Report No. 29694-AF Afghanistan Poverty, Vulnerability and Social Protection: An Initial Assessment March 7, 2005 Human Development Unit South Asia Region Document of the World Bank ACKNOWLEDGEMENTS This report was prepared by a team ledby Christine Allison (SASHD). Other core members o f the team were Renos Vakis (HDNSP) and Abi Masefield (consultant). Various members o f the Afghanistan Country team, most notably William Byrd, Paul Dorosh, Asger Christiansen, and Asta Olesen provided invaluable support and guidance to the team. The report has benefited from the parallel preparation of "State Building, Sustaining Growth and Reducing Poverty", the World Bank's economic report on Afghanistan (2004), and the Country Gender Assessment, under preparation (2004105). The report was prepared under the overall supervision of Alastair McKechnie (Country Director, SACAM).Gerhude Cooper helped with report production. The report draws heavily on the 2003 National Risk and Vulnerability Assessment, which was ledby the Ministryo f Rural Rehabilitation and Development (MRRD), with extensive support from the World Food Program, FA0 and others. Special thanks to Minister Haneef Atmar and Andrew Pinney (MRRD), and Susana Rico and Scott Ronchini (WFP). Inthe Bank, Kalpana Mehra undertook muchofthe data work. Valuable comments and advice were received from the Peer Reviewers, Kathy Lindert (LCSHD) and Johannes Hoogeveen (HDNSP), the poverty measurement advisor, Peter Lanjouw (DECRG), and a quality enhancement review panel, chaired by Margaret Grosh (HDNSP). The report also benefited from comments provided at the review o f the first draft (June 2004), and subsequent discussions in early 2005. The views and insights of a variety o f stakeholders (govemment, NGOs, the donor community, researchers and advisor) are gratefully acknowledged, and add to the completeness of the report. CONTENTS PageN o. EXECUTIVE SUMMARY .............................................................................................................................. i I. POVERTYANDVULNERABILITY:AHISTORICALPERSPECTIVE............................................ 1 Origins of Poverty and Vulnerability: A HistoricalPerspective .............................................................. 1 RiskManagement Strategies: HowFamilies Survived ............................................................................ 2 Recent Developments: Improving Livelihoods and Opportunities .......................................................... 5 I1. RURALLIVELIHOODS......................................................................................................................... 7 Understanding RuralLivelihoods ............................................................................................................ 7 The Diversity ofLivelihoods inRuralAfghanistan ................................................................................. 9 Insights from the NRVA: Occupational Structure ................................................................................. 15 The Structure o fWages.......................................................................................................................... 17 Migration and Remittances .................................................................................................................... 19 Gender Roles andRuralLivelihoods ..................................................................................................... 21 Determinants o f RuralPortfolios ........................................................................................................... 23 I11. A PROFILE OFPOVERTYAND VULNERABILITYIN RURALAFGHANISTAN .......................................................................................................... 27 Poverty ................................................................................................................................................... 28 The Correlates o fPoverty ...................................................................................................................... 29 Nutrition and Malnutrition ....................................................................................... Determinants o f RuralPoverty............................................................................................................... 35 Are Things Getting Better? .................................................................................................................... 35 The Impact o f Shocks on Welfare .......................................................................................................... 38 Shocks and Vulnerabilities ..................................................................................................................... 37 The Performance o f Some FormalSafety Nets...................................................................................... RiskManagementInstruments: Coping with Shock .............................................................................. 39 41 IV. URBANPOVERTY AND VULNERABILITY.................................................................................... 47 Introduction............................................................................................................................................ 47 Background............................................................................................................................................ 48 Dimensions o f UrbanPoverty and Vulnerability ................................................................................... 49 UrbanAsset Vulnerability ...................................................................................................................... 52 UrbanRisks and Shocks......................................................................................................................... 53 Who are the Vulnerable?........................................................................................................................ 53 Coping with Risks and Shocks............................................................................................................... 59 Variants, Sources and Determinants o fUrbanVulnerability ................................................................. 61 V. POLICY DIRECTIONS: TOWARDS A NATIONAL STRATEGY FOR POVERTY REDUCTION ..................................................................................................................... 63 Introduction............................................................................................................................................ 63 The Essential Characteristics o fthe Poor, Re-visited............................................................................. 63 Economic Growth: The EssentialPathfor Poverty Reduction .............................................................. 65 Equipping the Poor with Assets and Capacities for Growth: Public Investment Priorities ............................................................................................................ 67 A Safety Net for the Poorest?................................................................................................................. 72 The Social ProtectionProgram............................................................................................................... 68 Conclusion ............................................................................................................................................. 76 References ................................................................................................................................................. 91 Tables Table 2.1 Classificationo f the Main Components o f Livelihood Strategies inRural Afghanistan ...........8 Table 2.2 Classification o fHouseholds Engaged inOpium Cultivation.................................................. 11 Table 2.3 Livestock PopulationPer Family............................................................................................. 14 Table 2.4 Occupational Structure byFood Consumption Quintiles......................................................... 16 Table 2.5 Employment by Consumption Quintile ................................................................................... 16 Table 2.6 Comparative Estimates o fWage Rates for Various Types o f Informal Employment....................................................................................................... 18 Table 2.7 Reported Wage Rates ($/day) Disaggregated by Region for Various Activities ............................................................................................................. 18 Table 2.8 Daily Wages for men, women and children ($/day) ................................................................ 19 Table 2.9 Migration and Poverty, Afghanistan 2003 ............................................................................... 20 Table 2.10 24 Table 2.11 Farm Constraints, Rural Afghanistan 2003....... .................................................................. Major Sources ofIncome and Poverty.............. .................................................................. 25 Table 2.12 UnderstandingRural Employment Choices, (Probit Estimations) ........................................... 26 Table 3.1 Rural Poverty by Region.......................................................................................................... 27 Table 3.2 Various Household Characteristics by Consumption Quintiles ............................................... 30 Table 3.3 Access to Basic Infrastructure by Food Consumption Quintiles, Rural Afghanistan 2003 .................................................................................................... 32 Table 3.4 Access to Basic Facilities and Services, RuralAfghanistan 2003 ........................................... 32 Table 3.5 Net Enrollment, Rural Afghanistan 2003................................................................................. 32 Table 3.6 Health Status and Facility Access by Consumption Quintiles, Rural Afghanistan 2003 .................................................................................................... 33 Table 3.7 Poverty Correlates: Log o f Food Consumptionper Capita...................................................... 36 Table 3.8 Covariate Shocks by Food Consumption Quintiles ................................................................. 38 Table 3.9 Idiosyncratic Shocks by FoodConsumption Quintiles ............................................................ 38 Table 3.10 Coping Mechanisms and Recovery from Selected Covariate Shocks, by Food Consumption Quintiles ............................................................................................ 40 Table 3.11 Coping Mechanisms and Recovery from Selected Idiosyncratic Shocks, by FoodConsumption Quintiles ............................................................................................ 40 Table 3.12 Program Participationand Shocks Incidence........................................................................... 41 Table 3.13 Program Incidence 200212003, Rural Afghanistan .................................................................. 42 Table 3.14 44 Program Targeting Indicators under Various Eligibility Criteria (%)...................................... Program Participation and Distributionby Poverty Groups (%).............................................. Table 3.15 45 Table 4.1 Estimated UrbanPopulation inMajor Towns o fAfghanistan ................................................. 48 Table 4.2 UnderstandingDistinctive Aspects o fUrbanPoverty and Vulnerability inAfghanistan............................................................................................. 49 Table 4.3 Selected Indicators for Poverty and Vulnerability inUrban and Rural Afghanistan ............................................................................................................. 50 Table 4.4 Urban Service DeliveryDeficiencies ........ .................................................................... 51 Table 4.5 Kabul: IlliteracyRates by Age Range and r.................................................................. 52 Table 4.6 Factors affecting Kabul Households' Susceptibility to Threats Associated with Winter..................................................................................................... 56 Table 4.7 Results of Anthropometric Surveys inKabul(1995-2003) ...................................................... 57 Table 4.8 UrbanCase Studies to highlight the Coping Economy o f UrbanAfghanistan ............................................................................................................ 60 Table 4.9 Variants, Sources and Determinants o fUrbanVulnerability in Afghanistan.......... ...................................................................................................... 62 Table 5.1 The Essential Characteristics o fThree Groups o fRuralHouseholds...................................... 64 Table 5.2 Three Essential Successful National Programs and Poverty Reduction.................................. 69 Table 5.3 SelectedCharacteristics among "Vulnerable" Households...................................................... 73 Table AI1.1 Calculation o fFood Consumption Patterns for Rural Poverty Lines....................................... 84 Table AII.2 Rural Poverty Line Sensitivity ................................................................................................. 85 Table AIII.1 Calorie Gap Estimates ............................................................................................................. 87 Table AIII.2 Distribution o f Calories across Welfare Groups................................................................... 88 Figures Figure 2.1 10 EmploymentDistributionby Sector andWelfare Group (%) .............................................. Classification o f Land Holding ............................................................................................ Figure 2.2 17 Figure 3.1 Rural Poverty. by Regions.................................................................................................... 28 Figure 3.2 Rural Food Security "Problems Satisfying FoodNeeds Duringthe Last 12 months" ................................................................................................................ 29 Figure 3.3 34 Perceptions o f change "How do things compare with last year?' ........................................ Daily Caloric Distributionper FoodGroup. By Consumption Quintile ............................... Figure 3.4 35 Figure 3.5 Incidence o f Covariate Shocks (% o fHousehold Reporting)............................................... 37 Figure 3.6 Incidence o fIdiosyncratic Shocks ("A o f Households Reporting) ........................................ 37 Figure AI11.1 Food Consumption Distribution(lorenz curves). by Region................................................ 87 Figure AIII.2 Food Supply -Distribution o f District-level Caloric Intake Gaps ...................................... 89 Boxes Box 1.1 UnpackingDrought inAfghanistan: Extensive and ExtendedImpacts.................................. 3 Box 2.1 Diversity and Dilemma: Rural Livelihoods and OpiumPoppy Cultivation......................... Box 2.2 The Role o f Opiuminthe Livelihoods o f Itinerant Harvesters: Helmand Province ............12 13 Box 2.3 14 The `Winterization' o fRural Livelihoods ............................................................................ The Role o f Women inOpiumPoppy CultivationinAfghanistan....................................... Box 2.4 19 Box 2.5 Refugee Returnand Remittances ......................................................................................... 21 Box 3.1 Some Safety Net Programs................................................................................................... 43 Box 3.2 Oxfam's Cash for Work inHazarajat: Lessons Learnt for Labor ....................................... 46 UnderstandingUrban Livelihoods inAfghanistan ................ Intensive Public Works ...................................................... Box 4.1 ....................................... 47 Box 4.2 Afghanistan's UrbanWater Supply and SanitationCrisis.................................................... 51 Box 4.3 The Challenge o fUrban Land Management inAfghanistan ................................................ 53 Box 4.4 KabulVulnerability Mapping............................................................................................... 55 Box 4.5 RisingDrugAddiction inAfghanistan................................................................................. 58 Box 4.6 Child Labor inUrbanAfghanistan ....................................................................................... 60 Box 5.1 National Development Framework ...................................................................................... 66 Box 5.2 Emerging Lessons for Social Protection from Post-Conflict Countries Box 5.3 Key DesignFeatures o fa GoodPublic Works Program...................................................... 71 Box 5.4 Targeting .............................................................................................................................. 74 Box 5.5 76 Use o f Consumption Expenditures to Measure Welfare ...................................................... Food-Based Safety Nets: Getting the most out o f FoodAid ................................................ Box AII.l Box AII.2 Calculating a GeneralPoverty Rate and Future Poverty Estimates Comparisons ................80 81 Box AII.3 International versus NationalBasedPoverty Lines.............................................................. 83 Appendices Appendix I Data and Sampling Issues..................................................................................................... 78 Appendix I1 Measuring Poverty . Conceptual Issues............................................................................... 80 Appendix I11 Calorie Gap Analysis............................................................................................................ 