UNHCR Shelter Source

Evalua on of the UNHCR Shelter Assistance
Programme
Final Dra
The Maastricht Graduate School of Governance (MGSOG) is the Public Policy Graduate
School of Maastricht University, combining high-level teaching and research.
The institute provides multi-disciplinary top-academic training. Doing so, it builds on the academic
resources of the different faculties at Maastricht University as well as those of several foreign partners.
In January 2011, the School became part of the United Nations University, strengthening further its
international training and research network while building on the expertise of UNU-MERIT the
Maastricht based research institute of the UNU. One of the key areas of education and research is
Migration Studies, where MGSOG has gained a strong reputation.
Samuel Hall. (www.samuelhall.org) is a research and consulting company with headquarters in
Kabul, Afghanistan. We specialise in socio-economic surveys, private and public sector studies,
monitoring and evaluation and impact assessments for governmental, non-governmental and
international organisations. Our teams of field practitioners, academic experts and local interviewers
have years of experience leading research in Afghanistan. We use our expertise to balance needs of
beneficiaries with the requirements of development actors. This has enabled us to acquire a firm grasp
of the political and socio-cultural context in the country; design data collection methods and statistical
analyses for monitoring, evaluating, and planning sustainable programmes and to apply cross-
disciplinary knowledge in providing integrated solutions for efficient and effective interventions.
Acknowledgements
The research team would like to thank, first and foremost, the men, women, children who agreed to
participate in this research and share their experiences throughout the 15 provinces surveyed. Second,
we express our gratitude for the commitment of a team of 50 field staff who worked tirelessly over the
course of 8 weeks during the fall of 2012. The research teams in Afghanistan were led by Nassim Majidi
and Hervé Nicolle, project directors, Camille Hennion, project manager, Fatma Wakil and Lucile Martin,
project officers and Ibrahim Ramazani and Abdul Basir Mohmand, field coordinators. The research has
been jointly designed and co-authored by Melissa Siegel, Craig Loschmann, and Katrin Marchand at the
Maastricht Graduate School of Governance (MGSoG) and Nassim Majidi and Camille Hennion, at Samuel
Hall.
Cover Picture © Lucile Martin
2
TABLE OF CONTENTS
List of Boxes .......................................................................................................................... 4
List of Figures......................................................................................................................... 4
List of Acronyms .................................................................................................................... 7
1. Introduction...................................................................................................................19
1.1 Shelter for Displaced Populations in Afghanistan................................................................ 19
1.2 Objectives: Assessment of the Shelter Assistance Programme ............................................ 20
1.3 Key Concepts..................................................................................................................... 21
1.4 Shelter Assistance in a Changing Humanitarian Context: 2009-2011 and beyond ................. 22
1.5 Report Outline................................................................................................................... 25
2. Methodology .................................................................................................................26
2.1 Quantitative Data Collection and Analysis .......................................................................... 26
2.2 Qualitative Data Collection and Analysis ............................................................................ 35
2.3 Limitations and Constraints ............................................................................................... 38
3. Shelter Assistance Programme: Design, Support and Monitoring & Evaluation ...............40
3.1 Design and physical aspects of the shelter .......................................................................... 41
3.2 Support ............................................................................................................................. 72
3.3 Monitoring & Evaluation.................................................................................................... 77
4. Beneficiary Selection......................................................................................................82
4.1 The Selection Process and Actors Involved ......................................................................... 83
4.2 Selection Criteria on the ground: vulnerability side-lined .................................................... 86
4.3 Targeting of the Most Vulnerable: Main categories left out ................................................ 98
4.4 Factors explaining the weaknesses of the selection process .............................................. 110
5. Socio-Economic Impact of the Shelter Assistance Programme..........................................113

5.1 Impact on Beneficiary Households ................................................................................... 114
5.2 Impact on Communities ................................................................................................... 136
6. Partnerships with Other Stakeholders ..........................................................................140

6.1 Shelter Assistance in Afghanistan..................................................................................... 141
6.2 Assessing Partnerships..................................................................................................... 144
7. Conclusion: What is the impact of UNHCR’s shelter assistance on the sustainable

reintegration of returning refugees and IDPs? .....................................................................155
7.1 Overall Conclusion........................................................................................................... 155
7.2 Reflections on Reintegration and Sustainability ................................................................ 158
7.3 Assessing the SAP Guiding Principles ................................................................................ 167
7.4 Strategic evolutions of the SAP ........................................................................................ 171
8. Recommendations: Reinforcing SAP Guiding Principles ................................................172
3
Bibliography .......................................................................................................................187
Annex 1: Household Sampling by Districts and Categories of Respondents .........................194
Annex 2: List of Key Informant Interviews ..........................................................................196
LIST OF BOXES
Box 1: Access to Water…………………………………………………………………………………………………………………….… . 59

Box 2: Best Practices: Inter-community Relations ...................................................................................130
Box 3: Occupancy Rate – Methodological Considerations.......................................................................153
Box 4: Shelter – An Incentive for Return and Settlement? ......................................................................155
Box 5: Implementing SAP in Insecure Areas.............................................................................................157
LIST OF FIGURES
Figure 1: Donors Earmarked Contributions 2009-1011 (Million USD) ..................................................... 151
LIST OF TABLES
Table 1: Household Survey by Beneficiary Status and Migratory Status ................................................... 27
Table 2: Household Survey Sampling by Province...................................................................................... 29
Table 3: Household Survey Sampling by Type of Location......................................................................... 29
Table 4: Community Survey Sampling........................................................................................................ 30
Table 5: Multi-Dimensional Poverty Index ................................................................................................. 33
Table 6: Breakdown of KIIs per Province and Type of Respondent ........................................................... 37
Table 7: Composition of Focus Groups....................................................................................................... 38
Table 8: Type of Shelter by Location (UNHCR Beneficiaries) ..................................................................... 42
Table 9: Type of Shelter by Province (UNHCR Beneficiaries) ..................................................................... 43
Table 10: Dissatisfaction with Elements of the Shelter Programme.......................................................... 46
Table 11: Average Household Size by Province.......................................................................................... 47
Table 12: Dissatisfaction with Elements of the Shelter Programme by Location (UNHCR Beneficiaries) . 48
Table 13: Dissatisfaction with Elements of the Shelter Programme by Region (UNHCR Beneficiaries) ... 49
Table 14: Satisfaction with the Model (Community Representatives) ...................................................... 51
Table 15: Received Necessary Materials (UNHCR Beneficiaries) ............................................................... 53
Table 16: Procurement of Wood by Province (UNHCR Beneficiaries) ....................................................... 54
Table 17: Problems during Construction*.................................................................................................. 55
Table 18: Problems during Construction by EVI Status (UNHCR Beneficiaries)* ....................................... 56
Table 19: Problems during Construction by Location (UNHCR Beneficiaries)* .......................................... 57
Table 20: Amounts Paid by Beneficiaries in AFN by Location .................................................................... 58
Table 21: Percentage of Monthly Income Paid for Shelter by Beneficiaries by Location .......................... 58
4
Table 22: Amounts Paid in AFN by Province (UNHCR Beneficiaries).......................................................... 60
Table 23: Percentage of Monthly Income Paid For Shelter by Province (in %) (UNHCR Beneficiaries) ..... 62
Table 24: Hiring Labourers during Construction ........................................................................................ 64
Table 25: Hiring Labourers during Construction by Location (UNHCR Beneficiaries) ................................ 64
Table 26: Hiring Labourers during Construction by Province (in %) (UNHCR Beneficiaries)...................... 65
Table 27: Duration of Construction by Location (UNHCR Beneficiaries).................................................... 66
Table 28: Average Level of Debt by Location and Beneficiary Status ........................................................ 67
Table 29: Level of Debt by Province and Beneficiary Status ...................................................................... 69
Table 30: Impact of Shelter Programme on Household Debt (in %) .......................................................... 70
Table 31: Completion of Construction (UNHCR Beneficiaries)................................................................... 70
Table 32: Training Received in Conjunction with Shelter Assistance.......................................................... 73
Table 33: Training by Location (UNHCR Beneficiaries)*............................................................................. 74
Table 34: Training by Remoteness of Location (UNHCR Beneficiaries)*.................................................... 74
Table 35: Complementary assistance to EVIs (UNHCR Beneficiaries)*...................................................... 75
Table 36: Complementary assistance to EVIs by Location (UNHCR Beneficiaries)* .................................. 76
Table 37: Monitoring of Shelter ................................................................................................................. 78
Table 38: Monitoring practices by province (UNHCR Beneficiaries) .......................................................... 78
Table 39: Monitoring of Shelter by Region (UNHCR Beneficiaries)............................................................ 79
Table 40: Additional Shelter Owned by Household ................................................................................... 81
Table 41: UNHCR Beneficiary Categories by Migratory Status .................................................................. 87
Table 42: EVI Status of UNHCR Beneficiaries ............................................................................................. 88
Table 43: EVI Status by Province of UNHCR Beneficiaries........................................................................... 89
Table 44: Beneficiary Status of EVI Households ......................................................................................... 91
Table 45: Migratory Status of Non-Beneficiary EVI Households ................................................................ 92
Table 46: Failed Applications for Shelter Assistance by EVI Category and Migratory Status..................... 93
Table 47: Used Selection Criteria (Community Representatives) ............................................................... 94
Table 48: VRF Form and Shelter Assistance ............................................................................................... 95
Table 49: Time between Return/Arrival and Selection into Programme (UNHCR Beneficiaries) .............. 95
Table 50: Time of Return ............................................................................................................................ 96
Table 51: Time of Return by EVI Status (UNHCR Beneficiaries) ................................................................. 96
Table 52: Time of Displacement and Time of Return (UNHCR Beneficiaries) ............................................ 97
Table 53: IDP within each category ............................................................................................................ 98
Table 54: Multi-dimensional Poverty Index of IDPs ................................................................................... 99
Table 55: UNHCR Assisted IDP Households by Region ............................................................................... 99
Table 56: IDP UNHCR Beneficiaries by Province....................................................................................... 101
Table 57: Land Ownership before Building Shelter (UNHCR Beneficiaries) ............................................. 102
Table 58: IDP Integration (UNHCR Households)....................................................................................... 104
Table 59: Land Ownership prior to Shelter Assistance............................................................................. 104
Table 60: Multi-dimensional Poverty Index for Landless Households ..................................................... 106
Table 61: Land Allocation in Communities ............................................................................................... 106
Table 62: Land Allocation Solutions ......................................................................................................... 107
Table 63: Marital Status of Female-Headed Households......................................................................... 108
Table 64: Beneficiary Status of Female-Headed Households................................................................... 108
5
Table 65: Multi-dimensional Poverty Index of Female-Headed Households........................................... 109
Table 66: Households with Ill or Disabled Members................................................................................ 109
Table 67: Multi-dimensional Poverty Index for Households with Ill or Disables Members ..................... 110
Table 68: Subjective Impact of Shelter Assistance on Economic Situation of the Household (in %)....... 114
Table 69: Percentage Change in Housing Arrangement Before and After Assistance by Location (UNHCR
Beneficiaries) ............................................................................................................................................. 115
Table 70: Self-Employment as Main Income-generating Activity ............................................................ 119
Table 71: UNHCR Beneficiaries Change in Indebtedness after Receiving Assistance (in %) .................... 120
Table 72: Access to Safe Drinking Water.................................................................................................. 122
Table 73: Type of Electricity ..................................................................................................................... 122
Table 74: Type of Heating......................................................................................................................... 123
Table 75: Distance to nearest health facility ............................................................................................ 124
Table 76: Multi-Dimensional Poverty Index by Beneficiary Status .......................................................... 125
Table 77: Multi-Dimensional Poverty Index by Migration Status ............................................................ 125
Table 78: Multi-Dimensional Poverty Index by Beneficiary and Migration Status .................................. 126
Table 79: Multi-dimensional Poverty Index by Beneficiary Status (Marginal Effects) .............................. 127
Table 80: Multi-dimensional Poverty Index by Migration Status (Marginal Effects) ................................ 127
Table 81: Multi-dimensional Poverty Index by Beneficiary and Migration Status (1) (Marginal Effects). 128
Table 84: One Month before Assistance vs. After Assistance (Time of Survey) ...................................... 131
Table 85: UNHCR Beneficiaries vs. Non-Beneficiaries, First Returned vs. Time of Survey (Marginal
Effects)....................................................................................................................................................... 134
Table 86: UNHCR Beneficiaries vs. Non-Beneficiaries, Before vs. Time of Survey (Marginal Effects) ...... 135
Table 87: Impacts of Shelter Programme on the Community.................................................................. 136
Table 88: Economic Impacts of Shelter Programme ................................................................................ 138
Table 89: Shelter Assistance Programmes in Afghanistan ....................................................................... 141
Table 90: Attitude of Community towards Beneficiaries (in %) ............................................................... 160
Table 91: Plans to Stay in Current Place of Residence (in %) ................................................................... 163
6
LIST OF ACRONYMS
ACBAR Agency Coordinating Body for Afghan Relief
AFN Afghanis
ANDMA Afghanistan National Disaster Management Authority
BSC Beneficiary Selection Committee
CDC Community Development Council
CDP Community Driven Process
DANIDA Danish International Development Agency
DoRR Department of Refugees and Repatriations
DRRD Department of Rural Rehabilitation and Development
DRC Danish Refugee Council
ECHO European Commission
ES/NFI Emergency Shelter/Non Food Items
EVI Extremely Vulnerable Individual
FGD Focus Group Discussion
GoA Government of Afghanistan
HLP Housing, Land, Property
IDMC Internal Displacement Monitoring Centre
IDP Internally Displaced Persons
INGO International Non-Governmental Organisation
IOM International Organisation for Migration
IP Implementing Partner
JIPS Joint IDP Profiling Service
KII Key Informant Interview
LAS Land Allocation Scheme
MoE Ministry of Economy
MoPH Ministry of Public Health
MoRR Ministry of Refugees and Repatriation
MRRD Ministry of Rural Rehabilitation and Development
MGSoG Maastricht Graduate School of Governance
MUDH Ministry of Urban Development Affairs and Housing
NNGO National Non-Governmental Organisation
NSP National Solidarity Programme
NRC Norwegian Refugee Council
PSN Persons with Specific Needs
PSU Primary Sampling Unit
ROI Regions of Origin Initiative
SAP Shelter Assistance Programme
UNDP United Nations Development Programme
UN-OCHA United Nations – Office for the Coordination of Humanitarian Affairs
UNHCR United Nations High Commissioner for Refugees
VRF Voluntary Repatriation Form
WASH Water, Sanitation and Hygiene
WB World Bank
7
EXECUTIVE SUMMARY
With more than 2.7 million Afghan refugees in the region, and an estimated 3 million globally,
Afghanistan has the largest refugee population in the world. Since the fall of the Taliban, the country has
witnessed massive return, with 5.7 million refugees returning and 4.6 million assisted by the United
Nations High Commissioner for Refugees (UNHCR). Yet, Afghanistan in 2012 reflects drastically different
trends from the year the repatriation process started, a decade ago, in 2002. One common and
continuous trend, however, is the overwhelming need for shelter and land for displaced populations –
whether returning refugees or internally displaced persons (IDPs).
UNHCR’s Shelter Assistance Programme (SAP) has provided, since 2002, more than 220,000 units of
shelter to vulnerable returnees and IDPs throughout Afghanistan. The programme’s design and
implementation procedures have been improved over the years. To date, only one internal assessment
of the programme has been conducted by UNHCR – with a limited scope, in 2005. A 2012 evaluation of
the Danish Regions of Origin support to Afghanistan also touched upon the shelter programme.1 Several
other studies have researched the needs and vulnerability of returnees and IDPs in the country2, but the
SAP’s contribution to reintegration outcomes, defined as achieving sustainable return and parity
between returnees and other members of the local community, has not been researched.
The present study conducted by researchers at the Maastricht Graduate School of Governance (MGSoG)
and Samuel Hall Consulting aims at filling this important gap with the following objectives:
1. Assess the shelter program contribution to reintegration outcomes and in achieving parity
between returnees and others;
2. Evaluate the shelter program design in terms of performance at the beneficiary level and its
effectiveness according to UNHCR guidelines;
3. Assess the relevance and sustainability of the shelter programme in the broader context of
humanitarian assistance in Afghanistan.
The study’s scope covers UNHCR shelter beneficiaries and communities targeted between 2009 and
2011 only; but the relevance of its findings will inform future shelter strategies. The evaluation covers
the socio-economic aspects of shelter assistance on four levels – (1) household level, (2) community
level, (3) institutional level, and finally (4) the macro-level.
This executive summary presents the result of the evaluation – covering an overall positive assessment
of the SAP’s reintegration outcome for returnees, discussing key weaknesses and areas for future
improvements in developing improved guiding principles and strategy for the shelter programme in 2013
and beyond. The summary is organized in four parts: (1) UNHCR’S SAP: An effective contribution to
reintegration; (2) UNHCR’S position as a leader on shelter intervention-unmatched by other
stakeholders; (3) weaknesses of the shelter programme; and (4) conclusions and recommendations.
1
Cosgrave J, Bryld E, and Jacobsen, L (2012), Evaluation of the Danish Regions of Origin Support to Afghanistan. Copenhagen:
Danida.
2
See for example: CMI (2008); De BREE (2008); LUMP et al.(2004); BARAKAT et al. 2012); Brookings-Bern Project on Internal
Displacement - The Liaison Office (2010); MAJIDI (2011); Samuel Hall/NRC/IDMC/JIPS (2012).
8
1. UNHCR’S SAP: AN EFFECTIVE CONTRIBUTION TO REINTEGRATION
The socio-economic impact of the SAP has been assessed on a sample of 4,488 respondent households in
15 provinces and all regions of Afghanistan, including 2,035 UNHCR shelter beneficiaries, 1,990 non-
beneficiaries and of 463 shelter beneficiaries from other shelter agencies in the East – the region
representing the largest share of shelter activities.
Shelter beneficiaries fare better than non-beneficiaries

The results from multi-dimensional poverty and cross-sectional regression analyses show:
 Highest rates of poverty among non-shelter beneficiaries: 86 per cent of non-shelter

beneficiaries are multi-dimensionally poor, as opposed to 71 per cent of UNHCR shelter
beneficiaries and the lowest numbers, 68 per cent, among non-UNHCR shelter beneficiaries.
 A lower probability of being poor and lower degrees of poverty for beneficiaries compared
to non-beneficiaries.
Using a difference in difference analysis across time – comparing two periods: after return from abroad
and after receiving assistance – the research concludes that:
 The data collected provides solid evidence that UNHCR’s SAP has had a significant and
positive impact on reducing household deprivation along indicators of interest including
access to a house, electricity, sanitation and access to a mobile.
IDPs fare worse than returning refugees in the reintegration process

However, these positive results should be weighed against two important findings:
 First, over three quarters – 78 per cent – of the overall sample are multi-dimensionally
poor. Concerning the individual dimensions most households are deprived in education,
followed by economic well-being, social capital, housing and health. In terms of
reintegration, this means that returnees are reintegrating in deprived communities. They
fare better than non-shelter beneficiaries, but the context weighs in negatively on all. The
aim of reintegration being to achieve parity is therefore not systematically relevant – the end
result is that beneficiaries surpass the conditions seen in other groups; however, they also
remain multi-dimensionally poor.
 Second, main differences across groups show that the positive reintegration impact is
mainly true for returning refugees and not for IDPs. IDPs are noticeable more deprived than
any other group while refugee returnees are the least deprived.
Comparisons between shelter beneficiaries and non-beneficiaries show, across a range of models, that
the probability of being multi-dimensionally poor is significant and positive for nearly all non-beneficiary
sub-groups in relation to the reference group. This means that beneficiaries are less deprived on a range
9
of socio-economic indicators. However, IDPs remain a marginalized population whose reintegration has
not materialized as they fare worse than all other sub-groups in the sample.
Positive impact on intra-community relations

Overall, most stakeholders expressed their satisfaction with – and even sometimes their gratitude for –
the fact that the shelter programme had been implemented in their community. This positive
assessment is linked to two major effects that the programme has on the community as a whole: i)
supporting the development of villages, and ii) easing potential tensions within the community – as long
as the selection process is seen as fair and transparent.
Among the surveyed households less than 1 per cent do not perceive the impact on the community of
the programme as positive. Among the community leaders this number is higher at 10 per cent.
Nevertheless, the subjective opinion of the impact on the communities is very positive.
2. UNHCR’S POSITION AS A LEADER ON SHELTER INTERVENTION –

UNMATCHED BY OTHER STAKEHOLDERS
By looking at changes over time – a month before assistance and one month after assistance –
differences between UNHCR beneficiaries and non-UNHCR beneficiaries arise. Looking at access to
housing for example UNHCR beneficiary households are 36 per cent less deprived compared to before
receiving the assistance, as opposed to 20 per cent of non-UNHCR beneficiary households. Overall, a
higher percentage of households which were assisted by UNHCR are less deprived on nearly every
indicator, aside from “Heating” and “Flooring”, over the two periods in comparison to households
assisted from other organizations that were not UNHCR.
An important feature of shelter assistance conducted in the country is the very small number of actors
involved in it. No other organisation active in Afghanistan is able to implement a shelter programme that
is comparable in size and scope to UNHCR’s. There have been considerable changes since 2008. A large
shelter coordination system in 2002-2008 has decreased due to worsening security, funding cuts, lower
interest and the dominance of NNGOs as implementing partners. The Government of Afghanistan has
never had a housing policy for rural areas and has suffered from 7 ministerial changes since 2002, and
failing communications lines between provincial directorates and the Ministry of Refugees and
Repatriation. Partners have come and gone – UNDP, UN Habitat, have implemented shelter without
sufficient sustainability due to reductions in funding. UNHCR remains the only leader in the humanitarian
stakeholder landscape that has provided shelter to all regions of Afghanistan.
This means that UNHCR holds a distinct leadership in shelter provision and that any reduction of the
scope of UNHCR’s SAP will hardly be covered by other partners. The potential impact of UNHCR’s
changes in strategy on the lives of vulnerable households is therefore significant – an important factor of
consideration for donors, as well as for UNHCR’s strategic review of its SAP.
10
Furthermore, most other stakeholders (IOM, OCHA, CARE, ZOA, InterSOS and ACTED for instance) focus
on natural-disaster IDPs or on natural-disaster affected populations. Only NRC and UNHCR have a
specific focus on shelter assistance for conflict-induced IDPs and returnees. Limiting the scope of the
UNHCR shelter assistance programme would therefore have particularly negative impacts for conflict-
induced IDPs. Taking into account the growing numbers of conflict-induced IDPs throughout the country,
and the fact that IDPs were marginalized in the reintegration progress assessed in this sample, this is a
group that should be the focus of increased attention– and shelter a cornerstone of durable solutions for
IDPs.
Additionally, UNHCR is the organisation with the widest geographical coverage of its shelter programme,
through all regions of Afghanistan and especially in the South. Southern regions are where most IDPs are
located3, something that UNHCR acknowledged as the central level authorized a specific focus on IDPs
for the shelter programme in the South. This means that even more than anywhere else in the country,
the UNHCR shelter programme answered to specific local needs with the shelter programme. The gap
left by a reduced shelter programme will therefore be particularly acute in the South.
Clearly, UNHCR has a crucial role in terms of shelter assistance, that is unmatched by any other actors in
the country. Any evolution of the programme should take this central role and responsibility into
account.
3. WEAKNESSES OF THE SHELTER ASSISTANCE PROGRAMME

Stronger beneficiary selection and greater inclusion of the most vulnerable households are the two
priorities for ensuring the programme is inclusive of all Extremely Vulnerable Individuals (EVIs).
Financial burden of SAP on beneficiaries

Both quantitative and qualitative data showed a high level of satisfaction of beneficiaries with the shelter
package, which provides good quality material that most beneficiaries would otherwise not have been
able to afford. The distribution process works efficiently with 93 per cent of beneficiaries having received
all the necessary materials for their shelter. Yet, the construction process is a difficult and costly
process for beneficiaries as:
 972 of the beneficiary households (48%) ran into problems during construction.
 89 per cent of the households with problems ran out of money during construction (this
corresponds to 42% of all UNHCR beneficiaries) with
 Significant disparities in household contribution according to provinces/location.
 47 per cent of households that ran into problems (22 per cent of the beneficiaries) reported a
lack of sufficient access to water to build shelters and rely on costly solutions.
3
See for example, Brookings-Bern Project on Internal Displacement - The Liaison Office, (2010), Beyond the Blanket: Towards
more effective protection for internally displaced persons in Southern Afghanistan.
11
The lack of money was mentioned as by far the main challenge faced by beneficiaries when building their
shelters. Almost all beneficiaries mentioned that they had to take up loans to cover labour costs and
wall components. Additional costs were also necessary for buying stones for foundations and, depending
on the availability of material in a given area, bricks, cement or clay.
Household contributions
Almost all beneficiaries, 93 per cent, had to contribute to the shelter construction as per SAP guidelines.
However, the amount of funds contributed varies significantly with urban UNHCR beneficiaries spending
significantly more out of their own pockets than rural beneficiaries. The data shows a 13,000 AFN (260
USD) gap between urban and rural households, and a smaller, yet sizeable gap of 6 810 AFN (136 USD)
between urban and semi-rural households. This is due to the higher costs of materials and labour in
urban areas – higher costs that will have to be taken into account in developing an urban strategy for the
shelter programme, discussed in the recommendations chapter. Moreover, this is also due to the fact
that urban households on average earn a higher income than those of rural or semi-rural areas. To speak
in relative terms, the ratio of amount paid on the shelter to household monthly income of beneficiary
households, providing evidence that while UNHCR beneficiary households located in an urban context
spend more in absolute terms, semi-rural households spend slightly more in relative terms.
Indebtedness
When asked about the impact of the shelter assistance programme on household debt 47 per cent of all
beneficiaries indicated that it increased. This appears to be a bigger problem for beneficiaries of other
programmes (54.4%) than for the UNHCR beneficiaries (34.5%). However, no abandonment of shelters
due to debts contracted because of the programme were noticed in the field, a potential sign that this is
not a major threat to sustainability in the short term, but might become one if sufficient income
opportunities are not secured.
Community support
UNHCR mainly relies on ashar, or community assistance, to support the most vulnerable households.
However, this was not a practice noticed in the field, with community members mentioning ashar could
not be an option, as most villagers were faced with difficulties in sustaining their own household.
Community representatives, however, indicated that community members did assist the beneficiaries in
building their shelters in 60 per cent of cases. Mainly this was assistance in the form of unskilled labour;
in rare cases community members also provided skilled labour and materials.
It is important to stress the very high completion rates of the shelters; only 2 households out of a sample
of 2,035 beneficiaries sampled had not completed their shelter – although qualitative fieldwork urges to
be cautious as field observations reported a higher number of incomplete shelters, not covered in the
quantitative sample. This was mainly due to the incapacity of the beneficiaries to finish building the
shelter and earn a living at the same time.
12
Selection Process: Vulnerable groups sidelined
The selection process clearly appeared as the main weakness in the implementation of the shelter
program as it failed to integrate the most vulnerable. Many flaws in the process were identified during
qualitative fieldwork and confirmed by quantitative analysis. These include:
 Exclusion error: Insufficient focus on and inclusion of vulnerable groups as put forward in the
UNHCR Shelter Guidelines.
 Inclusion error: More than half of non-refugee returnees receiving UNHCR assistance are not
considered to be in the “extremely vulnerable” based on the EVI categories4, indicating a
misallocation of assistance as this group does not present the migratory profile nor signs of
vulnerability that would make them eligible.
The main factor explaining these failures is the significant gap between the SAP guidelines on paper and
the reality of selection as it is conducted on the ground, where the Voluntary Repatriation Form (VRF)
and land ownership take precedence over any other criteria of selection.
The overreliance on the VRF as the main basis for selection has in certain cases led to the under-
representation of particularly vulnerable displaced households. The Multi-dimensional Poverty Index
used in this study shows that the most vulnerable – “the vulnerable within the vulnerable” – are IDPs,
landless households, female-headed households and those households showing illness or disability.
i. IDPs were underrepresented in the selection process, or only included in small proportions:
just above 11 per cent of UNHCR beneficiaries in 2009-2011 were IDPs5. In our own sample, just
above 9 per cent of UNHCR beneficiaries were IDPs and 84 per cent of them are deprived on a
range of socio-economic indicators compared to 77 per cent of non-IDP households. We find
clear targeting of IDP households for shelter assistance in Takhar and Helmand, followed by
Faryab, Hirat and Kandahar. However in all other provinces this clear targeting of IDPs did not
take place.
ii. Landless vulnerable households were underrepresented in the selection process. This is not
surprising given land ownership is in most cases a requirement for receiving shelter assistance.
As such, only a small percentage of UNHCR beneficiaries, 17 per cent, did not own their land
before becoming a beneficiary. This is in contrast to other programs, where 27 per cent of
households were landless before assistance. Community representatives mentioned that the
proposed solutions for providing land to the landless were applied in a very limited number of
cases. In 27 of the 60 communities vulnerable landless people applied to the shelter assistance
4
In 2012 the term ‘Extremely Vulnerable Individuals’ has been replaced by People with Specific Needs (PSN).
According to the PSN guidelines PSN are defined as ‘persons who, due to their specific physical, psychological,
mental and social situation are not able to cope with new circumstances or be integrated or reintegrated without
external support’. These include women at risk, unaccompanied children, disabled persons… As UNHCR was still
using the EVI terminology between 2009 and 2011, we will keep this term throughout this report.
5
UNHCR Assisted IDP families: 2009, 2010, 2011”, UNHCR Data Unit, Kabul.
13
programme. Less than half of the communities were able to provide shelter assistance to those
without land ownership.
iii. Female-headed households: The fact that women were not included in the beneficiary selection
in several areas (only 13 out of 60 communities included women in their Beneficiary Selection
Committees) raises concern about the effective access to female-headed households, but also
about the assessment of the living conditions of potentially eligible families, since, as it was
noticed in the field for our own staff, only women are allowed to enter private areas. Aside from
including female IP staff in selection in some of the provinces, none of the procedures
mentioned in the guidelines for inclusion of females in the selection process have been
mentioned either by sub-offices, IPs or community members. When women were involved, it
was often in marginal roles.
iv. Households with Ill or Disabled members: households with a member who is physically,
mentally or chronically ill are also considered extremely vulnerable. Within our sample, we find
that 35 per cent of households of this type are UNHCR beneficiaries compared to the 38 per cent
from other organizations. While the representation of households with ill or disabled members is
on par with the entire sample, they are more likely to be deprived than the average shown again
by the Multi-dimensional Poverty Index.
Insufficient support to beneficiaries and EVIs

Almost one in three beneficiary households stated not having received any complementary training.
When comparing training provided by UNHCR vs. other shelter agencies, findings show that UNHCR
beneficiaries are worse off in terms of the support they receive. While one in three UNHCR shelter
beneficiaries did not receive any training, this number drops to one in seven in other shelter
programmes. Within the training sessions conducted, most concerning was the gap on hygiene
promotion. While 54 per cent of beneficiaries receive complementary training on construction, less than
20 per cent of them receive hygiene promotion training. UNHCR shelter beneficiaries are significantly
less likely than other shelter beneficiaries to receive any hygiene support.
The research team observed that the link between protection and the shelter programme is insufficient
at the sub-office level, as the mechanism in place to identify and provide additional support to EVIs is
inefficient. The programme guidelines and UNHCR’s EVI programme plan for additional cash assistance
for EVI beneficiaries, but the research team found very rare examples of this practice actually
implemented in the field.
Overall, complementary support to beneficiaries, and particularly to EVIs, is insufficient. The research
shows inefficient mechanisms to provide additional assistance to EVIs to build their shelter. There is
therefore no conclusive finding that shows that additional support and complementary training are
made available to EVIs more than the average beneficiary. This shows that the shelter programme can
aim to focus more on EVIs not only in the selection process, but in the support trainings provided as well.
14
Beneficiary complaints / Urban dissatisfactions
The main complaints raised – by all shelter beneficiaries, UNHCR and non-UNHCR alike – are:
 The quality of technical assistance – the highest level of dissatisfaction raised by 14 per cent of
UNHCR beneficiaries, and even more – 22 per cent – in other shelter programmes.
 The quality of latrines – raised by 12 per cent of UNHCR beneficiaries, less than the 18 per cent
of other programme beneficiaries.
 The size of the shelter – complaints were greater among UNHCR beneficiaries (11.5%) than
other shelter beneficiaries (5.8%).
Urban beneficiaries were more critical of the quality of technical assistance and the quality of latrines
provided by the shelter programme. Specifically, their dissatisfaction ranked twice as high as their rural
counterparts, and three times that of their semi-rural counterparts.
As such, the data underlines an expectations gap between what the shelter programme offers and urban
households’ needs. There is an added pressure in urban areas to have adequate housing – in terms of
quality but also in terms of appearance, to blend in more effectively within the urban landscape. The
UNHCR shelter model was seen as being too rudimentary for urban households. The latrines provided
proved ill-adapted and will be considered in the recommendations section. An added focus will be
needed in future shelter strategies on the ways the SAP can be adapted to an urban context that is
increasingly home to internal displacement and refugee return.
Beyond the urban specificities, semi-rural households also raised concerns – above that of their
counterparts – on the size of the shelter, the quality of windows and the design of the shelter.
Inadequate risk mitigation and prevention mechanisms

Risk mitigation measures are not properly integrated in the implementation of the shelter program,
limiting sustainability of the SAP. Preventive measures imposed by the programme’s guidelines are
limited and only cover earthquake-mitigation measures.
In earthquake-prone areas, risk mitigation is solely taken into account through the inclusion of wood-
bracing which were often removed by beneficiaries, due to a lack of awareness of their use. This
emphasizes the need for proper awareness training about the importance of such elements.
Preventive measures against floods are also seriously lacking, including proper risk assessments and the
possibility not to include a village in the SAP if it is located on flood-prone areas. This was notably the
case in Nangarhar, Hirat and Jawzjan. In the latter, despite high risks in the province, the only measure
recommended in practice by UNHCR was to build the shelters 60 cm above the ground, which was not
systematically implemented across the province and is insufficient in case of serious flooding. In
Kandahar, Hirat, Jawzjan, Parwan and Nangarhar inhabitants insisted on the need to build retaining walls
to support the sustainability of the shelters.
15
4. CONCLUSIONS & RECOMMENDATIONS
Security has deteriorated consistently in Afghanistan since 2005 making it more difficult – and costly –
for UNHCR to undertake extensive project monitoring, visiting the field and beneficiaries directly.
Related questions of access, costs, logistics and UNDSS restrictions are now part of the operational
framework limiting UNHCR in its work in Afghanistan.
In addition, the partnership opportunities of the shelter programme have not increased, to the contrary
– the Government of Afghanistan does not have a housing policy for rural areas and communications
between provincial directors and Kabul-level representations of ministries remain a key challenge;
coordination systems on shelter alone have been discontinued due to lower funding, interest and access
of international humanitarian actors; and national NGOs now dominate the implementation field.
UNHCR is the sole humanitarian actor with an uncontested leadership of shelter activities.
This study takes into consideration this changing humanitarian context, and the unchanged humanitarian
needs of the displaced in Afghanistan – with land and shelter being the priority needs for returning
refugees and IDPs. This study also takes into account decreasing funding available to UNHCR – and will
therefore frame recommendations that are deemed to be ‘implementable’ for UNHCR as of 2013.
UNHCR will have to focus on activities that are “on budget” and activities that have shown their success.
In an environment of limited funds, access and time, the overarching recommendation of this report is
to continue the SAP, on a nationwide scale, investing resources on shelter (with better targeting)
rather than on new initiatives that are not ‘on budget’ that have not proven their success, and that reach
lower numbers of beneficiaries, thereby potentially creating inter-community tensions.
The SAP focus on vulnerability in its guidelines has been to assist those that showed the greatest needs.
A qualitative rather than quantitative approach, a needs-based rather than location-based approach is at
the core of the objectives of the SAP – and should remain at its core in future strategies.
Given the key findings of this research and the evidence of SAP’s contribution to reintegration, the
question can no longer be ‘Should the shelter programme continue to be implemented in Afghanistan?’
but rather ‘How should the programme evolve to:
1. Better adapt to the current migratory trends of the country;

2. Better fit the needs of the most vulnerable;
3. Be more inclusive of IDPs and other vulnerable segments of the population;
4. Be sustainable in an increasingly complex humanitarian context?’
The study’s main findings point to the need for a more protection-focused, needs and evidence-based
approach to the shelter programme in Afghanistan. Current plans to limit the SAP to ‘reintegration
sites’ across the country are unrealistic and counter-productive, as they tend to increase tensions
between communities. The main recommendation of the research is to continue implementing the
Shelter Assistance Programme as a cornerstone of UNHCR’s activities, as a humanitarian agency, with a
necessary update of the existing UNHCR Shelter Programme Guiding Principles (2011).
16
REINFORCING SAP’S GUIDING PRINCIPLES
The recommendations of this research are based on the existing SAP Guiding Principles – elements
highlighted in bold/orange are the recommended additions to the Guiding Principles for 2013 and
beyond. From 8 main Guiding Principles, the research team proposes a set of 11 Guiding Principles.
These include two types of recommendations are presented:
First, the addition of new Guiding Principles – such as the need to integrate Impact and Needs
Assessments (item 1), IDPs’ direct participation (item 4), a Partnership Strategy (item 10) and a
Monitoring Framework (item 11). These additions are both the most relevant to the project
planning cycle and to the changing humanitarian context of Afghanistan. Impact, Needs
assessments and Monitoring frameworks are prerequisite for any accountable and transparent
implementation process, while IDPs’ direct participation and a solid Partnership Strategy are
requirements imposed by a humanitarian context defined by increasing internal displacement
and lack of access.
Second, the strengthening of already existing Guiding Principles – Our recommendations seek to
improve, and often breakdown in more detail, principles such as the Community-based approach
(item 2), Women’s direct participation (item 3), Access to land (item 5), Focus on vulnerability
(item 6), Environmental concerns (item 7), and the Preservation of cultural and regional
preferences (item 8). These are principles that were found, in our research and fieldwork, to be
weak in their implementation – and hence need to be strengthened by better adapting to the
challenges at the field level.
The proposed set of 11 SAP Guiding Principles below – detailed in the core recommendations chapter of
the report – is a “ready to use” revised set of guidelines for UNHCR’s 2013 programmatic review.
2013 SAP GUIDELINES – 22 PRINCIPLES
1. Impact and Needs Assessments

a. Baseline
b. Calendar and flexibility of construction process
c. Setting standard for household contribution
2. Community-based approach
a. Increasing the degree of transparency of the selection process
b. Impact on non-beneficiaries
c. Complementary assistance
3. Women’s direct participation

a. Include women and gender criteria in the selection of beneficiaries
b. Include gender criteria in the selection process of IPs
17
4. IDPs’ direct participation
a. Increasing the proportion of IDPs
b. Include IDPs in the selection of beneficiaries
5. Access to land
a. Evidence of land ownership or NOC
b. Legal assistance in cases of land dispute, inheritance, mahr
6. Focus on vulnerability
a. Beneficiary Selection Committee
b. No family overlooked
c. No contribution requirements for EVIs
d. Training / sensitization workshops
7. Environmental concerns
a. Alternative materials
b. Latrine per family
c. Hygiene and Sanitation Training
d. Adopting a regional risk mitigation approach
8. Preservation of cultural and regional preferences

a. Flexibility in design
b. Adopting an urban approach
9. Contribution to local economies

a. Reviving local economies
b. Local procurement of raw materials
10. Partnership Strategy

a. Involvement of local authorities
b. Involvement of CDCs
c. Linkages with civil society
d. Linkages with development actors
11. Monitoring Framework

a. Internal monitoring - Increase involvement of UNHCR staff
b. Community-based monitoring
c. Guidelines on corruption and fraud
d. Monitoring framework and follow-up mechanisms
18
1. INTRODUCTION
1.1 SHELTER FOR DISPLACED POPULATIONS IN AFGHANISTAN

With more than 2.7 million Afghan refugees in the region, and an estimated 3 million globally,
Afghanistan has the largest refugee population in the world. Since the fall of the Taliban, the country has
witnessed massive return, with 5.7 million refugees returning and 4.6 million assisted by the United
Nations High Commissioner for Refugees (UNHCR). Yet, in 2012 Afghanistan reflects drastically different
trends from the year the repatriation process started, a decade ago, in 2002.
First, the number of refugee returns has dropped to less than 70,000 in 2011. “For the first time since
2002 (…) the country has a negative migration rate: more Afghans are leaving than returning”6. Internal
displacement is now the growing humanitarian concern, with a population estimated at over half a
million individuals7. Given the deterioration of security in Afghanistan and the withdrawal of
international forces, the rise of internal displacement will continue to be a key trend in coming years –
and a key priority for the humanitarian community, and for UNHCR, the lead aid agency on conflict-
induced displacement. This context of increasing insecurity – especially since 2005 – is a reality with
which UNHCR has had to work in order to develop its programme, and will hence frame our analysis.
Second, returnees’ preferences for urban settings and their inability, or unwillingness, to return to their
province of origin have resulted in a massive influx of returnees and IDPs to urban areas. This raises
concerns about the absorption capacities of rapidly growing urban areas and access to livelihood
opportunities for newcomers8, and about the ability to provide durable solutions to displaced
populations.
One common trend – in this changing context – is the need for shelter and land; the lack of which
severely impacts the overall vulnerability, poverty levels and livelihood potential of the growing numbers
of displaced populations. Not having access to land or shelter, and lacking security of tenure, prevents
displaced populations from breaking an enduring cycle of poverty.
Under this premise, UNHCR, with the support of the Government of Afghanistan (GoA) and the
international community, established a shelter assistance programme targeting refugees and internally
displaced persons (IDPs). For ten years, the Shelter Assistance Programme (SAP) has been the
cornerstone of UNHCR’s assistance to voluntary returnees in Afghanistan, with more than 220,000
constructed shelters since 2002. After a decade of shelter assistance, key questions remain:
 Has the programme effectively contributed to reintegration outcomes for displaced populations?
 Has the programme adequately targeted the most vulnerable within the displaced populations
and has it been implemented according to its guiding principles?
6
Sean Carberry: “Afghans Begin New Exodus, Often At Great Cost”, NPR, December 2, 2012.
7
Samuel Hall/NRC/IDMC/JIPS (2012) The challenges of IDP Protection – Research Study on the Protection of Internally Displaced
Persons in Afghanistan.
8
MAJIDI Nassim, « Urban Returnees and Internally Displaced Persons in Afghanistan », Middle East Institute, Fondation pour la
Recherche Stratégique, January 25, 2011.
19
1.2 OBJECTIVES: ASSESSMENT OF THE SHELTER ASSISTANCE
PROGRAMME
UNHCR’s Shelter Assistance Programme has provided, since 2002, more than 220,000 units of shelter to
vulnerable returnees and IDPs throughout Afghanistan. The programme’s design and implementation
procedures have been improved over the years. To date, only one internal assessment of the programme
has been conducted by UNHCR – with a limited scope, in 2005. A 2012 evaluation of the Danish Regions
of Origin support to Afghanistan also touched upon the shelter programme.9 Several other studies have
researched the needs and vulnerability of returnees and IDPs in the country10, but the SAP’s contribution
to reintegration outcomes, defined as achieving sustainable return and parity between returnees and
other members of the local community, has not been researched.
The present study conducted by researchers at the Maastricht Graduate School of Governance (MGSoG)
and Samuel Hall Consulting aims at filling this important gap and its objectives are:
1. Assess the shelter program contribution to reintegration outcomes and in achieving parity
between returnees and others;
2. Evaluate the shelter program design in terms of performance at the beneficiary level and its
effectiveness according to UNHCR guidelines;
3. Assess the relevance and sustainability of the shelter programme in the broader context of
humanitarian assistance in Afghanistan.
The evaluation covers the socio-economic aspects of shelter assistance through a multi-dimensional
poverty analysis to answer key research questions, at four levels:
a) At the household level: A quantitative survey, direct field observation, focus group discussions
and qualitative interviews, to assess if the programme is efficiently targeting the most
vulnerable.
b) At the community level: A comparison of the situation of beneficiaries vs. non-beneficiaries in
communities, to assess the integration of returnees and IDPs and the socio-economic impact of
the programme on communities at large.
c) At the organizational and institutional level: An analysis of the responses of stakeholders, the
strengths, weaknesses and opportunities of existing shelter programs and partnerships in
Afghanistan.
d) At the macro-level: An evidence-based analysis of the context, incorporating information about
the evolution of the humanitarian context in Afghanistan.
9
Cosgrave J, Bryld E, and Jacobsen, L (2012), Evaluation of the Danish Regions of Origin Support to Afghanistan. Copenhage:
Danida.
10
See for example: CMI (2008); De BREE (2008); LUMP et al.(2004); BARAKAT et al. 2012); Brookings-Bern Project on Internal
Displacement - The Liaison Office (2010); MAJIDI (2011); Samuel Hall/NRC/IDMC/JISP (2012).
20
1.3 KEY CONCEPTS
The key concept at the heart of the study is sustainable reintegration, as the conclusion of this research
will be dedicated to analysing the impact of the shelter programme on the sustainable reintegration of
returning refugees and IDPs.
The notion of “sustainable return” is a long-term, contextual understanding of return incorporating

social and economic dimensions. It is possible to draw a distinction between:
1. Narrow indicators at the individual or household level, such as whether returnees subsequently
re-migrate.
2. Broader definitions, which understand sustainability as involving both the reintegration of
individual returnees in their home societies, and the wider impact of return. “The broader
definition suggested also draws attention to the idea that continued mobility after an initial
return – including circulation and the development of a “transnational” lifestyle – may be more
“sustainable” than a single and definitive return to the refugee’s place of origin.”11
Drawing from the definition according to which a livelihood is considered “sustainable” if it can be
maintained without external inputs, Black and Gent suggest a benchmark for “sustainable return” both
at the individual and aggregate (community, region) level according to the increased or reduced reliance
on external inputs (humanitarian and development aid) and vulnerability of economic, social and
political systems of the areas of return.12 The assessment of a “sustainable return” should therefore not
only prioritize outcomes for refugees, but, at a larger scale, take into consideration the impact on the
entire community.
Sustainable reintegration is understood as a process achieving parity with other community members.
The comparative measurement is between returnees and other community members, between
beneficiaries and non-beneficiaries, using one group as a control group to assess levels of reintegration.
The concept of reintegration lacks a standard international definition. However, the guidelines of the
UNHCR Handbook for Repatriation and Reintegration Activities provide elements constituting a starting
point to grasp the implications behind the concept: “Reintegration is a process that should result in the
disappearance of differences in legal rights and duties between returnees and their compatriots and the
equal access of returnees to services, productive assets and opportunities” leading to a “sustainable
return – in other words, the ability of returning refugees to secure political, economic (legal) and social
conditions needed to maintain life, livelihoods and dignity”13. The concept of reintegration therefore
places the emphasis on the disappearance of differences between the returnee and the host population,
the access to the same legal rights, equal services, productive assets and opportunities.
On the operational level, this means:
11
Richard BLACK, Saskia GENT, « Sustainable Return in Post-Conflict Context », Sussex Center for Migration Research, Brighton ,
UK, 2006.
12
R. BLACK, S GENT, op. cit.
13
Handbook for Repatriation and Reintegration Activities, UNHCR, Geneva, 2004, p. 4-5.
21
1. Taking into account the general context of return, i.e. not merely focusing on returnees but
taking into account the whole community in which the reintegration process is meant to take
place, with a relative comparison of returnees and non-returnees within communities.
2. A broader coordination of actors involved in reintegration activities, with a clear understanding
of the division of responsibilities to avoid gaps and overlaps.
3. Involvement of national authorities to mainstream reintegration as an institutional process.
1.4 SHELTER ASSISTANCE IN A CHANGING HUMANITARIAN CONTEXT:

2009-2011 AND BEYOND
This research is designed around these objectives and concepts to assist UNHCR in strategically assessing
the future of its shelter programme, by looking at lessons learned from the past. The shape of UNHCR’s
shelter programme should be determined taking into account the results of this study.
Several other stakeholders, such as the Norwegian Refugee Council (NRC) or UN-Habitat, have built on
their past experience to incrementally adapt their shelter assistance to fit with the new Afghan migration
context and to meet the evolving needs of their populations of concern. The central point of this study is
to evaluate the strengths and weaknesses of UNHCR’s shelter programme and to suggest ways to adapt
and improve a programme, which has proven essential for migration-affected populations in
Afghanistan. The stakes of the present evaluation are high as its findings will inform the strategic choices
of UNHCR and, more importantly, may considerably impact the life and opportunities of Afghan
returnees and IDPs.
1.4.1 Rationale of UNHCR’s Shelter Assistance Programme

As part of UNHCR’s initial reintegration assistance, the shelter assistance programme follows a self-build
model, which supports beneficiaries to construct their own accommodations. The programme aims to
have the widest geographic coverage possible with a focus on rural areas where return rates are high,
while additional efforts are made to target areas of possible future return.
Official eligibility for assistance requires that the beneficiary be a returned refugee or IDP, with access to
land on which to build a house. Nevertheless the programme is guided by a focus on vulnerability
allowing beneficiary selection to be wider than just returnees with access to land. In fact, staff members
are advised to ensure that no vulnerable families within the community are overlooked or rejected from
receiving assistance. The vulnerability criterion follows that of the “extremely vulnerable individual14”
definition including people who may be in life threatening situations, unable to help themselves, lacking
family and community support, or suffering from physical or mental trauma. Typically these include
female-headed households, disabled or elderly heads of households without external support, and large
families with insufficient income. Overall, special attention is paid to the relative situation of the
14
Extremely Vulnerable Individuals (EVI) are presently considered Persons with Specific Needs (PSN), however we utilize the EVI
definition throughout corresponding to the period we are evaluating.
22
individual within the family and the community in order to identify vulnerable beneficiaries. Moreover,
in the case of landless families in need of shelter, who meet the vulnerability criteria, there is the
possibility of land allocation in order to allow them to benefit from the programme. In sum, while the
programme explicitly targets vulnerable refugee and IDP returnees, the focus is on finding a shelter
solution for any community member, which meets the vulnerability criteria.
1.4.2 Guiding Principles of SAP

On the whole, the UNHCR shelter programme adheres to the eight following main guiding principles –
which will be tested and analysed throughout this report:
1. Community based approach

The UNHCR shelter programme is a community based, self-help programme. The community
takes primary responsibility for identifying eligible beneficiaries to receive shelter assistance,
while the Ministry of Refugees and Repatriation, authorities, implementing partners and UNHCR
play an advisory role.

Recognizing the challenges of facilitating female participation, UNHCR and implementing
partners involve women in selection, implementation, monitoring and management to the
greatest extent possible within regionally and culturally appropriate contexts.
3. Access to land
Only families with evidence of land ownership will be eligible for shelter assistance. However,
those who had a house on government owned land for a long time may also be eligible, provided
that the land is not disputed and the local authorities issue a no-objection certificate (NOC) for
them to construct a new house. In addition, a family who meets the vulnerability criteria and has
a lease or right to use the land from a landowner may also be eligible for assistance. However,
landless beneficiaries are not included in UNHCR’s shelter programme; they fall under the
responsibility of the Government of Afghanistan.
Beneficiary selection is based on the belief that vulnerable families would not be able to
establish shelters without external assistance. UNHCR recognizes that vulnerability is a relative
phenomenon in one targeted location or village as compared with another location. That is why
the Beneficiary Selection Committee (BSC) is tasked to play an important role in identifying
vulnerable beneficiaries. Extreme vulnerability can be identified during the beneficiary selection
process or during programme implementation. For vulnerable categories such as female-headed,
disabled or elderly heads of households without external support and large families with
insufficient income, all involved staff should ensure that no vulnerable families are overlooked or
rejected for assistance. If all the above efforts fail, as a last resort, an additional cash component
23
($25 for Standards A and B, $50 for Standard C are recommended but flexible) can be allocated
from a sub-office’s respective budget to assist individual cases to build their shelter. This can be
in the form of individual/family grants or through cash for work projects. The extent of the
assistance is at the discretion of each sub-office. Regional staff and BSC members are responsible
to ensure that all beneficiaries, especially the most vulnerable, are able to complete their
shelters. Families who are unable to complete (or who are ineligible for the programme because
they are too poor) should not be excluded, as these are the most vulnerable members of a
community.
Afghanistan’s forest is one of the most destroyed sectors of the environment. When
implementing shelter projects, regional offices should consider this fact and use alternative
materials in lieu of wood or, in cases where wood cannot be avoided, try to ensure that wood
products are either imported or are from sustainably harvested local sources. The UNHCR shelter
package therefore includes iron doors and windows for all shelters throughout Afghanistan.
Similarly, iron roof beams or dome roofs made of brick are promoted wherever possible. The
shelter package also includes one latrine for every family, increasing environmental hygiene in
beneficiary communities.

Recognizing the diversity of climatic conditions and cultural preferences in the design of houses
in each region, the UNHCR shelter programme provides a model design against which the in-kind
(material) and cash contribution are based. For instance, under the UNHCR shelter programme,
the dome type ceiling can be seen in west and north Afghanistan and the flat roof with beams in
central, east, southeast and south Afghanistan. The flexibility of the shelter programme should
allow for these variations.

The UNHCR shelter programme seeks to contribute to reviving local economies through its
implementation wherever possible. This includes using skilled and unskilled labour, and local
procurement of raw materials.
8. Involvement of local authorities

In 2003, MUDH, MORR and MRRD developed a national policy for shelter programme
harmonization with the help of other key players. Throughout the process of shelter
implementation, district officers should be encouraged to be actively involved. This is particularly
important during beneficiary selection, the most sensitive step in the shelter programme. The
BSC must include members of the Community Development Council (CDC) where present or the
provincial, district, or village shura (committee of elders and trustees), local authorities (district
authorities, provincial representatives of MORR), in addition to representatives from the
24
implementing partner (IP) and representatives from UNHCR (where field presence is possible).
Joint monitoring is also important, especially concerning communication channels with
beneficiaries, land disputes, ownership and other related issues.
These guiding principles can therefore be grouped under I) Selection process, II) Socio-economic impact,
and III) Partnership strategies – which will be discussed in Chapters 4, 5 and 6.
1.4.3 Changes in the Programme

UNHCR has adapted and revised the guidelines through periodic review. Revisions in 2008 featured an
expanded floor area, improved quality of shelter materials and enhanced sanitation components.
Revisions in 2009 focused on cost reduction, earthquake mitigation, climate and technology adapted
design and standardization of shelter kit components. Options were offered for beneficiaries to use the
shelter kit for a modified and more “module based” shelter concept that would help to open the door for
a phased implementation approach. This allowed for the possibility to build the shelter for expansion,
e.g. the option to start with a one room module and expand by adding additional modules to a two or
more room shelter, subject to need and availability of kits. The 2010 programme suggested to
complement the module tailored shelter packages with a third package, namely the repair and upgrading
kit. The tailoring gave the option of a more diversified range of shelter packages, which therefore helped
to better respond to the shelter needs of individual vulnerable beneficiaries. In the end, the 2010
strategy aimed to open the door for a more self-help based intervention in order to gradually achieve
better coverage of the needs in all locations of return and settlements. Finally, the 2011 shelter guideline
followed that of the years prior and focused on improved quality of shelter materials for better results.
This evaluation aims at providing further input to improve the strategic orientation of the shelter
programme – with specific recommendations provided in Chapter 8.
1.5 REPORT OUTLINE

- Chapters 1 and 2 introduce the background, objectives and methodology of this evaluation.
- Chapter 3 provides an overview of the physical aspects, support and monitoring of the SAP.
- Chapter 4 focuses on the selection process – reviewing beneficiary socio-economic profiles,
levels of vulnerability, and location.
- Chapter 5 analyses the socio-economic impact of the SAP – on beneficiaries, their communities
and on their access to services.
- Chapter 6 reviews the current state, strengths and weaknesses of SAP’s partnership strategy.
- Chapter 7 builds on the survey’s findings to draw conclusions on SAP’s potential to further
sustainable reintegration among beneficiaries.
- Chapter 8 then concludes and provides a set of key recommendations for UNHCR in its future
strategic orientations regarding SAP and assistance to its target population of concern.
25
2. METHODOLOGY
2.1 QUANTITATIVE DATA COLLECTION AND ANALYSIS
2.1.1 Household Survey

To draw conclusions about the efficiency of the selection process, the performance of the programme
and its socio-economic impact on beneficiary households, a large-scale quantitative survey was
conducted in 15 provinces of Afghanistan. An individual questionnaire of 113 closed questions, which
lasted approximately one hour, was conducted with a total of 4,488 individuals15 who belonged to three
categories:
 2,035 UNHCR Beneficiaries

 1,990 Non-Beneficiaries
 463 Beneficiaries of Shelter programmes other than UNHCR’s in the Eastern region only
(Beneficiaries of UN-Habitat, NRC, IOM, IRC and CHF shelter programmes in Nangarhar).
Despite considerable efforts to find beneficiaries from other programmes in the East this proved very
challenging for various reasons: a) most of them were scattered around the province, b) the short time-
line further impeded efforts to receive support from relevant stakeholders (NRC, IRC and IOM), reach the
areas of implementation and identify beneficiaries and c) the turn-over of shelter teams - for IRC and
IOM for instance.
Moreover, the sample can also be viewed by the migratory status of the household surveyed including:
 2,325 Refugee Returnees

 1,200 Non-refugee Returnees
 415 IDP
 548 No Mobility households
Table 1 provides the overall breakdown both by beneficiary status and migratory status.
15
An additional 60 surveys were completed, however, respondents did not know which organization they received shelter
assistance from. To prevent biasing the results, this group will be excluded from any further analysis in this evaluation.
26
Table 1: Household Survey by Beneficiary Status and Migratory Status
UNHCR Non-UNHCR Non-

Total
Beneficiaries Beneficiaries Beneficiaries
Refugee Returnees 1 355 243 727 2 325

% 66.58 52.48 36.53 51.80
Non-Refugee Returnees 390 134 676 1 200
% 19.16 28.94 33.97 26.74
IDPs 187 9 219 415
% 9.19 1.94 11.01 9.25
No Mobility 103 77 368 548
% 5.06 16.63 18.49 12.21
Total 2 035 463 1 990 4 488
% 100.00 100.00 100.00 100.00
27
Table 2 provides an overview of the composition of the final sample. A more detailed list of each district
sampled is provided in the annex (see Annex 1). In each province, the number of respondents mirrored
the distribution of shelter activities, for a statistically representative survey sampling. Within each
district, the research team adopted a cluster sampling scheme. Primary Sampling Units (PSUs) were
randomly selected as much as possible, based on the lists of locations provided by UNHCR and its
Implementing Partners (IPs). In some cases, a number of constraints (security, remoteness and necessity
to have a minimum number of shelters per location) reduced our ability to randomly selection locations.
In each PSU, both beneficiary and non-beneficiary respondents were randomly selected, when possible.
All of the selected shelter beneficiaries received assistance between 2009 and 2011 – as per the terms of
reference provided by UNHCR for this evaluation.
28
Table 2: Household Survey Sampling by Province
UNHCR Non-UNHCR Non- Total Total

Region Province
Beneficiaries Beneficiaries Beneficiaries Province Region
Kabul 197 2 185 384
Central 571
Parwan 101 1 85 187
Central
Bamyan 32 - 29 61 61
Highland
Laghman 162 - 138 300
East 2 368
Nangarhar 790 455 823 2 068
Balkh 50 - 51 101
Faryab 75 2 97 174
North 595
Jawzjan 118 2 100 220
Sari Pul 56 - 44 100
Kunduz 60 - 60 120
Northeast 190
Takhar 39 - 31 70
Helmand 56 - 52 108
South 263
Kandahar 75 1 79 155
Southeast Paktya 123 - 117 240 240
West Hirat 101 - 99 200 200
Total 2 035 463 1 990 4 448 4 488
Table 3 on provides an overview of the final sample by district type, with the majority of respondents,
62.7 per cent, residing in rural areas.
Table 3: Household Survey Sampling by Type of Location

Total
Urban 334 87 413 834

% 16.41 18.79 20.76 18.59
Semi-rural 424 5 411 840
% 20.84 1.08 20.66 18.72
Rural 1 277 371 1 165 2 813
% 62.75 80.13 58.57 62.69
Total 2 035 463 1 989 4 487
% 100.00 100.00 100.00 100.00
29
2.1.2 Community Survey
The household survey was complemented by a community survey conducted in each PSU. The
community survey aimed at collecting a mix of quantitative and qualitative data about the profile of the
community and about the modalities of the shelter programme and the consequences of its
implementation on the community.
A total of 60 community surveys were compiled as part of a second dataset. The provincial distribution
of these communities is shown in
Table 4. The team conducted this survey with the malik or the head of the shura or of the CDC in the
village16. In case either of these leaders was absent during the visit, the team interviewed their deputies
or other informed authorities in the village.
Table 4: Community Survey Sampling
Region Province N Total Region
Kabul 5
Central 6
Parwan 1
Central Highland Bamyan 4 4
Laghman 5
East 24
Nangarhar 19
Balkh 2
Faryab 3
North 11
Jawzjan 5
Sari Pul 1
Kunduz 2
Northeast 3
Takhar 1
Helmand 2
South 4
Kandahar 2
Southeast Paktya 4 4
West Hirat 4 4
Total 60 60
16
A malik is ‘the individual who represents community interests to formal government institutions. He is the village executive’,
while shuras are the traditional deliberative councils as defined by Brick (2008), the political economy of customary village
organizations in rural Afghanistan. Community Development Councils (CDCs) have been introduced by the Ministry of Rural
Rehabilitation and Development as the main deliberative – and elected – council through which the funds of the National
Solidarity Programme are channeled to the local communities.
30
2.1.3 Multi-Dimensional Poverty Analysis
This research aims at assessing 1) the socio-economic profiles of beneficiary households and their
communities and 2) the reintegration outcome of the shelter programme, i.e. assessing the level of
parity between returnees, IDPs and no-mobility households in SAP communities. Methodologically, this
requires an indicator able to compare different dimensions of wellbeing upon which to rate the poverty
of a household. The selected tool – a Multi-Dimensional Poverty Index (MPI) – reflects deprivations in
different dimensions that have an impact on the poverty of a household.
For the purpose of this study, the multi-dimensional poverty analysis allows us to conduct a more
comprehensive assessment of how deprivation relates to our sample than would be the case if using a
single monetary indicator like income. The multi-dimensional poverty index (MPI) is based on the idea
that the well-being of a person or a household is not only dependent on income or consumption, but
also on multiple other dimensions like health, education, security and standard of living. Combining all
the dimensions leads to the overall identification of poor households in the multi-dimensional sense. Our
approach follows that which was pioneered in UNDPs widely-recognized Human Poverty Index (HPI)
within their Human Development Reports (HDRs), and has since been developed further in recent years
by such authors like Alkire and Santos (2010)17 and Alkire and Foster (2007)18.
Methodologically we follow a step-by-step process, first analysing household deprivation by individual

indicators before scaling to the dimensional level, and concluding with an overall multi-dimensional
poverty rate. The first step in constructing the MPI is to assess household deprivation along individual
indicators within pre-defined dimensions. We therefore identify a range of relevant indicators with
specific thresholds in which an Afghan household can be considered deprived or not. While selection of
indicators may be criticized as arbitrary, identification was made following an exhaustive review of
related literature while also taking into account the contextual environment in question as well as data
at hand. In particular, conversations with our in-country research team allowed for a greater
understanding of which indicators and thresholds were appropriate.
The next step involves calculating poverty at the dimensional level. Here we apply a 30 per cent cut-off,
meaning a household deprived in nearly a third of the individual indicators, weighted equally, within that
dimension is characterized as dimensionally poor. The formal expression is:
1
=
=1 >
17
ALKIRE Sabina, SANTOS Maria Emma, “The Multidimensional Poverty Index: Contradictions and Analysis”, October 2011.
18
ALKIRE Sabina, FOSTER James, “Counting and Multidimensional Poverty Measurement”, OPHI Working Paper No.7, University
of Oxford, 2007.
31
where n represents the number of households; is the binary variable for dimensional deprivation for
house on dimension d, taking a value of 1 if the aggregated and weighted indicators, , is greater
than the cut-off, . As noted, each indicator within a dimension is weighted equally and sums up to 1.
Finally, we are able to repeat the exercise at the overall multi-dimensional level again using the cut-off of
30 per cent. While the procedure is the same, one notable difference is that dimensions are weighted
equally causing individual indicators to have relative weights depending on the number of indicators
making up each particular dimension. All told, a household deprived in 30 per cent of the individual
indicators with varying relative weights across dimensions is characterized as multi-dimensionally poor.
Formally:
1
=
=1 >
where n represents the number of households; is a binary variable for overall deprivation taking a
value of 1 if the aggregated and weighted dimensions, , is greater than the threshold, . As stated
prior, each dimension is weighted equally and sums up to 1 while each indicator is given a relative
weight. Table 5 provides an overview of both indicators as well as dimensional and relative weights.
The four dimensions used in our analysis include:
Dimension 1: Economic
Dimension 2: Education
Dimension 3: Health and Nutrition
Dimension 4: Housing
Table 5 presents the individual indicators of deprivation within each dimension, a description of the
thresholds used, as well as the dimensional and multi-dimensional weights applied for construction of
the dimensional and multi-dimensional indices. Moreover, the level of deprivation along each individual
indicator is shown in the last column as well as the dimensional and multi-dimensional poverty indices
for our entire sample.
32
Table 5: Multi-Dimensional Poverty Index
Dimensional MPI %
Dimension Variable Household is deprived if…
Weight Weight Deprived
Expenditure per capita is below

Expenditure per
the $1.25/day, $38.02/month, 20.00% 4.00% 22.00%
capita
$456.25/year poverty line
Household has less than 2 sources

Number of
of income (only working age 20.00% 4.00% 77.00%
income sources
adults)
Dimension 1:
Economic Household has at least one child
Child labour 20.00% 4.00% 4.00%
working
Household Household indebtedness is in top
20.00% 4.00% 15.00%
Indebtedness 20% of sample
Ratio of unemployed household
Dependency
members to employed household 20.00% 4.00% 53.00%
ratio
is below the sample mean (6.52)
Dimension 1 100.00% 20.00% 64.00%
Literacy Household respondent is illiterate 50.00% 10.00% 78.00%

Dimension 2: School At least one child does not attend
Education 50.00% 10.00% 50.00%
attendance school
Dimension 2 100.00% 20.00% 87.00%
Household does not have access
Access to health 14.29% 2.86% 13.00%
to a health facility
Household cannot satisfy food
Food security needs "sometimes" (3-6 times a 14.29% 2.86% 36.00%
week)
Household expenditure per capita
Food
on food is below 690 AFS, 14.29% 2.86% 52.00%
expenditure
monthly
Dimension 3: Household eats meat less than
Health & Food variety the sample median (1 time a 14.29% 2.86% 44.00%
Nutrition week)
Household reports a member who
Illness/ Disability 14.29% 2.86% 35.00%
is ill, disabled, or a drug addict
Immunizations Children are not immunized 14.29% 2.86% 2.00%
At least one child has passed

Child mortality 14.29% 2.86% 15.00%
away due to health reasons
Dimension 3 100.00% 20.00% 33.00%

Household lives with relatives,
Dimension 4: Access to
friends, or a temporary shelter 9.09% 1.82% 8.00%
Housing housing
(tent, shack, etc.)
33
Subjective
Quality of housing is worse than
relative quality of 9.09% 1.82% 35.00%
other households
housing
Electricity Household has no electricity 9.09% 1.82% 48.00%
Household has no access to safe

Drinking water 9.09% 1.82% 7.00%
drinking water
Household has no toilet or uses
Sanitation open field, bush, or area in the 9.09% 1.82% 8.00%
compound which is not a pit
Heating Household has no heating 9.09% 1.82% 32.00%
Household has a floor which is

Flooring 9.09% 1.82% 0.00%
dirt, sand or dung
Household own less than 2 assets
Asset ownership (radio, TV, telephone, bicycle, 9.09% 1.82% 31.00%
motorcycle, car or refrigerator)
Land Household owns no land 9.09% 1.82% 57.00%
Livestock Household owns no livestock 9.09% 1.82% 81.00%
Subjective
Household economic well-being is
economic well- 9.09% 1.82% 24.00%
worse than other households
being
Dimension 4 100.00% 20.00% 43.00%
Household does not own a mobile
Mobile phone 25.00% 5.00% 20.00%
phone
Household has no membership in
Membership 25.00% 5.00% 91.00%
a community organization
Dimension 5:
Social Capital Household has not received
Help network 25.00% 5.00% 41.00%
& Inclusion assistance since living in the area
Subjective
Household does not feel secure 25.00% 5.00% 4.00%
security
Dimension 5 100.00% 20.00% 50.00%
MPI - 100.00% 78.00%
Following the MPI construction, we are able to compare groups based on this index. First, we provide a
simple mean comparison suggestive of differences among categories, before a more complete cross-
sectional regression analysis. The regression analysis uses a probit model in order to estimate the
predicted probability of a household being multi-dimensionally deprived. Formally:
P (MPI i=1|Xi) = iXi
34
where MPI i indicates the binary dependent variable of household i taking the value of 1 if the MPI
analysis characterizes the household as multi-dimensionally deprived, and 0 otherwise; Xi is the binary
independent variable indicating treatment based on which category the households falls under; I
represents the regression parameter to be estimated; and  indicates the cumulative normal distribution
function. Moreover, a set of control variables are used including which province the household lives in,
whether the location is urban, semi-rural or rural, the size of the household, whether a household is
identified as an EVI, whether a household member is a current migrant, and whether the household
received remittances from abroad.
While the cross-sectional regression analysis gives us evidence of how groups differ, we are not able to
say whether this difference is due to the shelter assistance program or not. In order to estimate the
impact of the shelter assistance program we must go one step further, and perform a difference-in-
difference (DiD) analysis. Again we utilize a probit model yet look at differences over time, allowing us to
conclude how UNHCR-beneficiaries compare to a non-beneficiaries or non-UNHCR beneficiaries because
of the program. The formal expression of our probit DiD model is:
P (D i,t=1|Xi) = ∝ + βXi + γTt + δXi ∗ Tt + εi,t
where D i,t is the deprivation for household i in period t; Xi is the binary independent variable indicating
treatment taking a value of 1 if the household is a UNHCR Beneficiary, and 0 otherwise; Tt is the binary
variable indicating time taking a value of 1 if the time period is when the respondent was surveyed, and
0 otherwise; and Xi ∗ Tt is the interaction term representing actual treatment. Moreover ∝, β, γ, Pt and δ
are the regression parameters to be estimated while  is the cumulative standard normal distribution
function and εi,t represents the error term.
2.2 QUALITATIVE DATA COLLECTION AND ANALYSIS

In addition to the quantitative data, qualitative data was also collected for the purpose of the
programme evaluation. This was necessary to get a richer picture of the programme, its conception and
implementation.
2.2.1 Secondary Data

A thorough desk review of existing literature on the issues of return migration, internal displacement,
shelter as well as broader related issues and conceptual humanitarian debates was conducted. The
secondary research allowed for:
 A detailed overview of the different components and evolutions in the shelter assistance
programme through a large review of project documentation since 2009, including the shelter
guidelines and package details, aggregated data on shelter and lists of beneficiaries, in addition
to general UNHCR policy documents.
35
 A thorough comprehension of the trends and dynamics behind the concepts of return migration
and internal displacement, critical in understanding the issues at stake in the shelter programme.
This was done both at the international level and in the Afghan context. Special attention was
given to concepts of return and repatriation, reintegration, vulnerability and shelter.
 A review of existing literature on shelter and return migration, including past evaluations of
shelter programmes, so as to identify past and present issues and lessons learned.
 Placement of the programme in broader policy and humanitarian debates, such as access and
remote monitoring, partnership strategies and cash vs. non-cash assistance. This allowed us to
compare and assess their relevance in the Afghan context and identify what dilemmas and
strategic choices are appropriate and relevant for the programme.
2.2.2 Key Stakeholder Interviews

The research team conducted a total of 79 key informant interviews (KIIs) at the national and at the
sub-national level. These interviews aimed at:
 Grasping the practical modalities of implementation of the programme.

