership training related to citizens perceptions of local government overall planning and service delivery.

Table A.16: Regression Results for H10

Dependent variable: Local Government Overall H10 compare H10 test (1) (2) Leaders Trained (%) 0.137 (0.146) Female 0.115∗∗∗ (0.021) 0.115∗∗∗ (0.021) Age −0.002∗∗∗ (0.001) −0.002∗∗∗ (0.001) Minority Clan −0.224∗∗∗ (0.022) −0.225∗∗∗ (0.022) Caste −0.121∗∗ (0.053) −0.122∗∗ (0.053) Poor 0.022 (0.030) 0.022 (0.030) Better Oﬀ 0.054∗ (0.031) 0.054∗ (0.031) Rich 0.046 (0.078) 0.046 (0.078) Constant 2.072∗∗∗ (0.086) 1.953∗∗∗ (0.152) Observations 1,420 1,420 Log Likelihood −3,814.006 −3,813.564 Akaike Inf. Crit. 7,648.011 7,649.128 Bayesian Inf. Crit. 7,700.595 7,706.971 Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

100 Table A.17: Model Comparison for H10

Statistic N Mean St. Dev. Min Max Df 2 10.500 0.707 10 11 AIC 2 7,648.570 0.790 7,648.011 7,649.128 BIC 2 7,703.783 4.508 7,700.595 7,706.971 logLik 2 −3,813.785 0.312 −3,814.006 −3,813.564 deviance 2 7,627.570 0.624 7,627.128 7,628.011 Chisq 1 0.883 0.883 0.883 Chi Df 1 1.000 1 1 Pr(>Chisq) 1 0.347 0.347 0.347

101 Appendix B Appendix: Further Results

B.1 Further Model Results (Hogaan and Non-Imple- mentation Villages)

102 Elders: Priority Women Elders: Priority Women

Hogaan 85 (64,● 94)

−4 (−8,● 0)

Non−imp 89 (69,● 96)

Elders: Priority Youth Elders: Priority Youth

Hogaan 82 (62,● 94)

−3 (−7,● 1)

Non−imp 86 (65,● 95)

Elders: Priority Outcast Clans Elders: Priority Outcast Clans

Hogaan 75 (53,● 89)

1 (−5,● 9)

Non−imp 73 (52,● 88)

Relgious Leaders: Priority Women Relgious Leaders: Priority Women

Hogaan 94 (84,● 98)

−4 (−11,● −1)***

Non−imp 99 (93,● 100)

Religious Leaders: Priority Youth Religious Leaders: Priority Youth

Hogaan 91 (80,● 96)

−3 (−8,● 1)

Non−imp 95 (85,● 98)

Religious Leaders: Priority Outcast Clans Religious Leaders: Priority Outcast Clans

Hogaan 82 (59,● 94)

−7 (−15,● −1)*

Non−imp 90 (68,● 98) significance VC: Priority Women VC: Priority Women a● negative a● none Hogaan 86 (67,● 96)

−3 (−8,● 1)

Non−imp 90 (70,● 97)

VC: Priority Youth VC: Priority Youth

Hogaan 83 (65,● 94)

−2 (−6,● 3)

Non−imp 85 (67,● 95)

VC: Priority Outcast Clans VC: Priority Outcast Clans

Hogaan 74 (51,● 90)

1 (−5,● 7)

Non−imp 73 (50,● 90)

DC: Priority Women DC: Priority Women

Hogaan 85 (68,● 94)

−2 (−7,● 3)

Non−imp 86 (70,● 95)

DC: Priority Youth DC: Priority Youth

Hogaan 82 (64,● 92)

−1 (−7,● 4)

Non−imp 83 (66,● 93)

DC: Priority Outcast Clans DC: Priority Outcast Clans

Hogaan 73 (52,● 89)

−1 (−8,● 6)

Non−imp 74 (53,● 89)

0% 25% 50% 75% 100% −15% −10% −5% 0% 5% 10% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 50: Perceptions of Institutional Inclusion

103 Women Involved in Planning Process Women Involved in Planning Process

94 (84,● 98) 22 (12, 31)*** Other Other ● 96 (88, 99) CAP ● 23 (15, 33)*** CAP ● 93 (83, 97) Training ● 20 (11,● 29)*** 78 (64, 87) Training Community Entry ● 6 (−5, 16) 79 (68, 87) Community Entry ● None: Hogaan ●

72 (59, 81) 7 (−2,● 15) Non−imp ● None: Hogaan

Youth Involved in Planning Process Youth Involved in Planning Process

95 (84,● 98) 22 (13, 32)*** Other Other ● 92 (81, 96) CAP ● 19 (11, 28)*** CAP ● 85 (71, 92) Training ● 13 (4,● 21)** 72 (55, 82) Training Community Entry ● 0 (−11, 10) 80 (67, 87) Community Entry ● None: Hogaan ●

7 (1,● 15)* 72 (58,● 81) None: Hogaan Non−imp significance

● Minority Clans Involved in Planning Process Minority Clans Involved in Planning Process a none a● positive 94 (85, 98) 19 (10, 29)*** Other ● Other ● 92 (83, 96) CAP ● 17 (9, 26)*** CAP ● 95 (87, 98) Training ● 20 (12, 29)*** Training ● 71 (56, 81) Community Entry ● −3 (−14, 6) 81 (70, 88) Community Entry ● None: Hogaan ● 6 (−1, 13) 75 (63, 83) ● Non−imp ● None: Hogaan

Outcast Clans Involved in Planning Process Outcast Clans Involved in Planning Process 91 (78, 96) 17 (8, 26)*** Other ● Other ●

91 (80,● 96) 17 (9, 26)*** CAP CAP ● 88 (76, 94) Training ● 14 (6, 23)*** Training ● 64 (48, 77) Community Entry ● −9 (−19, 0) 77 (64, 85) Community Entry ● None: Hogaan ● 3 (−3, 9) 74 (60, 83) ● Non−imp ● None: Hogaan

0% 25% 50% 75% 100% −20% −10% 0% 10% 20% 30% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 51: Comparison of Non-Implementation with Levels of Hogaan Activity on Community Perceptions of Inclusion

Respondent Involved in Planning Process[Youth Respondents] Respondent Involved in Planning Process[Youth Respondents]

68 (48,● 82) 13 (−2, 28) Other Other ● 84 (68, 92) CAP ● 29 (16, 43)*** CAP ● 83 (65, 91) Training ● 27 (15, 41)*** 66 (43, 79) Training ● Community Entry ● 10 (−5, 25) 54 (36, 66) Community Entry ● None: Hogaan ● −1 (−12, 9) 55 (37,● 68) None: Hogaan ● Non−imp significance

● Youth Involved in Planning Process[Youth Respondents] Youth Involved in Planning Process[Youth Respondents] a none ● 94 (78, 98) 25 (11, 38)*** a positive Other ● Other ● 93 (79,● 98) 25 (13, 38)*** CAP CAP ● 85 (68, 92) Training ● 17 (3, 29)** Training ● 72 (51,● 84) Community Entry 4 (−13, 18) 81 (69, 87) Community Entry ● None: Hogaan ● 13 (3, 24)** 68 (53, 78) ● Non−imp ● None: Hogaan

0% 25% 50% 75% 100% 0% 20% 40% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 52: Youth Perception of their own and Youth General Involement in the Planning Process

104 Respondent Involved in Planning Process[Minority Clan Respondents] Respondent Involved in Planning Process[Minority Clan Respondents]

79 (57,● 91) 34 (16, 49)*** Other Other ● 84 (66, 93) CAP ● 39 (24, 53)*** CAP ● 86 (67, 94) Training ● 40 (24, 56)*** 84 (65, 93) Training ● Community Entry ● 39 (22, 54)*** 53 (36, 66) Community Entry ● None: Hogaan ● 8 (−4, 20) 44 (26,● 61) None: Hogaan ● Non−imp significance

● Minority Clans Involved in Planning Process[Minority Clan Respondents] Minority Clans Involved in Planning Process[Minority Clan Respondents]a none ● 92 (72, 98) 23 (6, 37)** a positive Other ● Other ● 86 (74,● 94) 18 (4, 32)** CAP CAP ● 96 (75, 99) Training ● 26 (9, 40)** Training ● 73 (51,● 85) Community Entry 3 (−17, 21) 76 (66, 83) Community Entry ● None: Hogaan ● 8 (−4, 20) 68 (52, 80) ● Non−imp ● None: Hogaan

0% 25% 50% 75% 100% −20% 0% 20% 40% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 53: Minority Clan Perception of their own and Minority Clan General Involement in the Planning Process

VC Meetings Consider my Household Viewpoint VC Meetings Consider my Household Viewpoint 94 (90, 97) Hogaan ● significance −1 (−5,● 6) ● none Non−implementation 95 (86,● 99) a

0% 25% 50% 75% 100% −6% −3% 0% 3% 6% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 54: Alternate Participation Measures: Comparison Village Type

Should youth have the same rights as elderly people? Should youth have the same rights as elderly people?

Hogaan 84 (77,● 88)

1 (−8,● 12)

Non−imp 83 (70,● 90)

significance Currently, do youth have the same rights as elderly people? Currently, do youth have the same rights as elderly people? a● none

Hogaan 79 (72,● 84)

6 (−5,● 17)

Non−imp 73 (61,● 82)

0% 25% 50% 75% 0% 10% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 55: Leader Perception of Youth’s Rights and Inclusion in Hogaan and Comparison Villages

105 Contribution: International Development Project Contribution: International Development Project

hogan 66 (51,● 76)

28 (18,● 37)*** comparison 38 (22,● 51)

Contribution: Social Services Contribution: Social Services

hogan 67 (51,● 79)

13 (6, ●21)*** comparison 54 (37,● 68)

Contribution: Poor and Needy Contribution: Poor and Needy 84 (68, 90) hogan ● significance 0 (−6, 5) ● ● a none ● positive comparison 85 (68,● 91) a

Contribution: Security Contribution: Security

hogan 83 (66,● 92)

7 (2, ●14)** comparison 76 (58,● 87)

Contribution: Blood Contribution: Blood

hogan 82 (69,● 88)

−6 (−12,● 1) comparison 87 (77,● 92)

0% 25% 50% 75% −10% 0% 10% 20% 30% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 56: Percentage Having Made Contributions of Diﬀerent Kinds – Further Models

106 Priority: Life Priority: Life

hogan 72 (57,● 82)

−8 (−14,● −2)**

comparison 80 (66,● 88)

Priority: Justice Priority: Justice

hogan 53 (49,● 56)

−3 (−11,● 5)

comparison 56 (48,● 63)

Priority: Equality Priority: Equality

hogan 15 (12,● 19)

7 (2, ●10)**

comparison 9 (5,● 13)

Priority: Education Priority: Education

significance hogan 59 (48,● 66) a● negative 9 (3, ●16)** ● 49 (37, 59) a none comparison ● a● positive

Priority: Health Priority: Health

hogan 71 (65,● 75)

−9 (−15,● −3)**

comparison 80 (73,● 85)

Priority: Consulted on Decisions Priority: Consulted on Decisions

hogan 7 (4,● 12)

3 (0,● 7)

comparison 4 (1,● 8)

Priority: Free Speech Priority: Free Speech

hogan 5 (2,● 11)

2 (−1,● 6)

comparison 2 (1,● 7)

0% 25% 50% 75% −10% 0% 10% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 57: Included in Top Three Rights Priorities in Project and Comparison Villages – Further Models

Women's Issues in Conflict Resolution Women's Issues in Conflict Resolution

Hogaan 79 (73,● 83)

9 (3, ●16)**

Non−implementation 70 (61,● 77)

Youth Issues in Conflict Resolution Youth Issues in Conflict Resolution

Hogaan 77 (68,● 83) significance 9 (3, 16)*** ● ● a none ● Non−implementation 68 (57,● 76) a positive

Minority Clan Issues in Conflict Resolution Minority Clan Issues in Conflict Resolution

Hogaan 76 (66,● 83)

3 (−2,● 10)

Non−implementation 73 (61,● 80)

0% 20% 40% 60% 80% 0% 5% 10% 15% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 58: Conﬂict Mechanisms in Project and Comparison Villages – Further Models

107 B.2 Further Model Results (Gender Interaction)

108 Elders Consider my Viewpoint Elders Consider my Viewpoint Elders Consider my Viewpoint

b. Hogaan 88 (75,● 93) Male b. Hogaan 0 (−4, 4) ● difference b. Hogaan 88 (75,● 92) Female −5 (−15, 3) ● a. Non−imp 86 (70,● 93) Male a. Non−imp 5 (−3, 13) ● difference a. Non−imp 90 (79,● 95) Female

Interact Elders Interact Elders Interact Elders

b. Hogaan 73 (62,● 81) Male b. Hogaan −20 (−26, −15)*** ● difference b. Hogaan 52 (40,● 62) Female 19 (6, 30)*** ● a. Non−imp 79 (66,● 88) Male a. Non−imp−39 (−49, −28)*** ● difference a. Non−imp 40 (27,● 52) Female

Elders Accessible Elders Accessible Elders Accessible

b. Hogaan 94 (85,● 97) Male b. Hogaan 1 (−2, 5) ● difference b. Hogaan 95 (87,● 97) Female −1 (−10, 6) ● a. Non−imp 93 (80,● 98) Male a. Non−imp 2 (−5, 10) ● difference a. Non−imp 95 (84,● 98) Female

Religious Leaders Consider my Viewpoint Religious Leaders Consider my Viewpoint Religious Leaders Consider my Viewpoint

b. Hogaan 96 (91,● 98) Male b. Hogaan −3 (−6, 0)* ● difference b. Hogaan 92 (88,● 95) Female −4 (−13, 3) ● a. Non−imp 91 (81,● 96) Male a. Non−imp 1 (−6, 9) ● difference a. Non−imp 93 (84,● 96)

Female significance

● none Interact with Relgious Leader Interact with Relgious Leader Interact with Relgious Leader a a● positive b. Hogaan 75 (61,● 82) Male b. Hogaan −16 (−22, −10)*** ● difference b. Hogaan 58 (45,● 69) Female 16 (4, 28)** ● a. Non−imp 80 (65,● 88) Male a. Non−imp −32 (−43, −21)*** ● difference a. Non−imp 47 (33,● 60) Female

Religious Leaders Accessible Religious Leaders Accessible Religious Leaders Accessible

b. Hogaan 97 (94,● 98) Male b. Hogaan 0 (−2, 2) ● difference b. Hogaan 96 (95,● 98) Female −2 (−9, 4) ● a. Non−imp 94 (84,● 98) Male a. Non−imp 2 (−4, 9) ● difference a. Non−imp 96 (89,● 98) Female

Respondent made request of local government Respondent made request of local government Respondent made request of local government

b. Hogaan 18 (14,● 21) Male b. Hogaan −2 (−6, 2) ● difference b. Hogaan 15 (12,● 18) Female 6 (−3, 16) ● a. Non−imp 19 (12,● 27) Male a. Non−imp −8 (−17, −1)* ● difference a. Non−imp 10 (6,● 15) Female

Respondent Involved in Planning Process Respondent Involved in Planning Process Respondent Involved in Planning Process

b. Hogaan 72 (56,● 82) Male b. Hogaan −10 (−15, −5)*** ● difference b. Hogaan 62 (46,● 75) Female 4 (−7, 15) ● a. Non−imp 67 (49,● 79) Male a. Non−imp −14 (−24, −4)** ● difference a. Non−imp 52 (36,● 67) Female

0% 25% 50% 75% 100% −40% −20% 0% −10% 0% 10% 20% 30% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 59: Comparison of Gender Groups on Participation Measures (4/4)

109 Do Women Have a Right to Protection Under the Law Do Women Have a Right to Protection Under the Law Do Women Have a Right to Protection Under the Law

b. Hogaan 98 (95,● 99) Male b. Hogaan 0 (−2, 2) ● difference b. Hogaan 98 (96,● 99) Female 0 (−7, 6) ●

a. Non−imp 98 (89,● 100) Male a. Non−imp 0 (−5, 7) ● difference a. Non−imp 99 (90,● 100) Female Should Women Have Same Rights as Men Should Women Have Same Rights as Men Should Women Have Same Rights as Men

b. Hogaan 68 (62,● 72) Male b. Hogaan 3 (−2, 8) ● difference b. Hogaan 71 (66,● 74) Female −8 (−20, 3) significance ● a● none a. Non−imp 70 (58,● 78) Male a. Non−imp 11 (1, 22)* ● difference a. Non−imp 81 (73,● 86) Female Do Women Have the Same Rights as Men Do Women Have the Same Rights as Men Do Women Have the Same Rights as Men

b. Hogaan 50 (41,● 58) Male b. Hogaan 1 (−5, 6) ● difference b. Hogaan 51 (42,● 58) Female 0 (−13, 13) ●

a. Non−imp 52 (40,● 63) Male a. Non−imp 1 (−12, 12) ● difference a. Non−imp 53 (41,● 62) Female

0% 25% 50% 75% 100% −10% 0% 10% 20% −20% −10% 0% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 60: Women’s Rights in Comparison to Men’s