87 Map:No.AFG33358 .................................................................................................................................... 94 Some KeyConceptsusedthroughout thisReport WorkingDefinition A condition that reflects physical, economic, social and political deprivation, as well as a lack of assets and income. In this report, an individual is classijied as poor if hidher per capitafood-consumption is below the food) poverty line, defined as the annual cost of consuming 2100 calories/day of the typical basket of food items among the poorest. Although an incomplete assessment of poverty due to the lack of information on non-food essential consumption items, this methodology for assessing poverty is widely acknowledged as robust in the poverty measurement community. With the availability of more comprehensive consumption data in the 2005 NRVA, a more complete assessment ofpoverty will bepossible. Whereaspoverty reflects an unacceptable level of well-being, vulnerability captures the exposure to uninsured risk leading to a socially unacceptable level of well-being in thefuture. Vulnerability is of growing concern topolicy makers since it refers to the idea that certain groups, particularly the poor and sub-groups within the poor, may be relatively more vulnerable to shocks, the consequence of which is deeper, chronicpoverty. Vulnerability hinges upon the notion of risks as uncertain events. Risks impact upon people through the changes in behavior brought about as a result of their anticipation. The presence of risks can induce behaviors that lead to or deepen poverty rather than the desirable opposite effect, i.e.poverty reduction. Shocks impact upon people through the realization of risks. Shocks often lead to fluctuations in income and werfare outcomes and are a key cause ofpoverty. Shocks can be idiosyncratic (affecting a single person or household) or covariate (affecting a group of individuals or households at once). Nobody can escape risks but signijicant differences exist in the ability of different individuals and households to manage risks and the impact of shocks. Despite their constraints, poor people frequently play an active and vital role in risk management, both in prevention and mitigation (ex ante) as well as coping with the negative impacts of shocks when they occur (expost). Risk management instruments can be informal (community / households), market based (such as insurance) or public sector led, and they can operate on an ex-ante basis (prevention of shocks, mitigation of impact) or ex-post (coping with the result). All public sector interventions that help individuals, households and communities to manage risk and deal with its consequences. Social Protection can also provide support to those in chronic poverty. The term "livelihood" is used to describe the capabilities and assets of individuals and households and the strategies used to secure a living. An individual is food insecure ifhe/she consumes less than the minimum requirea calories per day. For rural Afghanistan this is set at 2100 calories per day. The National Riskand Vulnerability Assessment* This report draws extensively on the Nationwide Risk and Vulnerability Assessment (NRVA) which was carried out in July-September 2003 under the umbrella o f the Ministry o f Rural Rehabilitation and Development (MRRD) with collaboration from other ministries, and with assistance from several international agencies including WFP, the World Bank, FA0 and others. The NRVA surveyed 1850 rural villages, 5,560 community groups and 11,200 households in almost all the districts in the country, with only ahandfulo fareas inthe Southnot covered due to security reasons. Questions were asked inrelation to many aspects o f rural life including food consumption, health, education, water, markets, transport, agriculture and livestock, migration, income generation, assets, risks and shocks. The resulting data include information on a variety o f subjects and topics that allow analyses at the individual, household and community level. While the absence o f a sampling frame and recent population estimates prevent the NRVA from being used to truly represent all of rural Afghanistan, and the data face other limitation (detailed below), invaluable and far reaching insights into the lives o f rural Afghans are available for the first time inthe post-Taliban era. Inorder to interpret the data accurately, it is important to recognize certainlimitations ofthe NRVA: The NRVA is not statistically representative of all rural Afghanistan: Because o f the absence o f accurate population data, the NRVA does not allow accurate conclusions to be drawn about exactly how many rural people are poor and which regions and provinces are better or worse off. This will be possible only after the completion o f the census. However, the sample size is sufficiently large to draw valuable insights into the experience o f rural poverty (economic and other kinds o f wellbeing) and capture significant patterns and trends to informdecision making. No conclusive Kuchi data: While efforts were made to include nomadic populations and to collect information from Kuchi leaders, the NRVA was not very successful inthis regard. A more comprehensive assessment o f the nomadic population (NMAK) was undertaken in the spring o f2004, andpreliminary data arejust becoming available. No urban coverage: The NRVA did not cover urban centers (such as main cities and provincial capitals). However, some initial insights into urban areas have arisen from qualitative data undertaken by various agencies. A one off `snapshot': The NRVA was undertaken during July and September 2003. Some of the questions asked related to the situation households faced at that time (after a good harvest and at a time when food prices were relatively low). Some questions seek to capture people's experience over the year prior to the survey (2002 - 2003), but the NRVA does not capture seasonality. At the household level, only food consumption data were collected. In addition to the absence o f non-food consumption data, the seven day recall method used to collect information on quantities o f food consumed and construct the welfare measure pose other challenges: scaling up from one week to a year; dividing the food among household members to arrive at a per capita estimate; and translating food quantities into monetary (expenditure) estimates. Notwithstanding these challenges, the approach usedin this report to construct the welfare measure (poverty) is within the bounds o f practice in the international poverty community and the results are considered acceptable. Over time, many o fthese data limitations canbe addressed, and the 2005 NRVA is likely to be a major step forward. * For more details on the NRVA data, sampleandwelfare measurement, see Appendices Iand 11. EXECUTIVESUMMARY 1. Throughout its history, Afghanistan has been impacted by political and economic shocks, and natural disasters. Following a succession o f despotic regimes, the collapse o f legitimate government institutions and their substitution with rule by force over the past quarter o f a century, it is o f little surprise that Afghanistan has emerged as a poor nation, whose people exhibit some o f the worst human and social indicators inthe world. Yet the poverty o f the people is not as one might expect. Suffering has been widespread and multi-dimensional, yet the people o f Afghanistan are not on the brink o f starvation, nor have they beeninrecent years. This is due mainly to widespread personal and private initiative, and the support o f non-governmental organizations. Afghans are resourceful, resilient, creative, opportunity-seeking and entrepreneurial people (as witnessed by the high incidence o f labor migration, entrepreneurial activity wherever they are located, trading networks and remittances). Their achievements inthe face of adversity are noteworthy. Giventhis spirit of survival and independence, significant opportunities lie ahead as political stability returns, the economy grows, and the institutions o f state begin to pick up the reins. Poverty and vulnerability in Afghanistan, while extensive, have always been and should continue to be managed principally by the poor themselves. As such, the challenge for future govemments will be to hamess and support the proven capabilities o fthe Afghan people towards collective national objectives for development and prosperity. 2. This report on poverty, vulnerability and social protection seeks to provide a first comprehensive study o f the situation facing the people o f Afghanistan, highlighting the relationship betweenpoverty, risk and vulnerability. Specifically, the objectives o f the report are: i. togainamorein-depthunderstandingofpovertyandvulnerability asexperiencedbydifferentsocio- economic groups; ii.toexaminetherangeofrisk-managementinstruments(informalandformal)thatareavailableandtobetter understand their effectiveness; and iii.tosuggesttheparametersofabroad-basedpovertyreductionstrategy. 3. The report draws extensively on the Nationwide Risk and Vulnerability Assessment (NRVA) which was carried out in July-September 2003 under the umbrella o f the Ministry o f Rural Rehabilitation and Development (MRRD) with collaboration fromother ministries, and with assistance from several international agencies including WFP, the World Bank, FA0 and others. The NRVA surveyed 1850 rural villages, 11,200 households inalmost all the districts in the country, with only a handful o f areas in the South not covered due to security reasons. The resulting data include information on a variety o f subjects and topics that allow analyses at the individual, household and community level. While the absence o f a sampling frame and recent population estimates prevent the NRVA from being used to truly represent all o f rural Afghanistan, and the data face other limitations, invaluable and far reaching insights into the lives o frural Afghans are available for the first time inthe post-Taliban era. A similar data effort is soon to be launched for the urban areas, thereby providing a complementary level o funderstandingo furban livelihoods. 4. This report is organized as follows. The first chapter provides a short historical overview o f poverty and vulnerability, and highlights some the ways in which the Afghan people have survived more than two decades o f conflict, recurring drought and other natural disasters. Chapters Two and Three focus on livelihoods, poverty and vulnerability inruralAfghanistan, where more than three-quarters o f the Afghan populationlive, while Chapter Four focuses on the same subject matter in urban areas. These three chapters provide the reader with a detailed insight into life in Afghanistan, how people work with the opportunities around them, manage risk and vulnerability, and the resultant outcomes in terms o f poverty and human welfare. The final chapter o f the report sets out the policy directions for poverty reduction, and suggests that the essential path for poverty reduction is sustained, equitable, broad-based economic growth, with the State playing the role o f facilitator, regular and financier o f high-retum infrastructure and services, but leaving provision largely inthe hands o f the private sector, communities and NGOs. Some directed provision of public resources may be justified for the extreme poorlmost vulnerable who are unable to participate in economic activity, and experience to-date suggests that working with community networks is the most effective way to reach these people. How do rural peoplelive? 5. The report confirms that livelihoods o f the approximate 17.5 million rural Afghans are highly complex. 11 People in different parts o f the country have very different survival strategies and different households livingin the same area live in very different ways. Even within the same household, family members draw on diverse income sources depending on their specific assets and abilities as well as on seasonal opportunities. This diversity appearsto be a crucial strategy for the ruralpopulation to manage the risks and uncertainties which they report as a central part of everyday life. Although living in very difficult circumstances, findings highlight the remarkable resilience o f Afghan people. 6. Rural income strategies depend on both farming (self employed farming and agricultural wage labor) and non-farm employment (self owned enterprises and non-agricultural wage labor). As expected, agriculture (including poppy) plays a major role inproviding a livelihood inrural Afghanistan, with better off households with land assets more likely to farm their own land, while poorer households, even those with some land, are more likely to seek additional employment opportunities andor sharecrop. Out o f the poorest 20 percent o f the NRVA sample rural population, only a small percentage engage infarming on their own land, whereas a much larger percentage o fbetter off households farm. The majority o f the sample rural poor are therefore net consumers o f food rather than producers. One quarter o f the NRVA households reportedbeing completely landless. These are among the poorest. 7. Perhaps surprisingly, the NRVA data highlight the extent to which the rural non-farm sector provides a significant share o f employment, both for richer and poorer households alike. Almost half o f all rural employment involves non-farm activities (such as construction, trade, manufacturing, transport, mining and other services) and the data suggest a strong linkage between farm and non-farm growth. Cross-border migratory strategies (mainly Iran and Pakistan) were reported by 14 percent o f rural households. Many more households reported highly mobile arrangements, and engage in opportunistic travel to neighboring districts and provinces in search o f seasonal work. Literacy and access to assets are key determinants o f non-farm employment, with important differentials in productivity and wages evident. 8. More work is needed to better understand the role that women play in household income generating strategies. While significant regional variations exist, the NRVA sample suggests that the poorest rural women can be economically active and that rural women ingeneral perceive education, skills development and access to savings and credit services as top priorities for improving both their own welfare and that o f their families. Child labor is also apparent throughout much o f rural Afghanistan. What do we know about the ruralpoor? 9. International experience in measuring poverty suggests that it is generally more reliable to look at what people spend or consume than what they earn. As one o f the approaches used to measure poverty in rural Afghanistan, the NRVA therefore collected data on food consumption over a recall period o f seven days. It should be stressed that no methodology i s totally accurate in assessing poverty and that the overall pattern is more significant than any precise figure -throughout this report the sample is divided into "quintiles" and much of the analysis compares patterns across the five quintile groups. The data suggest that around half o f the sample have food expenditures that are insufficient in order to purchase a basic food basket providing 2100 calories a day per person based on local prices. When in addition, households were asked directly whether they had experienced problems satisfying food needs over the year, the results were found to be very similar - around half o f the rural sample reporting that they had experienced some food insecurity. The poorest segment o f the NRVA sample (the poorest quintile) report food insecurity year round. Although a little better off, the bulk o f the rural population is living on the edge of poverty, and the on-set o f a large shock - another wide-spread drought, for example - could send many more people into poverty. In terms o f numbers o f people, the report estimates that around 3.5 million rural Afghans are extremely poor, another 10.5 million are vulnerable to extreme poverty, and the balance, another 3.5 million are less poor, butnonetheless still vulnerable to poverty. 10. Poverty in Afghanistan is multidimensional, involving a complex interplay between low assets (physical, financial and human), years o f insecurity and drought, indebtedness, poor infrastructure and public services, traditional roles and other factors. In summary, and drawing on analysis whereby different factors are simultaneously compared with the poverty indicator in an effort to distill their separate effects, the main determinants ofrural poverty emerge: e Location -the NRVA data suggest that some parts o fAfghanistan may have overall poverty levels that ... 