 Evaluating the coordination mechanisms in place for shelter assistance.
 Assessing the quality of the partnership between UNHCR and national authorities.
 Getting the perspective of other stakeholders on the programme.

 Comparing the various shelter programmes in place in the country.
At the provincial level, these KIIs were conducted in provinces directly visited by international research
staff - namely Kabul, Parwan, Kandahar, Nangarhar, Faryab, Jawzjan, Balkh and Hirat. A full list of the KIIs
completed in Kabul and in the provinces is provided in Annex 2.
The following categories of key stakeholders were covered by these interviews:
 UNHCR staff
o At the central and field level
o Previous UNHCR Afghanistan staff involved in the 2009 – 2011 SAP including:
 Management
 Protection officers
 Shelter programme officer
 Other UN agencies
 Donors
 Governmental authorities
 International NGOs
 National NGOs / Implementing Partners (IPs) working on shelter assistance in Afghanistan
36
Table 6: Breakdown of KIIs per Province and Type of Respondent
Other UN NNGOs/
Location UNHCR Donors GoA INGOs Total
agencies IPs
Central 7 4 4 3 1 3 22
East 3 3 0 5 3 3 17
South 3 0 0 2 1 1 7
North 3 4 0 5 3 7 22
West 4 1 0 0 3 3 11
Total 20 12 4 15 11 17 79
2.2.3 Focus Group Discussions

In order to grasp more personal and substantiated opinions about the shelter programme, 58 focus
group discussions (FGD) were conducted. These focus groups were based on semi-directive focus group
guides designed to foster the discussions and debates on a series of themes central to the programme
and its evaluation.
The following are some of the discussed themes:
 The effective modalities of the selection process

 The shelter package and material
 Participation of women in the programme

 Identification of potential issues relative to the implementation
 Impact on the household
 Impact on the community
 Perception of UNHCR and its partners
FGDs were conducted with a) UNHCR beneficiaries, b) non-beneficiary returnees, c) non-beneficiary non-
migrants and d) beneficiaries from other shelter programmes in the East. In order to guarantee a
representation of women’s opinions about the programme, the field team was asked to conduct
separate FGDs with women and with men. Yet, because of the difficulties in accessing women in certain
provinces (for example in Kandahar, Helmand, Faryab and Kunduz) and due to the differences in the level
of awareness of respondents, a large majority of FGDs were conducted with men. The following Table 7
shows the composition of the 58 focus group discussion.
37
Table 7: Composition of Focus Groups
UNHCR Non-UNHCR Non-Beneficiary Non-Beneficiary

Gender Total
Beneficiaries Beneficiaries Migrants Non-Migrants
Female 3 0 2 0 5
Male 19 5 22 7 53
Total 21 5 24 7 58
2.2.4 Field Observations

A qualitative field report was provided for each PSU visited by the research team – a collection of
provincial overviews is provided in Annex 3. These qualitative reports provided information about the
specific context and the particularities of each surveyed location.
The field reports were implemented as a way to go deeper into the context, the modalities of
implementation and into the analysis of the factors entering into play to explain the success or the
failure of the programme in each sampled area. For this report they are used to contextualise the
analysis of quantitative findings and provide UNHCR with a localized analysis.
2.3 LIMITATIONS AND CONSTRAINTS

Given the limitations imposed by security and other constraints on the sampling, a purely random
sampling methodology could not be used.
The main constraints encountered in the field included:
 Security: Given the size of the teams (10 to 20 interviewers each) and the time necessary to
conduct the survey in each location, the teams were very visible in the field and were therefore
asked to take precautions. This impacted the sampling especially in Faryab, Nangarhar, Jawzjan,
Kandahar, Helmand and Laghman provinces. In these provinces in particular, the team had to
either substitute the initial district selected to a safer one or to cover two or three districts
instead of one, so as to limit the risks.
 Geographical repartition of shelters: Villages with too small number of shelters had to be
excluded from the sampling to guarantee that the teams would meet their targets. Locations
with 20 and more shelters were privileged, restricting the randomness of the sampling.
 Selection of respondents: As much as possible, the teams relied on lists of beneficiaries and a
snowball method to find respondents but given the cultural context of the country, field teams
sometimes had to go through the community leader and/or the implementing partner to select
respondents. In rural areas, it is almost impossible and sometimes even dangerous, to enter a
38
village without the full endorsement of the community leaders. The mediation of community
leaders might have introduced a bias selection of respondents in some cases. In particular, this
might have reduced a bit the presence of respondents who were not the initial beneficiaries in
the selection, as community leaders sometimes feared that it would decrease their chances of
getting shelter assistance in the future. Yet, this bias is limited as in rural areas, the survey team
was often able to survey most or all of the beneficiaries listed by UNHCR, while in urban and
semi-urban settings, the team was not forced to rely as much on community leaders for their
sampling. In Southern and Eastern regions (Kandahar, Helmand and Nangarhar), the IPs
sometimes joined the field team while they conducted the survey, which might have introduced
some biases either in the selection of respondents or in the interviews with the community
leaders, even though those were not conducted in their presence.
 Awareness of respondents: In a lot of cases, men and heads of households were working while
our teams conducted the survey. Interviewers conducted the survey with the most informed
adult available in each household. Female interviewers conducted their interviews with female
members of the household. This should not have a major impact on the results of the survey.
Yet, it could have an impact on the quantitative data, as the level of awareness of respondents
could be lower than the one of the head of household. Women respondents in particular
sometimes found it challenging to answer questions about income, expenses or the construction
of their shelter. This is not an issue specific to this particular study but is a general constraint
when conducting survey in Afghanistan.
 Beneficiaries from other shelter programmes: It proved more difficult than expected to survey
beneficiaries from other programmes in the Eastern regions, mostly because, contrary to UNHCR
beneficiaries, these respondents were often scattered around urban areas or numerous villages.
39
3. SAP: DESIGN, SUPPORT AND MONITORING & EVALUATION
UNHCR’s SAP is a community-based, self-help programme whereby households build homes for
themselves. UNHCR supports them by providing a shelter package that includes essential construction
materials (tools, roofing beams, doors and windows), and by supervising in order to achieve minimum
standards of quality in accordance with the Sphere Standards. The first step to evaluate this assistance is
by looking at the design and physical aspects of the shelters, support towards and monitoring of the
construction. Key findings from this section include:
1. Design and physical aspects of the shelter
 High level of completion of shelters. Yet the state of shelters varied significantly and depended
on household economic profile and the level of investment they could dedicate to their shelter.
 High level of satisfaction of beneficiaries with the shelter package, with good quality material
that most beneficiaries would not have been able to afford without the assistance of UNHCR.
The distribution process works efficiently for 93 per cent of beneficiaries.
 Main complaints raised:
o Limited size of the shelter given the large size of beneficiary households
o Low quality of latrines, and insufficient technical assistance
o The quality of doors and windows was too low to be sustainable
 A difficult and costly construction process for beneficiaries, as:
o 972 of the beneficiary households (48%) ran into problems during construction.
o 89 per cent of the households with problems ran out of money during construction (this
corresponds to 42% of all UNHCR beneficiaries) with
o Significant disparities in household contribution according to provinces/location.
o 47 per cent of households that ran into problems (22 per cent of the beneficiaries)
reported a lack of sufficient access to water to build shelters and rely on costly solutions.
 UNHCR procedures for cash distribution are robust enough to avoid misallocation. Yet the
most vulnerable households use the cash for more immediate purposes than the construction of
shelters, e.g. prioritizing food over the purchase of glass panes.
 Risk-mitigation measures are not properly integrated in the implementation of the shelter
program, limiting sustainability of the SAP.
2. Support and additional assistance

 Inefficient mechanisms to provide additional assistance to EVIs to build their shelter
 Limited complementary training: 54 per cent of beneficiaries receive complementary training on

construction, while less than 20 per cent of them receive hygiene promotion training.
3. Monitoring and evaluation

 Overall, IPs ensured a satisfactory technical monitoring through regular field visits, yet
monitoring procedures do not ensure that the most vulnerable are targeted.
40
3.1 DESIGN AND PHYSICAL ASPECTS OF THE SHELTER
3.1.1 Shelter Design

The main type of shelter implemented across provinces was the standard two-room shelter, including a
corridor and latrine. Annual variations in the design and material provided were introduced nationwide
according to field observations and recommendations from IPs:
 The size of the rooms was progressively enlarged from 2009 to 2011.
 Wooden beams, reportedly subject to termite attacks, were replaced by iron beams in 2010 and
fire bricks were introduced in the roofing components.
 The three small windows in the 2009 shelters were replaced by two larger windows as of 2010,
following complaints about lack of light and ventilation.
Interiors of shelters: Wooden beams (Faryab Province; Andkhoy District); Iron beams (Jawzjan Province, Sheberghan
district).
In addition, evolutions in the programme have included the addition of one-room interventions, which
started out as a tool used by UNHCR in emergencies to support local communities to absorb displaced
persons by building families an extra room. This allowed the organization to increase its responsiveness
in the face of emergencies and allow for more flexibility.
Among the interviewed UNHCR beneficiaries the majority, 81.7 per cent, built two-room shelters, 17.9
per cent built a one-room shelter while less than 1 per cent built a completely different type of shelter.
41
The sizes of shelters built by beneficiaries of other programmes are similarly distributed, 82 per cent are
two rooms and 18 per cent are one room.
In most cases, beneficiaries did not have a say in the choice of the model of shelters that they would
build, as this was instead decided by UNHCR. Only 13 per cent of households that received shelter
assistance from UNHCR choose themselves.
As per the guidelines, there were differences in the standards across regions, with dome shaped roofs in
the West and flat roofs in the Central, Southern and Eastern regions. In the West and East, as well as in
the South, the shelter programme also comprised a more systematic implementation of “one-room
shelters for IDPs” (including a corridor and latrines), and “repair-kits” composed of one additional room
for to an existing house, reportedly as a mean to adapt to the wide variety of profiles of beneficiaries.
Breaking down the type of shelter by location, Table 8 highlights the higher uptake of one-room shelters
in urban areas, compared to semi-rural or rural areas that have the lowest proportion of one-room
shelters. This further underlines the need for flexibility of models in urban areas. One-room shelters can
be used as a tool to absorb displaced persons in their new environments by building families an extra
room. This allows upgrading or expanding of shelters that already house displaced family members, who
opt for living with host families. It also fits more realistically with the more limited space available in
urban areas compared to returnee townships or rural areas. Flexibility in shelter models is therefore an
asset for beneficiaries depending on their location.
Table 8: Type of Shelter by Location (UNHCR Beneficiaries)
Urban Semi-rural Rural Total
One room 121 98 144 363

% 36.34 23.11 11.29 17.86
Two room 212 323 1 126 1 661
% 63.66 76.18 88.31 81.74
Other 0 3 5 8
% 0.00 0.71 0.39 0.39
Total 333 424 1 275 2 032
% 100.00 100.00 100.00 100.00
Table 9 shows an unequal spread of one-room shelters due to decisions made at the regional office level.
In Bamyan (68.8%) and Helmand (87.5%), the majority of UNHCR beneficiaries was given one-room
shelters. In Nangarhar, one in four households was given the one-room option, above the sample
average. On the other hand, provinces such as Sari Pul, Kandahar, Takhar, Jawzjan, Kabul and Parwan
had less than 10 per cent of one-room shelters. One-room shelters were mostly used to provide shelters
for IDPs in an effort to quickly address the needs of IDPs without antagonizing governmental authorities.
This was particularly the case in Nangarhar where UNHCR and NRC used one-room shelters to provide
assistance to IDPs despite the strong reluctance of provincial authorities and in Helmand, where the
UNHCR sub-office was able to adapt to the high movements of intra-provincial displacements.
42
UNHCR field staff and IPs do not always support the option of one-room shelters as the implementation
is more complex when different models of shelters co-exist, and commonly goes against the will of
beneficiaries who ask for bigger shelters. Still, specific attention should be paid to the added value of
one-room shelters in urban contexts, in emergency contexts and their relevance to specific regions and
provinces. Although there is a more systematic use of one-room shelters in certain provinces of the
Central, East and South regions, these remain an exception and lessons learned should be shared to
analyse the adaptability to other provinces as well.
Table 9: Type of Shelter by Province (UNHCR Beneficiaries)

One room Two room Other Total
Kabul 13 178 6 197

% 6.60 90.36 3.05 100.00
Parwan 9 92 0 101
% 8.91 91.09 0.00 100.00
Bamyan 22 10 0 32
% 68.75 31.25 0.00 100.00
Laghman 19 142 0 161
% 11.80 88.20 0.00 100.00
Nangarhar 202 586 1 789
% 25.60 74.27 0.13 100.00
Balkh 5 45 0 50
% 10.00 90.00 0.00 100.00
Faryab 10 65 0 75
% 13.33 86.67 0.00 100.00
Jawzjan 1 117 0 118
% 0.85 99.15 0.00 100.00
Sari Pul 1 55 0 56
% 1.79 98.21 0.00 100.00
Kunduz 7 53 0 60
% 11.67 88.33 0.00 100.00
Takhar 0 38 0 38
% 0.00 100.00 0.00 100.00
Helmand 49 7 0 56
% 87.50 12.50 0.00 100.00
Kandahar 2 72 1 75
% 2.67 96.00 1.33 100.00
Paktia 7 116 0 123
% 5.69 94.31 0.00 100.00
Hirat 16 85 0 101
% 15.84 84.16 0.00 100.00
Total 363 1 661 8 2 032
% 17.86 81.74 0.39 100.00
43
3.1.2 State of Shelters
Shelters were found in a good state, being completed according to the guidelines and presenting no
major external signs of degradation. Many shelters had been improved by adding cooking spaces and
terraces as well as decorated and furnished rooms. Observations showed that the general state of the
shelters and latrines could vary greatly and was related to several factors:
i. The economic situation of beneficiaries
In cases where very vulnerable households were unable to provide higher investments to maintain the
general state of their habitat or to upgrade their shelters, there were observations of degradation of the
building. Vulnerable households were also more likely to have used the cash given by UNHCR for other –
more urgent - purposes than the construction of their shelters, including food and water. The shelters of
these households would generally be in a poorer state and lack glass windows, for example. In Kandahar
province, and to a smaller extent in Nangarhar and Parwan, some shelters were poorly constructed, with
walls of low quality and glass panes replaced by plastic sheets. The availability of appropriate material for
building and upgrading shelters and the availability of sources of income was also a factor determining
the capacity of the households to engage in further investments.
Disparities were observed related to the wellbeing of beneficiary households. The absence of window
panes and poor construction of the walls emphasize the fact that the initial economic situation of
beneficiaries has a substantial impact on their ability to build, maintain and rearrange their shelter.
Yet, it is important to stress that only in few cases, shelters were not completed (particularly in Qala-e
Nasro), due to the incapacity of the beneficiary to finish building the shelter and earn a living at the
same time. As detailed further below in Table 31, only 2 out of the 2,034 households surveyed had not
completed their shelter. Yet, it must be noted that this does not give a representative picture of the level
of completion of the programme as incomplete shelters would more likely be empty and are therefore
underrepresented in the survey. Qualitative observations reported a higher number of incomplete
shelters. Interestingly, the two households that reportedly quit the programmes were refugee returnees
who received assistance in 2010 and 2011 in the provinces of Nangarhar (Bahesod, Akhonzada) and
Laghman (Markaz Mehtarlam). They were located in rural and semi-rural locations. The fact that neither
of these households were reported living in a remote location should therefore allow follow-up on their
cases, and similar cases, to find out the reasons for their dissatisfaction or inability to cope with the
programme.
One of the 2010 shelters visited was not completed and missed all material provided in the shelter
package, which were lying in a neighbouring ground. The explanation given by the wife of the
beneficiary and confirmed by neighbours was the beneficiary had gone to Kabul to find daily work,
that he couldn’t come back for construction and couldn’t afford additional investment to complete the
shelter. 10 people were living in the two unfinished rooms. – Qala e Nasro, Paghman Province.
‘One of the 2011 shelter visited had no windows, no doors and the walls were unfinished. The
beneficiary reported that he could not afford completing the construction of his shelter. The
beneficiary family lived with relatives in the village. – Shobash Khorde Turkmenia, Jawzjan Province
44
Sometimes, exact replicas of UNHCR shelters had started to be built by non-beneficiaries, expecting to
receive further assistance through the shelter package, as for example in Aab Dara in Paghman. This
notably underlines the fact that despite complaints about the size of the rooms and quality of doors and
windows, the current design of shelters was considered as appropriate in meeting immediate needs of
beneficiaries and the population at large.
Appropriation of the shelter and its surrounding environment denoted a clear intention to stay, even in
cases where threats were placed on the sustainability of the settlement due to insufficient infrastructure
and lack of income opportunities. This was notably the case in homogeneous tribal environments, where
related families were grouped on the same compound, inside surrounding walls, according to a
traditional disposition of habitat around a common courtyard (in Parwan, Kabul and Hirat for instance),
allowing sharing of common living facilities, such as a tanur for cooking.
Surrounding walls are notably a major requirement: in cases where they could not be constructed,
especially in heterogeneous environments where neighbours were not related, absence of privacy and
security could lead to abandonment of shelters. This was for example the case in Pitawa (Qarabagh
district – Kabul Province) where the field team observed two shelters that lacked surrounding walls.
Beneficiary households preferred living with relatives and had left the shelters unoccupied.
ii. The main usage of the shelter: living space or storage?
In multiple cases in Jawzjan and Parwan, and occasionally in Nangarhar, shelters were not used as living
space per se, but had rather been turned into storage rooms, secondary or guesthouses and occasionally
shops. As beneficiary households had concentrated their investments on their main living space, the
general state of shelters used as storage space was relatively poor. Often they were missing doors and
windows, which had been used for other purposes on the premises where the family lived. In cases
where they were used as secondary or guesthouses, conversely, further investment had been made and
they were considered a source of pride. “Misuse” of shelters is disquieting as it stresses flaws in the
selection process: in such cases, shelters were not an immediate and essential need for beneficiary
households, putting into question the cost effectiveness of the programme.
Similar conclusions were drawn from observations of the use of latrines. The state and use of latrines
was highly related to the implementation of WASH programmes19. In cases where they were inexistent
(Kabul district aside from the reintegration site of Kuchi Abad, Khanjar Khil in Parwan), latrines were not
used, often constructed outside walled compounds or used for other purposes than hygiene, such as
storage. Conversely, in Nangarhar where WASH components had been implemented conjointly with the
shelter programme, latrines were not only used, but had been replicated and adopted by other members
of the community, emphasizing the importance of complementary programmes and awareness about
hygiene as an important component of sustainable reintegration. The importance of complementary
assistance and training will be discussed further below in section 3.2on Support.
19
The WASH programmes aim at saving lives and reducing illness through global access to safe water, adequate sanitation and
improved hygiene. The WASH programme’s long-term prevention and control measures reduce the severe impact of WASH-
related diseases by improving health, reducing poverty and increasing economic development.
45
3.1.3 Appropriateness of Shelter Design
Overall, beneficiaries were relatively satisfied with the shelter package they received, as they would not
have been able to purchase most of the materials provided themselves. This was notably the case with
I-beams, T-beams and ceiling bricks (since 2010), which are unavailable on the local market and/or are
unaffordable for beneficiaries. This suggests that the shelter package does answer the needs of
beneficiaries quite accurately, although complaints raised by respondents and community leaders are
important indicators to take into account for improvements to the shelter programme.
i. Main complaints of the shelter package
As illustrated in
Table 10, the top 3 complaints raised – by all shelter beneficiaries, UNHCR and non-UNHCR alike – are:
 The quality of technical assistance
 The quality of latrines

 The size of the shelter
The importance of support and additional assistance will be discussed in the section on support. The
analysis here focuses on the size of shelters – a recurrent complaint during the survey.
Table 10: Dissatisfaction with Elements of the Shelter Programme
UNHCR Non-UNHCR
Beneficiaries Beneficiaries
N % N %
Quality of technical assistance 285 14.01 104 22.46
Quality of latrine 238 11.70 84 18.15
Size of the shelter 233 11.45 27 5.83
Thermal isolation 177 8.70 53 11.45
Quality of door 146 7.18 36 7.78
Quality of lintels 134 6.59 50 10.80
Quality of windows 128 6.29 27 5.83
Design of the shelter 94 4.62 15 3.24
Quality of roof 79 3.88 45 9.72
46
A recurrent complaint of beneficiaries about the design of shelters was the size of the rooms,
repeatedly mentioned across all provinces, with the exception of Jawzjan, maybe due to the
traditionally smaller size of households in the North (see
Table 11). The difference is the average household size across province is significant with for example
Jawzjan counting on average 6.65 members per households as against Helmand where the average size
of households is above 10 members. These types of provincial differences could be better integrated in
future programming.
The level of dissatisfaction about the size of shelters was higher among UNHCR beneficiaries (11.5%)
than among the beneficiaries of other programmes (5.8%). The opposite is true for the quality of latrines,
where 18.2 per cent of other programme beneficiaries were not satisfied and 11.7 per cent of UNHCR
beneficiaries. For both groups the highest level of dissatisfaction concerns the quality of technical
assistance. 14.0 per cent of UNHCR beneficiaries were not satisfied in this aspect and even more than 22
per cent of the beneficiaries of other programmes (see section 3.2.1 for more on this issue).
Table 11: Average Household Size by Province
Region Province N Mean Min Max
Kabul 384 8.45 1 37

Central
Parwan 187 7.56 2 19
Central Highland Bamyan 61 7.52 1 23
Laghman 300 8.89 1 25

East
Nangarhar 2 067 10.10 1 55
Balkh 101 6.55 2 17
Faryab 174 7.57 2 30
North
Jawzjan 220 6.65 1 16
Sari Pul 100 6.73 2 22
Kunduz 120 6.78 2 16
Northeast
Takhar 70 6.41 2 15
Helmand 108 10.41 2 31
South
Kandahar 155 9.72 2 41
Southeast Paktya 240 10.75 2 63
West Hirat 200 6.43 1 19
Total 4 487 9.04 1 63
Urban dissatisfactions
47
Urban beneficiaries were more critical of the quality of technical assistance and the quality of latrines
provided by the shelter programme (Table 12). Their dissatisfaction ranked twice as high as their rural
counterparts, and three times that of their semi-rural counterparts.
As such, the data underlines a clear expectations gap between what the shelter programme offers and
urban household needs. There is an added pressure in urban areas to have adequate housing – in terms
of quality but also in terms of appearance, to blend in more effectively within the urban landscape. The
UNHCR shelter model was seen as being too rudimentary for urban households. The latrines provided
proved ill-adapted and will be considered in the recommendations section. An added focus will be
needed in future shelter strategies on the ways the SAP can be adapted to an urban context that is
increasingly home to internal displacement and refugee return.
Table 12: Dissatisfaction with Elements of the Shelter Programme by Location

(UNHCR Beneficiaries)

(N=334) (N=424) (N=1277) (N=2035)
Quality of technical assistance 84 40 161 285

% 25.15 9.43 12.61 14.00
Quality of latrine 69 27 142 238
% 20.66 6.37 11.12 11.70
Size of the shelter 38 73 122 233
% 11.38 17.22 9.55 11.45
Thermal isolation 54 25 98 177
% 16.17 5.90 7.67 8.70
Quality of door 34 35 77 146
% 10.18 8.25 6.03 7.17
Quality of lintels 46 21 67 134
% 13.77 4.95 5.25 6.58
Quality of windows 25 36 67 128
% 7.49 8.49 5.25 6.29
Design of the shelter 21 30 43 94
% 6.29 7.08 3.37 4.62
Quality of roof 14 12 53 79
% 4.19 2.83 4.15 3.88
Beyond the urban specificities, semi-rural households also raised concerns – above that of their
counterparts – on the size of the shelter, the quality of windows and the design of the shelter.
The data does not present any specific particularities for remote locations that did not indicate more or
less satisfaction than non-remote areas on issues of the quality of the equipment or the provision of
technical assistance.
48
Regional dissatisfactions
The main complaints raised differ across regions (Table 13). Respondents in the Central Highland and the
Eastern regions mainly raised the quality of technical assistance as a key issue. However, the quality of
latrines posed a problem mainly in the Eastern and Southern regions, which could indicate a certain
cultural inadequacy of the latrine models in Pashtun communities. Lastly, the size of shelter was an
obstacle more evenly shared by regions, with the Western region ranking highest, with almost one in five
households interviewed dissatisfied with the size of the shelter. Qualitative observations also confirmed
that it was a concern in the South and the East. This issue was the least problematic in the Central and
Central Highland regions.
Table 13: Dissatisfaction with Elements of the Shelter Programme by Region

Central North- South-

Central East North South West Total
Highland east east
(N=298) (N=952) (N=299) (N=131) (N=101) (N=2035)
(N=32) (N=99) (N=123)
Quality of
technical 16 10 233 1 0 13 11 1 285
assistance 5.37 31.25 24.47 0.33 0.00 9.92 8.94 0.99 14.00
%
Quality of
20 3 181 2 2 23 4 3 238
latrine
6.71 9.38 19.01 0.67 2.02 17.56 3.25 2.97 11.70
%
Size of the
30 3 79 45 16 19 21 20 233
shelter
10.07 9.38 8.30 15.05 16.16 14.50 17.07 19.80 11.45
%
Thermal
9 3 145 2 2 8 7 1 177
isolation
3.02 9.38 15.23 0.67 2.02 6.11 5.69 0.99 8.70
%
Quality of
23 11 79 5 3 9 3 13 146
door
7.72 34.88 8.30 1.67 3.03 6.87 2.44 12.87 7.17
%
Quality of
7 8 93 1 0 23 1 1 124
lintels
2.35 25.00 9.77 0.33 0.00 17.56 0.81 0.99 6.58
%
Quality of
34 11 52 12 0 8 4 7 128
windows
11.41 34.38 5.46 4.01 0.00 6.11 3.25 6.93 6.29
%
Design of
18 0 41 4 5 9 11 6 94
the shelter
6.04 0.00 4.31 1.34 5.05 6.87 8.94 5.94 4.62
%
Quality of
7 1 58 1 0 10 1 1 79
roof
2.35 3.13 6.09 0.33 0.00 7.63 0.81 0.99 3.88
%
The dissatisfaction about the model of shelter that was built in the respective community is also
confirmed by the community leaders, of which more than 58.6 per cent indicated that they were not
satisfied with the type of shelter built in the community as shown in
49
Table 14. The most common reason for this was the size of the shelter, perceived as being too small.
50
Table 14: Satisfaction with the Model (Community Representatives)
North Northeast South East West Central Total
Yes 5 0 3 12 0 4 24
% 8.62 0.00 5.17 20.69 0.00 6.90 41.38
No 6 3 4 11 4 6 34
% 10.34 5.17 6.90 18.97 6.90 10.34 58.62
Total 11 3 7 23 4 10 58
% 18.97 5.17 12.07 39.66 6.90 17.24 100.00
Complaints about the size of the rooms were particularly sensitive in Pashtun communities, with
traditionally large households, where beneficiaries often mentioned living in one shelter with over eight
and sometimes over ten people. Changes were introduced accordingly, through suppression of the
separation walls with the corridor to create a single room for instance. This was often the case in
Nangarhar and Kandahar. In such cases, the corridor in itself was deemed unnecessary, and at least one
wall was removed to create additional space to allow the family to gather. These types of regional
differences raise the question of the appropriateness of region-based approaches taking into account
cultural norms and practices tailored to regional needs and cultural practices. In cases where other
shelter programmes had been implemented and had provided larger rooms, such as UN-Habitat or CHF
in Nangarhar, UNHCR beneficiaries compared their shelters with those of other beneficiaries and
unanimously deemed the latter more appropriate considering cultural practices of gathering. There
appears to be little awareness about the rationale behind the existence of two separate rooms in the
shelter, both at the beneficiary and IP level. Education about the diffusion risks of propagation of
infectious diseases among members of a single household is therefore necessary.
It is important to note that there are regional differences in the satisfaction levels of the model. The
highest return area – the Eastern region – provides a balanced view of community satisfaction over the
type of shelters built. This is also the case in the South. This is partly explained by the fact that the design
of the shelters are better adapted to the warm climate conditions in the East and the South – and less
adapted to the Northern and Western areas. Although UNHCR has tried to adapt its shelter design to the
needs of the highest return areas, it should not be to the detriment of communities in the Northern,
Northeast and Western regions. A proper assessment of the climate, natural disaster risks and issues of
risk mitigation and prevention raised earlier, will need to be reinforced in future shelter programmes.
This can be a good opportunity for UNHCR to link up its technical assessment with that of engineers of
the Ministry of Refugee and Repatriation and the Ministry of Rural Rehabilitation and Development,
hence strengthening its partnership strategy.
Quality of doors and windows
Another recurrent complaint across all provinces, shared by beneficiaries, IPs and field engineers alike,
was the poor quality of the iron doors and window frames. Both are inadaptable to the weather
conditions (heat or cold) and subject to rust and deformation. Whenever their economic situation
allowed it, beneficiaries removed iron frames to replace them with wooden ones. In several locations,
51
iron doors had not been fixed and were used for other purposes, such as covering shacks or cooking
areas, or they were used as outside doors for compounds.
ii. Consequences of dissatisfaction: changes in the design post-handover
While most shelters were built according to UNHCR guidelines, some beneficiaries implemented changes
after the official handover, according to the capacity of the beneficiary family. As mentioned above, the
main change observed in the field was the removal of the corridor to increase the size of the two
remaining rooms. This was especially the case in the South (Kandahar & Helmand) and the East
(Nangarhar). In Hirat, the research team observed number of shelters significantly modified, with often
two or three shelters being joined one to another through the addition of a large common space and a
kitchen at the centre. In urban areas, beneficiary households often had to adapt the design of their
shelters to the size and shape of the land plot they occupy.
Changes resulted from different types of motivations:
 Whenever the design was considered inappropriate: enlargement of rooms, windows enlarged
for ventilation in Nangarhar and Kandahar, narrowed for protection from the cold in Hirat.
 Whenever the material provided in the package was deemed inappropriate: replacement of iron
doors and windows.
 Improvements that are indications of appropriation of the shelter and are positive sign in terms
of intention to settle.
A certain uniformity of changes was noted in specific areas, with entire communities adapting the design
according to specific regional or traditional needs (open kitchens in Hirat, enlargement of rooms in
Kandahar and Nangarhar). As long as they do not put extra economic pressure on beneficiaries or
endanger the general stability of the building, changes are not in themselves negative signs, but they are
rather an indication of an appropriation of the shelter according to the needs of beneficiaries, indicating
an intention to stay and settle. These adaptations call for technical monitoring to ensure that the
structure of shelters is preserved.
Changes in the design and poor use of risk-mitigation measures do call for stronger technical training
and awareness-raising initiatives to be conducted prior to the implementation of the programme in
order to contribute to its sustainability.
3.1.4 Construction Process
Receiving the material
Qualitative and quantitative observations showed that the provision of material to beneficiaries for
the construction of their shelter worked efficiently and that beneficiaries were satisfied with the
material they received.
52
UNHCR as well as other organizations provided all the necessary materials for building the shelters to
their respective beneficiaries in the large majority of sampled households (93.4% and 91.8%). Among the
UNHCR beneficiaries, differences are observed according to their location as shown in Table 15.
Respondents in urban areas reported that they did not receive all necessary materials in 13.2 per cent of
the cases, while this was the case significantly less in semi-rural (3.5%) and rural (5.9%) areas.
Table 15: Received Necessary Materials (UNHCR Beneficiaries)

(N=334) (N=424) (N=1276) (N=2034)
Yes 290 409 1 201 1 900

% 86.63 96.46 94.12 93.41
No 44 15 75 134
% 13.17 3.54 5.88 6.59
Total 334 424 1 276 2 034

% 100.00 100.00 100.00 100.00
More than 94 per cent of UNHCR beneficiaries and 92 per cent of other programme beneficiaries
reported that the quality of the materials they received was good. Among the UNHCR beneficiaries the
satisfaction with materials was a little lower in urban (88.6%) than in semi-rural (93.4%) and rural
(96.1%) areas. The majority of beneficiaries also reported receiving the materials on time (UNHCR:
94.9%; other programmes: 97.6%). Again, the reported conditions in urban areas are less satisfactory
with 8 per cent of respondents in this category indicating that they received their materials late. This
percentage is lower in semi-rural (4.0%) and rural (4.8%) areas.
Provincial differences in the procurement of material to beneficiaries were noticed. Quantitative findings
show that wood and wooden beams for example were distributed noticeably less in Laghman,
Nangarhar, Helmand and Kandahar compared to other provinces. The qualitative fieldwork also showed
indications for differences in the procurement of materials across regions. For instance, three iron doors
were provided to beneficiaries of two-room shelters in the West, whereas in the South, East and Central
regions, inside doors were wooden. Other variations included procurement of glass panes in the South
and East, whereas additional cash assistance was provided in the West and Central regions.
Procurement of wood: a challenge in the East and South
The quantitative data confirmed these observations, most notably on the procurement of wood. As seen
in
Table 16, over half of Laghman beneficiaries (59.9%) and almost half of Nangarhar beneficiaries (47.3%)
noted they did not receive wood as a material of the shelter package. Given that the Eastern region is
the highest area of return and of SAP interventions, the fact that procurement challenges were
specifically raised there should be remedied in future SAP strategies. Overall, the Eastern and Southern
region offices will need to improve their procurement of wood – as Southern provinces such as Helmand
53
(almost half of beneficiaries), Kandahar (one third of beneficiaries) and Paktya (one fifth of beneficiaries)
recorded the highest rates of challenges faced in wood procurement process.
Table 16: Procurement of Wood by Province (UNHCR Beneficiaries)
Yes No Total
Kabul 158 39 197

% 80.20 19.80 100.00
Parwan 76 25 101
% 74.25 24.75 100.00
Bamyan 30 2 32
% 93.75 6.25 100.00
Laghman 65 97 162
% 40.12 59.88 100.00
Nangarhar 416 374 790
% 52.66 47.34 100.00
Balkh 50 0 50
% 100.00 0.00 100.00
Faryab 69 6 77
% 92.00 8.00 100.00
Jawzjan 110 8 120
% 93.22 6.78 100.00
Sari Pul 56 0 56
% 100.00 0.00 100.00
Kunduz 59 1 60
% 98.33 1.67 100.00
Takhar 38 0 38
% 100.00 0.00 100.00
Helmand 29 27 56
% 51.79 48.21 100.00
Kandahar 50 25 75
% 66.67 33.33 100.00
Paktya 96 27 123
% 78.05 21.95 100.00
Hirat 91 10 101
% 90.1 9.9 100.00
Total 1 393 641 2 034

% 68.49 31.51 100.00
There were also disparities in the material used for the construction of walls, the main contribution of
beneficiaries to construction, according to the availability of material in specific areas. In Hirat for
54
instance, cement was preferred over mud bricks, due to the absence of clay in the region. The choice in
material therefore did not always result from the specific preference of beneficiaries or their economic
situation, but was also directly impacted by the availability of material, with repercussions on their level
of investment. This was taken into account in Hirat, with flexible cash grants, which was however not the
case in any other province. Other shelter agencies – such as NRC – now have adopted different methods
for the procurement of materials, meant to support local economies, decrease procurement hurdles and
give beneficiaries the responsibility to purchase construction materials. These different options will be
discussed at more length in the recommendations chapter.
Main problems encountered by beneficiaries during construction
The main problems encountered by beneficiaries during construction were:
 Lack of money
 Lack of water
 Lack of skilled labour
Construction of the shelters did not go smoothly in all cases. Slightly less than half (47.8%) of the UNHCR
beneficiaries reported that they ran into problems during construction of their shelters – an issue related
to the lack of technical assistance mentioned previously. A similar proportion of the beneficiaries of
other programmes had problems during the construction (50.0%). Table 17 shows the different types of
problems the beneficiaries encountered.
Table 17: Problems during Construction*
UNHCR Non-UNHCR
N=972 % N=236 %
Ran out of money 861 88.58 208 88.14
Insufficient access to water 457 47.02 121 51.27
Lack of skilled labour 287 29.53 65 27.54
Weather problems 278 28.60 70 29.66
Ran out of materials 203 20.88 54 22.88
Lack of unskilled labour 95 9.77 2 0.85
Materials of poor quality 74 7.61 20 8.47
Materials not delivered on time 55 5.66 8 3.39
Lack of technical knowledge 43 4.42 10 4.24

*Multiple answers per respondent were possible.
55
Looking specifically at the group of extremely vulnerable households, it shows that they encountered
problems during construction more often (52.1%) than non-EVI households (44.1%). Table 18 shows that
EVI households had more problems in all areas except for the timely delivery of materials. EVI
households in particular are significantly different than non-EVI households in terms of problems with
unskilled and skilled labour. While 21.7 per cent of non-EVI households experienced a lack of skilled
labour, this is the case for 37.2 per cent of EVI households. This confirms the need to provide extra
assistance to the most vulnerable during the construction process as they struggle more than others to
build their shelters. This also shows that – at least between 2009 and 2011 – the link between Protection
units and the implementation of the SAP was not strong enough to address this need efficiently.
Table 18: Problems during Construction by EVI Status (UNHCR Beneficiaries)*
EVI Not EVI
N % N %
Overall 492 52.06 480 44.08
-Ran out of money 442 89.84 419 87.29
-Insufficient access to water 238 48.37 219 45.63
-Lack of skilled labour 183 37.20 104 21.67
-Weather problems 142 28.86 136 28.33
-Ran out of materials 104 21.14 99 20.63
-Lack of unskilled labour 58 11.79 37 7.71
-Materials of poor quality 42 8.54 32 6.67
-Materials not delivered on time 26 5.28 29 6.04
-Lack of technical knowledge 26 5.28 17 3.54

The main problems are faced by households regardless of their location, however, the degree of the
problems vary between rural, semi-rural and urban households.
Quality of materials – low satisfaction in urban areas. Findings point to the inadequacy of the quality of
materials delivered to urban areas as they often do not match the quality available on the local market.
17.3 per cent of households in urban areas complained about the poor quality of materials, as opposed
to 6.7 per cent in semi-rural and 5.1 per cent in rural areas. Understandably, the more remote or rural
the beneficiary households are, the less critical they are of the quality of the materials. As a result, this
could inform future programming by considering vouchers or cash grants for households to buy their
own equipment in urban areas.
56
Lack of technical knowledge in urban areas. Although unskilled and skilled labour is easier to come by in
urban areas as compared to other locations, urban beneficiary households have insufficient technical
knowledge when it comes to building or supervising the construction of their shelter. This is also due to
the different landscape and requirements of urban shelter construction. An emphasis on developing an
urban approach to training and to support will therefore be necessary in future shelter programming.
Looking into how housing in Kabul and other urban areas can be improved, extended or expanded
support will contribute to greater protection of beneficiaries in urban areas.
The main problems in rural areas are the overall lack of labour and lack of access to water – further
developed in one of the sections below.
Table 19: Problems during Construction by Location (UNHCR Beneficiaries)*
Urban Semi-Rural Rural
N=168 % N=239 % N=565 %
Ran out of money 143 85.12 223 93.31 495 87.61
Insufficient access to water 67 39.88 104 43.51 286 50.62
Weather problems 56 33.33 50 20.92 172 30.44
Lack of skilled labour 39 23.21 76 31.80 172 30.44
Ran out of materials 34 20.24 73 30.54 96 16.99
Materials of poor quality 29 17.26 16 6.69 29 5.13
Lack of technical knowledge 16 9.52 10 4.18 17 3.01
Lack of unskilled labour 12 7.14 19 7.59 64 11.33
Materials not delivered on time 12 7.14 8 3.35 35 6.19

 Scarce financial resources
The lack of money was mentioned as by far the main challenge faced by beneficiaries when building their
shelters. Almost all beneficiaries mentioned that they had to take up loans to cover labour costs and
wall components. Additional costs were also necessary for buying stones for foundations and, depending
on the availability of material in a given area, bricks, cement or clay.
Household contributions: Higher expenditures in urban areas
Almost all beneficiaries, 93 per cent, had to contribute to the shelter construction as per SAP guidelines.
However, the amount of funds contributed varies significantly with urban UNHCR beneficiaries spending
significantly more out of their own pockets than rural beneficiaries as seen in Table 20. The data shows a
57
13,000 AFN (260 USD) gap between urban and rural households, and a smaller, yet sizeable gap of 6 810
AFN (136 USD) between urban and semi-rural households. This is due to the higher costs of materials
and labour in urban areas – higher costs that will have to be taken into account in developing an urban
strategy for the shelter programme, discussed in the recommendations chapter. Moreover, this is also
due to the fact that urban households on average earn a higher income than those of rural or semi-rural
areas. To speak in relative terms, Table 21 illustrates the amount beneficiary households paid on the
shelter as a percentage of their monthly income, providing evidence that while UNHCR beneficiary
households located in an urban context spend more in absolute terms, semi-rural households spend
slightly more in relative terms.
Table 20: Amounts Paid by Beneficiaries in AFN by Location
UNHCR Beneficiaries Non-UNHCR Beneficiaries
N Mean Min Max N Mean Min Max
Urban 319 46 088 900 700 000 78 28 244 5 000 100 000
Semi-rural 389 39 278 800 500 000 4 66 250 15 000 130 000
Rural 1 195 33 199 1 000 560 000 334 40 940 1 000 500 000
Total 1 903 36 602 800 700 000 416 38 803 1 000 500 000
Table 21: Percentage of Monthly Income Paid for Shelter by Beneficiaries by Location
Location UNHCR Beneficiaries Non-UNHCR Beneficiaries
N Mean Min Max N Mean Min Max
Urban 319 5.91 0.08 100.00 78 3.86 0.58 20.00
Semi-rural 384 6.18 0.06 138.89 4 17.02 3.75 26.00
Rural 1177 5.29 0.11 100.00 333 6.56 8.41 0.06
In terms of provincial differences illustrated in
58
Table 22, shows that beneficiaries in Hirat display the highest amount of household contribution at
48,870 AFN (977 USD) with the lowest expenses recorded in Sari Pul with 14,260 AFN (285 USD).
59
Table 22: Amounts Paid in AFN by Province (UNHCR Beneficiaries)
Province N Mean Min Max
Hirat 92 48 870 3 000 500 000
Paktya 112 48 000 2 000 300 000
Nangarhar 752 45 256 1 800 700 000
Kabul 176 40 582 2 000 250 000
Bamyan 30 33 720 900 200 000
Kandahar 67 30 184 800 400 000
Laghman 156 28 542 3 000 150 000
Helmand 50 24 340 3 000 85 000
Parwan 93 24 151 1 000 100 000
Jawzjan 110 23 773 1 000 410 000
Faryab 72 23 278 1 000 95 000
Takhar 36 19 556 1 000 50 000
Balkh 49 18 402 3 000 95 000
Kunduz 58 18 057 1 300 100 000
Sari Pul 50 14 260 1 000 45 000
Total 1 903 36 602 800 700 000
This difference in contributions – with a range covering a 700 USD difference – is better understood
when again looking at its relative burden when compared to household income. Shown in
60
Table 23, we see those households in Hirat spend by far the highest share of their monthly income on
the shelter, with Helmand having the lowest contribution. It is important to note the sub-office of Hirat
already reviews yearly the cash grant based on the costs of material and labour, a good practice that
should be generalized to all sub-offices.
61
Table 23: Percentage of Monthly Income Paid For Shelter by Province (in %) (UNHCR Beneficiaries)
Province N Mean Min Max
Hirat 89 11.14 0.30 138.89
Paktya 111 6.40 0.22 33.33
Nangarhar 751 6.25 0.08 100.00
Bamyan 30 6.07 0.08 40.00
Kabul 171 5.94 0.20 68.18
Balkh 47 4.96 0.75 95.00
Takhar 35 4.79 0.25 11.67
Parwan 90 4.76 0.11 100.00
Faryab 71 4.37 0.06 20.00
Laghman 155 4.35 0.38 21.67
Kunduz 57 3.83 0.22 20.00
Jawzjan 109 3.34 0.33 22.53
Kandahar 67 3.29 0.07 25.00
Sari Pul 47 2.97 0.22 22.50
Helmand 50 2.84 0.33 9.44
There are other important disparities across provinces. Typically regions of high return and high rates of
urbanisation, such as Hirat, Nangarhar and Kabul present significantly higher levels of household
contribution. This is unsurprising given the higher level of local prices and labour costs in these regions.
The material used for the construction of the shelter also enters into play, especially in Hirat province,
where beneficiaries had to use cement and burned bricks in the absence of clay, which significantly
increased the level of household contributions in this province.
The level of contribution expected from beneficiaries is not detailed in the SAP guidelines, which only
mention the fact that beneficiaries are expected to cover the costs of labour and of the construction of
walls. In general, stakeholders had a rough estimate of the level of contribution expected from
beneficiaries. NRC in Nangarhar estimates that it covered about 50 per cent of the costs of the shelter by
distributing a cash grant of $1,100. CARE on the other hand decided to cover the entire costs of the one-
room shelters that they built in the North, which is $900 more than UNHCR two-room shelters. The level
of household contribution should be more clearly defined by the organisation and should be included as
an indicator for the monitoring and evaluation of the programme as it plays an important role in its
impact and sustainability. Households’ contribution is important to guarantee a certain level of
62
involvement and commitment of beneficiaries to the process, hence playing a role in the sustainability of
the SAP. Yet, beneficiary households need to have a clearer idea of the costs and the total amount that
they will have to cover before starting the process to better plan the construction and reduce the
likelihood of unsustainable indebtedness. Data provided in table 22 can support this effort.
 Lack of water
Lack or limited access to water during the construction process is one of the main challenges during
construction for 40 per cent of urban household, 44 per cent of semi-rural households and 51 per cent of
rural households. This was notably the case in Chamtala and Sheikh Mesri in Nangarhar and in Northern
provinces. Water being a major requirement for elaboration of mud bricks, this placed a major burden
on beneficiary families especially in rural areas as it impacts half of the beneficiaries adversely.
Droughts during the summer were a major concern, as was the lack of fuel to allow water pumps to
function. In cases where the bulk of construction takes place in the summer, beneficiaries asked for
extensions of delays to wait for the rainy season. In some cases, beneficiaries were dependent on buying
water from water tanks, which were provided by local private companies for 500 AFN per week
(Kandahar) or paid for by UNHCR (Jawzjan). This type of differences calls for a more homogenized
approach and clearer guidelines about the support provided to beneficiaries in specific contexts. Starting
construction earlier in the spring would help reduce the risks of incompletion of shelters.
BOX 1: Access to Water
Lack or limited access to water during the construction process was mentioned as one of the main
challenges during construction. Water being a major requirement for elaboration of mud bricks,
this placed a major burden on beneficiary families. Droughts during the summer were a major
concern especially in Northern provinces which suffer regularly from acute drought, as were lack of
fuel to allow water pumps to function. In some cases, beneficiaries were dependent on buying
water from water tanks, an expensive resource provided by local private companies against 500
AFN per week (Kandahar) or paid for by UNHCR (Jawzjan). In Kunduz province (Sertak Sedarak),
some beneficiaries took on loans at the First Micro Finance Bank (FMFB) to cover water costs while
in another village of the province (Julgia Uzbekia) beneficiaries reported they had to pay 300 AFN
per hour to pump water from the river.
The issue of water did not affect exclusively drought-prone areas. In Helmand province, focus
groups highlighted similar issues and coping strategies: ‘My main problem during the construction
was the lack of water. I had to buy one water tanker and to pay 600 AFN. Overall I had to take on a
loan of 15,000 AFN from my relatives to be able to complete my shelter.’ (Twakal, Focus Group
Discussion with UNHCR Beneficiaries, Camp Mukhtar, Helmand).
The lack of water plays a role in the level of indebtedness of beneficiary households. These
additional costs should be taken into account when calculating the level of contribution expected
from beneficiaries.
63
 Lack of unskilled and skilled labour
Lack of unskilled labour was a specific concern for rural families (11.3%) while lack of skilled labour was a
concern throughout all locations – affecting 23 per cent of urban, 32 per cent of semi-rural and 30 per
cent of rural beneficiary households. As such almost one in four households in urban areas and one in
three households in semi-rural and rural areas lacked skilled labour for the construction of their shelter.
Not surprisingly, therefore, Table 24 shows the majority of UNHCR beneficiaries (68.4%) as well as
beneficiaries of other programmes (68.0%) had to hire labourers during the construction process – a
burden on beneficiary households but a positive repercussion on the local economy.
Most beneficiaries had to spend additional money when they did not have any skills in construction. The
mean cost UNHCR beneficiaries paid for labourers was 24,337 AFN, while beneficiaries of the other
programmes on average paid 18,369 AFN.
Table 24: Hiring Labourers during Construction
N % N %
Yes 1 392 68.44 315 68.03
No 642 31.56 148 31.97
Total 2 034 100.00 463 100.00
The hiring of additional labourers is a common trait throughout the sample, with limited geographic
variations according to location (Table 25) but more substantial differences across provinces (Table 26).
The households most dependent on external skilled labours were found in Bamyan, Laghman and Hirat,
with the least dependent in Kandahar, Parwan and Paktya. Provinces of high return, such as Kabul,
Nangarhar and Helmand, were close to average dependency rates.
Table 25: Hiring Labourers during Construction by Location (UNHCR Beneficiaries)
Urban Semi-Rural Rural Total
N % N % N % N %
Yes 249 74.55 285 67.22 858 67.24 1 392 68.44
No 85 25.45 139 32.78 418 32.76 642 31.56
Total 334 100.00 424 100.00 1 276 100.00 2 034 100.00
64
Table 26: Hiring Labourers during Construction by Province (in %) (UNHCR Beneficiaries)
Yes No Total
Bamyan 93.75 6.25 100.00
Laghman 82.72 17.28 100.00
Hirat 79.21 20.79 100.00
Sari Pul 78.57 21.43 100.00
Balkh 76.00 24.00 100.00
Takhar 73.68 26.32 100.00
Jawzjan 72.50 27.50 100.00
Kunduz 71.67 28.33 100.00
Nangarhar 68.76 31.24 100.00
Kabul 63.82 36.18 100.00
Helmand 62.50 37.50 100.00
Faryab 57.14 42.86 100.00
Paktya 56.91 43.09 100.00
Parwan 56.86 43.14 100.00
Kandahar 43.42 56.58 100.00
Total 68.36 31.64 100.00
 Delays: As per UNCHR shelter guidelines, beneficiaries are obligated to complete their shelters
within three months of signing the letter of undertaking, unless special circumstances cause
delays. While almost 70 per cent of UNHCR beneficiaries did complete construction within this
time frame, there are still 30 per cent that took longer, in the majority between three and six
months. Among the beneficiaries of other programmes, this number is slightly less (22.3%). Table
27 shows that construction by UNHCR beneficiaries in urban areas was more often completed
within the three-months timeframe (81.1%) than that of the beneficiaries in semi-rural (66.0%)
and rural areas (67.2%).
65
Table 27: Duration of Construction by Location (UNHCR Beneficiaries)
N % N % N % N %
Less than 3 months 271 81.14 280 66.03 857 67.17 1408 69.22
3 to 6 months 54 16.17 114 26.89 351 27.51 519 25.52
More than 6 months 9 2.70 29 6.84 67 5.25 105 5.06
Not yet finished 0 0.00 1 0.24 1 0.08 2 0.10
Total 334 100.00 424 100.00 1 276 100.00 2 034 100.00
The delays in finishing the construction in three months were often caused by the problems previously
detailed – the lack of resources, skills and water:
 Lack of skills: IPs sometimes mentioned that short delays put considerable pressure on
monitoring of the construction process. In Saracha, site engineers emphasized that unequal
construction skills of beneficiaries were a major challenge, as many did not have any prior
experience in construction and therefore required additional technical assistance.
 The need to sustain a living during the time of construction: most beneficiaries cannot afford to
focus on construction on a daily basis.
 Lack or limited access to water: In cases where the bulk of construction has to take place in the
summer, beneficiaries asked for extensions to wait for the rainy season. Starting construction
earlier in the spring would help reducing the risks of incompletion of shelters before winter.
In addition, other problems were raised during qualitative interviews:
 Lack of flexibility: Beneficiaries mentioned not having been able to introduce changes in the
design during construction, lest they should receive the final cash grant. Subsequently, changes
in structure were often introduced after completion. The inclusion of additional wooden beams
in order to enlarge habitable space can notably prove problematic, putting in danger the overall
structure of the building by introducing dissymmetry in the design of the shelter. This calls for
greater awareness behind the reasons for the design and additional technical advice on specific
points regarding the structure of the building.
In some cases, the size of the land plot required adaptation of the design of the shelter,
associated with an additional investment. In such cases (Saracha reintegration site, Kahdistan),
IPs allowed minor changes in the design, but mentioned no additional assistance was given to
beneficiaries, a problem given the necessity to extend surrounding walls.
66
 Indebtedness
The qualitative data collection showed that the level of indebtedness also depends on the
material used (burnt or mud bricks). The amount of debt varied from 50,000 AFN to 100,000 AFN
and was sometime even as high as 200,000 AFN. Interestingly, in Hirat province, cash grants
were adapted to fluctuations of labour costs year after year, a practice that was not noticed in
other areas.
More than 83 per cent of the surveyed households indicated that they had outstanding debt at
the time of interview. The national average level of debt was 99,208 AFN. Broken down by the
type of location, Table 28 shows that debt levels are highest in semi-rural and lowest in rural
areas. UNHCR beneficiaries as well as Non-UNHCR beneficiaries have higher debt in total, while
the average debt in urban areas is lower for UNHCR beneficiaries (111,905 AFN) than for Non-
beneficiaries (112,702 AFN). This might be due to the fact that households have to invest more
of their own resources into building a shelter when they do not receive the assistance by UNHCR.
Table 28: Average Level of Debt by Location and Beneficiary Status

Total
N Mean N Mean N Mean N Mean
Urban 283 111 905 73 136 712 350 112 702 706 114 865
Semi-rural 346 122 689 5 89 900 330 119 529 681 120 917
Rural 1 045 90 897 343 94402 967 83 175 2355 88 237
Total 1 647 101 019 421 101 685 1 647 96 734 3742 99 208
As for household debt by province,
67
Table 29 shows that among UNHCR beneficiaries, those in Helmand, Paktya, Kandahar, Kabul and
Nangarhar have the highest average levels. Compared with non-beneficiaries, we see in certain
provinces like Kabul, Balkh, Faryab, Sari Pul, Kunduz and Paktya that UNHCR beneficiaries have
noticeable lower overall debt. However the situation is just the opposite in other provinces like Bamyan,
Nangarhar, Jawzjan, Takhar, Helmand and Kandahar where UNHCR beneficiaries have higher debt
relative to non-beneficiary households.
68
Table 29: Level of Debt by Province and Beneficiary Status

Total
N Mean N Mean N Mean N Mean
Kabul 165 113 079 1 50 000 153 122 039 319 117 179
Parwan 78 61 615 0 - 66 61 212 144 61 431
Bamyan 27 79 741 0 - 21 56 319 48 69 494
Laghman 146 86 062 0 - 119 82 899 265 84 642
Nangarhar 670 112 618 415 101 786 712 100 555 1797 105 337
Balkh 37 49 189 0 - 44 61 298 81 55 767
Faryab 56 60 946 2 170 250 80 79 863 138 73 496
Jawzjan 63 68 540 2 54 000 65 51 031 130 59 562
Sari Pul 43 37 698 0 - 29 50 931 72 43 028
Kunduz 49 45 776 0 - 52 51 596 101 48 772
Takhar 38 54 026 0 - 29 40 621 67 48 224
Helmand 49 169 020 0 - 45 157 756 94 163 628
Kandahar 67 144 582 1 70 000 65 124 766 133 134 337
Paktya 95 155 347 0 - 87 168 023 182 161 407
Hirat 91 79 951 0 - 80 79 975 171 79 962
Total 1 647 101 019 421 101 685 3 742 99 208 3 742 99 208
When asked about the impact of the shelter assistance programme on household debt 47.2 per
cent of all beneficiaries indicated that it increased. This appears to be a bigger problem for
beneficiaries of other programmes (54.4%) than for the UNHCR beneficiaries (34.5%).
69
Table 30: Impact of Shelter Programme on Household Debt (in %)
UNHCR Non-UNHCR
Total
Decreased 34.51 32.90 34.24
Increased 45.80 54.40 47.24
Remained the same 11.04 8.81 10.66
No debt 7.05 2.85 6.35
I don’t know 1.61 1.04 1.51
Total 100.00 100.00 100.00
However, no abandonment of shelters due to debts contracted because of the programme were noticed
in the field, a potential sign that this is not a major threat to sustainability in the short term, but might
become one later on if sufficient income opportunities are not secured. This was notably a major
concern in Kandahar, where little job opportunities were available for beneficiaries.
3.1.5 Handover
Of the UNHCR beneficiaries interviewed for the purpose of this study, 2,026 indicated that they had
completed the programme entirely. The large majority (96.4%) did receive their handover certificate.
Yet, these figures are probably misleading, as beneficiaries who may have dropped out of the
programme were a lot less likely to be included in the sampling. While six households are still in the
process of completing the programme, two had dropped out along the way. A little over 3 per cent
completed building their shelter, but did not receive a handover certificate. Differences across different
types of locations or provinces were not noticed.
Table 31: Completion of Construction (UNHCR Beneficiaries)

N %
Yes, we have our handover certificate 1960 96.36
Yes, but we did not receive our handover certificate 66 3.24
No, the shelter is not yet finished 6 0.29
No, we dropped out of the program 2 0.10
Total 2 034 100.00
70
No particular problems were mentioned during handover, with beneficiaries noting they received the
cash grants after control of the completion of the shelter. In most cases, shelters were effectively
completed and no major issues were reported in this respect. The distribution of cash grants is often a
delicate stage in the implementation of a programme in Afghanistan. The absence of reported problems
and frauds at that stage of the shelter programme is therefore a positive finding in and of itself. A
notable difference in the allocation of cash grants was observed in Hirat, however, with the Sub-Office
allegedly adapting the final grant to yearly fluctuations of labour costs, a practice worth considering as
beneficiaries repeatedly mentioned strains implied by the level of indebtedness due to purchase of
material and costs for additional labour.
The involvement of DoRR representatives in handover varied according to the relationship of the UNHCR
Sub-office with the Directorate. In Nangarhar and Kandahar, for instance, mistrust between the UNHCR
and the DoRR lead to occasional absence of the later during handover. This issue will be raised again in
Chapter 6 on Partnerships.
3.1.6 Risk Mitigation and Prevention

One of the weaknesses in the design of the programme identified in the field was the lack of an
assessment of natural disaster risks conducted prior to construction. Preventive measures imposed by
the programme’s guidelines are limited and only cover earthquake-mitigation measures.
In earthquake-prone areas, risk mitigation is solely taken into account through the inclusion of wood-
bracing in the design of the shelter. However, in the East, DoRR reported that wood-bracing was often
removed by beneficiaries, due to a lack of awareness of their use. This emphasizes the need for proper
awareness training about the importance of such elements.
Preventive measures against floods are also seriously lacking. This was notably the case in Nangarhar,
Hirat and Jawzjan. In the latter ZOA issued practical recommendations to upgrade shelters and avoid
degradation, which had happened in the province in 2012. Despite high risks in the province, the only
measure recommended in practice by UNHCR was to build the shelters 60 cm above the ground, which
was not systematically implemented across the province and is insufficient in case of serious flooding. In
Kandahar, Hirat, Jawzjan, Parwan and Nangarhar inhabitants insisted on the need to build retaining walls
KIIs with UNHCR’s sub-offices, IPs and other stakeholders such as the ANDMA, confirmed that proper risk
assessments in flood-prone and earthquake-prone areas were absent. Coordination efforts between
ANDMA and UNHCR were scarce. Based on proper risk assessments, UNHCR could envisage adopting a
firmer prevention policy, which would include the non-inclusion of flood-prone areas in the programme.
At the central level, the MoRR raised serious concerns about risk-mitigation in UNHCR design and site
selection:
71
 Past large-scale destructions were not due to the strength of earthquakes, but to weak
construction.
 The regional and environmental context needs to be taken into account because availability of
material and poor weather conditions affect construction and building.
 Poor mapping capacity of ANDMA.
 No multilateral approbation committee for design.
Both in Hirat and Nangarhar, ANDMA insisted there had been no consultation about risks with UNHCR
prior to the implementation of the programme in the respective province. Interestingly, ANDMA
underlined being able to conduct such evaluations, but being only consulted in post-disaster situations,
highlighting they had better coordination and cooperation with other UN agencies such as the WFP. The
insufficient acknowledgement of these risks is highly problematic, as it threatens the sustainability of the
programme in specific areas, with a direct impact on its cost-effectiveness in cases where batches of
shelters are destroyed by natural disasters.
3.2 SUPPORT
3.2.1 Additional Assistance for Beneficiaries

UNHCR mainly relies on ashar, or community assistance, to support the most vulnerable households.
However, this was not a practice noticed in the field. Community members mentioned ashar could not
be an option, as most villagers were faced with difficulties in sustaining their own household. Community
representatives, however, indicated that community members did assist the beneficiaries in building
their shelters in 60 per cent of cases. Mainly this was assistance in the form of unskilled labour; in rare
cases community members also provided skilled labour and materials.
The distribution of cash assistance prior to completion of the shelters was mentioned in Hirat, but
seemed to be generally avoided in other provinces. Both IPs and UNHCR staff mentioned concerns about
potential misuse of the money. This calls for more consideration about effective measures to support
EVIs, notably through closely monitored cash assistance. The programme guidelines and UNHCR’s EVI
programme plan for additional cash assistance for EVI beneficiaries, but the research team found very
rare examples of this practice actually implemented in the field.
This shows that the link between protection and the shelter programme is still insufficient at the sub-
office level, as the mechanism in place to identify and provide additional support to EVIs is inefficient – a
lost opportunity for the programme to fully take into account and address the specific needs of the most
vulnerable among UNHCR’s target population.
72
3.2.2 Complementary Training
Most beneficiaries did receive some form of training in conjunction with shelter assistance – however
almost one in three beneficiary households indicated not having received any support or training.
As
Table 32 shows, 28.2 per cent of UNHCR beneficiaries did not receive any training, while fewer
beneficiaries of other programmes did not receive training (13.4%). The most common form of training
that was provided to beneficiaries was training on construction (UNHCR: 54.3%; other programmes:
65.2%), followed by maintenance training (UNHCR: 33.0%; other programmes: 41.0%) and training on
procurement issues (UNHCR: 26.2%; other programmes: 38.2%).
Table 32: Training Received in Conjunction with Shelter Assistance
N % N %
No training 574 28.22 62 13.39
Training 1 460 71.78 401 86.61
- Training on construction 1 105 54.33 302 65.23
- Training on maintenance 672 33.04 190 41.04
- Training on procurement issues 533 26.20 177 38.23
- Hygiene promotion 404 19.86 159 34.34
- Other training 2 0.10 0 0.00
When comparing training provided by UNHCR vs. other shelter agencies, findings show that UNHCR
beneficiaries fare worse off in terms of the support they receive. Other shelter beneficiaries
systematically received more training than UNHCR shelter beneficiaries as shown in
Table 32. While one in three UNHCR shelter beneficiaries did not receive any training, this number drops
down to one in seven in other shelter programmes.
Within the training sessions conducted, most concerning was the gap on hygiene promotion. UNHCR
shelter beneficiaries are significantly less likely to receive any hygiene support. Hygiene promotion was
indeed less common (UNHCR: 19.9%; other programmes: 34.3%) - a key finding of this study and a point,
which will be discussed in the recommendations of this report. Hygiene training and WASH assistance
should be improved since the research has shown that the state and use of latrines was highly related to
the implementation of such training.
73
3.2.3 Complementary Training by Location
The breakdown by location shows that rural beneficiaries are the least likely to receive training – and
urban beneficiaries the most likely. While almost 80 per cent of urban beneficiaries receive training, the
percentage drops to 71 per cent for semi-rural beneficiaries and 70 per cent for rural beneficiaries (
Table 33).
The most significant difference between locations is seen for the training on procurement issues and
hygiene promotion.
Table 33: Training by Location (UNHCR Beneficiaries)*