Should Women Have Same Rights as Men Should Women Have Same Rights as Men 72 (67, 76) Female ● significance 4 (−1,● 9) 68 (62, 72) Female ● none Male ● a

0% 20% 40% 60% 80% 0.0% 2.5% 5.0% 7.5% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 61: Women’s Rights in Comparison to Men’s

Trust VC Trust VC Trust VC

b. Hogaan 84 (74,● 91) Male b. Hogaan 5 (1, 10)** ● difference b. Hogaan 90 (82,● 94) Female 0 (−10, 10) ●

a. Non−imp 75 (61,● 86) Male a. Non−imp 5 (−4, 15) ● difference a. Non−imp 81 (68,● 89) Female Trust DC Trust DC Trust DC

b. Hogaan 84 (71,● 91) Male b. Hogaan 5 (1, 10)** ● difference b. Hogaan 89 (78,● 94) Female 3 (−7, 13) ●

a. Non−imp 77 (62,● 88) Male a. Non−imp 2 (−6, 11) ● difference a. Non−imp 80 (66,● 89) Female significance Trust Elders Trust Elders Trust Elders a● none

b. Hogaan 92 (88,● 94) Male b. Hogaan 2 (−1, 6) ● difference b. Hogaan 94 (92,● 96) Female 2 (−7, 10) ●

a. Non−imp 92 (79,● 96) Male a. Non−imp 1 (−7, 9) ● difference a. Non−imp 92 (83,● 96) Female Trust Religious Leaders Trust Religious Leaders Trust Religious Leaders

b. Hogaan 96 (93,● 98) Male b. Hogaan 0 (−2, 3) ● difference b. Hogaan 96 (94,● 98) Female 1 (−7, 7) ●

a. Non−imp 94 (84,● 98) Male a. Non−imp 0 (−6, 7) ● difference a. Non−imp 94 (87,● 97) Female

0% 25% 50% 75% 100% −5% 0% 5% 10% 15% −10% −5% 0% 5% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 62: Comparison of Gender Groups on Trust Measures

110 Does VC Provide Services Does VC Provide Services Does VC Provide Services

b. hogaan 72 (55,● 82) male b. hogaan 0 (−5, 5) ● difference b. hogaan 71 (56,● 81) female 2 (−8, 12) ● a. non−implementation 75 (58,● 86) male a. non−implementation −3 (−12, 7) ● difference a. non−implementation 73 (55,● 83) female Can VC Manage Projects Can VC Manage Projects Can VC Manage Projects

b. hogaan 80 (67,● 87) male b. hogaan 0 (−5, 4) ● difference b. hogaan 79 (67,● 86) female −3 (−15, 7) ● a. non−implementation 69 (53,● 80) male a. non−implementation 3 (−6, 13) ● difference a. non−implementation 72 (57,● 82) female Does VC Manage Money Transparently and Accountably Does VC Manage Money Transparently and AccountablyDoes VC Manage Money Transparently and Accountably

b. hogaan 75 (62,● 83) male b. hogaan 1 (−3, 6) ● difference b. hogaan 76 (63,● 84) female 2 (−9, 12) ● a. non−implementation 71 (57,● 81) male a. non−implementation 0 (−9, 10) ● difference a. non−implementation 70 (57,● 80) female significance Does DC Provide Services Does DC Provide Services Does DC Provide Services a● none

b. hogaan 75 (63,● 83) male b. hogaan 1 (−4, 5) ● difference b. hogaan 75 (64,● 84) female 0 (−11, 10) ● a. non−implementation 71 (55,● 81) male a. non−implementation 1 (−9, 11) ● difference a. non−implementation 71 (58,● 82) female Can DC Manage Projects Can DC Manage Projects Can DC Manage Projects

b. hogaan 76 (65,● 83) male b. hogaan −1 (−6, 4) ● difference b. hogaan 75 (65,● 82) female −5 (−15, 7) ● a. non−implementation 70 (56,● 80) male a. non−implementation 3 (−6, 14) ● difference a. non−implementation 73 (61,● 82) female Does DC Manage Money Transparently and Accountably Does DC Manage Money Transparently and AccountablyDoes DC Manage Money Transparently and Accountably

b. hogaan 73 (62,● 79) male b. hogaan 0 (−5, 6) ● difference b. hogaan 73 (63,● 79) female −2 (−13, 9) ● a. non−implementation 73 (60,● 82) male a. non−implementation 2 (−7, 12) ● difference a. non−implementation 76 (65,● 83) female

0% 25% 50% 75% −10% −5% 0% 5% 10% −10% 0% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 63: Comparison of Gender Groups on Service Delivery Measures – Further Models

111 Women Involved in Planning Process Women Involved in Planning Process Women Involved in Planning Process

b. hogaan 82 (69,● 90) male b. hogaan 1 (−3, 6) ● difference b. hogaan 83 (70,● 91) female 4 (−6, 14) ●

a. non−implementation 73 (59,● 84) male a. non−implementation −2 (−12, 6) ● difference a. non−implementation 71 (56,● 81) female Youth Involved in Planning Process Youth Involved in Planning Process Youth Involved in Planning Process

b. hogaan 81 (69,● 89) male b. hogaan 1 (−3, 6) ● difference b. hogaan 82 (71,● 89) female 5 (−5, 15) ●

a. non−implementation 74 (58,● 84) male a. non−implementation −3 (−12, 5) ● difference a. non−implementation 70 (56,● 81) female significance Minority Clans Involved in Planning Process Minority Clans Involved in Planning Process Minority Clans Involved in Planning Process a● none

b. hogaan 83 (70,● 90) male b. hogaan 1 (−3, 6) ● difference b. hogaan 84 (72,● 91) female −3 (−13, 7) ●

a. non−implementation 72 (55,● 83) male a. non−implementation 4 (−4, 14) ● difference a. non−implementation 77 (62,● 86) female Outcast Clans Involved in Planning Process Outcast Clans Involved in Planning Process Outcast Clans Involved in Planning Process

b. hogaan 78 (63,● 87) male b. hogaan 2 (−2, 7) ● difference b. hogaan 80 (66,● 89) female 1 (−9, 12) ●

a. non−implementation 72 (57,● 84) male a. non−implementation 1 (−9, 10) ● difference a. non−implementation 73 (59,● 84) female

0% 25% 50% 75% −10% −5% 0% 5% 10% 15% −10% −5% 0% 5% 10% 15% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 64: Comparison of Gender Groups on Community Perceptions of Inclusion – Further Models

112 Contribution: International Development Project Contribution: International Development Project Contribution: International Development Project

b. hogaan 67 (51,● 79) male b. hogaan −3 (−7, 3) ● difference b. hogaan 65 (48,● 76) female 14 (2, 25)** ● a. non−implementation 47 (29,● 61) male a. non−implementation −16 (−26, −6)*** ● difference a. non−implementation 31 (16,● 45) female Contribution: Social Services Contribution: Social Services Contribution: Social Services

b. hogaan 72 (55,● 83) male b. hogaan −7 (−12, −2)** ● difference b. hogaan 65 (47,● 77) female 7 (−4, 19) ● a. non−implementation 63 (44,● 78) male a. non−implementation −15 (−24, −4)** ● difference a. non−implementation 48 (31,● 63) female Contribution: Poor and Needy Contribution: Poor and Needy Contribution: Poor and Needy

b. hogaan 86 (70,● 92) male b. hogaan −3 (−7, 1) ● difference b. hogaan 83 (66,● 90) significance female 6 (−4, 16) ● none ● a ● positive a. non−implementation 91 (73,● 96) a male a. non−implementation −8 (−18, 0)* ● difference a. non−implementation 82 (63,● 90) female Contribution: Security Contribution: Security Contribution: Security

b. hogaan 89 (70,● 95) male b. hogaan −6 (−12, −2)** ● difference b. hogaan 82 (63,● 91) female −3 (−13, 8) ● a. non−implementation 79 (60,● 90) male a. non−implementation −4 (−14, 4) ● difference a. non−implementation 75 (55,● 87) female Contribution: Blood Contribution: Blood Contribution: Blood

b. hogaan 84 (73,● 90) male b. hogaan −3 (−8, 1) ● difference b. hogaan 81 (68,● 87) female 4 (−6, 14) ● a. non−implementation 92 (80,● 96) male a. non−implementation −7 (−16, 0) ● difference a. non−implementation 85 (71,● 91) female

0% 25% 50% 75% 100% −20% −10% 0% −10% 0% 10% 20% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 65: Percentage Having Made Contributions of Different Kinds – Further Models

113 Priority: Life Priority: Life Priority: Life

b. hogaan 72 (58,● 83) male b. hogaan −1 (−6, 4) ● difference b. hogaan 72 (57,● 82) female 10 (0, 20) ●

a. non−implementation 87 (72,● 94) male a. non−implementation −11 (−20, −2)* ● difference a. non−implementation 76 (60,● 86) female

Priority: Justice Priority: Justice Priority: Justice

b. hogaan 55 (49,● 59) male b. hogaan −2 (−9, 3) ● difference b. hogaan 52 (48,● 56) female 2 (−11, 14) ●

a. non−implementation 60 (47,● 69) male a. non−implementation −4 (−15, 6) ● difference a. non−implementation 55 (45,● 62) female

Priority: Equality Priority: Equality Priority: Equality

b. hogaan 16 (12,● 20) male b. hogaan −1 (−5, 3) ● difference b. hogaan 15 (12,● 18) female 0 (−7, 8) ●

a. non−implementation9 (4,● 17) male a. non−implementation 0 (−8, 5) ● difference a. non−implementation9 (4,● 14) female

Priority: Education Priority: Education Priority: Education

b. hogaan 59 (48,● 67) male b. hogaan 1 (−5, 6) ● difference b. hogaan 59 (49,● 67) female −5 (−17, 6) significance ● a● none a. non−implementation 46 (31,● 57) male a. non−implementation 6 (−6, 17) ● difference a. non−implementation 52 (38,● 62) female

Priority: Health Priority: Health Priority: Health

b. hogaan 70 (64,● 75) male b. hogaan 1 (−4, 6) ● difference b. hogaan 71 (66,● 75) female −3 (−14, 8) ●

a. non−implementation 78 (66,● 85) male a. non−implementation 4 (−6, 14) ● difference a. non−implementation 81 (73,● 87) female

Priority: Consulted on Decisions Priority: Consulted on Decisions Priority: Consulted on Decisions

b. hogaan8 (4,● 15) male b. hogaan −2 (−6, 1) ● difference b. hogaan6 (3,● 11) female 1 (−5, 9) ●

a. non−implementation6 (2,● 14) male a. non−implementation −3 (−11, 2) ● difference a. non−implementation3 (1,● 7) female

Priority: Free Speech Priority: Free Speech Priority: Free Speech

b. hogaan5 (2,● 12) male b. hogaan 0 (−3, 2) ● difference b. hogaan5 (2,● 11) female 1 (−4, 8) ●

a. non−implementation3 (1,● 11) male a. non−implementation −1 (−8, 3) ● difference a. non−implementation2 (0,● 7) female

0% 25% 50% 75% −20% −10% 0% 10% −10% 0% 10% 20% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 66: Comparison of Gender Groups on Top Three Rights Priorities – Further Models

114 Do Youth Have a Right to Protection Under the Law Do Youth Have a Right to Protection Under the Law Do Youth Have a Right to Protection Under the Law

b. hogaan 96 (91,● 99) male b. hogaan 0 (−2, 3) ● difference b. hogaan 97 (93,● 99) female −2 (−10, 4) ● a. non−implementation 94 (83,● 98) male a. non−implementation 2 (−3, 10) ● difference a. non−implementation 97 (88,● 99) female Should Youth Have Same Rights as Elderly People Should Youth Have Same Rights as Elderly People Should Youth Have Same Rights as Elderly People

b. hogaan 78 (73,● 81) male b. hogaan −1 (−6, 4) ● difference b. hogaan 77 (73,● 79) female −7 (−17, 2) significance ● a● none a. non−implementation 83 (72,● 89) male a. non−implementation 6 (−3, 15) ● difference a. non−implementation 89 (82,● 93) female Do Youth Have the Same Rights as The Elderly Do Youth Have the Same Rights as The Elderly Do Youth Have the Same Rights as The Elderly

b. hogaan 59 (51,● 65) male b. hogaan −3 (−8, 3) ● difference b. hogaan 56 (49,● 62) female −13 (−25, 0) ● a. non−implementation 49 (36,● 59) male a. non−implementation 10 (−1, 22) ● difference a. non−implementation 59 (49,● 67) female

0% 25% 50% 75% 100% 0% 10% 20% −20% −10% 0% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 67: Youth Rights in Comparison to Older People – Further Models

All Clans Equal Right Legal Protection All Clans Equal Right Legal Protection All Clans Equal Right Legal Protection

b. hogaan 98 (95,● 99) male b. hogaan 0 (−1, 3) ● difference b. hogaan 98 (97,● 99) female 4 (−4, 12) ● a. non−implementation 97 (87,● 99) male a. non−implementation −3 (−11, 4) ● difference a. non−implementation 93 (83,● 97) female Should Minority Clans Have Equal Rights with Majority Clans Should Minority Clans Have Equal Rights with MajorityShould Clans Minority Clans Have Equal Rights with Majority Clans

b. hogaan 95 (91,● 97) male b. hogaan −1 (−4, 2) ● difference b. hogaan 94 (89,● 96) female 2 (−5, 10) ● a. non−implementation 94 (85,● 98) male a. non−implementation −4 (−11, 3) ● difference a. non−implementation 90 (81,● 95) female Should Outcast Clans Equal Right with Majority Clans Should Outcast Clans Equal Right with Majority Clans Should Outcast Clans Equal Right with Majority Clans

b. hogaan 77 (53,● 91) male b. hogaan 2 (−2, 6) ● difference b. hogaan 79 (54,● 92) female 1 (−8, 10) significance ● a● none a. non−implementation 82 (58,● 95) male a. non−implementation 1 (−8, 9) ● difference a. non−implementation 83 (60,● 95) female Do Minority Clans Have Equal Rights with Majority Clans Do Minority Clans Have Equal Rights with Majority ClansDo Minority Clans Have Equal Rights with Majority Clans

b. hogaan 71 (66,● 75) male b. hogaan 1 (−4, 6) ● difference b. hogaan 72 (68,● 76) female 10 (−2, 21) ● a. non−implementation 78 (66,● 86) male a. non−implementation −9 (−18, 2) ● difference a. non−implementation 70 (58,● 78) female Do Outcast Clans Have Equal Rights with Majority Clans Do Outcast Clans Have Equal Rights with Majority ClansDo Outcast Clans Have Equal Rights with Majority Clans

b. hogaan 62 (57,● 67) male b. hogaan 2 (−3, 8) ● difference b. hogaan 65 (60,● 68) female 2 (−10, 14) ● a. non−implementation 63 (50,● 73) male a. non−implementation 0 (−11, 11) ● difference a. non−implementation 63 (53,● 71) female

0% 25% 50% 75% 100% −10% 0% 10% −10% 0% 10% 20% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 68: Minority and Outcast Clan in Comparison to Majority Clan – Further Models

115 Conflict Mechansims Effective Overall Conflict Mechansims Effective Overall Conflict Mechansims Effective Overall

b. hogaan 77 (63,● 86) male b. hogaan 7 (2, 12)*** ● difference b. hogaan 84 (71,● 91) female −4 (−15, 7) ● a. non−implementation 75 (59,● 85) male a. non−implementation 10 (2, 21)* ● difference a. non−implementation 86 (72,● 93) female Conflicts Resolved Peacefully Conflicts Resolved Peacefully Conflicts Resolved Peacefully

b. hogaan 92 (88,● 95) male b. hogaan 3 (0, 6) ● difference b. hogaan 95 (92,● 97) female 1 (−8, 10) ● a. non−implementation 88 (75,● 94) male a. non−implementation 1 (−7, 11) ● difference a. non−implementation 90 (81,● 94) female Women's Issues in Conflict Resolution Women's Issues in Conflict Resolution Women's Issues in Conflict Resolution

b. hogaan 79 (72,● 84) male b. hogaan −1 (−6, 4) ● difference b. hogaan 78 (72,● 83) female −1 (−13, 10) significance ● a● none a. non−implementation 70 (57,● 79) male a. non−implementation 1 (−10, 11) ● difference a. non−implementation 70 (59,● 78) female Youth Issues in Conflict Resolution Youth Issues in Conflict Resolution Youth Issues in Conflict Resolution

b. hogaan 76 (67,● 83) male b. hogaan 1 (−4, 6) ● difference b. hogaan 77 (69,● 84) female −1 (−13, 11) ● a. non−implementation 66 (53,● 76) male a. non−implementation 2 (−9, 13) ● difference a. non−implementation 69 (55,● 77) female Minority Clan Issues in Conflict Resolution Minority Clan Issues in Conflict Resolution Minority Clan Issues in Conflict Resolution

b. hogaan 72 (61,● 79) male b. hogaan 6 (1, 11)** ● difference b. hogaan 78 (69,● 84) female 6 (−6, 17) ● a. non−implementation 72 (59,● 82) male a. non−implementation 0 (−10, 10) ● difference a. non−implementation 72 (60,● 81) female