111 are higher than others, particularly the West, North and West Central. By contrast, the data indicate that the North East may have relatively lower levels o f poverty. However, while these findings are o f interest, caution i s urged for two reasons. First, without accurate population figures and a sampling frame, these patterns are highly tentative. Second, the data also highlight huge variation within regions, provinces, districts and communities. Inother words, even if one province seems to have a higher than average incidence o f poverty, an apparently better off province may contain some districts in which poverty is even higher. Very poor households exist inalmost all villages. Number of children -The averagehousehold inthe NRVA sample comprises 7.5 persons, comprising four children, and three to four adults. While there is little variation in household sue across food consumption quintiles, household composition does differ: the better off households have fewer children and more adults, suggesting a possible link between income generation capacity and dependency, although inthe Afghan context children are a source o f income as well dependents. Gender of householdhead -There is a strong relationship between the sex ofthe householdhead and household welfare: female headedhouseholds, defined as such by the households themselves, are more concentrated inthe lowest food consumption quintiles. Given the constraints that women face inrural Afghanistan, especially vis-a-vis working outside the home, ownership o f assets and the low levels o f literacy, it is not surprising to findfemale headedhouseholds amongst the poorest. Disability -About 14 percent o f rural households have a disabled family member, and some o f these households are among the very poorest. I t is important to distinguish, however, among households with a disabled member: not all are equally poor. Migration-Households who have migrated domestically (including IDPs) tendto bepoorer, butthose who have returned from abroad are relatively better off. Remittances play an important role in alleviating poverty; HumanCapital -Illiteracy is high inrural Afghanistan, and only about 30 percent o fhouseholdheads can read and write: among women it is only 10percent. Levels o f education are very low, on average only 1-2 years o f schooling. Nonetheless, household heads who can read and write have a lower incidence o f food consumption poverty. This is partly mediated through the labor market and access to non-agricultural employment, and may also reflect better farming practices. Migration abroad, usually more associated with the educated, i s also positively correlatedwith higher food consumption, whereas intra-rural migration is more closely aligned with the poor. Land - One quarter o f the households in the NRVA sample were landless, and some o f these households are extremely vulnerable. Among those with land, ownership o f irrigated land has a more positive relationship with higher food consumption that those deriving their livelihood from rainfed land. After water, access to livestock and tools for farming is the next most important consideration, and poor households report a high incidence o f production constraints arising from lack o f livestock and tools, as well as water. Opium Poppy -the cultivation o f opium poppy leads to higher levels o f income. In addition, households residing in communities where poppy is produced are also less likely to be poor. This is most likely due to the labor intensive nature o f poppy cultivation, involving large numberso f workers, especially during the harvest, as well as the general effect o f opium on the local economy. 11. The NRVA data underscore the fact that serious problems are faced inaccessing basic services byvirtually all rural households - rich and poor alike. Very few people have access to government services and infrastructure. One of the most striking indicators is lack o f access to health facilities - only 9 percent o f the communities surveyed inthe NRVA have accessto a healthfacility. While those who can affordto buyservices fromthe private sector are getting limited electrical power, medical services and transport, access to publicly provided services such as drinking water, roads, schools, clinics and hospitals is extremely low among almost all households. The welfare implications o f this are far reaching. Alarming data on sickness and mortality, especially maternal and child mortality (see MICS, 2003) suggest that a focus on household income poverty alone does not explain the extent of suffering and deprivation across all rural communities inAfghanistan. 12. The NRVA collected data on different `shocks' that affect rural households, with the view to an improved understanding o f how these impact householdpoverty, and why some people do better than others at managing life ina very risky environment. Four out of five households said they hadbeenaffected by a widespread shock during recent times, and the majority identified drought as the most serious shock. Seventy percent o f drought affected iv households said they had still not recovered income and asset losses, and 40 percent o f rural households said they had had to reduce food consumption as a coping strategy. Given poor dietary pattems to begin with, any further reduction in food quantity or diversity would have long-lasting implications for children in particular. These findings suggest that policies that reduce the likelihood o f shocks such as drought (improved water management) and support risk management instruments which help reduce their negative impact (crop diversity, crop insurance) are very important inrural Afghanistan. 13. After drought, the death or sickness o f a working member o f the family, and loss o f employment are prevalent among poor households. The poorest quintile o f sampled households is twice as likely to have experienced sickness and/or a death inthe family as other households, highlighting the seriousness o f the health situation inrural Afghanistan. Such shocks represent individual suffering and human tragedy. They also have knock on effects in terms o f their negative impact on the welfare o f all household members due to the resulting loss o f incomes and assets associated with ill-health, long-term disability, burial costs and lost human capital when children are withdrawn from school to work. Some poor households regretfully reported that such shocks lead to the marriage o f daughters at a younger age than would otherwise have been the case. Repeated shocks appear to trap households in poverty andpoor people may needto be helpedto escape from these traps to supplement their own efforts. 14. Contrary to expectations, only a small percentage (2 percent) o f households reported violence as a household level shock. There could be many reasons for this including the fact that violence is often localized so that, for example, a high incidence in a few districts is not apparent innational averages. The way the question was asked and the sensitivities around this issue (particularly with regard to political and household violence) may also explain under-reporting. 15. Despite this overall bleak picture, the rural NRVA sample reported that they thought things were better or the same inthe summer o f 2003 as compared to the previous year. For some households, the situation was clearly better, but for others things had worsened - these were typically the poorest households. This suggests that even at a time when the rural economy was benefiting from economic growth and a good harvest, there remains a core component of poor and extremely poor people for whom life is very difficult. 16. The NRVA explored the connection between poverty and participation in relief food and emergency employment programs (food for work and cash for work) during the year 2002-2003. The data show that almost half of all rural households had benefited at some time from participation in these programs. However, the data also indicate that the share o f poor and less poor households among those participating was quite similar. This was particularly true for the labor-based public works programs, yet even relief food was finding its way to the non-poor top quintile, pointing to the challenge o f targeting o f these programs. While these results could be viewed as `bad targeting' and an example o f `resource leakage', the issues raised are complex. As discussed, the majority o f rural households live in or close to poverty, and it is very hard to distinguish between those more in need from those a little better off. The challenge faced by labor-based, employment programs to directly involve more disabled people and higher numbers o f women - some o f the poorest people in rural Afghanistan - has been acknowledged. In addition, further work is required to measure the extent to which communities share available resources through local support networks and mutually benefit from newly created assets to improve the welfare o fpoor households. It may not be correct to assume that poor households can only benefit from direct participation and payments. However, clearly the findings o f the NRVA point to the requirement for further investigation o f this matter if scarce resources are to have maximumimpact onpoverty reduction among the poorest. What do we know about poverty inurban areas? 17. Inthe absenceo faccurate population data, the Government estimates that around 6.4 millionpeople (about 25% of the population) live inurban Afghanistan and that this figure will have doubled by 2015. InKabul alone, as many as one million returning refugees may have resettled since early 2002, and the return o f remaining refugees from Pakistan is likely to further swell the numbers o f the urban population. The 2003 NRVA collected no data in urban areas, but efforts are now underway to undertake a similar exercise with the urban population. This will add valuable insights into urban poverty and vulnerability. At this stage it may be helpfulto consider the main findings of existing qualitative assessments, mainly relating to Kabul, to highlight potential similarities and differences betweenpoverty inrural and urbanAfghanistan. Initial insights suggest the following: V e Because o f the dense concentration o fpeople, urban households are often more dependent on complex public delivery systems to meet their needs than their rural counterparts may be. Protracted conflict and instability has resulted in the collapse o f basic systems for urban management and investment in Afghanistan and data on disability and diarrhea prevalence inchildren under five years suggests little difference between rural and urban areas. e Employment opportunities are often highly insecure in Afghan cities and towns and the frequent movement o f households (often living in rented accommodation) can have a negative impact on community support networks, leaving the poor extremely vulnerable. e Urban areas are a magnet for extremely poor and destitute households and individuals who may be highly dependent on support for their survival. However, at the same time, with a higher proportion o f expenditure on, for example, rent and fuel costs inurban areas, many poor households may face food insecurity. e For many poor households, urban Afghanistan presents a high risk environment, with accidents (road, fire, etc.), personal violence and crowded unsanitary living conditions widely reported. e Nevertheless, an initial survey among the most vulnerable areas o f Kabul suggest that over half o f the sampled households reported having more to eat now than a year ago, compared to 19 percent who reportedthat they were eating less. Policy Directions 18. Sustained, equitable economic growth is the essential path for poverty reduction - for in the absence o f growth, there will be little to share among the poor. Although poor at present, there are many people inAfghanistan who can participate ina growing economy if given the endowments and opportunities - education and skills, health, jobs, business opportunities and credit, land and water. Even some o f the disabled, certain categories o f women, intemally displaced people - which include some o f the nomadic Kuchi - can, if equipped with assets and capabilities, enjoy a level o f participation in economic activity. Government has a key role to play in laying the foundations for sustained economic growth, equipping and assisting its citizens to participate and benefit from growth as the mainavenue for poverty reduction, andputting inplace a minimumsafety net for those who genuinely cannot help themselves. The latter, however, needs to build carefully on community solidarity and mutual support mechanisms, which have proventhe backbone o fAfghan society for centuries. Economic Growth: The Core of Poverty Reduction 19. The essential path for poverty reduction in Afghanistan is sustained, equitable economic growth. In the absence o f growth, and the conditions necessary to facilitate it, the scope for poverty reduction is severely constrained. Thus, the key challenge for Afghanistan is to put inplace the institutions, policies and services that will generate sustained, broad-based economic growth. As noted inthe report, SecuringAfghanistan's Future, inaddition to the level, the quality o f economic growth will be critical: "The quality o f growth is very important. What is needed is growth with improving social indicators (reflecting investments in human capital) and without a significant deterioration in income distribution. This ensures continuing broad consensus around the policies and decisions needed to create the enabling environment for economic growth. Inthe Afghan context, growth needs to be labor-intensive, sustainable macro-economically and financially, environment-friendly and conducive to social development. Growth also needs to be reasonably well-balanced ethnically and regionally in order to avoid exacerbating political tensions among different groups and regions." (SAF, p. 7). On this basis, growth will fully support poverty reduction. 20. Afghanistan has areas o f good potential for growth, ranging from agricultural production and processing to mining, construction, trade and other services. Exploiting Afghanistan's geographical position as a "land-bridge'' between Central and South Asia can also be a source o f growth, through trade, transit and joint development o f resources. The govemment's SAF report puts forward an ambitious goal o f 9 percent annual non-opium GDP growth over the next 12 years. If realized, this will have important positive implications for overall poverty reduction. However, Afghanistan faces some major constraints insustaining such a growth rate, which will have to be led by a strong, competitive private sector. Key challenges include: (i) improved security and rule o f law; (ii)a pro-private sector regulatory framework, free o f corruption; (iii) an effective, accountable state; (iv) provision o f essential support services such as infrastructure (power, roads, serviced land, water), finance, insurance, business support services, agriculture extension and marketing, and quality assurance for export; and (v) development o f vi human capital skills, both technical/professional/managerial for business leadership and a literate, skilled workforce. The Government has already made a very good start by maintaining macroeconomic stability and initiating some important structural reforms, as well as capacity building in the public administration. These very successful and important initiatives need to be complemented by capacity building at all levels o f government and delivery o f key infiastructural and social services. If successful, these policies will pave the way for wholesale poverty reduction throughout mucho f the country and for a large share o fthe population. Equippingthe Poor with Assets and Capacitiesfor Growth: Public InvestmentPriorities 21. Sustained over some years, economic growth can increase opportunities for most Afghan people, including poor people inboth rural and urban areas. Its impact inthis regard will depend on how well equipped poor people are to participate in growth, and to what extent growth helps the poor to build their assets - physical, financial, human and social. Having the ability to accumulate physical capital such as (imgated) land, livestock and small businesses, having access to (micro) finance and business support services, as well as finding employment opportunities will be fundamental to poverty reduction for people who are presently poor. Improvements in opportunities for poor people will also require buildingtheir humancapital - education, skills and health - which is especially important for people who are dependent on their labor for a livelihood - such as the landless. Asset building, on the part o f the poor, needs to be supported by a range o f well-targeted, cost-effective programs. While the overall trust o f the Government's development strategy provides a solid foundation for poverty reduction, greater emphasis will be needed in some areas to ensure that the poor are equipped to contribute and participate ina growing economy. Inso as the Government can focus scarce administrative and fiscal resources on these programs as a matter o f priority, this will be to the benefit o f poverty reduction and the poor. Those elements o f the government's program most important for poverty reduction are the following: 0 Education and training: access to quality schooling for girls and boys is of the highest priority for poverty reduction in the longer-run, as one o f the drivers o f growth. In the more immediate future literacy training for women, and skills development for a wide range o f people can play large dividends. Experimenting with different delivery models for schools and skills development training should be pursued. 0 Health and nutrition: improvements in morbidity and mortality will come from a variety o f interventions ranging from better maternal and child healthcare, a more varied diet, clean drinlung water and improved sanitation, and education. In the short-run the emphasis should be on potable water and provision o fbasic health services. As drought continues to plague much o f the country, the shortagelabsence o f dnnlung water in the rural areas has become a major concern, and warrants topmost attention o f the government. The provision o f drinking water and modem sanitation schemes are drivers for a reduction in infant and child mortality, improvements in health more generally, learning in school and worker productivity; as with education, provision o f health services can come from a variety o f different providers. Livelihoods and Social Protection: this consists of a number o f government programs, operating through a variety o f parallel interventions: community level asset creation, short-term employment generation, micro-finance, and participation and inclusion in local decision making. Core elements o f this program include National Emergency Employment Program, National Solidarity Program, rural community water and sanitation schemes, area based development projects and micro-finance. All these programs are o fthe highest important. 0 Natural resource management: investment in agriculture (half o f the Afghan economy) for the settled population and rangelands for the livestoek economy (important for the kuchi and women) will have important implications for the rural poor with some assets as well as the landless who work as agricultural laborers; provision o f rural finance, water managemedimgation infrastructure, tool and seeds, research and extension services, and marketing are all important inthe longer-run. 0 Urban management: urban services, especially water and sanitation, solid water disposal, and housing are o fthe highest priority for the urbanpoor. 