(N=334) (N=424) (N=1276) (N=2034)
No Training 69 122 383 574

% 20.66 28.77 30.02 28.22
Training 265 302 893 1 460
% 79.34 71.23 69.98 71.78
- Training on construction 186 251 668 1 105
% 55.69 59.20 52.35 54.33
- Training on maintenance 121 141 410 672
% 36.23 33.25 32.13 33.04
- Training on procurement issues 110 106 317 533
% 32.93 25.00 24.84 26.20
- Hygiene Promotion 97 46 261 404
% 29.04 10.85 20.45 19.86
- Other training 0 0 2 2
% 0.00 0.00 0.16 0.10
Furthermore, Table 34 shows that remote areas are the least likely to be covered by training
programmes, with the notable exception of training on procurement issues which is slightly more
prevalent in remote areas.
Table 34: Training by Remoteness of Location (UNHCR Beneficiaries)*
Remote Area Non-remote Area Total

(N=522) (N=1512) (N=2034)
No Training 165 409 574

% 31.61 27.05 28.22
Training 357 1 103 1 460
% 68.39 72.95 71.78
74
- Training on construction 283 822 1 105
% 54.21 54.37 54.33
- Training on maintenance 163 509 672
% 31.23 33.66 33.04
- Training on procurement issues 140 393 533
% 26.82 25.99 26.20
- Hygiene Promotion 89 315 404
% 17.05 20.83 19.86
- Other training 0 2 2
% 0.00 0.13 0.10
3.2.4 Complementary Training for EVIs

When focusing on the differences in training of UNHCR beneficiaries in terms of whether the household
is considered an EVI or not as presented in Table 35, we find that EVIs were slightly more likely to receive
training than non-EVIs, 73.8 per cent compared to 70.1 per cent. Still this difference is minimal and
supports the argument that the shelter programme is able to focus more on EVIs not only in the
selection process, but also in support training provided– whether in construction, maintenance,
procurement or hygiene promotion.
Table 35: Complementary assistance to EVIs (UNHCR Beneficiaries)*
EVI No EVI Total

(N=945) (N=1089) (N=2034)
No Training 248 326 574

% 26.24 29.94 28.22
Training 697 763 1 460
% 73.76 70.06 71.78
% 56.83 52.16 54.33
- Training on maintenance 355 317 672
% 37.57 29.11 33.04
- Training on procurement issues 267 266 533
% 28.25 24.43 26.20
- Hygiene Promotion 180 224 404
% 19.05 20.57 19.86
- Other training 0 2 2
% 0.00 0.18 0.10
When disaggregating by location, the survey shows that EVI beneficiaries in urban areas are much more
likely to receive training than semi-rural or rural beneficiaries. Indeed, only 17.4 per cent of UNHCR
75
beneficiaries in urban areas had not received any training, compared to one in four beneficiaries in semi-
rural areas and almost one in three in rural areas (29.0%).
However, it is important to note that hygiene promotion training for EVIs – which is across the board the
least well-covered training type – is lacking the most in semi-rural areas where only one in ten
beneficiary households have reported receiving hygiene promotion training, compared to one in five
rural households and over one in four urban households.
These geographical discrepancies highlight the overall lack of support training but its specific lack in non-
urban locations (
Table 36).
Table 36: Complementary assistance to EVIs by Location (UNHCR Beneficiaries)*

(N=167) (N=209) (N=569) (N=945)
No Training 29 54 165 248

% 17.37 25.84 29.00 26.24
Training 138 155 404 697
% 82.63 74.16 71.00 73.76
% 61.68 60.77 53.95 56.83
- Training on maintenance 76 76 203 355
% 45.51 36.36 35.68 37.57
- Training on procurement issues 61 53 153 267
% 36.53 25.36 26.89 28.25
- Hygiene Promotion 46 22 112 180
% 27.54 10.53 19.68 19.05
- Other training 0 0 0 0
% 0.00 0.00 0.00 0.00
3.2.5 Complementary assistance at the community level

In areas where it was made available, complementary assistance worked as a strong condition for the
sustainability of the program, providing communities with facilities and essential services or the creation
of a sustainable environment (water, schools, clinics, and roads) and contributing to diffuse tensions by
benefitting the entire community. Most villages had benefitted from NSP programmes through DRRD
(drilling of wells, cleaning of qarez) and assistance from other organizations (shelter, WASH
programmes).
The importance of complementary assistance is acknowledged by UNHCR in its guidelines and at the
Kabul level. Empowering the community and providing help to develop water points, schools and
76
infrastructure is therefore considered by UNHCR as one of the components of the program to enhance
its sustainability. However, outside of reintegration sites, additional assistance to communities seemed
to be more of a coincidence than the result of any form of coordination, and the shelter programme was
often a “stand alone” intervention. This was notably the case in areas where access was a problem, for
instance in Kandahar province and in Kahdistan. The “integrated” approach upheld by UNHCR prior to
2012 does not seem to have been systematically implemented. In Hirat, there were no regular patterns
for complementary assistance: WASH programs had not been implemented since 2008 and cash for work
has only been done in parallel to shelter in some cases. In Jalalabad conversely, IPs mentioned WASH
programmes were systematically included as part of the implementation of the shelter programme and
non-beneficiaries insisted on the benefits of such initiatives. Systematic implementation complementary
programmes (schools, clinics, WASH) appears as a good practice to be considered at the national level,
including through partnerships with other agencies and organizations, as well as the involvement of
provincial directorates. However, it should not be reduced to specific sites since needs are widely
present – a needs-based, rather than location-based, approach should therefore be adopted to ensure
that needs are covered.
3.3 MONITORING & EVALUATION
3.3.1 Systems of Monitoring and Evaluation
The majority of the community representatives reported that the technical advisers visited their
communities on a regular basis, on average between three and five times during the construction
process with a drop after the handover. This highlights that there is close to no follow-up of
beneficiaries, which limits any internal assessment of sustainable reintegration or longer-term impact
of the programme. This will be a key point to incorporate in internal field assessments to ensure a more
continuous M&E process.
In most cases field visits were conducted once per week or once every two weeks. Most beneficiaries
also mentioned receiving regular visits of IPs throughout the process. More than 99 per cent of the
beneficiaries of other programmes indicated that there had been monitoring of their shelter during the
construction process. This is an indication that the monitoring systems in place in other shelter
programmes (in the East) are more comprehensive than those of UNHCR overall, where more than 5 per
cent were not monitored at all.
77
Table 37: Monitoring of Shelter
N % N %
Yes 1927 94.74 459 99.14
No 107 5.26 4 0.86
Total 2034 100.00 463 100.00
There were no significant monitoring disparities between urban, semi-rural or rural locations.
Surprisingly, some of the provinces that rated lowest on monitoring were Sari Pul (85.2%), Parwan
(88.1%), Hirat (88.1%) and Kabul (89.9%) – the most secure provinces and where access is open,
therefore not justifying a lack of M&E due to security or other restrictions.
Faryab rates among the least monitored provinces, understandable given the security and access
conditions (see provincial overview for Faryab). Out of the list below (Table 38), a number of provinces
are limited due to their difficult access but IPs could be tasked to reinforce monitoring and follow-up.
Among these are, as mentioned above Kabul, Parwan, Sari Pul, and Hirat.
Table 38: Monitoring practices by province (UNHCR Beneficiaries)
Yes No Total
Kabul 177 20 197

% 89.85 10.15 100.00
Parwan 89 12 101
% 88.12 11.88 100.00
Bamyan 31 1 32
% 96.88 3.13 100.00
Laghman 158 4 162
% 97.53 2.47 100.00
Nangarhar 781 9 790
% 98.86 1.14 100.00
Balkh 47 3 50
% 94.00 6.00 100.00
Faryab 65 10 75
% 86.67 13.33 100.00
Jawzjan 104 14 118
% 88.14 11.86 100.00
Sari Pul 48 8 56
% 85.71 14.29 100.00
78
Kunduz 57 3 60
% 95.00 5.00 100.00
Takhar 37 1 38
% 97.37 2.63 100.00
Helmand 55 1 56
% 98.21 1.79 100.00
Kandahar 72 3 75
% 96.00 4.00 100.00
Paktya 117 6 123
% 95.12 4.88 100.00
Hirat 89 12 101
% 88.12 11.88 100.00
Total 1 927 107 2 034
% 94.74 5.26 100.00
This assessment is confirmed when looking at the broader regional reach of monitoring activities (
Table 39). The breakdown does not follow security or access points. The Western and Central regions
rate lower than average. Achievements in monitoring in the East are highest, followed by the South.
Table 39: Monitoring of Shelter by Region (UNHCR Beneficiaries)
Yes No Total
Central 266 32 298

% 89.26 10.74 100.00
Central Highland 31 1 32
% 96.88 3.13 100.00
East 939 13 952
% 98.36 1.37 100.00
North 264 35 299
% 88.29 11.71 100.00
Northeast 94 4 98
% 95.92 4.08 100.00
South 127 4 131
% 96.95 3.05 100.00
Southeast 117 6 123
% 95.12 4.88 100.00
West 89 12 101
% 88.12 11.88 100.00
79
Total 1 927 107 2 034
% 94.74 5.26 100.00
During implementation, close monitoring of construction seemed to be regularly carried out by the IP
staff following the construction throughout the process, with rates varying according to the scope of the
area visited and the level of access of areas (several times a week to two times a month). Both
beneficiaries and IPs reported that IP engineers would regularly visit the shelters, often three to five
times until handover. Beneficiaries mentioned that during the regular visits by IP staff throughout
construction they received basic explanations about the plan. In Kandahar and some areas of Nangarhar
a foreman was recruited in the community and hired by the IP to ensure monitoring and in some cases
appeared to be the only one involved in day-to-day monitoring. Hiring short-term local staff was usually
related to the difficulties of access, like in Kandahar for example. IPs in Nangarhar (Saracha) underlined
the fact that additional assistance needed to be given whenever beneficiary households were unskilled
and unable to hire skilled labour, placing a strain on the deadlines.
Overall, IPs seemed to have the required technical expertise to provide support, though flexibility in the
assistance provided derived more from their own willingness than from general UNHCR guidelines.
The involvement of UNHCR staff in monitoring of implementation varied according to:
 The degree of accessibility of areas due to security restrictions: very limited access in Kandahar
and Helmand. In Kunar and Laghman, all monitoring activities were sub-contracted to a specific
monitoring IP.
 Practices in sub-offices: In Hirat, UNHCR staff was regularly in the field and directly monitored
construction. Conversely, they were generally absent in Kabul and Parwan, despite the generally
safe context and high accessibility of PSUs.
In some cases, UNHCR staff relied on their own networks to monitor the situation in the field and to
triangulate information provided by the IP. It is mostly the cases in provinces where UNHCR has
experienced national staff able to work through their own information networks, like in Faryab for
example.
One major problem is the fact that there is no follow-up at all after handover. This causes a subsequent
serious lack of data on the outcomes of the programme, which is problematic in terms of measuring its
impact and assessing whether the envisaged objectives and outcomes have been met. It also reduces the
opportunity to check whether the adaptations to the design made by beneficiaries do not endanger the
soundness of the shelters.
3.3.2 Accountability of Programme Stakeholders

Complaint mechanisms appeared rather non-existent for beneficiaries, and mainly rely on the degree of
availability of the IPs and their willingness to address problems faced by beneficiaries. However, yearly
reviews of the programme done conjointly by UNHCR Sub-Offices and IPs to identify strengths and
weaknesses in the implementation of the programme proved useful, with subsequent changes
introduced in the design based on field observations. This was identified as a good practice, though the
80
UNHCR Sub-Office in Nangarhar expressed concern about the fact that recommendations were
sometimes not sufficiently taken into account at the central level.
Identification of Potential Cases of Fraud and Misallocation of Assistance
In general, it was observed that in some cases beneficiaries did not use shelters as intended. When
shelters are transformed into storage rather than living space, it is obvious that the selection process has
its flaws. Shelters are, in this case, not an immediate and urgent need for beneficiary households.
Two cases of fraud were identified in Khanaqa, with one household receiving two shelters (husband and
wife), and one other where the occupant had bought the VRF and the assistance “package” that goes
with it from a wealthier community member. Another indicator for misallocation is the fact that
respondents indicated that they currently own another shelter than the one they built through the
shelter assistance programme.
Table 40 shows that 21.0 per cent of UNHCR beneficiary households say they have at least one other
shelter. This is the case for an even higher percentage of beneficiaries of other programmes (33.5%).
Further cases of fraud are mostly related to the selection process and will therefore be presented in the
following section.
Table 40: Additional Shelter Owned by Household
UNHCR Non-UNHCR
N % N %
Yes 428 21.04 155 33.48
No 1 602 78.76 306 66.09
I don’t know 4 0.20 2 0.43
Total 2034 100.00 463 100.00
81
4. BENEFICIARY SELECTION
Most stakeholders described the selection process as the most sensitive stage of implementation, with a
high potential for tensions to be created inside the community, as well as between various stakeholders
(local authorities, IPs and beneficiaries). It also has a strong impact on the sustainability of the
programme. Yet, the selection process clearly appeared as the main weakness in the implementation
of the shelter program as it failed to integrate the most vulnerable.
Many flaws in the process were identified during qualitative fieldwork and confirmed by quantitative
analysis. These include:
 Irregular selection and participation of BSC members

 Misunderstanding of the selection criteria
 Insufficient focus on vulnerability as put forward in the UNHCR Shelter Guidelines
 Exclusion and under-representation of main vulnerable categories of displaced population:
o IDPs, female heads of households and landless people were largely excluded from
selection. Only 9 per cent of beneficiaries in our sample were IDPs, while only 2 per cent
were female-headed households.
o Households with health conditions and disabilities were under-represented in the
selection of eligible EVI households, with preference given to socio-economic and
demographic vulnerabilities, underlining a concern of exclusion of the ill and disabled in
the SAP.
 Error of inclusion:
o More than half of non-refugee returnees receiving UNHCR assistance are not considered
to be in the “extremely vulnerable” based on the EVI categories, indicating a
misallocation of assistance as this group does not present the migratory profile nor signs
of vulnerability that would make them eligible.
The main factor explaining these failures is the significant gap between the SAP guidelines on paper and
the reality of selection as it is conducted on the ground, where the Voluntary Repatriation Form (VRF)
and land ownership take precedence over any other criteria of selection. As it is, the selection process
does not allow the SAP to live up to some of its key guiding principles, such as women’s direct
participation or the focus on vulnerability.
82
4.1 THE SELECTION PROCESS AND ACTORS INVOLVED
According to the UNHCR shelter guidelines, ensuring the smoothness of selection in a given community
mainly lies in the hands of the Beneficiary Selection Committee (BSC), responsible for identifying
vulnerable households and facilitating the selection process in a transparent way.
However important variations in the selection procedure were observed in the field. Firstly, the inclusion
of the Beneficiary Selection Committee (BSC) was not always systematic or was not identified by the
beneficiary communities. In Kandahar for instance, the BSC appeared non-existent. Selection and
identification of beneficiaries were primarily conducted by community leaders, with only some
involvement of the IPs. Little or no oversight from BSC members (UNHCR, IP, DoRR) in specific areas
raises concerns about the capacity to effectively reach vulnerable people and leaves the door open for
favouritism to interfere in selection. A variety of practices have been observed in different provinces
according to the extent of involvement and interference of various stakeholders in the selection.
Irregular involvement of BSC members and lack of balance in the involvement of various actors in the
selection process appeared to have a direct impact on the transparency and effectiveness of the process
in reaching the most vulnerable.
The difference in practices observed resulted from the following, sometimes overlapping factors:
a) The degree of involvement of UNHCR staff
There is a strong correlation between the degree of involvement of UNHCR in the BSC and the
accessibility of the area of implementation. This is however not a systematic pattern: UNHCR
sometimes relies on its IPs for selection even in accessible areas, such as Parwan or Kabul - a
highly problematic trend in terms of monitoring of the selection process and of ensuring that the
guidelines and criteria are correctly implemented.
On the other hand, good practices were identified in Hirat, where UNHCR staff was present
throughout the selection process. IPs reported that the presence of UNHCR was a strong
component for the credibility of the committee in the eyes of the local authorities and that it
ensured the correct implementation of the guidelines.
b) The degree of reliance on the implementing partners (IPs)
As mentioned previously, over-reliance on IPs often directly resulted from the inaccessibility of
an area to UNHCR staff. In this case selection mainly rested in the hands of IPs and community
leaders, as for example in Kandahar. Instances of “remote selection” emphasized the need for a
strong monitoring procedure and follow-up. Lack of reliable monitoring posed a direct threat to
the transparency of the selection process and the effectiveness of the process in targeting the
most vulnerable.
In Nangarhar, a separate IP was specifically hired to monitor the selection procedure, but this
was not the case in most provinces. In general, remote selection and monitoring requires further
83
attention and follow-up by UNHCR staff, all the more as there were several allegations of
corruption at the IP level on behalf of beneficiaries and non-beneficiaries. The selection process
is one of the stages, when instances of frauds and corruption are the most likely to occur.
Despite the fact this is acknowledged as a problem at the central level, it does not seem to have
repercussions in the field. Though allegations need to be taken cautiously, given the high
sensitivity of the selection process and eventual resentments, they highlight the need for close
and transparent monitoring specifically tailored for the selection process and direct involvement
of UNHCR wherever access is not an issue.
c) The degree of reliance on community leaders (maliks, shura)
Inclusion of local authorities is a requirement for the smoothness of the process in a given
community and their transparent involvement in the process a guarantee for the success of
identification and selection of beneficiaries.
However, transparency often remained a serious issue: one of the major problems mentioned
in the field was the potential bias introduced in the selection process through direct
interference of community leaders. Favouritism was identified as a common practice, especially
when maliks are both responsible for identifying eligible community members (almost
exclusively VRF holders) and prioritizing the needs, with little or no oversight from UNHCR or
other actors in selection20. This was often the case in homogeneous single-tribe communities. A
recurrent complaint on behalf of beneficiaries was the need for “relationships” and privileged
contacts with community leaders to be included in the programme, which field visits sometimes
confirmed. This was the case in Majbur Abad (Nangarhar) and Khanaqa (Parwan), for instance. In
Shakalak e Islam (Jawzjan province), the deputy qariador (malik in uzbek areas) and his relatives
had received seven shelters in one compound, four of which were used as storage rooms or
guesthouses. In some areas of Jawzjan, UNHCR and its IPs lost access to implementation areas
and relied on community leaders for selection (Dashte Laily), which made it difficult to assess the
reliability and efficiency of the process.
In some cases, not all bodies of authorities in communities were taken into account and given a
role in the selection process. In these cases, the IPs strongly relied on a single authority in
locations where several were in charge of a given community, while others were left out (notably
women shuras). This opens the door for complaints, resentment and a strong feeling of
discrimination.
In other instances, due to an over-reliance on community leaders, the selection procedure

described in the guidelines does not seem to have been followed at all. In Kandahar and Kabul
provinces for instance, there were reports of the use of games of chance (Pitawa in Kabul
province and Laghman), where the malik allocated assistance by picking names written on pieces
of paper. This practice was also noticed in Kuchi Abad for the allocation of plots of land to
beneficiaries, which is disquieting as the site is directly under the supervision of UNHCR. In any
20
Practices of corruption and withholding of assistance by maliks are a regular complaint of villagers. This is an observation that
has also been reported in Kantor (2009).
84
case, the transparency of the process is a major requirement to alleviate doubts on behalf of
community members about the selection process, and special attention should be given to
reaching all members of a given community.
d) The degree of involvement of the DoRR
The role of the DoRR in the selection process fluctuated according to provinces, the
relationship of the sub-office with the directorate and the local influence of the DoRR.
In some cases involvement of the DoRR in the BSC is occasional and does not appear to be an
active one. Though IPs and UNHCR insisted a representative of the DoRR was systematically
present in selection, this was not always confirmed in the field and the degree of the influence of
his representative in beneficiary selection varied. In Nangarhar for instance, the sub-office and
DoRR cultivated a complex relationship. The director clearly expressed his frustration with not
being able to voice his opinion, while UNHCR and IPs reported being reluctant to allow him to
get too involved in selection due to suspicions of corruption and confessed making minimal
efforts to include him in the selection process. Similar comments were made in Kandahar, where
the DoRR openly expressed complaints about not being involved at all during selection,
expressing the feeling that he was side-lined on purpose by UNHCR and its IP. In Jawzjan, a
conflictive relation between UNHCR and the DoRR in 2009-2010 resulted in the exclusion of the
representative when UNHCR took the lead in selection. Relations however have improved
notably over the past year. In Balkh, Parwan and Kabul, the DoRR seemed to be only present as a
governmental caution, but without a particular say in the actual selection process. In Hirat,
interestingly, the presence of the DoRR was mentioned as essential to facilitate selection and
curb too much interference of local authorities in the process.
e) Female participation
Notably, women’s participation in the selection process does not seem to have been
implemented uniformly, especially in highly patriarchal communities. Women’s participation
was non-existent in Kandahar, and any mention of women’s participation in Parwan and
Nangarhar were often received with surprise or sarcasm by community members, though IPs did
mention having female employees for the WASH awareness programs. In Hirat, Jawzjan and
Faryab, IPs mentioned employing female staff to reach female-headed households, which cannot
be identified by male staff.
In the community survey, 13 out of the 60 community representatives indicated that women
were participating in the selection process in their respective communities. However, when
asked more specifically for their role, it became clear that local women were not involved in the
process in any instance. It was merely female staff from UNHCR and IPs that came to the
villages, not to assist in beneficiary selection but to inform local women about VRF forms and to
train them on hygiene and maintenance.
This lack of inclusion of women in the selection process had been previously highlighted by the
Danida ROI Evaluation (2012), which stated “the evaluation was concerned about some gender
85
aspects of UNHCR’s shelter programme. (…) It is UNHCR policy to have women representatives in
the shelter beneficiary selection committees. However, the impact of this is not documented,
and basic issues like women’s land rights are not addressed in the UNHCR documents.”
This is a key issue that must be further addressed and developed, by integrating female
representatives more uniformly in the selection process.
f) Community based approach
According to the UNHCR guidelines the selection process is meant to be implemented according
to a community based approach: “the community takes primary responsibility for identifying
eligible beneficiaries to receive shelter assistance, while the Ministry of Refugees and
Repatriation, local authorities, implementing partners and UNHCR play advisory and
coordination roles.”21 In practice however, involvement of the community was seldom
mentioned and was replaced by consultation of community leaders (shuras, maliks). As
mentioned earlier, the single focus on community leaders is not always effective in reaching all
vulnerable members in a community, due to potential interference of nepotism and/or
corruption. In Bez Akmalati, beneficiaries of a UN-Habitat shelter programme emphasized the
importance of inclusion of elected community members, which can be held accountable for the
selection of vulnerable beneficiaries, and praised this practice. Though the UN-Habitat process is
time-consuming and might not be applicable in the context of the UNHCR programme, further
attention is required in including representative members of the targeted community to ensure
fair selection of beneficiaries.
4.2 SELECTION CRITERIA ON THE GROUND: VULNERABILITY SIDE-

LINED
According to the guidelines of the
“Beneficiary selection will be based on vulnerability, the
programme, vulnerability should be the
most important criterion of the selection process (…).
cornerstone of the selection process.
Vulnerable groups are those without stable support from
While the official eligibility for assistance
income earning family members or without sufficient
requires that the beneficiary be a returned
income to meet household demands.”
refugee or IDP, with access to land on
2011 UNHCR Shelter Guidelines which to build a house, the programme is
guided by a focus on vulnerability allowing
beneficiary selection to be wider than just returnees with access to land. In fact, all involved staff
members are advised to ensure that no vulnerable families within the community are overlooked or
rejected for assistance. The vulnerability criterion follows that of the “extremely vulnerable individual”
21
UNHCR Shelter Guidelines 2008, p.8.
86
definition including people who may be in life threatening situations, unable to help themselves, lacking
family and community support or suffering from physical or mental trauma. Typically these include
female-headed households, disabled or elderly heads of households without external support and large
families with insufficient income. Overall, special attention is paid to the relative situation of the
individual within the family and the community in order to identify vulnerable beneficiaries. Moreover,
in the case of landless families in need of shelter and who meet the vulnerability criteria, there is the
possibility of land allocation in order to allow them to benefit from the programme. In sum, while the
programme explicitly targets vulnerable refugee and IDP returnees, the focus is on finding a shelter
solution for any community member which meets the vulnerability criteria. Yet the analysis of the
profiles of beneficiaries surveyed in the frame of this study showed that this priority on vulnerability
enunciated by the guidelines has yet to be operationalized in the field.
4.2.1 Migratory status

Among the UNHCR beneficiaries surveyed for this evaluation, and as indicated in Table 41, the majority,
66.6 per cent, were returning refugee households. Another 19.2 per cent were non-refugee returnees,
while IDPs represent only 9.2 per cent of the sample. The remaining 5.1 per cent are households that
never migrated.
This shows that UNHCR struggles to adapt the SAP to the changes in the migratory trends at play in the
country and is still overwhelmingly focusing on returnees, leaving IDPs aside.
Table 41: UNHCR Beneficiary Categories by Migratory Status
N %
Refugee Returnees 1 355 66.58
Non-refugee Returnees 390 19.16
IDPs 187 9.19
No Mobility 103 5.06
Total 2 035 100.00
87
4.2.2 Vulnerability: the uneven integration of EVIs in the programme
The selection of extremely vulnerable individuals (EVI) varied according to provinces and IPs, and there
seems to be no uniform national practice: in Hirat, the Protection Unit was reportedly systematically
involved in selection in order to identify EVIs, which was not the case in Nangarhar and in Kandahar, for
instance, where EVIs were referred to the Protection Unit but not included in the shelter programme.
This involvement seemed productive in Hirat as IPs and the arbab in Kahdistan for instance, displayed a
higher degree of awareness of the criteria for EVIs. Most of the time, the main criteria regarding EVIs
mentioned by communities and IPs were “widow” and “disabled”, occasionally “very low income”,
though no specific information was given as how to identify them. In some cases, there was recognition
of the need to focus more on vulnerable households and IDPs in Parwan/Kabul (ABR). In such cases,
recent documented voluntary returnees had the priority over more vulnerable households. In Jawzjan,
EVIs were not considered a priority, and were only identified and considered for potential additional
assistance a posteriori, whereas in Faryab, IPs did not appear to have been sensitized to EVIs. Clearer and
more uniform instructions as well as more flexibility and overview by UNHCR staff might be a good way
to ensure more fairness in selection.
Table 42 shows that among our sample, more than half of the UNHCR beneficiaries are not EVI
households. Only 46.4 per cent can be considered as such, even though criteria to define EVIs are loose.
This indicates a clear failure to target the most vulnerable.
Even more worrisome is the fact, that more than half of the non-refugee returnees that received
UNHCR assistance are not considered to be an EVI. This represents a misallocation of assistance as this
group is not addressed in the first place and does also not present the signs of vulnerability that would
make them eligible. The same is true for the no mobility group that is not extremely vulnerable, but did
receive UNHCR shelter assistance. While they represent only 2 per cent of all UNHCR beneficiaries
surveyed for the evaluation, this is a clear indication that misallocation of assistance does occur.
Table 42: EVI Status of UNHCR Beneficiaries
Not EVI EVI Total
Refugee Returnee 747 608 1 355

% 55.13 44.87 100.00
Non-Refugee Returnee 205 185 390
% 52.56 47.44 100.00
IDP 105 82 187
% 56.15 43.85 100.00
No Mobility 33 70 103
% 32.04 67.96 100.00
Total 1 090 945 2 035
% 53.56 46.44 100.00
88
Table 43 however shows differences in coverage of EVIs across provinces. Extremely vulnerable
households in Bamyan, Kunduz and Kandahar for example are included more than non-vulnerable
households, while just the opposite is the case in provinces like Parwan, Nangarhar, Balkh, Jawzjan,
Takhar and Paktya.
Table 43: EVI Status by Province of UNHCR Beneficiaries
No EVI EVI Total
Kabul 98 99 197
% 49.75 50.25 100.00
Parwan 60 41 101
% 59.41 40.59 100.00
Bamyan 11 21 32
% 34.38 65.63 100.00
Laghman 77 85 162
% 47.53 52.47 100.00
% 56.46 43.54 100.00
Balkh 28 22 50
% 56.00 44.00 100.00
Faryab 38 37 75
% 50.67 49.33 100.00
Jawzjan 76 42 118
% 64.41 35.59 100.00
Sari Pul 29 27 56
% 51.79 48.21 100.00
Kunduz 27 33 60
% 45.00 55.00 100.00
Takhar 23 16 39
% 58.97 41.03 100.00
Helmand 27 29 56
% 48.21 51.79 100.00
Kandahar 28 47 75
% 37.33 62.67 100.00
Paktya 71 52 123
% 57.72 42.28 100.00
Hirat 51 50 101
% 50.50 49.50 100.00
Total 1 090 945 2 035
% 53.56 46.44 100.00
89
90
 Excluding EVI households
Consequently, there are also households that do fall into the EVI categories but did not receive shelter
assistance.
Table 44 shows that significant shares of households defined as EVI were in fact not addressed by any
shelter assistance program. For example, 43.3 per cent of households with a chronically ill member were
not included in either UNHCR or other programs, while 45.4 per cent of households with very low
income were excluded.
Table 44: Beneficiary Status of EVI Households

Total
Beneficiary Beneficiary Beneficiary
N=4 488
N=2 035 N=463 N=1 990
Chronically ill 465 122 448 1 035
% 44.93 11.79 43.29 100.00
Very low income 319 44 302 665
% 47.97 6.62 45.41 100.00
Physically Disabled 274 86 271 631
% 43.42 13.63 42.95 100.00
Mentally disabled 134 31 118 283
% 47.35 10.95 41.70 100.00
Large family (5 or more children
141 10 119 270
and no livelihoods)
52.22 3.70 44.07 100.00
%
Unaccompanied Elderly (over 60) 56 16 44 116
% 48.28 13.79 37.93 100.00
Elderly-Headed Household 72 8 70 150
% 48.00 5.33 46.67 100.00
Female Head of household 38 22 56 116
% 32.76 18.97 48.28 100.00
Unaccompanied minor (under 18) 19 4 18 41
% 46.34 9.76 43.90 100.00
Single Parent 11 1 15 27
% 40.74 3.70 55.56 100.00
Drug addict 6 1 12 19
% 31.58 5.26 63.16 100.00
Child-Headed Household 1 0 3 4
% 15.00 0.00 75.00 100.00
91
Gender-based violence survivor 2 0 1 3
% 66.67 0.00 33.33 100.00
Of these household defined as extremely vulnerable yet were excluded from any program, a substantial
portion were also refugee returnees, as shown in Table 45, providing evidence of flaws in the selection
process. For the most common EVI causes, like chronic illnesses, disability, low income and large families,
around one third of non-beneficiary households were officially recognized refugee returnees.
Table 45: Migratory Status of Non-Beneficiary EVI Households
Refugee Non-refugee
IDPs No Mobility Total
Returnees Returnees
N=219 N=368 N=1990
N=727 N=676
% 31.92 31.70 13.62 22.77 100.00
Very low income 91 110 40 61 302
% 30.13 36.42 13.25 20.20 100.00
Physically Disabled 98 86 34 53 271
% 36.16 31.73 12.55 19.56 100.00
Large family (5 or more children
43 32 12 32 119
and no livelihoods)
36.13 26.89 10.08 26.89 100.00
%
Mentally disabled 44 41 13 20 118
% 37.29 34.75 11.02 16.95 100.00
% 37.29 34.75 11.02 16.95 100.00
Elderly-Headed Household 23 15 10 22 70
% 32.86 21.43 14.29 31.43 100.00
Female Head of household 10 21 7 18 56
% 17.86 36.50 12.50 32.14 100.00
Unaccompanied Elderly (over 60) 10 21 4 9 44
% 22.73 47.73 9.09 20.45 100.00
Unaccompanied minor (under 18) 6 4 1 7 18
% 33.33 22.22 5.56 38.89 100.00
Single Parent 7 3 3 2 15
% 46.67 20.00 20.00 13.33 100.00
Drug addict 3 4 1 4 12
% 25.00 33.33 8.33 33.33 100.00
Child-Headed Household 2 1 0 0 3
% 66.67 33.33 0.00 0.00 100.00
92
Gender-based violence survivor 0 1 0 0 1
% 0.00 100.00 0.00 0.00 100.00
The exclusion of EVI households is particularly a problem in those cases where the household did in fact
apply for shelter assistance, but was not chosen to receive it. Table 46 shows households categorized by
EVI categories and migration status that did apply for shelter assistance but were not chosen. In the
sample the largest vulnerable groups that were denied shelter assistance are physically disabled,
chronically ill and very low-income households.
Added attention to health and disability – as two key protection concerns – would allow UNHCR to target
vulnerable households that currently fall outside of the reach of the programme.
Of the 13 EVI categories used by UNHCR in its 2009-2011 programming:
 Vulnerabilities relating to health and disability (whether chronically ill households, physically
disabled, and the mentally disabled) were 3 of the top 4 vulnerabilities most often disregarded in
the selection process.
 Vulnerabilities relating to socio-economic household profiles – such as very low income and large
households – ranked second in terms of numbers.
 Demographic characteristics (targeting children, the elderly and women) were given priority.
Table 46: Failed Applications for Shelter Assistance by EVI Category and Migratory Status
Refugee Non-refugee No
IDPs Total
Returnees Returnees Mobility
Physically disabled 42 40 15 11 108
Very low income 37 50 9 11 97
Large family 16 12 2 6 36
Unaccompanied elderly 7 10 2 3 22
Elderly household head 10 5 4 2 21
Female household head 1 7 3 4 15
Single parent 4 2 1 1 8
Unaccompanied minor 2 2 0 2 6
93
Drug addict 1 2 0 1 4
Child HH head 1 1 0 0 2
Gender-based violence
0 1 0 0 1
survivor
All of the above shows that the selection process was not focused on the vulnerability of beneficiaries. In
reality the research showed that the main criteria of selection of beneficiaries used throughout the
country was the presentation of a Voluntary Repatriation Form (VRF), to the extent that in some cases,
holding a VRF was the only criterion mentioned by communities as effectively implemented in
selection, alongside with the requirement of land ownership. This was also indicated in the community
survey, where the representatives of 60.0 per cent of the communities indicated that the VRF was a
criterion for beneficiary selection. Findings presented in
Table 47 confirm observations from the field, which showed that awareness about the criteria to define
vulnerability was very low, and that only the criteria ‘widow/female-headed households’ and ‘very low
income’ were understood as signs of vulnerability, which were reported as criteria of selection by about
40 per cent of surveyed community representatives. Although these are clear categories of vulnerable
households, they should not be given more weight or precedence over otherwise eligible EVI
households.
Table 47: Used Selection Criteria (Community Representatives)

N=60 %
VRF 36 60.00
Families with very low or unstable income 26 43.33
Female household head 25 41.67
Disabled individual 18 30.00
Other 9 15.00
Large families of eight or more members 4 6.67
Chronically ill individual 1 1.67
Elderly household head 0 0.00
Underage household head 0 0.00
94
Another indication of the focus on the VRF form in the selection process for the UNHCR programme is
presented in
Table 48. It shows that of all returnees in the sample 74.4 per cent had a VRF form. The percentage of
those receiving UNHCR assistance is significantly higher than this average with 86.8 per cent. In contrast,
non-beneficiaries only have a VRF form in 58.3 per cent of the cases.
Table 48: VRF Form and Shelter Assistance

Total
N % N % N % N %
Yes 1 509 86.77 286 75.86 817 58.27 2 654 74.38
No 214 12.31 90 23.87 561 40.01 873 24.47
I don’t know 16 0.92 1 0.27 24 1.71 41 1.15
Total 1 739 100.00 463 100.00 1 402 100.00 3 568 100.00
UNHCR staff and IPs reported prioritizing recent returnees, based on the assumption that more ancient
returnees had had the time to install coping mechanisms, especially in terms of shelter, and were
therefore less vulnerable than the rest.
The date of return (written down on the VRF) was also mentioned as an additional means for selection,
with recurrent complaints of non-beneficiaries not having been considered eligible because their VRF
had “expired” (Nangarhar, Parwan, Jawzjan). There seems to be no particular pattern for selection in
these cases as detailed in
Table 4922. Quantitative data shows that 43 per cent of refugee returnee beneficiaries had received
shelter assistance within a year after their return. But a significant proportion of refugee returnee
beneficiaries, 28 per cent, had received shelter assistance more than three years after their return. The
assumption that the longer returned have had the time to find their own shelter was not always verified
in the field, with older returnees mentioning living with relatives or getting increasingly indebted with
rent with similar and sometimes greater needs than the actual beneficiaries.
Table 49: Time between Return/Arrival and Selection into Programme (UNHCR Beneficiaries)
Refugee Returnees Non-refugee Returnees IDPs Total
Less than a month 47 13 0 74
% 3.47 3.33 0.00 3.11
1 to 6 months 374 103 55 638
% 27.64 26.41 29.57 27.58
6 months to 1 year 167 39 22 228
% 12.34 10.00 11.83 11.82
22
This has also been reported in the DANIDA Report 2012.
95
1 to 3 years 389 80 50 519
% 28.75 20.51 26.88 26.91
3 to 5 years 172 50 19 241
% 12.71 12.82 10.22 12.49
More than five years 204 105 40 349
% 15.08 26.92 21.51 18.09
Total 1353 390 186 1929
% 100.00 100.00 100.00 100.00
UNHCR beneficiaries were more likely to have returned in more recent years, representing the largest
group of returnees from the 2009-2011 timeframe. Although 22.2 per cent were selected from the 2002-
2004 period, this is significantly less than in other programs, 33.2 per cent, or among non-beneficiaries,
33.7 per cent. In effect, it appears preference was given in UNHCR’s programme to the reintegration of
more recent returnees.
Table 50: Time of Return

Total
Before 2002 0 0 2 2
% 0.00 0.00 0.14 0.06
2002-2004 387 125 473 985
% 22.19 33.16 33.71 27.95
2004-2008 710 216 521 1 447
% 40.71 57.29 37.13 41.06
2009-2011 647 36 407 1 090
% 37.10 9.55 29.01 30.93
Total 1 744 377 1 403 3 524
% 100.00 100.00 100.00 100.00
The assumption among stakeholders is that protracted and new caseloads should be dissociated, with
duration of displacement becoming a criterion of selection for interventions. However, data from this
study shows that the rates of EVIs do not decrease with the duration of displacement. Among UNHCR
beneficiaries indicated in Table 51, EVIs are systematically right below the 50 per cent mark regardless of
whether they were displaced in 2002, 2004 or 2009. As a result, this data draws attention to the fact that
vulnerability, and not the timing of return, should be a key determinant in the selection process.
Table 51: Time of Return by EVI Status (UNHCR Beneficiaries)
96
Not EVI EVI Total
2002-2004 202 185 387

% 42.2 47.8 100.00
2004-2008 398 312 710
% 56.06 43.94 100.00
2009-2011 352 295 647
% 54.4 45.6 100.00
Total 952 792 1 744
% 54.59 45.41 100.00
The data on the timing of displacement matched with the timing of return highlights what has been
shown before: a preference in the selection process for returned refugees displaced prior to 2001. More
recent waves of conflict and displacement, specifically impacting IDPs, were only minimally captured in
the sample.
The overwhelming majority, 91.0 per cent, of beneficiaries were those displaced prior to 2001, as
illustrated in
Table 52.
Table 52: Time of Displacement and Time of Return (UNHCR Beneficiaries)
Before 2001 2001-2004 2005-2009 After 2009 Total
2002-2004 369 17 1 0 387

% 21.23 0.98 0.06 0.00 22.27
2004-2008 641 51 14 2 708
% 36.88 2.93 0.81 0.12 40.74
2009-2011 572 36 34 1 643
% 32.91 2.07 1.96 0.06 37.00
Total 1 582 104 49 3 1 738
% 91.02 5.98 2.82 0.17 100.00
97
4.3 TARGETING OF THE MOST VULNERABLE: MAIN CATEGORIES LEFT
OUT
The overreliance on the VRF as the main basis for selection has in certain cases led to the under-
representation of particularly vulnerable displaced households. Past research has shown that “the
vulnerable within the vulnerable” are IDPs, landless households, female-headed households and those
households with ill or disabled members who represent a growing humanitarian concern in the country.
We similarly find evidence that these groups are relatively more deprived using a Multi-dimensional
Poverty Index which incorporates a range of socio-economic indicators and whose construction is
explained further in Chapter 5. Each of these groups will be discussed in this section as main categories
of vulnerable and eligible beneficiaries.
4.3.1 IDPs
A fact commonly acknowledged by all stakeholders was that IDPs were underrepresented in the
selection process, or only included in small proportions: just above 11 per cent of UNHCR beneficiaries
between 2009 and 2011 were IDPs23. In our own sample, just above 9 per cent of UNHCR beneficiaries
were IDPs, as presented in Table 53. It’s important to note however that the selection of IDPs is biased
from the start, given locations for assistance were chosen for the presence of refugee returnees. This is
evident by the fact that IDPs are only 11 per cent of all non-beneficiaries as indicated in the table below.
As such, UNHCR may look to identify where IDPs have a high presence and include those locations in
future assistance programs so as to better cover this increasingly important group.
Table 53: IDP within each category

Total
Non-IDP 1 848 454 1 771 4 073

% 90.81 98.06 88.99 90.75
IDP 187 9 219 415
% 9.19 1.94 11.01 9.25
Total 2 035 463 1 990 4 488
% 100.00 100.00 100.00 100.00
This is a serious cause for concern as IDPs have been identified as one of the most vulnerable segments
of the population and are increasingly considered as a major humanitarian priority in Afghanistan
today24. Table 54 below supports this claim, showing that within our sample 83.6 per cent of IDP
households are deprived on a range of socio-economic indicators compared to 77.2 per cent of non-IDP
households.
23
UNHCR Assisted IDP families: 2009, 2010, 2011”, UNHCR Data Unit, Kabul.
24
UNHCR, The World Bank « Research on IDPs in Urban Settings », May 2011,,Kabul. NRC/JIPS/IDMC/Samuel Hall, Challenges of
IDP Protection, 2012, Kabul.
98
Table 54: Multi-dimensional Poverty Index of IDPs
Non-IDPs IDPs Total
Not Deprived 930 68 998

% 22.83 16.39 22.24
Deprived 3 143 347 3 490
% 77.17 83.61 77.76
Total 4 073 415 4 488
% 100.00 100.00 100.00
Across regions, the Southern provinces were a notable exception with a high number of inter-district
conflict-induced IDP beneficiaries provided with one-room shelters as indicated in Table 55. This was the
result of an active policy towards IDPs initiated by the sub-office and supported by the central office.
Table 55: UNHCR Assisted IDP Households by Region25

2009 2010 2011 Total
Central 20 44 26 90
Central Highlands - 70 34 104
East - - 214 214
North 171 652 302 1 125
Northeast 48 194 100 342
South - 623 950 1 573
Southeast - - - -
West 65 327 221 613
Total 304 1 910 1 847 4 061
25
From UNHCR Data Unit, Kabul.
99
Table 56 shows that within our own sample, we find clear targeting of IDP households for shelter
assistance in Takhar and Helmand, 92.3 per cent and 87.5 per cent respectively. Moreover, selection of
IDPs was noticeable in Faryab, Hirat and Kandahar. However in all other provinces this clear targeting of
IDPs did not take place.
100
Table 56: IDP UNHCR Beneficiaries by Province
Non-IDPs IDPs Total
Kabul 197 0 197

% 100.00 0.00 100.00
Parwan 89 12 101
% 88.12 11.88 100.00
Bamyan 32 0 32
% 100.00 0.00 100.00
Laghman 161 1 162
% 99.38 0.62 100.00
% 98.61 1.39 100.00
Balkh 45 5 50
% 90.00 10.00 100.00
Faryab 49 26 75
% 65.33 34.67 100.00
Jawzjan 113 5 118
% 95.76 4.24 100.00
Sari Pul 56 0 56
% 100.00 0.00 100.00
Kunduz 54 6 60
% 90.00 10.00 100.00
Takhar 3 36 39
% 7.69 92.31 100.00
Helmand 7 49 56
% 12.50 87.50 100.00
Kandahar 61 14 75
% 81.33 18.67 100.00
Paktya 122 1 123
% 99.19 0.81 100.00
Hirat 80 21 101
% 79.21 20.79 100.00
Total 1 848 187 2 035
% 90.81 9.19 100.00
In the case of IDPs, securing land tenure is highly problematic and was identified as one of the main
problems in their inclusion in the program, alongside their difficult identification due to lack of proper
documentation. In our sample, Table 57 shows that of UNHCR beneficiaries, IDP households were the
least likely, in relation to other groups, to own the land which their shelter was built on prior to
assistance.
101
Table 57: Land Ownership before Building Shelter (UNHCR Beneficiaries)
Yes No Total
Refugee Returnees 1 165 190 1 355

% 85.98 14.02 100.00
Non-refugee Returnees 306 84 390
% 78.46 21.54 100.00
IDP Returnees 126 60 186
% 67.74 32.26 100.00
% 87.38 12.62 100.00
Total 1 687 347 2 034
% 82.94 17.06 100.00
In some cases where they were provided with camp documents (“IDP VRFs” in Kandahar and Hirat), IDPs
were included in shelter assistance, though they remained a small proportion of beneficiaries as
compared to returnees. Moreover, their access to assistance in general is highly dependent on the
provincial context, and was mentioned as challenging given the high political nature of displacement.
With the rising importance of internal displacement in the Afghan context, this under-representation is
highly disquieting, putting into question the relevance of the programme.
The relatively low inclusion of IDPs in the program compared to returnees is one of the main
challenges of the shelter program, requiring specific attention to identify potential solutions as internal
displacement and the incapacity of most IDPs to return to their province of origin is an increasingly
important reality in Afghanistan. Not sufficiently taking them into account leaves out substantive
segments of populations in need of shelter. This was underlined as a serious concern by community
members (e.g. in Kanaqa in Parwan), reporting those who were not able to migrate abroad in the first
place were also those with fewer resources and that returnees from Iran and Pakistan returned with
more skills and were more able to secure a livelihood than returning IDPs. What is perceived as a
discriminating selection can turn into a potential driver of tension, with returning IDPs insisting they had
stayed in their country during the war, fought and lost their assets, and not received any form of
assistance, whereas better-off returnees were getting all the assistance.
Nonetheless, of those few IDP households which were assisted by the UNHCR programme there is little
evidence of difficulties in integrating into the local community.
102
Table 58 shows the overwhelming majority of IDP households have been positively received by fellow
households.
103
Table 58: IDP Integration (UNHCR Households)
Community Behaviour N %
Very welcoming – very supportive 70 37.63
Welcoming – supportive 113 60.75
Not welcoming – not supportive 3 1.61
Total 195 100.00
4.3.2 Landless Vulnerable Households

Similar to IDPs, those households who did not own land were underrepresented in the selection process.
This is not surprising given land ownership is in most cases a requirement for receiving shelter assistance,
notably in order to avoid later tensions around land tenure, a highly sensitive issue in Afghanistan. As
such, only a small percentage of UNHCR beneficiaries, 17.1 per cent, did not own their land before
becoming a beneficiary. This is in contrast to other programs, where 27.2 per cent of households were
landless before assistance.
Table 59: Land Ownership prior to Shelter Assistance
UNHCR Non-UNHCR
Total
Yes 1 687 337 2 024

% 82.94 72.79 81.06
No 347 126 473
% 17.06 27.21 18.94
Total 2 034 463 2 497
% 100.00 100.00 100.00
Similar to IDPs, those households without land are on the whole more likely to be deprived across a
range of socio-economic indicators.
104
Table 60 shows that 82.9 per cent landless households within our sample are deprived compared to 71.0
per cent of land owning households.
105
Table 60: Multi-dimensional Poverty Index for Landless Households
Landless
No Yes Total
% 28.98 17.12 22.24
Deprived 1,375 2,115 3 490
% 71.02 82.88 77.76
Total 1 936 2 552 4 488
% 100.00 100.00 100.00
The shelter guidelines identify four solutions for providing shelter to landless households: 1) allocation of
community or village public land; 2) allocation of private land; 3) allocation of land through the LAS and
4) other land allocation identified as viable and applicable by the BSC26. The fieldwork showed that none
of these solutions seemed to have been applied much in the field, aside from LAS in LAS townships,
which is also supported by Table 62. Neither UNHCR or IPs mentioned trying to resolve the issue of
landlessness for vulnerable families through the suggested solutions which was particularly sensible in
areas where such types of solutions had been implemented by other organizations (UN-Habitat in Bez
Akmalati, Nangarhar for instance). Compensating the absence of land documents through confirmation
of land ownership by the local shura was the only active procedure mentioned.
Community representative also only mentioned that the proposed solutions were applied in a very
limited number of cases. In 27 of the 60 communities vulnerable landless households applied to the
shelter assistance programme. Less than half of the communities were able to provide shelter assistance
to those without land ownership as can be seen in Table 61.
Table 61: Land Allocation in Communities
N %
No means 14 53.85
Allocation of public land 4 15.38
Allocation of private land 3 11.54
Other means 3 11.54
Land Allocation Scheme 2 7.69
Total 26 100.00
26
UNHC Shelter Guidelines, p. 8.
106
Beneficiary households had either inherited land or bought it, either collectively (land purchased by an
entire tribe) or individually. In the case of the latter, there are doubts that these households represent
the most vulnerable. However, there were a small percentage of landless households that received land
in order to be able to build their shelter. As Table 62 shows the most commonly applied solution
reported by beneficiaries in the field was the Land Allocation Scheme.
Table 62: Land Allocation Solutions
UNHCR Non-UNHCR
Total
Land Allocation Scheme 140 104 244

% 40.35 82.54 51.59
Other 112 9 121
% 32.28 7.14 25.58
Family/ land of relatives 50 10 60
% 14.41 7.94 12.68
Community or village 35 3 38
% 10.09 2.38 8.03
Rental land 10 0 10
% 2.88 0.00 2.11
Total 347 126 473
% 100.00 100.00 100.00
However, given the low level of prior explanation of criteria to community members, landless vulnerable
families were not brought to the attention of the BSC and were in many cases de facto excluded from
the programme. No occurrences of community assistance to landless people by providing private or
communal land in order for them to receive shelter were observed on the field.
These findings show that, even more than other shelter programmes implemented in the country,
UNHCR SAP is leaving out one of the most vulnerable segments of the displaced population and that no
functioning mechanism is in place to try to integrate this population.
107
4.3.3 Female-headed households
The third group of interest is that of female-headed households. In the data collected for the purpose of
this study, there were a total of 116 female headed households. Of this total, the majority, 79.3 per cent,
are widows while 15.5 per cent are married, as shown in Table 63.
Table 63: Marital Status of Female-Headed Households
N %
Widow 92 79.31
Married 18 15.52
Single 4 3.45
Engaged 1 0.86
Divorced 1 0.86
Total 116 100.00
Moreover, Table 64 shows that of all households assisted by UNHCR only 1.9 per cent are female-
headed households in comparison to the 4.8 per cent assisted by other organizations.
Table 64: Beneficiary Status of Female-Headed Households

Total
Yes 38 22 56 116
% 1.87 4.75 2.81 2.58
No 1 997 441 1 934 4 372
% 98.13 95.25 97.19 97.42
Total 2 035 463 1 990 4 488
% 100.00 100.00 100.00 100.00
The underrepresentation of female-headed households in the SAP is particularly worrisome when taking
into consideration how vulnerable they are in comparison to male-headed households. Again using the
Multi-dimensional Poverty Index, we see that 95.7 per cent of female-headed households within our
sample are deprived across a range of socio-economic indicators compared to only 77.3 per cent of
male-headed households.
108
Table 65: Multi-dimensional Poverty Index of Female-Headed Households
Female-Headed Household
No Yes Total
% 22.71 4.31 22.24
Deprived 3 379 111 3 490
% 77.29 95.69 77.76
Total 4 372 116 4 488
% 100.00 100.00 100.00
Moreover, our research shows that allocating assistance to widows was well understood in most
communities. However, they were in most cases the only women to be clearly included in the
programme as EVIs.
The fact that women were not included in the beneficiary selection in several areas raises concern
about the effective access to female-headed households, but also about the assessment of the living
conditions of potentially eligible families, since, as it was noticed in the field for our own staff, only
women are allowed to enter private areas. Aside from including a female IP staff in the selection process
in some of the provinces, none of the procedures mentioned in the guidelines for inclusion of females in
selection have been mentioned either by sub-offices, IPs or community members.
4.3.4 Households with Ill or Disabled Members

Finally, households with a member who is physically, mentally or chronically ill are also considered
extremely vulnerable. As indicated in Table 66 we find that within our sample 35.3 per cent of
households of this type are UNHCR beneficiaries compared to the 37.8 per cent from other
organizations.
Table 66: Households with Ill or Disabled Members

Total
No 1 316 288 1 305 2 909

% 64.67 62.20 65.58 64.82
Yes 719 175 685 1 579
% 35.33 37.80 34.42 35.18
Total 2 035 463 1 990 4 488
% 100.00 100.00 100.00 100.00
109
While the representation of households with ill or disabled members is on par with the entire sample,
they are more likely to be vulnerable than the average shown again by the Multi-dimensional Poverty
Index in
Table 67. Here we find that 82.0 per cent of these households are deprived across a range of socio-
economic indicators in comparison to the 75.5 per cent which do not have an ill or disabled member.
Table 67: Multi-dimensional Poverty Index for Households with Ill or Disables Members
Ill or Disabled
No Yes Total
% 24.51 18.05 22.24
Deprived 2 196 1 294 3 490
% 75.49 81.95 77.76
Total 2 909 1 579 4 488
% 100.00 100.00 100.00
4.4 FACTORS EXPLAINING THE WEAKNESSES OF THE SELECTION

PROCESS
Qualitative research helped identify some of the key factors weakening the robustness of the SAP’s
selection process:
1. Lack of awareness about the criteria of selection: There is a clear misunderstanding of the criteria
for selection on behalf of communities, and worryingly, IPs also emphasized the importance of the
VRF in selection. There is evidence that vulnerability was sometimes only considered as secondary
criteria, behind other guidelines of the programme. Only in rare cases was vulnerability mentioned
by both communities and IPs as playing a role in the selection process. UNHCR field staff themselves
were only partially aware of the fact that vulnerability should be central in the selection process. A
recurrent observation was that documented recent returnees were entitled to receive a shelter
regardless of their degree of vulnerability. The high priority of the VRF can further be explained by
the fact that it simplifies identification of returnees, with less subsequent effort put into assessing
the effective needs of potential beneficiaries.
Moreover, the often weak and fragmented awareness about the criteria for selection may have
potential repercussions for intra- or inter-community tensions (see Section 5). Misunderstanding of
the criteria, and the feeling that the programme discriminates against vulnerable households was
seen as sometimes leading to resentment on behalf of non-beneficiaries and is thus a potential
driver of conflict. This indicates the lack of clear communication about the criteria utilized for
assistance prior to the selection process.
110
2. Supply-driven process of selection: One of the main challenges to the effective selection of
vulnerable households was the fact that selection of areas of implementation was more supply-
driven rather than being based on a needs assessment. In cases where high return areas where
identified and allocated a high number of shelters, all VRF holders in a tribe indiscriminately received
assistance. This raises the issue of the quality of the preliminary assessment of needs done by
UNHCR. There were recurrent problems observed in Kabul, Hirat and Nangarhar, with quotas of
shelters having to be met whatsoever. This has a considerable impact on selection, as in such cases
shelters are distributed to less vulnerable households based on the VRF criteria, whereas
neighbouring communities are completely left out of the programme, fuelling potential tensions as
the case of Kuchi Abad in Kabul province clearly shows.
3. Quick selection process: There were conflicting comments about the time-span of the selection
process and no reliable information could be gathered about the effective duration of selection. In
several cases, the short amount of time effectively spent in the village for selection of beneficiaries
raised problems in targeting vulnerable households. In Nangarhar, cases were mentioned where
potentially eligible people could not take off work during selection and were therefore excluded
from the process. In areas where various programmes were implemented, beneficiaries compared
the selection processes of different organizations and deemed the UN-Habitat process fairer as it
was more focused on community participation.
4. Weakness of UNHCR selection tool: A possible explanation for the insufficient attention paid to
vulnerability in selection is the fact that the “Beneficiary Social Verification Check List” provided by
UNHCR to its IPs in order to identify eligible households uses broad and vague categories
inappropriate to assess the effective degree of vulnerability of a household. Stronger guidelines
about vulnerability and the inclusion of under-represented categories should come both from the
central level and from the active involvement of the protection unit.
5. Unclear connection between the SAP and UNHCR EVI project: Discussions with IPs and field staff
showed that there was confusion between the focus on vulnerability articulated in the guidelines of
the SAP and the extra assistance received by extremely vulnerable beneficiaries in the frame of
UNHCR’s EVI project in place since 2002 which gives small cash grants to EVIs. For example, Hirat and
Nangarhar sub-offices underlined the fact that imposed quotas of assistance to EVIs (5-10%) were
considered as a glass ceiling above which they could not provide further assistance. The overall
understanding of EVI categories as a framework for selection is therefore lacking.
111
5. SOCIO-ECONOMIC IMPACT OF THE SHELTER ASSISTANCE
PROGRAMME
This section presents the multi-dimensional poverty analysis used to assess the impact of SAP:
o Over three quarters, 78 per cent, of the overall sample are multi-dimensionally poor – in
education, economic well-being, social capital, housing and health.
o The highest rates of poverty are seen among non-shelter beneficiaries. 86 per cent of non-
shelter beneficiaries are multi-dimensionally poor, as opposed to 71 per cent of UNHCR
shelter beneficiaries and 68 per cent of non-UNHCR shelter beneficiaries. Other shelter
agencies have a lower poverty rate than UNHCR beneficiaries in the East.
o IDPs are the most deprived group among beneficiaries. Supporting the idea that IDPs are
‘the most vulnerable within the vulnerable’, within our sample they have the highest overall
rate of poverty at 81 per cent of all UNHCR beneficiaries. Refugee returnees are the least
deprived, receiving the core of the assistance.
Utilizing a cross-sectional regression analysis to measure differences in probability of deprivation across

groups, our results find a correlation between lower rates of poverty and inclusion in the UNHCR
shelter programme. Households are more likely to be poor if they are non-beneficiaries, as compared to
the reference groups UNHCR beneficiary and other shelter beneficiary.
While the results of the cross-sectional regression analysis are suggestive, they do not indicate causation
of deprivation since the most vulnerable beneficiaries were often left out of the beneficiary selection
(see Chapter 4). Those households chosen to be UNHCR beneficiaries or non-UNHCR beneficiaries may
have been selected precisely because they were less deprived.
However using a difference in difference analysis looking at differences across groups and across time –
after return from abroad and after receiving assistance – we find solid evidence that the UNHCR shelter
assistance programme had a significant and positive impact on reducing household deprivation along
certain indicators of interest including access to a house, electricity, sanitation and access to a mobile.
113
5.1 IMPACT ON BENEFICIARY HOUSEHOLDS
5.1.1 Preliminary Observations

Beneficiaries by-and-large perceived the impact of the shelter assistance on their economic situation as
positive. In the overwhelming majority of cases, beneficiaries were satisfied with their participation in
the programme, first and foremost because it addressed a priority need in their lives and second, as they
perceived it had significantly improved their living conditions and the socio-economic situation of their
household. UNHCR and non-UNHCR shelter beneficiaries alike assessed a positive impact of the shelter
programme on their well-being, highlighting that the type of aid offered – i.e. shelter assistance – was
best adapted to their needs upon return.
As indicated in Table 68, half of UNHCR beneficiaries indicated that participation in the shelter
programme improved their economic situation. About 6 per cent of the UNHCR beneficiaries indicated
that it was far better than before. A quarter of the respondents did not notice any change in their
economic situation, while around 18 per cent believed their economic situation to be worse or far worse
following assistance. The perception of community leaders is even more positive, with 98 per cent
indicating that the shelter assistance programme improved the situation of the beneficiaries. Only one
respondent denied such a positive impact.
Table 68: Subjective Impact of Shelter Assistance on Economic Situation of the Household (in %)
UNHCR Non-UNHCR
Total
Far worse 3.52 2.85 3.41
Worse 14.72 15.28 14.81
Same 25.65 27.98 26.04
Better 50.00 50.26 50.04
Far better 6.11 3.63 5.70
Total 100.00 100.00 100.00
In all the provinces visited, beneficiaries saw the programme as one of the most important support
systems they could receive upon return or displacement and did not doubt that their participation in the
programme had positively impacted their household, often significantly.
There are several aspects by which the programme improved the socio-economic situation of
beneficiaries:
114
i. Improvement of housing conditions
The first obvious effect of the programme on households was the direct improvement of
beneficiaries’ living conditions thanks to the access to a durable building to reside in. This is far
from a trivial impact, as their former housing conditions were often very precarious. Table 69
shows that UNHCR beneficiary households moved into single family houses either by themselves
or with other families. This is the case for households in rural as well as urban households. SAP
particularly provided households in semi-rural areas with the means to move out of temporary
and shared housing arrangements and into their own single family house. Beneficiaries also
share their new living space with other households.
Table 69: Percentage Change in Housing Arrangement Before and After Assistance by Location
Urban Semi- Rural Total
rural
(N=292) (N=404) (N=1234) (N=1930)
We own a single family house 24.22 126.16 89.23 81.22
We own a house that we share with other households 20.00 175.00 45.83 52.63
We own a single family apartment 0.00 - 100.00 50.00
We rent a house that we share with other households -100.00 -100.00 -76.47 -90.70
We rent a single family house -100.00 -95.92 -94.64 -95.87
We live with family -100.00 -97.50 -97.38 -97.52
We live in a temporary shelter (shack; tent) -91.67 -99.13 -97.40 -97.73
We live with friends -87.50 -100.00 -100.00 -98.11
Other - -100.00 -100.00 -100.00
I don't know - - -100.00 -100.00
Qualitative fieldwork showed without ambiguity the importance of getting a durable house in
the life of beneficiaries. This is linked to the type of housing that the beneficiaries had prior to
their participation in the programme. The qualitative fieldwork identified the following scenario:
 In a significant number of cases, the programme enabled a transition from tent and temporary
shelters to durable shelters. Field observations showed that some specific segments of the
returnee and IDP populations were more likely than others to be living in tents before receiving
their shelters. This was often the case for:
a) Former nomadic tribes like Kuchi or Arab tribes as for example in Aâb Dara in
Kabul province, in Saracha in Nangarhar or in Kandahar province;
115
b) New settlements where no housing structures pre-existed to host new influx of
populations like in Shogofan 1 in Hirat province;
c) In villages where the entire population had had to flee during the civil war or the
Taliban regime and found their village destroyed upon return as for example in
Shobash Khorde Turkmenia in Jawzjan or in Parwan province, which was a front
line during the conflict. Houses had been destroyed directly by fighting or simply
by prolonged period of absence.
d) IDP caseloads.
 Another typical situation observed in the field was returning households being hosted by
relatives. In these cases, families had to share small and over-crowded living spaces in the
original compound of their ancestors. Often, brothers would have got married while in Pakistan
and would have to share the unique compound of their ancestors with their siblings and their
families when they came back to their place of origin. This explains why a lot of the shelters
visited were built within familial compounds, next to a pre-existing house or the instances where
multiple shelters were built in one familial compound.
 A third common situation was that of returnee or IDP families renting their house before getting
shelter assistance. Perhaps contrarily to other contexts, renting a house is an issue in a country
like Afghanistan characterised by a very strong under-employment and irregular incomes.
Renting a house reinforces the risk of indebtedness and is in itself a form of socio-economic
insecurity for rural households.
This means that the shelter programme plays an important role in the protection of displaced
populations who avoid the multiple risks related to precarious housing conditions. The impact
of the programme in terms of protection is multifaceted:
 Firstly, a durable shelter is a real improvement in terms of protection of returnee and IDP
families in an environment where protection risks are numerous and reinforced by
displacements27. This came back in every focus group discussion with beneficiaries. Several
threats to protection arise from the fact of living in temporary shelters. Those include an
increased vulnerability towards the harsh weather conditions existing in the country.
Temporary shelters leave children and pregnant women in very precarious situations. This is
not a negligible protection risk as witnessed by the 2012 harsh winter which claimed the
lives of at least 30 children in the IDP informal settlements of Kabul28.
 The absence of permanent shelter during winter is a driver for secondary migrations from
the North of the country to the East and the Southeast, as it was mentioned in Aad Dara. The
programme therefore limits secondary displacements due to harsh living conditions.
27
For IDPs, see for example Samuel Hall/NRC/IDMC/JISP, “Research Study on the Protection of Internally Displaced Persons in
Afghanistan”, NRC, 2012.
28
For more information, see: NYT article: Rod Nordland, Afghan Refugee Children Perish in Harsh Winter, 9 May 2012, New York
Times.
116
 Several stakeholders also pointed out the various improvements that the shelters brought to
the lives of beneficiaries in terms of hygiene and sanitation. For example, the use of two
separate rooms reduces the risk of transmission of communicable diseases and improves the
ventilation conditions within the shelters. The obligation to build separate latrines is also a
real improvement for the hygiene of beneficiary households, even if these latrines were not
always used in all the communities visited for this survey.
 Getting a shelter was also a way to reduce the protection risks related to overcrowded living
conditions when several families were forced to share their living space. Several issues arose
from the fact of sharing private houses: a lack of intimacy hardly compatible with the cultural
norms at play in Afghan communities, and an increased risk of violence against women and
children linked to the extra stress caused by displacement and stressful living conditions.
ii. Social position of beneficiary households
 Fresh Start
Getting a permanent shelter was also perceived by beneficiaries as the necessary basis on which
to start re-building their lives upon return or following displacement. Beneficiaries and non-
beneficiaries alike indicated that shelter was one of their very first needs upon return and/or
displacement. It therefore appeared as the first asset that returnees and IDPs tried to secure
once they reached their place of origin or their place of displacement. Beneficiaries confirmed
without ambiguity that UNHCR assistance greatly helped them securing this, as a lot of
beneficiary households would not have been able to buy the material or to pay for the necessary
labour to build their shelters on their own financial means.
 Social Status
An important aspect related to the access to ownership and permanent housing was also the
feeling of beneficiary families to have gained – or in most cases recovered – an ‘honourable’
social position in their community. Sharing houses with relatives gave to a lot of beneficiaries the
feeling that they imposed an undue
burden on their relatives. Living in When guests come, now we can receive them.
tents or not owning one’s own house Before, we couldn’t welcome them properly, now
was perceived as a real loss of social we can. This is good for the entire community and
status. A lot of people interviewed for our honour
mentioned that living in a temporary Khord Agha, UNHCR Beneficiary – Beshood,
shelter – especially in tents – was
Saracha
equivalent to a life of beggars, often
object of despise in the Afghan
117
society29, as summarized by a local Afghan saying: ‘ask for bread but don’t ask for a house’,
meaning that begging for bread is acceptable, but not for a house30.
Regaining a proper house was therefore a way for beneficiary household to re-establish a
satisfying social position.
iii. Economic impacts of the programme
 UNHCR Shelters, a Base for Economic Activities?
The 2012 Danida Evaluation of ROI support to Afghanistan brought to light the interesting
argument that the shelters had an important economic impact on the lives of beneficiary
households, as they served as ‘home-based enterprises’ where beneficiaries – especially women
– could set up their own economic activities31.
Yet this research only marginally supported this argument. The field team observed, only in the
Northern region, that women were weaving carpets within their shelters. Moreover, the
quantitative data does not provide strong evidence for such an impact. Table 70 compares the
level of self-employment among beneficiaries and non-beneficiaries across the different phases
of migration and after integration in the shelter programme. The data shows that for UNHCR
beneficiaries, beneficiaries of other programmes and non-beneficiaries the level of self-
employment decreased when they returned from migration. Comparing the levels of self-
employment one month before receiving assistance and at the point of interview, it appears that
while no change is observed for beneficiaries of other programmes, the self-employment rate for
UNHCR beneficiaries increased by about 5 per cent. However, no causal relationship between
participation in the shelter programme and this increase can be assumed. This potential impact
of the shelter programme therefore should not be overestimated, particularly as the qualitative
fieldwork showed that a) the proportion of shelters used for economic purposes among the
shelters visited was very low and b) this economic function appeared to exist only in specific
provinces (North/Northeast).
29
See for example – Samuel Hall Consulting, (2011), Jogi and Chori Frosh communities: a Story of Marginalization.
Commissioned by UNICEF.
30
See FGD Bamyan UNHCR Beneficiaries Nawabad Sashpul: Mohammed Reza – ‘Gadaygar- e naan bash, gadaygar- e khana
nabash.’ in Dari.
31
DANIDA ROI Evaluation (2012), p. 74.
118
Table 70: Self-Employment as Main Income-generating Activity
1 Month prior
Before migration After return Time of survey
assistance
N % N % N % N %
UNHCR
546 28.28 390 20.20 464 24.03 592 29.13
Beneficiaries
Non-UNHCR
132 34.29 103 26.68 103 26.68 124 26.78
Beneficiaries
Non-Beneficiaries 480 29.59 350 29.59 - - 558 28.1
 Assessing the risks of increased indebtedness
For the most vulnerable, participating in the shelter programme was not an easy process as it
would undoubtedly place the economic resources of beneficiary households under considerable
pressure. The contribution required from beneficiaries included the material for the walls and
paying for skilled and unskilled labour necessary for the construction of the shelter. Table 71
shows that Laghman (66.1%), Bamyan (62.5%), Helmand (51.8%), Nangarhar (51.5%) and Hirat
(47.5%) are the provinces where the highest proportions of beneficiaries reporting increased
level of indebtedness. Faryab, Sari Pul, Kabul or Jawzjan provinces on the other hand seemed
relatively more immune to this issue. Various factors may explain the differences across
provinces:
 Qualitative observations found that indebtedness appeared more of an issue in areas

where beneficiaries had to procure burnt bricks, especially Kandahar, Nangarhar or
Hirat. In areas where shelters were built with mud bricks, like in Jawzjan or Kabul
provinces, procuring the building material was easier for beneficiaries. Beneficiaries
reported heavily relying on informal credit, borrowing money from multiple sources in
the community. In Hirat province as well, procuring for the burnt bricks was challenging
for beneficiaries, who sometimes had to buy their bricks on credit from the surrounding
brick kilns.
 The remoteness and insecurity of certain provinces may also enter into play as the prices
of labour and material are higher in area. This would for example be the case in Laghman
and in Bamyan provinces.
119
Table 71: UNHCR Beneficiaries Change in Indebtedness after Receiving Assistance (in %)
Decreased Increased Remained No debt I don't know Total
Kabul 32.99 37.06 21.83 5.58 2.54 100.00
Parwan 43.56 38.61 6.93 7.92 2.97 100.00
Bamyan 15.63 62.5 15.63 6.25 0.00 100.00
Laghman 23.46 66.05 4.94 4.94 0.62 100.00
Nangarhar 32.41 51.52 10.25 5.19 0.63 100.00
Balkh 40.00 38.00 6.00 12.00 4.00 100.00
Faryab 40.00 26.67 16.00 10.67 6.67 100.00
Jawzjan 40.68 36.44 6.78 15.25 0.85 100.00
Sari Pul 46.43 25.00 10.71 16.07 1.79 100.00
Kunduz 43.33 31.67 13.33 11.67 0.00 100.00
Takhar 21.05 39.47 21.05 13.16 5.26 100.00
Helmand 33.93 51.79 8.93 5.36 0.00 100.00
Kandahar 37.33 42.67 6.67 8.00 5.33 100.00
Paktya 34.96 40.65 11.38 10.57 2.44 100.00
Hirat 32.67 47.52 11.88 3.96 3.96 100.00
Total 33.87 45.97 11.06 7.33 1.77 100.00
 It must be added that distortions and flaws in the selection process significantly
increases the risk of indebtedness of beneficiaries and non-beneficiaries. In Hirat
province for example, IDP households sought to buy ‘fake’ VRF for the equivalent of
between 140 and 200 USD to increase their chances to get selected. In some instances, it
was also reported that beneficiaries had to pay additional amounts of money (reportedly
10,000 AFS or 200 USD) to the malik and the IP to access the programme. Such
complaints were also reported in Nangarhar province, where returnees in LAS townships
had to pay bribes to the DoRR in order to receive a plot. Unsurprisingly this increases
significantly the total contribution required from beneficiary households. It also
threatens non-beneficiary households who invest in these types of strategies with no
guarantee of being ‘paid back’ by at least accessing a permanent shelter.
120
 Debt and informal loans as a common coping strategy in Afghanistan
If indebtedness did increase in absolute terms following the participation in the programme, the
phenomenon cannot be assessed outside the very specific economic context of Afghanistan.
Informal loans are a common practice for households, which do not have access to any other
forms of outside funding, in the absence of a formal banking system or a redistributing welfare
state. As summarized by AREU in their research on the subject: ‘on the question of how
widespread the use of informal credit in rural livelihoods is, the answer is simple: all households
both take and give informal credit.’32 This was confirmed by most stakeholders who judged that
the programme was a way of limiting the extent of indebtedness of beneficiary households
rather than increasing it. In Bez Akmalati for instance, beneficiaries from other programmes and
the community leader expressly mentioned that non-beneficiaries had to go through higher
levels on indebtedness to build shelters. In several cases, non-beneficiaries mentioned they were
not able to acquire the same quality of material as the one provided by UNHCR (especially iron
beams).
Coping with high levels of indebtedness is highly dependent on the relative economic dynamism
of the location, offering potential opportunities for beneficiaries to reimburse their debts. In an
environment where every household incurs debts, beneficiary households are actually benefiting
from a very unusual access to in-kind support for the construction of their shelter. Unsustainable
indebtedness is therefore more likely for non-beneficiaries than for beneficiaries. The fact that a
large majority of beneficiary respondents said that they did not regret participating in the
programme despite the extra debt that they had to incur tends to suggest that the issue of
indebtedness was not a major challenge for beneficiaries.
iv. Access to Services
Focusing on fundamental services households have access to, we find modest differences
between UNHCR beneficiaries, non-UNHCR beneficiaries and non-beneficiaries. Generally –
whether looking at access to water, electricity, heating or health services – we see at least some
deprivation among UNHCR shelter beneficiaries, which leaves room for future areas of
improvement in the programme. Stronger linkages will need to be made, in future
programming, between housing and water, housing and electricity, housing and health, in
order to see shelter not only as a physical attribute but as a ‘home’ meeting beneficiaries’
expectations. Therefore, having stronger linkages with national priority programmes and service
delivery institutions and organizations is a way to vastly improve the access to services for
shelter beneficiaries.
32
AREU, (2007): ‘Finding the Money: Informal Credit Practices in Rural Afghanistan’ in Synthesis Paper Series. p. 19.
121
 Safe Drinking Water
In terms of safe drinking water, Table 72 shows a total of 7.5 per cent of those households
surveyed have no access, while only 2.4 per cent of those are non-UNHCR beneficiaries
compared to the 8.3 per cent of UNHCR beneficiaries.
Table 72: Access to Safe Drinking Water