0% 25% 50% 75% 100% −10% 0% 10% 20% −10% 0% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 69: Comparison of Gender Groups on Conflict Mechanisms – Further Models

116 Resolution: Clan Issues Resolution: Clan Issues Resolution: Clan Issues

b. hogaan 91 (84,● 94) male b. hogaan −1 (−4, 3) ● difference b. hogaan 90 (83,● 93) female 7 (−1, 16) ● a. non−implementation 94 (84,● 98) male a. non−implementation −8 (−17, −1)* ● difference a. non−implementation 86 (74,● 91) female Resolution: Water Issues Resolution: Water Issues Resolution: Water Issues

b. hogaan 91 (85,● 94) male b. hogaan 0 (−3, 3) ● difference b. hogaan 91 (85,● 93) female −2 (−10, 6) ● a. non−implementation 87 (74,● 93) male a. non−implementation 2 (−5, 9) ● difference a. non−implementation 89 (79,● 94) female Resolution: Grazing Issues Resolution: Grazing Issues Resolution: Grazing Issues

b. hogaan 87 (78,● 91) male b. hogaan 2 (−2, 6) ● difference b. hogaan 88 (82,● 93) female 3 (−5, 10) ● a. non−implementation 90 (78,● 95) male a. non−implementation −1 (−7, 7) ● difference a. non−implementation 89 (79,● 94) female significance Resolution: Land Issues Resolution: Land Issues Resolution: Land Issues a● none

b. hogaan 85 (75,● 91) male b. hogaan 2 (−2, 6) ● difference b. hogaan 87 (78,● 92) female 2 (−8, 10) ● a. non−implementation 89 (77,● 95) male a. non−implementation 0 (−7, 9) ● difference a. non−implementation 89 (79,● 94) female Resolution: Charcoal Issues Resolution: Charcoal Issues Resolution: Charcoal Issues

b. hogaan 78 (66,● 84) male b. hogaan 7 (2, 13)** ● difference b. hogaan 85 (77,● 89) female 3 (−8, 14) ● a. non−implementation 82 (68,● 90) male a. non−implementation 4 (−5, 14) ● difference a. non−implementation 86 (75,● 92) female Resolution: Leadership Issues Resolution: Leadership Issues Resolution: Leadership Issues

b. hogaan 87 (79,● 92) male b. hogaan 3 (−1, 7) ● difference b. hogaan 90 (83,● 94) female −6 (−17, 4) ● a. non−implementation 73 (58,● 82) male a. non−implementation 9 (0, 19) ● difference a. non−implementation 82 (69,● 89) female

0% 25% 50% 75% 100% −10% 0% 10% 20% −10% 0% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 70: Comparison of Gender Groups on Issue Resolution – Further Models

117 Elders: Priority Women Elders: Priority Women Elders: Priority Women

b. hogaan 83 (64,● 94) male b. hogaan1 (−3, 5) ● difference b. hogaan 84 (64,● 94) female −3 (−12, 4) ● a. non−implementation 86 (65,● 95) male a. non−implementation3 (−2, 12) ● difference a. non−implementation 90 (71,● 97) female

Elders: Priority Youth Elders: Priority Youth Elders: Priority Youth

b. hogaan 80 (61,● 92) male b. hogaan3 (−1, 8) ● difference b. hogaan 84 (66,● 94) female 0 (−8, 7) ● a. non−implementation 83 (63,● 95) male a. non−implementation3 (−3, 11) ● difference a. non−implementation 86 (68,● 96) female

Elders: Priority Outcast Clans Elders: Priority Outcast Clans Elders: Priority Outcast Clans

b. hogaan 73 (51,● 88) male b. hogaan1 (−4, 5) ● difference b. hogaan 74 (53,● 88) female −3 (−13, 7) ● a. non−implementation 69 (47,● 86) male a. non−implementation4 (−5, 13) ● difference a. non−implementation 74 (51,● 89) female

Relgious Leaders: Priority Women Relgious Leaders: Priority Women Relgious Leaders: Priority Women

b. hogaan 96 (86,● 99) male b. hogaan−2 (−6, 1) ● difference b. hogaan 94 (83,● 98) female −5 (−103, 3) ● a. non−implementation 100 (0,● 100) male a. non−implementation0 (−4, 100) ● difference a. non−implementation 99 (93,● 100) female

Religious Leaders: Priority Youth Religious Leaders: Priority Youth Religious Leaders: Priority Youth

b. hogaan 90 (78,● 96) male b. hogaan2 (−2, 6) ● difference b. hogaan 92 (81,● 97) female −1 (−9, 6) ● a. non−implementation 93 (79,● 98) male a. non−implementation2 (−2, 10) ● difference a. non−implementation 96 (86,● 99) female

Religious Leaders: Priority Outcast Clans Religious Leaders: Priority Outcast Clans Religious Leaders: Priority Outcast Clans

b. hogaan 82 (57,● 95) male b. hogaan0 (−3, 3) ● difference b. hogaan 82 (56,● 94) female −2 (−11, 6) ● a. non−implementation 89 (64,● 98) male a. non−implementation1 (−5, 11) ● difference a. non−implementation 91 (69,● 98)

female significance

● negative VC: Priority Women VC: Priority Women VC: Priority Women a a● none b. hogaan 87 (67,● 96) male b. hogaan−1 (−5, 3) ● difference b. hogaan 85 (67,● 95) female −7 (−17, 0)* ● a. non−implementation 85 (64,● 95) male a. non−implementation5 (0, 15)* ● difference a. non−implementation 91 (72,● 98) female

VC: Priority Youth VC: Priority Youth VC: Priority Youth

b. hogaan 81 (64,● 92) male b. hogaan2 (−2, 6) ● difference b. hogaan 83 (66,● 94) female −2 (−11, 6) ● a. non−implementation 82 (63,● 93) male a. non−implementation4 (−2, 11) ● difference a. non−implementation 86 (68,● 95) female

VC: Priority Outcast Clans VC: Priority Outcast Clans VC: Priority Outcast Clans

b. hogaan 74 (50,● 91) male b. hogaan0 (−4, 4) ● difference b. hogaan 74 (51,● 90) female −10 (−21, −1)* ● a. non−implementation 66 (44,● 86) male a. non−implementation10 (2, 19)** ● difference a. non−implementation 77 (53,● 93) female

DC: Priority Women DC: Priority Women DC: Priority Women

b. hogaan 85 (67,● 95) male b. hogaan0 (−4, 4) ● difference b. hogaan 85 (68,● 94) female −2 (−10, 5) ● a. non−implementation 85 (66,● 95) male a. non−implementation2 (−4, 9) ● difference a. non−implementation 87 (69,● 96) female

DC: Priority Youth DC: Priority Youth DC: Priority Youth

b. hogaan 79 (61,● 91) male b. hogaan 4 (0, 8) ● difference b. hogaan 84 (67,● 93) female −2 (−11, 7) ● a. non−implementation 79 (60,● 91) male a. non−implementation5 (−2, 14) ● difference a. non−implementation 85 (67,● 95) female

DC: Priority Outcast Clans DC: Priority Outcast Clans DC: Priority Outcast Clans

b. hogaan 72 (51,● 87) male b. hogaan1 (−3, 6) ● difference b. hogaan 74 (52,● 89) female −4 (−15, 5) ● a. non−implementation 70 (48,● 86) male a. non−implementation6 (−2, 15) ● difference a. non−implementation 76 (54,● 90) female

0% 25% 50% 75% 100% 0% 25% 50% 75% 100% −100% −75% −50% −25% 0% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 71: Perceptions of Institutional Inclusion – Further Models

118 B.3 Further Model Results (Clan Interaction)

VC: Priority Women VC: Priority Women 76 (53, 91) Outcast ● −11 (−21, −3)*** Outcast ● 84 (66, 95) Minority ● −3 (−7, 1) 88 (69, 96) Minority ● Majority ●

VC: Priority Youth VC: Priority Youth 74 (52, 90) Outcast ● −9 (−20, −2)** Outcast ● 81 (62, 92) Minority ● −4 (−8, 0) 85 (67, 94) Minority ● Majority ●

VC: Priority Outcast Clans VC: Priority Outcast Clans 68 (46, 87) Outcast ● −7 (−15, 1) Outcast ● 72 (50, 88) Minority ● −3 (−7, 1) 75 (52, 91) Minority ● Majority ● significance

● DC: Priority Women DC: Priority Women a negative a● none 76 (56, 90) Outcast ● −9 (−19, −2)** Outcast ● 84 (66, 94) Minority ● −2 (−6, 1) 86 (68, 95) Minority ● Majority ●

DC: Priority Youth DC: Priority Youth 75 (54, 88) Outcast ● −8 (−18, 0)* Outcast ● 81 (66, 92) Minority ● −2 (−6, 2) 83 (67, 93) Minority ● Majority ●

DC: Priority Outcast Clans DC: Priority Outcast Clans 66 (46, 83) Outcast ● −8 (−17, 0) Outcast ● 71 (51, 87) Minority ● −3 (−7, 1) 74 (55, 89) Minority ● Majority ●

0% 25% 50% 75% 100% −20% −15% −10% −5% 0% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 72: Perceptions of Institutional Inclusion, by Clan Groups

119 Trust VC Trust VC Trust VC

b. Hogaan 88 (78,● 93) Minority b. Hogaan −2 (−6, 2) ● difference b. Hogaan 90 (81,● 94) Majority 0 (−10, 10) ● a. Non−imp 79 (65,● 87) Minority a. Non−imp −2 (−11, 7) ● difference a. Non−imp 81 (69,● 88) Majority Trust DC Trust DC Trust DC

b. Hogaan 87 (74,● 93) Minority b. Hogaan −2 (−7, 2) ● difference b. Hogaan 89 (78,● 94) Majority −6 (−16, 2) ● a. Non−imp 83 (68,● 92) Minority a. Non−imp 4 (−4, 13) ● difference a. Non−imp 79 (65,● 87) Majority significance Trust Elders Trust Elders Trust Elders a● none

b. Hogaan 92 (89,● 95) Minority b. Hogaan −3 (−6, 0) ● difference b. Hogaan 95 (93,● 97) Majority 1 (−8, 12) ● a. Non−imp 90 (76,● 95) Minority a. Non−imp −3 (−13, 4) ● difference a. Non−imp 93 (85,● 97) Majority Trust Religious Leaders Trust Religious Leaders Trust Religious Leaders

b. Hogaan 95 (91,● 97) Minority b. Hogaan −2 (−6, 0)* ● difference b. Hogaan 97 (95,● 98) Majority −2 (−9, 6) ● a. Non−imp 94 (83,● 98) Minority a. Non−imp 0 (−8, 5) ● difference a. Non−imp 94 (87,● 97) Majority

0% 25% 50% 75% 100% −10% −5% 0% 5% 10% −10% 0% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 73: Comparison of clan Groups on Trust Measures

VC Meetings Consider my Household Viewpoint VC Meetings Consider my Household Viewpoint VC Meetings Consider my Household Viewpoint

b. Hogaan 95 (88,● 98) Minority b. Hogaan 0 (−6, 4) ●

b. Hogaan 95 (91,● 98) difference Majority 3 (−6, 21) significance ● a● none a. Non−implementation 94 (69,● 99) Minority a. Non−implementation −3 (−21, 4) ● difference a. Non−implementation 97 (87,● 100) Majority

0% 25% 50% 75% 100% −20% −15% −10% −5% 0% 5% 0% 10% 20% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 74: Participation: Comparison Clan – Further Models

120 Does VC Provide Services Does VC Provide Services Does VC Provide Services

b. Hogaan 70 (54,● 79) Minority b. Hogaan −3 (−8, 2) ● difference b. Hogaan 73 (57,● 82) Majority 5 (−5, 17) ● a. Non−implementation 68 (49,● 80) Minority a. Non−implementation −8 (−18, 1) ● difference a. Non−implementation 77 (62,● 86) Majority Can VC Manage Projects Can VC Manage Projects Can VC Manage Projects

b. Hogaan 78 (65,● 85) Minority b. Hogaan −3 (−7, 2) ● difference b. Hogaan 81 (68,● 87) Majority −1 (−13, 11) ● a. Non−implementation 70 (55,● 82) Minority a. Non−implementation −2 (−12, 8) ● difference a. Non−implementation 72 (58,● 82) Majority Does VC Manage Money Transparently and Accountably Does VC Manage Money Transparently and AccountablyDoes VC Manage Money Transparently and Accountably

b. Hogaan 73 (62,● 81) Minority b. Hogaan −4 (−9, 1) ● difference b. Hogaan 77 (66,● 84) Majority 2 (−9, 13) ● a. Non−implementation 67 (51,● 78) Minority a. Non−implementation −6 (−16, 5) ● difference a. Non−implementation 72 (60,● 82) Majority significance Does DC Provide Services Does DC Provide Services Does DC Provide Services a● none

b. Hogaan 72 (60,● 81) Minority b. Hogaan −4 (−9, 0) ● difference b. Hogaan 77 (66,● 84) Majority −3 (−14, 9) ● a. Non−implementation 71 (55,● 83) Minority a. Non−implementation −1 (−12, 9) ● difference a. Non−implementation 73 (60,● 82) Majority Can DC Manage Projects Can DC Manage Projects Can DC Manage Projects

b. Hogaan 71 (60,● 79) Minority b. Hogaan −7 (−12, −2)** ● difference b. Hogaan 78 (68,● 84) Majority −5 (−16, 6) ● a. Non−implementation 72 (58,● 81) Minority a. Non−implementation −1 (−12, 8) ● difference a. Non−implementation 73 (61,● 81) Majority Does DC Manage Money Transparently and Accountably Does DC Manage Money Transparently and AccountablyDoes DC Manage Money Transparently and Accountably

b. Hogaan 70 (60,● 76) Minority b. Hogaan −6 (−11, −1)** ● difference b. Hogaan 75 (68,● 81) Majority −5 (−15, 7) ● a. Non−implementation 74 (61,● 84) Minority a. Non−implementation −1 (−12, 9) ● difference a. Non−implementation 76 (66,● 82) Majority

0% 25% 50% 75% −10% 0% 10% −10% 0% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 75: Comparison of clan Groups on Service Delivery Measures – Further Models

121 Women Involved in Planning Process Women Involved in Planning Process Women Involved in Planning Process

b. Hogaan 83 (70,● 90) Minority b. Hogaan −1 (−5, 4) ● difference b. Hogaan 84 (71,● 91) Majority −1 (−12, 8) ●

a. Non−implementation 73 (58,● 85) Minority a. Non−implementation 1 (−9, 9) ● difference a. Non−implementation 72 (58,● 83) Majority Youth Involved in Planning Process Youth Involved in Planning Process Youth Involved in Planning Process

b. Hogaan 82 (69,● 89) Minority b. Hogaan −1 (−6, 4) ● difference b. Hogaan 83 (72,● 89) Majority 0 (−9, 11) ●

a. Non−implementation 72 (57,● 83) Minority a. Non−implementation −1 (−11, 8) ● difference a. Non−implementation 73 (59,● 83) Majority significance Minority Clans Involved in Planning Process Minority Clans Involved in Planning Process Minority Clans Involved in Planning Process a● none

b. Hogaan 81 (69,● 88) Minority b. Hogaan −6 (−11, −1)** ● difference b. Hogaan 87 (76,● 92) Majority 2 (−9, 14) ●

a. Non−implementation 70 (54,● 82) Minority a. Non−implementation −8 (−18, 1) ● difference a. Non−implementation 79 (64,● 87) Majority Outcast Clans Involved in Planning Process Outcast Clans Involved in Planning Process Outcast Clans Involved in Planning Process

b. Hogaan 79 (63,● 88) Minority b. Hogaan −3 (−8, 2) ● difference b. Hogaan 82 (67,● 89) Majority 4 (−8, 15) ●

a. Non−implementation 70 (52,● 83) Minority a. Non−implementation −7 (−17, 2) ● difference a. Non−implementation 77 (62,● 87) Majority

0% 25% 50% 75% −10% 0% 10% −10% 0% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 76: Comparison of clan Groups on Community Perceptions of Inclusion – Further Models