22. With these programs the Government o f Afghanistan has put inplace many o f the elements o f good policy aimed at equipping the poor with assets and capacities that will support their active participation in a growing economy. Many of these programs have the dual attribute o f being good for both growth and poverty reduction. The vii immediate challenge is to ensure adequate finding and effective implementation o f these programs. There is a danger that the government is taking on too much at once, and the launch o f new programs could result in limited fiscal and administrative capacity being spread over too many areas. At this stage, data are inadequate to ascertain which o f the national investmentprograms are most cost-effective as poverty reduction instruments - this i s an area o f analysis that should be developed over the next few years as part o fthe (Interim) NationalDevelopment Strategy. The Social ProtectionProgram 23. Social Protection is at the heart o f poverty reduction. Some o f the commendable features o f government's livelihoods and social protection policies and programs include favoring productive safety nets such as labor intensive public works and other employment creation activities over direct transfer programs, strengthening livelihoods through education, skills development and micro-finance, providing health services, drinking water and sanitation, and working with and through communities. These directions are largely in line with emerging good practice frompost-conflict countries. 24. Two major challenges face the design o f social protection inAfghanistan. First, how best to support both short-term emergency measures to deal with the immediate needs o f poor and vulnerable people as well as put in place the foundations for a longer-term social protection system, compatible with an economy undergoing reconstruction and development. Fortunately some programs, such as public works programs and the National Solidarity Program can serve both short and longer-term goals. For others - such as a social safety net - there is a more direct trade-off (see below). The second challenge refers to how to prioritize among the many competing immediate needs for support: traditional poor and vulnerable groups (some o f the kuchi, rural landless) versus new conflict-related vulnerable groups (such as demobilized soldiers, war disabled, war widows and orphans, displaced people) versus new vulnerabilities (such as drought related, or where opium poppy is being eradicated). As discussed above, although poor at present, many o f these people can participate in a growing economy if given the endowments and opportunities -education and skills, health, jobs, business opportunities and credit, landand water. Even some o f the disabled, certain categories o f women, internally displaced people - which include some o f the nomadic Kuchi - can, if equipped with assets and capabilities, enjoy a level o f participation in economic activity. Favoring social protection interventions that provide both "protection" in the short run and underpin equitable growth inthe longer runare to be preferred. 25. Although designed to address the needs o f poor and vulnerable people, thus far core elements o f Afghanistan's social protection program have proven more effective inreaching the broad population as compared to the poorest: a "welfare neutral" distribution o f participation is broadly observed. Given the importance o f getting programs up and running, severe administrative limitations and the absence o f data to guide resource allocation decisions, these outcomes are not unsurprising. Moreover, many o f the existing programs, by nature o f their design, are not well equipped to reach the very poorest and most vulnerable. These categories o f people cany physical, spatial or social attributes that may prevent their participation in mainstream economic activities, especially activities requiring a laboring effort outside o f the home (disabled and women). Little can be done to change those characteristics, but at the margin the programs do have the potential to be more effective inreaching out to some o f the poorest and most vulnerable. Indeedoperational manuals for programs such as NSP and NEEP make provision for greater outreach. For example, inaddition to the investments ininfrastructure, NSP block grants canbe used for (i) self-help savings and credit schemes for women and disabled, and (ii) productive asset transfers for vulnerable women and disabled. To make this effective, there is a needto develop clear methodologies for how this assistance canbe best targeted and provided drawing on best practice lessons from other projects. Some pointers for improved targeting o f labor-intensive public works programs include consideration o fthe types o f activities, their location, the wage rate andmode ofpayment. A Safety Net for the Poorest? 26. One o f the outstanding, more immediate - and most difficult - challenges for Afghanistan's social protection strategy lies ineffectively reaching the poorest and most vulnerable people who are not able to participate inasset-creating, income-generating programs, labor-based programs, yet who are least able to weather any further shocks. The first question to tackle is "who to protect". In recent times Afghanistan policy makers have designed programs for certain categories o f people in the belief that they are among the poorest and most vulnerable. These include people with physical disabilities, female-headed households, the nomadic kuchi, IDPs and returning ... Vlll refugees, and highly disadvantaged children. Analysis o f the NRVA data indicates that there are indeed people in Afghanistan who can be recognized as very poor and vulnerable, and indeed the Government's use o f targeting by characteristic (categorical targeting) has, for most part, been appropriate given its ease o f administration. However, it shouldbe acknowledged that not allpeople inthe categories-disabled, female-headed households, IDPs, refugees and Kuchi are equally poor, and that other people not included in these categories, such as the landless, are among the poorest. 27. Inconsidering a more "finely tuned" safety net for the poorest and most vulnerable, there are a number o f realities that cannot be ignored: Extreme poverty and vulnerability in Afghanistan are not well understood: their precise nature, the numbers o f people affected, their essential conditions and location are yet to be fully documented - this report has only begunto scratchthe surface o favery complex situation. In addition to economic dimensions o f poverty, social and political conditions afflicting the most marginalized and disempowered communities, households and individuals in Afghanistan need to be addressed. Institutional and fiscal capacities are extremely limited, thereby constraining the scope and affordability o f (new) programs. M u c ho f the implementation capacity lies with NGOs. The existing cash transfer scheme for war-related disabled persons and the families o f martyrs, awarded irrespective o fpoverty status, may have established a precedence for entitlement programs. In addition, some interest groups have established powerful lobbies, often fully backed by special interest ministries (e.g., disabled, martyrs, kuchi, women's affairs). N e w vulnerable groups, though arguably not among the poorest - such as retiring civil servants and displaced public sector workers - are also making demands o n the nationalbudget. Hitherto, many o f the programs designedto reachthe poorest and most vulnerable have been supported by UNagencies and implemented by NGOs - these typically exist outside the domain o f the national development program and the national budget, and often provide in-kindsupport. 28. Currently under consideration in Afghanistan is a National Vulnerability Program (NVP)which aspires to provide support to some o f the poorest and most vulnerable people. As its design moves forward, here are guiding principles for the shodmedium term: Focus o n data collection and furfher analysis o f extreme poverty and vulnerability before launching any new programs. There is an unresolved tension between special interest groups and a poverty-based assessment o f those most inneed. Review the design and effectiveness o f on-going programs under implementationbyNGOs, and assess the potential for scaling-up. Try and influence the design and implementation o f existing program to ensure that they are as inclusive as possible. Modify the existing cash transfer program for war-related disability and martyrs families as a matter o f priority to ensure consistency intarget groups and modalities. Only as a large resort launch new programs. Program design should be grounded in the very limited admmistrative capacity and flexible indemands o n the national budget. 29. With regard to the launching o f any new direct (targeted) support to the extreme poor and most vulnerable members o f society, questions remain as to how best to target these individuals and their families. Self-targeting, categorical targeting and working through community "solidarity" channels remain the preferred mechanisms, although the extent to which community networks can be relied upon - if at all in urban areas - to effectively provide for the poorest and most vulnerable remains an open question. This is an agenda item for the NVP as it develops a monitoring and evaluation capacity. More generally, continuing research o n the nature o f poverty and vulnerability, and linking findings to enhanced targeting o f program will be key to increasing the efficiency and effectiveness o f limitedresources ina very difficult delivery climate. 30. While there is a public policy gap vis-a-vis a safety net for the poorest and most vulnerable, which the NVP is tasked to address, there are a number o f schemes runby NGOs, and supported by UNand bilateral agencies which ix have provided humanitarian support over the years and helped prevent mass destitution and hunger. For example, food, cash and housing have beenprovided to returning refugees; relief food aid has beenprovidedto IDPs, female- headed households and disabled persons alongside food-for-work programs. One o f the key challenges for the government and the aid community is to increasingly bring humanitarian assistance into the framework o f the government's development program, and ensure that it is managed as an efficient and effective element o f the NVP. Food aid is o f particular importance to the country's safety net. In the short-term, food-based safety net programs canmake an important contribution to the welfare o f the poor ifproperly managed. Conclusion 31. The above has discussed an optimistic scenario in which sustained economic growth provides a "rising tide" inwhich the bulk o f the Afghan people can participate, so long as the complementary investments in enabling capacities, services and infrastructure are made. This coupled with a very carefully designed safety net for the poorest, left in large part to NGOs, self-targeting and community based arrangements, maximizing the utility o f the food aid resource and drawing very selectively on Afghan budget resources, can make for a comprehensive poverty reduction strategy. But what if there is a major downturn in the economy, or other covariate shocks - such as drought - which impact negatively on lives and livelihoods, or more idiosyncratic shocks resulting injob loses or ill- health of a family member. Afghanistan cannot prevent such shocks; it cannot insure against all such risks nor can it afford to provide transfer payments to compensate those affected when shocks hit. On the other hand, the Government would want to intervene and try and avoid the worst forms o f private coping arrangements that have prevailed in the past when adversity struck. Finding the right balance between supporting the upside - growth and prosperity - yet beingprepared for the down side is an important consideration. 32. A final word about the formulation o fa national development (poverty reduction) strategy. A good strategy is not only technically and economically sound - it needs to be fully "owned" by the poor as well as by other stakeholders in the society. In this regard, the role o f civil society groups, including the private sector, to identify key policy, institutional, and infrastructure constraints, as well as to suggest ways forward, will be vital to establishing a poverty reduction strategy that clearly identifies the problems that it seeks to address and builds ownership across the society. The role o f both national and international NGOs in this process will be very important for enhancing ownership. Inthis context, the Government is appropriately planning to engage inextensive consultations with different segments o f society as it develops its National Development Strategy, a process that is targeted to reach a first stage o f completion with the preparation o f an InterimNational Development Strategy later in2005. 1 I.POVERTYANDVULNERABILITY:AHISTORICALPERSPECTIVE Originsof Poverty andVulnerability: A HistoricalReview 1.1 Vulnerability begins with a notion o f risks as uncertain events, negatively impacting upon individual welfare through their realization as shocks. In macro-economic terms, Afghanistan performed modestly well inthe 1960s - 70s.' The most obviously apparent risk to both national growth and household well-being was drought, mainly in the early 1970s. However, the following decades o f conflict were to have a profound impact on the prospects for security and development. From 1978 onwards and for almost a quarter of a century the people o f Afghanistan have endured an extraordinary onslaught o f severe shocks. The devastating impact o f these resultedina war-shattered nation characterized by the near total collapse o f physical, economic infrastructure and functioning government institutions, widespread poverty and chronic erosion and abuse o f human rights, massive human displacement (by 2001 over 25% o f the population was displaced within the region) and the inability to cope effectively with recurrent natural disasters.* Protracted conflict introduced additional risks and shocks into the communication and transportation sectors, disrupting services and isolating communities as a result o f the miningo f roads, the ambushing o f convoys and the increased incidence o f banditry and theft. 1.2 Prior to the coup o f 1978, semi-feudal structures existed in most o f rural Afghanistan. The installation o f socialist policy involved efforts to implement massive land reform policies but essentially little land redistribution took place. Nonetheless, a significant proportion o f households from wealthier sections o f communities left the country at this time, and taking their capital assets with them, settled permanently overseas. The resultant negative impact on human and capital resources and opportunities for investment and growth within Afghanistan should not be underestimated. Traditional structures o f local government quickly disintegrated, although informal community structures proved more robust, and indeedwere usedto serve the interests o fthe socialist administration. 1.3 Withthe expansiono f local warlordism andrising instability after 1992, the prospects for national recovery were further set back. While the regulation o f cross border and domestic trade flows was not, for the main part, a priority for the Taliban regime, the national economy became increasingly underpinned by illicit cross border arms and narcotics trade. Local governance structures emerged from a web o f competing forces at the village level, involving returning landlords, the affiliates o f existing warlords, mullahs and elders. Rising levels o f insecurity shaped the day-to-day decisions made by Afghan households in different ways depending on the local context - limiting agricultural production, increasing the susceptibility o f civilians to attacks and increasing reliance on migration, remittances and opportunities for trade. 1.4 Population movement has historically been central to risk management strategies in Afghanistan. Opportunistic seasonalmovements insearch o f employment opportunities and cross border migrationhave occurred throughout the twentieth century. Increased levels o f politically and economically motivated out-migration took place both during the Soviet era and when the Mujahidin came to power in 1992, followed by several years o f intense conflict within Kabul. The regional dynamics o f migration and return by Afghan families over the last decade (particularly when the Taliban rose to power in 1996 and then when the coalition's military campaign took place in 2001) may have also been substantially influenced by the existence o f extended household networks and cross border social capital assets. 1.5 Nonetheless, and despite the diversity o f risks and shocks affecting Afghanistan over the years, widespread and protracted drought is perceived by Afghans as the most threatening to lives and livelihoods, with one or more negative outcomes (including reduced production, lost incomes, lost assets, unemployment and increased debt burdens) experienced by the majority o f household^.^ Over time, with high debt burdens leading to the loss o f land through mortgage and sale, the impact o f repeated shocks has contributed to the fact that poorer households tend to have fewer land assets and to derive a higher proportion o f their income from smaller, more diversified sources o f non-farm income (such as migration, carpet weaving, et^.^) Between 1998 and 2001, one o f the most severe I "Securing Afghanistan'sFuture". TISA (2004) Ibid. Lautze et aI, 2002 A historicalstudy of the carpetmarket(Pain 2001) highlightsthe sharpdeclineinterms oftrade for carpet weavers since the 1970sdue for exampleto increasingrelianceon importedwools and dyes, loss o f livestock duringthe 1970drought andthe 2 droughts in living memory further underminedthe asset base o f households with the effect that many o f the poorest (in both settled and nomadic communities) have been faced with little choice but to diversify their livelihood strategies beyond traditional farming activities due to irreparable asset depletion and destitution. 