Total
Yes, we boil water 42 1 53 96

% 2.06 0.22 2.66 2.14
Yes, free potable water 1 725 413 1 652 3 790
% 84.77 89.2 83.02 84.45
Yes, we buy potable
99 38 129 266
water
4.86 8.21 6.48 5.93
%
No 169 11 156 336
% 8.30 2.38 7.84 7.49
Total 2 035 463 1 990 4 488
% 100.00 100.00 100.00 100.00
 Power Supply
As for power supply, around the same percentage of households in each group, 48 per cent
have no electricity. Still, non-UNHCR beneficiary households are much less likely to have
access to public electricity and thus are more prone to use personal generators.
Table 73: Type of Electricity

Total
Public electricity 350 36 352 738

% 17.20 7.78 17.69 16.44
Personal generator 55 32 47 134
% 2.70 6.91 2.36 2.99
Public and personal generator 217 81 259 557
% 10.66 17.49 13.02 12.41
Solar electricity 428 87 393 908
% 21.03 18.79 19.75 20.23
No electricity 985 227 939 2 151
% 48.4 49.03 47.19 47.93
Total 2 035 463 1 990 4 488
% 100.00 100.00 100.00 100.00
122
 Heating
In terms of heating, UNHCR beneficiary households are the least likely of the three groups to
not have any form of heating at 28.7 per cent, in comparison with 54.9 per cent of non-
UNHCR beneficiary households. Moreover, UNHCR beneficiaries are much more likely to use
Sandali as a heating source, while non-UNHCR beneficiaries are more likely to use Bukhari.
Table 74: Type of Heating

Total
Central public heating 3 0 3 6

% 0.15 0.00 0.15 0.13
Bukhari 631 175 619 1 425
% 31.01 37.8 31.11 31.75
Electricity 10 2 13 25
% 0.49 0.43 0.65 0.56
Sandali 627 10 571 1 208
% 30.81 2.16 28.69 26.92
Tanur/ Tabakhana 132 3 96 231
% 6.49 0.65 4.82 5.15
Gas 46 19 63 128
% 2.26 4.10 3.17 2.85
None 584 254 618 1 456
% 28.70 54.86 31.06 32.44
Other 2 0 7 9
% 0.10 0.00 0.35 0.20
Total 2 035 463 1 990 4 488
% 100.00 100.00 100.00 100.00
 Health Facility
Lastly, UNHCR beneficiary households are much more likely to have no access to any health
facility in comparison to non-UNHCR beneficiary households, 15.0 per cent compared to 1.7
per cent. Still, non-UNHCR beneficiaries are more likely to live one or more hours away from
a health facility than UNHCR beneficiaries, and around the same percentage of households in
each group live within one hour.
123
Table 75: Distance to nearest health facility

Total
Less than one hour away 1 019 241 1 045 2 305

% 50.07 52.05 52.51 51.36
One or more hours away 708 213 667 1 588
% 34.79 46.00 33.52 35.38
None accessible 306 8 275 589
% 15.04 1.73 13.82 13.12
I don’t know 2 1 3 6
% 0.10 0.22 0.15 0.13
Total 2 035 463 1 990 4 488
% 100.00 100.00 100.00 100.00
5.1.2 Multi-dimensional Poverty Analysis

As stated in Section 2, the multi-dimensional poverty analysis allows us to conduct a more
comprehensive assessment of how deprivation relates to our sample than would be the case if using a
single monetary indicator like income. Our approach follows that which was pioneered in UNDPs widely-
recognized Human Poverty Index (HPI) within their Human Development Reports (HDRs), and has since
been developed further in recent years by such authors like Alkire and Santos (2011)33 and Alkire and
Foster (2007)34.
As Error! Reference source not found.shown in the Multi-Dimensional Poverty Index (Section 2 – table
5), 78 per cent of households within our sample are multi-dimensionally poor. In terms of poverty by
dimension, we find that 66 per cent of households are deprived in Dimension 1: Economic, 87 per cent of
households are deprived in Dimension 2: Education, 33 per cent of households are deprived in Dimension
3: Health and Nutrition, 43 per cent of households are deprived in Dimension 4: Housing, and 50 per cent
of households are deprived in Dimension 5: Social Capital and Inclusion. In regards to poverty along
individual indicators, some of the highest levels of deprivation occur in all dimensions. For example, 91
per cent of households have no membership in a community organization, 81 per cent own no livestock,
78 per cent of household respondents are illiterate, 77 per cent have less than 2 sources of income, and
52 per cent of households spend less than 690 AFS/month on food.
33
ALKIRE Sabina, SANTOS Maria Emma, “The Multidimensional Poverty Index: Contradictions and Analysis”, October 2011.
34
ALKIRE Sabina, FOSTER James, “Counting and Multidimensional Poverty Measurement”, OPHI Working Paper No.7, University
of Oxford, 2007.
124
5.1.3 Mean Difference across Groups
By comparing the average multi-dimensional poverty rate across groups, we identify differences in
deprivation. Table 76, we compare those households which were assisted by UNHCR with those
household which were assisted by other organizations as well as with those which were not assisted at
all. The comparison shows that non-beneficiaries report the highest overall multi-dimensional poverty
at 86 per cent. What’s more, UNHCR beneficiaries are slightly worse off overall than their non-UNHCR
counterparts, 71 per cent against 68 per cent.
Table 76: Multi-Dimensional Poverty Index by Beneficiary Status
Mean SE [95% Conf. Intervals]
UNHCR Beneficiaries 0.71 0.01 0.70 - 0.73
Non-UNHCR Beneficiaries 0.68 0.02 0.64 - 0.73
Non-Beneficiaries 0.86 0.01 0.85 - 0.88
Prob > F = 0.0000 N=4 488
Furthermore, a breakdown by whether the household has an official refugee returnee, a non-refugee
returnee, an IDP or no members who have ever moved helps to identify how different groups within our
sample fare. Here, IDPs are noticeable more deprived than any other group while refugee returnees are
the least deprived.
Table 77: Multi-Dimensional Poverty Index by Migration Status
Refugee Returnee 0.74 0.01 0.72 - 0.76
Non-Refugee Returnee 0.82 0.01 0.80 - 0.84
IDP 0.84 0.02 0.80 - 0.87
No Mobility 0.80 0.02 0.76 - 0.83
Prob > F = 0.0000 N=4 488
Finally we are able to combine those two previous breakdowns to provide a more detailed comparison
of overall multi-dimensional poverty by sub groups. aries alike.
Table 78 shows that among UNHCR beneficiaries, IDPs have the highest overall rate of poverty at 81 per
cent. On the other hand, non-refugee returnees have the highest overall rate of poverty among non-
UNHCR beneficiaries and non-beneficiaries alike.
125
Table 78: Multi-Dimensional Poverty Index by Beneficiary and Migration Status
Refugee Returnees 0.70 0.01 0.67 - 0.72

Beneficiaries
Non-Refugee Returnees 0.72 0.02 0.67 - 0.76

UNHCR
IDPs 0.81 0.03 0.75 - 0.86
No Mobility 0.76 0.04 0.67 - 0.84

Beneficiaries
Non-UNHCR
IDPs 0.67 0.17 0.34 - 0.99
No Mobility 0.62 0.06 0.51 - 0.73

Beneficiaries

Non-
IDPs 0.87 0.02 0.82 - 0.91
No Mobility 0.85 0.02 0.81 - 0.88
Prob > F = 0.0000 N=4 488
5.1.4 Cross-Sectional Regression Analysis

While these mean comparisons of overall multi-dimensional poverty are suggestive of differences among
groups, we are able to perform a cross-sectional regression analysis, explained in Section 2, which more
completely measures differences in deprivation between groups.
The results of the regression analysis based on a 3-group breakdown are presented in Table 79, with
each separate model using a different reference group by which to compare the other two groups
against. Because coefficients of a probit model with binary dependent variables are inherently
problematic for interpretation, the marginal effect is presented.
Models 1 and 2 estimate that the probability of being multi-dimensionally poor is significant and
increases by 15 per cent and 17 per cent simply by being a non-beneficiary in comparison to the
reference beneficiary groups. Likewise, Model 3 predicts the probability of being multi-dimensionally
poor decreases by 16 per cent and 21 per cent by being a UNHCR beneficiary and non-UNHCR
beneficiary, respectively, in comparison to a non-beneficiary.
126
Table 79: Multi-dimensional Poverty Index by Beneficiary Status (Marginal Effects)
Model 1 Model 2 Model 3
UNHCR Beneficiaries 0.02 -0.16***
Non-UNHCR Beneficiaries -0.02 -0.21***
Non-Beneficiaries 0.15*** 0.17***
N 4 485 4 485 4 485
The results of the regression analysis based on the 4-group breakdown are presented in
Table 80, with again each separate model using a different reference group by which to compare the
other three groups against. Here Model 1 estimates the probability of being multi-dimensionally poor is
significant and increases by 7 per cent, 5 per cent and 5 per cent by being a non-refugee returnee, IDP
and having no mobility in relation to the reference group, refugee returnee. Similarly,
Models 2-4 support this result indicating the probability of being multi-dimensionally poor is
significant and decreases by 7 per cent, 6 per cent and 5 per cent by being a refugee returnee, in
relation to the reference groups of non-refugee returnee, IDP and no mobility, respectively.
Table 80: Multi-dimensional Poverty Index by Migration Status (Marginal Effects)
Model 1 Model 2 Model 3 Model 4
Refugee
-0.07*** -0.06* -0.05**
Returnees
Non-Refugee
0.07*** 0.02 0.02
Returnees
IDPs 0.05* -0.02 0.01
No Mobility 0.05** -0.02 -0.01
N 4 485 4 485 4 485 4 485
Finally, the results of the regression analysis based on the more detailed 12-group breakdown are
presented in
Table 81 -
Table 83, with again each separate model using a different reference group by which to compare the
other 11 groups against. In Models 1-4, the reference group is always a sub-group of UNHCR beneficiary.
In each, the estimated probability of being multi-dimensionally poor is significant and positive for each of
the non-beneficiary sub-groups in relation to the reference group. The marginal effect shows this
127
estimation is at its most extreme when comparing the non-beneficiary, non-refugee sub-group to each
of the UNHCR beneficiary sub-groups.
Table 81: Multi-dimensional Poverty Index by Beneficiary and Migration Status (1) (Marginal Effects)
Refugee Returnees -0.02 -0.05 -0.05

Beneficiaries
Non-Refugee Returnees 0.02 -0.03 -0.02

UNHCR
IDPs 0.04 0.02 0.00
No Mobility 0.04 0.02 0.00
Refugee Returnees -0.04 -0.06 -0.09 -0.09

Beneficiaries
Non-UNHCR
Non-Refugee Returnees 0.04 0.02 0.00 0.00
IDPs 0.04 0.02 0.00 0.00
No Mobility -0.04 -0.07 -0.1 -0.09
Refugee Returnees 0.14*** 0.12*** 0.10** 0.10**

Beneficiaries
Non-Refugee Returnees 0.16*** 0.15*** 0.13*** 0.13***

Non-
IDPs 0.12*** 0.11*** 0.09* 0.09*
No Mobility 0.11*** 0.10*** 0.08* 0.08*
N 4 485 4 485 4 485 4 485
Likewise, Models 5-8 use each of the Non-UNHCR Beneficiary sub-groups as a reference group. Again,
the estimated probability of being multi-dimensionally poor is significant and positive for nearly all
non-beneficiary sub-groups in relation to the reference group. Only in Model 7 are the estimations not
significant, most likely due to the low number of observations for the reference group non-UNHCR
beneficiary, IDPs. Moreover, Model 6 shows that the estimated probability of being multi-dimensionally
poor is significant and negative for non-UNHCR beneficiaries, refugee returnees in relation to the
reference group non-UNHCR beneficiaries, non-refugee returnees.
128
Refugee Returnees 0.03 -0.04 -0.05 0.04

Beneficiaries
Non-Refugee Returnees 0.05 -0.02 -0.03 0.06

UNHCR
IDPs 0.07 0.00 0.00 0.08
No Mobility 0.07 0.00 0.00 0.08
Refugee Returnees -0.09* -0.09 0.01

Beneficiaries
Non-UNHCR
Non-Refugee Returnees 0.07* 0.00 0.07
IDPs 0.07 0.00 0.08
No Mobility -0.01 -0.09 -0.09
Refugee Returnees 0.16*** 0.10** 0.10 0.16***

Beneficiaries
Non-Refugee Returnees 0.18*** 0.13*** 0.13 0.18***

Non-
IDPs 0.14*** 0.09* 0.09 0.15***
No Mobility 0.14*** 0.08* 0.08 0.14***
N 4 485 4 485 4 485 4 485
Finally, Models 9-12 use each of the non-beneficiary sub-groups as a reference group. Supporting those
earlier results, the estimated probability of being multi-dimensionally poor for nearly all sub-groups of
both UNHCR and non-UNHCR beneficiaries is significant and negative in relation to the reference group
in each model.
Model 9 for example shows that for refugee returnees exclusively, the probability of being multi-
dimensionally poor is significant and decreases by 17 per cent and 23 per cent, respectively, by being a
UNHCR beneficiary and non-UNHCR beneficiary in relation to a non-beneficiary.
Likewise, Model 10 shows that for non-refugee returnees in particular, the probability of being multi-
Focusing solely on IDPs, Model 11 shows that the probability of being multi-dimensionally poor is
significant and decreases by 12 per cent by being a UNHCR beneficiary in relation to a non-beneficiary.
And lastly,
Model 12 shows that for those households who have never moved the probability of being multi-
129
Refugee Returnees -0.17*** -0.21*** -0.16*** -0.14***

Beneficiaries
Non-Refugee
-0.16*** -0.20*** -0.15*** -0.13***
UNHCR
Returnees
IDPs -0.13** -0.17*** -0.12* -0.10*
No Mobility -0.13** -0.18*** -0.12* -0.10*
Refugee Returnees -0.23*** -0.28*** -0.22*** -0.20***

Beneficiaries
Non-UNHCR
Non-Refugee
-0.13** -0.18*** -0.12* -0.10*
Returnees
IDPs -0.13 -0.18 -0.12 -0.10
No Mobility -0.24*** -0.29*** -0.23*** -0.21***
Refugee Returnees -0.04 0.01 0.02

Non-Refugee
Beneficiaries
0.04 0.04 0.06*

Returnees
Non-
IDPs -0.01 -0.05 0.01
No Mobility -0.02 -0.06* -0.01
N 4 485 4 485 4 485 4 485
While the results of the cross-sectional regression analysis just presented are suggestive, they do not
indicate causation of deprivation due to a problem with selection – as reviewed in chapter 4 which
detailed how most vulnerable beneficiaries were often left out of the beneficiary selection. As a result,
those households chosen to be UNHCR beneficiaries or non-UNHCR beneficiaries may have been
selected even if they were less deprived. In fact, the principle criteria for selection into the UNHCR
shelter assistance programme are formally that an individual has a Voluntary Repatriation Form (VRF)
and has land in which to build a shelter on. Therefore there is good reason to believe that those selected
were originally less deprived than those not selected, hence not surprisingly, they are also less deprived
after receiving shelter assistance. While other criteria were also incorporated into the programme in
certain cases including vulnerability-related characteristics like if the household is female-headed,
landless, or internally displaced, those criteria were not followed uniformly across all locations. As such,
we should be careful not to conclude based on these results that the shelter assistance programme
has led to less deprivation, only that those benefiting from assistance, by UNHCR or other
organizations, are by-and-large less deprived than those not benefitting.
130
5.1.5 Changes over Time
While the cross-sectional analysis does not allow for any firm conclusions regarding the impact of the
shelter assistance programme, we are able to assess changes over time for an individual household given
that the questionnaire asked specific questions for multiple periods, including both before and after
assistance (time of survey). Even though an analysis based on an overall MPI is not possible, gauging
deprivation on a select choice of individual indicators is still indicative to whether households are better
off or not following participation in the programme.
Table 84 compares the simple difference in household deprivation of UNHCR beneficiaries and non-
UNHCR beneficiaries, along individual indicators similar to those used in the MPI analysis. The results
vary across indicators, however certain changes are noticeable. Looking at access to housing for
example, UNHCR beneficiary households are 35.5 per cent less deprived compared to before receiving
the assistance, as opposed to 20.0 per cent of non-UNHCR beneficiary households. Overall, a higher
percentage of households which were assisted by UNHCR are less deprived on nearly every indicator,
aside from “Heating” and “Flooring”, over the two periods in comparison to households assisted from
organizations that were not UNHCR.
Table 84: One Month before Assistance vs. After Assistance (Time of Survey)
UNHCR Non-UNHCR
Total
Access to House (%)
Less Deprived 35.47 19.95 32.73
No Change 64.01 79.79 66.77
More Deprived 0.52 0.26 0.51
Avg. Monthly Income (%)
Less Deprived 14.15 10.58 24.35
No Change 81.57 80.78 68.76
School Attendance (%)
Less Deprived 7.87 5.44 5.38
No Change 88.92 90.41 73.43
Child Labour (%)
No Change 95.63 96.08 95.19

131
Assets (%)
No Change 74.84 77.97 71.88
More Deprived 13.66 12.31 21.57
Access to Land (%)
No Change 85.16 82.07 72.08
More Deprived 10.47 14.04 10.87
Access to Livestock (%)
No Change 84.18 75.81 77.63
Access to Mobile (%)
No Change 79.31 87.04 59.63
Safe Drinking Water (%)
No Change 97.99 98.70 95.28
Electricity (%)
No Change 83.88 86.61 70.50
Heating (%)
No Change 92.78 82.51 81.22
132
Sanitation (%)
No Change 81.97 87.90 84.27
Flooring (%)
No Change 98.82 96.54 99.04
Security (%)
No Change 95.38 97.62 95.66
The above table is interesting in that it provides evidence of how individual households’ level of
deprivation along certain indicators changed following participation in the shelter assistance
programme, both UNHCR’s and others. However once again, we are not able confidently say whether
assistance from UNHCR is was what caused the change due to the fact that these households may have
changed over time even without it. For this, we need to measure the differences between groups across
time using a difference-in-difference approach.
5.1.6 Difference-in-Difference Analysis

As explained in Section 2, because we have information across time we are able to measure the impact
of being a UNHCR-beneficiary compared to a non-beneficiary or non-UNHCR beneficiary by looking at
the difference of the groups both before and after assistance. In this way, the difference-in-difference
approach allows for an explicit conclusion regarding the effect of shelter assistance program.
The results of the probit DiD regression analysis, comparing UNHCR beneficiaries to non-beneficiaries
(those households which did not receive assistance) between when they first returned from abroad and
after being surveyed is presented in
Table 85. The results show that the UNHCR shelter assistance programme unsurprisingly led to less
deprivation in terms of access to a house. More specifically, UNHCR beneficiaries are 31 per cent less
likely to be deprived on this indicator in comparison to non-beneficiaries because of receiving assistance.
133
Similarly, UNHCR beneficiaries are 7 per cent less likely to be below the poverty line ($1.25/day), with
the estimate significant at the 5 per cent level. The only other indicators in which the effect is significant
are electricity, sanitation and access to a mobile (significant at 10 per cent level), all of which show
beneficiary households are less deprived following assistance and compared to when first returned from
abroad.
Table 85: UNHCR Beneficiaries vs. Non-Beneficiaries, First Returned vs. Time of Survey
(Marginal Effects)
Access House Avg. Monthly Income School Attendance
Time -0.19*** -0.20*** -0.14***

UNHCR Beneficiaries 0.04*** 0.07*** 0.02
Interaction -0.31*** -0.07** -0.03
N 7 084 7 084 7 084
Child Labour Assets Access Land

Time -0.03*** -0.10*** -0.08***
UNHCR Beneficiaries -0.02* -0.05** -0.09***
Interaction 0.02 0.01 0.03
N 7 059 7 084 7 084
Access Livestock Access Mobile Water
Time -0.07*** -0.35*** -0.02*

UNHCR Beneficiaries -0.04** 0.05*** 0.01
Interaction 0.00 -0.05* 0.00
N 7 084 7 084 7 084
Electricity Heating Sanitation
Time -0.23*** 0.00 -0.08***

UNHCR Beneficiaries 0.09*** -0.02 0.06***
Interaction -0.09*** -0.02 -0.17***
N 7 084 7 084 7 084
Flooring Security
Time -0.01** -0.06***
UNHCR Beneficiaries 0.00 0.00
Interaction -0.01 -0.02
N 7 084 7 084
134
The results of the probit DiD regression analysis, comparing UNHCR beneficiaries to non-beneficiaries
(those households which did not receive assistance) between before migrated and after being surveyed
is presented in Table 86. Once again, the results show that the UNHCR shelter assistance programme led
to less deprivation in terms of access to a house with UNHCR Beneficiaries 16 per cent less likely to be
deprived on this indicator in comparison to non-beneficiaries because of receiving assistance. The only
other indicators in which the effect is significant are sanitation and security (significant at the 5 per
cent level) all of which show beneficiary households are less deprived following assistance and
compared to before migration.
Table 86: UNHCR Beneficiaries vs. Non-Beneficiaries, Before vs. Time of Survey (Marginal Effects)
Access House Avg. Monthly Income School Attendance
Time 0.00 -0.32*** -0.32***

UNHCR Beneficiaries -0.01 0.01 -0.02
Interaction -0.16*** 0.00 0.01
N 6 440 6 440 6 283
Child Labour Assets Access Land

Time 0.00 0.03 -0.02
UNHCR Beneficiaries 0.00 -0.04** -0.04*
Interaction 0.00 0.01 -0.02
N 6 307 6 440 6 440
Access Livestock Access Mobile Water
Time 0.19*** -0.77*** -0.01

UNHCR Beneficiaries -0.03 0.00 0.07
Interaction -0.02 0.00 -0.06
N 6 440 6 440 4 998
Electricity Heating Sanitation

Time -0.43*** 0.07 -0.20***
UNHCR Beneficiaries 0.05 -0.04 0.02
Interaction -0.05 0.00 -0.16***
N 4 959 4 960 6 437
Flooring Security
Time -0.01* -0.57***
UNHCR Beneficiaries -0.01 0.10**
135
Interaction 0.00 -0.14**
N 4 973 4 936
Overall, we find solid evidence that the UNHCR shelter assistance programme had a significant and
positive impact on reducing household deprivation, of varying degree and depending on differences
between particular periods in time, along certain indicators of interest including access to a house,
electricity, sanitation, access to a mobile and subjective security.
5.2 IMPACT ON COMMUNITIES
5.2.1 Impact on Intra-community Relations

Overall, most stakeholders expressed their satisfaction with – and even sometimes their gratitude for –
the fact that the shelter programme had been implemented in their community. Among the surveyed
households less than 1 per cent does not perceive the impact on the community of the programme as
positive. Among the community leaders this number is a bit higher with 10 per cent. Nevertheless, the
subjective opinion of the impact on the communities is very positive. This is linked to a series of effects
that the programme has on the community as a whole as described in this section.
Table 87: Impacts of Shelter Programme on the Community
Households Community leaders
N % N %
Very negative 0 0.00 0 0.00
Negative 5 0.11 1 1.67
Neutral 28 0.62 5 8.33
Positive 1 798 40.06 38 63.33
Very positive 2 654 59.14 16 26.67
I don’t know 3 0.07 0 0.00
Total 4 548 100.00 60 100.00
 Supporting the development of villages
Community representatives noted that the shelter programme improved the availability of
housing in their area. In some cases, like in Aâb Dara in Kabul or in Shogofan in Hirat province,
the settlements of returnees and IDPs were recent and the village lacked the proper housing
136
structures and services to receive important influxes of population. In these cases, the
programme was seen as a welcome assistance to the development of the village. In Jawzjan
province, one community leader noted that the programme was useful ‘for the beauty of the
village’, because permanent shelters replaced tents and destroyed houses. This may seem trivial
but it does mean that the programme played a role in the ‘normalisation process’ of the life of
these returnee and displaced communities, who often found their villages entirely destroyed
upon return.
Furthermore, it was mentioned that in some cases the programme had a sort of ‘domino effect’
on the living condition of the rest of the community by raising the standards in housing or in
hygiene. In Hirat province for example, it was noted that the demand for higher quality I-beams
like the ones distributed by UNHCR increased following the implementation of the programme in
Kahdistan. This was a phenomenon observed regularly in the field.
The programme also impacted the awareness of the whole community about hygiene and
sanitation issues. In Nangarhar, for example, it was observed that some non-beneficiaries had
copied hygiene facilities and practices of beneficiaries. This positive impact is yet a lot stronger
when WASH training complements the shelter programme, as it was the case in some parts of
Nangarhar province.
This impression of a shelter programme benefitting the entire community was reinforced by the
fact that some communities felt that the returns of refugees brought the attention of
international and national stakeholders to their villages and that ‘help called help’ to the
community. Yet, this was not always the case as in many villages the shelter programme was
implemented on its own, without any other sort of complementary programme coming along.
 Internal Tensions
Asked about tensions in the community caused by the shelter assistance programme, the
majority of community representatives, 73 per cent, denied such developments. Qualitative
fieldwork further indicated that in many instances the shelter programme played a role in easing
the potential tensions raised by the arrival of returnees, linked to the fact that new comers are
seen as exerting additional pressure on scarce resources, especially water, electricity or housing.
By reducing the number of families depending on the resources of other households and
reducing the pressure to share their living space, the shelter programme helped diffusing these
potential tensions by mitigating the negative impact of returns on the overall community,
especially in terms of housing.
In rare instances, the shelter programme appeared to be stirring up some internal tensions and
divisions in the community. Among the sample of the community survey this was the case in 27
per cent of the 60 communities. Interestingly, this only happened when the selection process
was really perceived as unfair or corrupted like for example in Mazra in Kandahar.
The potential impact of the programme on tensions between beneficiaries and non-
beneficiaries varies according to the profile of communities though. Homogeneous tribal
137
communities developed specific coping mechanisms through solidarity networks, which were
non-existent in heterogeneous communities. Despite the fact that communities often welcomed
the assistance provided with the arrival of returnees, which represented the majority of
beneficiaries, tensions aroused whenever assistance was too concentrated. The assumption that
host communities will support the settlement of beneficiaries is based on the assumption that
they are either less vulnerable or have better coping mechanisms, which did not appear to be
the case in most locations.
Appropriateness of selection in terms of vulnerability therefore appears essential, as well as the

inclusion of a percentage of members of the host community in order to diffuse tensions. This
was noticed as an effective practice in Jawzjan.
In some cases, the host community felt that it did not have the capacity to support the return of
migrants. Need for further assistance or support was therefore highly emphasized, repeatedly
mentioned in Jalalabad, but also in Aab Dara, where further arrivals from other members of the
tribe were expected.
There were more significant tensions in areas where the selection process was problem-
ridden. In Kandahar, for example, where most of the selection process had to rely on community
leaders, tensions surrounding the programme were more acute. In areas of Hirat province where
the frauds or corruption were reported non-beneficiaries expressed their anger and
disappointment at the way the programme was implemented. In Jawzjan on the other hand,
where no corruption cases were reported, tensions surrounding the programme were non-
existent. This shows that:
a) The programme in itself is not divisive but rather much appreciated by communities as a
whole,
b) Yet, the selection process is a very sensitive issue, which may lead to conflicts if not handled
properly.
5.2.2 Impact on the Local Economy

The impact of the shelter programme on the local economy of beneficiary communities was rarely very
significant and most of the time it was only short term. As presented in Table 88 the community survey
did show that 60 per cent of community leaders felt that the programme had impacted the labour
market.
Table 88: Economic Impacts of Shelter Programme
N =60 %
Local prices 29 48.33
Labour market 36 60.00

138
Access to natural resources 20 33.33
The programme entailed a short-term employment boost, with beneficiary households typically hiring
two to three extra workers for one month to build their shelters. In Jawzjan for example, community
leaders indicated that each beneficiary household would employ three workers for the construction of
walls and for the roofing. They would typically be paid between 150 and 200 AFS per day of work.
About half of community leaders acknowledged that the programme had had an impact on local prices.
Yet, qualitative fieldwork showed that the impact was limited. In many implementation sites, the
availability of material is sufficient to meet the increase in demands triggered by the programme. This is
especially the case in areas like Deh Sabz or in the surrounding of Hirat province, where the volume of
production of the numerous brick kilns largely outweighs the rise in demand caused by the programme.
In remote rural areas, it was noted that the programme could impact prices of construction material.
The shelter programme was sometimes pointed at as playing a role in the rise of land prices that
characterised a lot of the implementing sites visited, especially in Kabul or Nangarhar provinces, where
the pressure on available land is important. But the rise in land prices is a very complex issue in
Afghanistan, linked to multiple structural factors, including the prevalence of land grabbing and the low
availability of land. The shelter programme represent a marginal element in a much larger phenomenon.
5.2.3 Impact on Inter-community Relations

While tensions caused by the implementation of the programme within communities were only
observed in a few cases, it happened that the programme often caused tensions between various
neighbouring communities. This was especially the case in contexts, where communities from various
ethnic backgrounds would co-exist, as it was for example the case in Faryab or in Kabul province. The
field observations did not show that the shelter programme in itself would stir a lot of tensions. Yet,
when the programme was part of one of the reintegration sites, where several other assistance
programmes were provided, then neighbouring communities clearly failed to understand the logic of
such a concentration of assistance in one sole location. This was the case in Kochi Abad, Kabul province.
The issues in these cases are that a) other communities will not be able to understand the rationale
behind the solutions and reintegration strategy and b) it might reinforce existing ethnic tensions as in
Kochi Abad, where only Kuchi populations received assistance, while neighbouring Hazara communities
did not. In this case, assistance can become a factor of instability, with frequent rivalries and violent
fighting between Hazara and Kuchi communities in the Central region. Keeping in mind “Do No Harm”,
this potential impact should be given more consideration.
BOX 2: Best Practices: Inter-community Relations
In Faryab Province, UNHCR had instructed the Implementing Partners to use the flexibility offered
by the guidelines of the shelter programme as an instrument to diffuse potential ethnic and inter-
community tensions. The programme allows the selection committee to select 5 to 10 per cent of
beneficiaries among the ‘host community’. In Faryab these shelters were in some cases allocated to
neighbouring communities of different ethnic background, in order to mitigate the inter-community139
tensions born out of the selection. This is an interesting practice in areas characterised by a complex
ethnic context.
6. PARTNERSHIPS WITH OTHER STAKEHOLDERS
UNHCR requested that the research team look into issues of partnership and the role of stakeholders in
the shelter assistance programme. UNHCR also specifically asked one question – is a handover of the SAP
possible to the Government of Afghanistan?
Before delving into the findings of the team’s fieldwork, it is necessary to take a step back and review
evolutions that have considerably impacted the partnership opportunities open to UNHCR on shelter
intervention and coordination. First, considerable changes have occurred, with a transition from a large
shelter coordination mechanism from 2002-2008, to a lack thereof in the following years. Coordination
and interest in shelter interventions gradually diminished due to insecurity, lowered interest among
larger NGOs, decreasing funding, and the dominance of national NGOs as implementing partners.
Second, the Government of Afghanistan has worked on urban housing policies but has not to date
developed a housing policy for urban areas. Third, shelter agencies have come and gone in Afghanistan,
with UNDP and UN Habitat reducing their activities due to lack of sustainability, cost effectiveness and
funding. UNHCR, by the scope of its intervention nationwide, has found itself to be in a position of
leadership on shelter assistance, especially given that in many parts of Afghanistan UNHCR was often the
only service provider. Moreover in recent years other changes have occurred. For one, there is the
increasing involvement of international organizations and national NGOs – however, many parts of rural
Afghanistan, especially in the conflict-ridden South, have only been consistently covered by a few
humanitarian agencies – including UNHCR and WFP just to name some of the major actors.
As a result, this evaluation’s aim is to assess the efficiency of current partnerships and mechanisms of
coordination upon which the shelter programme is based and to analyse the impact of these
partnerships on the implementation of the programme and its sustainability in the future. Partnership
strategies are a key debate in context of transition and limited humanitarian access, given that with
increasing insecurity and conflict and the related decrease in access, how are international agencies like
UNHCR going to continue doing their work and how can they do it effectively? Who should be and who
can be the reliable partners in this context of constraints? Given the history of changes with
partnerships, and the increasing insecurity since 2005 in Afghanistan, this section will seek to review
some of the gaps in partnerships over shelter that prevent any possible handover strategy at this stage.
This analysis is based on key informant interviews conducted at the central and provincial levels with a
range of stakeholders involved in shelter assistance and/or working on return and reintegration of
refugees and IDPs.
140
6.1 SHELTER ASSISTANCE IN AFGHANISTAN
6.1.1 Actors in Shelter Assistance

An important feature characterising shelter assistance conducted in the country is the very small number
of actors involved.
Table 89 gives an overview of the main recent programmes of shelter assistance in the country based on
the information that was acquired for the purpose of this evaluation. As this table shows, no other
organisation active in Afghanistan is able to implement a shelter programme that is comparable in size
and scope to UNHCR’s.
This means that UNHCR holds distinct leadership in this specific area of intervention and that any
reduction in the scope of the UNHCR programme will hardly be covered by other partners. The
potential impact of UNHCR’s changes in strategy on the lives of vulnerable households is therefore
significant.
Table 89: Shelter Assistance Programmes in Afghanistan

Size of Regions of
Organisation Dates Unit Cost Remarks
Programme Operations
UNHCR 220,000  All 2002-2012 $1,800 to $2000
Stopped
 Northern
Only provides
IOM 1538  Western 2009 to 2011 $1,900
emergency
 Northwest
shelter
400 Central  Central Community-
UN-HABITAT 2011 to 2014 $1,100
200 Eastern  Eastern based approach
 Central Stopped due to
UNDP $9,000
(Aliceghan) unsustainability
Emergency
UN-OCHA  Northern Humanitarian
Fund
ACTED  Northwest
1025 shelters
(2009-2011),  Central (Alice Flood-affected
incl. 200 shelters Ghan) populations
CARE 2009 to 2011 $2,700
in Jawzjan.  Northern IP of UNDP in
30,000 shelters (Jawzjan) Alice Ghan
since 1999
1,700 shelters Urban focus
 Eastern
CHF 1,500 shelter 2009 to 2011 $5,300 Funded by
(Jalalabad)
kits USAID
Programme
 Eastern 2009 to 2011 stopped due to
IRC 715 shelters
(Jalalabad) Stopped interruption of
funding
141
 Northwest ERF
InterSOS (Faryab) 2012 Flood affected
 Central populations
Natural-disaster
 Northern affected
Islamic Relief 175 shelters 2012 $1,200
(Balkh) population
New Programme
 Central
14,000 shelters Community-
 Eastern
NRC since 2002; 2012 2012 $1,200 based
 Western
2,075 shelters Programme
 Northern
 Central
30,000 shelters
Shelter Now  Northeast 1998-2009 Stopped
in eleven years
 Western
 North (Sari Flood-affected
ZOA 363 shelters 2012 $1,900
Pul) populations
Furthermore, most other stakeholders (IOM, OCHA, CARE, ZOA, InterSOS and ACTED for instance) focus
on natural disaster-affected populations including IDPs. Only NRC and UNHCR have a specific focus on
shelter assistance for conflict-induced IDPs and returnees. Limiting the scope of the UNHCR shelter
assistance programme would therefore have particularly negative impacts for conflict-induced IDPs.
Taking into account the growing numbers of conflict-induced IDPs throughout the country, this is a group
that should be the focus of increased attention and intervention – and shelter is a cornerstone of any
durable solution.
Additionally, UNHCR has the widest geographical coverage of its shelter programme, with presence in all
regions of Afghanistan and especially in the South. NRC might extend its activities in the South in 2013
but as of yet that has not been the case. Because the Southern regions are where most IDPs are
located35, UNHCR authorized a specific focus on IDPs for the shelter programme in the South. This means
that even more than anywhere else in the country, the UNHCR shelter programme has answered
adequately to specific local needs with the shelter programme. The gap left by a reduced shelter
programme will therefore be particularly acute in the South.
All this is evidence that UNHCR has a crucial role in terms of shelter assistance, one that is unmatched by
other actors in the country. Any reform of the programme should take this central role and responsibility
into account.
35
See for example, Brookings-Bern Project on Internal Displacement - The Liaison Office, (2010), Beyond the Blanket: Towards
more effective protection for internally displaced persons in Southern Afghanistan.
142
6.1.2 Coordination mechanisms
Although the years 2002-2008 witnessed a large shelter coordination mechanism, this gradually
diminished due to security, the development and frequent meetings of the cluster system, lowered
interest among larger NGOs, the dominance of national NGOs as IPs, and very importantly, reduced
funding from donors. In 2012/2013, at the time of this field research, coordination mechanisms for
permanent shelter assistance are almost non-existent, as there is no proper forum dedicated to this
question. There are, however, three main forums where the issue of permanent shelter could be – and
sometimes are – discussed:
1) The Emergency Shelter and NFI cluster (ES/NFI)

2) The IDP task force
3) The Housing, Land and Property (HLP) task force
Yet, permanent shelter is a very secondary question in each of these, which only indirectly address
permanent shelter-related questions. This has several explanations:
 The cluster system established by UN-OCHA was set up specifically for humanitarian and
emergency concerns. It was designed as an ad-hoc coordination structure to be called upon in
case emergency matters. Permanent shelter does not particularly fit into this framework. Both
the ES/NFI and the IDP Task force are therefore humanitarian and emergency oriented. Some
stakeholders at times use the ES/NFI cluster to discuss issues related to permanent shelter
assistance on an ad-hoc basis but it is not systematic and chairing organisations do not see the
cluster as an appropriate forum to cover these issues, which are not purely humanitarian.
 The HLP task force could seem as a relevant forum as it is less inherently linked to emergency
interventions and its main focus is on structural issues. But the HLP task force is still at its initial
stage and was conceived as an advocacy tool to tackle in particular the question of land. The IDP
task force, on the other hand, has showed that it could support interesting processes, including
the articulation of an IDP national policy. Linking up the HLP task force and the IDP task force
could help addressing in a coordinated manner the connected issues of lack of access to land and
absence of shelters, as both are major protection issues for IDPs.
 Given the small number of actors involved in shelter assistance, most stakeholders feel more
comfortable addressing these issues through bilateral and ad-hoc relations, especially between
UNHCR and NRC.
 There is a clear reluctance on the UNHCR side to ‘clusterize refugees’36, that is to subject the
main populations of concern under the mandate of the organisation to a cluster system, which
would de facto lead to a division of responsibilities with other actors, including OCHA.
The consequence of these various factors is that permanent shelter and shelter assistance are issues that
are rarely addressed collectively by humanitarian actors. UNHCR has the complete lead on shelter
assistance in the country but works rather in isolation from other organisations and actors on this issue.
36
See declaration of UNHCR Dep. Rep. in KII- Kabul Programme Unit.
143
6.2 ASSESSING PARTNERSHIPS
6.2.1 Government of Afghanistan
 National Level
Given its mandate and that returnees and IDPs are ultimately a national responsibility, UNHCR’s first
partner in the country should be its line ministry, the Ministry of Refugees and Repatriation (MoRR). This
is particularly the case since one of the responsibilities of UNHCR is to build the capacities of national
authorities so as to develop favourable conditions for a potential handover in the mid-term. This
requires a satisfactory working relationship between the ministry and UNHCR both at the central and at
the provincial level, and proper capacity building initiatives as they relate to shelter activities.
Unfortunately, at the moment, the conditions for a sustainable partnership are not in place.
This is linked to several issues weakening the relationship between the two institutions:
 Lack of a national housing policy for rural areas: The Government of Afghanistan has never
had a housing policy for rural areas, even though for urban areas a number of ministries are
involved. It remains, however, complicated to know who has authority, whether the
Municipality, or the Ministry of Housing and Urban Development, with provincial differences
to be noted.
 Low capacity of MoRR: Initially, MoRR was designed as a temporary ministry, with a limited
budget and operational role. UNHCR supported advisors in MoRR for almost ten years (they
were removed at the request of MoRR in early 2011). In 2008, UNHCR partnered with the
Civil Service Reform Commission for a six-month investment in the Ministry’s reform. A key
question then became whether funding advisers actually helped in terms of building
capacity. UNHCR’s efforts to build the capacities of the ministry and to set up a solid
partnership with the ministry were limited to the provincial levels, on practical issues and
delivery of services, such as with the selection of beneficiaries for the shelter assistance
programme. However, the capacity building component – whether technical, financial or
material – remained limited as compared with other types of partnerships built between
international institutions and their line governmental institutions in the country, such as
IOM’s capacity building programmes for MoRR and ANDMA, or the World Bank’s for MRRD.
In part due to this low investment in the national structures of assistance for returnees, and
in part due to limited funding and the lack of a mandate for implementation as well as high
staff turnover and poor overall management, MoRR has not developed enough capacities to
address the colossal issues that fall under its mandate, seriously weakening the prospects of
any sort of handover to be operationalized soon. Furthermore, MoRR’s budget has
historically been among the lowest in the government, not reaching the minimum of 40 per
cent of budget expenditures in 2012 for example. Nonetheless, UNHCR has effectively
maintained an on-going dialogue and cooperation with MoRR, on policy matters and
strategy but not on funding or capacity building. Direct relations on implementation with
144
provincial departments were stronger in some provinces than others, but missing at the
national level.
It must be added that the high turnover of ministers – 7 since 2002 – meant that with every
ministerial change at MoRR, came a change of advisers and technocrats, making continuity
and institutional knowledge hard to achieve. This is partly why the feeling of ‘being left in the
dark’ is a consequence of systemic weaknesses. The situation is even more extreme in the
provinces where there is high turnover, continuity only at low ranking civil service positions
and rampant corruption, not the least with the Land Allocation Scheme where MoRR and
DoRR officials are reported to have allocated land plots to themselves.
 MoRR does not represent a reliable partner to take over or continue UNHCR’s shelter
programme as instances of corruption, inefficiency and mishandling of funds are numerous.
Unsurprisingly, the level of distrust of UNHCR towards MoRR is also very high, further
complicating any sustainable partnership between the two institutions. On top of this, the
lack of human resources and capacity at the ministry is also apparent with very few
counterparts able to showcase technical or thematic knowledge on the populations falling
under their responsibility.
The overall political economy, systemic weaknesses and levels of mistrust challenge the
efficiency of a strong partnership between MoRR and UNHCR. The tensions surrounding the
authorisation to conduct the present study were striking evidence of the tense relationship
in 2012, and the power dynamics established between UNHCR and the MoRR. If it seems
anecdotal, the episode was symptomatic of the type of interactions in place between the
two institutions and showed the blatant lack of communication and information sharing that
characterise them. UNHCR made the choice to increasingly focus its effort on sub-national
levels by bypassing the central level when possible – for understandable reasons as reviewed
in our discussion of UNHCR’s relationship with DoRRs. However understandable, this has
added to the tense relationship of UNHCR and MoRR.
 Shelter programme: is MoRR in the dark or is there disconnect with provincial

departments? During conversations at MoRR in Kabul, the level of awareness about the
shelter programme appeared extremely low. It was difficult to judge whether this low-level
of knowledge was the result of a lack of interest, engagement or the lack of initiative by
UNHCR. To some extent, all of these factors come into play - however, it seems that the
biggest obstacle was the lack of coordination or disconnect between MoRR and its provincial
counterparts, DoRRs. There were no effective transmission systems from Kabul to the
provinces or vice versa – whether for policy information, for documents, for data or other
strategic developments at the national level. Heads of DoRRs were often in the dark with
regards to developments in Kabul, while MoRR was in the dark regarding actual
implementation of programs for the displaced at the provincial level. MoRR staff only very
rarely travel to the field – an exception rather than a rule. Either funding is lacking, or the
interest to visit field sites is low if not accompanied by an extra compensation for leaving
Kabul.
145
Various interviews with staff at the central level proved that the MoRR did not even have
basic information on the nature and implementation of the programme, such as design,
locations of implementation or programme planning. However, it should be underlined that
UNHCR also made the choice to short-cut the central level for the implementation of the
programme – due to issues of corruption, mismanagement, lack of structure at MoRR –
which, although understandable, may endanger the sustainability of the programme. MoRR
staff complained about a ‘complete disconnection of the information cycle between the two
organisations’37. The research team did not find, in 2012, any pre-planning coordination or
information-sharing mechanisms on shelter between the two institutions at the Kabul-level.
MoRR staff was resentful to be left in the dark and denied their status of responsible
authority on these issues – at the national and sub-national levels.
 Reintegration Strategy – a new bone of contention? Finally, the recent attempt of UNHCR
to shift its strategy in the country and the implementation of the Solutions Strategy focusing
on 48 sites (12 were actually implemented in 2012) raised new concerns within the MoRR.
Its representatives did not appreciate the fact that the amount of money ($1 billion) which
was supposed to be allocated to the assistance of one million Afghan returnees finally ended
up covering only 48 sites throughout the whole country. It is not surprising that this new
strategy is at odds with the interest of the ministry, which builds part of its credibility and
influence by being associated with UNHCR programmes in the field. A significant reduction of
the number of beneficiaries of UNHCR’s assistance programme has an important impact on
the visibility of the MoRR and its potential constituencies. Furthermore, MoRR
representatives do not fully agree with the deeper objectives that this strategy is supposed
to fulfil in the frame of the regional ‘Solution Strategy’38. As summarized by a senior MoRR
representative: ‘We don’t believe that UNHCR should focus on reintegration. We believe that
its expertise in on movements, not on stabilization’39.
These various factors explain why the partnership currently in place between UNHCR and MoRR is
inefficient. The choice made by UNHCR to bypass the central institution and work almost exclusively with
the sub-national levels, and MoRR’s inability and unwillingness to be involved on the shelter assistance
portfolio, prevent any possible handover strategy. MoRR does not have the capacity to take on a greater
role in the control over and coordination of shelter assistance at the moment. At the same time, the
levels of corruption and mismanagement of resources of MoRR at the central level – as highlighted on
the management of the Land Allocation Scheme (LAS) for example – and the lack of financial resources
of the ministry make it an unreliable and uncertain partner.
37
KII – MoRR Kabul/GMU.
38
UNHCR Afghanistan (2012): Afghans Solution Strategy. http://www.unhcr.org/afghanistan/solutions-strategy.pdf
39
KII – MoRR Kabul/GMU.
146
 Sub-national Level
Directorates of Refugees and Repatriation (DoRRs)
In general, UNHCR has stronger relationships with DoRRs at the provincial level than it has with the
ministry at the central level. In recent years, following the closure of camps in Pakistan and engineered
return from Iran, UNHCR and DoRRs have increased their cooperation on humanitarian responses –
which included large shelter components. In these emergency settings, the need to work more closely
with DoRRs was acute because of the need for rapid service delivery for returning refugees. This is also
one of the key humanitarian reasons why engaging with MoRR’s bureaucracy at the central level
provided more disadvantages than solutions, contrary to direct cooperation between sub-offices and
DoRRs. In other terms, beneficiaries could not wait for capacity building of MoRR, especially with such
hurdles as financial management being difficult to solve.
The level of interaction between the two institutions is higher as the DoRRs are meant to play a direct
role in the selection of beneficiaries and in the monitoring of the programme. Yet, despite the guidelines
requiring a strong involvement of the DoRRs, it appeared that its actual level of involvement and the
quality of its relationship with UNHCR varied greatly across provinces:
 In provinces like Hirat or Balkh for example, DoRR was present in the field along with UNHCR
or the staff of implementing partners and actively took part in the selection process. UNHCR
field staff noted that this inclusion of a governmental body was a great support to the
legitimacy of the selection process.
 In other provinces, like Nangarhar or Kandahar, there were several issues poisoning the
relationship between the two actors. DoRRs repeatedly accused UNHCR of its lack of
communication and transparency, while UNHCR resented DoRRs for its lack of capacities, its
questionable reliability and the constant demands for further incentives and stipend.
 Understaffing and lack of competence of DoRR staff were mentioned as a problem in

building reliable partnerships with UNHCR in the field. Turnover of directorate staff
according to nominations of Directors, nepotism and insufficient expertise were emphasized
as central problems impeding the efforts for capacity building.
According to the provincial context, UNHCR has been more or less able to develop strong relationships
with or to exert real leverage on governmental authorities. In Nangarhar for example, where the weight
of UNHCR programmes is considerable, UNHCR has gained an important political leverage over the
governor office and the DoRR. In Hirat on the other hand, the DoRR has the support of important
political leaders, which gives more independence towards UNHCR. In both of these cases, and in other
provinces, the relationship between DoRRs and UNHCR is very political and sensitive. In Nangarhar for
example, the accusations of corruption on both sides showed a very tense relationship, through which
the shelter programme becomes more an opportunity to raise tensions and to gain bargaining power
than a necessity to work together for a greater objective.
147
Other sectoral ministries
Linkages with other governmental departments are most often non-existent. This is not due to a lack of
efforts throughout the past ten years – UNHCR staff was seconded to MoRR, MRRD and MUDH in the
past. UNHCR also had strong contacts with NSP and MISFA on micro-finance during the years of focus of
this evaluation (2009-2011). However, since then, UNHCR’s main interactions with other line ministries
occurred through coordination mechanisms dealing with issues other than permanent shelters, such as
the Provincial Development Committees or other working groups. At the provincial level, there is an
absence of interaction with some authorities that could benefit the sustainability of the programme,
including:
 No formal coordination between DRRD, involved in NSP programmes in rural areas, and
UNHCR. Both operate on parallel levels, with programmes sometimes coincidently
implemented in the same locations. DRRD representatives met in Nangarhar for instance
underlined the fact they had no particular contact with UNHCR. Given the relative
experience and competence of DRRD in the implementation of assistance programmes in
rural area, this is a surprising gap in the design of the programme.
 No contacts between the municipality or DUDH with UNHCR in urban and semi-urban
areas where urban planning falls under their responsibility. In Nangarhar for instance, both
UN-Habitat and CHF operating in the same locations relied strongly on cooperation with the
Municipality. This could be considered a) to solve issues linked to access to land and b) in the
perspective of inscribing the shelter programme in an urban upgrading scheme to ensure its
sustainability.
 No interaction with ANDMA at the provincial level. The fieldwork showed that the
prevention of natural disaster is at the moment poorly integrated in the design and the
implementation of the shelter assistance programme. To prevent the risk of flooding,
systematically discussing the locations of implementation with ANDMA could be an option
that would benefit both organisations.
 No consultation in the selection and evaluation of implementing partners. In Nangarhar,

the Directorate of Economy complained about not being consulted or followed in the
selection and evaluation of implementing partners, even though the core of its mandate is
the evaluation of NNGOs. This was a notable source of tension with UNHCR giving rise to
potential allegations of corruption.
148
6.2.2 Partnerships with Implementing Partners
UNHCR has remained the foremost actor on shelter assistance. Since then, many agencies have come
and gone. Others have stayed but stopped shelter interventions. For example, actors such as UNDP had
such high unit costs that they were not sustainable, or in the case of UN Habitat were unable to sustain
funding for large-scale shelter programmes. In addition, few actors, except UN humanitarian agencies,
have the geographic coverage that UNHCR has throughout Afghanistan. Hence handover possibilities
have not only been scarce with the government, but also with other stakeholders.
 INGOs
While some INGOs had been IPs for UNHCR in the first years of the programme (e.g. ACTED), INGOs were
mostly out of the implementation of the programme between 2009 and 2011. There are a few reasons
for this: a) the overhead costs of any INGOs operating in the country, in a context of budgetary
restrictions, b) a lowered interest in shelter activities among larger NGOs, and c) a strategic direction of
UN agencies to partner with national NGOs in order to build their capacities in the long run and to allow
for future handover strategies.
In the frame of the 2012 reintegration strategy, a few INGOs were selected for the implementation. The
Danish Refugee Council (DRC) was selected to take over the logistics of the shelter and winterization
programmes. DRC has the responsibility to receive, store and release the material sent by the central
level for the shelter programme. Yet, in 2012 as well, most INGOs could not be selected to implement
some of the projects of the reintegration sites because of their costs. The RFP stipulated that one
proposal had to be submitted for each different component, forcing INGOs to factor in their overhead
costs in each proposal rather than levelling them in a general proposal.
 NNGOs and local NGOs
The shelter programme has increasingly relied on national and local NGOs both for the implementation
of the programme and for its monitoring. Partnering with NNGOs has several advantages for UNHCR:
 It greatly increases the reach of the programme, as NNGOs are not subject to the security
restrictions of UN-agencies. Relying on NNGOs guarantees a greater acceptance of the
programme in the field, as these organisations are able to establish direct links to the
communities.
 NNGOs are usually able to adopt an implementation strategy that is a lot closer to the field and
beneficiaries. In a lot of cases, NNGOs establish a temporary office directly in the
implementation site, like in Faryab or Kabul province. This guarantees a closer monitoring of the
progress of the construction of the shelters.
149
 NNGOs have cheaper operational costs than INGOs and therefore ensure a greater cost-
efficiency to the programme, enabling higher numbers of shelters to be built.
 The reliance of the programme on NNGOs has helped developing their expertise in shelter
assistance. Some NNGOs, like CHA or DHSA for example in the North or APA in Kandahar, now
have more than ten years of experience working with UNHCR and have reached a satisfactory
level of capacities, especially in terms of technical monitoring of the construction of shelters.
Yet, the tendency of UN agencies to rely on NNGOs is increasingly questioned in the Afghan context, as
there are numerous shortcomings that put the sustainability of the programme at risk:
 Despite the efforts to build the capacities of NNGOs, the inherent issue is that only rarely these
organisations have developed a not-for-profit mentality and function rather as small
construction companies. These organisations have no long-term plan to deliver assistance to a
province according to the needs, but rather a ‘business logic’ aiming at getting the ‘contracts’ of
UN programmes. This was for example the case in Hirat, where the sub-office had chosen to
change IPs every year to mitigate the risks of fraud. UNHCR staff mentioned that all the NNGOs
there (i.e. SSOAR; CRDSA; ERSA) also had for-profit activities, typically in the construction sector.
 Given this mind-set, numerous cases of corruption of IPs were reported in the field, seriously
undermining UNHCR’s credibility and the sustainability of the programme in the field.
 If the technical expertise of the NNGOs is overall satisfying, the level of awareness about the
specific requirements of the programme in terms of beneficiary selection and, most importantly,
of protection issues is still concerning and low. Field staff was notably unaware about the criteria
of vulnerability to be used in the selection process, or of protection principles, as shown for
example in discussions with the field staff of the DHSA in Faryab or WESTA in Jalalabad.
 Yearly rotation of IPs for implementation of the programme (notably in Nangarhar and Hirat)
does not allow follow-up and building of expertise in the shelter programme. The rationale
behind this system appeared rather unclear, underlines gaps and insufficient focus on expertise
in the selection process of IPs.
 Given these difficulties, the question remains, how NNGOs can, at this time, deliver larger-scale
programmes that would include other interventions such as livelihoods. This is a central question
in the context of the current Solutions Strategy. Other interventions are more complex than
shelter – yet this shelter evaluation concerns facts about the UNHCR-NNGO relationship. Instead,
the more successful local interventions have been community-level development projects, which
UNHCR funded for several years (such as micro-hydropower in the East).
The main issue with an over-reliance on NNGOs to implement the programme is that it should come
with a very strong monitoring system to avoid the pitfalls detailed above. Yet, UNHCR is increasingly
relying on other NNGOs for monitoring activities as well. Even in provinces where access is satisfactory,
UNHCR staff tends to entirely hand over monitoring to these NNGOs. It was for example the case in the
Northern region (NPO), in Kabul province or in Nangarhar, where the monitoring IP for remote areas was
150
also used to monitor implementation sites accessible to UNHCR. Furthermore, some of the NNGOs
contracted to conduct monitoring activities are also contracted by sub-offices for the implementation of
other programmes in the province, leading to conflicts of interest. For example:
 APA, the IP used for monitoring in Nangarhar, was also the implementing IP for several projects
in the reintegration site of Kuchi Abad in Kabul province.
 NPO, monitoring IP for all Northern regions (8 provinces) had also been contracted for several
UNHCR projects in the North, including the construction of a school in Sholgara district (Balkh),
some road rehabilitation projects and several projects in the reintegration site of Mohjer Qeslaq.
A good M&E system is based on an efficient triangulation of information from the field, including
external and independent oversight over the programme. The present system does not guarantee this
triangulation as most of the information is gathered by under-staffed NNGOs depending on UNHCR
contracts for their existence.
6.2.3 Donors
The following figure summarizes the earmarked contributions of donors to the shelter programme
between 2009 and 2011. The SAP is also funded through the considerable unmarked contributions of
donors like USAID (BPRM) and CIDA.
Figure 1: Donors Earmarked Contributions 2009-1011 (Million USD) 40
4
2009
3 2010
2 2011
40
Based on figures communicated by UNHCR Data Unit Kabul.
151
Most donors acknowledged having very little information on the shelter programme itself, including
objectives and levels of success. This is caused in part by a lack of transparency or information sharing,
but also because of a lower level of interest and funding from donors. Because of the security
restrictions that they are subjected to in the country, donors rely exclusively on UNHCR documentation
and reports to monitor the allocation of their funds. Donors themselves have very little access to the
field to conduct their own evaluation of UNHCR programmes. Thus this system is less conducive to a
proper oversight of UNHCR’s activities in the country.
More importantly for donors, this evaluation should provide sufficient support as to the impact of the
shelter programme on the lives of returnees and IDPs. With such information at hand, donors can now
judge of the necessity to disburse funds to support shelter activities, whether with UNHCR or with
other shelter agencies given the positive results revealed by the data collected from beneficiaries of both
groups. To be able to operationalize the recommendations from this study, UNHCR will require the
support of donors to ensure that shelter programmes remain ‘on budget’ nationwide.
The introduction of the recent Regional Solutions strategy changed the approach of some donors, who
increasingly ask for more accountability and transparency in the choices of UNHCR for the country. ECHO
and CIDA were particularly defiant about the new reintegration approach. If donors still lack a proper
mechanism to oversight UNHCR programmes, their requests for accountability have increased. The
present evaluation has already raised interest among donors, who are keen on having more research-
based knowledge about UNHCR activities in the country.
6.2.4 Humanitarian and Development Actors

The shelter programme occupies an odd position on the humanitarian-development spectrum, which
does not help in defining efficient partnerships and coherent interventions. The whole concept of
reintegration is in itself at the crossroad of humanitarian, early recovery and development approaches.
As underlined prior, if shelter assistance is a response to one of the primary needs of returnee and
displaced populations, the sustainability of shelter intervention is strongly determined by two main
factors: availability of basic services in the area on the one hand and access to livelihoods on the other.
Linkages with other types of interventions are therefore an important factor in the success of the shelter
programme on the ground.
Linkages with other humanitarian/early recovery programmes
Discussions at the central level with UNHCR showed that there is currently no systematic mechanism to
guarantee the existence of linkages with other humanitarian programmes covering basic needs. UNHCR
used to support an ‘integrated approach’. This meant that several different programmes could be
implemented in the same area, especially WASH programmes. Yet, this was not formally articulated at
the central level, which meant that it did not percolate easily to sub-national levels, where very varied
levels of integration between programmes were observed.
152
Furthermore, the implementation of humanitarian programmes provided by other actors than UNHCR as
complements to the shelter programme appeared to be more the result of chance than of a systematic
policy fostering linkages with other actors. The programme was largely designed as a stand-alone
operation, which would not particularly necessitate linkages with other interventions. This is further
reinforced by the absence of a proper forum, where the issues related to permanent shelter could be
properly addressed.
Linkages with development programmes
This gap is more blatant when it comes to early recovery/development programmes, which fall out of
the mandate of UNHCR. The missing link between humanitarian and development actions is a recurrent
problem and a broader weakness of the international intervention in Afghanistan. This shortcoming is
explained by a) the weakness of the main actors in charge of development in the country, especially
UNDP and b) the focus of development funds on governance and state building rather than on sub-
national economic development.
Importantly, the Regional Solutions strategy, initially designed to bring other actors on board to support
UNHCR’s effort concerning reintegration, ends up being as isolated a strategy as the shelter programme.
Because of UNHCR reluctance to share its responsibilities in terms of reintegration and its specific
populations of concern, and of the scarcity of early recovery/development actors ready to intervene side
by side with UNHCR on reintegration, reintegration – including its shelter component – remains out of
any efficient coordination mechanism. ‘There is no strong forum to discuss reintegration; everything we
do, we do it in isolation. (…) We should also coordinate issues that are not emergencies. Reintegration is
not about emergency, it is about opening opportunities for development.’41 This illustrates how the
reintegration strategy is weakened by its own contradictions that is building hastily a strategy based on
partnerships without any strong partnerships or linkages in place.
Humanitarian and development actors in the field
At the provincial level, UNHCR relies on bilateral relationships with relevant humanitarian actors to
coordinate on shelter assistance because of the absence of appropriate coordination forum and the
small number of actors involved. The most important point of this coordination is to reduce the risk of
duplication of shelter assistance. Overall, the risk of duplication appeared to have been properly
mitigated. Only in rare instances had UNHCR implemented a shelter programme in the same area than
other actors (e.g. IOM or UN-Habitat) and this did not lead to overlap in the selection of beneficiaries.
This issue of duplication therefore did not appear as a major challenge for the implementation of the
shelter programme and it was properly addressed through bilateral coordination.
As mentioned above, the coordination with other humanitarian and development actors in the field is
very limited. Both the shelter programme and the 2012 reintegration pilot sites were implemented in
relative isolation from other actors. This leads to two main shortcomings:
41
KII – UNHCR HIRAT – Associate Reintegration Officer.
153
1) There is no specific forum to discuss about the relevance and adequacy of shelter guidelines and
design at the provincial level and to share lessons learned about shelter assistance. This only
happens on an ad-hoc basis when another actor takes the responsibility to share its conclusions on
the shelter package with UNHCR. Even in the specific case of Nangarhar province, where monthly
multilateral shelter meetings are held between stakeholders under the supervision of UNHCR, these
are not used for information sharing and exchange about practices, but merely to share figures and
avoid duplication.
2) Secondly, as mentioned previously, opportunities to link the shelter programme with other
programmes of assistance are often missed, especially for the provision of basic services (clean
water, schools or health facilities) or for the implementation of livelihood programmes. Coexistence
of other assistance programmes in PSUs were often the result of coincidence linked to the
accessibility of the area, or its identification by several actors as an area in need of assistance due to
intensive destruction (Parwan), or massive influx of migrants (Nangarhar).
6.2.5 Military Actors

In the field, the shelter programme is sometimes implemented in areas where various military and
insurgent actors are active. In some instances, the programme was implemented in areas where PRT had
their own distribution programmes. In Nangarhar, UNHCR has an informal relationship with ISAF in the
frame of a reinforced civil-military partnership by which humanitarian actors are informed of upcoming
military operations to be ready to respond to the consequent needs. If this approach can be beneficial in
terms of humanitarian action, there are implications to such interactions: risks of jeopardizing their
neutrality through interactions with actors engaged in the conflict - including the military or the
insurgency - is something humanitarian actors need to keep in mind. In the mid- and long-run, the
question of neutrality of humanitarian actors is likely to grow in importance in a transition context where
balances between AOGs and governmental forces may shift swiftly. In order to avoid putting at stake
their neutrality, humanitarian actors in Afghanistan today must be in discussions with all parties to the
conflict - including insurgents - in order to safeguard their image of neutrality.
154
7. CONCLUSION: WHAT IS THE IMPACT OF UNHCR’S SHELTER
ASSISTANCE ON THE SUSTAINABLE REINTEGRATION OF
RETURNING REFUGEES AND IDPS?
In this question, and in this report, sustainable reintegration is defined as achieving parity with other
community members, in terms of socio-economic security. The Shelter Assistance Programme was
reviewed on its socio-economic potential, not on purely technical aspects. The comparative
measurement used by the research team allows for a multi-dimensional poverty analysis and a
comparison between shelter beneficiaries, community members, and returning refugees and IDPs who
did not receive shelter assistance.
7.1 OVERALL CONCLUSION
7.1.1 Assessing the shelter programme’s reintegration impact

The multi-dimensional poverty index used in this research showed that non-beneficiaries report the
highest overall multi-dimensional poverty at 86 per cent, compared to 71 per cent for UNHCR shelter
beneficiaries and 68 per cent for beneficiaries of other programmes. Moreover, through regression
analysis we find that being a non-beneficiary increases the probability of being multi-dimensionally
poor – by an additional 15 to 17 per cent – a significant figure.
These two findings put together show that UNHCR shelter beneficiaries fare better than non-
beneficiaries or other community members. In terms of reintegration, it means they have not only
achieved, but in most cases exceeded the objective of parity. Section 7.2 of this chapter will review the
elements contributing to or limiting reintegration in more detail. The overall context matters and both
factors contributing to, and potential obstacles to, sustainability and reintegration should be highlighted.
Extending the analysis further, comparisons over time, from the moment of exile, to return, and
before/after assistance provide solid evidence that the UNHCR shelter programme has had a significant
and positive impact on reducing household deprivation. This is true on a set of indicators – housing,
electricity, sanitation, use of mobile phones, subjective security.
Furthermore, there are only rare traces of any intra-community tensions caused by the shelter
programme. The norm is one of acceptance and appreciation, except in cases where beneficiary
selection was seen as problematic by the communities. There were observations however of tensions
across communities, especially where neighbouring communities were made up of various ethnic
backgrounds. Specifically when the programme was part of a reintegration site, where several other
assistance programmes were provided, it was common that neighbouring communities did not
understand why assistance was concentrated in one sole location.
155
Given the findings that shelter assistance is an important element to support sustainable reintegration,
Chapter 8 will recommend alternative beneficiary selection or distribution modalities that would be
more appropriate in order to diminish some of the weaknesses of the programme. These aspects of the
programme – beneficiary selection and targeting the most vulnerable – are reviewed next and are crucial
to a positive reform of the shelter programme that needs to adapt to a changing humanitarian context.
Most importantly the programme needs to focus on the inclusion of IDPs, who are so far mostly
marginalized.
7.1.2 Targeting the most vulnerable?