Contribution: International Development Project Contribution: International Development Project Contribution: International Development Project

b. Hogaan 65 (49,● 75) Minority b. Hogaan −1 (−6, 4) ● difference b. Hogaan 66 (51,● 77) Majority 15 (2, 28)* ●

a. Non−implementation 23 (9,● 39) Minority a. Non−implementation −17 (−28, −4)** ● difference a. Non−implementation 41 (24,● 54) Majority Contribution: Social Services Contribution: Social Services Contribution: Social Services

b. Hogaan 66 (50,● 78) Minority b. Hogaan −2 (−8, 2) ● difference b. Hogaan 69 (52,● 80) Majority −1 (−12, 11) ●

a. Non−implementation 53 (34,● 68) Minority a. Non−implementation −2 (−12, 9) ● difference a. Non−implementation 55 (37,● 69) Majority Contribution: Poor and Needy Contribution: Poor and Needy Contribution: Poor and Needy

b. Hogaan 84 (65,● 90) Minority b. Hogaan −1 (−5, 3) ● difference b. Hogaan 85 (66,● 90) significance Majority 5 (−3, 16) ● none ● a ● positive a. Non−implementation 80 (59,● 89) a Minority a. Non−implementation −7 (−16, 1) ● difference a. Non−implementation 87 (68,● 93) Majority Contribution: Security Contribution: Security Contribution: Security

b. Hogaan 83 (65,● 91) Minority b. Hogaan −2 (−7, 1) ● difference b. Hogaan 85 (67,● 93) Majority 3 (−7, 12) ●

a. Non−implementation 72 (52,● 85) Minority a. Non−implementation −5 (−14, 4) ● difference a. Non−implementation 77 (60,● 88) Majority Contribution: Blood Contribution: Blood Contribution: Blood

b. Hogaan 79 (65,● 87) Minority b. Hogaan −4 (−9, 1) ● difference b. Hogaan 83 (70,● 88) Majority −4 (−13, 7) ●

a. Non−implementation 87 (72,● 94) Minority a. Non−implementation 0 (−9, 8) ● difference a. Non−implementation 87 (75,● 92) Majority

0% 25% 50% 75% −20% −10% 0% 10% −10% 0% 10% 20% 30% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 77: Percentage Having Made Contributions of Different Kinds – Further Models

122 Priority: Life Priority: Life Priority: Life

b. Hogaan 72 (57,● 82) Minority b. Hogaan 0 (−5, 4) ● difference b. Hogaan 72 (58,● 83) Majority 7 (−3, 18) ● a. Non−implementation 75 (57,● 86) Minority a. Non−implementation −7 (−17, 2) ● difference a. Non−implementation 83 (68,● 90) Majority

Priority: Justice Priority: Justice Priority: Justice

b. Hogaan 55 (49,● 59) Minority b. Hogaan 2 (−3, 8) ● difference b. Hogaan 52 (48,● 56) Majority 17 (3, 31)** ● a. Non−implementation 46 (33,● 56) Minority a. Non−implementation −15 (−27, −3)** ● difference a. Non−implementation 61 (52,● 69) Majority

Priority: Equality Priority: Equality Priority: Equality

b. Hogaan13 (10,● 17) Minority b. Hogaan −2 (−6, 2) ● difference b. Hogaan 16 (12,● 19) Majority −10 (−19, −1)** ● a. Non−implementation 14 (7,● 22) Minority a. Non−implementation 7 (0, 16)* ● difference a. Non−implementation7 (3,● 11) Majority

Priority: Education Priority: Education Priority: Education

b. Hogaan 57 (45,● 66) Minority b. Hogaan −3 (−9, 2) ● difference significance b. Hogaan 60 (50,● 68) Majority 2 (−11, 16) a● negative ● a● none a. Non−implementation 45 (31,● 57) ● positive Minority a. Non−implementation −6 (−18, 6) a ● difference a. Non−implementation 51 (39,● 61) Majority

Priority: Health Priority: Health Priority: Health

b. Hogaan 72 (66,● 77) Minority b. Hogaan 2 (−4, 7) ● difference b. Hogaan 70 (65,● 74) Majority −5 (−16, 6) ● a. Non−implementation 85 (73,● 91) Minority a. Non−implementation 7 (−3, 16) ● difference a. Non−implementation 77 (69,● 83) Majority

Priority: Consulted on Decisions Priority: Consulted on Decisions Priority: Consulted on Decisions

b. Hogaan8 (4,● 13) Minority b. Hogaan 2 (−1, 5) ● difference b. Hogaan6 (3,● 10) Majority −3 (−12, 3) ● a. Non−implementation7 (2,● 16) Minority a. Non−implementation 5 (0, 13) ● difference a. Non−implementation2 (1,● 6) Majority

Priority: Free Speech Priority: Free Speech Priority: Free Speech

b. Hogaan3 (1,● 7) Minority b. Hogaan −2 (−5, 1) ● difference b. Hogaan5 (2,● 11) Majority −4 (−13, 1) ● a. Non−implementation3 (1,● 13) Minority a. Non−implementation 2 (−2, 10) ● difference a. Non−implementation2 (0,● 6) Majority

0% 25% 50% 75% −20% −10% 0% 10% −20% −10% 0% 10% 20% 30% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 78: Comparison of clan Groups on Top Three Rights Priorities – Further Models

123 All Clans Equal Right Legal Protection All Clans Equal Right Legal Protection All Clans Equal Right Legal Protection

b. Hogaan 98 (95,● 99) Minority b. Hogaan −1 (−3, 1) ● difference b. Hogaan 99 (98,● 99) Majrity −2 (−9, 6) ● a. Non−implementation 96 (84,● 99) Minority a. Non−implementation 1 (−7, 7) ● difference a. Non−implementation 95 (87,● 98) Majrity Should Minority Clans Have Equal Rights with Majority Clans Should Minority Clans Have Equal Rights with MajorityShould Clans Minority Clans Have Equal Rights with Majority Clans

b. Hogaan 94 (90,● 97) Minority b. Hogaan −1 (−4, 2) ● difference b. Hogaan 95 (92,● 97) Majrity −6 (−14, 3) ● a. Non−implementation 96 (85,● 99) Minority a. Non−implementation 5 (−3, 12) ● difference a. Non−implementation 91 (82,● 95) Majrity Should Outcast Clans Equal Right with Majority Clans Should Outcast Clans Equal Right with Majority Clans Should Outcast Clans Equal Right with Majority Clans

b. Hogaan 78 (56,● 93) Minority b. Hogaan −2 (−6, 2) ● difference b. Hogaan 80 (58,● 94) Majrity −3 (−12, 8) significance ● a● none a. Non−implementation 82 (57,● 95) Minority a. Non−implementation 1 (−9, 10) ● difference a. Non−implementation 81 (58,● 94) Majrity Do Minority Clans Have Equal Rights with Majority Clans Do Minority Clans Have Equal Rights with Majority ClansDo Minority Clans Have Equal Rights with Majority Clans

b. Hogaan 70 (65,● 74) Minority b. Hogaan −5 (−10, 0) ● difference b. Hogaan 74 (70,● 77) Majrity −2 (−14, 10) ● a. Non−implementation 72 (56,● 81) Minority a. Non−implementation −2 (−14, 8) ● difference a. Non−implementation 74 (64,● 81) Majrity Do Outcast Clans Have Equal Rights with Majority Clans Do Outcast Clans Have Equal Rights with Majority ClansDo Outcast Clans Have Equal Rights with Majority Clans

b. Hogaan 62 (57,● 67) Minority b. Hogaan −3 (−8, 2) ● difference b. Hogaan 65 (61,● 69) Majrity −1 (−13, 13) ● a. Non−implementation 62 (48,● 72) Minority a. Non−implementation −2 (−15, 9) ● difference a. Non−implementation 65 (54,● 72) Majrity

0% 25% 50% 75% 100% −15% −10% −5% 0% 5% 10% −15% −10% −5% 0% 5% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 79: Minority and Outcast Clan in Comparison to Majority Clan – Further Models

124 Resolution: Clan Issues Resolution: Clan Issues Resolution: Clan Issues

b. hogaan 91 (85,● 93) reference b. hogaan −1 (−4, 3) ● difference b. hogaan 90 (82,● 93) Majority_Clan.3Minority 11 (1, 22)* ●

a. non−implementation 92 (83,● 96) reference a. non−implementation −11 (−23, −2)** ● difference a. non−implementation 80 (64,● 89) Majority_Clan.3Minority Resolution: Water Issues Resolution: Water Issues Resolution: Water Issues

b. hogaan 91 (86,● 94) reference b. hogaan −1 (−5, 2) ● difference b. hogaan 89 (82,● 93) Majority_Clan.3Minority 2 (−6, 14) ●

a. non−implementation 89 (79,● 94) reference a. non−implementation −4 (−15, 4) ● difference a. non−implementation 85 (69,● 92) Majority_Clan.3Minority Resolution: Grazing Issues Resolution: Grazing Issues Resolution: Grazing Issues

b. hogaan 88 (81,● 92) reference b. hogaan 0 (−5, 3) ● difference b. hogaan 88 (80,● 92) Majority_Clan.3Minority 0 (−9, 9) ●

a. non−implementation 90 (82,● 95) reference a. non−implementation 0 (−8, 7) ● difference a. non−implementation 90 (79,● 96)

Majority_Clan.3Minority significance

● none Resolution: Land Issues Resolution: Land Issues Resolution: Land Issues a a● positive b. hogaan 86 (79,● 91) reference b. hogaan 2 (−2, 6) ● difference b. hogaan 88 (80,● 92) Majority_Clan.3Minority 1 (−8, 10) ●

a. non−implementation 89 (80,● 94) reference a. non−implementation 1 (−7, 9) ● difference a. non−implementation 90 (78,● 96) Majority_Clan.3Minority Resolution: Charcoal Issues Resolution: Charcoal Issues Resolution: Charcoal Issues

b. hogaan 84 (74,● 88) reference b. hogaan −2 (−6, 3) ● difference b. hogaan 82 (72,● 87) Majority_Clan.3Minority −3 (−13, 8) ●

a. non−implementation 85 (74,● 91) reference a. non−implementation 1 (−9, 10) ● difference a. non−implementation 86 (71,● 93) Majority_Clan.3Minority Resolution: Leadership Issues Resolution: Leadership Issues Resolution: Leadership Issues

b. hogaan 90 (83,● 93) reference b. hogaan 0 (−5, 3) ● difference b. hogaan 89 (82,● 93) Majority_Clan.3Minority 0 (−10, 11) ●

a. non−implementation 79 (66,● 87) reference a. non−implementation −1 (−11, 9) ● difference a. non−implementation 78 (63,● 88) Majority_Clan.3Minority

125 Elders: Priority Women Elders: Priority Women Elders: Priority Women

b. Hogaan 81 (60,● 92) Minority b. Hogaan −4 (−9, −1)** ● difference b. Hogaan 86 (67,● 95) Majority −1 (−9, 8) ● a. Non−implementation 86 (67,● 95) Minority a. Non−implementation −3 (−11, 2) ● difference a. Non−implementation 90 (72,● 97) Majority

Elders: Priority Youth Elders: Priority Youth Elders: Priority Youth

b. Hogaan 82 (63,● 93) Minority b. Hogaan 0 (−5, 3) ● difference b. Hogaan 83 (64,● 93) Majority 1 (−7, 8) ● a. Non−implementation 85 (66,● 95) Minority a. Non−implementation −1 (−8, 5) ● difference a. Non−implementation 86 (69,● 95) Majority

Elders: Priority Outcast Clans Elders: Priority Outcast Clans Elders: Priority Outcast Clans

b. Hogaan 74 (51,● 88) Minority b. Hogaan −1 (−6, 3) ● difference b. Hogaan 75 (52,● 89) Majority 1 (−8, 11) ● a. Non−implementation 71 (49,● 88) Minority a. Non−implementation −2 (−12, 6) ● difference a. Non−implementation 74 (52,● 89) Majority

Relgious Leaders: Priority Women Relgious Leaders: Priority Women Relgious Leaders: Priority Women

b. Hogaan 91 (78,● 96) Minority b. Hogaan −5 (−11, −2)*** ● difference b. Hogaan 96 (88,● 99) Majority −3 (−10, 4) ● a. Non−implementation 98 (87,● 100) Minority a. Non−implementation −1 (−8, 1) ● difference a. Non−implementation 99 (95,● 100) Majority

Religious Leaders: Priority Youth Religious Leaders: Priority Youth Religious Leaders: Priority Youth

b. Hogaan 88 (76,● 94) Minority b. Hogaan −5 (−9, −1)** ● difference b. Hogaan 93 (84,● 97) Majority −3 (−9, 5) ● a. Non−implementation 93 (80,● 98) Minority a. Non−implementation −2 (−9, 3) ● difference a. Non−implementation 95 (85,● 98) Majority

Religious Leaders: Priority Outcast Clans Religious Leaders: Priority Outcast Clans Religious Leaders: Priority Outcast Clans

b. Hogaan 79 (55,● 94) Minority b. Hogaan −3 (−7, 1) ● difference b. Hogaan 82 (58,● 95) Majority −6 (−18, 3) ● a. Non−implementation 93 (72,● 99) Minority a. Non−implementation 3 (−5, 14) ● difference a. Non−implementation 89 (67,● 98) Majority significance VC: Priority Women VC: Priority Women VC: Priority Women a● none

b. Hogaan 85 (67,● 94) Minority b. Hogaan −2 (−7, 1) ● difference b. Hogaan 88 (69,● 96) Majority 2 (−5, 11) ● a. Non−implementation 87 (68,● 96) Minority a. Non−implementation −5 (−13, 1) ● difference a. Non−implementation 92 (74,● 98) Majority

VC: Priority Youth VC: Priority Youth VC: Priority Youth

b. Hogaan 81 (64,● 92) Minority b. Hogaan −3 (−8, 1) ● difference b. Hogaan 85 (68,● 94) Majority 2 (−6, 11) ● a. Non−implementation 80 (62,● 92) Minority a. Non−implementation −5 (−14, 1) ● difference a. Non−implementation 87 (69,● 95) Majority

VC: Priority Outcast Clans VC: Priority Outcast Clans VC: Priority Outcast Clans

b. Hogaan 75 (51,● 90) Minority b. Hogaan −2 (−6, 2) ● difference b. Hogaan 76 (53,● 91) Majority 6 (−4, 15) ● a. Non−implementation 68 (46,● 86) Minority a. Non−implementation −7 (−16, 0) ● difference a. Non−implementation 76 (54,● 91) Majority

DC: Priority Women DC: Priority Women DC: Priority Women

b. Hogaan 85 (68,● 94) Minority b. Hogaan −1 (−5, 3) ● difference b. Hogaan 86 (70,● 95) Majority 5 (−2, 15) ● a. Non−implementation 82 (63,● 93) Minority a. Non−implementation −6 (−14, 0)* ● difference a. Non−implementation 89 (72,● 96) Majority

DC: Priority Youth DC: Priority Youth DC: Priority Youth

b. Hogaan 82 (64,● 93) Minority b. Hogaan −1 (−5, 4) ● difference b. Hogaan 83 (66,● 93) Majority 4 (−4, 13) ● a. Non−implementation 80 (61,● 92) Minority a. Non−implementation −5 (−13, 2) ● difference a. Non−implementation 85 (68,● 94) Majority

DC: Priority Outcast Clans DC: Priority Outcast Clans DC: Priority Outcast Clans

b. Hogaan 72 (50,● 88) Minority b. Hogaan −3 (−7, 1) ● difference b. Hogaan 75 (53,● 89) Majority 1 (−9, 11) ● a. Non−implementation 72 (51,● 88) Minority a. Non−implementation −3 (−12, 5) ● difference a. Non−implementation 76 (54,● 90) Majority

0% 25% 50% 75% 100% −10% 0% 10% −10% 0% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 81: Perceptions of Institutional Inclusion – Further Models

126 B.4 Further Model Results (Age Interaction – Ref: 35+)

127 Currently Attends Community Meetings Currently Attends Community Meetings Currently Attends Community Meetings

b. Hogaan 58 (45,● 68) Youth b. Hogaan −5 (−10, 0)* ● difference b. Hogaan 63 (50,● 73) 35+ −6 (−17, 5) ● a. Non−imp 55 (42,● 68) Youth a. Non−imp 0 (−9, 11) ● difference a. Non−imp 55 (41,● 67) 35+ Should Attend Community Meetings Should Attend Community Meetings Should Attend Community Meetings

b. Hogaan 85 (63,● 95) Youth b. Hogaan 1 (−3, 4) ● difference b. Hogaan 84 (63,● 94) 35+ 0 (−9, 9) ● a. Non−imp 82 (60,● 93) Youth a. Non−imp 1 (−7, 9) ● difference a. Non−imp 81 (57,● 93) 35+ significance Attend VC Meeting This Year Attend VC Meeting This Year Attend VC Meeting This Year a● none

b. Hogaan 82 (70,● 88) Youth b. Hogaan −4 (−10, 1) ● difference b. Hogaan 86 (76,● 91) 35+ 0 (−13, 12) ● a. Non−imp 80 (62,● 90) Youth a. Non−imp −3 (−15, 7) ● difference a. Non−imp 84 (70,● 91) 35+ Respondent Involved in Planning Process Respondent Involved in Planning Process Respondent Involved in Planning Process

b. Hogaan 63 (47,● 76) Youth b. Hogaan −3 (−8, 2) ● difference b. Hogaan 66 (50,● 78) 35+ −2 (−13, 8) ● a. Non−imp 57 (39,● 70) Youth a. Non−imp −1 (−10, 8) ● difference a. Non−imp 57 (39,● 71) 35+

0% 25% 50% 75% 100% −10% 0% 10% −10% 0% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 82: Comparison of Youth Groups on Participation Measures (1/3)