1.6 Also associated with the 1999-2001 drought was a dramatic decline in domestic grain production (off set byincreased levels o f cereal imports) anda significant reductioninthe livestock p~pulation.~The impact ofdrought, together with location specific incidents o f ethnic discrimination,6 compelled many additional households (many extremely poor and acutely vulnerable) to `displace' their livelihood base to differentparts o f the country, rural and urban, inorder to survive andretain what few assets they may still have possessed. RiskManagement Strategies: How Families Survived 1.7 The remarkable resilience, responsiveness and strength o f Afghan people are confirmed by the notable absence of any large-scale famine inthe face of adverse circumstances over the last twenty five years.7 This would appear to be borne out by nutritional data from different parts o f the country.' A key contributing factor is undoubtedly the diversity o f economic activities across Afghanistan, within communities and even within most households, facilitating a wide range o fprivate and informal risk management arrangements, utilized by Afghans in the face o f adversity.' Undoubtedly the rich fabric o f community-based networks and institutions has also played a critical role in supporting households and individuals. Another factor may be the role o f unregulated market institutions in facilitating the substitution o f locally produced wheat by commercially imported cereal, resulting in the relative stabilization o f grain (and flour) prices throughout most o fAfghanistan. 1.8 Afghanistan's geographical landscape and the range o f climatic zones contained within its national boundaries give rise to a variety o f household economy zones. As a result, farm production (cash crops - both licit and illicit, food crops for local consumption and livestock), seasonal labor migration, remittances and trade, have each come to represent a significant component infacilitating access to food, maximizing household well-being and mitigating the negative impact o f shocks on household assets." However, despite frequent acknowledgment o f the importance o f informal risk management capacities and strategies for effective policy reform and investment, this area appears to remain poorly understood. In the context o f protracted instability, such understanding is further constrained by the fact that a significant share o f livelihood strategies i s pursued outside the formal economy. The Afghan economy is dominated by small-scale, informal enterprises in all sectors. In the context o f pervasive insecurity, restricted assets and limited capacity, entrepreneurial dynamism is been channeled into trade, small-scale service and processing activities and informal and illicit activities. These include sharecropping (often tied to conditional opium production), and illicit trafficking o f narcotics or humans, employment as local combatants and a host of other criminal activities. The majority o f financial transactions continue to be provided by informal money dealers (hawah). increasednumber ofhouseholdsresortingto carpet weaving s an ex-post risk management strategy. FAO(2003). "For example, manyhouseholdsofPashtunorigin left NorthemAfghanistanduringthe period 1992-1996,retumedunder Talibanrule andhaveagainbeen forced to move since the end of 2001. The Hazarasexperiencedsevere economic blockadesby theTalibanduringthe late 1990s. The rise of theTalibanenabledthenomadicKuchisto regaincontroloftheir traditional pasturesinHazarajat(lost after 1978)but the Kuchismayhave once again lost access to these lands since the political shift in 'late2001 ." (Painand Lautze, 2002) "Although malnutrition andhungerwere reported, thisdid not degenerate into catastrophicsituation, for althoughthe infrastructuredevelopedby agriculturalproductionsystems in many areas has beendegradedor destroyed, the basic elements of landandwater remain." (UNDP, 1993 AfghanistanRehabilitationStrategy-Action Planfor ImmediateRehabilitation) 'Whilethe nutritional situationinAfghanistan is characterizedby an extremelyhigh prevalenceof chronic malnutrition (stunting), estimatedto bebetween45-59%, levelsof acute malnutrition (wasting) remain relatively low, estimated to bebetween 6-10%. Analysis o f aparticular economic crisis resultingfrom a blockadeby the Talibanin 1997/8combined with a severe winter and rising grainpricesconcludesthat householdsavertedfamine by employingarangeof complexlivelihood strategiesincluding sale of livestock, increasedlabor, migration, growing indebtedness,use of socialsupport mechanismsand accessto distress foods. The overallimpact of severeshock was heightened impoverishmentfor manyhouseholds, but not outright famine. (Semple, 1998). lo The onset of the cold season for much o f the countrymay severely restrict opportunitiesfor incomegeneration and local mobility for severalmonthsof the year, andimposeslivelihood diversification as an ex-ante risk managementlsurvivalstrategy for manyhouseholds. 3 Box 1.1: UnpackingDrought in Afghanistan: Extensive and Extended Impacts The three year period 1999- 2001 has beenreportedbyAfghans nationwide as the worst drought inliving memory, exacerbating existing vulnerabilities arising from protracted conflict. Ex post household risk management strategies included sale o f assets (ranging from livestock, land, farm tools and inthe most extreme cases, family members such as underage child marriage); the generation o f additional household incomes through migration strategies, increased reliance on the labor o f women and children, employment with local militia, and informal credit (inboth cash and kind). The overall severity o f protracted drought can be quantified in terms o f rainfall data and cereal production. However, the actual impact o f the shock on people's lives i s refracted through livelihood strategies and risk management arrangements, with a variety o f outcomes at the household level. Field reports highlight that agricultural productionis not necessarily a good proxy for the overall welfare o frural households given the diversity of livelihoods and riskmanagement strategies. Among the most significant outcomes for rural Afghans are: Indebtedness:Overall debt burdens appear high with the consequences o f debt crises profound and disturbing." Not only the poor, but even middle income rural households often faced no choice but to borrow money, often against their lands, houses, orchards, water rights, etc.. The lender can often use the asset until full repayment has been made. Individuals unable to repay money can be forced to sell the asset in order to clear debts but this appears relatively rare. Insome cases families have beenpressuredinto marrying their young daughters to money lenders. Greater Insecurity of land and water rights: In order to cope with losses o f income, households asset sales have often included land, thereby increasing the inequitable systems o f land ownership and heightening social inequality more generally inrural communities. Shocks also impact on informal community-based institutions, which play an important role inregulating household access to key resources on which livelihoods are based. For example, deep concems are widely reported inmany villages that the drought has increased the ploughing up o f traditional pasture areas for rain-fed cultivationby local 'power holders', thereby depriving poorer households o fthe access to common pasture that they have traditionally relied upon." Scarcity o f resources can also heighten tension and the threat o f armed conflict between different villages or specific ethnic groups, realized for example, by the introduction of informal and unfair restrictions on access to water points and in the collapse o f management systems and the traditional regulation o f irrigationdistribution systems.I3 Livestock Sales: Livestock losses were profound for both the pastoralist Kuchi and for the rural population in general who reduced or lost access both to sources o f animal traction (oxen) and household consumption o f dairy products. Recent reports o f the reduced profitability o f sharecropping arrangements highlight the significance of ownership o f animal traction, even for landless households. Environmental Damage: The protracted drought involved an increased dependence on the harvesting o f trees and shrubs as a coping strategy at the same time as supplies were dwindling. The collection of fuel wood became an important source o f income for the poor, hit by reduced farm and shepherding opportunities. Inaddition, the market demand for fuel wood increased as a result o f the widespread decimation o f livestock herds and associated losses of dung for fuel. Health, nutrition and education, especially o f children and youth. Children may be forced to take on the role as primary household breadwinner where adults are confined to home due to disability, debt, insecurity or restricted mobility for cultural reasons (for many women). Parents may decide to withdraw children from school. Widespread water insecurity also has a negative impact on human capital by significantly increasing the time spent fetching water, reducedbathing to less than once a month and a general rise inwater-related health problems and water-borne disease (people may share the same source o fdrinkingwater as livestock). 1.9 Migration has been a central livelihood strategy for Afghan households for decades. In general, better-off Afghans have found various opportunities in Europe and North America, while young men have often looked for I'Lautzeet a1reportthat around 80% ofthe householdssurveyedfaced serious levels ofdebt insecurity, with debt a source of shame anddeep anxiety for the household. l2Favre, R (2003) Grazing LandEncroachment. l3Pain(2004), UnderstandingVillage Institutions. 4 work inthe Middle East, and others have sought refuge inPakistan and Iran. Internal displacement during times o f conflict has consisted o f rural to rural, rural to urban and urban to rural movement depending upon the specific risks and shocks being managed and the location o f existing social networks. The historical emergence o f a remittance economy in Afghanistan has involved significant financial flows from the diaspora community which has undoubtedly played a key role in poverty reduction. However, it is extremely difficult to estimate the relative significance o f remittance-derived earnings for Afghan households and communities. For asset poor refugees with little prospect o f earning more than a minimum survival wage, the primary motivation for out-migration by many young men may have been simply to ease the burden o f consumption expenditure at home. Nevertheless, refugee and migratory movements may simultaneously have promoted access to new experience, skills, financial capital and, where remittances were possible, to sustain households and rebuilderoded assets. Urbanization has been accelerated by the refugee returns since 2002 (an estimated 2.6 million), with the majority o f returnees being childredyoung adults (with little or no experience o f agricultural management), attracted by the expansion of urban reconstruction activities, buoyant wage rates and enhancedprospects for furthering education. Around 40 percent o f all returnees to date have returned to Kabul and many more to urban centers across the co~ntry.'~ 1.10 The diversity of income generating activities is extensive in Afghanistan, and highly adapted to local conditions. For example, nomadic communities pastoral livelihoods have evolved as a strategy to handle demographic changes, climatic trends and cyclical household dynamics. In addition, the self-reliance o f Afghans, many living inremote areas, inaccessible for parts o f the year, may also explain the inability o f successive regimes to assert their absolute control over communities and local livelihood strategies. Both rural and urban households frequently engage in multiple occupations, involving multi-location residences to enhance access to opportunities for income generation and to reduce the concentration o f household consumption requirements. Nowhere is the strength to cope with externalities and the pervasive erosion o f human rights more evident than inthe daily lives o f Afghan women. On the basis o f the scant empirical work that has been undertaken in this area, it i s also apparent that the economic contribution o f female household members throughout the country, both with regard to farm and non-farm activities, should not be underestimated." 1.11 However, despite its life saving significance, the defacto unregulated self reliance o f communities has also involved negative consequences for the prospects o f recovery and reconstruction at micro and macro levels, While the role o f sharing mechanisms within kinship and community wide networks should not be underestimated in the Afghan context, nevertheless, certain community-based informal risk management strategies such as sharecropping and related credit based instruments have historically served as exploitative mechanisms, perpetuating and exacerbating inequalities between different groups. As is the case with most protracted complex emergencies, there are those that manage to retain or even increase their wealth, while others fall deeper into poverty. Growing debt burdens for poorer households have generated a windfall for informal moneylenders and land owners, yet frequently result in loss o f access to land, orchards, houses and water rights for borrowers as a result o f defaults on repayment. The risk o f exploitation is heightened during periods where public institutions are effectively non- functional and fundamental civil and humanrights are most at risk. The incidence o f widespread poverty under such conditions in the Afghan context has involved the heightened vulnerability o f marginalized groups to conscription, sexual and physical assault, arbitrary detention, indentured labor, prostitution and human trafficking. The reported defacto lowering o f the age for marriage o f daughters as a result o f household insecurity vividly highlights the way that survival mechanisms can carry unacceptable social costs, although the prevalence o f such practices remains difficult to quantify. A t the bottom line, the stark reality o f human development indicators, especially for women and children, can leave no doubt as to the personal costs bornby Afghan households. 1.12 In this context, perhaps one of the most effective and widespread household and community based strategies for risk management to emerge in Afghanistan involves the spread o f informal and illegal economic activities such as cultivation o f opium poppy. It is increasingly clear that an expanding drug based economy carries highly negative implications for the reconstruction o f the nation state, national growth and security inthe long run. At the same time, the Government is acutely aware that a counter-narcotics strategy, driven by an immediate, widespread and effective opium poppy eradication campaign, could constitute a serious shock to the livelihoods of the rural poor, thereby potentially undermining achievements inthe immediate and short term with regardto poverty and vulnerability reduction. l4TISA (2004) 15 AREU (2004), Gender RolesinAgriculture. 5 1.13 There has been substantial debate regarding the significance o f poppy cultivation as a strategy for capital accumulation and managing risk employed by the rural poor and the better off alike. On the one hand, poppy production would appear to simultaneously provide land owners, sharecroppers and landless households (through the demand for seasonal employment) with highly lucrative opportunities for income generation. However, as has also been argued, closer investigation o f who cultivates opium poppies and why, highlights that for many vulnerable landless and land poor rural households, the net financial returns are often relatively modest. For the poor, cultivation may have been undertaken primarily to gain access to land and credit to repay debt, meet minimum subsistence needs and smooth consumption. Moreover, evidence is emerging that entrenched "adverse incorporation" arrangements are not confined to the poppy economy but have spread to other crops as well as the rural labor market.I6 RecentDevelopments:ImprovingLivelihoodsand Opportunities 1.14 As a result of renewed hope for the prospect o f regaining overall enhanced peace and stability, new opportunities for growth, security and poverty reduction are emerging. There are encouraging signs o f progress. Afghanistan's economy has performed very strongly inthe past two years, with non-drug GDP increasing by more than 50% (29% in 2002 and 23% in 2003), albeit starting from a very low base. Farmers have been in a better position to take advantage o f the excellent rainfall in 2002 and 2003, resulting inbumper harvests, although 2004 was not such a good year. The last three years have also seen a rebound in opium production to near-record levels, with poppy now growing in all 34 provinces. The opium economy in 2003 comprised about a third o f total drug- inclusive GDP and accounted for 314 o f global illicit opium production. Growth in the agricultural sector, together with post war expansion in construction and commercial services has contributed to the generation o f new employment opportunities with a positive impact on wage rates. With the improved political and economic situation, an estimated 2.6 million refugees have returned to Afghanistan, and 600,000 internally displaced people have returned to their homes. 1.15 Recent good growth performance has been supported by the Government's sound macroeconomic policies -ahighlysuccessfulcurrency reforminlate2002, aprudent"no-overdraft'' policyprohibitingdomestic financing o f the budget deficit, conservative monetary policy which has brought inflationdown to around 10% per annum, and good management o f the exchange rate. Progress has also been made in mobilizing domestic revenue, which rose from negligible levels in 2001 to over 4% o f GDP in 2003; in trade reform - import duties have been rationalized and customs administration reforms are underway; and in financial sector reform where the Central Bank has been made legally autonomous and several foreign banks have set up inKabul. 