This study shows that the main weakness of the shelter programme is beneficiary selection and
targeting of the most vulnerable – since our findings shows that the most vulnerable are often also
those marginalized from the programme - whether IDPs, female headed households, the landless, or the
ill and disabled.
Data collected from our sample shows that, overall, more than half of the UNHCR beneficiaries did not
qualify as EVI households. Only 46 per cent can be considered as such, even though criteria to define
EVIs are loose. This indicates a failure to target the most vulnerable. More worrisome is the fact that
more than half of the non-refugee returnees that received UNHCR assistance are not considered to be an
EVI.
Problems of beneficiary selection are due partly to the irregular selection and participation of Beneficiary
Selection Committee members, misunderstanding of the selection criteria and of vulnerability criteria,
and misallocation of assistance to households that did not qualify as the most vulnerable in their
communities.
As will be reviewed in the recommendations, an over-reliance on IPs for selection and monitoring, even
in non-remote areas, has meant that UNHCR staff has not been as present as they could have been in
the field, thereby limiting the effectiveness of the process. In addition, interference of community
leaders, the uneven participation of DoRRs and the weak inclusion of female participants can all be areas
of further improvement to strengthen the outcomes of the programme.
Most notably, considerations for the situation of IDPs, female-headed households, landless
households and households with ill and disabled members can be significantly improved. Although the
evaluation is based on households who received assistance in 2009-2011, and that improvements have
to date been achieved to better integrate these vulnerable populations in the programming, the shelter
programme still has to evolve from a refugee returnee-focused programme, to one that is equally – if
not more – needed for internally displaced persons. The development of a National IDP Policy and the
increasing trends in conflict-induced displacement should constitute major push factors for UNHCR to
more closely collaborate with the Government of Afghanistan to use shelter as a means to achieving
durable solutions – especially when findings on its contribution to reintegration are positive, as shown in
this evaluation. This mention takes us to the next conclusion on a viable partnership strategy for the SAP.
156
7.1.3 Partnership – is a handover of the Shelter Assistance Programme possible?
One of the main questions set forth by UNHCR at the start of this study asked to know which ministry of
the Government of Afghanistan could have the mandate and capacity to address shelter issues for the
most vulnerable returning refugees. The research team’s findings showed that there are – at the
moment – no government entities or any other organization active in Afghanistan able to implement a
shelter programme of the same size and scope as UNHCR, and with the same target population in mind.
Looking specifically at the Government of Afghanistan, the line ministry with the mandate over such a
programme would be the Ministry of Refugees and Repatriation – however, the ministry lacks the
capacity, and is weakened by other factors (lack of resources, corruption, lack of leadership) that would
prevent it from stepping up to this task. Furthermore, other ministries, such as the Ministry of Rural
Rehabilitation and Development, focus only on rural areas whereas the evaluation of the shelter
programme – and an element recommended for its future strategy – highlights specific urban and semi-
rural (or semi-urban) needs.
A stronger partnership strategy will need to be carefully developed to further support the shelter
programme, with funding support from donors, requiring additional advocacy efforts. At this stage, and
based on this evaluation, it is clear that interactions with the Government of Afghanistan on the shelter
programme are mainly taking place at the sub-national level. National-level relationships of coordination
of the shelter activities are often non-existent. This is partly due to the limited space given to shelter
assistance in the cluster and task force systems – resulting in a weak coordination system over shelter.
Furthermore, since international organizations have been gradually diminishing their involvement on
shelter and stepping out of the implementation framework of UNHCR’s SAP, the only implementing
partners remain national NGOs, with their strengths (access, flexibility, low costs) but the weakness of
not being able to take on the role of leadership on questions of shelter. Instances of misallocation and
fraud at the local level have highlighted the need to be cautious of the relationship of IPs with
communities, and the over-reliance on IPs for implementation and monitoring, sometimes leading to
conflicts of interests. These factors limit any handover of the shelter programme in the near future.
Furthermore, the geographic coverage provided by UNHCR’s shelter programme remains to this date
one of its key strengths – with all regions of Afghanistan and 26 provinces covered in total in 2009-2011.
This level of breadth and depth is a unique asset that UNHCR should hold on to – especially given the
overall positive results of its programme on the reintegration process of its beneficiaries.
157
7.1.4 Strategic direction of the shelter programme: How to do shelter in the
future?
UNHCR holds a distinct leadership position over a well-functioning (in terms of reintegration impact) yet
improvable shelter assistance programme. Any reduction of the size or scope of the programme will
not be covered by other actors and would result in a negative impact on vulnerable populations. This is
specifically true of conflict-induced IDPs.
Access to land and shelter has been documented as the top priority need of displaced populations in
Afghanistan. The achievements of the shelter programme have now been documented in this
evaluation. Based on needs and achievements, the research team therefore recommends holding on to
the Shelter Assistance Programme as the cornerstone of UNHCR interventions in Afghanistan. Potential
improvements will be detailed in Chapter 8 – Recommendations.
However, although UNHCR is currently framing the importance of shelter activities in light of its
reintegration potential, it is important to remember that shelter is not in and of itself the key to a
sustainable reintegration, shelter being just one of many components that can contribute to
reintegration. The rest of this chapter will therefore reflect on the factors and indicators contributing or
threatening reintegration and sustainability, before concluding on the relevance of the current shelter
guidelines.
7.2 REFLECTIONS ON REINTEGRATION AND SUSTAINABILITY

The concept of reintegration lacks a standard international definition. The guidelines of the UNHCR
Handbook for Repatriation and Reintegration Activities provide elements constituting a starting point to
grasp the implications behind the concept. “Reintegration is a process that should result in the
disappearance of differences in legal rights and duties between returnees and their compatriots and
the equal access of returnees to services, productive assets and opportunities” leading to a
“sustainable return – in other words, the ability of returning refugees to secure political, economic
(legal) and social conditions needed to maintain life, livelihoods and dignity”42. “Sustainable
reintegration” is therefore understood as a process achieving parity with other community members in
terms of socio-economic security and ensuring they have access to decent living conditions. The
emphasis is placed on the disappearance of differences between the returnee and the host population,
the access to the same legal rights, equal services, productive assets and opportunities.
In the Afghan context, the existing standards for reintegration can be traced back to the National Return,
Displacement and integration strategy of 1382 (2003), which mainstreamed reintegration in government
and development programmes and identifies UNHCR as the main agency in charge of strengthening the
capacity of MoRR. However, the concept of “sustainable re-integration” of returnees in Afghanistan
appears ill-conceived and highly questionable: high levels of instability and dire economic conditions set
42
UNHCR (2004), Handbook for Repatriation and Reintegration Activities, Geneva, p. 4-5.
158
a low level for the conditions and context of return, incompatible with the idea of a sustainable
environment. Rather than on reintegration, sustainability is therefore what should be focused on:
ensuring long-term livelihood opportunities, maintained without external inputs, for all members of the
community, so as to draw the vulnerable out of misery.
Furthermore, as stated prior it is important to remember that shelter is not in and of itself the key to a
sustainable reintegration, but just one of many components that can contribute to reintegration. If the
survey results show that reintegration levels are low, this does not mean that the shelter program would
have failed, but instead that there could be other explanations, e.g. that the planning of SAP was weak
and could be strengthened.
This section reviews the determinants of reintegration and of sustainability, as two separate processes to
underline the fact that reintegration can happen today, without being necessarily sustainable given the
evolving Afghan context and the dynamic migratory trends. This section also underlines the steps that
can be taken – with specific examples from the qualitative and quantitative data – to shed light on what
can be done to ease reintegration and sustainability.
7.2.1 Determinants of Reintegration
 Intra/Inter community factors
Reintegration is a process that needs to be understood as a relative indicator – comparing the

situation of the target population (here beneficiaries) with that of local residents or non-
beneficiaries. By creating new groups within the community, by excluding some and including
others, it is important to ensure that the programme does not create tensions within the
community.
In this study, the implementation of the programme sometimes – albeit only rarely – raised
intra- and inter-community tensions, seen as threats to reintegration. The perception of the
degree of relevance of the selection process by the communities at large has strong implications
in the mitigation/exacerbation of tensions between beneficiaries and non-beneficiaries. For
example, neighbouring villages which had been completely neglected in the design of the
assistance programme for the reintegration site of Kuchi Abad, non-beneficiaries left out of
selection in Majbur Abad and IDPs not targeted in Khanaqa strongly resented the shelter
programme. This reinforces unstable environments, and places strain on programme’s success.
In terms of feeling part of the community, the quantitative data suggests that UNHCR
beneficiaries and those of other programmes feel welcome in their respective host communities.
Table 90 shows that between 2 and 2.5 per cent of different beneficiaries did not feel welcome
and supported, but only very small proportions encountered aggressive or violent behaviour by
other community members. As a result inter/intra community tensions remained the exception,
positively impacting reintegration processes.
159
Table 90: Attitude of Community towards Beneficiaries (in %)
UNHCR Non-UNHCR
Total
Very welcoming – very supportive 42.23 42.49 42.27
Welcoming – supportive 55.18 53.89 54.97
Not welcoming – not supportive 2.23 2.59 2.29
Aggressive 0.26 0.52 0.30
Violent 0.10 0.52 0.17
Total 100.00 100.00 100.00
 Natural disasters
One determinant of reintegration is the adequacy of the environment for a sustainable living –
for long-term settlement. In this regard, the occurrence of natural disasters can jeopardize both
reintegration and sustainability. In a number of locations, the sustainability of the reintegration
entailed by the programme was threatened by potential occurrence of natural disasters in the
area of implementation of the shelters:
o In Jawzjan, at least five shelters were noticed to have been destroyed by a flood in 2012.
o In Jebrail, Hirat, inhabitants expressed serious concern about the fact that an entire
section of the town, including shelters, had been constructed on a flood prone area,
while in Shogofan villagers feared the risk of floods related to the recent construction of
a canal, which construction was part of a ‘reintegration package.’
o In Nhia 1 (Jalalabad, Nangarhar) and Khanaqa (Parwan), urgent need for retaining walls
was mentioned where shelters had allegedly been destroyed by floods in 2010.
The lack of acknowledgement of risks of natural disaster is a weakness in the design of the
programme. Risks are solely taken into account for the design of the shelter itself, through
inclusion of wood bracing in earthquake prone areas. Preliminary assessments were largely
overlooked by most stakeholders, UNHCR staff and IPs mentioning relying on sight-checks by
engineers, common sense and village knowledge. In all of the locations visited and despite
specific examples mentioned above, there seems to be no culture of prevention of risks in
implementing areas. There is no technical expertise in this field and there appears to be no
awareness of IPs about such risks and no form of coordination or consultancy with other
stakeholders on this issue: ANDMA for risk assessment and municipalities or Directorates of
Urban Development and Housing in urban areas. Other organisations (e.g. ZOA or NRC) had
developed stronger procedures to mitigate natural risks in their areas of implementation.
160
 Secondary displacement as a last resort
Respondents interviewed considered secondary displacement as a last resort, as migration is

perceived as a painful, tiresome and costly process. Overall, beneficiaries and non-beneficiaries
alike insisted on their intention to stay, provided they had access to services in the following
years. Even in extremely remote areas, like in Khoje Sabz Posh in Faryab province, characterised
by no access to basic services and by very scarce sources of livelihoods, beneficiaries confirmed
that they had no intention to move again. In the trade-off between livelihood/migration on the
hand and land and shelter ownership/stability on the other, beneficiaries often chose to keep
their shelter. Still, information collected here was at the household level and individual
responses may vary, so we must be cautious of over-interpreting this finding.
 Resilience and coping mechanisms
Resilience is an important factor of reintegration – seen through coping mechanisms adopted by

beneficiary households. Interestingly, even in cases where the conditions for sustainability were
low, beneficiaries had developed coping mechanisms against the absence of livelihood through
temporary migration, indicating shelters are perceived as a central component of family life and
considered as an essential need and hub from which to plan other activities and develop coping
mechanisms:
 Seasonal work migration for herding and seasonal crop picking (Aab Dara in Deh Sabz,
Kabul, and Kuchi Abad in Kabul). In such cases, shelters were closed down and secured
for the time of migration, with yearly patterns of return of the beneficiary household.
This was further corroborated by comments from members of the local community.
 Migration of the head of household while the rest of the family remained in the shelter
(Jawzjan, Nangarhar, Hirat), either in the place of exile or to larger cities with better
employment opportunities, for example Kabul, Mazar-e-Sharif, Deh Sabz brick kilns,
Pakistan or Iran. In some cases, men commuted back and forth on a weekly basis to a
city or area with higher employment opportunities (mentioned in Kandahar, Nangarhar,
Parwan).
These observations tend to confirm that shelter is a cornerstone in the reintegration of returnees and
IDPs. It also shows that returnees develop more complex responses to the challenge of livelihood than
definitive secondary displacements. It further shows that returnees and IDPs develop appropriate
coping strategies when they do not have to invest their resources in land and housing – their priority
need – to instead focus on their next protection needs – livelihood and food.
161
7.2.3 Determinants of Sustainability
Focusing on the programme itself, assessing how it contributes to the sustainability of reintegration of
beneficiaries in a given environment implies taking into account three different sets of indicators:
1) Household-level indicators, such as occupancy rates and rates of secondary

displacement.
2) Community-level indicators taking into account the reintegration of beneficiary
households at the aggregate level (community at large, including neighbouring
communities), according to factor such as economic and social environment in the area
of return/settlement, the ethnic and economic homogeneity/heterogeneity.
3) Environmental indicators.
This section will challenge some of the assumptions on the sustainability of return. Although not
exhaustive, these findings are important to draw attention to in the conclusions as they were seen as key
findings based on field visits.
High occupancy and low secondary displacement rates
The occupancy of the shelters was high or very high, with few trends of secondary displacement.
Beneficiaries explicitly expressed a desire to settle and stay in the area of implementation of the shelter
programme. In specific cases, such as the land allocation sites of Sheikh Mesri and Chamtala
(Nangarhar), or reportedly insecure areas which could not be accessed (in Jawzjan for instance),
occupancy rates were lower, with secondary displacement and abandonment of shelters due to lack of
livelihood opportunities and lack of access to services.
BOX 3: OCCUPANCY RATE – METHODOLOGICAL CONSIDERATIONS
Assessing the occupancy rate in each location is a complex process and the figures found cannot lead
to any generalisation of results as a systematic check of all beneficiary households could not be
undertaken in each location. No assessment could be done in inaccessible areas, where rates were
sometimes reported to be lower. Remarks about the occupancy rate in this field report are based on:
- Systematic checks in villages with a small concentration of shelters.

- Direct observation by the team on a proportion of shelters in areas where thorough
checks could not be conducted due to dissemination of habitat.
- Assessment based on secondary sources: community leaders, inhabitants’ observations
(both beneficiaries and non-beneficiaries).
Even a systematic check of all the shelters built in an area is not as straightforward as it seems as
some shelters seem to be unoccupied but show signs of life. As reported by neighbours, this could be
a sign of a) short-term visits to relatives, b) temporary migration of the family to other regions of
Afghanistan or c) migrations to Pakistan or Iran. In these cases, it is difficult to draw conclusions, as it
is impossible to know whether these migrations will only last a few months or longer.
162
In most areas, empty shelters were the exception and remaining community members had a clear
perception of the reasons why the shelters had been abandoned. Being compelled to leave a shelter
allocated by UNHCR was allegedly considered as a failure. In Kahdistan (Hirat) for instance, inhabitants
insisted on the fact that reasons for abandonment of a shelter by one of their community member were
not linked to the program in itself, but to personal difficulties. Of the surveyed population, Table 91
shows that only a very minimal proportion had plans to move from their current place of residence, 0.17
per cent. The number is slightly higher for non-beneficiaries of which 2 per cent indicated that they plan
to migrate (again) internally.
Table 91: Plans to Stay in Current Place of Residence (in %)

Total
Yes 99.84 100.00 97.78 99.83
No, I have plans to move again internally 0.1 0.00 2.22 0.13
No, I have plans to migrate internationally 0.05 0.00 0.00 0.04
Total 100.00 100.00 100.00 100.00
Plans to settle
A strong sign indicating an intention of durable settlement was shelters were often appropriated by
beneficiaries, and sometimes improved, depending on the economic situation of the family. In several
cases, beneficiaries had made substantive investment to improve their shelters to turn them into proper
houses, decorating the inside of the rooms, upgrading the immediate environment by adding terraces
and plants for instance, and introducing minor changes after construction to adapt the shelter to their
needs (e.g. opening kitchens in Hirat or adding annexes). Whenever possible, additional investment had
been made to dig wells inside compounds. As mentioned earlier, in the majority of locations visited,
beneficiaries expressed no intention to migrate again and the investments made in the upgrading of
shelters is a clear indicator shelters are perceived a durable component of family life.
Despite strong aspirations to settle permanently in locations with high occupancy rates, communities
expressed considerable concern about the availability of jobs and facilities, a potential factor for later
secondary displacement if these issues are not addressed. Lack of livelihood opportunities on the long-
term places a substantive threat on the sustainability of the program for beneficiary families.
Unemployment was repeatedly mentioned as one of the main problems faced by most communities,
with beneficiaries and non-beneficiaries alike relying on daily work and unstable incomes. Coping
mechanisms such as temporary work migration are mostly viable on the short–term and households
expressed high concern about the long-lasting consequences of lack of income-generating opportunities
in the neighbourhood.
163
BOX 4: SHELTER –
AN INCENTIVE FOR RETURN AND SETTLEMENT?
In some communities, the shelter programme was expressly mentioned as an incentive for return,
sometimes triggering further arrivals. This was the case in Aab Dara, where a part of the tribe that
had remained in Pakistan came after shelter implementation in the hope to receive a shelter and
more were mentioned to arrive in case the shelter programme was pursued. These types of
comments were recurrent in tribal communities where strong links were kept with other members
of the tribe which had remained in displacement (Parwan, Kabul).
In both Aab Dara and Kahdistan, interestingly, shelter worked as an incentive despite low
livelihood opportunities and lack of services. The case of Kahdistan is significant in the sense that
returnees were mainly IDP returnees unable to return to their place of origin (Maimana in Faryab
in the case of the Moghol beneficiaries interviewed). Both the availability of land, which they had
purchased, and the implementation of the shelter programme were perceived as an opportunity
to start a new life and were met with great expectations, but also increasing concern about the
future and the improvement in terms of access to public health and education facilities. Housing,
and therefore the shelter programme, were clearly seen as a first step to (re-)integrate in a new
environment. The male members of beneficiary households were ready to cope with difficulties
linked to employment opportunities and expecting further development of the area in the
following years. However, women beneficiaries insisted on the importance of bazaar, health and
schooling facilities to address daily needs of the household.
7.2.4 Immediate Threats to Sustainability

In the minority of cases where shelters were abandoned, main reasons mentioned for secondary
displacement were the lack of privacy in the design of the shelters, land disputes and access to basic
services.
 At the household level, the incapacity to build surrounding walls, which is considered as a major
component for protection and privacy of families in Afghan society. This was noticed in Pitawa
(Kabul) and was reportedly the case in Kunar (KII APA, Nangarhar). Lack of surrounding walls
were mentioned as a critical problem in heterogeneous tribal environment, where privacy of the
family is considered all the more essential as neighbours are not related to the household
members. Surrounding walls are therefore necessary in the LAS where various communities
are mixed. The all too famous failure of Alice Ghan settlement illustrated this43.
43
See for example MAJIDI Nassim, « Home Sweet Home! Repatriation, Reintegration and Land Allocation in Afghanistan »,
paper for REMMM, forthcoming December 2012.
164
 Land disputes were a factor for secondary displacement observed in LAS in Nangarhar. This was
reportedly due to misallocation and inappropriate distribution of land, a phenomenon that is
common in LAS.
 Lack of access to basic services was also identified as a factor contributing to secondary
displacement. Transportation and water having the highest impact in the short-term, with a
direct correlation to the ability to sustain a decent livelihood, like in the LAS of Andkhoy in
Faryab province.
7.2.5 Long-Term Threats to Sustainability
The main factors determining the sustainability of reintegration (location; access to basic services and
access to livelihood) are closely interlinked. Our study showed that the poor selection of locations for the
shelter programme prevented, from the start, a sustainable return and reintegration process.
 Location and accessibility: a structural factor
The location of the shelter implementation sites was one of the main factors driving the success or
potential failure of a settlement in a particular area, with direct consequences on sustainability deriving
from:
- The availability of basic services addressing human needs (water, electricity, health facilities).
- Accessibility to employment opportunities and basic health and schooling facilities.
PSUs visited therefore faced unequal conditions for sustainability based on their accessibility, availability
of transport infrastructure and economic environment prior to implementation. In cases where returnees
come back to a village they had left for several years, often decades, in areas affected by war or
degradation due to long absence, basic infrastructure including water and transport or a favourable
environment for employment was most of the time non-existent. This is also the case in areas where
access to land was made possible by low prices, which directly derived from the absence of basic
facilities and the lack of accessibility of the area (Aab Dara, Kahdistan). In such cases, inhabitants
mentioned relying on the potential development of the area in the coming years and the need for
support from the government or assistance programmes in the meantime.
 Lack of transportation infrastructure
This was notably a primary concern for all communities as it has a direct impact on the access to urban
centres with livelihood opportunities, bazaars, health and schooling facilities. Considerable differences
were noticed between locations according to the degree of accessibility of the village. “Roads” were
therefore repeatedly mentioned as a major need by community leaders and members. The proximity of
a main circulation axis or proper roads facilitates mobility of active members of the community to find
165
income opportunities, which could be further enhanced by the dynamism of the economic environment.
That was notably the case in Bakhtyaran, semi-urban areas around Jalalabad and, to a lower extent,
Saracha. In locations with a higher connectivity (Jalalabad, Bakhtiayran in Deh Sabz), as in areas where
shelters were implemented inside or in the immediate vicinity of the agglomeration, immediate access
to the city/ dynamic economic environment meant higher employment opportunities (as rickshaw
drivers for example), but also the ability to reach health and banking facilities in town.
Conversely, in more remote areas, lack of access and transport infrastructure implies a substantive
investment for beneficiaries both in money and time. This was reportedly one of the major reasons for
temporary labour migration, with inhabitants not being able to afford commuting daily to places for
work. High concern about this was expressed in Kahdistan and Aab Dara for instance. Under harsh
weather conditions keeping inhabitants from using deficient roads or streets, lack of proper road
infrastructure had very practical negative consequences: not reaching local bazaars for job opportunities
and purchase of basic goods, impossible transportation of injured and sick people to health facilities in
the winter, and the inaccessibility of schooling facilities for children (Allah Abad, Aab Dara, Majbur Abad).
The proximity of an urban centre does not automatically imply access to school and hospitals and
enhances the need for proper roads and streets. This was notably a major issue in Allah Abad,
neighbouring the reintegration site of Kochi Abad, and in Majbur Abad.
BOX 5: IMPLEMENTING SAP IN INSECURE AREAS
Despite the difficulty to assess on site the occupancy of shelters in insecure areas, which could
not be accessed by the field team, two elements advocate the unsustainability of programme
implementation:
1. No proper monitoring can be conducted, therefore considerably reducing the visibility

over the programme.
 Accessibility Kabul central office confirmed that monitoring remote areas was
of basic services
a major issue.
2. Secondary displacements are more likely to occur due to the instability of the situation
in the area, as was mentioned in Qaysar district.
Directly deriving from the availability of transport infrastructures is the accessibility of basic services.
Despite the strong expression of intentions to settle permanently in the shelters, very serious concerns
were expressed on numerous occasions about the availability of health and schooling facilities. The fact
that returnees had had access to schooling and health facilities abroad (Iran and Pakistan) made them
more aware of their necessity.
Increasing demographic pressure in areas of high return placed strains on public schooling, health
facilities and access to water even whenever those were previously available. In Jalalabad return rates
are high and in areas where the programme was implemented the need for additional assistance
programs was emphasized in order to fight against degradation of roads and guarantee access to water
and availability of health facilities. Not taking into account these elements puts at stake the programme
as, in the long term, this might lead to secondary displacement of beneficiaries, and non-beneficiaries
develop an acute feeling of being neglected.
166
 Absence of Livelihood and Unemployment
Absence of livelihood opportunities is the issue of highest concern for beneficiaries and non-beneficiaries
alike. In areas of high return, the arrival of new comers has a direct impact on the availability of jobs,
which needs to be supported by a dynamic economic environment. This was mentioned as a factor of
tension between beneficiaries and non-beneficiaries when demographic pressure was combined with
the feeling non-beneficiaries had been neglected. This was allegedly the case in Khanaqa (Parwan),
where beneficiaries were mainly returnees from Iran and Pakistan who arrived with higher skills
compared to returning IDPs who had not been selected for the programme. This was also the case in
Bagram, with returnees from Iran could secure skilled jobs in the neighbouring Air Base. The combination
of two disadvantages, not taken into account in the selection process, triggered resentment putting at
stake the success of the programme in the village.
The previous occupation in the area of displacement therefore has a significant impact on the economic
reintegration of returning migrants. Beneficiaries who either returned to their area of origin or settled in
new settlements after long periods of displacement had acquired specific skills in exile/ displacement.
Most returnees, who were the bulk of beneficiaries, stuck to the daily urban occupation as workers or
street vendors they had practiced in Pakistan or Iran. Only in a very few cases was agriculture mentioned
as a means for livelihood and was usually a complementary strategy to cope with lack of income through
raising of livestock and extensive agriculture, even in areas such as Parwan where land is available and
agriculture a developing sector. The importance of incomes deriving from urban occupations further
emphasize the need for connectivity to a neighbouring bazaar or urban centre and the importance of
linking urban and rural development.
7.3 ASSESSING THE SAP GUIDING PRINCIPLES

As seen in the introduction, on the whole, the UNHCR shelter programme adheres to eight main guiding
principles – this conclusion returns to these to assess their relevance.
1. Community based approach

“The UNHCR shelter programme is a community based, self-help programme. The community
takes primary responsibility for identifying eligible beneficiaries to receive shelter assistance,
while the Ministry of Refugees and Repatriation, authorities, implementing partners and UNHCR
play an advisory role. In effect this approach has proven to set limits on an effective beneficiary
selection process, on support during the construction process and on monitoring and follow-up.”
One of the findings of this study is the insufficient supporting activities or training activities
provided by UNHCR to its beneficiaries. The construction process is difficult and costly for
beneficiaries – with 89 per cent of households running out of money during the construction
process, significant disparities in household contribution and lack of access to water in almost
half of the sample surveyed. While UNHCR emphasizes community help, in practice relying on
167
the community to support beneficiaries during construction is not a sustainable option given
most community members face difficulty in sustaining their own household.
Mechanisms to provide additional assistance to EVIs to build their shelter are inefficient – and
hence need to go beyond the community-based support. In addition, limited complementary
training – notably on hygiene promotion – is an obstacle to more positive results of the
programme. These require a more direct intervention by UNHCR and its IPs, during the
construction process for support and following the construction process for monitoring.

“Recognizing the challenges of facilitating female participation, UNHCR and implementing
partners involve women in selection, implementation, monitoring and management to the
greatest extent possible within regionally and culturally appropriate contexts.”
This target was met in less than one third of the communities surveyed – with only 13 out of 60
communities surveyed showing a female presence during the beneficiary selection process, and
this participation came from UNHCR itself, not the local community. In addition, the survey
shows that female-headed households are not sufficiently integrated in the programme, and
their results, based on the MPI analysis, are lower than those of other beneficiaries. Women are
included but their direct participation is still limited and where they were involved it was in no
meaningful way.
3. Access to land
“Only families with evidence of land ownership will be eligible for shelter assistance. However,
those who for a long time had a house on government owned land may also be eligible, provided
that the land is not disputed and the local authorities issue a no-objection certificate (NOC) for
them to construct a new house. In addition, a family who meets the vulnerability criteria and has
a lease or right to use the land from a landowner may also be eligible for assistance. However,
landless beneficiaries are not included in UNHCR’s shelter programme; they fall under the
responsibility of the Government of Afghanistan.”
It is acknowledged that the shelter programme is designed for the most vulnerable “landowners”
whilst the most vulnerable returning refugees are those with no access to land (or insufficient
access to land to accommodate the increased family size during years of exile). However, given
the seemingly contradiction in terms, i.e. targeting landowners yet aiming to target the most
vulnerable, UNHCR will need to find ways to target the most vulnerable returnees to secure
access to land and shelter. This will be addressed in the recommendations section of the report.
“Beneficiary selection is based on the belief that vulnerable families would not be able to
establish shelters without external assistance. UNHCR recognizes that vulnerability is a relative
phenomenon in one targeted location or village as compared with another location. That is why
168
the Beneficiary Selection Committee (BSC) is tasked to play an important role in identifying
vulnerable beneficiaries. Extreme vulnerability can be identified during the beneficiary selection
process or during programme implementation. For vulnerable categories such as female-headed,
disabled or elderly heads of households without external support and large families with
insufficient income, all involved staff should ensure that no vulnerable families are overlooked or
rejected for assistance.”
This research finds that some categories are included, but still to some extent overlooked,
namely IDPs, female-headed households, the landless and the ill and the disabled. The latter
category is often times more difficult to research on and target, in part due to the tendency of
families to hide the disabled from public view. This study shows that proper vulnerability
targeting will require more sensitized and specific training – of UNHCR local staff, IPs, community
leaders and other BSC members to ensure that they are included before other non-EVI
community members. This will constitute a priority for UNHCR’s future programmatic changes.
“Afghanistan’s forest is one of the most destroyed sectors of the environment. When
implementing shelter projects, regional offices should consider this fact and use alternative
materials in lieu of wood or, in cases where wood cannot be avoided, try to ensure that wood
products are either imported or are from sustainably harvested local sources. The UNHCR shelter
package therefore includes iron doors and windows for all shelters throughout Afghanistan.
Similarly, iron roof beams or dome roofs made of brick are promoted wherever possible. The
shelter package also includes one latrine for every family, increasing environmental hygiene in
beneficiary communities.”
The study finds that in effect, wood beams have been set aside for the preferred iron beams
although provincial differences in the procurement of wood, when needed, still remain and can
be addressed. However, environmental considerations have trickled down to the regional
offices. Regarding the use of latrines, this study finds that the level of hygiene and sanitation can
drastically be improved with beneficiaries not showing positive impact rates of the shelter
assistance on their level of hygiene. Providing a latrine alone per family does not suffice –
qualitative observations showed latrines were the last piece of the construction process,
sometimes uninstalled, or used for other purposes. The provision of training and support will
prove crucial on this aspect.

“Recognizing the diversity of climatic conditions and cultural preferences in the design of houses
in each region, the UNHCR shelter programme provides a model design against which the in-kind
(material) and cash contribution are based. For instance, under the UNHCR shelter programme,
the dome type ceiling can be seen in west and north Afghanistan and the flat roof with beams in
central, east, southeast and south Afghanistan. The flexibility of the shelter programme should
allow for these variations.”
169
Shelter design remains flexible but the level of household contribution remains unequal, with
important geographic disparities. 93 per cent of households had to contribute to their shelter,
however gaps remained between rural, semi-rural and urban locations. This is due to the higher
costs of materials and labour in urban areas – higher costs that will have to be taken into
account in developing an urban strategy for the shelter programme, discussed in the
recommendations chapter. However, evidence shows that while UNHCR beneficiary households
located in an urban context spend more in absolute terms, semi-rural households spend slightly
more in relative terms. There are other important disparities across provinces. Typically regions
of high return and high rates of urbanisation, such as Hirat, Nangarhar and Kabul present
significantly higher levels of household contribution.

“The UNHCR shelter programme seeks to contribute to reviving local economies through its
implementation wherever possible. This includes using skilled and unskilled labour, and local
procurement of raw materials.”
The requirement for skilled and unskilled labour, although a burden on beneficiary households,
has a positive repercussion on the local economy. However, the impact of the programme on the
local economy of beneficiary communities was rarely very significant and most of the time only
short term. The community survey showed that 60 per cent of community leaders felt that the
programme had impacted the labour market but in most cases, this was a short-term
employment boost, with beneficiary households hiring two to three extra workers for one
month. The impact on local prices was similarly limited. Finally, the shelter programme was
sometimes pointed at as playing a role in the rise of land prices, especially in Kabul and
Nangarhar. However, this is related to a much larger phenomenon than reintegration assistance.
8. Involvement of local authorities

“In 2003, MUDH, MORR and MRRD developed a national policy for shelter programme
harmonization with the help of other key players. Throughout the process of shelter
implementation, district officers should be encouraged to be actively involved. This is particularly
important during beneficiary selection, the most sensitive step in the shelter programme. The BSC
must include members of the Community Development Council (CDC) where present or the
provincial, district, or village shura (committee of elders and trustees), local authorities (district
authorities, provincial representatives of MORR), in addition to representatives from the
implementing partner (IP) and representatives from UNHCR (where field presence is possible).
Joint monitoring is also important, especially concerning communication channels with
beneficiaries, land disputes, ownership and other related issues.”
This remains one of the weak points of the process with MoRR being “kept in the dark”, or being
incapable of linking up with the SAP at the national or sub-national level. The harmonization
policy has not led to greater coordination between sectoral ministries, MoRR and UNHCR, or
with IPs, each working on their own due to structural, systemic, financial and staffing constraints.
170
7.4 STRATEGIC EVOLUTIONS OF THE SAP
Given the key findings of this research and the evidence of SAP’s contribution to reintegration, the
question can no longer be ‘Should the shelter programme continue to be implemented in Afghanistan?’
but rather ‘How should the programme evolve to:
a. Better adapt to the current migratory trends of the country;

b. Better fit the needs of the most vulnerable;
c. Be more inclusive of IDPs and other vulnerable segments of the population, and of potential
partners of the programme; and
d. Be sustainable in an increasingly complex humanitarian context?’
These questions will be addressed in the final chapter of this report.
171
8. RECOMMENDATIONS:
REINFORCING SAP GUIDING PRINCIPLES
The recommendations of this research are based on the existing SAP Guiding Principles – elements
highlighted in bold/orange are the recommended additions to the Guiding Principles for 2013 and
beyond. From 8 main Guiding Principles, the research team proposes a set of 11 Guiding Principles.
These include two types of recommendations are presented:
First, the addition of new Guiding Principles – such as the need to integrate Impact and Needs
Assessments (item 1), IDPs’ direct participation (item 4), a Partnership Strategy (item 10) and a
Monitoring Framework (item 11). These additions are both the most relevant to the project
planning cycle and to the changing humanitarian context of Afghanistan. Impact, Needs
assessments and Monitoring frameworks are prerequisite for any accountable and transparent
implementation process, while IDPs’ direct participation and a solid Partnership Strategy are
requirements imposed by a humanitarian context defined by increasing internal displacement
and lack of access.
Second, the strengthening of already existing Guiding Principles – Our recommendations seek to
improve, and often breakdown in more detail, principles such as the Community-based approach
(item 2), Women’s direct participation (item 3), Access to land (item 5), Focus on vulnerability
(item 6), Environmental concerns (item 7), and the Preservation of cultural and regional
preferences (item 8). These are principles that were found, in our research and fieldwork, to be
weak in their implementation – and hence need to be strengthened by better adapting to the
challenges at the field level.
The proposed set of 11 SAP Guiding Principles below is a “ready to use” revised set of guidelines for
UNHCR’s 2013 programmatic review.
2013 SAP GUIDELINES – 22 PRINCIPLES
1. Impact and Needs Assessments

a. Baseline
c. Setting standard for household contribution
2. Community-based approach

a. Include women and gender criteria in the selection of beneficiaries
4. IDPs’ direct participation

5. Access to land
a. Evidence of land ownership or NOC
b. Legal assistance in cases of land dispute, inheritance, mahr
a. Beneficiary Selection Committee
b. No family overlooked
c. No contribution requirements for EVIs
a. Alternative materials
b. Latrine per family
c. Hygiene and Sanitation Training

c. Flexibility in design
d. Adopting an urban approach

a. Reviving local economies
b. Local procurement of raw materials
10. Partnership Strategy

b. Involvement of CDCs
11. Monitoring Framework

a. Internal monitoring - Increase involvement of UNHCR staff
173
1. IMPACT AND NEEDS ASSESSMENTS
a. Baseline
Introducing systematic baselines to build program specificities
Measuring the impact of the programme is not possible without a proper baseline of needs identifying
what the initial economic and social conditions in the area of implementation were. Baselines can also
help in providing a clear perception of community relations to avoid creating tensions through the
programme.
Improving needs assessments
One of the main weaknesses identified in the programme was the absence of needs assessments prior to
implementation. This could be improved through data collection and inclusion of complaint mechanisms
at the start of the process. The NRC model, with involvement of the M&E team at the very beginning of
implementation, can be taken into consideration to develop stronger needs assessments.
Baselines to drive the selection process
One of the main challenges to the effective selection of vulnerable households was the fact that
selection of areas of implementation was more supply-driven than based on the needs of the
community. In cases where high return areas where identified and allocated a high number of shelters,
all VRF holders received assistance. This raises the issue of the quality of the preliminary assessment of
needs. There were recurrent problems observed in Kabul, Hirat and Nangarhar, with quotas of shelters
having to be met. This has a considerable impact on selection, as in such cases shelters are distributed to
less vulnerable households based on the VRF criteria, whereas neighbouring communities are left out of
the programme, fuelling potential tensions.
Defining a set calendar for selection and construction
In most cases, the timing of the construction can increase the challenges of the process. This study
shows that if the programme starts late and the bulk of the construction work takes places during
summer, at a time when water is not available, this increases household debt, poverty and increases the
chances of an unsustainable outcome.
Based on field evidence, in cases where the bulk of construction had to take place in the summer,
beneficiaries asked for extensions and waited for the rainy season. Furthermore, the lack of water plays
a role in the level of indebtedness of beneficiary households.
Starting construction earlier in the spring – setting specific calendars for selection, then construction –
would help reduce the risks of incompletion of shelters before winter, and would help reduce levels of
indebtedness. It would also increase coordination within UNHCR to know that:
174
i. Fall months = selection of shelter beneficiaries
ii. Winter months = training / sensitization of beneficiaries, communities, IPs
iii. Spring months = construction
iv. Summer months = follow-up and monitoring
Allowing flexibility during the construction process
The one-type-fits-all approach of the shelter design does not meet the needs of the beneficiaries.
Additionally, flexibility in construction with the approval of IP and/or UNHCR staff would prove more
efficient and cost-effective, as beneficiaries otherwise tend to change the original design after handover.
The inclusion of additional wooden beams in order to enlarge habitable space can notably prove
problematic, putting in danger the overall structure of the building by introducing dissymmetry in the
design of the shelter. Beneficiaries tend to change the design of their shelters after the handover to
avoid losing their cash grant, putting at risk the sustainability of their shelters. More flexibility during the
construction process would allow UNHCR to monitor these changes, decreasing the risks for
beneficiaries.
More involvement of UNHCR staff – Required presence in accessible areas
Solutions requiring additional effort or involvement were often overlooked due to insufficient
implication of UNHCR staff or weak remote monitoring processes. The fact that even in accessible areas,
UNHCR staff did not participate in all beneficiary selection is a major area of concern.
c. Setting a standard for household contribution
The level of contribution expected from beneficiaries is not detailed in the SAP guidelines, which only
mention the fact that beneficiaries are expected to cover the costs of labour and of the construction of
walls. It is recommended that future guidelines incorporate a standard for contributions. The level of
household contribution should be more clearly defined by the organisation and to be included as an
indicator for the monitoring and evaluation of the programme as it plays an important role in the impact
and sustainability of the programme.
The differences in assistance allocated to beneficiaries in the construction process across provinces calls
for a more homogenized approach and clearer guidelines about the support provided to beneficiaries in
specific contexts. The study showed great disparities according to location (urban, semi-rural and rural)
and provinces, with at times a difference in contributions covering a range of 700 USD difference. The
relative burden – rather than the absolute amount – of contribution should be studied, relative burden
when compared to household income. The study finds that those households in Hirat have the highest
contribution to monthly income ration, with Helmand having the lowest. It is important to note the sub-
office of Hirat already reviews yearly the cash grant based on the costs of material and labour, a good
practice that should be generalized to all sub-offices.
175
Typically regions of high return and high rates of urbanisation, such as Hirat, Nangarhar and Kabul
present significantly higher levels of household contribution. This is unsurprising given the higher level of
local prices and labour costs in these regions. The material used for the construction of the shelter also
enters into play, especially in Hirat province, where beneficiaries had to use cement and burned bricks in
the absence of clay, which significantly increased the level of household contributions in this province.
To increase the level of control over the construction process and its outcomes, UNHCR could develop a
standard for contributions to the shelter construction. This standard will have to be based on other
agencies’ work, on economic assessments of the cost of shelter in Afghanistan, and as a factor of income.
This would help raise awareness among families of the costs involved in the shelter process. In the case
of the most vulnerable households, the same process as adopted by NRC – i.e. of “no contribution for
EVIs” – could be considered at this stage by UNHCR to avoid doing more harm than good. However,
these will have to be the focus of a separate study looking at cost-effectiveness issues and budgetary
constraints facing the organization.
2. COMMUNITY-BASED APPROACH
The recommendations below are all based on factors that contributed, in the locations surveyed, to
creating tensions within and between communities. Implementing these guidelines will therefore serve
to ease community tensions and to avoid potentially adverse effects of the shelter programme.
Increasing the degree of transparency of the selection process and paying specific attention to
communicating about the criteria and refining them prior to and during selection are essential to
ensure all members of a given community are reached, to avoid allegations of misallocation of assistance
and misunderstanding of the selection criteria. Publicity campaigns about the selection were mentioned
by ABR in Parwan, but seemed to have had little impact in the communities visited. These type of
initiatives need to be strengthened and standard and clear complaint mechanisms need to be
introduced.
The “Shelter +” model has had negative socio-economic impacts on the area of implementation. Based
on observations in reintegration sites, the “shelter +” model as part of the new Regional Solutions
strategy contributes to increasing tensions that arise if shelter and complementary assistance are limited
to specific sites, enhancing feelings of discrimination and differences with neighbouring communities.
Further taking into account non-beneficiaries in the planning of the programme is important to avoid
fuelling tensions. It is crucial for the SAP not to have negative secondary effects on non-beneficiary
communities, which endangered sustainability of reintegration in Kochi Abad for instance. A strong
factor of sustainability is to make sure a) the programme does not put back-to-back beneficiaries and
non-beneficiaries and b) the programme benefits the entire community, including neighbouring
villages/blocks.
176
In areas where it was made available, complementary assistance worked as a strong condition for the
sustainability of the program. The “integrated” approach upheld by UNHCR prior to 2012 does not seem
to have been systematically implemented. In Hirat, there were no regular patterns for complementary
assistance: WASH programs had not been implemented since 2008 and cash for work has only been
done in parallel to shelter in some cases. In Jalalabad conversely, IPs mentioned WASH programmes
were systematically included as part of the implementation of the shelter programme and non-
beneficiaries insisted on the benefits of such initiatives. Systematic implementation complementary
programmes (schools, clinics, WASH) appears as a good practice to be considered at the national level,
including through partnerships with other agencies and organizations, as well as the involvement of
provincial directorates.
However, it should not be reduced to specific sites since needs are widely present – a needs-based,
rather than location-based, approach should therefore be adopted to ensure that needs are covered.
3. WOMEN’S PARTICIPATION
a. Include women and gender criteria in the selection process
Though cultural considerations in terms of gender inclusion in selection cannot be overlooked,

systematic efforts should be carried out to include women in selection, through women’s shuras for
instance. This study found that the level of participation of women at every stage of the implementation
was low. Awareness needs to be raised about gender considerations in terms of vulnerability so that IPs
can go beyond the “widow” criterion to determining gender-sensitive vulnerabilities.
Based on field evidence, a ‘gender criteria list’ should be developed by UNHCR and handed to its IPs. This
list could include – but not limited to – the following criteria:
i. Female-headed households due to the absence of the head of household (migration

dynamics),
ii. Women living in households where the head of household is unable to work due to
disability or illness,
iii. Women who are mothers to newborns / infants,
iv. Legal disputes over land due to inheritance or mahr cases involving displaced
women – the resolution of which would help displaced women and their families to
be entitled to possessing land and being eligible for shelter assistance.
This study showed that the staffing of IPs, i.e. the presence of women in their field staff, had an impact
on the inclusion of women in the SAP. The number of women in the staff of IP candidates should be
taken into account as a criterion to select IPs by UNHCR sub-offices.
177
4. IDPS’ DIRECT PARTICIPATION
Given the current humanitarian trends in the country, conflict-induced displacement is and will remain
on the rise. A focus on IDPs can be argued on purely humanitarian grounds in Afghanistan. The research
team advises for a greater inclusion of IDPs in the shelter programme since conflict-induced IDPs are a
key population of concern for UNHCR – as only they and NRC cover this population in their targeted IDP
programming. Although the research team is aware that, under international mandate, UNHCR’s
primarily responsibility is for refugees, and while it has no formal mandate for IDPs, the understanding in
Afghanistan at least is that conflict-induced IDPs fall under the responsibility of UNHCR. Although a focus
on IDPs may arouse negative comments from refugee-hosting countries, given the lower rates of return
and the humanitarian concerns for IDPs, this emphasis should be taken on by UNHCR – regardless of
political and constitutional arguments to the contrary, with the need to give precedence to UNHCR’s
humanitarian role.
The shelter programme should be targeted to these IDPs by introducing a new guiding principle of IDP
participation and inclusion.
I. Adjust the programme guidelines to make the objective of IDP participation and
inclusion clear to all stakeholders, IPs and communities,
II. Raise awareness of field staff on the necessity to increase the number of IDPs
selected in the programme, in regions where it is relevant (e.g. Hirat),
III. Integrate IDPs in displacement and IDPs in their areas of return, further supporting
different forms of durable solutions, and
IV. Use one-room shelters for the specific purpose of integrating IDPs and landless
people in the programme.
Similar to efforts to include women in the selection process – whether through beneficiary selection
committees or through local shuras – IDPs can also be further represented in communities where their
voices are often more marginal to that of returning refugees or no-mobility households. Without
creating tensions between groups, the requirement to have IDP representative in selection committees
should be reinforced.
178
5. ACCESS TO LAND
b. Legal assistance in case of land disputes
Building on NRC’s expertise, it is important to frame the shelter programme within a broader operational
priority to resolving land disputes involving displaced populations. As seen with the inclusion of women’s
right to land through inheritance and mahr, additional legal assistance in case of land disputes plays an
important role in the reintegration process, especially for women, as land disputes are a very common
challenge for families displaced due to exile, conflict or natural disasters. UNHCR’s cooperation and
partnership with NRC needs to be sustained to assist in the resolution of land disputes, and its
cooperation with AIHRC in identifying the cases of land dispute.
Increasing the proportion of landless beneficiaries
The majority of returnees and IDPs do not own land. Solutions suggested in the guidelines for allowing
access to land to vulnerable households need to be implemented in the field. Despite the high sensitivity
of this issue at the government level, collaborating with governmental authorities on this issue is a
requirement. Raise awareness of the field staff on the guidelines.
6. FOCUS ON VULNERABILITY
c. No contribution requirement for EVIs
Increase the involvement of the Protection Unit to systematically identify beneficiaries facing
difficulties to contribute to the construction of their shelter.
o Activate the EVI extra assistance mechanism in a more systematic manner.

o Identify a community representative to be made responsible for the identification and
the support to specifically vulnerable beneficiaries.
Additional support for the most vulnerable needs to be further taken into account during the
construction process through:
i. Tailored cash-assistance through involvement of the BSC / Protection Unit,

ii. By including a shelter model with no contribution for EVIs.
Heighten involvement of the Protection Unit in the selection process to identify EVIs and suggest
complementary assistance procedures.
Explaining the criteria for vulnerability through community-based workshops to communities

(beneficiaries, non-beneficiaries and community leaders alike), so that IPs, DoRR representatives and
community members are sufficiently sensitized about the issue. The “Beneficiary Social Verification
179
Check List” for instance, insufficiently addresses the issue and does not provide IPs with any clear means
to identify vulnerable households.
o Post-selection, during the winter, specific community-level sensitization to vulnerability

should be carried out as a precursor to the construction process. This will also ensure
greater community involvement and contribution to the shelter construction of the most
vulnerable households.
o It will also allow communities to understand the benefits to all – even non-beneficiaries
– of the introduction of a shelter programme in their location.
Introducing staff and training IPs to the developments of the PSN approach and its practical
ramifications. The survey showed that staff was still entrenched in the EVI categorizations and less
knowledgeable, if at all, on the new PSN guidelines. What is the purpose of these new guidelines? Why
were they developed and which gaps do they address? How can they improve the work of UNHCR staff
and of its IPs?
o This should also be done through workshops during the winter months, through each
regional office.
7. ENVIRONMENTAL CONCERNS
c. Hygiene and sanitation training
The current guidelines highlight the importance of 1 latrine per family as part of the environmental
concerns. The guidelines can be improved by adding a qualitative priority component: through
increased hygiene and WASH training for beneficiaries, with specific sessions for women, men and
children.
A key finding of this study is the gap on hygiene promotion. UNHCR shelter beneficiaries are significantly
less likely to receive any hygiene support. Hygiene promotion was indeed less common (UNHCR: 20%;
other programmes 34%). Hygiene training and WASH assistance should be improved since the research
has shown that the state and use of latrines was highly related to the implementation of such training.
It is important to note that hygiene promotion training for EVIs – which is across the board the least well-
covered training type – is lacking the most in semi-rural areas where only 1 in 10 beneficiary households
have reported receiving hygiene promotion training, compared to 1 in 5 rural households, and over 1 in 4
urban households.
One of the weaknesses in the design of the programme identified in the field was the lack of assessment
of natural disaster risks conducted prior to construction. Preventive measures imposed by the
programme’s guidelines are limited and only cover earthquake-mitigation measures.
180
Regional differences in preferences also calls for a region-based approach leaving more room in
designing the program to specific provincial contexts and to specific risks of natural disasters – whether
floods or earthquakes, for example.
i. A proper assessment of the climate, natural disaster risks, and issues of risk mitigation
and prevention raised earlier, will need to be reinforced in future shelter programmes.
This can be a good opportunity for UNHCR to link up its technical assessment with that
of engineers of the Ministry of Refugee and Repatriation, and the Ministry of Rural
Rehabilitation and Development, hence strengthening its partnership strategy.
ii. In earthquake-prone areas, risk mitigation is solely taken into account through the
inclusion of wood-bracing in the design of the shelter. However, in the East, DoRR
reported that wood-bracing was often removed by beneficiaries, due to a lack of
awareness of their use. This emphasizes the need for proper awareness training about
the importance of wood-bracing and other risk mitigation measures.
iii. Preventive measures against floods are lacking. This was notably the case in Nangarhar,
Hirat and Jawzjan. In the latter ZOA issued practical recommendations to upgrade
shelters and avoid degradation, which had happened in the province in 2012. Despite
high risks in the province, the only measure recommended in practice by UNHCR was to
build the shelters 60 cm above the ground, which was not systematically implemented
across the province and is insufficient in case of serious flooding. In Kandahar, Hirat,
Jawzjan, Parwan and Nangarhar inhabitants insisted on the need to build retaining walls
iv. Proper risk assessments in flood- prone and earthquake-prone areas were absent.
Coordination efforts between ANDMA and UNHCR were scarce – and could be
significantly improved in future guidelines.
8. PRESERVATION OF CULTURAL AND REGIONAL PREFERENCES

b. Adopting an urban approach
This is the result of strategic evolutions introduced by UNHCR management in 2009-2011, and which
should be further strengthened and continued in 2013 and beyond.
i. Breaking down the type of shelter by location highlights the higher uptake of one-room
shelters in urban areas, compared to semi-rural or rural areas that have the lowest
proportion of one-room shelters. This further underlines the need for flexibility of
models in urban areas – and the way one-room shelters can be used as a tool to absorb
displaced persons in their new environments by building families an extra room. This
allows upgrading or expanding of shelters that already house displaced family members,
who opt for living with host families. It also fits more realistically with the more limited
181
space available in urban areas compared to returnee townships or rural areas. Flexibility
in shelter models is therefore an asset for beneficiaries depending on their location.
ii. Further adapting to the current migratory context, taking into account trends of urban
migration. Fighting against the flow and insisting on implementing settlements in remote
areas, which lack of access and livelihood opportunities and basic services is counter-
productive. This is even more sensitive as most returnees come back with urban skills
and are incapable of sustaining a livelihood in rural areas. Inability to access centres of
employment, reach health and education facilities were high areas of concern for
communities and potential drivers of secondary displacement. In cases where these
opportunities are insufficient, complementary programmes needs to be considered.
9. CONTRIBUTION TO LOCAL ECONOMIES

Generally, we do find evidence for short-term impacts on the local economies due mainly to the demand
for skilled and unskilled labour, however this is a short-lived effect. Another key area where gains could
be made in the local economy is through the procurement of materials at the local level. While doing
procurement at a local level may not have an economies of scale effect, it would do more to aid the local
economy. Currently, there seems to be a trend away from local procurement of materials, which is
indicated in the guidelines. However, in order to contribute to the local economies this should be
considered again.
Development of the local economy should not be a main priority of SAP. However, some modifications
could be considered to help have a more positive effect in this respect. Additionally, linkages with other
organizations could also be strengthened to provide livelihood opportunities that can lead to
employment opportunities and establishment of businesses, which will have positive aspects on the
economy in the long-term.
10. PARTNERSHIP STRATEGY

UNHCR can improve the quality of its relationship with the MoRR at the central level by setting up
systematic mechanisms of information-sharing and regular lines of communication on shelter activities
between the two institutions at the national level.
Also holding a coordination meeting with ANDMA prior to planning and implementation to help
mitigate the natural risks in areas of implementation and develop a necessary culture of prevention
within both organisations.
182
Involvement of DoRRs
UNHCR does not systematically cultivate a close relationship with DoRRs in all provinces. This is a
reflection partly of the varying capacity and openness of the heads of these offices. However, if UNCHR
seeks to have a coherent program, it needs a coherent approach to selection that involves local
authorities in an open manner. If not, these authorities could eventually make the process more difficult
for UNHCR by intervening indirectly with communities, exerting pressure on potential beneficiaries and
their community leaders, such incidents having been uncovered in the field.
Increase involvement of sectoral ministries to link up the shelter activities with other on-going rural and
urban programmes, and to enhance the provision of basic services and community-based programming.
The provision of such services is not responsibility of UNHCR – but it is the responsibility of UNHCR to
advocate and coordinate by involving sectoral line ministries to inform them of the locations, baseline
studies and needs assessments in shelter programme locations.
b. Involvement of Community Development Councils (CDCs)
Partnership with CDCs
CDCs have been effective entities through which to deliver MRRD’s NSP programme and are now being
used by other counterparts as well. They can play a key role in linking the shelter programme with access
to basic services. In this case, UNHCR does not need to fund these CDCs for implementation of activities
but instead focus on capacity building to make sure that these CDCs understand the shelter programme,
its selection and objectives, and why proper linkages to services is a crucial determinant of the
sustainability of the programme. To achieve this, NNGOs can be tasked with meetings and trainings of
CDCs, and organizing meetings between CDCs and BSCs to better integrate the shelter programme within
broader initiatives and NPPs at the community level.
Partnerships with NNGOs
Rather than on technical capacity building, UNHCR can work on raising the awareness of its national
partners in several fields:
 On baseline studies and needs assessments
 On vulnerability: The ability of national IPs to identify vulnerable households and on

communicating about vulnerability criteria with communities should be drastically improved.
 On humanitarian values and objectives: UNHCR should work closely with its IPs at the
provincial level to progressively change ‘business like’ organisations into proper NGOs
serving longer term humanitarian objectives
These partnerships should exclude one organization from being both implementer, and monitoring
agent, even if these responsibilities are split in different provinces.
183
Partnership with INGOs
In case UNHCR decides to wind down its role in the implementation of shelter assistance, it should start
by fostering the involvement of INGOs in this type of activities. Double partnerships, whereby a NNGO
and an INGO work together on the implementation and the monitoring of the shelter programme, could
be envisaged to prepare for the progressive reduction of UNHCR direct role.
UNHCR should focus on its mandate as a humanitarian agency and on addressing the issue of
vulnerability instead of engaging in development programmes it does not have the structural framework
to implement.
 Here shelter plays a pivotal role on the humanitarian-early recovery – development

spectrum, a transitional role that can allow UNHCR to build the basis on which other actors
can then provide livelihoods, services etc.
Also strengthen partnership mechanisms and cooperation with development organizations and
agencies to effectively link assistance and development.
 As pointed above, shelter can be seen as a “transitional activity” that will allow UNHCR, as
a humanitarian actor, to effectively link up its work with those of national authorities and
development actors.
11. MONITORING FRAMEWORK

a. Internal monitoring - Increased involvement of UNHCR staff
Especially in less risky areas as a key factor in the monitoring and evaluation of the SAP. The over-
reliance on IPs and local partners presents advantages but can also be detrimental in cases where they
are not supervised by UNHCR in the field. The presence of IPs should not replace UNHCR’s direct
involvement in the field – especially in accessible areas like the Central region. Instead, the greater
access can be used as a means to pilot new monitoring systems to constantly evolve. Monitoring is an
organic and interactive process that needs constant revision and adapting to the context. Rules are
curbed and avoided in implementation – one way to avoid that is to strengthen the rules continuously.
Community-based monitoring systems have been tested in Afghanistan and proven their efficiency on
humanitarian and development programs, including on large-scale infrastructure programs as seen with
the work done by the NGO Integrity Watch Afghanistan.
Community based monitoring is not exclusive – it should be seen as a complementary tool to other types
of monitoring such as internal monitoring by UNHCR and external monitoring and evaluations, such as
the present evaluation, undertaken by independent and objective observers and experts. Community-
based monitoring relies on focal points within each community that can liaise directly with UNHCR staff
184
to report problems – whether directly over the phone, through meetings or through the availability of
complaints boxes in each community. The latter has been used as a tool by organizations such as NRC in
Afghanistan as part of their improvements in M&E standards.
Community-based monitoring provides a sense of ownership to communities who have an interest and
stake in making sure that programmes are well implemented – whether looking at proper selection
processes, proper implementation or proper delivery of construction materials in this case. Communities
can be provided with kits or checklists so that they know which elements of the process are important:
are the most vulnerable being targeted? Are shelters compliant with quality standards? Are there proper
risk mitigation procedures in place in case of floods? A full checklist or quality control kit – to be handed
to communities with an initial training – will not only raise their capacity but also their ability to build
better shelters.
Such community-based monitoring mechanisms are particularly useful to ensure that communities have
a voice – especially the most vulnerable within each communities, those whose voices are often the least
heard when going through more routine monitoring procedures.
These mechanisms should be replicated by UNHCR as they provide several advantages: i) strengthening
the link and relationship between UNHCR and the communities it serves, ii) avoiding an over-reliance on
implementing and local partners, as well as local authorities for information, iii) giving a voice to the
more vulnerable in M&E procedures and iv) increasing UNHCR’s capacity to triangulate and cross-check
information, especially in less accessible areas.
These mechanisms will allow UNHCR to identify the difficulties faced during construction and elaborate
adapted responses in a more interactive manner – making UNHCR also more accountable to
communities.
Stricter guidelines should be applied to limit the risks of corruption and fraud. The proximity between
UNHCR provincial staff and implementing IPs should be counter-balanced by external and independent
monitoring partners. Monitoring should be properly budgeted so as to a) avoid under-staffing and b)
increase the scope of monitoring to include qualitative analysis about the programme and responsibility
to identify potential frauds.
As recent investigations have shown, corruption and fraud are a risk for agencies like UNHCR in the field.
A recent MEC report has requested that UNHCR undertakes a “vulnerability to corruption” analysis. This
type of analysis can be extended to the SAP – a large nationwide program that involves a range of actors,
from the national to the local level. Given lessons learned in the Afghan context and the importance of
land and shelter, guidelines on corruption and fraud should be reinforced.
185
Follow up on the beneficiaries after handover with a checklist of M&E indicators
The lack of such a monitoring mechanism causes a subsequent serious lack of data on the outcomes of
the programme. This is problematic in terms of measuring its impact and assessing whether the
envisaged objectives and outcomes have been met. This should be done longitudinally with a set
calendar of field visits every 3 months during the 1st year following the shelter construction.
Deriving from the previous point is the necessity to have strong needs-assessment and follow-up
mechanisms to make sure the needs and expectations of beneficiaries are addressed. The case of Aab
Dara in Kabul province is a strong indication that an opportunity has been missed.
186
BIBLIOGRAPHY
KEY DOCUMENTS
Shelter
ASHMORE, Joseph et. al, “Diversity and Adaption of Shelters in Transition Settlements for IDPs’ in
Afghanistan”, Disasters, 2003, 27(4): 273–287.
ASHMORE, Joseph & FERRETTI, “Shelter evaluation Afghanistan Final Report”, NRC, 2010.
BUGNION & LAURENT, “External Evaluation of the UNHCR Shelter Program in Rwanda 1994 –
1999”, 1994
CRISP, Jeff, & TENNANT, Vicky, “Banking on Solutions: a real-time evaluation of UNHCR’s shelter
grant programme for returning displaced people in Northern Sri Lanka”, PDES, 2010
FOLEY, Pat, “Shelter Programme Monitoring and Evaluation Final Report”, UNHCR, 2005.
Mamadou DIAN BALDE, Jeff CRISP, Ewen MACLEOD, Vicky TENNANT “Shelter from the storm: A
real-time evaluation of UNHCR’s response to the emergency in Côte d’Ivoire and Liberia”, PDES,
2011.
Ministry of Foreign Affairs of Denmark, “Evaluation of the Danish Regions of Origin Support to
Afghanistan”, Tana, May 2012.
Humanitarian Access
BARNETT, Michael N. and FINNEMORE, Martha, “The Politics, Power, and Pathologies of
International Organizations”, International Organization, Vol. 53, No. 4 (Autumn, 1999), pp. 699-
732.
BERNARD, Vincent, “The Future of Humanitarian Action”, International Review of the Red Cross, pp.
891-897, December 2011.
Felix SCHMENDIMANN, “The Legal Framework of Humanitarian Access in Armed conflict”,

International Review of the Red Cross, 93, December 2011, pp. 993-1008.
FERRIS, Elizabeth, “Megatrends and the future of humanitarian action”, International Review of the
Red Cross, pp. 915-938, December 2011.
ICRC, “The Future of Humanitarian Action”: International Review of the Red Cross, 93, issue 884,
187
December 2011.
ICRC, “Discussion: What are the future challenges for humanitarian action?”, International Review
of the Red Cross, 93, pp. 89991, 2011.
SAMUEL HALL, “Redefining Humanitarian Assistance in Afghanistan – A Contextual Analysis”, for the
World Food Programme (WFP) Afghanistan, 2011.
TERRY, Fiona, “The International Committee of the Red Cross in Afghanistan: reasserting the
Neutrality of Humanitarian Action”, International Review of the Red Cross, 93, March 2011, pp. 173-
188.
Return and reintegration
ARNE, STRAND, et al., “Return in Dignity, Return to What? Review of the Voluntary
Return Programme to Afghanistan”, CMI Report #6, 2008.
BALLARD Brett: Reintegration programmes for Refugees in South-East Asia, Lessons learned from
UNHCR’s Experience, April 2002.
BIALCZYK, Agata, “Voluntary Repatriation and the Case of Afghanistan, Critical Examination”,
Refugee Studies Centre Working Paper no. 46, Oxford University, January 2008.
BLACK, Richard & GRANT, Saskia, “Sustainable return in post-conflict contexts”, International
Migration Vol. 44 (3) 2006.
CRISP, Jeff “Mind the gap! UNHCR, humanitarian assistance and the development process”,
Working Paper No. 43, New Issues in Refugee Research, PDES, 2001.
De BREE, June, “Return Migration to Afghanistan, Monitoring the Embeddedness of Returnees”,

CIDIN/ AMIDSt, Nijmegen, January 2008.
JAMAL Arafat: Refugee Repatriation and Reintegration in Guatemala, Lessons Learned From
UNHCR’s Experience, UNHCR, September 2000
MAJIDI Nassim, « Home Sweet Home! Repatriation, Reintegration and Land Allocation in
Afghanistan », in Guerre de Terre, Terre de Guerre (Land and War) published in Revue des Mondes
Musulmans et de la Méditerranée (REMMM), Forthcoming.
IDPs
188
Brookings-Bern Project on Internal Displacement, “Protecting Internally Displaced Persons: A
Manual for Law and Policymakers”, 2008.
Brookings, NRC-iDM, “IDPs in protracted displacement: Is local integration a solution?”, Report from
the Second Expert Seminar on Protracted Internal Displacement », 19-20 January 2011, Geneva
Brookings-LSE Project on Internal Displacement, “From Responsibility to Response: Assessing

National Approaches to Internal Displacement”, 2011.
IASC, (2007), handbook for the protection of internally displaced persons
IASC – Brookings-Bern project on Internal Displacement, (2010), Framework on durable solutions

for internally displaced persons,
MAJIDI, Nassim, “Urban Returnees and Internally Displaced Persons in Afghanistan”, in 2011
Responding to Conflict-Induced Displacement in Protracted Refugee Situations: Middle East
Institute (MEI) and the Foundation pour la Recherché Stratégique (FRS), “Pathways to Enhancing
Transatlantic Cooperation”, January 2011.
NRC/IDMC/JIPS/Samuel Hall, “Research Study on the Protection of Internally Displaced Persons in

Afghanistan”. NRC, 2012.
SAMUEL HALL, “Sustaining the Working Poor in Kabul Informal Settlements”, Solidarités
International Afghanistan, 2012.
SAMUEL HALL, “Cash based assistance for IDPs in the Kabul Informal Settlements”, for the Danish
Refugee Council (DRC) Afghanistan, forthcoming (2013).
The World Bank / UNHCR, “Research Study on IDPs in Urban Settings”, 2011.
LAS, land and housing
ALDEN WILY, Liz, “Land Rights in Crisis: Restoring Tenure Security in Afghanistan”, AREU, 2003.
ALDEN WILY, Liz, “Land and the Constitution: Current Land Issues in Afghanistan”, AREU, 2003.
KELLY, Jeremy, “Afghan project failing in a town called AliceGhan”, The Australian, June 15 2010
MACDONALD, Ingrid, “Afghanistan’s reintegration challenges land and housing”, Norwegian

Peacebuilding center, 2010.
MACDONALD, Ingrid, “Landlessness and Insecurity: Obstacles to Reintegration in Afghanistan”,

Middle East Institute, 2011
189
MAJIDI Nassim, « Home Sweet Home! Repatriation, Reintegration and Land Allocation in
Musulmans et de la Méditerranée (REMMM), Forthcoming (2013).
Presidential Decree No. 104, “On Land Distribution for Settlement of Eligible Returnees and IDPs”,
December 2005.
REED, & FOLEY, “Land and Property: Challenges and Opportunities for Returnees and IDPs in
Afghanistan”, unpublished report prepared for NRC (June 2009).
UNDP, “Lessons Learned Report AliceGhan 1 September 2006 – 31 December 2009”.
Cash and vouchers
HARVEY, Paul, “Cash and Vouchers in emergencies”, Humanitarian Policy Group, London, 2005.
HARVEY, Paul, “Cash-based response in emergencies, Humanitarian Policy Group, London, 2007.
HARVEY, Paul & HOLMES, Rebecca, “The potential for joint programmes for long-term cash
transfers in unstable situations”, Report commissioned by the UK Department of International
Development, London, 2007.
HOFMANN, Charles Antoine, “Cash transfer programmes in Afghanistan: a desk review of current
policy and practice”, HPG Background paper, June 2005.
LUMSDEN, Sarah: Cash for Work Programming, A Practical Guide, OXFAM, Great Britain,
SAMUEL HALL, “Humanitarian Assistance through Mobile Cash Transfer in Northern Afghanistan”,
for the Department of International Development UK (DFID), forthcoming (2013).
SAMUEL HALL, “Cash based assistance for IDPs in the Kabul Informal Settlements”, for the Danish
Refugee Council (DRC) Afghanistan, forthcoming (2013).
TROGER, Franziska and TENANT Vicky: The Use of Cash Grants in UNHCR Voluntary Repatriation
Operations, Report of a Lessons Learned Workshop, PDES, 2008/09.
Monitoring and Remote Management
A.STODDARD, A. HARMER, J.S. RENOUF, “Once Removed, Lessons and Challenges in Remote
Management of Humanitarian Operations for Insecure Areas”, Humanitarian Outcomes, February
2010.
NORMAN, B, “Monitoring and Accountability practices for remotely managed projects implemented
190
in volatile operating environments”, Humanitarian Innovation Fund, Tearfund, 2012.
Office for the Coordination of Humanitarian Affairs, “To Stay and Deliver, Good Practice for
humanitarians in complex security environments”, Policy and Studies Series, 2011.
SAMUEL HALL, “Humanitarian Assistance through Mobile Cash Transfer in Northern Afghanistan”,
for the Department of International Development UK (DFID), forthcoming (2013).
Vulnerability
AHMED Maha, “Defining, Measuring and Addressing Vulnerability: The Case of Post Conflict
Environments”, Maastricht University, November 2010.
ALKIRE Sabina, FOSTER James, “Counting and Multidimensional Poverty Measurement”, OPHI
Working Paper No.7, University of Oxford, 2007.
ALKIRE Sabina, SANTOS Maria Emma, “The Multidimensional Poverty Index: Contradictions and
Analysis”, October 2011.
Afghanistan
ABBASI-SHAVAZI Mohammad Jalal, GLAZEBROOK Diana, JAMSHIDIHA Gholamreza, MAHMOUDIAN

Hossein and SADEGHI Rasoul : Return To Afghanistan? A Study of Afghans Living in Tehran, AREU,
Faculty of Social Sciences – University of Tehran, June 2005.
BIALCZYK, Agata, “Voluntary Repatriation and the Case of Afghanistan, Critical Examination”,
Refugee Studies Centre Working Paper no. 46, Oxford University, January 2008.
CMI, “Return with Dignity, Return to What? Review of the voluntary return program to
Afghanistan”, June 2008.
De BREE, June, “Return Migration to Afghanistan, Monitoring the Embeddedness of Returnees”,

CIDIN/ AMIDSt, Nijmegen, January 2008.
GHOBADI, Negar, KOETTLE, Johannes and VAKIS, Renos, “Moving out of Poverty: Migration Insights
from Rural Afghanistan”, AREU, 2005.
LUMP et. al, “Voluntary Repatriation to Afghanistan Key Features”, Refugee Survey Quarterly 23(3):
149-17, 2004.
MAJIDI Nassim, « Home Sweet Home ! Repatriation, Reintegration and Land Allocation in
Musulmans et de la Méditerranée (REMMM), Forthcoming (2013).
MAJIDI, Nassim, “Returnee Reintegration – What are the standards?” in Reintegration Workshop,
organized by the Ministry of Refugees and Repatriation (MoRR), December 13-14, 2011.
MAJIDI, Nassim, “Evaluation of the IOM programme on the Socio-economic reintegration of

191
Afghans Returning from Iran and Pakistan (RARIP)”, Internal Report not for dissemination, IOM
Afghanistan, 2009.
MAJIDI, Nassim, “Research Study on the Coping Strategy of Returned Refugees in Urban Settings,”
Norwegian Refugee Council Afghanistan, Altai Consulting, March 2010.
MAJIDI, Nassim, “Research Study on Afghan Deportees from Iran”, Altai Consulting for ILO-UNHCR,
March 2008.
MARDSEN, Peter and TURTON, David, ‘Taking Refugees for a Ride? The Politics of Refugee Return to
Afghanistan’, AREU, 2002
MONSUTTI, Allesandro, “Afghan transnational Networks: Looking Beyond Repatriation”, AREU,

2006.
SAITO, Mamiko, ‘Dilemmas Between Borders: Experiences of young Afghans returning “Home” from
Pakistan and Iran, AREU, 2009.
SAITO, Mamiko: From Disappointment to Hope: Trnasforming Experiences of Young Afghans

Returning “Home” From Pakistan and Iran, AREU, 2008.
SCHMEIDL, S. “Repatriation to Afghanistan: durable solution or responsibility shifting?” Forced

Migration Review, Issue 33, September 2009.
OZERDEM& SOFIZADA, “Sustainable Reintegration to Returning Refugees in Post-Taliban

Afghanistan: Land and Related Challenges,” Conflict, Security and Development 6, no. 1 (2006): 75–
100.
General Bibliography
United Nations Guiding Principles on Internal Displacement, 1998.