Trust VC Trust VC Trust VC

b. Hogaan 91 (83,● 95) Youth b. Hogaan 5 (2, 9)** ● difference b. Hogaan 86 (76,● 91) 35+ 2 (−6, 11) ● a. Non−imp 81 (68,● 89) Youth a. Non−imp 3 (−6, 11) ● difference a. Non−imp 78 (65,● 87) 35+ Trust DC Trust DC Trust DC

b. Hogaan 91 (80,● 96) Youth b. Hogaan 6 (2, 10)*** ● difference b. Hogaan 85 (71,● 92) 35+ 6 (−3, 15) ● a. Non−imp 80 (63,● 89) Youth a. Non−imp 0 (−8, 8) ● difference a. Non−imp 79 (64,● 89) 35+ significance Trust Elders Trust Elders Trust Elders a● none

b. Hogaan 94 (91,● 96) Youth b. Hogaan 0 (−2, 3) ● difference b. Hogaan 94 (91,● 95) 35+ 7 (−1, 15) ● a. Non−imp 89 (77,● 95) Youth a. Non−imp −6 (−15, 1) ● difference a. Non−imp 95 (87,● 98) 35+ Trust Religious Leaders Trust Religious Leaders Trust Religious Leaders

b. Hogaan 96 (94,● 98) Youth b. Hogaan 0 (−2, 2) ● difference b. Hogaan 96 (94,● 98) 35+ 5 (−1, 13) ● a. Non−imp 91 (81,● 96) Youth a. Non−imp −5 (−13, 0) ● difference a. Non−imp 97 (90,● 99) 35+

0% 25% 50% 75% 100% −15% −10% −5% 0% 5% 10% −5% 0% 5% 10% 15% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 83: Comparison of Youth Groups on Trust Measures

128 Women Involved in Planning Process Women Involved in Planning Process Women Involved in Planning Process

b. Hogaan 85 (72,● 92) Youth b. Hogaan 3 (−1, 7) ● difference b. Hogaan 82 (69,● 90) 35+ 9 (−1, 19) ●

a. Non−implementation 68 (53,● 81) Youth a. Non−implementation −6 (−15, 2) ● difference a. Non−implementation 74 (60,● 85) 35+ Youth Involved in Planning Process Youth Involved in Planning Process Youth Involved in Planning Process

b. Hogaan 84 (71,● 90) Youth b. Hogaan 3 (−1, 8) ● difference b. Hogaan 81 (67,● 88) 35+ 8 (−2, 19) ●

a. Non−implementation 69 (54,● 81) Youth a. Non−implementation −4 (−14, 4) ● difference a. Non−implementation 74 (59,● 84) 35+ significance Minority Clans Involved in Planning Process Minority Clans Involved in Planning Process Minority Clans Involved in Planning Process a● none

b. Hogaan 86 (74,● 92) Youth b. Hogaan 4 (0, 8)* ● difference b. Hogaan 82 (69,● 89) 35+ 2 (−7, 12) ●

a. Non−implementation 76 (61,● 86) Youth a. Non−implementation 2 (−7, 11) ● difference a. Non−implementation 74 (59,● 84) 35+ Outcast Clans Involved in Planning Process Outcast Clans Involved in Planning Process Outcast Clans Involved in Planning Process

b. Hogaan 81 (66,● 88) Youth b. Hogaan 2 (−2, 6) ● difference b. Hogaan 79 (64,● 87) 35+ 1 (−9, 11) ●

a. Non−implementation 75 (58,● 85) Youth a. Non−implementation 1 (−9, 10) ● difference a. Non−implementation 74 (59,● 84) 35+

0% 25% 50% 75% −15% −10% −5% 0% 5% 10% −10% 0% 10% 20% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 84: Comparison of Youth Groups on Community Perceptions of Inclusion – Further Models

129 Contribution: International Development Project Contribution: International Development Project Contribution: International Development Project

b. Hogaan 63 (48,● 74) Youth b. Hogaan −4 (−9, 0) ● difference b. Hogaan 68 (52,● 78) 35+ 1 (−11, 12) ● a. Non−implementation 34 (18,● 50) Youth a. Non−implementation −5 (−16, 5) ● difference a. Non−implementation 40 (24,● 54) 35+ Contribution: Social Services Contribution: Social Services Contribution: Social Services

b. Hogaan 63 (45,● 76) Youth b. Hogaan −7 (−12, −3)*** ● difference b. Hogaan 70 (54,● 82) 35+ 4 (−6, 16) ● a. Non−implementation 48 (30,● 63) Youth a. Non−implementation −12 (−23, −3)** ● difference a. Non−implementation 60 (43,● 74) 35+ Contribution: Poor and Needy Contribution: Poor and Needy Contribution: Poor and Needy

b. Hogaan 83 (66,● 90) Youth b. Hogaan −1 (−5, 2) ● difference b. Hogaan 85 (68,● 91) 35+ 5 (−4, 15) significance ● a● none a. Non−implementation 81 (62,● 90) Youth a. Non−implementation −6 (−16, 2) ● difference a. Non−implementation 88 (72,● 94) 35+ Contribution: Security Contribution: Security Contribution: Security

b. Hogaan 83 (64,● 91) Youth b. Hogaan −1 (−5, 2) ● difference b. Hogaan 85 (67,● 92) 35+ −1 (−10, 9) ● a. Non−implementation 75 (54,● 87) Youth a. Non−implementation 0 (−9, 8) ● difference a. Non−implementation 76 (57,● 87) 35+ Contribution: Blood Contribution: Blood Contribution: Blood

b. Hogaan 82 (68,● 88) Youth b. Hogaan 0 (−4, 4) ● difference b. Hogaan 82 (70,● 88) 35+ 4 (−4, 13) ● a. Non−implementation 85 (72,● 91) Youth a. Non−implementation −4 (−13, 3) ● difference a. Non−implementation 89 (78,● 94) 35+

0% 25% 50% 75% −20% −10% 0% −10% 0% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 85: Percentage Having Made Contributions of Different Kinds – Further Models

130 Priority: Life Priority: Life Priority: Life

b. Hogaan 71 (56,● 83) Youth b. Hogaan 1 (−4, 5) ● difference b. Hogaan 71 (56,● 82) 35+ 11 (1, 21)** ● a. Non−implementation 74 (58,● 86) Youth a. Non−implementation −10 (−19, −2)** ● difference a. Non−implementation 85 (71,● 92) 35+

Priority: Justice Priority: Justice Priority: Justice

b. Hogaan 52 (47,● 55) Youth b. Hogaan −2 (−7, 3) ● difference b. Hogaan 54 (49,● 57) 35+ 9 (−3, 21) ● a. Non−implementation 50 (39,● 59) Youth a. Non−implementation −11 (−22, 0) ● difference a. Non−implementation 62 (51,● 70) 35+

Priority: Equality Priority: Equality Priority: Equality

b. Hogaan 15 (12,● 19) Youth b. Hogaan 0 (−4, 4) ● difference b. Hogaan 16 (12,● 19) 35+ −1 (−9, 7) ● a. Non−implementation9 (5,● 15) Youth a. Non−implementation 1 (−5, 8) ● difference a. Non−implementation8 (4,● 14) 35+

Priority: Education Priority: Education Priority: Education

b. Hogaan 62 (52,● 70) Youth b. Hogaan 7 (1, 12)** ● b. Hogaan difference 55 (44,● 64) significance 35+ −7 (−18, 5) ● none ● a ● positive a. Non−implementation 56 (41,● 67) a Youth a. Non−implementation 13 (3, 24)** ● difference a. Non−implementation 42 (29,● 54) 35+

Priority: Health Priority: Health Priority: Health

b. Hogaan 70 (64,● 75) Youth b. Hogaan −1 (−6, 3) ● difference b. Hogaan 71 (65,● 76) 35+ 3 (−6, 13) ● a. Non−implementation 78 (68,● 84) Youth a. Non−implementation −5 (−14, 4) ● difference a. Non−implementation 83 (75,● 88) 35+

Priority: Consulted on Decisions Priority: Consulted on Decisions Priority: Consulted on Decisions

b. Hogaan7 (3,● 11) Youth b. Hogaan −1 (−3, 2) ● difference b. Hogaan7 (4,● 12) 35+ 0 (−5, 6) ● a. Non−implementation3 (1,● 9) Youth a. Non−implementation −1 (−6, 4) ● difference a. Non−implementation4 (1,● 10) 35+

Priority: Free Speech Priority: Free Speech Priority: Free Speech

b. Hogaan5 (2,● 11) Youth b. Hogaan 0 (−3, 2) ● difference b. Hogaan5 (2,● 12) 35+ −1 (−7, 4) ● a. Non−implementation3 (0,● 9) Youth a. Non−implementation 1 (−4, 6) ● difference a. Non−implementation2 (0,● 7) 35+

0% 25% 50% 75% −20% −10% 0% 10% 20% −20% −10% 0% 10% 20% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 86: Comparison of Youth Groups on Top Three Rights Priorities – Further Models

131 Do Women Have a Right to Protection Under the Law Do Women Have a Right to Protection Under the Law Do Women Have a Right to Protection Under the Law

b. Hogaan 99 (97,● 100) Youth b. Hogaan 1 (0, 3) ● difference b. Hogaan 98 (95,● 99) 35+ 1 (−5, 8) ● a. Non−implementation 98 (87,● 100) Youth a. Non−implementation 0 (−7, 6) ● difference a. Non−implementation 99 (88,● 100) 35+ Should Women Have Same Rights as Men Should Women Have Same Rights as Men Should Women Have Same Rights as Men

b. Hogaan 70 (64,● 74) Youth b. Hogaan 0 (−5, 5) ● difference b. Hogaan 70 (65,● 74) 35+ −4 (−14, 7) significance ● a● none a. Non−implementation 79 (69,● 85) Youth a. Non−implementation 3 (−6, 12) ● difference a. Non−implementation 76 (66,● 82) 35+ Do Women Have the Same Rights as Men Do Women Have the Same Rights as Men Do Women Have the Same Rights as Men

b. Hogaan 49 (41,● 57) Youth b. Hogaan −2 (−7, 3) ● difference b. Hogaan 51 (43,● 58) 35+ −1 (−13, 11) ● a. Non−implementation 52 (41,● 63) Youth a. Non−implementation −1 (−12, 11) ● difference a. Non−implementation 53 (42,● 63) 35+

0% 25% 50% 75% 100% −10% −5% 0% 5% 10% −10% −5% 0% 5% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 87: Women’s Rights in Comparison to Men’s – Further Models

Do Youth Have a Right to Protection Under the Law Do Youth Have a Right to Protection Under the Law Do Youth Have a Right to Protection Under the Law

b. Hogaan 97 (93,● 99) Youth b. Hogaan 1 (−1, 3) ● difference b. Hogaan 96 (92,● 98) 35+ −1 (−9, 5) ● a. Non−implementation 97 (89,● 99) Youth a. Non−implementation 2 (−3, 9) ● difference a. Non−implementation 95 (85,● 98) 35+ Should Youth Have Same Rights as Elderly People Should Youth Have Same Rights as Elderly People Should Youth Have Same Rights as Elderly People

b. Hogaan 77 (73,● 80) Youth b. Hogaan 0 (−5, 4) ● difference b. Hogaan 77 (73,● 80) 35+ 6 (−3, 15) significance ● a● none a. Non−implementation 83 (74,● 89) Youth a. Non−implementation −6 (−14, 2) ● difference a. Non−implementation 90 (81,● 94) 35+ Do Youth Have the Same Rights as The Elderly Do Youth Have the Same Rights as The Elderly Do Youth Have the Same Rights as The Elderly

b. Hogaan 58 (52,● 63) Youth b. Hogaan 2 (−3, 7) ● difference b. Hogaan 56 (50,● 62) 35+ 0 (−12, 13) ● a. Non−implementation 57 (45,● 66) Youth a. Non−implementation 2 (−9, 13) ● difference a. Non−implementation 55 (44,● 63) 35+

0% 25% 50% 75% 100% −15% −10% −5% 0% 5% 10% −10% 0% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 88: Youth Rights in Comparison to Older People – Further Models

132 All Clans Equal Right Legal Protection All Clans Equal Right Legal Protection All Clans Equal Right Legal Protection

b. Hogaan 98 (96,● 99) Youth b. Hogaan −1 (−2, 1) ● difference b. Hogaan 99 (97,● 99) 35+ −7 (−16, 0)* ● a. Non−implementation 98 (89,● 100) Youth a. Non−implementation 6 (0, 15)* ● difference a. Non−implementation 92 (80,● 96) 35+ Should Minority Clans Have Equal Rights with Majority Clans Should Minority Clans Have Equal Rights with MajorityShould Clans Minority Clans Have Equal Rights with Majority Clans

b. Hogaan 93 (89,● 96) Youth b. Hogaan −1 (−3, 2) ● difference b. Hogaan 94 (90,● 96) 35+ −7 (−16, 0)* ● a. Non−implementation 95 (87,● 98) Youth a. Non−implementation 6 (0, 14) ● difference a. Non−implementation 89 (77,● 94) 35+ Should Outcast Clans Equal Right with Majority Clans Should Outcast Clans Equal Right with Majority ClansShould Outcast Clans Equal Right with Majority Clans

b. Hogaan 80 (54,● 93) Youth b. Hogaan 0 (−4, 4) ● difference b. Hogaan 80 (56,● 93) significance 35+ −5 (−15, 4) ● negative ● a ● none a. Non−implementation 85 (62,● 96) a Youth a. Non−implementation 5 (−3, 14) ● difference a. Non−implementation 81 (55,● 93) 35+ Do Minority Clans Have Equal Rights with Majority Clans Do Minority Clans Have Equal Rights with Majority ClansDo Minority Clans Have Equal Rights with Majority Clans

b. Hogaan 70 (66,● 73) Youth b. Hogaan −4 (−9, 0) ● difference b. Hogaan 74 (70,● 77) 35+ −3 (−14, 8) ● a. Non−implementation 73 (61,● 81) Youth a. Non−implementation −1 (−11, 9) ● difference a. Non−implementation 74 (63,● 82) 35+ Do Outcast Clans Have Equal Rights with Majority Clans Do Outcast Clans Have Equal Rights with Majority ClansDo Outcast Clans Have Equal Rights with Majority Clans

b. Hogaan 61 (56,● 65) Youth b. Hogaan −4 (−10, 0) ● difference b. Hogaan 66 (61,● 69) 35+ −11 (−23, 1) ● a. Non−implementation 66 (55,● 75) Youth a. Non−implementation 6 (−5, 17) ● difference a. Non−implementation 60 (48,● 69) 35+

0% 25% 50% 75% 100% −10% 0% 10% −20% −10% 0% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 89: Minority and Outcast Clan in Comparison to Majority Clan – Further Models

133 Conflict Mechansims Effective Overall Conflict Mechansims Effective Overall Conflict Mechansims Effective Overall

b. Hogaan 82 (69,● 89) Youth b. Hogaan 1 (−3, 5) ● difference b. Hogaan 81 (67,● 88) 35+ −3 (−12, 5) ● a. Non−implementation 85 (71,● 93) Youth a. Non−implementation 4 (−4, 12) ● difference a. Non−implementation 81 (67,● 89) 35+ Conflicts Resolved Peacefully Conflicts Resolved Peacefully Conflicts Resolved Peacefully

b. Hogaan 94 (91,● 96) Youth b. Hogaan −1 (−3, 2) ● difference b. Hogaan 95 (92,● 97) 35+ −5 (−14, 3) ● a. Non−implementation 92 (82,● 96) Youth a. Non−implementation 4 (−3, 12) ● difference a. Non−implementation 87 (78,● 93) 35+ Women's Issues in Conflict Resolution Women's Issues in Conflict Resolution Women's Issues in Conflict Resolution

b. Hogaan 79 (72,● 84) Youth b. Hogaan 1 (−3, 5) ● difference b. Hogaan 78 (72,● 83) 35+ 5 (−6, 17) significance ● a● none a. Non−implementation 68 (56,● 77) Youth a. Non−implementation −4 (−15, 5) ● difference a. Non−implementation 72 (61,● 79) 35+ Youth Issues in Conflict Resolution Youth Issues in Conflict Resolution Youth Issues in Conflict Resolution

b. Hogaan 77 (68,● 83) Youth b. Hogaan 0 (−5, 4) ● difference b. Hogaan 77 (68,● 83) 35+ 0 (−10, 11) ● a. Non−implementation 67 (54,● 77) Youth a. Non−implementation 0 (−10, 9) ● difference a. Non−implementation 68 (55,● 76) 35+ Minority Clan Issues in Conflict Resolution Minority Clan Issues in Conflict Resolution Minority Clan Issues in Conflict Resolution

b. Hogaan 77 (67,● 84) Youth b. Hogaan 1 (−3, 6) ● difference b. Hogaan 76 (65,● 82) 35+ 7 (−4, 18) ● a. Non−implementation 69 (56,● 79) Youth a. Non−implementation −6 (−16, 5) ● difference a. Non−implementation 75 (62,● 83) 35+

0% 25% 50% 75% 100% −10% 0% 10% −10% 0% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 90: Comparison of Youth Groups on Conflict Mechanisms – Further Models