1.16 Peace and strong economic performance have been accompanied by improvements in some social indicators. Restoration and expansion o f social services have been initiated, notably primary education and immunization - more than 4 million children now attend school o f which one third are girls.I7 And the end o f major conflict, combined with national programs like the National Solidarity Program (NSP), has enabled communities and other groups to begin buildingsocial capital. Despite this progress, Afghanistan remains a very poor country, with extremely low social indicators (e.g., infant mortality o f 115 and under five mortality o f 172 per 1,000 live births)and very low access to most public services (e.g., safe drinking water)." 1.17 Although double-digit economic growth is expected to have continued in2004, the drivers o frecent growth are to a considerable extent temporary and recovery-related (e.g., the boost ingrain output and construction) and are running their course. In order to achieve sustained rapid growth - necessary for poverty reduction, phasing out dependence on opium, and maintaining political buy-in for reforms - other, sustainable growth drivers will be needed. Afghanistan has areas o f good potential for growth, ranging from agricultural production and processing to mining, construction, trade, and other services. Exploiting Afghanistan's geographical position as a "land-bridge'' between Central and SouthAsia also can be a source o f growth. Enhancing the role o f women ineconomic life will helpensure that growth is broad-based. l6Mansfield(2004): Diversityand Dilemma: UnderstandingRural Livelihoods andAddressingthe Causesof opiumPoppy Cultivation inNagarharand Laghman, EastemAfghanistan. l7TISA (2004) The NationalHumanDevelopmentReportfor 2004 placesAfghanistan inthe list o flow human development countries, just above some ofthe very poorestWest African countries. 6 1.18 The challenge to reconstruct the political, physical and socio-economic infrastructure o f Afghanistan is enormous and the current Government has restricted capacity to reduce or mitigate the macro and micro risks faced by its people. At the individual and household level, the prevalence o f poverty is high, formal insurance and credit markets are virtually non-existent and social protection interventions have only recently been launched. At the national level the government has very low revenue-generating capacity and substantial dependence on international assistance is anticipated for many years to come. In certain parts o f the country (particularly the South and South East), the last couple o f years have seen heightened physical insecurities, to the point that in some cases programs have been suspended and agencies withdrawn. Ongoing political volatility has given rise to new challenges and strategic directions. Recent developments involving the engagement o f military forces inprovincial reconstruction are being closely monitored by humanitarian workers. Whilst enhanced security i s a precondition for much needed engagement with many rural communities and a fundamental requirement for improved coverage o f assistance, many observers remain concerned that such a shift may carry new risks for agency field staff, contractors and community based counterparts. 1.19 A process o fprofound economic, political and social changes is unfoldinginAfghanistan and inthis fragile post-conflict environment, Afghans remain highly vulnerable to fkture shocks. Ensuring that the opportunities generated by growth are accessible to poor women, men and children alike will be crucial if they are to impact positively on the welfare o f ordinary people. The identification o f specific groups who may require targeted support i s a growing priority for public policy. Equally important will be to support and strengthen livelihood strategies to protect gains at the household level and guard against a recurrent slide back into poverty. 7 11. RURAL LIVELIHOODS UnderstandingRuralLivelihoods 2.1 Rural livelihoods in Afghanistan are complex arrangements, with a large degree o f diversity, and with multiple strands and strategies. They are not restricted to village space. They involve agriculture - farmer-owned and worked as well as wage labor and share-cropping, off farm employment -both self-employment inthe home as well as employment elsewhere, migration - seasonal and longer-term, trading, licit and illicit, formal and informal sector activities. In order to be sustainable, livelihoods need to be able to cope with and recover from stresses and shocks, maintain or develop capabilities and assets and avoid undermining the resource base. Afghanistan's livelihoods are extraordinary well adapted to weather the many stresses and strains imposed on them, and despite large and frequent shocks, which have impoverishedpeople, mass starvationhas been avoided. 2.2 More specifically, strategies may include: opportunistic movements by household members seeking optimum employment, both farm and non farm, often to supplement micro farm production and contingent on climatic conditions and household circumstances; unwritten sharecropping contracts based on in-kind exchanges (including meals, clothing, inputs, etc.) and informal credit arrangements guided by prevailing community n o m ; unpredictable remittance flows through the traditional hawala system o f financial transfer networks, refracted through extended family relationships; the exchange of (non-monetized) labor and services among households, usually involving female household members (including children) in economically significant productive and reproductive activities, both inagricultural and non agricultural sectors; and, the increasing prevalence o f unreported illicit activities such as opium poppy production, processing and marketing. While these examples highlight the challenges o f undertaking livelihood analysis (i.e., what different people are doing and why), they also serve to underscore the potential advantages of adopting a livelihood perspective as the basis for understanding poverty and vulnerability inruralAfghanistan. 2.3 The aim o f this chapter is to characterize the importance o f various economic sectors in rural livelihoods, provide information on rural employment opportunities and explore the capacity o f individuals and households to take advantage o f various income generating activities and strategies available to them. This chapter provides an important back-drop for Chapter 3, which presents a profile o f poverty and vulnerability in rural Afghanistan. Drawing on the NRVA, a number o f usehl insights related to labor markets and income-generating opportunities in rural Afghanistan are summarized below and complemented by additional data where available. The challenge associated with classification and quantification o f livelihood strategies and labor markets, including the relative incidence and value o f activities, should not be underestimated. An initial classification o f the main components of livelihood strategies inruralAfghanistan is presented inTable 2.1. The Diversityof LivelihoodsinRuralAfghanistan 2.4 Agriculture. Contributing an estimated 53% to GDP (70% if the opium economy is included in the calculation), agriculture is central to the Afghan economy." Severely damaged infrastructure, services and investor confidence have had a negative impact on agricultural development. Trend growth in agricultural production (including all crop and livestock products) slowed dramatically from 2.2% per annum inthe pre-conflict period of 1961-78 to 0.2% per annum between 1978-2001. Cereal production actually declined at a rate o f 2% per annum between 1978-2001. Since 2002 there has been a strong recovery in agricultural productivity with the return o f normalrainfall and improved availability o f seeds and fertilizers. Total cereal production increased by 82% and 50% for 2002 and2003, respectively.*' While the precise contribution o f the agricultural economy to GDP is extremely difficult to measure (depending on exactly where the line is drawn between agricultural and non-agricultural eamings and on analysis o f the causal relationshipbetween both sectors) the apparent impact of agricultural growth due to post drought recovery combined with increased poppy production inmanypartso fthe country have had a remarkable acceleratedinfluence on the non-farmeconomy. 2o TISA (2004) 8 Table 2.1 :Classification of the M a i n Components of Livelihood Strategies inRural Afghanistan Strategy -- Examples I Self Wage 1 LicitI Illicit Employment Emnlovment Ownproduction and sale o f Agricultural commodities produced Yes cashcrops and livestock primarily for sale on the market, e.g., cotton, h i t s , timber, oil seeds, poppy, cereals, vegetables Productionand home Subsistencefarming, e.g., wheat, Yes consumption o f food crops meat and dairy, horticulture Processing and sale o f Milling o f wheat, butter narcotics, Yes agricultural commodities dried h i t s processing, etc. Trade o f agricultural Including pettytrade and firewood Yes N o Yes Yes commodities collection, etc. Exchange o f food crops in Through informal community based Yes N o Yes No kindfor localconsumption support networks, sharecropping Sale o f labor for private I Payment incash or kind,- e.g., - - N o Yes Yes Yes farm based employment livestock tending, weeding I within community harvesting, poppy based employment Public employment Onstate owned farm enterprises N o Yes Yes N o Migration - sale o f labor Seasonalor longer term, regular or N o Yes Yes Yes for farm based employment opportunistic, domestic or cross border, licit or illicit Remittance Receipt o f income by household N o Yes Yes Yes arising from migrationo fhousehold member(s)to work on farm based activities (licit or illicit) Non-agricultural production I Mining, construction, transport, Yes N o I Yes I Yes Trade o f non-agricultural I Includingpetty trade Yes N o 1 Yes Yes commodities I Informal service sector I Health services, food preparation, Yes N o I Yes III Yes enterprises bakery, hair cutting, etc. ~ Salaried employment in the Civil service, militia N o Yes Yes N o service sector Sale o f labor for non-farm Construction work, carpet weaving N o Yes Yes Yes employment. carrying water, domestic employment Migration Sale o f labor for non farm-based N o Yes Yes Yes employment (seasonal or longer term, regular or opportunistic, domestic or cross border, licit or illicit) Migration Enterprise established inplace o f Yes No Yes Yes migratory settlement Remittance Receipt o f income by household Yes Yes Yes Yes arising from migration o fhousehold member(s) to work on non farm based activities (licit or illicit) 9 2.5 The agricultural economy in Afghanistan is diverse as a result o f geographical variations. I t is estimated that only 12% o f the total land area is suitable for arable production (with 74% concentrated inthe north, northeast and west regions) o f which around half is cultivated annually.2' Wheat is the main cereal crop (68% o f total cultivated area in2002), grown by most farmers for subsistence and animal fodder, and predominantly grown inthe north, north-east and west o f the country. Most wheat production is from irrigated land with rainfedwheat area (and yield) varying considerably with seasonal conditions. Other cereals cultivated include rice, maize and barley (16% of total cultivated area in 2002). The most important non-cereals are pulses, oilseeds, fruits, vegetables, fodder and opiumpoppy.22 2.6 Land Ownership and Agriculture. National data on rural landownership is incomplete and a provisional correlation o f land ownership and welfare, based on NRVA findings is presented later in Chapter 3. It should be stressed that regional differences o f land distribution are significant, with surveys suggesting that ownership is highly skewed with an estimated 2.2% o f people owning 19% of total land (AREU, 2002).23 The land gini ~oefficient~~(a measure o f inequality) among farm households using the NRVA is 57 percent, suggesting significant inequality. Moreover, by their nature, landholding relations in Afghanistan are complex and challenge simple binary assumptions o f land ownership or landlessness. In order to understand different socio-economic groups' access to and control over assets such as land, it i s crucial to undertake analysis inthe broader context o f institutions and power relations. A basic five-fold classification has been suggested to facilitate understanding o f such relationships inthe Afghan context as outlined inFigure 2.1.25 2.7 Sharecropping. Sharecropping appears to take place inall rural villages, though to varying degrees. The number o f landowners who sharecrop out land seems to be associated with factors such as the size o f land holdings, household size and composition, and the availability o f male labor. Most o f the sharecroppers are landless or land poor (Grace 2004). The terms o f sharecropping vary primarily according to who supplies farming inputs (seeds, farming tools and oxen) and the localmarket. Where households sharecropping inland are able to supply only labor, they will receive a relatively small share o f the final harvest. Other terms exist where the sharecropper may live with the landowner and receive food and clothing (and consequently even less share o fthe final harvest). Extremely poor households may also receive an advance o f wheat seed at the outset o f the agricultural year to facilitate sharecropping contracts.26 2.8 Landlessness. Outright landlessness is a very real feature in rural Afghanistan with a significant proportion o f the rural population either not owning farmland or owning farms too small for survival (Wiley, 2003). The differential access to landbetweenhouseholds emerges as a key finding from various sources o f data, including the NRVA, which finds 24% o f households landless. It is widely reported that sale o f land was a frequent coping strategy during the drought years and that this has further contributed to the concentration o f land assets with wealthier ho~seholds.~~ 2'Ibid. 22World Bank (2004) Afghanistan- State Building, SustainingGrowthandReducingPoverty. 23There are fivejeribs to one hectare. 24A gini coefficientvariesbetween0 and 100with greater inequalitybeingcloser to 100. 25AREU (2004) RuralLivelihoodsMonitoringprogram. 26Inarecent study(Grace, 2004) peoplefrom five villagesreportedashortageof labordueto the availability ofaltemativeand more lucrativeemployment opportunitiesfor menoutside the village, bothwithin and outside Afghanistan. Many peoplewere reportedto bemovingout of sharecroppingto engage inwage employment to enhance economic retums(despitethe fact that such work is generally accordedlower socialstatus), as aresult ofthe increasingwages on offer, the relatively low price of wheat this year, and the reducednumber of oxen owned by householdsafter the drought. Some householdsappear to beinvolved in bothsharecroppingandwaged employment given the requirement to repaydebts accumulatedduring the drought. 27See for exampleMansfield(2004), CopingStrategies,AccumulatedWealth and Shifting Markets:The Story of Opium Poppy Cultivation in Badakhshan2000-2003. 10 Figure2.1: Classificationof LandHoldingz8 H o u s e h o l d s o w n i n g a n d c u l t i v a t i n g t h e i r o w n l a n d H o u s e h o l d s o w n i n g a n d c u l t i v a t i n g t h e i r o w n l a n d , i n a d d i t i o n t o s h a r e c r o p p i n g i n a d d i t i o n a l l a n d L H o u s e h o l d s o w n i n g a n d s h a r e c r o p p i n g I- o u t l a n d , e i t h e r p a r t i a l l y o r t o t a l l y E o u s e h o l d s w i t h n o l a n d m a y s h a r e c r o p i n l a n d f o r c u l t f v s t i o n R o u s e h o l d s w i t h o u t l a n d m a y n o t c u l t i v a t e a n y l a n d 2.9 The proportion o f land and landless households differs significantly from village to village (research inthe same district has identified virtually no landless insome villages while 70 percent incidence inothers) but ingeneral the prevalence o f families that can be classified as landless and land poor would appear to suggest a majority. Interestingly however, the NRVA data suggests that farming tools are widely perceived as a more significant farming constraint than landby all households, but especially the poorest. 2.10 Opiumproduction. Opiumpoppy is by far the most significant cash crop cultivated inAfghanistan. After decades o f conflict and with the collapse o f the formal economy and governance structures, opium i s now Afghanistan's leading economic activity and provides around one third o f (opium inclusive) GDP despite the fact that poppy is grown on only 1% o f arable land (UNODC). Moreover, the opium economy is expanding and poppy production now occurs inall 34 provinces in Afghanistan. Key factors drivingthis expansion include Afghanistan's natural and political comparative advantage (weak governance), high (but fluctuating) prices, rural poverty, indebtedness from the drought years and the need to re-establish livestock herdsand other assets, andpressures fIom criminal networks. Rising labor costs and initial attempts at eradication in 2003 do not appear to have reduced producer incentives with 69% o ffarmers interviewed reporting intentions to increase poppy cultivation inthe 200314 season. (UNODC, 2003):' At the end o f 2004, however, there are signs that Government efforts to contain the spread o f poppy production and support phased eradication are beginning to have impact. Early indications also suggest falling farm-gate prices. 28Inadditionto thisbroad classification, other arrangements such as germ, where land is provided as security against a loan, and the lender has the right to cultivate the land untilthe loan is fully repaid though does not claim title, and ejuru, aleasehold arrangement, also remain important inparts o frural Afghanistan. 29Almost 30% more than the previous highest (2004, US State Department). 