UNHCR Shelter Programme Monitoring and Evaluation, 2005.
UNHCR, Handbook for Repatriation and Reintegration Activities, May 2004.
“Persons with Specific Needs (PSN) Assistance Guidelines”, August 2012.
UNHCR Shelter Guidelines 2008
Other case studies

BALLARD Brett: Reintegration programmes for Refugees in South-East Asia, Lessons learned from
UNHCR’s Experience, April 2002.
BARRTCH & DUALEH “The protracted refugee situation in Eastern Sudan”, in Refugee Cooperation,
192
Middle East Institute, Recherché pour la Foundation Stratégique, March 1, 2011
CRISP Jeff, TENANT Vicky: Banking on Solutions: A Real-Time Evaluation of UNHCR’s Shelter Grant
programme for returning displaced People in Northern Sri Lanka, PDES, march 2010.
FAGEN WEISS, Patricia, “Refugees and IDPs after Conflict Why They Do Not Go Home”, USIP, 2011.
HABIBI, Gulbadin and HUNTE Pamela, “Afghan Returnees from NWFP, Pakistan to Nangarhar
Province”, AREU, 2006.
HAMMOND, L, The Discourse of Repatriation, in “The End of the Refugee Cycle” edited by Richard
Black and Khalid Koser, 1999.
JAMAL Arafat: Refugee Repatriation and Reintegration in Guatemala, Lessons Learned From
UNHCR’s Experience, UNHCR, September 2000.
KRUSE Stein-Erik, SUJI Omeno, RUGADYA Margaret: Review of Livelihoods and Economic Recovery
in Northern Uganda (LEARN) , Nordic Consulting Group, December 2009
THOMAS-JENSEN Colin: Crisis and opportunity; Protracted Displacement in Sudan, Middles East
Institute, Fondation pour la Recherche Stratégique, May 2011.
193
ANNEX 1: HOUSEHOLD SAMPLING BY DISTRICTS AND CATEGORIES OF RESPONDENTS
UNHCR Non - Non-UNHCR Total Number of

Region Province District
Beneficiaries Beneficiaries Beneficiaries Respondents
Dih Sabz 117 130 -
Kabul Paghman 48 34 - 384

Central
Qarabagh 32 21 2
Parwan Baghram 101 85 1 187
Central Highland Bamyan Bamyan District 32 29 - 61
Mihtarlam 68 72 -
Laghman 300
Qarghayi 94 66 -
Jalalabad 248 313 90
Behsud 377 190 20

East
Surkh Rod 120 233 164
Nangarhar 2 068
Mohmandara 13 25 43
Khewa 14 30 67
Kama 18 32 71
Sholgara 21 23 -
Balkh Dawalatabad 21 14 - 101
Nahree Shahee 8 14 -
Qaysar 28 25 2
North Faryab Khoje Sabz Posh 24 35 - 174
Andkhoy 23 37 -
Shibirghan 24 34 2
Jawzjan 220
Khawja du Khow 94 66 -
Sari Pul Sozma Qala 56 44 - 100
Kunduz Kunduz District 60 60 - 120

Northeast
Takhar Taloqan 39 31 - 70
Helmand Lashkar Gah 56 52 - 108
South Kandahar 26 21 1
Kandahar 155
Arghandab 49 58 -
Southeast Paktya Gardez 123 117 - 240
West Hirat Injil 101 99 - 200
Total 2 035 1 990 463 4 448
195
ANNEX 2: LIST OF KEY INFORMANT INTERVIEWS
Province Level Name Organisation Position
CENTRAL REGIONS
Kabul UN-Agency /IO Tomislav Babic & Ivan IOM
Ceko
Kabul NGO Haqman Munib & Eng. CARE Provincial Programme
Khawani Officer
Kabul Donor Jacques Dailoux ECHO Technical Adviser
Kabul NGO Subhash Jadhav NRC Shelter Programme

Manager
Kabul UNHCR – Central Iljia Todorovic & Stijn UNHCR Deputy Representative for
Afghanistan
Kabul UNHCR – Sub Office Gulham Sediqqi UNHCR Associate Reintegration
Kabul Officer
Kabul UN Agencies/IO Eng. Rahman UN-Habitat Project Manager
Kabul Donor Gul Joya Jafri CIDA First Secretary

(Development)
Kabul Donor Ted Jasik & Zmaryalai USAID/BPRM Refugee Coordinator
Farahi
Kabul UNHCR – Central Saradarwali UNHCR Associate Data Manager
Kabul Donor David Stewart Australian First Secretary

Embassy
Kabul NGO Jolien Veldwijk ZOA Programme Advisor
Kabul UNHCR-Central Jasmine Ketabchi UNHCR Protection Officer
Kabul UN-Agency/IO Jan Turkstra UN-Habitat Advisort Urban

Development
Kabul UNHCR – Central Robin Ellis & Malang UNHCR Head of Programme
Ibrahimi Reintegration Unit
Kabul UNHCR – Central Migena Tane UNHCR Associate Programme
Officer
Kabul UNHCR – Kabul Sub Douglas DiSalvo UNHCR Protection Officer
Kabul UN Agency Sara Otuku OCHA Humanitarian Assistance

Officer
Kabul UNHCR-Central Nematullah UNHCR Programme Unit
Kabul GoA Abdul Rahman Shams DoRR Director of DoRR Kabul
Parwan IP Ibrahim Haydari ABR Director
Parwan GoA Aghai Shirin DoRR Director
EASTERN REGIONS
Nangarhar UNHCR Sub Office Eng. Shirzai UNHCR Shelter Unit Officer
Mohammad Eamal Liaison Officer
Salamat Alieva Associate Programme
officer
Nangarhar UN Agency Dr. Rahimi UN-Habitat Senior Provincial Manager
Nangarhar GoA AlHaj Ghulam Haidar DoRR Director

Faqirzai
Nangarhar GoA Harzat Bilal Nurestani DoRR Director
Nangarhar UN Agency/IO Sadruddin Hasma Safi IOM Officer in Charge
Nangarhar IP Eng. Mahboob AGHCO
Nangarhar NGO Abdur Rauf IRC Humanitarian Programme

Assistance
Nangarhar IP Eng. Hafiz APWO Director
Nangarhar NGO Mr. Granagah CHF Former Programme

Manager
Nangarhar GoA Sayeed Qias Saeedi Directorate of Director
Economy
Nangarhar GoA Amir Mohammad Zabid Provincial Council Member of PDC
Nangarhar IP Eng. Shafiq APA Programme Manager

Abdul Wali Field Officer
Naqibullah Admin. Officer
Nangarhar NGO Mark Nagle NRC Area Coordinator
Nader Surgand Shelter Programme
Sabawoon Saba Manager
Field Engineer
Nangarhar UNHCR Sub-Office Minako Kakuma UNHCR Protection Unit
Nangarhar UNHCR Sub-Office Eng. Shirzai UNHCR Shelter Unit

Mohammad Eamal Liaison Officer
Nangarhar UNOPS Regional Office Sayeed Khaksar Padshah UNOPS Senior Regional Manager
Nangarhar GoA Eng. Mohammad Zakhi DoUDH Director
NORTHERN REGIONS
Balkh NGO Fardeen Hafizi NRC Shelter PC NRC Sari Pul
Balkh IP Eng. Mirwais NOP Programme Officer
Balkh IP Dr. Shahab DHSA Director

Fatma Jawzjan Field Supervisor
Balkh UN Agency Ms Evelyn Viedbock OCHA Head of Sub-Office
Mr. Njedkounkosse HAO
Mr. Hemmat
Balkh GoA Mr. Sabor DoRR Director
Balkh NGO Sayeed Mansoor Islamic Relief Programme Officer
Balkh UNHCR Sub-office Martin Bucumi UNHCR Head of Sub-officer
Faryab UNHCR Abdul Basir Sediktyar UNHCR Provincial Focal Point
Faryab UN Agency Hajj Azziz OCHA Focal Point
197
Faryab NGO Sean Mc GIrk ACTED Project Development
Advisor
Faryab GoA Mohammad Rahim DoRR Head of DoRR
Faryab IP Ilhaya DHSA Field Officer
Faryab UN Agency Mohammad Sakhi UNAMA National Deputy Head of

Office
Faryab NGO Anna Cervi NRC Area Coordinator
Jawzjan GoA Eng. Karim DoRR Director
Jawzjan GoA ENg. Mohammad Yusuf MRRD Head of MRRD
Jawzjan NGO Ali Ahmadi ZOA Shelter Project Manager
Jawzjan NGO Zarmina ACTION AID Project Manager
Jawzjan GoA ANDMA Assistant to the director
Jawzjan UN Agency/IO Azizullah Hashim UN-Habitat PUR
Jawzjan NGO Eng. Munir CARE Field Office Supervisor
SOUTHERN REGIONS
Kandahar IP Pashtun Zarak HAPA Programme Manager
Haji Mohammad Director
Kandahar UNHCR Sub-Office Mr. Jelev UNHCR Head of Sub-Office
Kandahar UNHCR Sub-Office Ahmad Dost UNHCR Programme Associate

Protection Unit
Kandahar GoA Mohammad Azim DoRR Director
Nawabi Head of Reintegration
Abdul Hamid General Manager
Abdul Farooq
Kandahar GoA Timor Sha Ayubi NSP Snr. Provincial Manager
Zarmina Data Entry
Farid Monitoring Officer
Kandahar NGO Ezatullah Save the Children Programme Manager
Kandahar UNHCR Sub-Office Abdul Jalil UNHCR Programme Associate
WESTERN REGIONS
Hirat NGO Azim DRC Protection Officer
Hirat UNHCR Sub-Office Wakil Ahmad Habibi UNHCR Senior Field Associate
Zahar Field Associate
Hirat UNHCR Sub-Office Inge Collijn UNHCR Head of Sub-office
Hirat UNHCR Sub-Office Abdul Karim UNHCR Programme Associate

Programme Unit Officer
Hirat UNHCR Sub-Office Pascal Dhieu UNHCR Associate Reintegration
Officer
Hirat IP Eng. Naeem Karim SSOAR Director
198
Hirat IP Dr. Aziz CRDSA Director
Hirat IP Eng. Abdul Rahman ERSA Director

Sahabee
Hirat NGO KHadeem Hussein NRC Shelter Team Leader
Rahimi
Hirat NGO Naseer HELP Reintegration Officer
Hirat UN-Agency/IO Nematullah Merrikhi IOM Senior Programme Assistant
199
ANNEX 3: PROVINCIAL OVERVIEWS
KABUL – PROVINCIAL OVERVIEW
CONTEXT
Kabul is the province with most assisted return, as UNHCR registered 1,217,624 assisted returnees to
Kabul between 2002 and 2012. 2818 shelters were completed by UNHCR in the entire Kabul province
between 2009 and 2011, with a strong focus on Deh Sabz and Paghman districts. Districts visited by the
research team were Deh Sabz, Paghman, and Qara Bagh.
As Table A.1 shows, our provincial sample of Kabul comprises 384 households, of which more than half,
197 are UNHCR beneficiaries. One household received shelter assistance from the Norwegian Refugee
Council. Moreover, 58 per cent of our sample are official refugee returnees, while 32 per cent are non-
refugee returnees, 4 per cent are IDPs and 6 per cent never migrated. Among the UNHCR beneficiaries
refugee returnees are the largest group with 78 per cent.
Table A.1: Provincial Sample, Kabul

Total

% 77.66 0.00 38.38 58.33
Non-Refugee Returnee 42 0 81 123
% 21.32 0.00 43.78 32.03
IDP 0 0 14 14
% 0 0.00 7.57 3.65
No Mobility 2 2 19 23
% 1.02 100.00 10.27 5.99
Total 197 2 185 384
% 100.00 100.00 100.00 100.00
Given its important political weight, Kabul province is unsurprisingly a very specific case in the Afghan
humanitarian context. There are several features of the province, which impact or should impact
favourably the implementation of the shelter programme and the sustainability of the reintegration of
returnees in the province:
 Kabul province receives by far the biggest amount of humanitarian and development assistance.
The level of investments in infrastructure development is unmatched in the rest of the country.
One of the implementation sites (Bakhtyaran) for example is located in the premises of the
200
project of ‘new Kabul city’ and beneficiaries should benefit from the fallout of this project. The
high number of actors present in the province opens up numerous possibilities for partnerships.
 Most of Kabul province is accessible and secure, facilitating greatly the implementation process,
as recourse to remote management and monitoring is not necessary. It is therefore surprising to
see that despite this favourable security context, the sub-office still relies on remote
management in the province.
 Kabul has benefited from the important economic fallouts related to the presence of the
international community, both humanitarian and military. The labour market and the economic
context are more dynamic in Kabul than anywhere else in Afghanistan, offering large
opportunities of skilled and unskilled employment. The sustainability of this economic dynamism
is yet highly questionable given the planned withdrawal of large parts of the international
community.
Yet, these general comments on the profile of the province only give part of the story as a range of
specific local factors vastly determines the sustainability of the reintegration of returnees. As every other
province, Kabul province is characterised by important intra-provincial disparities in terms of wealth and
development. As will be developed below, the success or failure of shelter programme is highly
dependent on the local context in which it is implemented.
DESIGN
Among the 197 beneficiaries interviewed, the majority, 90 per cent, had built a two-room shelter, while
7 per cent built a one-room shelter and 3 per cent built a different type. Compared to other provinces, it
appears that a large share of respondents, 22 per cent, got to choose the size of their shelter compared
to the national average of 13 per cent.
In most cases, shelters were in a relatively good state, with disparities according to the well-being of
beneficiaries and that of the community as a whole. In Kuchi Abad (Paghman), a reintegration site,
rooms were painted, decorated and well furnished, whereas in poorer and more remote communities
like Aab Dara, rooms only had basic furniture, though the outside appearance of shelters was good.
Despite complaints about the small size of the rooms and the iron door and window frames, which were
described as not adapted to the environment and were enlarged whenever possible, there was an
overall satisfaction with the design of shelters (cf. the case of Aab Dara, where exact replicas of shelters
had been constructed by non-beneficiaries).
Shelters were constructed inside surrounding walls according to a traditional housing disposition with
several related households living around a courtyard on the same premises: a sign that shelter is
considered as permanent housing and integrated in a broader living environment (shared oven (tanur),
common courtyard). In two cases where surrounding walls could not be constructed (Pitawa in Qara
Bagh district), the shelters were not occupied and beneficiaries were living with relatives.
201
The quality of material was praised by beneficiaries as well as non-beneficiaries, which often mentioned
roofing components (iron beams and roofing tiles) could not have been purchased without assistance.
Among the surveyed beneficiaries there was an overall satisfaction with the quality of the material, 91
per cent, and materials were received on time in most cases, 94 per cent.
Problems during construction: Less than half of the beneficiaries, 44 per cent, indicated that they
encountered some kind of problem during the construction of their shelter. Table A.2 shows the
different types of problems the beneficiaries encountered, the main problems being not having enough
financial capital and insufficient access to water.
Table A.2: Problems during Construction, in Kabul

UNHCR Beneficiaries
N=87 %
Ran out of money 74 85.06
Insufficient access to water 36 41.38
Ran out of materials 28 32.18
Lack of skilled labour 27 31.03
Weather problems 19 21.84
Lack of unskilled labour 9 10.34
Lack of technical knowledge 8 9.20
Materials of poor quality 5 5.75
Materials not delivered on time 5 5.75
More specifically, the observations showed that challenges during the building process depended on
environmental and economic conditions: lack of water in dry areas, lack of fuel for the functioning of
water pumps, lack of skills. Beneficiaries did not mention receiving particular support during the
construction process, which was sometimes a cause for indebtedness, as it required hiring additional
skilled labour. There was one mention of bribe to receive material from the IP in Ghazi Abad.
In many cases, latrines were not used. In Aab Dara, they had been constructed outside the walled
compounds, off housing premises, a sign they were not intended for regular use. Conversely, in Kuchi
Abad, where WASH programs were implemented and hygiene classes provided, latrines were used and
in good state, which emphasizes the need for complementary assistance programs.
202
SELECTION OF BENEFICIARIES
The main criteria mentioned for beneficiary selection were land ownership and VRF. Vulnerability as
such seemed not to be properly taken into account as only half of the beneficiaries can be classified as
extremely vulnerable. A remarkable 44 EVI households did apply for assistance but were rejected.
However, the selection process was not homogeneous in various locations visited, and BSCs were only
mentioned in one case:
- In a heterogeneous community like Bakhtyaran (Deh Sabz), the local Malik was not included in
beneficiary selection, allegedly because of risks of factionalism and corruption. The IP played the
major part in the selection process, though BSCs and inclusion of UNHCR representatives was not
mentioned.
- Conversely, in homogenous tribal communities (Aab Dara, Kuchi Abad), the Malik was
responsible for gathering community members and participated actively in the BSC for
identification of eligible people according to their needs and the quota of shelters allocated by
UNHCR (Aab Dara).
- In one case (Pitawa), the Malik was entirely responsible for selection of beneficiaries and
proceeded through a game of chance to choose beneficiaries, which raises concern about the
reason why neither UNHCR nor the IP were involved in the process and illustrates some of the
pitfalls of remote management of the shelter programme. This is especially preoccupying in a
province like Kabul where access is not an issue. It also shows that the rationale behind the
selection criteria of beneficiaries is not properly explained to communities, which sometimes
come up with their own selection arrangement, as it is the case in Pitawa.
In Kuchi Abad, returnees prior to 2007 were not selected for the programme. As illustrated in Table A.3,
this is a claim that is recurrent on behalf of non-beneficiaries, and has been explained elsewhere (e.g.
Nangarhar) by the limited amount of shelters allocated and the decision to focus on recent returnees
based on the assumption that ancient returnees have developed coping mechanisms in the meantime.
Table A.3: Date of Return by Beneficiary Status, in Kabul

UNHCR Non-
Total
2002-2004 17 37 54
% 8.72 24.34 15.56
2004-2008 65 46 111
% 33.33 30.26 31.99
2009-2011 113 69 182
% 57.95 45.39 52.45
Total 195 152 347
% 100.00 100.00 100.00
203
This assumption is not always verified in the field, as returnees can live several years with relatives
and/or migrate regularly without being able to settle. Furthermore, the choice of the year (in this case
2007) seems to be random as there are no specific guidelines in this respect.
UNHCR raised expectations through provision of shelter and winterization assistance, and there is a
general misunderstanding of why the programme was interrupted in 2012, despite on-going return
from abroad. This is a recurrent observation, a sign of a breach in the communication plan of UNHCR
towards communities. This is also a problem in cases where neighbouring villages have not received any
assistance, despite the needs of their inhabitants (Sayed Abad and Sar Allah, neighbouring Kuchi Abad),
creating a feeling of favouritism which can be a potential driver of tensions, especially as assistance was
provided to a Pashtun tribe who had recently settled, while the neighbouring population of returnees
and conflict-driven IDPs, Hazara and Tajik, were left out.
There were no mentions of a complaint mechanism and no follow-up after implementation.
SOCIO-ECONOMIC IMPACT
The primary impact of the programme mentioned was that it provided beneficiaries with a roof, which is
particularly sensitive for recent migrants previously living in tents, which were forced to migrate during
the winter due to harsh weather conditions. Overall, 34 per cent of the beneficiaries indicated that they
lived in a temporary shelter prior to receiving assistance. A shelter therefore appears as a permanent
and safe living facility-allowing people to settle, which seems to be one of the strongest aspirations of
the communities visited. Moreover, beneficiaries reported that due to the shelter assistance their
economic situation was better or far better off in 64 per cent of the cases, indicating a subjective positive
impact on the households.
Compared to non-beneficiaries, our multi-dimensional poverty index (MPI) presented in Table A.4 shows
that those households which received shelter from UNHCR are much less deprived.
Table A.4: Multi-dimensional Poverty Index by Beneficiary Status, in Kabul

Not Deprived Deprived Total
UNHCR Beneficiaries 51 146 197

% 25.89 74.11 100.00
Non-UNHCR Beneficiaries 0 2 2
% 0.00 100.00 100.00
Non-Beneficiaries 29 156 185
% 15.68 84.32 100.00
Total 80 304 384
% 20.83 79.17 100.00
204
The comparative impact between beneficiaries and non–beneficiaries varies according to the profile of
the community. In the case of Kuchi Abad, it appeared crucial to survey neighbouring communities.
Indeed, though the overall impact of the programme on the Kuchi community targeted appeared very
positive thanks to provision of housing, WASH components, a road, school and clinic, the neighbouring
villages had not only been completely left out of the reintegration programme, but suffered from
negative side-effects of construction, had a feeling assistance had been taken away from them, and that
they had been cheated. These considerations in turn affect the sustainability of the programme in the
area.
More generally, at the community level, one of the major effects of shelter assistance was the rise of
land prices due to increasing settlements in areas of implementation of the programme. This was
especially the case in Deh Sabz where the relatively low price of land compared to Kabul attracted
settlement.
SUSTAINABILITY
In most locations visited, there were high rates of occupancy. The beneficiaries interviewed indicated
that they were living in the shelter in all but three cases. A systematic check could not be undertaken in
all location, mainly due to their dispersion amongst existing habitat. In single tribe villages where habitat
was mainly composed of shelters, a more systematic check could be undertaken:
- In Aab Dara, 20 households had allegedly temporarily left their shelters for seasonal labour
migration, according to one of the inhabitants.
- In Kuchi Abad, some beneficiaries had temporarily left for herding.
In both these cases, shelters were in good state, with sealed doors and windows, a sign of an intention to
return rather than abandonment.
- In Pitawa, Qara Bagh, 3 unoccupied shelters were identified, because of the inability of the
beneficiary to construct surrounding walls and the case of a widow unable to complete shelter,
which raises the question of provision of additional assistance to EVIs.
Beneficiaries expressed an overall satisfaction with the programme, which allowed them to access
property and housing, a strong incentive for settlement. Even in cases with low accessibility and lack of
services, shelter was described as an essential need. Seasonal migration was mentioned as a compulsory
mechanism to cope with cold weather conditions in cases where solid housing is not available, and lack
of income opportunities in the vicinity. In such cases however, shelter remained a base for the
household and a point of return for the rest of the year. When interviewed, non-beneficiaries expressed
the same type of opinion. Accordingly, almost all respondents do not have plans to move again from the
current community.
In Aab Dara, the shelter programme had played an important role in the decision of this community to
return from Pakistan. The return of the community was based on a highly organised strategy of re-
settlement, with some representatives sent beforehand to buy the land.
205
Complementary assistance works as a strong condition for the sustainability of the programme, as
villages, which received assistance from various organizations, appeared better off. A more coordinated
approach between various organizations could work as a strong factor for sustainability (avoiding
potential tensions around access to water for instance, providing communities with a school, a clinic,
etc.).
The sustainability of the shelter programme varied according different sets of factors:
- Economic environment: connectivity of the village, facilities, employment opportunities, access

to relatively cheaper land plots, and assistance received. Complementary assistance seems to
have a significant impact on the sustainability of the shelter programme as it provides
communities with facilities and services essential for the creation of a sustainable environment
(water, schools, clinics, roads).
- Social environment: relationship to neighbouring villages. This was an area of concern noticed in
Kuchi Abad: despite the apparent success of the shelter programme for the community, the fact
that neighbouring villages had received no assistance and had been side-lined in the
reintegration activities was a potential driver of tension communities, which is an important
challenge to the apparent success of the programme. This is especially concerning knowing the
history of tensions existing between Kuchi and host communities in the province. The divisive
impact of such assistance strategies must not be underestimated.
According to these sets of factors, three types of profiles were identified:
- In Bakhtyaran, shelters were implemented inside an existing habitat, in a well-connected area

with access to services and a potential labour market. The village was also covered by several
complementary assistance programs from various organisations (ZOA, DACCAR, CARE). The
combination of these different factors contributed to the apparent success of the shelter
programme: beneficiaries were well integrated in the socio-economic environment of the town,
with shelters hardly differentiable from the rest of the habitat and the expression of an intention
to settle on land which was either inherited, or bought given the relatively lower price of land
compared to Kabul.
- In Aab Dara, a tribe representative sent from Pakistan 10 years prior to assistance had acquired
land. Community members expressed a strong will to settle and presented the shelter
programme as a strong incentive for return. Non–beneficiary families looked forward to
potential assistance in order to settle permanently and avoid seasonal migration in the winter
due to harsh weather conditions. This indicates the strong ‘stabilizing effect’ that the shelter
programme can have on communities who traditionally use temporary migrations as a key
coping mechanism. Kuchi tribes are a case in point. However, the lack of connectivity of the
village, off the main axis, in an area offering little job opportunities exerted pressure on the
inhabitants, especially as more families were planning to return from Pakistan. Access to water
was a major concern, as well as insecure job opportunities, with brick kilns closing in the winter
and daily work in agriculture depending on weather conditions.
206
In Aab Dara, shelter assistance was explicitly mentioned as a driver for return. Interestingly, new
returnees had started building their houses according to UNHCR design, and were expecting to
receive the shelter package in order to complete it despite the fact they had not been identified
as beneficiaries. Without UNHCR assistance, community members reported not being able to
purchase the remaining elements to complete their house, especially the roofing material (iron
beams and ceiling bricks). This also shows that the need for shelter assistance is real, as
households were not able to complete their shelters without the support of UNHCR. In areas of
high return, especially in similar cases where shelter assistance is described as an incentive for
return of other tribe members, lack of follow-up after completion might endanger the
sustainability of the programme, as increasing pressure is exerted on natural resources.
Assessment of needs and follow-up after completion of the programme would prevent risks of
secondary displacement.
- The reintegration site of Kuchi Abad is a homogenous Pashtun community which received
substantial assistance as part of the reintegration programme, but is located in a heterogeneous
environment with neighbouring communities receiving no assistance. The latter’s impression of
having been side-lined is a potential driver of tensions:
“We didn’t receive anything. In our village, we are deprived of everything: no road, no clinic, and
no school. They have influence (i.e. The Kuchi), through their Kuchi networks in Parliament and at
the district level. They included us in their statistics to get help, said there were 5000 poor people
in the area. That’s us. Our names were included in a list, but we didn’t get help. For example,
there is a mobile clinic, but it goes only to Koch Abad, we don’t benefit from it. It doesn’t come
here.” (FGD non-beneficiaries, Sayed Abad).
Disparities in allocation of assistance therefore introduce a major breach in the “reintegration” scheme.
In the case of Kuchi Abad, inhabitants from neighbouring Sayed Abad mentioned the construction work
as part of the reintegration scheme had negative effects on their village, corrupting water for instance.
Sayed Abad inhabitants who had no direct access to it deemed the road to Kuchi Abad, presented as a
programme meant to benefit all neighbouring communities, unusable. This strongly emphasizes the
risks when communities are taken individually into account for assistance: taking into account the
insertion of a community in a general environment in order to contribute to the well-being of the
entire area is a requirement. The sustainability born out of the comprehensiveness of the assistance
received by returnees through the reintegration strategy must be weighed against the risks of inter-
community tensions that it fuels.
PARTNERSHIPS
Apart from a contribution of InterSOS on WASH programs in Kuchi Abad, there seem to be no particular
links between UNHCR and other NGOs/agencies in the field in Kabul.
In Aab Dara, ZOA for WASH and German Cooperation (temporarily) for a school had been present, but
there seems to be no cooperation with UNHCR.
207
Kabul DoRR complained about the lack of coordination and relationship with UNHCR, claiming that the
DoRR had very little involvement either in the selection of beneficiaries or in the monitoring of the
programme.
208
PARWAN – PROVINCIAL OVERVIEW
CONTEXT
Parwan province is located on the North of Kabul province and is easily accessible through the Salang
Road, the main trading route to the North of the country. The area has been affected by heavy fighting
during the war and was consequently depopulated. Though the security context is degrading in Sheikh
Ali, Shiwari and Ghorband districts since the handover of the security by ISAF to the Afghan government
in December 2011, the visited Bagram and Charikar districts are generally secure. The province has
witnessed significant return of refugees44 since the fall of the Taliban, as well as an influx of IDPs from
other provinces escaping insecurity (Bamyan, Laghman, Kapisa and Nuristan). The central valley is
occupied by farmlands with a diverse agricultural production, which is traded in Kabul or Charikar, the
provincial capital. There is however a significant difference in access and employment opportunities in
the Central Highland Range in the north of the province, which presents a dry steppe-like landscape with
scarce accessibility to water due to the permeability of its soil, and which is off the main axis of
circulation.
Both the accessibility of the area and its strategic location can explain for the high concentration of
assistance provided to the villages visited, both from INGOs, international donors, and MRRD. The
province had previously benefitted from a land allocation scheme of 1100 plots directed at returnees
(2005). However, some communities complained about not having received specific forms of assistance
due to lack of networks and tribal favouritism (e.g. Qala-e Nasro and Khanaqa).
The proximity of Bagram Air Base has an economic impact on the area as it offers highly remunerated job
opportunities to qualified inhabitants, and provides others with unqualified jobs which are highly valued
and are subject to high competition. There is however considerable concern about the sustainability of
these jobs after the retreat of foreign troops. Another consequence of the proximity of Bagram Air Base
was the provision of assistance to some villages by Provincial Reconstruction Teams (PRTs).
In total, 1,315 shelters were completed by UNHCR over the entire province in 2009-2011. As Table A.5
shows, our provincial sample of Parwan comprises 187 households, of which more than half, 101, are
UNHCR beneficiaries. One household received shelter assistance from IOM. Moreover, some 49 per cent
of our sample are official refugee returnees, while 37 per cent are non-refugee returnees, 10 per cent
are IDPs and 4 per cent never migrated.
44
163,581 assisted returns between 2002-2012, according to UNHCR data from encashment Center.
209
Table A.5: Provincial Sample, Parwan
Total
% 66.34 0.00 29.41 49.20
% 19.8 0.00 57.65 36.90
IDP 12 1 6 19
% 11.88 100.00 7.06 10.16
No Mobility 2 0 5 7
% 1.98 0.00 5.88 3.74
Total 101 1 85 187
% 100.00 100.00 100.00 100.00
DESIGN
Most beneficiaries, 90 per cent did not personally choose the size of their shelter. The main design
implemented was the standard two-room shelter, 91 per cent, while a minority build one-room shelters,
9 per cent. In one case, a beneficiary mentioned having received “half a shelter” as she did not have a
VRF and had been selected as a woman head of household, which raises questions about the criteria
used for allocation of a specific design of shelter. The main complaints about the design were the size of
the rooms, considered too small for a traditional family, and the size of windows.
Materials were received on time in most cases, 95 per cent, and there was an overall satisfaction with
the quality of the material, 93 per cent. The roofing components were especially appreciated, as
beneficiaries mentioned they would not have been able to purchase iron beams and roof tiles
themselves. This was mentioned as a great advantage for beneficiaries, as non-beneficiaries either had
to get more indebted to have the same material, or to use lower quality items. Shelter assistance
therefore has a positive impact on the economic situation of beneficiary households, and further
attention is probably worth being paid to debt practices as a coping mechanism in Afghanistan. Wooden
doors and glass panes for windows were said not to have been provided with the package.
There is a disparity in the state of shelters according to the well-being of the households: the better off
had furnished and painted rooms, whereas the most vulnerable had not been able to buy glass window
panes preferring using the cash grant to address more immediate needs, including food and water, and
covered the iron frames with plastic sheets. This can be understood, as a side effect of the fact
vulnerability is not mainstreamed as a criterion in selection, with the initial situation of the beneficiary
(more or less vulnerable) having a consequent impact on their capacity to build, maintain, and
rearrange the shelter. In one case (Qala-e Nasro), the beneficiary had not been able to complete the
construction of his shelter, as he was busy supporting his household, which raises the question of
possibility of additional assistance for the most vulnerable – especially since ashar is seldom considered
as an option.
210
A significant proportion of shelters were used for purposes other than living space: storage of crops
and farming tools, and in one case, the shelter had been turned into a shop in addition to an existing
house. This is a sign beneficiaries selected were not necessarily the ones in urgent need of shelter.
Overall, however, shelters were clearly an essential component of the household’s living environment,
and were upgraded whenever the economic situation of the beneficiary allowed him/her to do so,
displaying an intention of long-term settlement.
Changes in the design: Whenever possible, iron frames were replaced by wooden frames and rooms
were enlarged. In Khanaqa, there were two cases where the shelters had entirely been dismantled to use
the package material for building another larger house. The original purpose of components of the
shelter package had been changed in some cases: latrines, for instance, were in many cases transformed
into tanur-houses (for cooking) or storage space for agricultural products or combustibles. In one case,
they had been constructed outside the compound and exclusively used for storage.
Problems with the construction: More than one third of the beneficiaries (34.7%) indicated that they
encountered some kind of problem during the construction of their shelter. Table A.6 shows the
different types of problems the beneficiaries encountered, the main problems being having enough
financial capital and a lack of skilled labour.
Table A.6: Problems during Construction, in Parwan

UNHCR Beneficiaries
N=35 %
Monitoring is done throughout the construction process and until handover (five times according to the
DoRR) by the IP in partnership with DoRR representatives. The survey data shows that 12 per cent of the
UNHCR beneficiaries in Parwan indicated that there was no monitoring of their shelter by the
organization that assisted them. There seems to be no follow-up after completion.
211
 Selection Process
The selection on the field seems to have been mainly conducted by ABR, the implementing partner, in
partnership with DoRR. This is an area of concern since Parwan is a secure and accessible area, which
does not require remote management or remote monitoring. The Kabul sub-office in charge of
administering the province, which mentions UNHCR staff rarely attends BSCs due to “security
restrictions”, confirms this. This is problematic, since, as mentioned earlier, the security situation was
good in the districts visited during evaluation. This calls for a better understanding of the conditions
under which UNHCR conducts remote management.
Non-beneficiaries repeatedly mentioned networks and influence as a requirement for selection,

apparently because of a heavy reliance on maliks for gathering of eligible inhabitants and beneficiary
selection. In Khanjar Khil, where habitat is scattered, some mentioned not having been warned of the
existence of the programme and not having been able to apply for selection on time. This is revealing of
a breach in the selection process as remote areas are not sufficiently taken into account and
insufficient attention is paid to communication about the shelter programme. This contributes to a
great confusion about the criteria of eligibility and the selection process.
In homogenous communities, reliance on the Malik for selection was not perceived as a driver of
conflict, thanks to solidarity mechanisms. In Khanaqa however, the fact that only one of the Malik out of
the two tribes in the community had been consulted created tensions as one part of the village received
substantial assistance – indiscriminately- and the other didn’t receive any, despite the fact, according to
local authorities and community members, that some would have been eligible for assistance.
Mention of women participation in the selection process was met with incomprehension/ regarded as
unnecessary by community members. However, ABR did specify a female member of their staff
participated in selection and conducted hygiene classes as part of the shelter programme.
Two cases of fraud have been noticed in Khanaqa, with one household receiving two shelters (husband
and wife), and one other where the occupant had bought the VRF and the assistance “package” that
goes with it from a wealthier community member.
 Criteria of Selection
The main criterion used for selection was ownership of a VRF, on a “first come first serve” basis. More
than 80 per cent of the beneficiaries indicated having a VRF form and 66 per cent were recognized
refugee returnees. Furthermore, there was no explicit mention of a beneficiary selection committee in
the field, and most of those who were warned and had a VRF received a shelter, regardless of their
economic situation, in all villages visited. The vulnerability criterion was seldom mentioned - which can
be an explanation for the fact that some shelters are used for purposes other than living. The
quantitative data shows that 74 households or 40 per cent of the sample fulfil one of the EVI criteria.
212
More than half of these (41 households) are beneficiaries of the shelter assistance programme. 15 EVI
households applied for assistance, but were not chosen as beneficiaries.
The fact that returning IDPs were not included in shelter assistance, as they did not have a VRF raised
concern in the communities, as those who were not able to migrate abroad were also those with fewer
resources. In Khanaqa for instance, though land did not seem to be an issue, as both returnees and IDPs
were returning to their village of origin, no shelter was allocated to returning IDPs. There is therefore no
clear understanding of why they were not considered as eligible in the programme. Failure to effectively
select IDPs weakens the ability of the programme to reach the most vulnerable. IDPs from other
provinces are excluded from the programme, allegedly since they do not own land and are seen by
authorities as temporary migrants.
 Complaint Mechanism
There was no complaint mechanisms mentioned. Several non-beneficiaries mentioned having gone to
the IP’s office but not having been given any clear information. ABR reports having conducted campaigns
prior to the selection process, warning the population of distribution of assistance and information about
the selection criteria. However, the fact that community members continue to believe that VRF is the
main requirement for receiving assistance shows the message has not been understood. More attention
needs to be paid to the publicity of the selection process and communication plan of UNHCR before,
during and after selection.
The socio-economic impact of the shelter programme in the different locations visited depends on the
profile of beneficiaries, which appears to come across a wide range of situations, as the main criteria is
less vulnerability than holding a VRF.
Overall, the economic situation of beneficiaries did not differ much from that of non-beneficiaries, and
was as diverse. The sole focus on VRF therefore raises concern about the efficiency and cost
effectiveness of the programme, which leaves out other vulnerable members of the community in need
of shelter (IDPs, non migrants, undocumented returnees), but integrates people who either have the
capacity to build their own house and use the shelter for purposes other than living (guesthouse, storage
room). In one case in Khanaqa, a beneficiary had previously benefitted from shelter assistance and was
using his shelter for storage. Now he was building an additional two-room house opposite the UNHCR
shelter, indicating flaws in the selection process. There is therefore a direct link between the impact of
the programme and the fairness of the selection process.
In our surveyed sample, 72 per cent of households noted their economic situation had been better or far
better following receiving shelter assistance. Shelter was explicitly mentioned as an essential need as
housing had been destroyed during the war or deteriorated due to long periods of absence, and as there
is an explicit will to settle all the more since the security situation is stable. Shelter assistance therefore
upgraded the living conditions of returnees who previously lived in tents (35.35% of UNHCR beneficiaries
213
in our survey sample), and in some cases (Qala-e Nasro) was explicitly mentioned as an incentive for
return for members of the tribe who arrived from Pakistan later.
Compared to non-beneficiaries, our multi-dimensional poverty index (MPI) presented in Table A.7 shows
that those households which received shelter from UNHCR are less deprived.
Table A.7: Multi-dimensional Poverty Index by Beneficiary Status, in Parwan


% 17.82 82.18 100.00
% 0.00 100.00 100.00
% 7.06 92.94 100.00
Total 24 163 187
% 12.83 87.17 100.00
As mentioned earlier, the fact that in some cases (Khanaqa, Khanjar Khil) one specific area of the village
received assistance whereas the other did not is perceived as a form of discrimination by non-
beneficiaries, a sign of the negative social impact of the programme.
SUSTAINABILITY
o Occupancy
A thorough assessment of the occupancy of all shelters could not be conducted in Khanjar Khil and Qala-
e Nasro. However, all respondents in the province indicated that they plan to stay in the current
community permanently. Of the 101 UNHCR beneficiaries only one indicated that they were not
currently living in a UNHCR shelter. This was confirmed by the community leaders and by field visits,
which showed that a high rate of occupancy, though, as mentioned earlier, some were used for purposes
other than living. In Khanaqa, one shelter was unoccupied as the beneficiary had returned to Iran to
make a living, and three other shelters were used for storage, and two had been dismantled to recycle
the package components in the construction of another house. Most shelters were therefore used, but
not all were occupied for the purpose they had been designed for. This can be interpreted as a sign of
flaws in the selection process and misallocation of assistance, which the IP recognized, though they
mentioned trying to avoid such situations as much as possible. ABR further mentioned a small proportion
of shelters were allocated to the most vulnerable, including those who did not have VRF forms, but that
identifying vulnerable individuals in the community was sometimes problematic.
o Economic Opportunities as the main factor determining the sustainability of the programme
The access to job opportunities and sources of income is one of the main issues for communities:
whether by accessing a plot of land, seeds, material and fertilizer for agriculture, or finding opportunities
as daily labourers in neighbouring towns. This was especially the case in the valley, where agriculture is
214
one of the main income-generating activities for inhabitants. In the Highlands, inhabitants had
subsistence agricultural activities but mainly depended on other sources of income (daily wages in town).
However, the fact that some shelters are used to house family members while the head of household
spends the week in Kabul for work is a sign that shelter assistance contributes to the economic well-
being of the household.
There is a disparity between the needs and situation of communities depending on their location
(valley/ highlands). In the valley, the accessibility of the area explains for the relatively good provision of
services (schools, pharmacies, proximity of a market), but also a better access to job opportunities. In the
Highlands, inhabitants mentioned struggling to make a living and considering going to Kabul for work,
and coming back on weekends, as several members of the community were already doing.
o The proximity of the Bagram Air Case: a risk for sustainability?
In Khanjar Khil, a significant portion of inhabitants were mentioned to work directly (20 households) or
indirectly (80 households) for the Bagram Air Case, raising serious concern about the retreat of the
troops and the horizon of 2014 in terms of settlement of skilled workers, which are expected to move
out of the village. However, recent returnees have fewer opportunities to access these jobs, as they
either do not have the relevant skills, or the necessary network. In Khanjar Khil, villagers had received
direct assistance from the base, including winterization assistance, water pumps, and tractors for
agriculture. Aside from the impact on the population, the important question for UNHCR is the impact
on the sustainability of the programme and on UNHCR assistance in the area, as there is a risk of
confusion between humanitarian and military assistance, which may jeopardize the neutrality of UNHCR.
Parwan is a fragile area, where security balances may change rapidly, and where neutrality will be crucial
for organisations to maintain their presence
o Natural Risks
There was no mention of risk assessment prior to the construction of shelters, despite the fact that in at
least one location (Khanaqa), the inhabitants insisted they were subject to yearly floods.
PARTNERSHIPS
DoRR Parwan mentioned having received capacity building for is staff from UNHCR Central Office, which
is also their main donor for funds and furniture. The proximity of Kabul allows direct communication with
both the line ministry and UNHCR central office. However, the relationship described by the DoRR was
more paternalistic on behalf of UNHCR than truly cooperative. The DoRR deplored the change of focus of
UNHCR to reintegration sites, as well as the site chosen to implement the reintegration strategy (Khanjar
Khil). With the arrival of returnees from abroad in an area where most of the previous housing has been
destroyed, a focus on specific sites for shelter assistance is perceived as not tailored to the effective
needs in the province: “Returnees are dispersed; if you create a heaven for one village and what do you
with other people? And we told them this and we continue saying this to them: the distribution of any
aid should be fair, even if it is limited it should be for everyone”.
All villages visited had benefitted from several types of assistance:

215
- There was notably a high presence of MRRD as part of NSP: wells, improvement/fixing of
irrigation systems.
- UN Habitat provided assistance for rehabilitation of roads, and provided electric generators in
Khanaqa – though those were said not to be useful since the population didn’t have the means
to purchase fuel.
- NPO (Norwegian Project Office) had previously provided shelters in two villages (Qala-e Nasro
and Khanaqa)
- The Dutch Committee constructed a veterinary clinic and conducted a vocational training for a
member of the village in Khanaqa.
- EAU funded installation of water pumps in Qala-e Nasro.
- Several mentions were made of “Americans” providing assistance, without further specification.
Both, the fact that the area suffered from intensive destruction during the war and remains highly
strategic explain the high concentration of humanitarian assistance. However, apart from partnerships
between UN Habitat and MRRD, there seemed to be little coordination between different types of
intervention. More linkages between humanitarian actors would probably have a positive effect on the
sustainability of the programme.
216
BAMYAN – PROVINCIAL OVERVIEW
CONTEXT
Bamyan is part of the Central Highlands region, characterized by its remoteness from Afghan main urban
centres given its altitude and severe weather conditions. The province is the heartland of the Hazara
community and has recently enjoyed a relative peaceful environment. Still Bamyan’s security situation is
deteriorating as the road linking the province to Kabul is getting more dangerous. Overall the province is
under-developed, with residents mainly sustaining themselves with agriculture activities, in particular
the production of potatoes. Some modest efforts have been made to develop tourism in the province.
Bamyan is neither a province of high returns nor an important destination of internal displacement. Most
of the returns occurred between 2002 and 2007, when about 22,000 voluntary returnees were
registered. UNHCR estimates the number of IDPs present in the province to be near zero.
The scope of the UNHCR shelter programme was therefore limited in Bamyan with a total of 285 shelters
built between 2009 and 2011. Our research team surveyed 61 households in Bamyan, of which 52 per
cent were UNHCR beneficiaries. Moreover, 84 per cent of those households surveyed were official
refugee returnees, 7 per cent were non-refugee returnees as well as IDPs and 3 per cent never migrated.
Table A.8: Provincial Sample, Bamyan

UNHCR Non-
Total
Refugee Returnees 32 19 51
% 100.00 65.52 83.61
% 0.00 13.79 6.56
IDPs 0 4 4
% 0.00 13.79 6.56
No Mobility 0 2 2
% 0.00 6.90 3.28
Total 32 29 61
% 100.00 100.00 100.00
DESIGN
While beneficiaries and communities expressed their overall satisfaction with UNHCR shelter assistance,
they pointed out at a few problems with the design of shelters. Beneficiaries complained that the design
of the shelters and material were not adapted to Bamyan’s very harsh winter. Specifically, the number of
beams included in the package was insufficient and caused the roofs to collapse during important
snowfalls as the distance between beams was too important to support important quantity of snow.
217
Many beneficiaries had added some wooden beams to reinforce the structure of their shelters,
increasing the overall costs of the shelters.
Beneficiaries also complained about the low quality of the metal frames of doors and windows provided
by UNHCR. These were considered to be particularly ill adapted to the weather conditions of the
province, as they froze and leaked during the winter. This explains a lower rate of satisfaction of
beneficiaries with the quality of the materials received compared to other provinces, with 81 per cent of
our sample satisfied. Overall, problems beneficiaries had during construction are outlined in Table A.9.
Table A.9: Problems during Construction, in Bamyan

UNHCR Beneficiaries
N=35 %
Materials were of poor quality 5 17.86
Materials were not delivered on time 3 5.71
Most beneficiaries and community leaders noted that the size of the shelters and the numbers of rooms
were insufficient to accommodate beneficiary families. This is linked with the fact that in several villages
visited by the research team, beneficiaries received one-room shelters only (Kata Sang; Shash Pul; Quli
Tubchi). Within our sample of households surveyed, 31 per cent of beneficiaries received one-room
shelters, a proportion that is significantly higher than in other provinces.
Overall beneficiaries and community leaders noted that they were appreciative of the support provided
by UNHCR staff. Monitoring visits took place during the construction phase every week or every two
weeks when the villages were too remote, with 97 per cent of our sample receiving monitoring visits
during construction.
While community leaders did not report particular problems with the selection process, acknowledging
that the poorest had been selected, further discussions with members of communities suggested
otherwise. They reported that only a few beneficiaries were legitimately selected following the criteria,
218
while most were relatives of local CDC members and of UNHCR staff. This would require further
investigation from UNHCR.
As for those households surveyed, 80 per cent of UNHCR beneficiaries were official refugees with a VRF
in hand, and 84 per cent owned the land their shelter was built on before. Moreover some 41 per cent of
UNHCR beneficiaries were extremely vulnerable while none were IDPs.
All communities expressed their satisfaction about the programme and noted that it greatly helped
returnee households by reducing significantly the level of expenses that they would have had to put into
their housing. It was also noted that the programme prevented returnee families to migrate again. Of
those households surveyed, 43 per cent considered their current economic situation as better, while 31
per cent thought they were worse.
Compared to non-beneficiaries, our multi-dimensional poverty index (MPI) presented in Table A.10
shows that those households which received shelter from UNHCR are less deprived.
Table A.10: Multi-dimensional Poverty Index by Beneficiary Status, in Bamyan

% 62.50 37.50 100.00
% 51.72 48.28 100.00
Total 35 26 61
% 57.38 42.62 100.00
Still it appears that the positive impact of the programme remained mostly at the household level.
Community leaders noted that the programme had little impact at the community level, where the
economic fallout was only short-term and limited. Nonetheless, when asked how the programme had
impacted the community as a whole, all households surveyed responded positively or very positively.
SUSTAINABILITY
The occupation rate of UNHCR shelters in Bamyan was very high and all the shelters visited by the
research team were occupied. The sustainability of the programme was supported by interesting
complementary projects of assistance aiming at developing Bamyan’s basic services and infrastructures,
including:
 Micro hydropower stations had been built in several villages (e.g. Meana qad and Shash pul
villages).
 UNHCR distributed solar panels in various villages.
219
 UNHCR helped building a proper canalization and water pipe systems in the communities, which
received shelter assistance.
In fact, of those households surveyed all planned to permanently remain in their current community.
Three main threats for the sustainability of the programme in the province were identified:
a) A high level of unemployment, especially among young people.

b) Lack of drinkable water, which was noted as a major issue by every community leader.
c) A deteriorating security context, reducing the relations between Bamyan and Kabul, hence
rendering trade more difficult for the province.
220
LAGHMAN – PROVINCIAL OVERVIEW
CONTEXT
Laghman remains a rural province, with poor infrastructure and a volatile security context. A large part
of the rural areas are out of bound for the international assistance. Within Laghman, the research team
visited the districts of Mehterlam, and Qarghayi.
The two main regions of return in Afghanistan – Nangarhar and Kabul – border Laghman province. Due
to its remoteness and insecurity, the province saw fewer returnees than neighbouring provinces, with a
total of about 128,000 voluntary returnees registered in the province. Given the proximity of the border
with Pakistan and of two of the main urban centres in the country, the province is still characterized by
the high mobility of its inhabitants who still travel back and forth between Afghanistan and Pakistan. The
number of IDPs present in the province is limited, due to the attraction of nearby Kabul and Jalalabad.
The scope of the UNHCR shelter programme was extensive in Laghman with a total of 2,187 shelter built
between 2009 and 2011. Our research team surveyed 300 households in Laghman, of which 54 per cent
were UNHCR beneficiaries. Moreover, 59 per cent of those households surveyed were official refugee
returnees while 19 per cent were non-refugee returnees, 2 per cent were IDPs and 20 per cent never
migrated.
Table A.11: Provincial Sample, Laghman

UNHCR Non-
Total
Refugee Returnee 128 50 178

% 79.01 36.23 59.33
% 12.35 26.81 19.00
IDPs 1 4 5
% 0.62 2.90 1.67
% 8.02 34.06 20.00
Total 162 138 300
% 100.00 100.00 100.00
DESIGN
Nearly all, 88 per cent, shelters visited were two-room shelters. Beneficiaries complained about the size
of rooms, saying that the rooms were very small, which is problematic in a Pashtun area like Laghman,
where the size of households is relatively higher. Often beneficiaries altered the design of their shelters
221
and removed the corridor to make one bigger room. The research team noted that often the interior of
the shelters was properly painted and decorated.
In most cases, beneficiaries had built some surrounding walls for their shelters. Only in Chardeh (Shahid
Baba), there was no surrounding wall and the design of the shelters had not been changed because the
community could not afford such investments.
The quality of material was praised by beneficiaries, who often mentioned roofing components (iron
beams and roofing tiles) as particularly useful. The beneficiaries did complain about the poor quality of
the metal doors however. In some villages instead of iron beams and roofing tiles, wooden beams and
wooden ceiling boards had been distributed, and beneficiaries were not happy about the quality of these
components, saying that wooden beams and wooden ceiling boards were not good.
During the building process, beneficiaries noted that they did not receive any particular support even
though they encountered a number of challenges including:
 lack of water in dry areas (49% of households surveyed)
 lack of petrol for water pumps,

 lack of skilled labours (22% of households surveyed),
 lack of soil to make mud bricks and
 lack of money (87% households surveyed).
In most cases, the selection of beneficiaries was based on land ownership and the presentation of a VRF,
with 90 per cent of households surveyed owning the land the shelter was built on before assistance, and
74 per cent having a VRF. In some villages the selection of beneficiaries lied in the hands of the Maliks
who were responsible of gathering community members and identifying eligible participants. UNHCR
and IPs employees were not always involved in the process. In some villages though, poverty and
vulnerability entered into play in the selection, with 52 per cent of those households selected being
characterized as extremely vulnerable
There was also a problem in those cases where neighbouring villages had not received any assistance,
despite their needs for shelters (e.g. Qala e Mollayan in chahardeh village) creating a feeling of
favouritism among non-beneficiary communities.
The original beneficiaries occupied all shelters visited because shelter was perceived as an essential need
by beneficiaries who had no alternative. Moreover the majority of beneficiary households surveyed, 67
per cent, considered their current economic situation better of far better compared to before assistance,
while 23 per cent considered it the same. Compared to non-beneficiaries, our multi-dimensional poverty
222
index (MPI) presented in Table A.11 shows that those households which received shelter from UNHCR
are less deprived.
Table A.12: Multi-dimensional Poverty Index by Beneficiary Status, in Laghman


% 24.07 75.93 100.00
Non- Beneficiaries 14 124 138
% 10.14 89.86 100.00
Total 53 247 300
% 17.67 82.33 100.00
All households surveyed believed the programme had been positive for the community as a whole. The
primary impact of the programme was to increase work opportunities. Communities also had the feeling
that the programme increased the level of assistance received by the village as it attracted population,
and therefore the attention of governmental institutions and aid organisations. As the programme
fuelled a growth in the local population, basic services were brought to the area, including schools,
clinics and roads.
In terms of housing, the programme made a huge difference as many beneficiaries were previously living
in tents. In fact 1 month before receiving assistance, 46 per cent of beneficiaries did not own a house.
SUSTAINABILITY
The field team reported a high level of occupancy of the shelters they visited, with almost all shelters
being occupied. Only two shelters had been temporarily left vacant as the owners had gone to Pakistan
and to Jalalabad for work, confirming the important mobility characterizing the province.
Insecurity is one of the main risks threatening the sustainability of the programme in Laghman. Given the
high mobility of Lagman’s population, further deterioration in the security context might lead to
secondary displacement. Still, all beneficiary households surveyed planned to stay in their current
community permanently.
223
NANGARHAR – PROVINCIAL OVERVIEW
CONTEXT
The province of Nangarhar holds a specific position in the shelter assistance programme as it is the
province with the second highest number of returnees after Kabul with 910,076 assisted returnees
screened at Momandara Encashment Center between March 2002 and August 201245. Between 2009
and 2011, a total of 12,433 shelters were completed by UNHCR in the entire province.
Due to the high number of returnees and the increasing number of internally displaced people migrating
to the province from the neighbouring Kunar, Laghman and Nuristan, Nangarhar is facing increasing
demographic pressure and there are subsequent highly political issues at stake when dealing with
returnees, secondary displacement and IDPs. According to a UNHCR Internal Displacement Country
Profile, Nangarhar is the province with the highest number of conflict-induced IDPs with 63,603 people
living in displacement as of May 201246. A substantial number of returnees originally from other
provinces (returnee-IDPs) choose to stay in Nangarhar due both to their incapacity to resettle in their
province of origin because of insecurity, and the attractiveness of Jalalabad agglomeration, which offers
livelihood opportunities and services unavailable in neighbouring provinces, according to governmental
authorities and various stakeholders. The growing number of IDPs, the settlement of returnees from
other provinces, and the increasing demographic pressure are growing concerns for government
authorities and the municipality of Jalalabad who view informal settlements as highly problematic.
A correlative to the high number of returnees/ IDPs in the province is the high concentration of
humanitarian assistance and funding, with an important presence of international organizations
providing both relief and development assistance. Shelter assistance has been implemented by a wide
range of organizations in the past years, with a variety of practices: UN Habitat in urban areas, mainly in
Jalalabad, UNHCR, NRC, IOM until 2008, and IRC until 2011. This was reflected on the field with locations
visited receiving assistance from different types of organizations. As table A.13 shows, our provincial
sample of Nangarhar comprises 2,068 households, of which 38 per cent are UNHCR beneficiaries, 22 per
cent are beneficiaries of other shelter assistance programmes and 40 per cent are non-beneficiaries.
Moreover, 53 per cent of our sample are official refugee returnees, while 33 per cent are non-refugee
returnees, 2 per cent are IDPs and 13 per cent never migrated. Among the UNHCR beneficiaries refugee
returnees are the largest group with 63 per cent. The share of no mobility households is smaller among
UNHCR beneficiaries than those of other programmes.
45
UNHCR “Assisted Voluntary Repatriation to Afghanistan – Return by Province of Destination 02 March 2002 – 31 August
2012”.
46
UNHCR, “Conflict-Induced Internally Displaced Persons in Afghanistan. Interpretation of Data as of 31 May 2012”, July 2012.
224
Table A.13: Provincial Sample, Nangarhar
Total
Refugee returnees 495 241 349 1 085
% 62.66 52.97 42.41 52.47
Non-refugee returnees 223 134 332 689
% 28.23 29.45 40.34 33.32
IDPs 11 6 18 35
% 1.39 1.32 2.19 1.69
% 7.72 16.26 15.07 12.52
Total 790 455 823 2 068
% 100.00 100.00 100.00 100.00
The issue of IDPs was reported as critical by most stakeholders, and strongly linked to the accessibility of
land and the problem of housing. There seems to be a strong belief among governmental authorities, but
also IPs, that development assistance should not be directed at IDPs and returnees from other provinces,
as their migration is considered as temporary. This analysis is challenged by NRC, which emphasizes the
need for long lasting solutions for IDPs incapable to return to their province of origin due to high
concerns about insecurity47. There is however a wide agreement on the fact that the issue of IDPs is
critical and is increasingly acknowledged as one of the main challenges for development organizations.
Both UNHCR and NRC stated having tried to get around the argument of land and informal settlement
brought up by authorities when dealing with IDPs by implementing one-room shelters presented as “an
improvement for a temporary shelter” or transitional temporary shelters. However, a very limited
number of IDPs seem to have effectively been reached by the programme, as showed by Table A.13, as
IPs seemed to consider IDPs as not eligible.
Insecurity is gaining ground in some districts (Surkh Rod and Beh Sud) as Taliban elements from
neighbouring provinces are penetrating rural areas. There is therefore an increasing reliance on remote
management on behalf of UNHCR, and some areas of high shelter implementation falling out of their
reach (Khales Family, the area with the highest concentration of shelters in Jalalabad district is currently
not accessible to UNHCR after the assassination of a member of their staff and increasing influence from
Taliban elements).
Issues related to returnees and IDPs are highly political in the region, with considerable economic
stakes due to funding of assistance directed at these populations. Allegations of corruption of
stakeholders have been reported. These need to be taken very cautiously, but were nonetheless present
throughout fieldwork and KIIs.
47
Cf. “IDP Protection Study”, Samuel Hall, NRC, IDMC, 2012.
225
DESIGN
UNHCR beneficiaries surveyed in Nangarhar mainly built two-room shelters, 74 per cent, while the rest,
26 per cent, built one-room shelters. The majority, 83 per cent, of the beneficiaries did not choose the
size of shelter themselves. More specifically, three different types of shelter were implemented, which
seem to be specific to Nangarhar:
 A “regular” two-room shelter + latrines + $100, for returnees with land and property.
 A one-room shelter + corridor + latrines + $50, originally designed as a solution for
provision of assistance to IDPs. This was however not reflected either in the field or
through conversations with implementing partners, which unanimously insisted that the
UNHCR shelter programme was primarily directed at returnees. The criterion for
allocating one-room shelter in the field was mainly the number of members of the
household (less than 8 members).
 A “repair-kit”: one room with no corridor as an extension to an existing house. Repair
kits are allocated to large households with insufficient room for housing, sometimes in
addition to a shelter.
There were reforms in the design of shelters over the years, based on UNHCR team observations and IP
recommendations: while in 2009 two-room shelters included wooden beams and three small windows,
two larger windows were included in the package in 2010 following field observations and complaints
about lack of light and ventilation. Additionally iron beams replaced wooden poplar beams affected by
termites. IPs and the UNHCR sub-office reported that yearly evaluations and recommendations to the
sub-office based on their field observations and complaints received from beneficiaries were at the
root of these changes in design. The UNHCR sub-office shelter team emphasized they had brought up
issues at the central level, though communication problems with the head office slowed down changes.
Changes introduced in Nangarhar were later reflected in the design at country level, according to them.
Table A.14 illustrates the main problems beneficiaries of both UNHCR and other programmes had during
participation.
226
Table A.14: Dissatisfaction with Elements of the Shelter Programme, in Nangarhar
UNHCR Non-UNHCR
Quality of technical assistance 175 104