134 Resolution: Clan Issues Resolution: Clan Issues Resolution: Clan Issues

b. Hogaan 89 (82,● 93) Youth b. Hogaan −2 (−6, 1) ● difference b. Hogaan 91 (84,● 94) 35+ −2 (−10, 7) ● a. Non−implementation 89 (78,● 94) Youth a. Non−implementation 0 (−8, 7) ● difference a. Non−implementation 89 (79,● 94) 35+ Resolution: Water Issues Resolution: Water Issues Resolution: Water Issues

b. Hogaan 91 (86,● 94) Youth b. Hogaan 1 (−3, 4) ● difference b. Hogaan 90 (84,● 93) 35+ 3 (−5, 12) ● a. Non−implementation 87 (74,● 93) Youth a. Non−implementation −3 (−11, 4) ● difference a. Non−implementation 89 (79,● 95) 35+ Resolution: Grazing Issues Resolution: Grazing Issues Resolution: Grazing Issues

b. Hogaan 89 (81,● 92) Youth b. Hogaan 2 (−1, 5) ● difference b. Hogaan 87 (79,● 91) 35+ 0 (−8, 8) significance ● a● none a. Non−implementation 91 (80,● 95) Youth a. Non−implementation 2 (−6, 9) ● difference a. Non−implementation 89 (79,● 94) 35+ Resolution: Land Issues Resolution: Land Issues Resolution: Land Issues

b. Hogaan 88 (79,● 92) Youth b. Hogaan 2 (−2, 6) ● difference b. Hogaan 86 (76,● 91) 35+ −2 (−10, 6) ● a. Non−implementation 92 (80,● 96) Youth a. Non−implementation 4 (−3, 11) ● difference a. Non−implementation 87 (76,● 93) 35+ Resolution: Charcoal Issues Resolution: Charcoal Issues Resolution: Charcoal Issues

b. Hogaan 84 (75,● 88) Youth b. Hogaan 1 (−3, 6) ● difference b. Hogaan 83 (74,● 88) 35+ −1 (−11, 9) ● a. Non−implementation 86 (75,● 93) Youth a. Non−implementation 2 (−7, 11) ● difference a. Non−implementation 84 (72,● 90) 35+

0% 25% 50% 75% 100% −10% −5% 0% 5% 10% −10% −5% 0% 5% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 91: Comparison of Youth Groups on Issue Resolution – Further Models

135 Elders: Priority Women Elders: Priority Women Elders: Priority Women

b. Hogaan 87 (67,● 95) Youth b. Hogaan 4 (0, 9)* ● difference b. Hogaan 82 (60,● 93) 35+ −3 (−13, 4) ● a. Non−implementation 93 (76,● 98) Youth a. Non−implementation 8 (1, 16)** ● difference a. Non−implementation 85 (65,● 95) 35+

Elders: Priority Youth Elders: Priority Youth Elders: Priority Youth

b. Hogaan 86 (67,● 95) Youth b. Hogaan 5 (2, 10)** ● difference b. Hogaan 80 (62,● 92) 35+ −5 (−15, 3) ● a. Non−implementation 92 (73,● 98) Youth a. Non−implementation 10 (3, 20)*** ● difference a. Non−implementation 81 (57,● 93) 35+

Elders: Priority Outcast Clans Elders: Priority Outcast Clans Elders: Priority Outcast Clans

b. Hogaan 74 (54,● 89) Youth b. Hogaan 0 (−3, 4) ● difference b. Hogaan 74 (53,● 89) 35+ −11 (−21, −2)** ● a. Non−implementation 79 (54,● 93) Youth a. Non−implementation 11 (3, 21)** ● difference a. Non−implementation 67 (45,● 84) 35+

Relgious Leaders: Priority Women Relgious Leaders: Priority Women Relgious Leaders: Priority Women

b. Hogaan 95 (85,● 98) Youth b. Hogaan 2 (−1, 5) ● difference b. Hogaan 94 (82,● 98) 35+ 5 (−4, 102) ● a. Non−implementation 100 (0,● 100) Youth a. Non−implementation 0 (−99, 5) ● difference a. Non−implementation 98 (91,● 100) 35+

Religious Leaders: Priority Youth Religious Leaders: Priority Youth Religious Leaders: Priority Youth

b. Hogaan 93 (83,● 97) Youth b. Hogaan 2 (−1, 5) ● difference b. Hogaan 91 (79,● 96) 35+ −2 (−9, 4) ● a. Non−implementation 97 (90,● 99) Youth a. Non−implementation 4 (0, 11) ● difference a. Non−implementation 93 (81,● 97) 35+

Religious Leaders: Priority Outcast Clans Religious Leaders: Priority Outcast Clans Religious Leaders: Priority Outcast Clans

b. Hogaan 81 (58,● 94) Youth b. Hogaan −1 (−5, 2) ● difference b. Hogaan 82 (60,● 95) 35+ −3 (−13, 5) ● a. Non−implementation 91 (68,● 98) Youth a. Non−implementation 1 (−6, 10) ● difference a. Non−implementation 89 (66,● 98)

35+ significance

● negative VC: Priority Women VC: Priority Women VC: Priority Women a a● none b. Hogaan 88 (70,● 96) Youth b. Hogaan 4 (1, 9)** ● difference b. Hogaan 83 (66,● 94) 35+ −2 (−11, 6) ● a. Non−implementation 93 (76,● 98) Youth a. Non−implementation 6 (1, 15)* ● difference a. Non−implementation 86 (68,● 96) 35+

VC: Priority Youth VC: Priority Youth VC: Priority Youth

b. Hogaan 86 (70,● 95) Youth b. Hogaan 4 (1, 9)** ● difference b. Hogaan 81 (63,● 92) 35+ −9 (−19, 0) ● a. Non−implementation 93 (76,● 98) Youth a. Non−implementation 13 (6, 23)*** ● difference a. Non−implementation 78 (61,● 91) 35+

VC: Priority Outcast Clans VC: Priority Outcast Clans VC: Priority Outcast Clans

b. Hogaan 74 (50,● 90) Youth b. Hogaan −1 (−5, 2) ● difference b. Hogaan 75 (51,● 91) 35+ −10 (−18, −1)* ● a. Non−implementation 77 (53,● 92) Youth a. Non−implementation 8 (1, 16)* ● difference a. Non−implementation 69 (47,● 87) 35+

DC: Priority Women DC: Priority Women DC: Priority Women

b. Hogaan 87 (69,● 96) Youth b. Hogaan 4 (0, 8)* ● difference b. Hogaan 83 (66,● 94) 35+ −1 (−9, 6) ● a. Non−implementation 90 (73,● 97) Youth a. Non−implementation 5 (−1, 13) ● difference a. Non−implementation 84 (67,● 94) 35+

DC: Priority Youth DC: Priority Youth DC: Priority Youth

b. Hogaan 85 (69,● 94) Youth b. Hogaan 5 (1, 10)** ● difference b. Hogaan 80 (63,● 91) 35+ −1 (−10, 7) ● a. Non−implementation 88 (69,● 96) Youth a. Non−implementation 6 (0, 14) ● difference a. Non−implementation 80 (61,● 92) 35+

DC: Priority Outcast Clans DC: Priority Outcast Clans DC: Priority Outcast Clans

b. Hogaan 74 (52,● 89) Youth b. Hogaan 0 (−4, 4) ● difference b. Hogaan 74 (52,● 89) 35+ −6 (−15, 3) ● a. Non−implementation 78 (54,● 92) Youth a. Non−implementation 6 (−2, 14) ● difference a. Non−implementation 72 (50,● 87) 35+

0% 25% 50% 75% 100% −100% −75% −50% −25% 0% 25% 0% 40% 80% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 92: Perceptions of Institutional Inclusion – Further Models

136 B.5 Further Model Results (Age Interaction – Ref: <25)

137 Currently Attends Community Meetings Currently Attends Community Meetings Currently Attends Community Meetings

b. Hogaan 54 (42,● 66) Youth b. Hogaan −8 (−14, −2)** ● difference b. Hogaan 62 (50,● 72) 25+ −2 (−17, 13) ● a. Non−implementation 50 (32,● 65) Youth a. Non−implementation −6 (−19, 8) ● difference a. Non−implementation 56 (44,● 67) 25+

Should Attend Community Meetings Should Attend Community Meetings Should Attend Community Meetings

b. Hogaan 84 (63,● 94) Youth b. Hogaan 0 (−5, 4) ● difference b. Hogaan 84 (64,● 94) 25+ 5 (−6, 18) ● a. Non−implementation 76 (51,● 92) Youth a. Non−implementation −5 (−18, 5) ● difference a. Non−implementation 82 (59,● 93) 25+

Attend VC Meeting This Year Attend VC Meeting This Year Attend VC Meeting This Year

b. Hogaan 84 (73,● 90) Youth b. Hogaan 0 (−7, 6) ● difference b. Hogaan 84 (74,● 89) 25+ 3 (−13, 24) ● a. Non−implementation 80 (53,● 92) Youth a. Non−implementation −2 (−24, 11) ● difference a. Non−implementation 83 (70,● 90) 25+

VC Meetings Consider my Household Viewpoint VC Meetings Consider my Household Viewpoint VC Meetings Consider my Household Viewpoint

b. Hogaan 97 (89,● 100) Youth b. Hogaan 3 (−2, 7) ● difference b. Hogaan 93 (89,● 97) 25+ 17 (3, 45)** ● a. Non−implementation 83 (47,● 95) Youth a. Non−implementation −13 (−41, −2)* ● difference a. Non−implementation 97 (85,● 100) 25+

VC Considers my Viewpoint VC Considers my Viewpoint VC Considers my Viewpoint

b. Hogaan 86 (71,● 93) Youth b. Hogaan 3 (−2, 8) ● difference b. Hogaan 83 (68,● 91) 25+ 9 (−2, 22) ● a. Non−implementation 73 (53,● 86) Youth a. Non−implementation −6 (−19, 4) ● difference a. Non−implementation 80 (63,● 88)

25+ significance

● none Interacted with VC Member Interacted with VC Member Interacted with VC Member a a● positive b. Hogaan 49 (36,● 60) Youth b. Hogaan −14 (−20, −8)*** ● difference b. Hogaan 63 (51,● 73) 25+ −9 (−23, 6) ● a. Non−implementation 39 (21,● 53) Youth a. Non−implementation −5 (−19, 8) ● difference a. Non−implementation 44 (31,● 55) 25+

VC Members Accessible VC Members Accessible VC Members Accessible

b. Hogaan 93 (77,● 97) Youth b. Hogaan 1 (−2, 5) ● difference b. Hogaan 91 (76,● 96) 25+ 0 (−8, 12) ● a. Non−implementation 86 (66,● 96) Youth a. Non−implementation 1 (−10, 9) ● difference a. Non−implementation 86 (69,● 94) 25+

DC Considers my Viewpoint DC Considers my Viewpoint DC Considers my Viewpoint

b. Hogaan 87 (70,● 94) Youth b. Hogaan 5 (0, 10)* ● difference b. Hogaan 82 (63,● 90) 25+ 15 (3, 29)** ● a. Non−implementation 67 (46,● 83) Youth a. Non−implementation −10 (−24, 0) ● difference a. Non−implementation 79 (63,● 88) 25+

Interacted with DC Member Interacted with DC Member Interacted with DC Member

b. Hogaan 34 (21,● 47) Youth b. Hogaan −14 (−20, −9)*** ● difference b. Hogaan 49 (35,● 61) 25+ −2 (−16, 11) ● a. Non−implementation 22 (9,● 38) Youth a. Non−implementation −13 (−24, 1) ● difference a. Non−implementation 35 (22,● 47) 25+

DC Members Accessible DC Members Accessible DC Members Accessible

b. Hogaan 85 (67,● 93) Youth b. Hogaan 1 (−4, 5) ● difference b. Hogaan 84 (67,● 92) 25+ −1 (−12, 10) ● a. Non−implementation 83 (63,● 94) Youth a. Non−implementation 2 (−8, 11) ● difference a. Non−implementation 81 (64,● 91) 25+

0% 25% 50% 75% 100% −40% −30% −20% −10% 0% 10% −25% 0% 25% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 93: Participation: Comparison Youth – Further Models 138 Does VC Provide Services Does VC Provide Services Does VC Provide Services

b. Hogaan 72 (55,● 82) Youth b. Hogaan 1 (−4, 6) ● difference b. Hogaan 71 (55,● 81) 25+ 6 (−7, 20) ● a. Non−implementation 70 (48,● 83) Youth a. Non−implementation −5 (−18, 7) ● difference a. Non−implementation 75 (59,● 84) 25+ Can VC Manage Projects Can VC Manage Projects Can VC Manage Projects

b. Hogaan 82 (70,● 89) Youth b. Hogaan 3 (−2, 7) ● difference b. Hogaan 79 (67,● 86) 25+ 12 (−2, 28) ● a. Non−implementation 63 (43,● 78) Youth a. Non−implementation −9 (−23, 3) ● difference a. Non−implementation 73 (59,● 82) 25+ Does VC Manage Money Transparently and Accountably Does VC Manage Money Transparently and AccountablyDoes VC Manage Money Transparently and Accountably

b. Hogaan 79 (66,● 86) Youth b. Hogaan 5 (−1, 9) ● difference b. Hogaan 75 (62,● 82) 25+ 20 (6, 34)** ● a. Non−implementation 58 (39,● 73) Youth a. Non−implementation −15 (−28, −2)* ● difference a. Non−implementation 73 (60,● 82)

25+ significance

● none Does DC Provide Services Does DC Provide Services Does DC Provide Services a a● positive b. Hogaan 78 (65,● 86) Youth b. Hogaan 3 (−2, 8) ● difference b. Hogaan 75 (63,● 83) 25+ 14 (1, 29)* ● a. Non−implementation 62 (43,● 76) Youth a. Non−implementation −12 (−26, 1) ● difference a. Non−implementation 74 (60,● 82) 25+ Can DC Manage Projects Can DC Manage Projects Can DC Manage Projects

b. Hogaan 80 (70,● 86) Youth b. Hogaan 5 (0, 11)* ● difference b. Hogaan 74 (64,● 81) 25+ 19 (4, 35)** ● a. Non−implementation 61 (42,● 75) Youth a. Non−implementation −13 (−28, 0) ● difference a. Non−implementation 75 (62,● 82) 25+ Does DC Manage Money Transparently and Accountably Does DC Manage Money Transparently and AccountablyDoes DC Manage Money Transparently and Accountably

b. Hogaan 78 (69,● 85) Youth b. Hogaan 7 (1, 12)* ● difference b. Hogaan 72 (63,● 78) 25+ 15 (2, 30)* ● a. Non−implementation 68 (49,● 79) Youth a. Non−implementation −9 (−23, 3) ● difference a. Non−implementation 76 (66,● 83) 25+

0% 25% 50% 75% −30% −20% −10% 0% 10% 0% 10% 20% 30% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 94: Comparison of Youth Groups on Service Delivery Measures – Further Models

139 Women Involved in Planning Process Women Involved in Planning Process Women Involved in Planning Process

b. Hogaan 83 (70,● 90) Youth b. Hogaan 0 (−5, 5) ● difference b. Hogaan 83 (71,● 90) 25+ 16 (3, 29)** ●

a. Non−implementation 58 (40,● 73) Youth a. Non−implementation −16 (−29, −4)** ● difference a. Non−implementation 75 (60,● 85) 25+ Youth Involved in Planning Process Youth Involved in Planning Process Youth Involved in Planning Process

b. Hogaan 82 (67,● 89) Youth b. Hogaan −1 (−6, 4) ● difference b. Hogaan 83 (69,● 89) 25+ 5 (−8, 18) ●

a. Non−implementation 67 (47,● 81) Youth a. Non−implementation −6 (−19, 7) ● difference a. Non−implementation 73 (58,● 83)

25+ significance

● none Minority Clans Involved in Planning Process Minority Clans Involved in Planning Process Minority Clans Involved in Planning Process a a● positive b. Hogaan 84 (71,● 91) Youth b. Hogaan 0 (−5, 5) ● difference b. Hogaan 84 (71,● 90) 25+ 12 (−1, 26) ●

a. Non−implementation 65 (47,● 79) Youth a. Non−implementation −11 (−24, 1) ● difference a. Non−implementation 77 (63,● 85) 25+ Outcast Clans Involved in Planning Process Outcast Clans Involved in Planning Process Outcast Clans Involved in Planning Process

b. Hogaan 79 (64,● 88) Youth b. Hogaan 0 (−6, 5) ● difference b. Hogaan 80 (65,● 88) 25+ 17 (4, 33)** ●

a. Non−implementation 58 (37,● 74) Youth a. Non−implementation −18 (−32, −5)** ● difference a. Non−implementation 77 (62,● 86) 25+

0% 25% 50% 75% −30% −20% −10% 0% −10% 0% 10% 20% 30% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 95: Comparison of Youth Groups on Community Perceptions of Inclusion – Further Models