11 Classification of Households Engaged Opium~u~tivation~' ResourcePoor Farmers Laborers 0 Opium cultivation Few, ifany assets other than labor. Highlabor intensity o f poppy i s source o fprofit and Productive assets such as livestock production (harvesting capital for other and landmay have been sold during considered a semi-skilled economic activities. past shocks, e.g., drought. activity, hence higher value Access to land, Mustwork as laborers or enter wage). water, labor and sharecropping arrangements (for A huge numbero fAfghans working capital which opium cultivation is move seasonally, following the 0 Risksassociated precondition inmany areas - on poppy harvest: an estimated with opium landlords and lenders insistence). 1.26 millionpeople take part in encourageowners to Opium has become principal means the harvest. sharecrop out land for o f accessing land and credit, which Availability o f labor is a key opium. locks poor farmers into the opium constraint for future expansion Opiumproduction economy. o fopium cultivation. does not appear to be Opiumcultivation is ameansof Women can earn cash from crowding out wheat survival for the resource poor farmer opium too, both by providing production as the area -higher prices may have little labor, and from the paschin, the under both crops has impact onhouseholdincome. Net right to the last pass through the risen inrecent years. returns for sharecroppers only a poppy field at harvest time. fraction o f theoretically high returns. Cash income contributes to purchase o f (affordable and available) food from market (majority o f poor households are net food consumers). Many resource poor households, negatively affectedby drought, have a debt overhang and need to access producer/consumer credit. Isalaam credit (i.e., advance against crop) is available for opiumproducers. 2.11 Overall, opium poppy presents itself as a wonder crop for Afghanistan's otherwise devastated rural economy. Poppy has a higher gross value o f output per hectare than any other crop inAfghanistan. While cost data are limited, it would appear that while h i t and horticultural produce are more profitable than wheat, they cannot compete with poppy. In 2000 opium poppy was estimated to be 8 times the value o f wheat (ICARDA, 2002) and this may have risen to 28 times that o f wheat by 2003. (Molla, 2003). Due to high profitability, relative drought resistance as a crop, nonperishability o f the commodity, poppy is therefore perceived as a low risklivelihood option by a significant proportion o f rural farmers and the cultivation of opium has emerged as a means o f survival for many rural communities. 30Based on AREU Rural Livelihoods MonitoringProgram. 12 2.12 NRVA data confirmthat households involved inpoppy production tendto be better offthan those who are not. Additionally, households in communities where poppy production is prevalent were more likely to have improved welfare, highlighting the positive impact o f poppy on the local economy. A significant proportion o f the Afghan population benefits fiom the production, harvesting, processing and trafficking o f opium. Poppy is a highly labor intensive crop requiring around 350 person daysiha. (compared to 41 days per hectare for wheat). Overall it is estimated that production alone may involve 1.7 million people (around 7% o f the total population), while several millionparticipate overall inprofits derived fiom the opium economy, ina broad network o f protection and pay offs, therefore highlighting the significance o f the poppy contribution to household income at multiple levels (World Bank, 2004). Women also appear to make up a significant component o f the opium labor force and the impact o f such opportunities on gender relations and empowerment has yet to be studied indetail (See Box 2.3 below). 31This analysisdraws mainly on the report of Mansfield, D. (2004), Diversity andDilemma 13 2.13 Multiple linkages o f the opium economy to the non-drug economy are also apparent with regard to access to finance and credit. The opium economy provides a considerable amount o f financing to the rural economy, and is therefore attractive to farmers with few other credit options. However, implicit interest rates appear to be high and with advancesrequired inopium, the long term dependency o n opiumproduction is e~acerbated.~~ 2.14 Livestock activities. Livestock is a major source o f food and income for Afghan farmers, traditionally being an integral part o f most farming systems inthe country. Range pasture land covering about 45% o f total land area has Table 2.3: Livestock Population Per Family traditionally supported a large livestock population, with 1995 1998 2003 animals being the only source o f income for some groups. Cattle 3.7 2.5 1.2 Livestock population declined inthe 1980sbut had recovered 21.9 14.2 2.9 to pre-war levels by the mid-1990s. However, the years o f Goats 9.4 5.8 2.4 Donkeys 1.1 0.7 0.5 drought from 1999-2001 have had a serious adverse impact. Camels 0.4 0.2 0.1 As per a 2003 census (MAAWFAO) the estimated livestock poul 11.6 6.8 4.0 population is about 34 million (including 12 million Of Source: F z , National Livestock Census, InterimReport,2003. poultry), which is a sizeable decline from estimates in the 1990s. B y 2003 the number o f cattle and poultry per family had fallen to one-third of the 1995 figure, with numbers o f sheep and goats per family suffering an even steeper decline (Table 2.3). 32Not only does opiumbring in one third ofthe total nationalincome, but opiumrevenuesinthe hands offarmers, laborersand pettyagents are highly likely to be spent on consumption goodsproduceddomestically,as well as investmentssuchas housing. 14 2.15 Historically, livestock are known to have been an integral component o f farming systems throughout much o f the country. With regard to the K ~ c h i there~ are the Pashtun pastoral nomads whose migratory patterns take , ~ them from the south and the east (inthe winter months) to the central highlands for differing lengths o f time inthe summer months, in accordance with their reliance on pastoral nomadism versus other sedentary sources o f incomes such as agriculture and trade. Similar patterns o f nomadism can be found inthe north among the Turkmen and the Arabs. Then there are other livestock producers, mainly inthe north among the Uzbeks and the Turkmen, who own large flocks o f livestock and settle in villages in the winter and move with their livestock to pasture lands in the summer months. These different categories o f pastoral nomads have their own specific problems today due to the depletion of their flocks, widespread disputes over landrights and failing access to grazing lands. At the time o f the NRVA and NMAK surveys (2003 and 2004), a proportion o f Kuchi households are reported to have abandoned their migratory way o f life as a result o f the 1999-2001 drought and the associated death or distress sales o f all livestock assets. Out o f a Kuchi population o f 2.4 million, around one million had ceased to migrate. While some have adopted alternative livelihoods in both the rural and urban areas, others are currently surviving within IDP camps. 2.16 Non-agricultural activities. The non-agricultural sector is often suggested as a potential direction for poverty reduction, and empirical literature indicates a strong relationship between the First, by diversifying their sources o f income, rural households can augment their incomes and minimize the extent to which they are affected by adverse income shocks from farm activities. Second, as non-agricultural incomes increase the households' cashliquidity, they canbuy farm inputs more easily, thus increasingtheir farm productivity. Finally, the non-agricultural sector offers landless households (which are unable to engage in farm activities because o f their lack o f land) another option for generating income. 2.17 The rural labor market in Afghanistan is characterized by: (i) a high degree o f fragmentation due to 33The literalmeaningofKuchi(toshift) relatesto the regularmovementofhouseholdsandhouseholdmembers. 34For arecentsurvey, see Lanjouwand Lanjouw(2001). 15 geographic, ethnic, cultural, political and socio-economic diversity; (ii) seasonal and market driven dynamics strong (such as reduced employment opportunities and declining wage rates during the winter season); and (iii) a high dependence on the informal (and frequently insecure) sector for incomes and employment. Rural non-farm employment inAfghanistan is also highly diverse, both with regardto the type o f economic activities and the returns available. At one end o f the spectrum returns can be high (e.g., enterprises involving significant investments in technology), while at the other end, where entry barriers are low and the poorest quintile are mainly represented, traditional handicraft skills, petty trading and manual /domestic labor are very poorly paid and the returns highly irregular. 2.18 Non-farm employment (as with farm employment) is therefore characterized by a high degree o f structural inequality, with many activities undertaken by households inrural areas generating low and highly insecure incomes (often with associated conditions that fail to meet basic human rights). Migration from isolated villages, while providing a crucial livelihood strategy, can also contribute to the acceleration o f social change, impacting in many ways on intra-household and intra-community relations. However, within the context o f Afghanistan the overall significance o f non-farm employment for poor households in rural Afghanistan is clearly apparent. Non-farm activities with a secondary or tertiary linkage to agricultural production (processing, marketing, etc.) also have a critical role to play inadding value to and stimulating investment inthe farm sector. Such activities typically require a start up capital investment and therefore tend to be initiated by higher income groups, but nevertheless remain key to promoting development o f the broader rural economy through cross sectoral synergies, the provision o f opportunities for skill acquisition and enhancedentry to new markets. Insightsfrom the NRVA: OccupationalStructure 2.19 Complementing the above, the NRVA suggests that the labor market in rural Afghanistan is characterized by a largely informal nature as well as its diversity in agriculture and non-agricultural opportunities. Almost 75 percent of the active individuals in the NRVA are self-employed, while another 22 percent work in small private firms, typically with fewer than five employees (Table 2.4). Similarly, more than 40 percent o f employed individuals are daily workers, while only 10 percent are salaried workers. Only 5 percent o f the rural sample works inthepublic sector. 2.20 Agriculture plays a major role interms o f employment inrural Afghanistan. Just over half of the employed individuals in the NRVA work in agriculture, either as subsistence farmers or as agricultural laborers (Table 2.4). Examining agricultural employment rates across consumption q ~ i n t i l e ssuggests at first that there i s not a ~ ~ systematic difference in the rates between the poorer and richer quintiles. Nonetheless, dividing the agricultural employment between labor (wage earners) and farming (self-employed) shows that there is a strong positive relationship between agricultural wage employment and poverty and an inverse relationship between farming and poverty. Specifically, almost 30 percent o f households in the lowest consumption quintile work in daily wage agricultural jobs, compared to only 15 percent among households in the highest consumption quintile (Table 2.5). Similarly, while only 20 percent o f the poorest households fann, almost twice as many do so among households in the highest quintile. These pattems may suggest the significance o f asset ownership (land, livestock, tools, etc.) in determining the extent to which agricultural activities contribute to poverty reduction and enhanced food security. They also indicate that further analysis of the potential constraints relatedto agricultural incomes is important. 35 A standardapproach to identify characteristics associated with poverty is to divide the sample o fhouseholds surveyed into "quintiles" - i.e., five groups each comprising 20% of the sample -based on aselected welfare indicator. For a detailed explanation o fhis methodology, see Appendix 2. 16 Table 2.4: OccupationalStructure by Food ConsumptionQuintiles 4 1 4 2 43 4 4 45 All active (better (poorest) off) people Type of Work (YO) Public 4 4 6 6 7 5 Private 29 24 20 20 15 22 Self-employed 67 72 74 74 78 73 Total 100 100 100 100 100 100 Type of Activity (YO) Agriculture 51 54 54 53 53 53 Manufacturing (textilekrafts) 10 7 5 6 5 7 Construction 12 14 14 16 12 14 Trade 2 4 5 6 9 6 EducationlHealth 3 3 4 5 6 4 Other (administrative, mining, transport) 22 18 18 14 15 16 Total 100 100 100 100 100 100 Type of payment(YO) Salary 7 8 10 10 14 10 Daily wage 51 42 38 37 31 40 Self-employed 34 42 48 50 53 45 Paid in-kind 7 6 3 2 1 4 Unpaid 1 2 1 1 1 1 Total 100 100 100 100 100 100 Note: Active individuals > 6 years age. Source: NRVA2003. Table 2.5: Employment by ConsumptionQuintiles 4 1 (Lowest) 4 2 43 4 4 45 All (Highest) workers Non-agricultural wage 39 38 34 37 36 37 Non-agricultural self-employed 9 9 12 10 11 10 Agricultural wage 30 23 19 16 15 20 Farming 22 30 35 37 38 33 Total 100 100 100 100 100 100 Source: NRVA 2003. Note: active individuals > 6 years. 2.21 The NRVA data reveal that the rural non-agricultural sector in Afghanistan is also significant and diverse, Inparticular, almost half o f the rural employed population in the NRVA works invarious non-agricultural activities: 14 percent in construction, 7 percent in manufacturing o f labor intensive activities and 6 percent in trade (Table 2.4). Various other sectors such as transport, miningor other services (unspecified) also exist and employ the remaining 20 percent o f the rural employed population. More than 35 percent o f the non-agricultural employment is waged, while around 10 percent o f the NRVA sample individuals are entrepreneurs - this is the reverse pattern from agriculture (Table 2.5). 17 Figure2.2: EmploymentDistributionby Sector andWelfareGroup (%) 100% 80% 60% 40% 20% 0% Q1 (Lowest) Q2 Q3 Q4 Q5 (Highest) Source: NRVA 2003. 2.22 The role o f both the agricultural and non-agricultural sectors is significant among all levels o f welfare (consumption groups). Infact both farm and non-farm sectors are remarkably similar across the groups. (Figure 2.2) The overall reported level o f household employment derived from the agricultural sector is 53%, ranging between 52% and 54% across the quintiles. Likewise, between 46% and 49% o f all rural employment was reported to be derived from non agricultural activities - either through self-employment or wage based labor.36Furthermore, the NRVA data highlight that the level o femployment derived specifically from own farmproductionis below 40% for all quintiles (only 22% for the poorest quintile). The single most significant non-agricultural activity across all consumption quintiles is construction. Trade is more likely to be associated with better-off households, (only 2% o f the poorest quintile work inthe trade and commerce sectors as opposed to 9% of the least poor quintile) and work in manufacturing (textileslcrafts) with poorer households. Better-off households are twice as likely to work in health and education services. The StructureofWages 2.23 While it may be possible to record the incidence o f such employment activities, it is extremely difficult to measure the significance o f income arising from various strategies in such a context and to gauge the returns to investments (both crucial determinants for livelihood choices). Information from WFP and a number o f NGOs suggestedthat wage rates for unskilled labor during the winter months was as low as 50 Afs. ($l)/day in some parts o f the country one to two years ago, while the poppy harvest season, involving semi skilled agricultural labor inputs (and frequently reported to attract even teachers and civil servants), can see the rise o f wages up to Afs. 500($10)/day. The reported range o f incomes arising from carpet weaving and other handicrafts (such as hat making), often undertaken by women, involves daily earnings that are commonly less than Afs. 50lday and can even be as low as Afs. 101day(Tables 2.6 and 2.7). 36An earlier assessment ofthe contribution ofagriculture sector opportunities for domestic employment had suggested that an estimated 67% o f the labor force were engaged in farm-based activities. FAO, (2003) Compendium o f Food and Agricultural Indicators. 18 Assumptions: $1=49Afs.One month=20 working days. Source: World Bank staffestimates. 2.24 The data in Table 2.7 highlight the regional variations in remuneration for similar activities across the country. For wages paid to men, women and children, for both farm and non- farm activities, a broadpattern can be discerned. Wages in the North East and South o f the country tend to exceed those in other areas, and wages in the West are relatively lower. The fact that farm and non-farm wage levels appear to be positively correlated in all regions may also suggest that the impact o f opium poppy on labor markets can apply to both farm and non-farm activities. 2.25 An analysis o f the purchasingpower of rural wages (based crudely on the ratio o f daily wages to the price o f wheat) suggests that there has been a general increase inpurchasing power. The average rise across regions was from a range o f 7 - 9 kg. of wheat per daily wage (2002) to 14 - 15 kg. o f wheat per day in 2004. The wheat equivalent o f agricultural wages rose especially fast in the North and Northeast in 2003, with an increase from around 5-10 kg./day (2002) to 20-22 kg.1day (early 2004). These figures reflect the broader trend o f increased wages and relatively stable wheat prices in 2002/2003/early 2004, but with a reversal in the latter part o f 2004 as wheat prices have risen. 2.26 The differential between payment levels for the labor provided by men, women and children is significant. That women appear to be less well paid than children for some activities, such as planting, handicrafts, weaving, domestic labor and other, unspecified activities is o f particular note.40 The social stigma associated with the employment o f women (as an indicator o f household hardship) may be a factor underlying this pattern. (This i s discussed inmore detail below.) 