% 22.15 22.86
Quality of latrine 132 83
% 16.71 18.24
Thermal isolation 117 53
% 14.81 11.65
Quality of lintels 77 49
% 9.75 10.76
Quality of door 71 36
% 8.99 7.91
Size of the shelter 66 26
% 8.32 5.72
Quality of roof 49 55
% 6.20 9.89
Quality of windows 44 27
% 5.57 5.94
Design of the shelter 31 15
% 3.92 3.40
The main complaint of beneficiaries, relayed by IPs and field engineers, was the small size of the rooms.
In many cases, the wall between the corridor and one of the rooms was removed to enlarge one of the
rooms. Other complaints included the iron window frames and doors, which required additional
treatment for rust. As a result, many UNHCR beneficiaries changed the design of the shelter after
handover by removing the corridor to enlarge the rooms. In such cases, a large one-room shelter (UN
Habitat, CHF) was unanimously considered as more appropriate considering cultural practices of
gathering and high number of family members. Yet, this solution might not be the most appropriate,
especially when considering health and hygiene issues.
Beneficiaries mentioned not having been able to introduce changes in the design prior to completion, as
they were not authorized to do so and would not receive the final cash grant if they did anyway. This lack
of flexibility sometimes proved problematic as some rooms were enlarged by adding wooden beams,
introducing dissymmetry in the design with a potential impact on the solidity of the overall structure. In
one case (Angur Bagh, case 1) the beneficiary waited to be able to purchase adequate material for his
house before starting to build it according to his own will, with IP approval. The fact that the most
vulnerable beneficiaries are unable to undertake changes after completion because of a lack of funds
advocates for more flexibility during the construction process.
Problems during construction: Almost half, 47 per cent, of the beneficiaries indicated that they
encountered some kind of problem during the construction of their shelter. This was the case for people
227
across urban and rural areas alike. Table A.15 shows the different types of problems the beneficiaries
encountered by the type of location they built the shelter in. It shows that the main problems were
having enough financial capital and insufficient access to water. The lack of water or limited access to
water during the construction phase was mentioned as a major burden on the beneficiaries especially in
Sheikh Mesri and Chamtala.
Table A.15: Problems during Construction by Location, in Nangarhar

N=119 % N=2 % N=246 % N=367 %
Ran out of money 100 84.03 1 50.00 215 87.40 316 86.10
Insufficient access to water 38 31.93 0 0.00 120 48.78 158 43.05
Weather problems 42 35.29 2 100.00 92 37.40 136 37.06
Ran out of materials 32 26.89 0 0.00 62 25.20 94 25.61
Lack of skilled labour 27 22.69 1 50.00 62 25.20 90 24.52
Materials of poor quality 27 22.69 0 0.00 14 5.69 41 11.17
Lack of unskilled labour 10 8.40 0 0.00 13 5.28 23 6.27
Materials not delivered on time 10 8.40 0 0.00 13 5.28 23 6.27
Lack of technical knowledge 7 5.88 0 0.00 4 1.63 11 3.00
Another problem noticed was the inadaptability of the design to the size/shape of the land plot, forcing
the beneficiary to arrange the design. In such cases, witnessed in Saracha reintegration site, UNHCR
engineers and IPs allowed the beneficiary to introduce small changes in the design, though no additional
support was provided for additional costs.
Extremely vulnerable individuals who do not have the skills or capacity to build a shelter (elderly,
widows, women heads of households or disabled) are not provided with additional assistance as part of
the shelter programme. In Chamtala for instance, non-beneficiaries stated that they did not apply for
shelter assistance because they knew beforehand that they would not be able bear the additional costs
associated with building the shelter. While knowing that they could receive some assistance through the
shelter package, they still however decided not to apply and rather continued to stay with their relatives
or neighbours. In such cases, returnees are theoretically supposed to receive help from the UNHCR
protection branch under the EVI programme, though this was never witnessed nor mentioned in the
locations visited, and very vulnerable individuals had to rely either on community assistance (scarce), get
increasingly indebted, or let their shelter deteriorate (case 7, Majbur Abad). The UNHCR Protection unit
stated having only a minor involvement in the shelter programme and selection of beneficiaries, and
emphasized the current lack of inter-unit approach. The Protection officer notably underlined that most
EVIs are not reached through the shelter programme as they cannot access land and are therefore not
eligible for shelter.
228
the quality of the material, 93 per cent. Procurement of material for shelter construction was done on a
partly decentralized model until 2012: the central office would procure the high beam, T-beam, door,
windows, and the IP be responsible for local procurement of lintels and ceiling bricks. In 2012, central
procurement was introduced, with all material being centrally procured by Kabul. There are concerns
about the quality of material provided, the additional costs (as material transits through Nangarhar and
is sent back from Kabul with additional transportation costs) and on-time delivery. The model introduced
by NRC in 2011 through its community driven process, relying mainly on direct procurement by
beneficiaries, is in this regard interesting, though it requires a strong monitoring and technical support,
which might not be applicable on a large scale basis – such as the model implemented by UNHCR until
2012.
Earthquake mitigation measures and flood assessments were largely overlooked by most stakeholders
(whether UNHCR, NGOs involved in shelter, government authorities or IPs). UNHCR staff and IPs solely
rely on a sight-check by engineers (who do not seem to have any qualification in risk assessment),
pointing to the fact that previously inhabited areas are not prone to floods – which is highly
questionable, as large scale floods in 2009 and 2010 in Jalalabad had a considerable impact on housing,
as proved by the visit in Nahiya 1. There is no form of coordination or consultancy with municipality or
Directorate of Urban Development and Housing in urban areas, or ANDMA for risk assessment.
There is no regular assessment of occupancy of shelters, allegedly because of the one-year budget cycle
of UNHCR.
 Profile of communities targeted by the shelter programme:
Areas of implementation are decided by UNHCR according to the screening at the border and the area
mentioned by returnees as their place of return. The first criterion for selection is therefore “high return
areas”, both in rural areas and inside the boundaries of the rapidly expanding Jalalabad agglomeration.
Profiles of beneficiary communities varied according to the location (urban/ rural) and the profile of the
community (homogeneous/ heterogeneous/ rate and stages of return). Some of the consequences of
high return rates are an increasing pressure on housing, but also on job opportunities, access to water,
and whenever services and health facilities were available, on the capacity of these facilities to support
the increasing demographic pressure.
UNHCR said the selection process for beneficiaries systematically went through a Beneficiary Selection
Committee (BSC), composed of a UNHCR member, a DoRR representative, the IP, and local authorities in
accessible areas. In non- accessible areas, the BSC is composed of the same members, without UNHCR
staff. There is no female participation in the selection process. However, the DoRR insisted that they
were often side-lined for beneficiary selection because they were not informed nor included in some
cases. UNHCR and IPs argued that there was a lack of good will on behalf of DoRR representatives, either
229
not available for selection or asking for additional support (cars), which neither UNHCR nor the IP were
willing or prepared to provide.
Reliance on the Malik was mentioned to introduce a major bias in selection with strong allegations of
nepotism on behalf of non-beneficiaries in various locations. In homogenous communities with a large
number of shelters allocated, the reliance on the Malik seemed to work and there were no complaints.
Selection was done primarily through VRF and most VRF holders were selected (Sediq Abad).
Interestingly, in those cases, selection was considered as fair even by non-beneficiaries, since the VRF
criterion was considered as transparent. In more heterogeneous contexts, there were widespread
allegations of favouritism. IPs also acknowledged the issue, while they mentioned not having relied on
the Malik for selection. In many cases however, the Malik is the one who gathers VRF holders for
selection and appears to have influence in the selection process.
The short time-span for selection of beneficiaries raises problems in targeting the most vulnerable.
Cases were mentioned where potentially eligible people were out of the village for work during selection
and were therefore left aside. In areas where other types of selection were carried out (UN Habitat in
Bez Akmalati) there was an emphasis on the relative effectiveness of the UN Habitat programme, which
takes several months but is more community-driven. In areas where different programmes were
implemented beneficiaries compared the selection processes of both organizations and deemed the UN
Habitat process, for instance, “fairer” than that of UNHCR, which had been quick but discriminating.
There appears to be issues in preliminary assessment of needs in targeted areas, with areas of high
return targeted with no consideration of the effective needs of the people. In such cases, the quota of
shelters allocated having to be constructed whatsoever; less vulnerable VRF holders were included,
whereas vulnerable individuals deprived of VRF were left out. Accordingly, 57 per cent of the surveyed
households cannot be classified as extremely vulnerable, while 43 per cent can. Additionally, the
quantitative data shows 139 EVI households applied for shelter assistance, but were not chosen. In
Sheikh Mesri, both AGHCO and WSTA mentioned having had issues for finding eligible individuals and
implementing the number of shelters allocated. WSTA also reported many incidents of forged VRF cards
and doubts about the legality of their land ownership.
 Criteria of Selection
Moreover, there was a discrepancy between UNHCR discourse and the reality on the field, where there
seemed to be a widespread misunderstanding of the criteria by communities. Despite the assurance by
UNHCR that all criteria for selection were taken into account, fieldwork showed shelters were primarily
given to VRF holders, as a way to distinguish eligible beneficiaries from others. The majority, 83 per cent,
of the beneficiaries did report that they did have a VRF, even though only 63 per cent are recognized
refugee returnees). There is a deep-rooted belief both among communities and IPs that UNHCR is to
bring assistance to VRF holders, and that other NGOs and organizations will provide assistance to others
(UN Habitat, CHF, NRC), though this is not the case. As a result, IDPs were not reflected at all in the
selection. Only 1 per cent of the UNHCR beneficiaries in our survey sample were in fact IDPs. Though
UNHCR mentioned having developed the one-room shelter model specifically for targeting IDPs, this was
not reflected in the field.
230
Consequentially, one of the side effects of the high focus on VRF mentioned by community members
were fraud and trafficking of forms, with undocumented returnees buying VRF from others. In some
cases, community members mentioned having been opposed the “expiration” of their VRF for not
receiving assistance. However, according to UNHCR there are no guidelines about duration of return for
assistance. According to them, UNHCR’s assistance may have been limited only because of UNHCR
budget limitations. They further mentioned that, due to those budget restrictions, recent returnees were
primarily targeted as they are considered to be more in need of shelter assistance than those who have
settled many years ago and had been able to establish coping mechanisms. Accordingly the quantitative
data shows that 71 per cent of the UNHCR beneficiaries had returned within the past five years.
However, field observations showed that in some cases, ancient returnees were as much, and
sometimes more in need than some recent returnees, and had been living with relatives for a long
period of time, which is considered as a great burden by the host community. Though UNHCR insisted it
made its selection criteria clear through posts and distribution of leaflets, this was not reflected on the
field, with no one mentioning having received notifications other than those of the Malik and the visits of
“engineers”.
The issue of land seems to be discriminating in the selection process as well as highly political. Land
ownership is a requirement for shelter in the UNHCR programme, though in most cases people do not
own proper documentation and the BSC has to rely on testimony of local authorities. In the survey, 82
per cent of the beneficiaries indicated that they did own the land the shelter was built on already before.
Among those that did not almost have received it through the LAS (48.3%), while others built on their
relatives’ land (15.4%) or on community/public land (9.8%).
Inside the boundaries of Jalalabad agglomeration, many shelter beneficiaries had benefitted from a land
allocation, when land was made available by municipality at lower prices (20 000 AFS). A high number of
returnees also benefited from the LAS in Sheikh Mesri, a formal LAS township, and Chamtala, which used
to be an informal settlement. In other cases (Bez Akmalati), access to land for the most vulnerable was
secured by UN Habitat through negotiation with the municipality. Conversely, beneficiaries from the
UNHCR programme, which does not include these sorts of mechanisms, were all landowners, raising
complaints about the selection in reaching the most vulnerable. In Bez Akmalati, which used to be an
informal settlement, a fix number of shelters was allocated by UNHCR, but many were not eligible due to
lack of access to land, and shelters were therefore not necessarily allocated to those who were most in
need. Securing land tenure is highly problematic, especially for returnees outside their province of origin:
those who can purchase land are therefore not the most vulnerable. In such cases, other programs, such
as those of UN Habitat, which rely on collaboration with relevant government authorities (municipality,
MUDH for urban areas) to provide eligible people with land, was perceived by the communities as more
effective in reaching the most vulnerable.
However, in LAS sites (Chamtala) Malik and beneficiaries were very critical about the involvement of the
government, especially the MUDH that they accused of bribery and nepotism. According to the Malik
and the community members, the DoRR and the MUDH asked for bribes for land allocation, holding
returnee’s VRFs and asking for 10.000 AFS upon request for a plot. The municipality and MUDH were
also accused of misallocation and mismanagement of plots, which were originally destined to be public
spaces -plots for schools, parks or clinics. This led to many tensions within the community as in certain
231
areas, clinics or schools could not be built or used because people had established their own houses on
the plots of land they were meant to be constructed on.
 Complaint Mechanism
There are no direct complaint mechanisms for UNHCR beneficiaries and targeted communities. Issues
are reported through field engineers and IPs. Conversely, NRC complaint mechanisms were the main
driver of an evolution to a community driven process and introducing more flexibility in the construction
process.
 At the Household Level
Shelter is unanimously seen by all stakeholders as a major need for returnees and a way to allow them
to focus on other important issues. Most beneficiaries underlined the fact that they planned to stay, as
they had “nowhere else to go” and insisted on ownership as a strong component of a durable
settlement. In one case (Saracha focus group), a beneficiary mentioned planning to leave in order to find
better job opportunities, though he emphasized the shelter would remain within his family as a base for
return.
The quantitative data shows that 49 per cent of the beneficiaries considered the impact of the shelter
assistance on their economic situation as positive, while 29 per cent did not see any change and 22 per
cent considered it worse than before. Beneficiaries interviewed in focus groups unanimously underlined
that shelter had had a very positive impact on their households by providing them with a roof (for those
previously living in tents) or allowing them to have a space of their own instead of having to live with
relatives, especially since they wouldn’t have been able to find the funds for purchasing the material
for building a house themselves, and needed technical assistance. The fact that the programme
alleviated pressure on relatives and other community members who previously had to host incoming
migrants was stated as one of the major positive effects of the programme.
Table A.16: Multi-dimensional Poverty Index by Beneficiary Status, in Nangarhar


% 32.15 67.85 100.00
% 31.87 68.13 100.00
% 14.34 85.66 100.00
Total 517 1 551 2 068
% 25.00 75.00 100.00
232
As in other provinces, indebtedness was repeatedly mentioned as one of the major issues faced by
beneficiaries who had to pay for labour work (as most do not have construction skills), bricks and stones
for foundations. Of the surveyed beneficiaries more than half indicated that their household debt
increased due to the participation in the shelter programme. At the same time 32 per cent indicated that
their debt level decreased. There were discrepancies in the amount of debt according to the material
used: 50.000 AFS-70.000 AFS for sun-dried bricks, over 100.000 AFS for cooked bricks. In this case,
technical support could have proved useful, as cooked bricks, though they are considered as a sign of
modernity, are less isolating and shock-resistant than traditional sun-dried bricks. Despite debts
contracted, there was a general agreement on behalf of beneficiaries and non-beneficiaries that shelter
assistance allowed beneficiaries to construct a house with less expenditure. It was mentioned that once
provided with housing, beneficiaries could work to reimburse their debt, and focus on other important
issues in family life, which they would not have been able to take into consideration without shelter
assistance.
 At the community level
However, shelter assistance is generally perceived as positive by the community, non-beneficiaries

insisting that “help calls help”. Returnees are often seen as a good thing in a village, as they draw the
help to them, which in return benefits the entire community (alleviation of pressure on host families,
roads, WASH programmes). In particular, one point that was underlined several times was the
introduction of latrines and hygiene classes, which changed the practices of the entire community, with
non-beneficiaries reproducing the same kind of facilities and adapting the same practices as beneficiaries
(Nahiya 1, Bez Akmalati). This emphasizes the importance of multiple assistance programmes and
greater coordination between agencies/NGOs involved in assistance. The programme was all the more
perceived as positive when digging of wells, draining of canals, construction of schools and clinics had
been implemented alongside shelter in the neighbourhood.
Side effects of the high concentration of assistance on areas of high return include the increase of the
price of land with a consequence on job opportunities (as shop keepers) and accessibility of housing for
other community members.
SUSTAINABILITY
 Occupancy
A minor group of those beneficiaries interviewed, 1 per cent, indicated that they are currently not
living in the UNHCR shelter. Also, there was a high rate of occupancy in most locations visited. In some
cases where the beneficiary himself had returned to Pakistan or gone to Kabul for work, his family
(mother, wife, children) were living in the shelter (cf. case 2 Angur Bagh): shelters are therefore
considered as a strong component of family life, even in cases when the lack of job opportunities to not
allow the member of the household responsible for bringing income to stay. This is underlined by the
fact, that while 11 non-beneficiary households currently plan to move from their current community, no
UNHCR beneficiary households have such plans.
233
However, in the LAS township of Sheikh Mesri, many unoccupied shelters were spotted. Though a
thorough assessment of occupancy rates could not be conducted due to the large size of the township,
in the area visited, 10 completely abandoned UNHCR shelters were spotted in a small perimeter were
16 shelters in total were visited. This leads to the assumption the rates of occupancy are low in the
township. This is corroborated by the fact that many allocate plots of land seemed abandoned.
Beneficiaries and non-beneficiaries mentioned people had left to live somewhere else because of lack of
employment opportunities and limited access to services. Moreover, misallocation of land plots led to
disputes over land, which discouraged people to settle on these plots. Access to safe water was
reportedly another important issue.
 Technical support
Technical support appears as a strong component in the sustainability of the shelter model, as it provides
beneficiaries but also their relatives and other members of the community, with lacking basic
construction skills. However, technical assistance mainly relied on the IP and its willingness to provide
additional help. A regular model for technical support as part of the UNHCR programme in Nangarhar
was the provision of plans, visiting several times a week during construction to check the progression of
the work and the quality of material used, and conducting a final check for handover. One IP (APWO)
mentioned they did provide additional help to unskilled beneficiaries, but did so on a voluntary basis, as
they had received no specific instructions about these cases. The same IP emphasized that deadlines for
completion of shelters (2 months), were particularly difficult to meet in cases where beneficiaries had no
skills at all and had to be explained the basics of construction. Ashar, or traditional community
solidarity, which UNHCR relies upon for construction, seems in most cases not to be a reality: other
members of the community being on daily wages, could not afford to take days off to assist others while
they have difficulties making ends meet for their own households. The issue of technical support as a
strong component of the programme deserves being looked into, as proper capacity building during
construction not only ensures the good quality of the shelter, but provides beneficiaries with skills that
they will later be able to use and share with other community members (NRC identified this as a major
aspect of its “community driven programme” as a mean to empower communities through
beneficiaries).
The case of latrine construction in the UNHCR shelter programme provides a good example of
dissemination of learning, as in several locations the introduction of latrines as part of the shelter
programme triggered the construction of latrines in the community through sharing of the construction
plan. Additional technical support and flexibility could also prevent post-construction changes
weakening the structure of the shelter (cf. above).
 Major threats to sustainability
Major concerns for sustainability were the lack of accessibility of jobs due to increasing demographic
pressure, access to safe water, roads, and availability of clinics (due to widespread diseases, specifically
in urban areas because of poor living conditions and pollution). UNHCR insisted a WASH component
(water and hygiene) was systematically included in the programme. Though in several areas,
beneficiaries denied having received hygiene trainings, in communities where either UNHCR or another
234
organization had provided WASH assistance, it benefitted the community in addition to beneficiaries of
the shelter programme.
 Access (roads) is a primary concern as it allowed inhabitants to work in Jalalabad as daily

workers. The case of Jalalabad is specific as shelters were implemented by UNHCR inside the
agglomeration, with the opportunity to go in town for daily work (construction, rickshaw
drivers), though there was considerable insistence on the difficulty to secure a livelihood due
to high demographic pressure.
 There are potential tensions in the relationship between beneficiaries and non-
beneficiaries due to the lack of transparency of beneficiary selection – especially in
communities where some areas were targeted and not others – cf. Majbur Abad. This is
illustrated by the fact that some non-beneficiaries first refused to talk to us until they
understood we were not coming to provide assistance but to get their opinion on the
programme.
REINTEGRATION SITES
The reintegration sites and concentration on some specific areas of high return is unanimously
considered (aside from APA) as a negative trend and discriminating development in the selection of
beneficiaries, leaving out the greatest bulk of vulnerable people which are settled outside targeted
areas (cf. UNHCR, AGHCO, APWO, WSTA, NRC, IRC, governmental authorities KIIs).
 Moreover, the UNHCR sub-office underlined intra-agency coordination remained strictly
theoretical and was not a reality in the field. UNHCR’s one-year budget cycle being framed for
assistance, the danger is therefore to side-line any comprehensive development strategy. In
the absence of any long-term development projects, reintegration remains a lure (SHEIKH
MESRI and FATIH ABAD).
 The Protection Unit underlined high risks of exclusion, as reintegration sites are selected
according to their accessibility, and not to the vulnerability criterion. The UNHCR Field unit
and maliks further mentioned proximity to the district centre and prior presence of
development organizations working in that area as criteria for selection. AGHCO and WSTA
explained not having been able to identify the right number of eligible people in Sheikh
Mesri, having referred it to UNHCR and having been instructed to target neighbouring
villages, which highlights problems in preliminary assessments.
 There is no formal coordination mechanism for reintegration sites.
PARTNERSHIPS
 Remote Management and Overreliance on IPs
UNHCR solely relies on one NGO, APA, to monitor non-accessible areas. Given the widespread
allegations of corruption in the province, this is a serious area of concern. All the more as APA is the NGO
involved in well-funded project in Kuchi Abad (Kabul), one of the most mediatised and a highly funded
235
reintegration site, which is a very clear conflict of interest and raises serious questions about the actual
functioning of the monitoring mechanism. Another major issue is that APA was also asked to monitor
accessible areas, where no need for remote management is theoretically needed, increasing the scope of
inaccessible areas. There is a serious concern with the fact that the area with most implemented shelters
is Khales Family, which is considered as non-accessible.
Due to high return and migration rates and increasing pressure on land and housing, Nangarhar is the
province, which received the highest concentration of shelter assistance in the country in the past years,
with consecutive high political and financial stakes. There are currently three on-going shelter
programmes: UNHCR, UN Habitat (as part of their urban upgrading programme) in urban areas, and NRC
in rural areas. IRC interrupted its shelter programme for returnees in 2011, after their main donor
rejected their proposal ECHO (no further explanation was given on this point: “political” and
“economical” problems were mentioned). IOM had a programme on going until 2009, interrupted due to
lack of funding, but is expected to restart a similar programme in the perspective of evictions from
Pakistan. CHF, a USAID funded organisation, had a programme in 2009-2011.
UNHCR heads a monthly shelter meeting held in its office, which is a specificity of Nangarhar as in other
provinces, the only information-sharing platform is the ES/NFI cluster. These are the main information-
sharing platform about shelter, the main objective reportedly being to avoid duplication. However UN
Habitat reported there was no actual information sharing about practices, and that the meeting
consisted of a “routine exchange of figures”. All three on-going programs are implemented in a distinct
way, and there seem to be disagreements on how shelters should be implemented and to whom they
should be directed.
Interestingly, though UNHCR and UN Habitat denied any overlap between their programs, their shelters
were often found in the same areas, and though the selection processes are sensibly different, criteria
for selection are theoretically similar (IDPs, returnees, vulnerable), with the notable difference that the
UN Habitat urban upgrading programme includes a process for solving land ownership issues in recent
settlements, a good practice of which UNHCR could get inspiration. In the field, beneficiaries from these
two organisations were effectively not the same, as UN Habitat put less focus on VRF, and land
ownership was not a prior requirement.
There was an insistence on behalf of the Protection unit for more intra –sectorial coordination,
coordination between UN agencies, and various stakeholders. KIIs with different UNHCR, UN Habitat
and UNOPS stakeholders showed there was lack coordination. The main focus for coordination is
emergency. Shelter coordination is done through a monthly meeting between various stakeholders,
though there seems to be longstanding disagreements on the way to implement shelter between NRC,
UN Habitat and UNHCR.
Governmental participation: One of the main complaints of all governmental authorities was the lack of
communication on behalf of UNHCR and its reluctance to include governmental partners in its
programmes. The main UNHCR partner in the shelter programme is the DoRR of Nangarhar, which is the
most staffed in Afghanistan and receives most of its funds from UNHCR. The current director was
mentioned to have considerable political influence as the brother in law of the speaker in Parliament.
236
There are strong tensions between UNHCR and DoRR however, with DoRR accusing UNHCR of pushing
them aside, and UNHCR recognizing being highly suspicious of the good will of the DoRR to implement
the programme according to UNHCR regulations. High levels of tension are reflected in allegations of
corruption on both sides. Miscommunication and suspicion are therefore considerable obstacles to
cooperation. Moreover there is reluctance on behalf of UNHCR to include governmental authorities in
the selection process. Despite a wide network of staff, the problem of incompetence of DoRR staff was
raised by several interviewees (i.e. representatives of IPs, IOM, UNHCR), due to widespread nepotism.
DRRD is a partner in development programs in rural areas and seen as more independent as it receives
funds from various donors.
The Directorate of Economy deplored being left out of UNHCR selection of IPs. With the DoE being
responsible for assessment and evaluation of NGOs, this is perceived as a major issue and a bypassing of
rules and regulations, especially in avoiding selection of “black-listed” NGOs.
The ANDMA is not included in the UNHCR shelter programme, and is not considered by any organization
involved in shelter as a potential partner for shelter. As underlined above, there is little to no risk
assessment of construction land when building shelters other than an in-sight check by engineers, and
reliance on popular experience. ANDMA is contacted in case of wide scale disasters in emergency
programs. AGHCO insisted it was the most under-funded and under-staffed government agency and had
little influence, let alone capacity to conduct prevention activities. The ANDMA itself, however,
considered they were very capable and had identified areas of high risks. They further stated all other
UN-agencies excluding UNHCR coordinate with them. They mentioned being a part of the emergency
coordination committee with other UN agencies and having coordinated emergency responses in the
past.
All governmental authorities insisted on the need for capacity building.
One of the main issues in Nangarhar was the apparent poor relationship between UNHCR, IPs and
governmental authorities and subsequent mutual allegations of corruption. The DoRR explicitly accused
UNHCR of being corrupted and of misallocating shelters, whereas UNHCR and IPs conjointly reported not
trusting governmental authorities and intentionally side-lining them because of their repeated attempts
to allocate assistance to their relatives and draw extra funds. On his side, the Directorate of Economy
deplored the fact UNHCR did not follow the “rules and regulations” by not asking for his approval for the
selection of IPs, which had previously been blacklisted by the Ministry of Economy due to identification
of corruption practices. There is therefore an apparent lack of trust between the various stakeholders
involved in shelter in the province.
237
BALKH – PROVINCIAL OVERVIEW
CONTEXT
Balkh province has seen the return of about 120,000 returnees over the past ten years. The number of
conflict-induced IDPs in Balkh is limited (below 5,000) but, as other provinces in the North, the province
is prone to various natural disasters, including drought, flooding and bank erosion alongside the Amu
River.
The scope of the UNHCR shelter programme was limited in Balkh with a total of 470 shelters built
between 2009 and 2011 Shelter assistance was concentrated in the district of Sholgara with 331 shelters
distributed there alone. Our research team surveyed 101 households in Balkh, of which nearly 50 per
cent were UNHCR beneficiaries. Moreover, 56 per cent of those households surveyed were official
refugee returnees while 9 per cent were non-refugee returnees, 13 per cent were IDPs and 22 per cent
never migrated.
Table A.17: Provincial Sample, Balkh

UNHCR Non-
Total
% 82 31.37 56.44
% 4 13.73 8.91
IDPs 5 8 13
% 10 15.69 12.87
No Mobility 2 20 22
% 4 39.22 21.78
Total 50 51 101
% 100.00 100.00 100.00
DESIGN
Nearly all shelters visited in Balkh, 90 per cent, were two-room shelters. Most beneficiaries, 96 per cent,
did not personally choose the size of the shelter and did not make particular changes to the design
during construction. The field team noticed that shelters in Balkh were in a better state than those
visited in other provinces. All shelters had proper glass windows. Beneficiaries expressed overall
satisfaction with the shelters received, and no particular complaints were raised. The size of shelters was
mentioned, with some 14 per cent of those households surveyed expressing dissatisfaction about size,
but it did not appear to be as much of an issue as in other provinces.
238
The main challenge for the construction of shelters was the lack of water as some of the areas of
implementation are very dry (cf. Sholgara for instance). In fact, 75 per cent of households surveyed
noted the lack of water as a problem during construction, a proportion that is significantly higher than in
other provinces. In this regard, UNHCR intervened and paid for the supply of water during construction
of shelters. Also, beneficiaries mentioned the difficulties they had to hire extra workers because of their
poor financial capacities, with some 90 per cent of households surveyed noting that they ran out of
money during construction. Overall, it appears beneficiaries relied on each other’s help to finish the
construction of shelters.
Contrary to some other provinces, the selection process in place in Balkh seems to have complied with
the programme’s guidelines, with a BSC composed of employees from the DoRR, as well as heads of
shuras, and representatives from UNHCR and the IP, all present for the selection of beneficiaries.
 Criteria of selection
The VRF and land ownership were the main criteria of selection used in the field, with 98 per cent of
beneficiaries having a VRF and 94 per cent owning the land their shelter was built on before assistance.
Discussions with the sub-office also showed that vulnerability had not been mainstreamed as one of the
priority criterion for selection of beneficiaries, yet still some 44 per cent of beneficiary households
surveyed are characterized as extremely vulnerable. Moreover, IDP and landless families were not
particularly included in the programme with only 10 per cent and 6 per cent of each, respectively, in the
sample.
The team observed that in some areas the selection had been made following ethnic lines, with
beneficiaries being selected exclusively from one ethnic group. A more balanced selection approach
might be necessary to avoid fuelling ethnic tensions within villages. In some areas, a few large families
shared all the shelters that had been distributed. Moreover, the selection process was not immune to
pressure from local powers. Some local commanders for example had received shelters, which they had
turned into big houses.
As in other contexts shelter assistance was perceived as responding to a major need in helping
households access proper forms of housing for which they would not have been able to pay for on their
own. Numerous households, 90 per cent of those surveyed, had been living in tents upon return. The
programme therefore greatly improved the living conditions of those beneficiaries. Moreover some 76
per cent of households surveyed noted their economic situation had been better or far better following
receiving shelter assistance. Compared to non-beneficiaries, our multi-dimensional poverty index (MPI)
presented in Table A.18 shows that those households which received shelter from UNHCR are less
deprived.
239
Table A.18: Multi-dimensional Poverty Index by Beneficiary Status, in Balkh
% 30.00 70.00 100.00
% 7.84 92.16 100.00
Total 19 82 101
% 18.81 81.19 100.00
Overall, the shelter programme was perceived as a good first basis on which returnees could build their
life and start focusing on other problems of their daily life. In Sholgara, the programme helped building
an important settlement on governmental land, which had been distributed to members of one
community. It is difficult to measure the impact of the shelter programme in itself as the site was
selected as a ‘reintegration site’ in 2012 and benefitted from a number of initiatives supported by
UNHCR.
SUSTAINABILITY
In those locations visited by the survey team in Balkh, the sustainability of the shelters built with the
support of UNHCR did not seem under particular threat. In fact, of those households surveyed all plan to
permanently remain in their current community. Moreover occupancy rates were high in all locations
and the areas chosen for implementation of the programme benefitted from a relatively good access to
services and often from a relatively satisfying access to employment.
PARTNERSHIPS
As far as IPs are concerned, UNHCR relies on its long-term partnership with DHSA for the whole Northern
and North western regions. DHSA appeared to be a reliable partner for UNHCR. The organization
includes women staff involved in the management of the programme in the field.
DHSA is monitored through UNHCR and through another IP called NPO (Norwegian Project Office). There
are some issues with this type of monitoring as: a) NPO monitored the implementation of the shelter
programme in all the North, including areas where UNHCR staff should have access; b) NPO was under-
staffed and could only have one staff member covering the monitoring for several UNHCR programmes
at a time in 3 provinces; c) there is a conflict of interest as NPO is also an implementing partner for
UNHCR in the 2012 reintegration site (Sholgara).
The DoRR noted that the relationship with UNHCR sub-office had significantly improved recently. The
DoRR reported being consulted at different stages of the implementation of the programme. The only
criticisms related to the absence of coordination with UNHCR regarding the selection of their IPs, upon
which they feel they have no say.
240
Regarding UNHCR’s relations with other partners, Mazar is the regional centre for most organisations
and UN-agencies working in the Northern region. There is a certain complementarity between
organisations providing shelter assistance in the North, as UNHCR got the lead on returnee and conflict-
induced IDPs, while OCHA coordinated the response for the shelter assistance of flood-affected
populations. Most of the organisations present in the region (ZOA, Islamic Relief, NRC) focus on flood-
affected population, leaving to UNHCR all other cases, except for NRC’s presence in Sari Pul. Given the
high presence of UN-agencies, INGOs and NNGOs in Mazar-e-Sharif, the province would be a good place
to pilot some partnerships with organisations in order to link shelter assistance with early recovery and
development activities.
The Emergency and NFI cluster served as a venue to discuss non-emergency sheltering among
stakeholders. In particular, discussions took place regarding the design of UNHCR shelters and its
adequacy with the regional context, taking into account the specific natural risks characterizing the
North. The cluster, as well as good bilateral relations between organisations, allowed UNHCR and others
to avoid duplication. While the cluster system has not been designed to address non-emergency issues,
having this sort of ad-hoc coordinating meetings – especially in areas prone to natural risks – is a good
practice that may be replicated in other regions.
241
FARYAB – PROVINCIAL OVERVIEW
CONTEXT
Faryab is not a province of very high return. UNHCR registered 73,458 individual assisted returns to
Faryab between March 2002 and August 201248. On the other hand, Faryab belongs to the top 10
provinces of internal displacement, with 13,819 conflict-induced IDPs reportedly living in displacement in
Faryab as of end of June 201249. Between 2009 and 2011, 1,568 shelters were built by UNHCR in the
province.
The high percentage of IDPs is reflected in the sample drawn for this study as there are relatively more
IDPs included than in most other provinces. As table A.19 shows, our provincial sample of Faryab
comprises 174 households, of which 41 per cent are official refugee returnees and 35 per cent are IDPs,
while 13 per cent are non-refugee returnees and 12 per cent never migrated. 43 per cent (75
households) of the surveyed households are UNHCR beneficiaries. Two surveyed households received
shelter assistance from IOM. The higher degree of inclusion of IDPs in Faryab province shows a welcome
adaptation of the programme to the migratory profile of the province. Yet, discussions with UNHCR staff
in Faryab showed that the IDPs included were mostly returnee-IDPs, as IDPs themselves were still rarely
included in the programme.
Table A.19: Provincial Sample, Faryab

Total
Refugee Returnees 48 0 24 72
% 64.00 0.00 24.74 41.38
Non-refugee Returnees 1 0 21 22
% 1.33 0.00 21.65 12.64
IDPs 26 2 32 60
% 34.67 100.00 32.99 34.48
% 0.00 0.00 20.62 11.49
Total 75 2 97 174
% 100.00 100.00 100.00 100.00
The security context in Faryab is rapidly deteriorating. The Norwegian PRT was shut down last spring
and Afghan national security forces struggle to take over and guarantee security in the whole province.
Most stakeholders agreed that they had lost direct access to large parts of the province. Insurgent
activities, IEDs, assassinations and kidnapping are now very common in the province.
48
UNHCR Afghanistan.
49
UNHCR Afghanistan, 06.2012, Statistical Summary of Conflict-Induced Internal Displacement. As of 30 June 2012.
242
Insecurity and loss of access have various consequences on the shelter programme:
 Due to insecurity, the province has an important number of intra-provincial IDPs.

 In terms of monitoring, numerous locations where the shelter programme was
implemented in the past are now out of reach.
 In terms of sustainability, further displacements – especially intra-provincial
displacements - are likely to take place in the coming years and some of the
beneficiaries of the shelter programme might be forced to leave their shelters due to
insecurity. This has already occurred in some districts.
 The presence of UN agencies in Maimana might be called into question if the context
further deteriorates.
Despite the insecure context, UNHCR maintains its presence in the province through an experienced
national staff. This is not the case for all UN-agencies, as the WFP has already closed down its own office
in Maimana.
For 2012, UNHCR has selected one reintegration site in the province in the village of Baimoghly located
in Khoje Sabz Posh district. The site is directly accessible to UNHCR staff. There, UNHCR concentrated its
activities and developed several different components including WASH, school, road construction, public
latrines, shelter and ‘protection training’. Protection training included: hygiene promotion, environment
protection, and peaceful coexistence.
DESIGN
The beneficiaries and the various stakeholders interviewed in Maimana were satisfied with the design
of UNHCR shelter. Most beneficiaries surveyed, 91 per cent, did not personally choose the size of their
shelter. The main design implemented was the standard two-room shelter, 87 per cent, while a minority
built one-room shelters, 13 per cent.
the quality of the material, 97 per cent.
Some points are worth noticing:
 ACTED made the remark that beneficiaries were asking for a kitchen to be included in
the design of the shelters.
 In several cases, people had used the cash allocated for glass windows to buy water
and food items. A lot of shelters had only plastic sheets put directly on the window
frames provided by UNHCR.
 In the North (Andkhoy district) of the province beneficiaries complained about the size
of the shelters. There several communities are specialized in carpet weaving, usually
within Uzbek or Turkmen communities. Beneficiaries claimed that it was impossible for
them to put the frame necessary to weave carpets inside their shelter. In most cases,
243
they were forced to use smaller weaving frames to adapt to the size of the room. This
specificity of carpet weaving communities is also important from a gender perspective,
as women are in charge of weaving carpets and might have difficulties weaving outside
their homes. (This issue was also mentioned in Jawzjan province).
Only 28 per cent of the surveyed beneficiaries indicated that they encountered some kind of problem
during the construction of their shelter, a figure that is remarkably low compared to other provinces.
The main problem UNHCR beneficiaries in Faryab mentioned was running out of money during
construction (26.7%). Additional encountered problems were a lack of skilled labour (12.0%) and
unskilled labour (9.3%) and insufficient access to water (12.0%). The monitoring of the construction
process seemed to have been less tight than in other provinces, especially in remote areas. In Aaq Dar
for example, beneficiaries were only rarely visited during the implementation, and the community leader
noted that he was in charge of monitoring the process at the end.
Strangely earthquake mitigation measures, as well as flood prevention measures, did not seem to
mean anything to most stakeholders in Faryab province, including UNHCR staff. Faryab province has a
low risk of earthquake according to OCHA’s map, but an earthquake took place in 2009 and destroyed
numerous houses, requiring the assistance of ACTED at the time. UNHCR staff also acknowledged that its
shelters would not resist flooding as traditional shelters built in the province would. Yet they did not
seem to consider taking preventing measures.
Stakeholders considered that implementing a cash-based shelter programme would be very tricky in
Faryab province. Because of the level of dire poverty of most households, the cash would necessarily be
used for other purposes as it was for example observed in Aq Mazar village (Khoje Sabz Posh district)
where most beneficiaries had used the cash provided by UNHCR to buy food and water instead of
purchasing glass windows. Cash-based programmes require very close monitoring mechanisms, which
would be very difficult to implement in a context like Faryab, where UNHCR has no access to numerous
implementation sites.
UNHCR speaks of ‘remote selection’ when it comes to selecting beneficiaries in Faryab province. From
what we have seen, the selection process lies mainly in the hand of the IPs, sometimes without prior
discussions with the community leaders (as it has been reported in Nisai Khona, Andkhoy district). The
criteria of vulnerability were not really mainstreamed in the process. Discussions with the IPs showed
that the IP staff was not particularly sensitized to EVI guidelines or other vulnerability criteria. UNHCR
mentioned that 3 to 4 per cent of beneficiaries are widows or female-headed households.
This also shows in the quantitative data collected for this study. While 40 per cent of UNHCR
beneficiaries did not have a VRF form, the majority of 60 per cent did. This proportion is yet significantly
lower than in other provinces, confirming a more flexible selection process. At the same time 78
households or 45 per cent of the Faryab sample fulfil one of the EVI criteria. Around half of these (37
households) are beneficiaries of the shelter assistance programme. Nine EVI households applied for
assistance, but were not chosen as beneficiaries.
244
The fact that vulnerability is not properly taken into account in the selection of beneficiaries leads to
some ‘misuse’ of the shelters. For example in Nisai Khona (Andkhoy district), in 7 or 8 shelters out of the
23 that could be directly visited, beneficiaries did not use their shelters as living spaces. Instead, the
shelters were part of larger compounds (with often 3 or 4 houses) and used as storage rooms or for
other various purposes.
NOTE for UNHCR: in Nisay Khona (Andkhoy district), the team could only find 23 of the 30 shelters listed
by UNHCR. Neither inhabitants of the village nor the Malik recognised the names of the 7 missing
beneficiaries.
As mentioned above, a lot of the inhabitants of Faryab province are former IDPs who came back to their
place of origin, where their houses had been destroyed. Faryab has an important caseload of conflict-
induced IDPs. Of the 60 IDP households interviewed, 43 per cent were shelter assistance beneficiaries.
Among those IPDs not included in the programme, half have an extremely vulnerable individual within
the household. Yet, the field team observed that in most cases, only returnees had been selected
despite the fact that IDPs had land and were living in their place of origin. All of the IDPs from Faryab
included in the quantitative data collection did in fact also designate this as their province of origin.
UNHCR confirmed the fact that IDPs were not included in the programme, despite their numerical
importance in the province. UNHCR’s main explanation was their lack of land but the team in the field
did not find land ownership to be an issue for IDPs in Faryab. In fact, all of the IDPs that did receive
shelter assistance reported that they did own the land prior to receiving assistance. Among the 32 non-
beneficiary IDPs, more than 81 per cent also did have access to land at the time of the survey. The issue
might be more related to a low level of awareness of IPs regarding specific IDP protection concerns and
the possibility to include them in the programme.
The ethnic context of Faryab is very complex with frequent tensions between neighbouring communities
of different ethnic background. In this context UNHCR staff smartly used the 10 per cent of shelters
allocated to the host community to diffuse tensions when necessary. One mean to do so was to
allocate these 10 per cent of the shelters to the neighbouring village of a different ethnicity. This is an
interesting practice that could be replicated in other contexts, where ethnic tensions are high and the
allocation of shelters might turn divisive.
Numerous beneficiary households, 43 per cent, had been living in tents upon return. The programme
therefore greatly improved their living conditions. In our surveyed sample, some 63 per cent of
households noted their economic situation had been better or far better following receiving shelter
assistance. The programme also helped conflict-induced IDPs to re-settle rapidly after being forced to
suddenly leave their village, as it was the case in Qaysar district, where families from the host
communities had had to host newly arrived IDP families.
245
Table A.20: Multi-dimensional Poverty Index by Beneficiary Status, in Faryab
% 21.33 78.67 100.00
% 50.00 50.00 100.00
% 13.40 86.60 100.00
Total 30 144 174
% 17.24 82.76 100.00
At the community level, the main impact of the programme was an increased pressure on water
resources, which are particularly scarce in the province. Yet, community leaders did not report any
particular tensions born out of this issue of water.
SUSTAINABILITY
A big concern for the sustainability of the shelter programme in Faryab province is the very poor access
to water of many beneficiary communities. The province is very dry and a provincial water distribution
system is yet to come. Of the surveyed 174 households, 32 per cent indicated that they do currently not
have access to safe drinking water as can be seen in Table A.21. Several villages visited by the team did
not have any water point accessible. In Aq Mazar and Mursha ghal (Khoje Sabz Posh district), people had
to walk 2.5 hours to reach the district centre in order to get water. The government pays for a water
tanker to supply the villages with water. But when it rains or during winter, the village is completely
inaccessible and cannot receive this water supply and the sustainability of governmental services in
Faryab province is not guaranteed given the deterioration of the security context. In the LAS located in
Northern Faryab (Andkhoy district), the absence of water has left the site mostly unoccupied.
Table A.21: Access to Drinking Water across Districts, in Faryab

Khoje Sabz
Andkhoy Qaysar Total
Posh
Boil water 0 7 8 15
% 0.00 11.86 14.55 8.62
Free potable water 13 6 29 48
% 21.67 10.17 52.73 27.59
Buy potable water 47 9 0 56
% 78.33 15.25 0.00 32.18
No safe drinking water 0 37 18 55
% 0.00 62.71 32.73 31.61
Total 60 59 55 174
% 100.00 100.00 100.00 100.00
246
The absence of water sources caused some specific challenges to the construction process as it:
 Slows down the building process, as beneficiaries have to find solutions to bring water in
order to make mud bricks. Some beneficiaries delayed the construction of their shelters
to wait for some rain falls (Khoje Sabz Posh district). UNHCR approved this delay in the
construction of shelters,
 Increases the costs of shelter building for beneficiary households. In Aq Mazar,
beneficiaries had gathered in groups of 4 or 5 families to be able to buy water and pay
for the transportation costs collectively, and
 Creates additional stress on a very rare resource for the community.
Perhaps surprisingly though, communities insisted to say that they had no intention to move anywhere
else as they owned their land and had shelters, stressing the stabilizing effect that the programme has
on certain communities.
The second main obstacle to the sustainability of the programme is the increasing insecurity. While
occupancy rates directly observed by our field team were satisfying, various stakeholders reported
some dynamics of secondary displacements due to insecurity. For example, the team observed this in
Qaysar district, one of the most insecure of the province, in the village of Bazardjay where one member
of a beneficiary household had been killed, triggering the departure of the whole family. Of the surveyed
households however all respondents in the province indicated that they plan to stay in the current
community permanently. Of the UNHCR beneficiaries two indicated that they were not currently living in
a UNHCR shelter.
PARTNERSHIPS
The coordination of stakeholders in Faryab province is made easier by the small number of actors
present in the province and by their proximity to one another. UNHCR has clearly endorsed a leading
role in the province due to the experience of its national staff there and to the relative weakness of
other organisations. For example, it appeared that UNHCR is supporting IOM in its efforts to assist
natural-disaster affected populations, despite them being out of its mandate. Moreover IOM supported
UNHCR in assessing and assisting conflict induced IDPs in 2010-2011.
The shelter programme – and other UNHCR Programmes in Faryab – is based on the work of reliable
implementing partners. Beneficiaries and other humanitarian actors praised the work of UNHCR IPs in
the province (DHSA; ORD; CHA). The satisfaction of beneficiaries in the field with the programme also
suggests that these IPs are satisfying partners, especially DHSA which had been in charge of
implementation for years. The IPs have several characteristics, which explain this reported reliability:
 Multiyear experience of programme implementation (more than 3 years for DHSA for
example).
 Multiple projects and various funding agencies. Contrary to what was observed in the
Central region where the IPs were dependent on the UNHCR shelter programme, which was
247
their unique programme and source of funding, IPs in the North/Northeast work for several
different donors and have multiple expertise. They also work across the region in various
provinces. DHSA for example works in Faryab, Jawzjan, Balkh and Sari Pul. CHA also has a
multi-provincial coverage in Afghanistan and counts CIDA, USAID and WFP among its donors.
 Their running costs are a lot cheaper than INGOs, which have to factor in every proposal
important fixed costs, including salaries for international staff and security measures.
 These IPs work with a pool of experienced engineers and field officers, who are re-hired
every year for the implementation of the shelter programme.
The flip side of this partnership system is that the IPs have learned how to work within the UNHCR
system and to maintain themselves in the structure. When the Central level introduced new procedures
to increase the transparency of the selection of IPs, the organisation selected changed (from DHSA to
ORD) but the staff operating in the field did not change. The same field officers are currently
implementing the shelter programme in the reintegration site of Faryab. It is probably a good thing that
experienced partners remain in place, but whether it is in line with the objectives of transparency that
drove the change in the selection procedure needs to be verified. The selection committee is based in
Mazar, and UNHCR staff in Faryab was perfectly aware of this practice.
There seemed to be quite a strong grip of the IPs on the programme in Faryab province. Two other types
of actors complained about their low involvement in the programme:
 Community leaders in Andkhoy district, who claimed that they had not been involved in the
selection of beneficiaries.
 The district governor of Andkhoy who claimed that he had requested the lists of
beneficiaries of the shelter programme but never received this information.
As the programme is largely based on remote management and monitoring in the province, it would be
better to multiply the checks on the implementation of the programme. Involving other stakeholders,
such as governmental authorities or community leaders, more fully would help monitoring the
programme.
UNHCR staff expressed a clear preference for NNGOs over INGOs arguing that NNGOs’ staff is a) more
experienced, b) more present in the field and c) cheaper. That might explain why no INGOs were
selected to implement UNHCR programmes on the 2012 reintegration site. According to INGOs, the
biding requirements made it difficult for INGOs to compete, as organisations had to submit one proposal
for each component of the reintegration site. Therefore INGOs had to integrate their fixed costs in each
of their proposals, whereas they could have levelled these costs in a global proposal covering several
projects of the reintegration site.
248
JAWZJAN – PROVINCIAL OVERVIEW
CONTEXT
Jawzjan is a rural province of Northern Afghanistan bordering Turkmenistan. UNHCR registered 126,536
individual assisted returns in Jawzjan between 2003 and August 201250. Among the Northern provinces,
Jawzjan is an important destination for return. According to UNHCR statistics, the caseloads of conflict-
induced Internally Displaced Persons (IDPs) are limited in the North compared to other regions of
Afghanistan. Yet, in 2012 with over 8,000 IDPs, Jawzjan province counted one of the most important
caseloads of IDPs in the North-Northeast regions along with Faryab51.
Jawzjan is less of a priority for national and international stakeholders than other key provinces in the
country. The province therefore presents a low number of stakeholders and depends quite extensively
on Mazar-e-Sharif (Balkh province) for decision-making, where most organisations’ sub-offices are
located. For the shelter programme, this means:
a) That there are a limited number of actors on which to rely or with which to partner in the
province.
b) That a diminution in shelter assistance or in assistance to returnees has potentially more impact
than in other provinces more extensively covered by international and national actors52.
In terms of security, Jawzjan presents the same kind of profile than other Northern provinces, where the
security landscape is complex and fluid. Yet, contrarily to Faryab province, a majority of Jawzjan districts
are rather quiet and accessible to humanitarian actors. Even if the security situation is deteriorating, it is
easier to implement assistance programmes in Jawzjan than in other neighbouring provinces in the
North. A few areas are now out of access, even for non-UN staff: for example Mangajik district or the
desert of Dashte Laily, near the provincial capital Sheberghan.
UNHCR distributed a bit more than 1,800 shelters in Jawzjan between 2009 and 2011. As table A.22
shows, our provincial sample of Jawzjan comprises 220 households, of which more than half, 118, are
UNHCR beneficiaries. Two households received shelter assistance from IOM. Moreover, 66 per cent of
our sample are official refugee returnees, while 16 per cent are non-refugee returnees, 2 per cent are
IDPs and 16 per cent never migrated. Among the UNHCR beneficiaries refugee returnees are the largest
group with 85 per cent.
50
UNHCR Afghanistan, Voluntary Assisted Returns – 03.2003 to 08.2012.
51
UNHCR, Afghanistan estimated IDP Population by district of displacement – 30.04.12 .
52
This is also true of neighbouring Sari Pul province.
249
Table A.22: Provincial Sample, Jawzjan
Total

% 84.75 100.00 43.00 65.91
% 10.17 0.00 24.00 16.36
IDP 5 0 0 5
% 4.24 0.00 0.00 2.27
% 0.85 0.00 33.00 15.45
Total 118 2 100 220
% 100.00 100.00 100.00 100.00
DESIGN
All the shelters visited by our team in Jawzjan province were 2-room shelters, with one-corridor and
iron beams. Among the 118 beneficiaries interviewed one had built a one-room shelter, while the rest
was all two-room shelters. Three respondents indicated that they got to choose the size of their shelter,
while for 98 per cent this was not the case. In numerous cases, these shelters were built next to other
UNHCR shelters or to normal houses within the family compound, surrounded by a protecting wall that
pre-existed the construction of UNHCR shelters.
Most of the shelters visited were in a satisfactory state. In a lot of cases, the shelters had clearly been
progressively turned into proper houses. The state of the shelters depended on a) the wealth of the
beneficiary household; b) on the main usage of the shelter. In multiple cases, the shelters served as a
secondary house or guesthouse. In these cases, the shelters were in a rather poor state, as most of the
investments of the households had focused on their main living space. These kinds of shelters were
often divided in two parts: one ‘reception’ room properly carpeted and furnished and one ‘storage’ room
with no floor or proper glass window. Glass windows were sometimes missing on the shelters; about a
bit more than a half of the shelters visited had their glass windows installed. In most cases, the
beneficiaries had preferred spending the money on food and water.
Shelters in Jawzjan were often used for carpet weaving. Carpet weaving is one of the only economic
activities women have access to in the North. In these cases, the frame used to weave carpets occupied
most of one of the two rooms, while the other room was typically used as the main living space.
Beneficiaries in Jawzjan had mostly respected the UNHCR design. A few families made some
arrangements to their shelters, including adding a bathroom in the corridor or extending the size of the
rooms either by getting rid of the corridor or by adding beams to the initial package. Interestingly
though, complains about the size of the shelters were not as prevalent as in other regions. Several
shelters were occupied families with less than 6 members. This is probably linked to the size of
250
households in the North, smaller than in Pashtun areas. The average household size for the whole
sample is 9.0, with large differences between households in the Northern region (6.9) compared to the
South (10). In Jawzjan an average household has 6.6 members. This sort of regional differences calls for a
region-based approach to shelters rather than a ‘one size fits all’ package. The OCHA sub-office insisted
on the importance of designing programmes tailored to the specific context of each province.
Overall, the visit of shelters in Jawzjan showed that the design of the shelters suited the needs of the
beneficiaries.
Problems during construction: More than one third, 38 per cent, of the surveyed beneficiaries indicated
that they encountered some kind of problem during the construction of their shelter. Table A.23 shows
the different types of problems the beneficiaries encountered, the main problems being not having
enough financial capital and insufficient access to water.
Table A.23: Problems during Construction, Jawzjan

UNHCR Beneficiaries
N=45 %
In the entire province, a main problem encountered by beneficiaries to build their shelters was the
difficult access to water (48.9%). As water is necessary to make mud bricks, beneficiaries struggled to
get the building material they were required to provide. This is especially true as the programme often
started during summer, making it impossible for beneficiaries to find water. Two solutions were found:
 Either the beneficiaries or the IPs asked for some time extensions so as to wait for
rainfalls. UNHCR usually agreed on these delays.
 Or the water was brought to the village by rented tankers, paid by UNHCR or by the
community. One litre of water in the region costs between 7 and 10 AFS.
251
Two points are worth noticing regarding the issue of water:
a. Both the IP and the beneficiaries were able to find solutions and to overcome the
difficulty without too much trouble, showing a welcome flexibility and adaptability in the
implementation of the programme.
b. The timing of construction is important. It appears that in most cases, the programme
started late and that the bulk of the construction work took place during summer when
water is not available. Starting earlier during spring could make more sense for water-
deprived regions like Northern Afghanistan. It would also help reducing the risk of
entering winter with incomplete shelters.
A major challenge for the shelter programme in Jawzjan is the high risk of floods affecting the
province. The 2012 floods destroyed some shelters built in 2011 in the district of Khawaja du Khoh. The
mud used to build both traditional houses and UNHCR shelters is not resistant at all against flooding,
leaving beneficiaries and non-beneficiaries very vulnerable to this type of natural disasters. Despite
these high risks, preventive measures are very scarcely taken into account in the programme. The only
measure that was implemented is to recommend beneficiaries to build their shelters higher up, onto
60cm high foundations. Yet, this was not systematically implemented in the province. This measure
would reduce the risks in case of limited flooding but, as the foundations in Jawzjan are also in mud and
cannot be built with stones due to the shortage of stones in the area, this measure will only have a
limited impact in case of major floods. ANDMA confirmed that there had not been any consultations on
risks with UNHCR prior to the implementation of the programme in the province and that ANDMA and
UNHCR had very little relations.
Discussions with the management of UNHCR in Mazar showed that the 2012 flood served as a wakeup
call for stakeholders in the North. Mazar programme staff intended to discuss the introduction of some
preventive measures in the shelter programme during the Shelter working group discussions in January
2013 in Kabul.
ZOA’s TECHNICAL RECOMMENDATIONS TO UPGRADE UNHCR SHELTER PACKAGE

Acc. to ZOA’s Shelter PM
1. Include a sidewalk minimum 60cm wide to properly evacuate water and avoid damaging foundations
in case of heavy rain and snowfall.
2. Include gutters in the package for water evacuation.
3. Include anti-rust painting for the beams for greater sustainability.
4. Factor in the costs of the painting of doors and windows for greater sustainability.
5. Be more specific in the guidelines about the construction of walls. Specifically strengthen anti-
earthquake measures: include wall bracing and anti-earthquake square bricks (35x35x7cm).
6. Factor in all the costs in the shelter package including costs for the ‘thatch mat’ or for ‘karnes’.
Price per unit of shelters including these improvements: $2400. (According to ZOA’s estimation).
252
The communities confirmed that they had received a close technical assistance from the IP during the
construction process. All the shelters were visited once a week in the province by the IP project manager
(DHSA permanent staff) and by the engineers that DHSA had hired for the project. The quantitative
fieldwork shows 12 per cent of beneficiaries that had not experienced any monitoring of their shelter.
According to all actors, the selection of beneficiaries was conducted by a beneficiary selection
committee, which included the community leaders, representatives from DHSA (including one woman),
representatives of UNHCR (in safe areas) and representatives of DoRR in most cases. The SAP guidelines
seem to have been respected more systematically in the province. Yet, there were conflicting reports on
the involvement of DoRR in the selection process in Jawzjan. While the head of DoRR complained about
his department being side lined and excluded from the selection process by UNHCR, beneficiaries
confirmed the presence of some representatives of DoRR when the selection was conducted in their
village. The IP declared that DoRR was always involved when they were in charge of the selection of
beneficiaries, but that it was not always the case when UNHCR had the lead on the selection. UNHCR and
the DoRR of Jawzjan had conflictive relations in 2009 and 2010, which explain this exclusion. The
relationship has improved greatly since 2011 according to both parties.
In a few areas of the province, UNHCR and its IP recently lost access and had to rely on community
leaders for the implementation of the programme. This is notably the case for Dashte Laily. In these
cases, it is very difficult to assess the efficiency and the reliability of the selection process. The head of
DoRR complained heavily about these ‘grey areas’, where the programme was implemented without
any overview, and reported some cases of corruption occurring in these areas. These accusations could
not be crosschecked, as our team could not visit these areas either. Yet, this type of situation raises the
question of the appropriateness of remote monitoring for the shelter assistance programme, as the
selection of beneficiaries often proves controversial and conflict-ridden.
 Selection Criteria
The selection criteria were not clear to most members of the beneficiary communities. Beneficiaries,
non-beneficiaries and community leaders alike considered that the VRF and the date of return were the
main criteria to be selected in the programme. More than 85 per cent of the beneficiaries in Jawzjan are
recognized refugee returnees. Discussions with non-beneficiaries showed that the criterion of recent
return was one of the prime criteria used by the IP to make the selection on the ground. Yet, other
factors, such as vulnerability were not perceived by communities to play a role at all. Around 38 per cent
of our sample in Jawzjan can be classified as extremely vulnerable and slightly less than half of these did
receive shelter assistance. At the same time six EVI households that applied were rejected assistance.
This fuels the impression that any recent returnee would be entitled to receiving a shelter rather than
those most in need.
253
The selection of beneficiaries appeared to be one of the weakest points of the programme in Jawzjan.
This was evidenced by the following observations:
 In Shobash Khorde Turkmenia, a majority of households visited did not present any signs of
particular vulnerability. On the contrary, the number of assets, the presence of livestock,
the size of properties and the fact that a lot of them included cultivated land plots suggested
that a significant proportion of people selected by the programme might have been able to
build their shelters without external assistance. This is might be linked to the practice of
UNHCR to allocate a certain number of shelters to an area prior to any local need
assessment.
 A high proportion of the shelters visited were used for other purposes than living. Slightly
more than 20 per cent of the shelters visited in Jawzjan for the purpose of this study (13 out
of 64) were not used as a living space by beneficiary households. Typically, these shelters
would be used as a) storage rooms, b) barns for livestock and harvest and c) a guesthouse
next to the family house. In one case, the shelter was even used as a private mosque for one
family. In these cases, it is legitimate to wonder whether shelter assistance was really
needed. It also suggests that there is room for reducing the scope of the programme, hence
its cost, by tightening the selection process.
 The level of awareness of actors in charge of the selection of beneficiaries about the
necessity to prioritize the most vulnerable households was strikingly low in the North. Even
for the management of UNHCR there, it appeared that the fact of being a returnee was the
determining factor and that vulnerability was very secondary. ‘All returnees who come back
to the North are in great need of shelter; they all need shelter’. The three most important
criteria used to select beneficiaries were a) the presentation of the VRF, b) the date of
return and c) access to land. Interviews with community leaders and with the IP (DHSA)
showed that the integration of EVIs in the selection process was not considered a priority. If
they happened to have been selected, EVIs were identified later on in the process and
considered for potential extra-assistance. UNHCR’s lists showed that another criterion used
to select beneficiaries is the type of sheltering that potential beneficiaries had at the time of
selection (tents, rented house, living with relatives). If it is undoubtedly an interesting
criterion to take into account in the selection, the main criteria should be the ability of the
household to build a shelter by its own means rather than the pre-existence of house.
 The programme targeted almost exclusively returnees. IDPs and members of the host
community were very rarely integrated in the programme, as confirmed by the quantitative
survey (cf. Table A.22). Some ‘returnee-IDPs’ were also included based on the fact that they
had a VRF. Yet, as the protection of IDPs does not represent as a major challenge in Jawzjan
as they do in other provinces, the focus on returnees may be more justified there. The DoRR
just began distributing land plots to IDPs in recently opened LAS, which might be worth
considering for the future inclusion of IDPs in the shelter programme.
254
The fact that beneficiary families did not use the shelters as their main living space suggests that
shelter assistance did not respond to a compelling humanitarian need for these families. A lot of
evidence indicated that selecting the most vulnerable was not properly mainstreamed in the
implementation of the programme in the field and that a tighter selection process could lead to a
reduction of the scope of the programme without endangering its positive humanitarian impact on the
reintegration of returnees and IDPs.
At the household level, the shelter programme helped protecting returnee families from the protection
risks related to over-crowded houses and to temporary sheltering. Most beneficiaries surveyed, 35 per
cent, used to live in tents or to share houses with their relatives, 46 per cent before starting the
construction of their own shelters. Several protection issues, especially for children and women, have
been linked to the fact of living in over-crowded houses and to the stress related to migration. The
shelter programme helps reducing these risks.
Beneficiaries also perceived the impact of the shelter assistance programme on their economic situation
as positive. While 73 per cent said they were now better or far better off, only 9 per cent considered
themselves worse off than before. These figures are significantly higher than in other provinces.
shows that those households which received shelter from UNHCR are much less deprived.
Table A.24: Multi-dimensional Poverty Index by Beneficiary Status, in Jawzjan


% 33.05 66.95 100.00
% 0.00 100.00 100.00
% 8.00 92.00 100.00
Total 47 173 100
% 21.36 78.64 100.00
Interestingly, the issue of indebtedness was not central for beneficiaries in Jawzjan. Community
leaders noted that these people would have got indebted anyway and that the programme was rather
decreasing the risk of indebtedness for these households than it was increasing it. Observations from
the field tend to suggest that participating in the shelter programme tends to reduce the risks of
unsustainable indebtedness. Indebtedness has to be analysed comparatively and to be considered in the
wider social context of rural Afghanistan, where informal credit and recourse to informal loans are
255
prevalent and a basis of social relations53. This means that households would take out loans to build their
shelters – or for other purposes – regardless of their participation in the programme, as it is a common
practice in rural Afghanistan. The quantitative data shows that 36 per cent of beneficiaries indicated that
participating in the shelter assistance programme increased their household debt. However, 41 per cent
reported that it actually decreased their debt level and 7 per cent said that there was no change. 15 per
cent indicated not having any household debt. This therefore suggests that the programme does not
necessarily increase debt, but might also be able to limit indebtedness in a context where loans are
common for the purpose of building a shelter.
Overall, field observations suggested that beneficiaries ended up being better off than the rest of the
communities in the North because a) they were not always particularly worst-off in the first place given
the loose selection process and b) in communities almost exclusively formed of returnees and migrants,
they did not have to get as indebted as others to build their shelters.
At the community level, the programme sustained a short-term boost in employment, which was
confirmed by community leaders. Typically, each shelter would require hiring 2 to 3 unskilled labourers
for a period of one month.
Finally, the programme seemed to play a role in the reintegration process of returnees in their host
communities. Even if members of the community are usually reluctant to report internal tensions within
the communities, there were reports of the additional pressure that the arrival of returnees posed on
the host community. This was especially the case for housing – with a lot of families having to share their
houses or their compounds with relatives coming back from abroad – and for access to key resources,
especially water in the context of Jawzjan. The shelter programme helped release this pressure and
diffuse the potential tensions that could have arisen from important influx of returnees in the
communities.
SUSTAINABILITY
Based on our field observations, the shelter programme had supported a sustainable reintegration of
returnees in Jawzjan, at least in the PSUs directly visited by our team. There, the occupancy rate was
high or very high. Of the 118 beneficiaries interviewed all indicated that they were currently living in the
UNHCR shelter. Out of the 64 shelters directly visited by the programme manager, only 2 shelters were
completely empty and showed no sign of life at all. The shelters had clearly been appropriated and quite
a lot of investments had often been made to turn them into proper houses. According to community
leaders, if temporary work migration is a common strategy for the heads of households, secondary
displacements involving entire families are very marginal in these areas of Jawzjan. Our field
observations and the high level of occupancy that we observed confirmed this. The fact that some
shelters were not used as the main living spaces of beneficiary household is not a relevant indicator of
the sustainability of the reintegration supported by the programme; rather it is an indicator of some
flaws in the selection process. Additionally, almost all respondents in the province indicated that they
53
See for example AREU (2007), Finding the Money: Informal Credit Practices in rural Afghanistan, AREU (2009), From Access to
Impact: Microcredit and rural livelihoods in Afghanistan.
256
plan to stay in the current community permanently. Only one household reported having plans to move
again due to a lack of work opportunities.
Yet, it must be noted that this observation does not apply to the entire province of Jawzjan, as our team
was not able to access some insecure areas, where the occupancy rates were likely and reported to be
lower. This is especially the case for Dashte Laily and for the Pashtun villages surrounding Shri Abad
Charpical in Sheberghan district. We had reports of low occupancy rates in these areas, but this
information could not be directly verified and might have been biased by the poor relationship existing in
the past between the head of DoRR and UNHCR in Jawzjan. Yet, this tends to confirm that implementing
the shelter programme in insecure areas is unsustainable because a) secondary displacements are more
likely to occur and b) no proper monitoring can be conducted reducing the visibility of the programme to
almost nothing.
Despite this satisfying assessment of the sustainability of the reintegration entailed by the programme in
Jawzjan, natural disaster – flooding in particular – represents an important threat to the sustainability
of the shelter programme - a threat that UNHCR should take more into account in the design of its
programme. Our team could observe at least 5 shelters, which had been partially or entirely destroyed
by 2012 floods. Without factoring in the overhead costs, this represents a net loss of between $6000 and
$10,000 of UNHCR investments. Despite all stakeholders admitting that the risk of flood is important
every year in the province, the culture of prevention is surprisingly low among the relevant stakeholders,
especially ANDMA or DHSA (IP). Other organisations working in similar contexts, such as ZOA or NRC,
have made the choice of simply forbidding the construction of shelter in areas subject to flooding.
UNHCR should consider this kind of measures and greatly enhance the level of awareness of its IPs about
this sort of risk.
PARTNERSHIPS
UNHCR relied on the same IP – DHSA – for the five Northern provinces of Afghanistan. This partner has a
long experience in the implementation of UNHCR’s programmes, as it has been working with the
organisation since 2002. DHSA has a good pool of experienced staff, especially engineers, who are hired
on a project-basis. We received no complaints about the practices of DHSA in the North, which was
widely praised by beneficiaries and community leaders for its frequent visits to the field and for the
technical assistance it provided to beneficiaries. It must be noted that, following the change in the
procedure of selection of IPs, DHSA is no longer the IP of UNHCR for the shelter programme in the North.
ORD was preferred to DHSA for 2012. DHSA was told that the costs they calculated for the procurement
of material were too high. If opening the process of selection of IPs has the advantage of a more
transparent and fair procedure, it also implies weakening some long-term relationships with local NGOs,
which had progressively been capacitated. DHSA is a case in point.
Since 2011, the Northern sub-office relies on a monitoring IP. The introduction of this new system raises
some questions. UNHCR relied on a local NGO called NPO to monitor 8 provinces of the North and
Northeast. The system has several shortcomings, which endangers the quality and the sustainability of
the programme:
257
 The monitoring is exclusively based on ‘quantitative’ monitoring, by which the monitoring
officer must fill in a ‘checklist’ focusing mainly on the progress of the construction process.
They are neither required to gather some qualitative data in the field nor to produce any sort
of analysis on the way the programme is implemented in the field. When asked to reflect on
the strengths and weaknesses of the programme, the IP proved unable to provide any
analytical observations, showing the need for some serious capacity building with monitoring
IPs. Overall, this leads to a very poor monitoring system when not conducted directly by
UNHCR staff, through which a share of crucial information is never reported to UNHCR.
 There is a conflict of interest as NPO is also an implementing partner for UNHCR on some of
its most expensive projects, including 4 projects in the reintegration site of Mohjer Qeshlaq
in 2012.
 The budget received by the monitoring IP covers only for one monitoring officer per
province. This officer is often in charge of monitoring several different programmes of
UNHCR at a time and the monitoring IP is required to monitor not only insecure areas
inaccessible to UNHCR staff but all the shelter implementation sites. UNHCR has taken the
habit of outsourcing monitoring even when it is not out of necessity and this despite the
obvious shortfalls that come with such a system.
 Finally, this monitoring system is not complete as some areas remain out of reach even for
the monitoring IP, as it is the case in Dashte Laily and Mangajik (according to the IPs; check
with UNHCR). In these areas, the programme relies only on remote monitoring.
UNHCR had a poor relationship with the DoRR in Jawzjan in the past. Both parties declared that the
relationship had improved a lot with more involvement from the DoRR and more constant
communication from UNHCR. Yet, the plans of the DoRR and of UNHCR do not match as DoRR is
expecting UNHCR to help in the new LAS that it develops in Jawzjan, while UNHCR chose a different site
for its 2013 reintegration site. There are some linkages possible in Jawzjan province that could prove
interesting for the future of the shelter programme:
 Due to the disastrous nature of 2012 flooding, some actors have decided to tackle the issue
and have started developing Disaster Risk Reduction (DRR) programmes in Jawzjan. Save the
Children and Action Aid are particularly involved in these activities. It would be particularly
relevant to extend this type of activities to the beneficiaries of the shelter programme.
Despite its low capacities, coordinating with ANDMA, as the relevant governmental body in
charge of risk mitigation and response, before implementing the shelter programme should
be necessary for UNHCR. Some programmes of cash-for-work could also focus on
strengthening collective preventive measures against floods (construction of dykes, cleaning
of canals, digging of canals…).
 In this highly rural province, agriculture-enhancement activities are also particularly relevant.
Welthungerhilfe and Action Aid, both active in the province, would be interesting partners to
link up with to support the livelihoods of shelter beneficiaries. Distribution of selected seeds,
258
of appropriate fertilizers and basic processing and/or business skills could support the
economic development of beneficiary communities.
259
SARI PUL – PROVINCIAL OVERVIEW
Areas visited: Zeri Chaghat; Sabzi Kalan (Suzma Qala district)
CONTEXT
Sari Pul received fewer than 40,000 returnees over the past ten years. The Northern districts of Sari Pul
(Sari Pul and Sayyad) also registered the arrival of significant numbers of IDPs. The total population of
IDPs in the province amounts to more than 10,000 individuals. As other in provinces in the North there is
a high risk of flooding in the province, meaning specific preventive measures must be taken into account
for the implementation of any shelter programme.
UNHCR built 804 shelters between 2009 and 2011 in Sari Pul. Our research team surveyed 100
households in the province, of which 56 per cent were UNHCR beneficiaries. Moreover 66 per cent of
those households were official refugee returnees, 17 per cent were non-official refugee returnees, 1 per
cent was IDPs and 16 per cent never migrated.
Table A.25: Provincial Sample, Sari Pul