Do Women Have a Right to Protection Under the Law Do Women Have a Right to Protection Under the Law Do Women Have a Right to Protection Under the Law

b. Hogaan 99 (95,● 100) Youth b. Hogaan 1 (−1, 3) ● difference b. Hogaan 98 (96,● 99) 25+ 92 (−6, 101) ●

a. Non−implementation0 (0, ●100) Youth a. Non−implementation−91 (−99, 6) ● difference a. Non−implementation 98 (91,● 100) 25+ Should Women Have Same Rights as Men Should Women Have Same Rights as Men Should Women Have Same Rights as Men

b. Hogaan 69 (62,● 74) Youth b. Hogaan −1 (−7, 4) ● difference b. Hogaan 70 (65,● 74) 25+ 2 (−11, 18) significance ● a● none a. Non−implementation 74 (57,● 85) Youth a. Non−implementation −3 (−18, 8) ● difference a. Non−implementation 77 (69,● 83) 25+ Do Women Have the Same Rights as Men Do Women Have the Same Rights as Men Do Women Have the Same Rights as Men

b. Hogaan 48 (39,● 56) Youth b. Hogaan −2 (−8, 4) ● difference b. Hogaan 51 (42,● 57) 25+ 6 (−9, 22) ●

a. Non−implementation 46 (30,● 61) Youth a. Non−implementation −8 (−22, 6) ● difference a. Non−implementation 54 (44,● 63) 25+

0% 25% 50% 75% 100% −100% −75% −50% −25% 0% 0% 25% 50% 75% 100% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 96: Women’s Rights in Comparison to Men’s – Further Models

140 Do Youth Have a Right to Protection Under the Law Do Youth Have a Right to Protection Under the Law Do Youth Have a Right to Protection Under the Law

b. Hogaan 99 (95,● 100) Youth b. Hogaan 2 (0, 5) ● difference b. Hogaan 97 (92,● 98) 25+ 2 (−4, 15) ● a. Non−implementation 96 (77,● 99) Youth a. Non−implementation 0 (−13, 6) ● difference a. Non−implementation 96 (89,● 99) 25+ Should Youth Have Same Rights as Elderly People Should Youth Have Same Rights as Elderly People Should Youth Have Same Rights as Elderly People

b. Hogaan 74 (68,● 79) Youth b. Hogaan −3 (−9, 2) ● difference b. Hogaan 78 (74,● 80) 25+ 6 (−5, 22) significance ● a● none a. Non−implementation 79 (61,● 89) Youth a. Non−implementation −9 (−25, 1) ● difference a. Non−implementation 88 (81,● 92) 25+ Do Youth Have the Same Rights as The Elderly Do Youth Have the Same Rights as The Elderly Do Youth Have the Same Rights as The Elderly

b. Hogaan 55 (47,● 62) Youth b. Hogaan −3 (−9, 3) ● difference b. Hogaan 58 (51,● 63) 25+ 3 (−12, 19) ● a. Non−implementation 50 (34,● 64) Youth a. Non−implementation −6 (−21, 9) ● difference a. Non−implementation 57 (47,● 63) 25+

All Clans Equal Right Legal Protection All Clans Equal Right Legal Protection All Clans Equal Right Legal Protection

b. Hogaan 98 (95,● 99) Youth b. Hogaan 0 (−3, 1) ● difference b. Hogaan 98 (97,● 99) 25+ −2 (−9, 12) ● a. Non−implementation 96 (75,● 99) Youth a. Non−implementation 2 (−12, 8) ● difference a. Non−implementation 94 (86,● 98) 25+ Should Minority Clans Have Equal Rights with Majority Clans Should Minority Clans Have Equal Rights with MajorityShould Clans Minority Clans Have Equal Rights with Majority Clans

b. Hogaan 91 (84,● 94) Youth b. Hogaan −4 (−8, 0)* ● difference b. Hogaan 95 (91,● 97) 25+ −5 (−14, 6) ● a. Non−implementation 93 (76,● 98) Youth a. Non−implementation 1 (−10, 8) ● difference a. Non−implementation 92 (83,● 96) 25+ Should Outcast Clans Equal Right with Majority Clans Should Outcast Clans Equal Right with Majority Clans Should Outcast Clans Equal Right with Majority Clans

b. Hogaan 74 (49,● 90) Youth b. Hogaan −6 (−11, −2)** ● difference b. Hogaan 80 (57,● 93) 25+ −3 (−14, 12) significance ● a● none a. Non−implementation 79 (52,● 94) Youth a. Non−implementation −3 (−17, 6) ● difference a. Non−implementation 84 (58,● 95) 25+ Do Minority Clans Have Equal Rights with Majority Clans Do Minority Clans Have Equal Rights with Majority ClansDo Minority Clans Have Equal Rights with Majority Clans

b. Hogaan 65 (59,● 70) Youth b. Hogaan −8 (−14, −3)** ● difference b. Hogaan 74 (70,● 77) 25+ −10 (−23, 7) ● a. Non−implementation 75 (56,● 86) Youth a. Non−implementation 2 (−13, 14) ● difference a. Non−implementation 73 (63,● 81) 25+ Do Outcast Clans Have Equal Rights with Majority Clans Do Outcast Clans Have Equal Rights with Majority ClansDo Outcast Clans Have Equal Rights with Majority Clans

b. Hogaan 57 (51,● 63) Youth b. Hogaan −8 (−14, −2)** ● difference b. Hogaan 65 (61,● 68) 25+ −15 (−29, 2) ● a. Non−implementation 68 (51,● 81) Youth a. Non−implementation 6 (−8, 21) ● difference a. Non−implementation 62 (52,● 70) 25+

0% 25% 50% 75% 100% −10% 0% 10% 20% −30% −20% −10% 0% 10% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 98: Minority and Outcast Clan in Comparison to Majority Clan – Further Models

141 Elders: Priority Women Elders: Priority Women Elders: Priority Women

b. Hogaan 88 (69,● 96) Youth b. Hogaan 4 (0, 10)* ● difference b. Hogaan 83 (64,● 94) 25+ 2 (−8, 13) ● a. Non−implementation 91 (69,● 98) Youth a. Non−implementation 2 (−7, 11) ● difference a. Non−implementation 88 (70,● 96) 25+

Elders: Priority Youth Elders: Priority Youth Elders: Priority Youth

b. Hogaan 87 (67,● 96) Youth b. Hogaan 5 (0, 10)* ● difference b. Hogaan 81 (62,● 93) 25+ −1 (−11, 9) ● a. Non−implementation 91 (67,● 98) Youth a. Non−implementation 6 (−3, 15) ● difference a. Non−implementation 84 (64,● 94) 25+

Elders: Priority Outcast Clans Elders: Priority Outcast Clans Elders: Priority Outcast Clans

b. Hogaan 74 (52,● 90) Youth b. Hogaan −1 (−5, 4) ● difference b. Hogaan 75 (53,● 90) 25+ −10 (−22, 3) ● a. Non−implementation 80 (55,● 95) Youth a. Non−implementation 9 (−3, 20) ● difference a. Non−implementation 71 (49,● 87) 25+

Relgious Leaders: Priority Women Relgious Leaders: Priority Women Relgious Leaders: Priority Women

b. Hogaan 94 (82,● 98) Youth b. Hogaan −1 (−5, 3) ● difference b. Hogaan 95 (84,● 98) 25+ 2 (−6, 101) ● a. Non−implementation 100 (0,● 100) Youth a. Non−implementation 0 (−100, 3) ● difference a. Non−implementation 99 (94,● 100) 25+

Religious Leaders: Priority Youth Religious Leaders: Priority Youth Religious Leaders: Priority Youth

b. Hogaan 92 (79,● 96) Youth b. Hogaan 0 (−4, 3) ● difference b. Hogaan 91 (81,● 96) 25+ 1 (−6, 13) ● a. Non−implementation 94 (78,● 99) Youth a. Non−implementation −1 (−12, 5) ● difference a. Non−implementation 95 (86,● 98) 25+

Religious Leaders: Priority Outcast Clans Religious Leaders: Priority Outcast Clans Religious Leaders: Priority Outcast Clans

b. Hogaan 79 (55,● 93) Youth b. Hogaan −3 (−8, 1) ● difference b. Hogaan 83 (59,● 95) 25+ −6 (−18, 3) ● a. Non−implementation 94 (70,● 99) Youth a. Non−implementation 3 (−6, 14) ● difference a. Non−implementation 90 (66,● 98) 25+ significance VC: Priority Women VC: Priority Women VC: Priority Women a● none

b. Hogaan 87 (69,● 97) Youth b. Hogaan 2 (−2, 7) ● difference b. Hogaan 85 (67,● 95) 25+ 0 (−9, 10) ● a. Non−implementation 91 (67,● 98) Youth a. Non−implementation 2 (−7, 11) ● difference a. Non−implementation 89 (70,● 97) 25+

VC: Priority Youth VC: Priority Youth VC: Priority Youth

b. Hogaan 85 (66,● 95) Youth b. Hogaan 2 (−2, 6) ● difference b. Hogaan 83 (65,● 93) 25+ −5 (−16, 6) ● a. Non−implementation 91 (70,● 98) Youth a. Non−implementation 6 (−3, 16) ● difference a. Non−implementation 83 (65,● 94) 25+

VC: Priority Outcast Clans VC: Priority Outcast Clans VC: Priority Outcast Clans

b. Hogaan 73 (50,● 90) Youth b. Hogaan −3 (−8, 1) ● difference b. Hogaan 76 (52,● 92) 25+ −7 (−19, 5) ● a. Non−implementation 76 (52,● 93) Youth a. Non−implementation 4 (−7, 14) ● difference a. Non−implementation 72 (49,● 89) 25+

DC: Priority Women DC: Priority Women DC: Priority Women

b. Hogaan 88 (70,● 96) Youth b. Hogaan 4 (0, 9)* ● difference b. Hogaan 84 (66,● 94) 25+ 5 (−4, 18) ● a. Non−implementation 85 (63,● 96) Youth a. Non−implementation −1 (−12, 7) ● difference a. Non−implementation 87 (68,● 95) 25+

DC: Priority Youth DC: Priority Youth DC: Priority Youth

b. Hogaan 85 (68,● 94) Youth b. Hogaan 3 (−2, 8) ● difference b. Hogaan 82 (64,● 92) 25+ 1 (−9, 12) ● a. Non−implementation 85 (64,● 95) Youth a. Non−implementation 2 (−9, 11) ● difference a. Non−implementation 83 (67,● 93) 25+

DC: Priority Outcast Clans DC: Priority Outcast Clans DC: Priority Outcast Clans

b. Hogaan 72 (51,● 87) Youth b. Hogaan −2 (−7, 3) ● difference b. Hogaan 74 (52,● 88) 25+ −3 (−15, 9) ● a. Non−implementation 76 (50,● 92) Youth a. Non−implementation 2 (−9, 13) ● difference a. Non−implementation 74 (53,● 90) 25+

0% 25% 50% 75% 100% −100% −75% −50% −25% 0% 25% −25% 0% 25% 50% 75% 100% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 99: Perceptions of Institutional Inclusion – Further Models

142 Trust VC Trust VC Trust VC

b. Hogaan 91 (82,● 95) Youth b. Hogaan 3 (−1, 7) ● difference b. Hogaan 88 (79,● 92) 25+ 9 (−1, 23) ● a. Non−implementation 74 (55,● 85) Youth a. Non−implementation −6 (−20, 4) ● difference a. Non−implementation 80 (69,● 88) 25+ Trust DC Trust DC Trust DC

b. Hogaan 91 (78,● 96) Youth b. Hogaan 3 (0, 8) ● difference b. Hogaan 87 (74,● 93) 25+ 6 (−5, 20) ● a. Non−implementation 78 (58,● 89) Youth a. Non−implementation −2 (−16, 8) ● difference a. Non−implementation 80 (66,● 89) 25+ significance Trust Elders Trust Elders Trust Elders a● none

b. Hogaan 95 (91,● 97) Youth b. Hogaan 1 (−2, 4) ● difference b. Hogaan 93 (91,● 95) 25+ 3 (−7, 18) ● a. Non−implementation 90 (70,● 97) Youth a. Non−implementation −2 (−17, 7) ● difference a. Non−implementation 92 (84,● 96) 25+ Trust Religious Leaders Trust Religious Leaders Trust Religious Leaders

b. Hogaan 97 (94,● 99) Youth b. Hogaan 1 (−2, 3) ● difference b. Hogaan 96 (94,● 97) 25+ 3 (−4, 18) ● a. Non−implementation 92 (74,● 98) Youth a. Non−implementation −2 (−16, 5) ● difference a. Non−implementation 95 (88,● 97) 25+

0% 25% 50% 75% 100% −20% −10% 0% 0% 10% 20% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 100: Comparison of Youth Groups on Trust Measures – Further Models

Contribution: International Development Project Contribution: International Development Project Contribution: International Development Project

b. Hogaan 62 (47,● 74) Youth b. Hogaan −4 (−10, 1) ● difference b. Hogaan 66 (52,● 77) 25+ −16 (−30, −1)* ● a. Non−implementation 48 (26,● 65) Youth a. Non−implementation 11 (−3, 25) ● difference a. Non−implementation 37 (20,● 49) 25+ Contribution: Social Services Contribution: Social Services Contribution: Social Services

b. Hogaan 58 (40,● 72) Youth b. Hogaan −11 (−17, −5)*** ● difference b. Hogaan 69 (51,● 81) 25+ −9 (−24, 5) ● a. Non−implementation 54 (31,● 71) Youth a. Non−implementation −1 (−15, 12) ● difference a. Non−implementation 55 (37,● 69) 25+ Contribution: Poor and Needy Contribution: Poor and Needy Contribution: Poor and Needy

b. Hogaan 79 (59,● 87) Youth b. Hogaan −6 (−12, −1)** ● difference significance b. Hogaan 85 (69,● 91) 25+ 4 (−9, 20) a● negative ● a● none a. Non−implementation 76 (52,● 88) ● positive Youth a. Non−implementation −10 (−25, 0) a ● difference a. Non−implementation 87 (71,● 92) 25+ Contribution: Security Contribution: Security Contribution: Security

b. Hogaan 81 (63,● 91) Youth b. Hogaan −2 (−8, 2) ● difference b. Hogaan 84 (67,● 92) 25+ 10 (−2, 24) ● a. Non−implementation 64 (43,● 82) Youth a. Non−implementation −13 (−26, −1)* ● difference a. Non−implementation 78 (61,● 89) 25+ Contribution: Blood Contribution: Blood Contribution: Blood

b. Hogaan 81 (67,● 88) Youth b. Hogaan −1 (−6, 4) ● difference b. Hogaan 82 (70,● 88) 25+ 17 (5, 35)** ● a. Non−implementation 72 (50,● 84) Youth a. Non−implementation −18 (−35, −6)*** ● difference a. Non−implementation 90 (81,● 94) 25+

0% 25% 50% 75% −20% 0% 20% −20% 0% 20% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 101: Percentage Having Made Contributions of Different Kinds – Further Models

143 Priority: Life Priority: Life Priority: Life

b. Hogaan 70 (55,● 81) Youth b. Hogaan −2 (−8, 3) ● difference b. Hogaan 72 (58,● 82) 25+ 8 (−5, 22) ●

a. Non−implementation 71 (49,● 84) Youth a. Non−implementation−10 (−25, 1) ● difference a. Non−implementation 81 (67,● 90) 25+

Priority: Justice Priority: Justice Priority: Justice

b. Hogaan 51 (45,● 57) Youth b. Hogaan −2 (−8, 4) ● difference b. Hogaan 53 (50,● 57) 25+ 13 (−5, 28) ●

a. Non−implementation 44 (29,● 59) Youth a. Non−implementation−15 (−29, 1) ● difference a. Non−implementation 59 (50,● 66) 25+

Priority: Equality Priority: Equality Priority: Equality

b. Hogaan 16 (11,● 21) Youth b. Hogaan 1 (−4, 5) ● difference b. Hogaan15 (12,● 18) 25+ 0 (−13, 9) ●

a. Non−implementation9 (3,● 21) Youth a. Non−implementation 1 (−6, 13) ● difference a. Non−implementation9 (5,● 13) 25+

Priority: Education Priority: Education Priority: Education

b. Hogaan 67 (55,● 75) Youth b. Hogaan 10 (4, 16)** ● b. Hogaan difference 57 (45,● 66) significance 25+ −5 (−19, 12) ● none ● a ● positive a. Non−implementation 60 (42,● 76) a Youth a. Non−implementation 14 (−2, 28) ● difference a. Non−implementation 47 (35,● 57) 25+

Priority: Health Priority: Health Priority: Health

b. Hogaan 71 (63,● 76) Youth b. Hogaan 0 (−6, 5) ● difference b. Hogaan 71 (65,● 75) 25+ 14 (0, 30)* ●

a. Non−implementation 68 (51,● 80) Youth a. Non−implementation−14 (−29, −1)* ● difference a. Non−implementation 83 (75,● 87) 25+