UNHCR Non-
Total
% 87.5 38.64 66.00
% 10.71 25 17.00
IDPs 0 1 1
% 0 2.27 1.00
No Mobility 1 15 16
% 1.79 34.09 16.00
Total 56 44 100
% 100.00 100.00 100.00
DESIGN
Nearly all of those shelters visited, 98 per cent were two-room shelters, and all had surrounding walls.
Moreover, the state of shelters was satisfactory. Beneficiaries were satisfied with the quality of the
material they received to build their shelters, especially the iron beams. In Zeri Chaghat village, NRC had
also constructed some shelters but their package had wooden door and windows. Beneficiaries noted
that they preferred UNHCR shelter package.
The lack of water was a problem for the construction of the shelters, with 67 per cent of those
households surveyed mentioning insufficient access to water for construction. Many of these households
therefore had to rent water tankers.
260
Beneficiaries received a visit of engineers one month after starting to build their shelter suggesting that
the level of monitoring of the shelter programme in Sari Pul was low. However of those households
surveyed 86 per cent said monitoring from UNHCR had taken place.
Community leaders played an important role in the selection of beneficiaries. In both villages, IDPs were
totally left out of the selection, which was almost exclusively based on the presentation of a VRF and on
land ownership. This is confirmed by the quantitative data which show that there were no IDPs included
in the selection for the SAP. In fact, all beneficiary households surveyed had a VRF while nearly all, 98 per
cent, owned the land their shelter was built on beforehand. Singling out returnee households created
tensions within some of the villages. The inclusion of the host community in the programme is indeed
quite low, as we found only one non-migrant beneficiary despite the fact that this group represented 16
per cent of the overall sample.
Non-beneficiaries did not really understand the rationale and the criteria behind the selection of
beneficiaries and were sometimes resentful to see that only returnees had been included in the
programme. This calls for a more fine-tuned approach to selection, which would adapt to the profile of
each location. Moreover, some non-beneficiaries complained about the selection process, arguing that
the members of the CDCs selected their relatives in priority.
Beneficiaries noted that without the assistance of UNHCR, it would have been impossible for them to
build their shelters because of their limited financial capacities. The programme improved the living
conditions of beneficiaries. In fact, 73 per cent of beneficiary households noted that their economic
situation was better or far better compared to before assistance. Compared to non-beneficiaries, our
multi-dimensional poverty index (MPI) presented in Table A.26 shows that those households which
received shelter from UNHCR are less deprived.
Table A.26: Multi-dimensional Poverty Index by Beneficiary Status, in Sari Pul

UNHCR Beneficiary 11 45 56
% 19.64 80.36 100.00
Non-Beneficiary 5 39 44
% 11.36 88.64 100.00
Total 16 84 100
% 16.00 84.00 100.00
The programme did not guarantee the parity between beneficiaries and non-beneficiaries, as the initial
vulnerability of households was not taken into account in the selection. EVIs were left out of the
261
selection, leading to an economic gap between beneficiaries and non-beneficiaries after completion of
the programme.
At the community level, all households surveyed considered the programme had a positive impact on
the community.
SUSTAINABILITY
The sustainability of the shelter programme implemented in Sari Pul depended on the location of the
settlements. While in Sabzi Qala beneficiaries were satisfied and enjoyed proper access to basic services,
the situation in Zeri Chaghat was different. There, the absence of drinking water, employment and basic
services clearly endangers the sustainability of the programme. About half of the shelters were either
empty or destroyed, as people had left to the nearest urban centres (Mazar-e-Sharif or Sari Pul). The lack
of water was a major threat as it continues to trigger secondary displacement in the province. Moreover,
Sari Pul is a flood-prone province where prior risk assessment and specific preventive measures should
be in place. Overall however, all households surveyed planned to stay in their current community
permanently.
262
KUNDUZ – PROVINCIAL OVERVIEW
Note: The research team only spent a few days in Kunduz province. The surveyed locations were close to
Kunduz city. The following remarks are therefore indicative and specific to the implementation areas
close to the provincial centre.
CONTEXT
Kunduz is one of the most volatile provinces of the North with a significant presence of insurgent groups,
which increasingly constrains the access of aid organisations. In terms of migration, Kunduz is one of the
more important regions of return in the North, with about 278,000 voluntary returnees registered over
the past ten years. The province however only counts a limited number of IDPs.
The scope of the UNHCR shelter programme was quite important in Kunduz with a total of 1,045 shelters
built between 2009 and 2011. Our research team surveyed 120 households in Kunduz, of which exactly
50 per cent were UNHCR beneficiaries. Moreover, 60 per cent of those households surveyed were official
refugee returnees while 13 per cent were non-refugee returnees, 15 per cent were IDPs and 12 per cent
never migrated.
Table A.27: Provincial Sample, Kunduz

UNHCR Non-
Total
% 81.67 38.33 60.00
% 1.67 25 13.33
IDPs 6 12 18
% 10.00 20 15.00
No Mobility 4 10 14
% 6.67 16.67 11.67
Total 60 60 120
% 100.00 100.00 100.00
DESIGN
Nearly all, 88 per cent, shelters visited by the research team were two-rooms and were usually in a good
state. Beneficiaries were by-and-large satisfied with the quality of the materials they received including
iron beams, doors and windows. What’s more, most shelters had surrounding walls.
The main complaint related to the size of the rooms of the shelters, which beneficiaries found too small
for their needs. Some beneficiaries wanted to change the design of their shelters during the construction
but could not because the IP required them to respect UNHCR’s guidelines. Some changed the design
after the handover and removed the corridor to get more space.
263
Beneficiaries encountered several challenges during the construction of their shelters. The lack of water
in particular rendered construction particularly challenging, with 43 per cent of those households
surveyed noting it as a problem. Beneficiaries had to thus invest significant amount of their own money
to get water. In Julga Uzbekia, beneficiaries reported that they had had 300Afs per hour to pump water
from the river, while other people took on a loan from the FMFB to pay for the water and build their
shelter. In fact 32 per cent of households surveyed said debt had increased following participation in the
programme.
Beneficiaries also did not have the financial means to hire skilled and unskilled labour to build their
shelters. The majority, 90 per cent, noted the lack of money necessary to build their shelters. In some
cases, the women participated in the construction by bringing the water from the river and preparing the
clay for the bricks. Children also had to give a hand. Another issue was the lack of straw to make clay,
which some households also had to pay for. Overall, many households had to spend their own money to
pay to build the shelter, including 97 per cent of those surveyed.
Some reported that the maliks and community leaders mostly conducted the selection of beneficiaries.
In Chartaq Sedarak a few people complained about the selection process claiming that community
leaders had chosen relatives.
The main criteria of selection were a) the VRF and b) land ownership. Of those assisted households
surveyed, 96 per cent had a VRF while 87 per cent owned the land their shelter was built on beforehand.
Non-beneficiaries noted that people who do not have land were not selected for the programme. Yet, as
threats of evictions and land conflicts were reported in the area (Chartaq Sedarak), it made sense for
UNHCR to prioritize the sustainability of the shelters in these areas of implementation by making sure
that beneficiary households would not be under the threat of eviction.
At the household level, the programme adequately responded to an urgent need as a part of
beneficiaries had reportedly been living in tents for years before receiving UNHCR assistance. In fact,
some 32 per cent of households surveyed did not have access to a home 1-month before receiving
assistance. Beneficiaries also noted that the programme greatly improved their living conditions and the
protection of their children against harsh summers. Economically, 60 per cent of households surveyed
said they were better off following assistance, while 18 per cent reported being the same. Compared to
non-beneficiaries, our multi-dimensional poverty index (MPI) presented in Table A.28 shows that those
households which received shelter from UNHCR are less deprived.
264
Table A.28: Multi-dimensional Poverty Index by Beneficiary Status, in Kunduz
% 33.33 66.67 100.00
% 15.00 85.00 100.00
Total 29 91 120
% 24.17 75.83 100.00
At the community level, the programme helped developing the settlement into a proper village as the
shelter assistance triggered further assistance, including the provision of electricity by the government.
The DRRD also provided assistance through the NSP, which helped digging wells and paving the roads in
these villages. Overall, all respondents thought the programme had a positive or very positive impact on
the community as a whole.
SUSTAINABILITY
The sustainability of the shelters built by UNHCR in the areas surveyed by the research team is not under
threat thanks to their good location. Because these areas are near Kunduz city, beneficiaries were able to
commute and find daily work in the city. All the shelters were occupied, and all households surveyed
planned on staying in their current community permanently. Even though, there were worrying reports
of threats of eviction and violence in the area, all beneficiaries had bought their land and were not
directly concerned by these threats.
265
TAKHAR – PROVINCIAL OVERVIEW
Note: the research team only spent a short time in Takhar. The following remarks are therefore only
indicative and refer mostly to the area of Qulbarsi Tojarbashi.
CONTEXT
Located in the Northeast of the country, Takhar is a very rural province, characterized by a high reliance
on subsistence agriculture and livestock, with an established tradition of meat and milk production.
While the province is well connected to Kunduz and to Badakhshan thanks to a good road, accessing the
rest of the province requires off-road driving and villages are very easily cut-off during winter. One of the
major issues in Takhar is the risk of natural disasters as numerous villages are under the threat of
landslides and flooding.
Takhar has not experienced significant movements of internal displacement and received a total of
approximately 75,000 returnees over the past ten years. It is therefore not a region of high return,
especially when compared with the neighbouring provinces of Kunduz and Baghlan.
The scope of the UNHCR shelter programme was limited in Takhar, with 641 shelters being built
between 2009 and 2011. Our research team surveyed 70 households in Takhar, of which 56 per cent
were UNHCR beneficiaries. Moreover, 80 per cent of those surveyed were IDPs while only 7 per cent
were official refugee returnees, 3 per cent were non-refugee returnees and 10 per cent never migrated.
Table A.29: Provincial Sample, Takhar

UNHCR Non-
Total
Beneficiary Beneficiary
% 7.69 6.45 7.14
% 0.00 6.45 2.86
IDPs 36 20 56
% 92.31 64.52 80.00
No Mobility 0 7 7
% 0.00 22.58 10.00
Total 39 31 70
% 100.00 100.00 100.00
DESIGN
All shelters visited in Takhar were standard two-room shelters. The research team observed that shelters
there were not in a good state, with many shelters lacking windows or glass panels. Beneficiaries praised
the quality of the iron beams received but complained about the quality of wooden doors and windows,
which deteriorated rapidly due to the high temperature changes they were subjected to. Beneficiaries
noted that building the shelters proved difficult because most did not have the financial means to hire
266
skilled worker (masons) or unskilled workers. In fact, 94 per cent of households surveyed noted they ran
out of money during the construction process. Thus, male members of beneficiary households had to
build the shelters themselves, and could not work and earn money during construction putting some
households in a precarious financial situation. Most beneficiaries had to borrow money to cover the
living costs of their household during the construction of their shelter, with 40 per cent mentioning their
debt had increased following participation in the programme.
SELECTION PROCESS
The selection process uncovered some land conflicts and local ethnic tensions between the Tajik and the
Gujur communities, which cohabit with difficulty in the area. According to local leaders, there was no
real selection process; rather all the members of a single community (Gujur community) received shelter
assistance. This specific tribe is originally from Farkhar district, where they owned land and livestock. As
they fought with the Taliban against the Jamia’t forces during the civil war, the community fled to
Pakistan and its property were grabbed by the Tajik community. Upon return, this community could not
get its land back and finally decided to buy land around Qulbarsi Tojarbashi, where UNHCR helped them
building their shelters. Surrounding Tajik villages did not receive assistance, as their economic situation
was relatively better. The specific profile of this beneficiary community explains why of those selected,
92 per cent were IDPs. Moreover, 74 per cent owned the land the shelter was built on beforehand, while
41 per cent were characterized as extremely vulnerable.
The programme considerably facilitated the return of the Gujur community and helped them re-adapt in
a difficult environment, where their land and property had been taken. Thus, the programme possibly
diffused some potential inter-ethnic land conflicts by offering an alternative to this community. Yet
singling out one specific community in a multi-ethnic context might fuel further pre-existing inter-ethnic
tensions.
Only 33 per cent of beneficiaries considered their current economic situation better compared to before
assistance, however all households were of the opinion that the programme was positive for the
community as a whole. Compared to non-beneficiaries, our multi-dimensional poverty index (MPI)
presented in Table A.30 shows that those households which received shelter from UNHCR are less
deprived.
267
Table A.30: Multi-dimensional Poverty Index by Beneficiary Status, in Takhar
% 2.56 97.44 100.00
% 3.23 96.77 100.00
Total 2 68 70
% 2.86 97.14 100.00
SUSTAINABILITY
The sustainability of the shelter programme in this specific area is questionable. The occupancy rate in
this area of Takhar was quite poor, with a number of shelters left empty and unoccupied. This is due to
the poor location of the village, far from the cities of Taloqan and Kunduz where most of the men go to
find daily work. The village suffers from a lack of livelihood and employment, which threatens the
sustainability of this shelter programme. The community regrets having bought the land in this specific
area and has seen a lot of departures. The community leader noted that it is likely that this drain will
continue because of the absence of work in the area. Still, of those households surveyed, all planned to
stay in their current community permanently.
268
HELMAND – PROVINCIAL OVERVIEW
CONTEXT
Helmand province has been characterized by large intra-provincial displacement, due to the intense
fighting in the province especially since the beginning of the military surge in 2009. Helmand is therefore
one of the top three provinces in Afghanistan in terms of a) number of people currently living in
displacement (about 63,000) and b) province of origin of IDPs (about 47,000). One of the locations
surveyed was Mukhtar Camp, where about 7,000 IDPs live after being forced to leave their place of
origin, either in Helmand or in the surrounding provinces including Badghis, Ghor and Uruzgan.
The scope of the UNHCR shelter programme was relatively important in Helmand with a total of 812
shelters built between 2009 and 2011. 364 families received assistance in the Mukhtar Camp alone. Our
research team surveyed 108 households in Helmand, of with 52 per cent were UNHCR beneficiaries.
Moreover, 77 per cent of those households surveyed were IDPs, 10 per cent were non-refugee
returnees, 9 per cent never migrated and 4 per cent were official refugee returnees.
Table A.31: Provincial Sample, Helmand

UNHCR Non-
Total
% 1.79 5.77 3.70
% 5.36 15.38 10.19
IDPs 49 34 83
% 87.50 65.38 76.85
No Mobility 3 7 10
% 5.36 13.46 9.26
Total 56 52 108
% 100.00 100.00 100.00
DESIGN
Because nearly all beneficiaries surveyed in Helmand were IDPs, 88 per cent of those shelters in the
province were one-room shelters. The research team noticed that all shelters visited were in a good
state, with most having been decorated and painted indicating that the appropriation process was well
under way. Moreover, all shelters had a surrounding wall, which they had constructed themselves.
The majority of beneficiaries, 96 per cent of those surveyed, were satisfied with the quality of the
material they received for their shelters. All the shelters had iron beams, iron doors and iron windows.
The only complaint by beneficiaries was related to the size of the room. In a Pashtun context like
Helmand, one-room shelters are insufficient to accommodate very large families. Indeed the average
269
size of households in Helmand was greater than 10 members, the most of the 15 provinces sampled.
Some beneficiaries wanted to adapt the design of the shelters and to remove the corridor however the
IP did not give permission to do so.
Some beneficiaries also noted a few problems encountered during the construction such as the lack of
water, with 51 per cent of households surveyed mentioning insufficient access to water for construction
as a problem. Where water was an issue, beneficiaries had to pay for the costs of a water tanker and
some were obliged to take loans to do so. Overall, beneficiaries noted that building a one-room shelter
did not represent a lot of work, and women were not involved in the construction of the shelters except
to bring water from the well.
UNHCR’s sub-office in the South adapted to the particular migratory profile of the province by
prioritizing conflict-induced IDPs in the selection of beneficiaries and providing them one-room shelters.
In fact, 88 per cent of the UNHCR beneficiaries surveyed were IDPs. This is a welcome adaptation to the
specific needs of a province where intra-provincial displacement increased suddenly.
The VRF was therefore discarded as a relevant criterion of selection. Instead, the selection was reported
to be based on vulnerability, taking into account poverty, lack of housing and the absence of a male head
of household. With that said, only 18 per cent of beneficiary households surveyed were characterized as
extremely vulnerable due to very low income, while none were female-headed households and 43 per
cent lacked housing 1-month before being assisted. If the flexibility introduced in the process is a good
thing, procedures to select beneficiaries without the VRF should be strengthened to ensure that the
most vulnerable IDPs are selected.
In the case of Camp Mukhtar, beneficiaries obtained an official permission letter from the DoRR to
occupy the land and build their shelter. In fact within our own sample throughout the province, 84 per
cent of beneficiaries did not own the land their shelter was built on.
At the household level, beneficiaries noted that the shelter programme greatly improved their living
conditions. Some beneficiaries were living under tents before receiving assistance and noted that they
would not have had the financial capacities to build the shelters on their own. In fact of those
households surveyed, 43 per cent did not have access to a home 1-month before participating in the
shelter programme. What’s more, just above 50 per cent of those households surveyed considered their
economic situation as better or much better after receiving assistance, while 39 per cent noted no
change. Compared to non-beneficiaries, our multi-dimensional poverty index (MPI) presented in Table
A.32 shows that those households which received shelter from UNHCR are less deprived.
270
Table A.32: Multi-dimensional Poverty Index by Beneficiary Status, in Helmand
% 26.79 73.21 100.00
% 9.62 90.38 100.00
Total 20 88 108
% 18.52 81.48 100.00
The construction of shelters was also an opportunity for these families to live in a way that was more in
line with their cultural norms. It meant notably that women could live in proper houses, within
surrounding walls, as opposed to tents where they were a lot more visible to the rest of the community.
Interestingly, beneficiaries also suggested that the implementation of the shelter programme helped
facilitate relations with the DoRR and governmental representatives, which accepted the presence of the
IDPs and stopped trying to evict them. The shelter programme seemed to have legitimized their
presence.
At the community level, non-beneficiaries noted that the implementation of the shelter programme
attracted more people to come and settle in the area, increasing the pressure on services and natural
resources, especially as there is an overall shortage of drinkable water in the area. Yet, they also noted
that it increased the attention of aid organisations, counterbalancing the additional pressure on natural
resources.
SUSTAINABILITY
All shelters visited were occupied and the team observed clear signs of appropriation of shelters by their
occupants. Most people clearly expressed their desire to stay and live in their shelters. In fact all
respondents to our survey planned to stay in the current location permanently.
One of the main factors that will affect the sustainability of the shelter programme in Helmand will be
the evolution of the security context, whilst the international military forces complete their withdrawal.
PARTNERSHIPS
The IDP camp of Mukhtar received the combined efforts of several organisations, with INTERSOS and
WFP also intervening in the area. INTERSOS and DRC built some wells in the camp, while WFP distributed
food.
271
KANDAHAR – PROVINCIAL OVERVIEW
CONTEXT
Located in the southern region, Kandahar borders the provinces Zabul, Uruzgan and Helmand, as well a
large border crossing to Pakistan. Most of its districts are contested areas challenged by the government
and Taliban, which has led to a great number of heavy military operations and insurgent attacks in recent
years. Due to the general decrease of the security in the southern region since 2008, Kandahar who has a
relatively better security situation, has received many IDPs from neighbouring provinces such as Uruzgan
and Helmand. According to UNHCR data, the top ten provinces of displacement are currently led by the
South with Kandahar having the largest protracted caseload of IDPs in the country- with an estimated
7.021 families and 48.886 individuals54, and 171,089 assisted individual returns55.
The recent military operations56 and the continuing attacks by the Taliban have also resulted in a large
number of intra-provincial conflict-induced IDPs, who only move to a different district or even to a
nearby village57. These military operations and the increase of the insecurity have reduced the
accessibility of government officials and international agencies in the province. Kandahar sub-office for
example does not have direct access to its sites and has implemented its shelter programme primarily
through its IP HAPA in Kandahar city, and three other districts namely Panjwayhi, Zhari and Arghandab.
In 2009 and 2010 these districts became temporarily inaccessible even for HAPA due to military
operations carried out by the ANA and foreign forces. As UNHCR is not able to have direct access, it
primarily relies on remote management through the establishment of so-called ‘liaison officers’. There
are two types of liaison officers: one which is on a UN contract (with UN insurance and security
restrictions) and acts on behalf of the UN and participates in the beneficiary selection process and
shelter implementation, while the other one is on a so-called ‘service contract’ which is contracted by an
Afghan NGO and responsible for shelter implementation in insecure areas where UNHCR cannot have
access through its IPs. However, our field research showed that this appears to be more theoretical as it
is primarily the IP and the community leader (Malik) who are involved in the selection process of the
beneficiaries.
The security situation in Kandahar remains therefore a critical issue as it not only affects the situation of
IDPs and returnees, but also very significantly influences whether the government, UN agencies and
NGOs are able to provide assistance. If the armed conflict spreads in Kandahar, it is expected that the
number of IDPs in Kandahar will increase in the coming year, further leading into exacerbation of their
already vulnerable situation.
UNHCR’s sub-office response to IDPs in Kandahar: UNHCR’s sub-office Kandahar has only been able to
operate in certain areas primarily through its IP HAPA, and started working recently with two other IPs;
OHW and HRDA, in the reintegration site Baba Wali village, in Arghandab district. When in 2010
54
Followed by Helmand, Uruzgan and Zabul, source UNHCR IDP Data Report – May 2012.
55
Assisted Voluntary Repatriation to Afghanistan Return by Province of Destination - 02 March 2002 - 31 August 2012, UNHCR
Branch Office Kabul.
56
Operation Hamkari.
57
Also called ‘battle- affected’ IDP’s.
272
thousands of people got displaced in Panjwayhi, Arghandab and Zhari, the sub-office decided to provide
shelter assistance to 600 IDPs, after conducting an assessment which showed that from the 2300 who
got displaced, 600 people were willing to move back to their previous locations if they were supported
by the UNHCR. In 2011, the sub-office allocated 600 two-room shelters in the districts of Panjwayhi,
Zhari and Arghandabad58. The implementation of 200 shelters was completed on 31 December 2011,
leaving 400 to be implemented throughout 2012. Yet despite this achievement, the overall attention to
IDPs has been limited as UNHCR only assisted those who were able to move back to their original places,
while a large number of IDPs still remain in displacement.
In total, UNHCR built 1,213 shelters between 2009 and 2011. As Table A.33 shows, our provincial sample
of Kandahar is comprised of 155 households, of which slightly less than half, 75, are UNHCR
beneficiaries. Moreover, 22 per cent of our sample are official refugee returnees, while 34 per cent are
non-refugee returnees, 25 per cent are IDPs and 19 per cent never migrated. Among the UNHCR
beneficiaries refugee returnees are the largest group with 40 per cent, closely followed by non-refugee
returnees with 39 per cent.
Table A.33: Provincial Sample, Kandahar

Total
% 40.00 0.00 5.06 21.94
% 38.67 0.00 30.38 34.19
IDP 14 0 25 39
% 18.67 0.00 31.65 25.16
% 2.67 100.00 32.91 18.71
Total 75 1 79 155
% 100.00 100.00 100.00 100.00
DESIGN
Most beneficiaries (82.7%) in Kandahar did not personally choose the size of their shelter. The main
design implemented was the standard two-room shelter (96.0%), while a minority build one-room
shelters (2.7%) or another type of shelter (1.3%). Materials were received on time in most cases, 89 per
cent, and the level of overall satisfaction with the quality of the material was lower at 84 per cent than
the national average, 94 per cent.
The beneficiaries interviewed in Kandahar seemed to be generally satisfied with the design of UNHCR
shelter. The most common complaints reported by beneficiaries included the small size of the rooms and
58
226 families were selected in Arghandab, 200 families in Panjwayhi and 174 families in Zhari, HAPA phase 2 weekly report
shelters 2011.
273
window frames, causing a lack of daylight during the day. Beneficiaries therefore combined the two
windows they received together to have more daylight. As in other provinces, beneficiaries also criticized
the shelter package because of the low quality of the iron doors and window lintels. Beneficiaries
complained that the iron doors were not durable as they only lasted for a couple of months or were
quickly damaged after the first rain falls as the doors did not contain anti-rust painting. Suggestions for
improvements for future programming mentioned by beneficiaries, community leaders and IP’s included
enlarging the rooms of the shelters. In few cases beneficiaries also complained about the lack of a
cooking space.
Since beneficiaries are not allowed any flexibility in the shelter design, IPs prescribed the beneficiaries to
adhere to the original shelter design. However, in most cases, beneficiaries tended to improve and
change their shelters after the official handover, sometimes with the approval and acknowledgement of
the IP. Main adaption techniques used after the official handover included the removal of the corridor to
create additional space, as most beneficiaries thought that the two-room shelter did not provide them
with enough space for their families, as well as they thought the corridor to be useless. However, our
field observations also showed that these beneficiaries had not received further technical assistance or
guidance from the IPs on how to maintain sufficient seismic resistance measures that offer adequate
protection from risks. As IPs only monitor during the construction process, it is recommended that follow
ups are conducted also after the final handover to monitor whether such shelters will resist shocks and
earthquakes.
Furthermore, in most cases, beneficiary’s latrines lacked doors. While most of the beneficiaries our
team visited had completed their shelters, in two cases (Naqilian and Munara) several shelters lacked
latrines while they already had received the 100 USD cash grants during the final handover. This leads to
questions on how the final verification and handover have been conducted, as per UNHCR guidelines
shelters are only qualified for handover after the latrines have been completed.
Compared to other provinces, shelters in Kandahar seemed to be less sophisticated and in some cases
in very bad conditions. Most shelters were 2-room shelters with mud walls, some lacking the mud
plaster. The walls from the inside and the outside were very roughly done and most of the shelters
visited lacked glass windowpanes. Beneficiaries reported that they had not received the windowpanes.
Instead plastic sheets or big pieces of cloths were used. Beneficiaries reported that they could not afford
to buy glass, and used the 100 USD mostly to pay off their incurred debts.
Overall, most beneficiaries had built surrounding walls. Only in a few cases, beneficiaries complained
about not being able to build surroundings walls that provided them with privacy and allowed females to
move freely within their yard. In these cases, they used big pieces of cloths to create a makeshift
surrounding wall.
Overall, 44 per cent of the beneficiaries indicated that they encountered some kind of problem during
the construction of their shelter. Table A.34 shows that common problems faced by beneficiaries during
the building process included: the limited access to water and insufficient money to cover the additional
costs for skilled labour. In one instance the lack of available earth to make mud appeared to be
problematic, as most of the plots in the village were privately owned.
274
Table A.34: Problems during Construction, in Kandahar
UNHCR Beneficiaries
N=33 %
The lack of female staff in the shelter programme, as well as the lack of female participation in the
community, was confirmed in KII with UNHCR staff, DoRR as well as with beneficiaries to be a chronic
problem. However, from our field observations it showed that UNHCR had done little to promote the
participation of women in the implementation of the shelter programme. Women in the community
have a passive role and seem to be primarily involved in the construction process of their own shelter.
Women’s activities included cooking meals for daily labourers, carrying water and in a few cases
preparing the mud for the walls. They were not included in the decision-making process or
implementation process of the shelter assistance programme in their community.
Furthermore, while UNHCR relies upon the Afghan tradition of Ashar for extra-assistance to be provided
to the neediest beneficiaries, based on our field visits and discussion with beneficiaries, we found that
that in all cases the principle of Ashar was rather non-existent and beneficiaries were primarily on their
own. The absence of any support mechanism is problematic especially for the EVIs, who because of their
vulnerable situation may face greater risks (such as indebtedness) than others. This calls into question
the appropriateness and relevance of the idea of relying on the Ashar.
Since UNHCR’s shelter assistance does not have any complaint mechanism, both beneficiaries as well as
non-beneficiaries often have nowhere to go. Especially returnees with VRFs, who were left out of shelter
assistance because they were not present during the day of the selection of beneficiaries, seemed to be
confused about the selection procedure.
275
The BSC appeared to be non-existent; beneficiary selection and identification was carried out primarily
through the Malik. Both beneficiaries as well as the Malik noted that it was primarily the Malik and the
IP who were involved in the selection process. Though the UNHCR sub-office staff explained that the
beneficiary selection process included a representative of the DoRR, UNHCR employee and a
representative from the community, our team found out that almost in all cases the beneficiary selection
and identification of ‘eligible’ households were primarily chosen by the Malik, without the direct
presence of the UNHCR or even the IP. Furthermore, in two cases (Munara and Baba Wali village) it is
doubtful whether the most vulnerable have been targeted, as shelters were allocated based on the
number of existing mosques in the village. In Munara, the total number of shelters was allocated by the
IP among 52 mosques, while in Baba Wali village each selected village (in total 7) received 40 shelters,
which were subsequently distributed by the Malik.
Such procedures of beneficiary selection are problematic because a) it means that the selection is
conducted with little oversight and coordination by the UNHCR, its IPs, or the DoRR and b) it is doubtful
that the criteria of vulnerability were properly taken into account given that a specific set of shelters was
allocated to each mosque rather than based on direct needs assessment conducted by UNHCR or its IP.
Furthermore, the over reliance on the Malik for the beneficiary selection without any checks and
balances by third parties, strengthens people’s negative perception of the Malik as implying nepotism
and favouritism in the distribution of aid in the community. Clearly, without a clear mechanism of
oversight and accountability, this over reliance also opens more opportunities for favouritism, nepotism
and sometimes corruption.
 Selection Criteria
The selection of beneficiaries has been primarily based on the VRF, the ownership of land and whether
households were ‘eligible’, defined by the Maliks as households with low income. In many cases, the
provision of shelter assistance to only a small number of VRF holders resulted in selecting only a small
number of beneficiaries among a big group of returnees (those with and without VRF), leading to feelings
of discrimination among the non-beneficiaries. In several cases, non-beneficiaries expressed their
frustration and anger about the discriminatory selection process. Beneficiaries felt the Malik was
corrupted and gave the priority to those who were closely linked to himself, believing the shelter
selection process to be unfair and not benefitting the poor but rather those who are already rich and
well connected. The survey found that 19 per cent of UNHCR beneficiaries were IDPs, a proportion which
seems too low to properly take into account the profile of Kandahar as the first province of displacement
of conflict-induced IDPs in the country.
Most shelter beneficiaries interviewed owned the land, either individually or collectively with other
relatives (brothers, brother-in- law or uncle). While 58 per cent owned the land their shelter was built on
already before receiving assistance, 25 per cent received their land through LAS. Only in few cases, such
276
as in Naqilian village, some community members had received plots from the Mullah Omar during the
Taliban regime, and in Mazra from a wealthy ruling Malik Haji Din Mohammad.
The vulnerability criteria seemed not to be properly taken into account in the selection of beneficiaries.
While 64 per cent of the beneficiaries can be classified as extremely vulnerable, the remaining 37 per
cent cannot. At the same time, 41 EVI households are not beneficiaries. Nineteen of these even applied
for assistance but were not chosen.
The fact that vulnerability was not used as a selection criterion was also particularly observed in Mazra
village and in the reintegration site Baba Wali village. Here, all visited newly built shelters were not used
as living space, but were rather used as storage rooms or to keep livestock such as cows or chickens.
Furthermore, in both villages beneficiaries seemed to have received the shelters additional to their
already existing houses which were in good state. Furthermore, all visited shelters in Mazra included
more than one shelter, as there were several beneficiaries relatives living in the same compound. In
several cases the shelter rooms were nicely furnished, but locked at the time of visit, showing that it was
primarily used as a guesthouse (mehman khana) instead of an actual living space.
At the household level, our research found that all beneficiaries were unable to raise money to cover
the additional cost of construction of their shelters and to hire daily labourers. Most beneficiaries
therefore resorted to a variety of ways to pay for the costs, including asking loans from multiple relatives
and shopkeepers they knew well. Incurred debts varied considerably; beneficiaries reported to be
indebted for between 70.000 AFS- and 100.000 (1.400 USD- 2.000 USD). Beneficiaries reported that this
placed a big burden on their household as most of the heads of households were unemployed or did not
earn enough to pay off their debts in the short term. In the long term this could prevent the household’s
ability to establish a secure livelihood, and based on the interviews conducted, it appeared rather
unclear whether and how these households would be able to pay off their debts. Accordingly, compared
to other provinces a rather small share of beneficiaries, 54 per cent, indicated that due to the assistance
their economic situation was better or far better.
277
Table A.35: Multi-dimensional Poverty Index by Beneficiary Status, in Kandahar
% 24.00 76.00 100.00
% 0.00 100.00 100.00
% 12.66 87.34 100.00
Total 28 127 155
% 18.06 81.94 100.00
At the community level, beneficiaries themselves emphasized that due to the increasing number of
returnees coming back, the population in their village had increased, further leading to an increase of the
price of local goods such as oil and wheat. Others suggested that the prizes of burnt bricks had also
increased due the implementation of the shelter programme in their village.
In all villages visited, the community consisted primarily of returnees, meaning that all community
members had either experienced international or national migration once in their lives. In all cases the
resettlements of the returnees were not seen as a source of conflict in itself and returnees were thought
not to have caused problems in the community. However, our field team observed that the unsystematic
procedure for the distribution of shelters and primarily the lack of transparency caused much discontent
and sometimes conflicts between the Maliks and the community members.
SUSTAINABILITY
Beneficiaries’ ability to improve their socio-economic condition is highly dependent on the accessibility
of water, electricity, health services, employment and the debt incurred during the shelter
construction. Interestingly, despite the hardship people faced in their daily lives, all beneficiaries were
keen to stay and had no regrets about their decision to return back to Afghanistan. This was primarily
because they were satisfied about the fact that they now owned a house and did not have to live with
other relatives anymore. All but one of the respondents in the province indicated that they plan to stay
in the current community permanently. The one that has plans to move again is a UNHCR beneficiary
and reported not being happy in the current community (Mazra).
A big concern for the sustainability of the shelter programme in terms of sustainable reintegration in
Kandahar province is the lack of access to water for many beneficiary communities. Beneficiary families
in almost all villages visited complained about the lack of safe drinking water. Many water points were
too far and in several cases the limited access to water slowed down the construction process. Access to
electricity and basic health services were also absent in most villages visited.
All of the 75 beneficiaries indicated that they were currently living in the UNHCR shelter. While the
occupancy rates directly observed by our team were satisfying, interviews with the Malik and
278
beneficiaries showed that secondary displacements were primarily caused because of the lack of
employment, as well as the lack of basic services such as water, electricity and health. In several cases
the owner of the shelter and their entire household had moved back to Pakistan to find employment
there. Whenever male household heads are able to find employment in the village, they tend to
continue to be engaged in the same low-paid job they had in Pakistan. These jobs included mostly
masonry and construction, and in few instances farming. In the absence of employment in the village,
most men tended to find work elsewhere, primarily in Kandahar city. While some men commuted back
and forth, others were forced to live in the city, leaving their families in the village, because of the high
transportation costs and low income, and only coming back once in three or four weeks.
PARTNERSHIPS
 Inter-units and inter-agencies Coordination
One particular challenge experienced by the sub-office has been the lack of coordination and
alignment between the protection unit and programme unit. Staff at the sub-office noted that the
recent change in the head of office as well as change in higher management at the central level has led
to a substantial shift towards more attention for coordination between the protection and programme
unit. One potential critical problem arising from the lack of coordination between the units is that there
will be insufficient or a lack of attention to vulnerability and to the identification and integration of EVIs
in the selection process.
The sub-office staff found that the central office did not pay too much attention to the southern office,
and this presented certain challenges for the sub-office in responding to the needs of the IDPs in the
southern areas. The sub-office in Kandahar welcomed the shift from humanitarian to development
programming in shelter assistance, as it believed that returnees need more than shelters to reintegrate.
In terms of coordination with other stakeholder, most coordination mechanisms are based on monthly
or ad-hoc working groups.
 UNHCR IPs
UNHCR has been working solely with the Afghan NGO Humanitarian Action for the People of Afghanistan
(HAPA) in the southern region since 2003, and has always had a good relationship with them. HAPA
works for different UN-agencies such as the WFP and IOM, as well as other donors such as CIDA and
JICA. Whilst the sub-office praised HAPA for its reliability and professionalism, it decided in 2012 to work
with two new IPs, OHW and HRDA. Although we did not receive any complaints about HAPA, it must be
noted that in several cases the behaviour of the HAPA employees seemed sometimes suspicious as in
some cases they were hesitating in showing us more shelters and stated that the shelters were too far
for the research team to visit.
As already mentioned, UNHCR’s monitoring in the southern- area, and implementation in areas which
are not accessible to the IPs, are all dependent on the so-called ‘liaison officers’. These liaison officers
are young men recruited locally who have received a short training and salary from UNHCR and are able
to travel to insecure areas because of their local knowledge, established social network and access to the
279
areas. Liaison officers’ tasks include both implementation and monitoring, such as monitoring the
distribution of assistance and monitoring beneficiaries’ construction process. While the sub-office in
Kandahar seemed to be satisfied with this approach, it should be acknowledged that relying on remote
management without direct supervision also brings a number of important risks and disadvantages, and
could weaken the quality and efficiency of the shelter programme in the long run.
 Relationship with National Authorities
The DoRR representatives in Kandahar were highly critical about the lack of coordination,
communication and information exchange with the UNHCR. The DoRR argued that they are completely
sidelined by the UNHCR during the beneficiary selection, implementation and distribution of materials.
The DoRR argued that they had offered help and advice in several instances to select certain
beneficiaries, but that UNHCR appeared to be non-responsive and acted separately with only informing
the DoRR after the selection and implementation of the shelters. Not involving the DoRR in the selection
and implementation programme, was confirmed by the sub-office, which argued that the DoRR were
only side-lined because of their incapacity and incapability to do anything. This general discontent with
the government and their lack of capacity was also felt by the returnees interviewed, who expressed
their frustration with the government and political leaders for failing to provide assistance both to the
returnees themselves, and to the community as a whole.
280
PAKTYA – PROVINCIAL OVERVIEW
CONTEXT
Located in the Southeast, Paktya is highly volatile and insecure, especially in its rural and mountainous
districts where governmental forces have little access. Paktya received more than 97,000 returnees over
the past ten years, while the province saw only a very small number of internal displacements. Given the
proximity with Nangarhar and Pakistan, the province is characterized by large movements of temporary
migration, especially during the winter.
The scope of the UNHCR shelter programme was important in Paktya, with a total of 1,372 shelters built
between 2009 and 2011. Our research team surveyed 240 households in Paktya, of which 51 per cent
were UNHCR beneficiaries. Moreover, 60 per cent of those households surveyed were official refugee
returnees while 26 per cent were non-official returnees, 2 per cent were IDPs and 12 per cent never
migrated.
Table A.36: Provincial Sample, Paktya

UNHCR Non-
Total

% 73.98 44.44 59.58
% 18.70 34.19 26.25
IDPs 1 4 5
% 0.81 3.42 2.08
No Mobility 8 21 29
% 6.50 17.95 12.08
Total 123 117 240
% 100-00 100.00 100.00
DESIGN
Nearly all shelters visited in Paktya, 94 per cent, were standard two-room shelters with iron beams,
windows and doors. Beneficiaries for the most part had no complaints about the material received, with
97 per cent noting the materials were of good quality. All beneficiary households had built their own
surrounding walls.
Most beneficiaries noted that the design and the size of the shelters were not adapted to the needs of
their families. A lot of beneficiaries changed the design, removed the corridor and enlarged the rooms
with the permission of the IP who monitored the process. Some households added extra rooms with
their own money.
281
The main problem that beneficiaries encountered during the construction was the lack of water, with 53
per cent of those surveyed mentioning the issue. Most had to have the water brought by a tanker, which
cost a significant amount of money.
Women helped during the construction of the shelter. Their main task was to bring the water and to
make the clay for the mud bricks.
Representatives of the IP and community leaders conducted the selection of beneficiaries. The main
criteria of selection were a) VRF, b) land ownership, and c) vulnerability. Of those beneficiary households
surveyed 96 per cent had a VRF, 94 per cent owned the land their shelter was built on beforehand, and
42 per cent were characterized as extremely vulnerable. Non-beneficiaries perceived the selection
process as fair and confirmed that it had targeted the most vulnerable in the community. There were no
reports of fraud or misallocation in the province.
At the household level, the shelter programme significantly improved the living conditions of beneficiary
households as many lived under tents or had to share a house with their relatives before receiving
shelter assistance. Of those households surveyed, 30 per cent did not own a house 1 month before
receiving assistance.
Many beneficiaries noted that they had to take on significant loans to build their shelters with 41 per
cent of those surveyed stating debt had increased following participation. What’s more, only 37 per cent
of beneficiaries considered their economic situation better after participation, while 38 per cent thought
it was the same and 20 per cent believed it to be worse. This suggests that the economic impact of the
programme on households was less positive in Paktya than in other surveyed provinces. Compared to
non-beneficiaries, our multi-dimensional poverty index (MPI) presented in Table A.37 shows that those
households which received shelter from UNHCR are less deprived.
Table A.37: Multi-dimensional Poverty Index by Beneficiary Status, in Paktya

% 24.00 76.00 100.00
% 0.00 100.00 100.00
% 12.66 87.34 100.00
Total 28 127 155
% 18.06 81.94 100.00
282
At the community level, community leaders and members of communities mentioned that the
programme had had a positive effect on intra-community relationships, as people were no longer forced
to share houses. Moreover, nearly all, 97 per cent, believed the programme had a positive or very
positive impact on the community as a whole. Some however complained about the pressure that the
programme exerted over natural resources and services, especially water.
SUSTAINABILITY
The research team noted a certain number of empty shelters, especially in the village of Levan, where
half were empty. However these shelters were visited at the beginning of December and it is possible
that their tenants had already left to Jalalabad for their seasonal winter migration.
Most community leaders noted that their villages suffered from a lack of basic services, including the
absence of clinics, the lack of paved road and the lack of water for consumption and agricultural
purposes. The general economic situation of these villages and this absence of basic services represent a
threat for the sustainability of the programme in the area.
All of the beneficiary households noted that they had no intention to move from their place of origin.
Still the sustainability of the programme in Paktya will depend heavily on the evolution of the security
context in the province in the coming years.
PARTNERSHIP
Only a few organizations (including IRC) are active in Paktya, as large parts of the province are too
insecure. The main institution complementing UNHCR activities has been the DRRD through CDCs.
UNHCR provided hygiene training for women in areas where the shelter programme was implemented.
283
HIRAT – PROVINCIAL OVERVIEW
CONTEXT
The migratory context in Hirat is a quite specific one for Afghanistan, with a wide variety of migratory
profiles including voluntary returnees from neighbouring Iran, undocumented returnees, deportees, but
also returnee-IDPs from Iran and Pakistan unable to return to their province of origin (Faryab, Ghor,
Baghdis), conflict-induced and economic IDPs. This very mixed and complex migratory context poses a
direct challenge for identification of eligible people for the shelter programme in the field and allocation
of assistance, especially since the rate of return from Iran is steadily decreasing, whereas the number of
IDPs coming to the province is increasing. The usual categories for identification of beneficiaries
(returnee, VRF holder, a small but usually minimal proportion of IDPs) therefore do not seem to be
relevant in this type of context, which directly implies serious considerations about how to reach
vulnerable households in need of shelter.
The volume of IDP presence in the province gives the issue of IDPs considerable political weight.
Settlements inside Hirat city and in surrounding IDP camps of Maslaq, Minaret and Shahidayee, as well
as informal settlements are a ground for political tensions, with a clear reluctance of authorities to
address the issue59.
One of the issues in the West was the strong focus of humanitarian assistance on Hirat province, with a
small proportion of programmes implemented in Farah province and only a few in Baghdis or Ghor. The
majority of the provinces in the Western region are therefore left out of most assistance and
reintegration programmes for migrants (this is the case both for UNHCR and NRC).
UNHCR supported the construction of 1,298 shelters in the province between 2009 and 2011. As Table
A.38 shows, our provincial sample of Hirat comprises 200 households, of which more than half (101) are
UNHCR beneficiaries. Moreover, some 49 per cent of our sample are official refugee returnees, while 15
per cent are non-refugee returnees, 29 per cent are IDPs and 8 per cent never migrated. Considering the
large amount of IDPs in the province, it appears that our sample is not representative. They represent 21
per cent of surveyed UNHCR beneficiaries.
59
The governor of Hirat issued a letter stating there were officially no registered IDPs in Hirat province, and they were therefore
not entitled to any form of assistance.
284
Table A.38: Provincial Sample, Hirat
UNHCR Non-
Total
% 67.33 29.29 48.50
% 7.92 21.21 14.50
IDP 21 37 58
% 20.79 37.37 29.00
No Mobility 4 12 16
% 3.96 12.12 8.00
Total 101 99 200
% 100.00 100.00 100.00
DESIGN
Until 2012, there was a high degree of variations in the shelter package, both in the types of models
used, the standards and the material provided, the latter varying on a yearly basis based on field
observations and recommendations from the IPs, with wooden planks for roofing replaced by fire brick
for instance. In 2009 two-room, one-room (for IDPs and returnee households with less family members)
and repair-kits were implemented. Throughout 2009-2011, two types of two-room shelters were
constructed: type C (including a dome roof, adapted to the local cultural norms) and type A (flat roof). In
2012 the package was uniform with a standard type C shelter. Among the surveyed beneficiaries 84 per
cent built a two-room shelter and the rest, 16 per cent, one room. Nobody got to choose the size of their
shelter.
Differences with other provinces were also noticed in the material provided: two-room shelters were
handed out three iron doors (as opposed to one iron door and two inside wooden doors in other
provinces), and the latrines were provided with ceramic toilets, which had not been observed yet in any
other province.
Both the UNHCR Field Associate and IPs mentioned that the previous flexibility in the standard of shelter
(one-room, two-rooms, repair kit) provided was a challenge in the field, in cases where decisions had to
be made in terms of allocation of a specific type of shelter to the beneficiaries, creating
misunderstanding as to the reasons why which type of shelter was allocated to whom. This was a driver
of conflict inside the communities as well as between beneficiaries and IPs.
Flexibility in the design was noticed in the field. Some changes had been validated by field engineers
during construction. Main changes included opening the corridor, used as an open space between two
rooms (systematically implemented in Kahdistan), or using it as a kitchen (Shogofan), adaptation to the
land plots (Shogofan and Kahdistan), and enlargements, the latter being done after handover. In itself,
this is a positive sign of the way beneficiaries consider their shelters as a living space and adapt it to their
needs. Flexibility was also allowed in adapting the shelter to the size and shape of the land plots, with
285
doors installed on the side of the shelters as an extra entrance. Surrounding walls were also
systematically constructed around plots where one to several shelters was built for related families.
The overall state of shelters was relatively good. In most shelters visited, the latrines were built and
used. There was general satisfaction about them, which can also be explained by the good quality of the
material used (ceramic for standing toilets).
A recurrent complaint on behalf of all stakeholders (beneficiaries, IPs and UNHCR) was the low quality of
the iron doors and window frames provided with the shelter, subject to rust and deformation. More
generally though, the interviewed beneficiaries in contrast indicated that they received their materials
on time and that they were of good quality in 94 per cent and 96 per cent of the cases respectively.
Problems during construction: Almost two thirds, 64 per cent, of the surveyed beneficiaries indicated
that they encountered some kind of problem during the construction of their shelter. Table A.39 shows
the different types of problems the beneficiaries encountered, the main problems being having enough
financial capital and insufficient access to water.
Table A.39: Problems during Construction, in Hirat

UNHCR Beneficiaries
N=65 %
Flexibility of the cash grant: There was a relatively high degree of variability as well in the amount of
cash grant provided to the beneficiaries during handover, which varied every year according to
fluctuations in the market prices for labour. The cash grants were sensibly higher than those provided in
other provinces (between 100-300), which can be explained both by the high cost of labour, and the
unavailability of clay in the area which requires the use of cement and burned bricks in most cases.
UNHCR staff reported that assessments of labour costs were conducted every year by UNHCR, though no
286
specific figure was provided for the annual cash grants handed out to the beneficiaries. The figures
mentioned in the field and by the IPs were between $110 for a two-room shelter to $250.
Risk assessment: As in other provinces, there appears to be no prior assessment of flood risks. In two
locations, inhabitants mentioned this as a major concern. In Shogofan, beneficiaries were worried
because the canal they built through the cash-for-work programme of UNHCR was not paved, which
increased the risks of over-flooding. In Jebrail, inhabitants and shura leaders mentioned an entire
neighbourhood, which included shelters (2008 and one 2009 shelter visited) and was built on a flood
prone area, due to land speculation and influence of the land mafia. This calls for a better assessment
beforehand.
Monitoring: The UNHCR sub-office in Hirat has no monitoring IP and conducts the monitoring of
construction themselves, with regular visits to the field in all accessible areas. There have been some
recent security issues in Gozarah, but access in general does not seem to be an issue. This is a good sign
that monitoring by UNHCR staff is possible in secure areas, which raises question about the regularity of
the recourse to external monitoring IPs across provinces (especially in Kabul and Parwan where
monitoring by UNHCR staff appeared elusive). The survey data shows that 12 per cent of the UNHCR
beneficiaries in Hirat indicated that there was no monitoring of their shelter by the organization that
assisted them.
The wide variety of migratory profiles in the province has a strong impact on beneficiary selection,
complicating identification of eligible people. Furthermore, this puts into question both the relevance
of the proportion of returnees included and the focus on the VRF criteria for selection, given the
comparatively high number of IDPs in the province. The majority of the beneficiaries in our sample, 67
per cent, were recognized refugee returnees. Comparing the percentage of EVI households of this group
to IDPs shows that 45 per cent of refugee returnees can be classified as extremely vulnerable compared
to 50 per cent of IDPS. Yet, IDPs are much less chosen for assistance. While three quarters of the
vulnerable refugee returnee households did receive assistance, only 38 per cent of the IDPs did.
The assistance therefore targeted people, who were not necessarily the most vulnerable, and left aside a
wide range of people in need of shelters who are under the mandate of UNHCR (conflict-induced IDPs).
However, the difficulty to identify IDPs and the high political stakes behind the assistance they are
allocated places considerable strains on their selection (high reluctance of local authorities to deal with
IDPs) and the proportion of IDPs selected in Hirat is higher than in most of the rest of the country.
This calls into questions the existing categories and creates several challenges on the field with:
 High complexity in identification of beneficiaries, with repercussions in the tensions inside

the community and complication of the work of IPs potentially accused of corruption or
favouritism.
 High rates of fraud and VRF trafficking in the province.
287
A side effect of the high focus on VRFs is the high rates of fraud and VRF trafficking in order to receive
assistance observed on the field. In both locations visited in Shogofan ‘cluster’ (Shogofan 1 and Esaq
Abad), it was confirmed that beneficiaries were not the original holders of the VRFs, which they
reportedly had purchased in Jebrail. Prices of VRF on this black market oscillated between 7,000 and
10,000 AFS (135 to 200 USD). In other cases (Shogofan), the beneficiaries rented out their shelters to
other families and continued living in some of the other IDP settlements surrounding Hirat. The existence
of this type of parallel “shelter-market” is a worrying dynamic. The problem has been acknowledged by
UNHCR, which decided in 2012 to place a higher focus on vulnerability for the selection in a welcome
move to tackle this issue. Recognizing the importance of the phenomena, NRC also changed its approach
by prioritizing vulnerability over documentation and accepting any kind of proof of existence abroad.
More than in other provinces, the discussions with IPs, local representative and UNHCR mentioned the
regular inclusion of 5 to 10 per cent of local community members as EVI, through the involvement of
the Protection unit. This was deemed as necessary to ease tensions inside the community, notably given
the problems raised in the identification of beneficiaries. The quantitative data does however show that
of the four beneficiary households without a migration experience only one can actually be classified as
extremely vulnerable. This assistance therefore also does not seem to target those most in need.
It appeared that the degree of involvement of local authorities (maliks, arbab, shura) in the selection
process evolved over time. In 2009-2010, it was allegedly reduced in order to avoid preferential
selection. Though their presence in the BSC was mandatory in order to avoid tensions in the community,
their decisional role was minimal to non-existent at that time. In one case (Jebrail), the shura was not
aware of the existence of the programme in 2009, though the IP (CRDSA) denied not having involved
them. The IP emphasized the fact that they tried to rely more on community members, including
women, than the official authorities, in order to access more eligible people and avoid nepotism. This
seems to have changed recently, especially in the reintegration sites, where local shuras play a significant
role throughout the implementation process. There were strong allegations of corruption on behalf of
communities and beneficiaries in the reintegration site of Shogofan, with shelters reportedly bought
from the Malik/IP. Several non-beneficiaries denounced this practice and one beneficiary confirmed that
he had paid 10,000 to the Malik and the IP to be selected in the programme. If corruption is effectively
widespread, cash-based assistance might be tricky to implement.
There appears to be a strong involvement of both, UNHCR field team as well as DoRR, in the selection
process. IPs mentioned the systematic presence of both of them in the BSC was particularly positive for
strengthening the BSC in the eyes of local authorities and facilitating the selection process. The DoRR in
particular was seen as a necessary governmental caution.
Another issue mentioned in selection was the focus on specific villages identified by UNHCR
beforehand. This was repeatedly mentioned by all three IPs as a concern, as neighbouring villages
included eligible people, which could not be reached. The method of selection of areas for shelter
assistance, putting the stress on numbers rather than vulnerability is therefore problematic, with less
vulnerable people being selected to complete the quota of shelters allocated, whereas vulnerable
households in neighbouring areas are completely left out.
288
NRC ‘s Selection Process
Recognizing the difficulty in identifying eligible beneficiaries in Hirat province, NRC decided to
tighten its selection process significantly for the 2012 selection. The following measures were
introduced:
 NRC is not restricted to some implementation locations beforehand.
 The focus of the selection was put on vulnerability rather than migratory status.
 While in the past the programme focused on returnees, this year 80 per cent of the
beneficiaries were IDPs, while 20 per cent were returnees.
 Rather than the VRF only, NRC accepts any sort of documents showing that the
individual/family has lived abroad.
 In case a fake VRF is identified, the household is immediately excluded from the selection.
 When people declare renting a house, necessity to present appropriate documentation to
be selected.
 Multiplication of cross-checks of information through:
o Closer house to house visits
o Double checking
o Inclusion of NRC’s M&E staff in the selection process
o Multiple visits
o Delayed decision in case of doubts
o Early morning visits to the field (7am).
Overall this selection process led to the exclusion of approximately 75 per cent of potential
beneficiaries. The process was also time-consuming and took several months. At the end, NRC was
satisfied with this process, which conducted to a much higher occupancy rate at the handover
period than previous years.
 At the household level
The socio-economic impact of the programme was reportedly positive, as 64 per cent of the beneficiaries
indicated that their economic situation after receiving the assistance was better or far better than
before. Beneficiaries explicitly mentioned the improvement in housing (solid housing versus tents, cf.
FGD women beneficiaries Kahdistan) as a major positive effect on their standard of living as a mean to
protect them from the natural environment. Around 40 per cent of the beneficiaries lived in tents or
other temporary shelter before they received the assistance. Therefore, as in other provinces, access to
property and housing was widely praised – a notable improvement for returnee-IDPs who mentioned
having lost their property in their region of origin. Interestingly, nomad tribes mentioned the change of
social status housing provided them with: “when we arrived, we had nothing, of course our life became
better. At the time our husbands were begging for work, we were just people living in tents. People said
“you don’t have anything, not even a house!” so now it’s better.” (Sibid, FGD women beneficiaries,
Kahdistan). In line with this positive reception of the programme is the fact that no beneficiary currently
has plans to move from the community.
289
However, though the extent of the phenomena could not been assessed, flaws in selection leading to
frauds and the existence of a VRF black market represented a considerable additional investment to get
a shelter, which could have been avoided and is another indicator of a need to question existing
categories (cf. above). This is also a sign that shelters are a response to a genuine need as people are
willing to invest substantive sums of money in order to access housing.
Table A.40: Multi-dimensional Poverty Index by Beneficiary Status, in Hirat


% 25.74 74.26 100.00
% 11.11 88.89 100.00
Total 37 163 200
% 18.50 81.50 100.00
 At the community level
It appears that the implementation of the shelter programme – especially when combined to other
programmes of assistance like in Shogofan – in specific areas triggers further displacement with
migration to the areas targeted for assistance. In new settlements, the programme supports the general
development of the village and increases the availability of housing in areas, which received large influx
of population in short periods of time.
Short-term effects:
 Short term employment boost, though this depends on the level of wealth of the
community and the capacity of beneficiaries to hire labour: this was the case in
Shogofan, but was not in Kahdistan, for instance, where beneficiaries constructed their
shelters themselves or with the assistance of relatives.
 Impediment to work and income: very-low income beneficiaries mentioned not being
able to hire extra labour, and that construction takes time on their work hours and has a
negative impact on their income. This further explains delays in construction (5-6
months for construction of a 2-room shelter).
Impact on local prices:
 Rise in the price of land, with speculation on land strongly linked to allocation of
assistance in a specific area.
 Differentiated effects on local prices:
- Price of material drops because of an influx of free material (Shogofan).
290
- In Kahdistan, the arbab mentioned that the price of beams had increased
because of a rise in demand for similar beams to the ones distributed by UNHCR.
This is also a sign that the quality of the material provided by UNHCR has
repercussions on the standards required by the community. One of the effects of
the shelter programme is therefore an overall potential improvement in the
quality of housing in the community.
SUSTAINABILITY
The rates of occupancy were high in all locations visited. Of the 101 beneficiaries four indicated that they
were not currently living in the UNHCR shelter. In some cases, the lack of livelihood had led to the
abandonment or selling of the shelter (Esaq Abad). A more widespread way to cope with the issue of
livelihood was the migration of the head of household in order to find better employment opportunities,
while the rest of the family lived inside the shelter. This indicates that shelter remains a strong
component of family life. There was therefore little secondary displacement, and beneficiaries insisted
they had no intention to migrate again, as migration was described as a costly and tiresome process:
“with a house that has been built now, why would we go anywhere? It’s so much effort and so difficult to
move! We don’t want to do that again.” (Said Bibi, FGD women beneficiaries Kahdistan).
Threats to sustainability: In the long run, the absence of access to safe water, the lack of livelihood and
the absence of basic services, especially governmental schools and health facilities, which were lacking in
most of the sites visited, might be an impediment to the sustainability of the programme. The case of
Saadat, the LAS where shelters had been implemented in 2007 and that is now mostly empty, is a sign of
the negative consequences of not taking into account basic needs in parallel to shelter. In reintegration
sites, the short-term nature of efforts implemented by UNHCR keeps assistance from addressing the
problems of beneficiaries in the long term. In Shogofan for example, most of the complementary
income-generating programmes implemented by UNHCR were short-term initiatives and could not
address the deeper issue of absence of employment and livelihood. Most of these initiatives were cash-
for-work programmes (road rehabilitation, canal), which will stop at the end of 2012.
There was no regular pattern for complementary assistance: no WASH programme has been
implemented by UNHCR since 2008, and cash for work has only been done in parallel to shelter in a few
cases.
PARTNERSHIPS
As noticed in other provinces, the partnership strategy of UNHCR is strongly lacking linkages with other
organisations, which could provide basic services (clinics and schools which are recurrently mentioned as
a great need by community members), WASH components or employment programmes. In Hirat, the
UNHCR has no partnerships with NGOs or other agencies.
The Hirat sub-office privileged one-year partnerships with the IPs implementing the shelter programme.
The method of IP selection with a regular rotation every year is problematic in terms of capacity building
and follow-up. A UNHCR Field Assistant provided no clear explanation about the selection process and
291
the reason for rotation, casting doubts about the reliability of the IP given contradictory information
gathered on the field. UNHCR staff suggested though that the rotation of IPs was a method to limit the
risk of corruption of IPs. As mentioned above, the fact that the Hirat sub-office has no monitoring IP and
is able to conduct monitoring itself raises the question of why a sub-office like Kabul needs to rely on
monitoring IPs.
However, the Hirat sub-office seems to cultivate strong links with governmental authorities (as
opposed to Nangarhar for instance), particularly with DoRR representatives reportedly present
throughout the implementation process, from selection to monitoring, as per the guidelines. The
Directorate of Economy was also responsible and present for monitoring of all development projects. It
must be underlined that governmental authorities could not be met in Hirat due to the presence of an
Iranian BIAFA delegation during our stay in the province and conflicting schedules. Despite these
observations, HELP mentioned the difficulty to work with the current DoRR as well as the strong political
stakes behind the issue of IDPs and the reluctance of local authorities to allow them to receive
assistance.
292

UNHCR Shelter Source

Uploaded by

Copyright:

Available Formats

UNHCR Shelter Source

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

UNHCR Shelter Source

Uploaded by

Copyright:

Available Formats

Evalua on of the UNHCR Shelter Assistance

Samuel Hall. (www.samuelhall.org) is a research and consulting company with headquarters in

Cover Picture © Lucile Martin

5. Socio-Economic Impact of the Shelter Assistance Programme..........................................113

6. Partnerships with Other Stakeholders ..........................................................................140

7. Conclusion: What is the impact of UNHCR’s shelter assistance on the sustainable

8. Recommendations: Reinforcing SAP Guiding Principles ................................................172

Box 1: Access to Water…………………………………………………………………………………………………………………….… . 59

Figure 1: Donors Earmarked Contributions 2009-1011 (Million USD) ..................................................... 151

Shelter beneficiaries fare better than non-beneficiaries

 Highest rates of poverty among non-shelter beneficiaries: 86 per cent of non-shelter

IDPs fare worse than returning refugees in the reintegration process

Positive impact on intra-community relations

2. UNHCR’S POSITION AS A LEADER ON SHELTER INTERVENTION –

3. WEAKNESSES OF THE SHELTER ASSISTANCE PROGRAMME

Financial burden of SAP on beneficiaries

Insufficient support to beneficiaries and EVIs

Inadequate risk mitigation and prevention mechanisms

1. Better adapt to the current migratory trends of the country;

2013 SAP GUIDELINES – 22 PRINCIPLES

1. Impact and Needs Assessments

3. Women’s direct participation

8. Preservation of cultural and regional preferences

9. Contribution to local economies

10. Partnership Strategy

11. Monitoring Framework

1.1 SHELTER FOR DISPLACED POPULATIONS IN AFGHANISTAN

The notion of “sustainable return” is a long-term, contextual understanding of return incorporating

On the operational level, this means:

1.4 SHELTER ASSISTANCE IN A CHANGING HUMANITARIAN CONTEXT:

1.4.1 Rationale of UNHCR’s Shelter Assistance Programme

1.4.2 Guiding Principles of SAP

1. Community based approach

2. Women’s direct participation

6. Preservation of cultural and regional preferences

7. Contribution to local economies

8. Involvement of local authorities

1.4.3 Changes in the Programme

1.5 REPORT OUTLINE

2.1 QUANTITATIVE DATA COLLECTION AND ANALYSIS

2.1.1 Household Survey

 2,035 UNHCR Beneficiaries

 2,325 Refugee Returnees

UNHCR Non-UNHCR Non-

Refugee Returnees 1 355 243 727 2 325

UNHCR Non-UNHCR Non- Total Total

Sari Pul 56 - 44 100

Southeast Paktya 123 - 117 240 240

West Hirat 101 - 99 200 200

Total 2 035 463 1 990 4 448 4 488

Table 3: Household Survey Sampling by Type of Location

UNHCR Non-UNHCR Non-

Urban 334 87 413 834

Table 4: Community Survey Sampling

Region Province N Total Region

Central Highland Bamyan 4 4

Methodologically we follow a step-by-step process, first analysing household deprivation by individual

The four dimensions used in our analysis include:

Dimension 3: Health and Nutrition

Expenditure per capita is below