Priority: Consulted on Decisions Priority: Consulted on Decisions Priority: Consulted on Decisions

b. Hogaan5 (3,● 11) Youth b. Hogaan −2 (−5, 2) ● difference b. Hogaan7 (4,● 12) 25+ −1 (−11, 6) ●

a. Non−implementation3 (0,● 15) Youth a. Non−implementation −1 (−6, 10) ● difference a. Non−implementation4 (1,● 9) 25+

Priority: Free Speech Priority: Free Speech Priority: Free Speech

b. Hogaan4 (1,● 10) Youth b. Hogaan −1 (−4, 2) ● difference b. Hogaan5 (2,● 11) 25+ −3 (−102, 5) ●

a. Non−implementation0 (0, ●100) Youth a. Non−implementation −1 (−6, 99) ● difference a. Non−implementation2 (1,● 7) 25+

0% 25% 50% 75% 100% 0% 50% 100% −80% −40% 0% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 102: Comparison of Youth Groups on Top Three Rights Priorities – Further Models

144 Conflict Mechansims Effective Overall Conflict Mechansims Effective Overall Conflict Mechansims Effective Overall

b. Hogaan 82 (69,● 90) Youth b. Hogaan 1 (−4, 6) ● difference b. Hogaan 81 (70,● 89) 25+ 9 (−2, 24) ● a. Non−implementation 75 (55,● 88) Youth a. Non−implementation −8 (−22, 2) ● difference a. Non−implementation 83 (70,● 91) 25+ Conflicts Resolved Peacefully Conflicts Resolved Peacefully Conflicts Resolved Peacefully

b. Hogaan 94 (90,● 97) Youth b. Hogaan 0 (−3, 3) ● difference b. Hogaan 94 (92,● 96) 25+ −3 (−12, 9) ● a. Non−implementation 92 (77,● 98) Youth a. Non−implementation 3 (−9, 11) ● difference a. Non−implementation 89 (81,● 94) 25+ Women's Issues in Conflict Resolution Women's Issues in Conflict Resolution Women's Issues in Conflict Resolution

b. Hogaan 77 (68,● 83) Youth b. Hogaan −2 (−8, 3) ● difference b. Hogaan 79 (73,● 83) 25+ 13 (−3, 29) significance ● a● none a. Non−implementation 57 (39,● 71) Youth a. Non−implementation −16 (−30, −1)* ● difference a. Non−implementation 72 (62,● 79) 25+ Youth Issues in Conflict Resolution Youth Issues in Conflict Resolution Youth Issues in Conflict Resolution

b. Hogaan 73 (63,● 81) Youth b. Hogaan −4 (−10, 1) ● difference b. Hogaan 78 (69,● 84) 25+ 12 (−4, 28) ● a. Non−implementation 53 (36,● 69) Youth a. Non−implementation −16 (−31, −3)* ● difference a. Non−implementation 70 (60,● 79) 25+ Minority Clan Issues in Conflict Resolution Minority Clan Issues in Conflict Resolution Minority Clan Issues in Conflict Resolution

b. Hogaan 75 (64,● 82) Youth b. Hogaan −2 (−7, 4) ● difference b. Hogaan 76 (67,● 83) 25+ 8 (−6, 23) ● a. Non−implementation 64 (45,● 78) Youth a. Non−implementation −10 (−24, 3) ● difference a. Non−implementation 74 (62,● 82) 25+

0% 25% 50% 75% 100% −30% −20% −10% 0% 10% −10% 0% 10% 20% 30% percentage percentage difference percentage difference of differences (a) Percentage (%) (b) Difference in % (c) Diff. in Diff. in % Figure 103: Comparison of Youth Groups on Conflict Mechanisms – Further Models

145 Resolution: Clan Issues Resolution: Clan Issues Resolution: Clan Issues

b. Hogaan 89 (80,● 93) Youth b. Hogaan −2 (−6, 2) ● difference b. Hogaan 91 (83,● 94) 25+ 4 (−6, 20) ● a. Non−implementation 84 (63,● 93) Youth a. Non−implementation −6 (−21, 3) ● difference a. Non−implementation 90 (80,● 94) 25+ Resolution: Water Issues Resolution: Water Issues Resolution: Water Issues

b. Hogaan 90 (83,● 94) Youth b. Hogaan 0 (−5, 3) ● difference b. Hogaan 91 (85,● 94) 25+ 7 (−2, 20) ● a. Non−implementation 81 (64,● 90) Youth a. Non−implementation −8 (−20, 1) ● difference a. Non−implementation 89 (80,● 94) 25+ Resolution: Grazing Issues Resolution: Grazing Issues Resolution: Grazing Issues

b. Hogaan 89 (81,● 93) Youth b. Hogaan 1 (−3, 6) ● difference b. Hogaan 88 (80,● 92) 25+ 5 (−5, 19) ● a. Non−implementation 86 (70,● 95) Youth a. Non−implementation −3 (−16, 5) ● difference a. Non−implementation 90 (82,● 94) 25+ significance Resolution: Land Issues Resolution: Land Issues Resolution: Land Issues a● none

b. Hogaan 85 (75,● 91) Youth b. Hogaan −1 (−6, 3) ● difference b. Hogaan 87 (79,● 91) 25+ −1 (−10, 14) ● a. Non−implementation 88 (71,● 96) Youth a. Non−implementation −1 (−14, 7) ● difference a. Non−implementation 89 (80,● 94) 25+ Resolution: Charcoal Issues Resolution: Charcoal Issues Resolution: Charcoal Issues

b. Hogaan 85 (74,● 90) Youth b. Hogaan 2 (−3, 7) ● difference b. Hogaan 83 (72,● 87) 25+ 2 (−9, 16) ● a. Non−implementation 85 (66,● 94) Youth a. Non−implementation 0 (−13, 10) ● difference a. Non−implementation 85 (75,● 91) 25+ Resolution: Leadership Issues Resolution: Leadership Issues Resolution: Leadership Issues

b. Hogaan 88 (79,● 93) Youth b. Hogaan −2 (−6, 2) ● difference b. Hogaan 90 (83,● 93) 25+ −5 (−16, 9) ● a. Non−implementation 81 (63,● 91) Youth a. Non−implementation 3 (−10, 13) ● difference a. Non−implementation 78 (67,● 86) 25+

146 B.6 Institutional Support for Services and Rights

147 Elders: Roads Elders: Roads

Hogaan 11 (7,● 16)

0 (−5,● 5)

Non−imp 11 (6,● 17)

VC: Roads VC: Roads

Hogaan 27 (18,● 37)

8 (2, ●15)**

Non−imp 19 (11,● 27) significance a● negative DC: Roads DC: Roads a● none a● positive Hogaan 47 (40,● 53)

−17 (−24,● −10)***

Non−imp 64 (55,● 71)

RL: Roads RL: Roads

Hogaan 16 (1,● 31)

0 (−3,● 2)

Non−imp 16 (0,● 32)

0% 20% 40% 60% −20% −10% 0% 10% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 105: Comparison of Hogaan and Non-Implementation Villages on Institutions Responsible for Roads

Elders: Resources Elders: Resources

Hogaan 23 (15,● 32)

0 (−7,● 6)

Non−imp 24 (15,● 33)

VC: Resources VC: Resources

Hogaan 37 (29,● 44)

11 (3,● 19)**

Non−imp 26 (18,● 34) significance a● negative DC: Resources DC: Resources a● none a● positive Hogaan 28 (22,● 33)

−14 (−21,● −6)***

Non−imp 41 (33,● 49)

RL: Resources RL: Resources

Hogaan 4 (3,● 5)

2 (0,● 4)

Non−imp 2 (0,● 4)

0% 10% 20% 30% 40% 50% −20% −10% 0% 10% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 106: Comparison of Hogaan and Non-Implementation Villages on Institutions Responsible for Resource Management

148 Elders: Water Elders: Water

Hogaan 11 (8,● 13)

−4 (−10,● 2)

Non−imp 14 (9,● 20)

VC: Water VC: Water

Hogaan 38 (29,● 46)

14 (4, ●22)***

Non−imp 25 (16,● 32) significance a● negative DC: Water DC: Water a● none a● positive Hogaan 37 (33,● 39)

−18 (−26,● −10)***

Non−imp 55 (47,● 62)

RL: Water RL: Water

Hogaan 16 (1,● 31)

0 (−3,● 1)

Non−imp 17 (1,● 32)

0% 20% 40% 60% −20% −10% 0% 10% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 107: Comparison of Hogaan and Non-Implementation Villages on Institutions Responsible for Providing Clean Water

Elders: Right Justice Elders: Right Justice

hogan 23 (19,● 27)

−4 (−10,● 3)

comparison 26 (20,● 33)

VC: Right Justice VC: Right Justice

hogan 20 (15,● 25)

7 (0,● 13)

comparison 13 (8,● 19) significance DC: Right Justice DC: Right Justice a● negative a● none hogan 10 (8,● 14)

−7 (−15,● 0)*

comparison 17 (10,● 25)

RL: Right Justice RL: Right Justice

hogan 49 (38,● 58)

1 (−7,● 10)

comparison 48 (37,● 57)

0% 20% 40% 60% −15% −10% −5% 0% 5% 10% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 108: Rights to Justice: Comparison Village Type

149 Elders: Right Equality Elders: Right Equality

hogan 21 (15,● 27)

−3 (−9,● 3) comparison 24 (16,● 32)

VC: Right Equality VC: Right Equality

hogan 28 (21,● 35)

14 (8, ●21)*** comparison 14 (9,● 20) significance a● negative DC: Right Equality DC: Right Equality a● none a● positive hogan 12 (9,● 14)

−9 (−17,● −3)** comparison 21 (13,● 28)

RL: Right Equality RL: Right Equality

hogan 42 (28,● 54)

−6 (−14,● 1) comparison 49 (34,● 60)

0% 20% 40% 60% −10% 0% 10% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 109: Institutions and Equality Rights: Comparison Village Type

Elders: Right Education Elders: Right Education

hogan 15 (10,● 21)

1 (−4,● 7) comparison 13 (8,● 20)

VC: Right Education VC: Right Education

hogan 35 (27,● 43)

12 (5, ●19)*** comparison 23 (15,● 31) significance a● negative DC: Right Education DC: Right Education a● none a● positive hogan 41 (36,● 45)

−15 (−24,● −6)*** comparison 56 (46,● 63)

RL: Right Education RL: Right Education

hogan 11 (8,● 13)

0 (−4,● 4) comparison 10 (7,● 14)

0% 20% 40% 60% −20% −10% 0% 10% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 110: Institutions and Education: Comparison Village Type

150 Elders: Right Health Elders: Right Health

hogan 14 (11,● 18)

1 (−5,● 7) comparison 13 (8,● 19)

VC: Right Health VC: Right Health

hogan 36 (28,● 44)

13 (6, ●20)*** comparison 23 (16,● 31) significance a● negative DC: Right Health DC: Right Health a● none a● positive hogan 45 (39,● 51)

−15 (−23,● −7)*** comparison 61 (52,● 68)

RL: Right Health RL: Right Health

hogan 6 (4,● 7)

1 (−3,● 4) comparison 4 (2,● 8)

0% 20% 40% 60% −20% −10% 0% 10% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 111: Institutions and Health: Comparison Village Type

Elders: Right Consult on Decisions Elders: Right Consult on Decisions

hogan 32 (28,● 36)

−2 (−9,● 4) comparison 35 (28,● 41)

VC: Right Consult on Decisions VC: Right Consult on Decisions

hogan 36 (30,● 42)

14 (7, ●20)*** comparison 22 (16,● 29) significance a● negative DC: Right Consult on Decisions DC: Right Consult on Decisions a● none a● positive hogan 24 (17,● 32)

−11 (−18,● −4)** comparison 35 (26,● 45)

RL: Right Consult on Decisions RL: Right Consult on Decisions

hogan 9 (7,● 11)

0 (−4,● 4) comparison 9 (6,● 13)

0% 10% 20% 30% 40% −10% 0% 10% 20% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 112: Institutions and Consultation: Comparison Village Types

151 Elders: Right Speech Elders: Right Speech

hogan 23 (19,● 28)

−1 (−8,● 6) comparison 24 (17,● 31)

VC: Right Speech VC: Right Speech

hogan 31 (24,● 38)

11 (4, ●18)*** comparison 20 (14,● 26) significance a● negative DC: Right Speech DC: Right Speech a● none a● positive hogan 29 (21,● 37)

−12 (−20,● −6)*** comparison 42 (31,● 51)

RL: Right Speech RL: Right Speech

hogan 19 (11,● 28)

0 (−6,● 6) comparison 19 (11,● 29)

0% 20% 40% −20% −10% 0% 10% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 113: Institutions and Free Speech: Comparison Village Type

Elders: Right Life Elders: Right Life

Hogaan 45 (40,● 49)

−10 (−19,● −2)**

Non−imp 55 (46,● 63)

VC: Right Life VC: Right Life

Hogaan 25 (21,● 29)

8 (2,● 15)*

Non−imp 16 (11,● 23) significance a● negative DC: Right Life DC: Right Life a● none a● positive Hogaan 16 (13,● 20)

−2 (−9,● 4)

Non−imp 19 (12,● 25)

RL: Right Life RL: Right Life

Hogaan 16 (12,● 22)

3 (−3,● 9)

Non−imp 13 (7,● 20)

0% 20% 40% 60% −20% −10% 0% 10% percentage percentage difference (a) Percentage (%) (b) Diﬀerence in %

Figure 114: Institutions and Right to Life: Comparison Village Type

152 Does VC Provide Services Does VC Provide Services

no social/international contribution 70 (54,● 79)

4 (−1,● 9)

some social/international contribution 74 (60,● 82)

Can VC Manage Projects Can VC Manage Projects

no social/international contribution 73 (60,● 82)

8 (3, 13)***●

some social/international contribution 81 (70,● 87)

Does VC Manage Money Transparently and Accountably Does VC Manage Money Transparently and Accountably

no social/international contribution 70 (58,● 79)

7 (2, ●13)**

some social/international contribution 78 (67,● 84) significance Does DC Provide Services Does DC Provide Services a● none a● positive no social/international contribution 69 (57,● 78)

9 (4, 14)***●

some social/international contribution 78 (67,● 85)

Can DC Manage Projects Can DC Manage Projects

no social/international contribution 67 (58,● 75)

11 (6, ●16)***

some social/international contribution 79 (70,● 84)

Does DC Manage Money Transparently and Accountably Does DC Manage Money Transparently and Accountably

no social/international contribution 68 (58,● 74)

9 (3, 14)***●

some social/international contribution 76 (69,● 81)

0% 25% 50% 75% 0% 5% 10% 15% percentage percentage difference (a) Proportions (b) Diﬀerence in Proportions

Figure 115: Government Service Perceptions and Social Services and International Contributions

153 Appendix C Appendix: Diagramatic Representation of Original and Revised Theory of Change

154 Citizens from a cross-section of the community (including minority clans, women and youth) have better understanding of the roles and responsibilities of formal institutions, their relationship to citizens (rights and responsibilities) Outcome 1: Citizen Participation in Citizens learn participation and meeting management skills Decision Making and hold village institutions to account

Citizens have access to formal mechanisms for en- gaging with village institution (both informing decisions and holding duty-bearers to account)

Roles and responsibilities of village institutions are Outcome 2: vil- agreed on and understood by institutions and communities, including consultation and accountability lage institutions mechanisms and recognised in regional policies demonstrate management of development funds Capacity of Village Councils to plan and manage services (in collaboration with other local institutions) in a transparent in a way that is transparent and equitable is improved and equitable way enabling e↵ective social development/ Service delivery / social development initiatives are provided by Village Councils service delivery at village level

the capacity of Village Councils to prepare and implement Advocacy Plans is improved Outcome 3: Village Impact: Local institutions advocate Government more e↵ectively to Village Councils advocate for resources inﬂuence district Responsive, level policy, resourc- Accountable ing and practice and E↵ective Participation of Village Councils in key District Committees (Safety, Planning, Education, Health, WASH) is improved

Village Council priorities reflected Outcome 4: District in District Development Plans level organs or institutions respond to the needs

Village Councils demonstrate readi- represented by ness to implement development activities village institutions

Outcome 5: Both Village level institutions (formal and informal) and citizens and insti- citizens have understanding and skills in conflict management and have identified different approaches tutions have acted and mechanisms for dealing with conflict at different stages (from addressing root causes, inter-personal to prevent (address responses and formal conflict resolution mechanisms) root causes) and resolve conﬂicts Within a district, villages have collectively identified conflict issues (particularly related to natural through appro- resource management) and agreed ways to address these priate channels according to the conﬂict context

Figure 116: Original Hogaan Theory of Change

155 Leaders trained in advocacy or conﬂict resolution

Leaders trained to provide transparent, accessible Leaders trained and account- in civic education able services to communities

Outcome 2: Village Level Institutions able to Plan, Manage and Co-ordinate

Impact: Local Government Responsive, Accountable and Eﬀective

Outcome 1:Citizen Participation in Decision Making and Conﬂict Mitigation

Citizens trained Citizens trained to coordinate to participate priorities with in collaborative government discussions about stakeholders local development

Figure 117: Replacement Hogaan Theory of Change

156