37Datafor 2004 suggest that mid-seasonunskilledcasual labor wages are retaininga similar value to high seasonwages, suggestingless seasonality inwage rates. 38These refer to cash wages only. In-kind payment such as mealswould beover and abovethese amounts. 39Maletta(2003) 40Harvestwages representthe peak agriculturalwages, with poppy beingby far the highest paid activity ($lO-l2/day). Other farm work wages (for men) range between$1.42/dayfor ashepherdto $1.69 for planting. 19 Table 2.8: Daily Wages for Men, Women and Children ($/day) Source: Maletta (2003). Migration and Remittances 2.27 Livelihood strategies in rural Afghanistan are spatially dispersed. It is increasingly common for a household member to be physically absent from the household for significant periods (often more than three months) of the year. Whether migration strategies are seen ina positive light as ex-ante risk management strategies, or in a negative light as potentially disruptive ex-post coping strategies, depends on a number o f assumptions. 4'Maletta(2003) "Winters of Discontent". 20 Nevertheless it needs to be recognized that Afghanistan's strategic location as a land bridge, together with the in- country diversity o f agro-ecological zones, implies that population movements should not automatically be associated with shock induced displacement and enforced rehgee status. Where the returns are favorable, migration opportunities can often provide a crucial source o f income for investment in both household welfare and farm or non-farm enterprises that may ultimately be instrumentalinmoving a household out o fpoverty. 2.28 The NRVA data suggest that 22 percent o f households reported having a household member who had migrated over the last year (Table 2.9), and over the past five years, more than one third o f the sample had experienced migration. There did not appear to be a significant difference betweenpoor and non-poor households in the reliance on migration strategies, again implying significant diversity o f opportunities for different socio- economic groups. Seventy-eight percent o f the individuals that migrated reported that this was due either to shortage o f employment opportunities or household income. Of those households reporting migration destinations inthe last year, around one half involved cross border movement (primarily Pakistan and Iran), and the other half was within Afghanistan. Of significance, is the strong prevalence o f external migration among the upper income quintile (both cause and effect), whereas for the other quintiles, within country migration dominates. There is considerable variation inmigrationpatterns within Afghanistan: as a region, the South dominates among migratory households. Table 2.9: Migration and Poverty, Afghanistan 2003 4 1 42 43 4 4 45 AllRural (Lowest) (Highest) Household member migrated last year (%) 21 21 25 23 20 22 Main destination Other parts Afghanistan 51 53 56 52 28 48 Abroad 49 47 44 48 72 52 Total 100 100 100 100 100 100 Main reasonfor migration Insufficient income 26 42 33 23 28 31 Insufficient work 52 41 47 58 53 50 Study 1 2 3 7 6 3 Marriage 2 1 1 5 3 3 Sickness 4 1 1 1 1 1 Lack o f security 9 7 7 2 4 6 Other 6 6 8 4 5 6 Total 100 100 100 100 100 100 Source: NRVA 2003. 2.29 Local migration by male household members often involves their return every two weeks or so with household provisions. One recent study found that the remittances sent back to villages by young men working in Iran provide a significant source o f income for many households. Another study undertaken in the same area noted an increasing trend for men to seek work locally rather than migrating to Iranas a result o f the greater availability of attractive employment opportunities closer to home. However, it was noted inall villages participating inthe study that the overall incidence o f labor migration, both within Afghanistan and cross border, had significantly increased inrecent years. This could be attributable to a number of factors including (i)displacement and refugee experience over the past 20 years resultinginthe emergence o f a class o f young men less likely to be exposed to agricultural skills and more pre-disposed to urban lifestyles together with wider spatial coverage o f informal social capital networks (Box 2.4); (ii) greater population pressure on agricultural resources; (iii)concentration o f land resources among fewer households as a result o f continued asset depletion o f the poor; and (iv) depleted livestock population after successive droughts. Where workfare interventions are being implemented, they are a factor in reducing the probability o fhousehold members migrating (AREU/Grace 2004). 21 Gender Roles and Rural Livelihoods43 2.30 A more detailed analysis o f gender roles within rural households and inthe context o f livelihood strategies can provide important insights into the constraints faced by different categories o f women and men, and different household compositions. In view of the scarcity o f historical and empirical data regarding the livelihoods of Afghans, it is hardly surprising that there is little comprehensive analysis o f the roles that different groups of women and men play in productive activities, both farm and non-farm based. Despite the interest in gender politics associated with the fall of the Taliban, few studies have been conducted o n the specifics o f gender roles in Afghanistan and the focal unit o f most assessment, policy and programming activities remains the household. On the basis o f international experience, it is clear that such a focus can severely restrict understanding of the extent to 42Stigter (2005) TransnationalNetworks and Migration from Herat to Iran (AREU, Case Study). 43For further insights into gender issues, see the World Bank's Gender Assessment for Afghanistan (draft, 2004) 22 which individuals are able to access different livelihood opportunities. However, more recently, a preliminary study of gender roles undertaken in rural Afghanistan has attempted to generate increased awareness and interest in this area, given its critical significance for policy reform and program design (AREU, 2004). The main findings o f the study are reviewedbelow.44 2.31 Landowning Households. Where landholdings are reported as part o f a household's asset base, the property rights o f all household members should not be automatically assumed. The findings highlight that women inthe villages studied rarely inherit land inpractice despite the fact that under Shari'a law they do have property rights. While both unmarried and marriedwomen alike may be able to access the land and the relatedproduce, they do not have full control o f land assets. While widows may retain rights to sharecrop out their land (mostly to male relatives) insome parts o f the country, rights o f ownership are rarely inherited from their deceasedhusbands (as it is intended as the inheritance for male relatives). 2.32 Landowning households are generally wealthier than those without land. The extent to which women from fanning households engage directly inagricultural production was found to relate significantly to the relative wealth and status o f the household, the availability o f male labor as well as the age o f the women. The study identified a degree o f cultural stigma attached to women working on the land, but this varies from place to place. However, it is known that women are involved in agricultural activities, whether crops such as wheat, or less visible kitchen gardens and home-based livestock rearing that make a significant contribution to the rural economy. Women from poorer households are relatively more likely to be engagedinagricultural activities. 2.33 Giventhe significance of mobility restrictions for Afghan women, where households had vegetable plots close to home and where agricultural activities could be undertaken within the house (e.g., seed preparation and processing activities), women were often found to be active in related activities. In some cases, where households had no sons to undertake male labor within the household, women were occasionally found to be involved inmany stages o f agriculture (including sewing, harvesting and processing). Therefore it would be appear that adherence to conventional gender roles is to some extent dependent upon household assets. The market sale o f agricultural produce however, was found to be primarily the domain o f men. 2.34 Sharecropping Households. All sharecroppers appear to be male given the requirement to demonstrate fanning experience and expertise as a sharecropper. The reported incidence o f increased labor shortages within villages and the resulting trend for households to s h f t from sharecropping to higher return cash based wage employment may have implications for the extent to which women are engaged inagricultural production. 2.35 Households with Livestock. Women play a prominent role in livestock production, not only in the by- products such as dairy products and wool, but also incaring for animals and the collection o f fodder. But the extent to which women have de facto ownership rights over livestock is limited. As discussed above, households with livestock often employ seasonal migration pattern (during the spring) to access pasture, and both women and men are involved in migration. The decision regarding who migrates and who stays at home is based on a context specific assessment. For example, younger women may stay at home to engage in carpet weaving activities. Where women migrate, they are involved in activities such as the construction o f mobile shelters, the drying o f skins and the processing o f dairy products. 2.36 Households and Farm Labor. Women were found to be working in h i t production during the spring season, guarding, weeding or harvesting crops. In post-harvest activities, women play a very significant role with rice, horticulture products, fruits, dairy and other livestock relatedproducts. 2.37 Households and Non-Farm Labor. It may be too early to asses the impact o f increased wages for employment opportunities on gender roles and gender relations. While men are working away from the household women may rely significantly on other male relatives to carry out productive activities. Women's activities innon- farm labor primarily involve carpet weaving, sewing and tailoring, although activities can differ between villages. Carpet weaving has taken on a heightened significance since the drought as many households are now more dependent on this as a source o f income and had significant debts to repay.45Girls were reported to begin carpet 44Grace(AREU, 2004). Gender Roles inAgriculture: Case Studies of Five Villages inNorthemAfghanistan. 45Often advancesofbetweenUS$50-100are providedby carpet traders, dependingon the size of the carpet. 23 weaving from as young as eleven years old and for some landless households, carpet weaving was reported to be the primary source o fincome. 2.38 Two key factors appear to influence gender dynamics in Afghanistan: (i) restricted mobility outside the domestic compound, and (ii) restricted formal entitlements/control with regard to productive assets. These variables vary significantly both within and between households depending, for example, on overall livelihood assets and opportunities and the age o f female household members. Empirical analysis highlights the extent to which cultural ascriptions regarding gender roles differ significantly in varying contexts as well as being dynamic over time, with economic hardship forcing many households to risk social standing inorder to survive and retain productive assets. DeterminantsofRuralPortfolios 2.39 Analysis of rural livelihood strategies highlights a remarkable degree o f diversity among both poor and non-poor households. This diversity canbe identified between (i) agro-ecological zones, (ii) households living inthe same community with different attributes and assets, and (iii) household members. The findings and pattems from the previous sections raise a number o f questions related to opportunities and constraints facing rural households. Findings from the NRVA suggest the following: 0 Participation in agricultural wage employment is in many ways a job o f last resort. Besides participation inthe poppy sector, agricultural wages are the lowest in rural Afghanistan, inaddition to being characterized by a high degree o f insecurity and seasonality. Indeed, the correlates o f agriculture wage earners suggest that thesejobs are strongly associated with extremely poor households that have very few assets, low human capital base and minimal access to basic services such as electricity (Tables 2.10 and 2.12). In addition, the probability to work in these jobs is positively correlated with households that reside in rainfed agro-ecological areas, suggesting the significance o f regional marginality. While these results cannot be interpreted as causal, they indicate that the lack of availability o f other opportunities may be a key impediment in allowing people to take advantage o f other higher return opportunities. 0 For the most part, households who diversify their incomes from various sources are less likely to be poor than those who do not. But households who have diversified income portfolios across both agricultural and non-agricultural wage employment have high poverty rates (Table 2.10). This again underscores that dependenceuponwage labor typically produces low incomes. The NRVA includes some limited information on farming constraints. In particular, a third o f the households report that lack o f irrigation is a key agricultural constraint (Table 2.11). Access to water is one o f the most crucial factors inrural Afghanistan. Inaddition, lack o f assets such as land and tools are also reported as important constraints to farming (18 and 26 percent, respectively). Further, lack o f access to inputs like seed and credit seem to impede households from engaging in more intensive agricultural production. These pattems do suggest the existence o f many market failures in rural Afghanistan although the data does not allow for analysis o f whether this is primarily a problem of effective demand or supply 24 Table 2.10: Major Sources of Income and Poverty Below food Household income derived from (major source): expenditure threshold* Agriculture Non-agriculture Wage Farmer Wage Self-employed YO Yes No Yes No 62 Yes No No No 59 N o No Yes No 50 Yes Yes No No 48 No No Yes Yes 43 No Yes No Yes 43 No Yes Yes No 42 No Yes Yes Yes 41 No Yes No No 41 No N o No Yes * OverallNRVA 35 povertyrate, based on food consumption, is 48%. Source:NRVA 2003. e The NRVA indicates that there is a mix o f low and high return opportunities in the non-agricultural sector (Table 2.4 and 2.5). While the high return opportunities tend to be more accessible by those households with a strong asset base, lower return non-agricultural opportunities may nevertheless provide a crucial poverty reducing option for rural households, providing them multiple potential sources for daily andor seasonal incomes to meetbasic needs. In addition to the gender differences, education can play an important role in accessing wage employment in the non-agricultural sector (Table 2.12). A literate person has an eight percent higher probabilityo f finding ajob inthe non-agricultural wage employment sector. Reflecting the general state o f the rural economy, the NRVA collected information on small enterprises (employing less than 5 people). The most important attributes correlated to these non- agricultural firms are those related to assets and access infrastructure: individuals with larger assets are more likely to own a firm. In addition, access to basic services such as electricity is also positively correlated to non-agricultural self-employment. Furthermore, geographical attributes such as the existence o f a market in the community suggest that economic integration can enhance access to non- agricultural self-employment. e Finally, the existence o f poppy producers within the community is positively correlated with higher non-agricultural self-employment. One hypothesis for this is that the higher incomes associated with the poppy industry may provide enhanced opportunities for local expansion o f non-agricultural activities. 25 Table 2.11: Farm Constraints, Rural Afghanistan 2003 Q1(Lowest) 4 2 43 44 QS (Highest) All households Inigationlaccess to water 24 32 34 35 29 31 Farming tools 37 27 24 20 23 26 Land 19 19 21 16 15 18 Seed 8 10 14 19 20 14 Rainfall 4 5 3 5 6 5 Credit 4 4 2 1 2 3 Fertilizers 3 2 2 3 4 3 Labor 1 1 0 1 1 1 Total 100 100 100 100 100 100 Source:NRVA 2003. 2.40 To summarize, the employment model estimates (presented in Table 2.12) provide a number o f useful insights. In particular, by distinguishing among various employment activities, the empirical analysis suggests a number of areas where potential constraints exist inaccessingvarious employment opportunities: (i) importance the o f human capital and asset accumulation in accessing non-agricultural employment; (ii) the role that geography, access to infrastructure and basic services play in allowing rural households to engage in non-agricultural self- employment; (iii) the high correlation between women and non-agricultural wage occupation^^^, and (iv) the strong influence o f the poppy industry inlocal employment dynamics inthe non-agricultural self-employment sector.47 46Recallingthe caveat that women's actual involvementinagriculturalproductionis less likely to be recognized(and is generallynot waged) thereforemore likely to be significantly under-reported. 47Nonetheless,much of this analysis is indicativeanddoes not allow causalinferences. Despite the uniquenessof the NRVA as one of few data instrumentsfor rural Afghanistan, better data suchas longitudinal datawould benecessaryto fully understand the dynamicsof ruralpovertyand livelihoods as well as explorethepotentialdirect and indirect linkagesbetweenagricultural andnon-agriculturalsector. For example, exploringfurther the region-specificdynamics, distinguishinghrtherbetweenlow and high retumnon-agriculturalemployment or analyzingthe intra-householdmulti-occupationaldynamicsmay allow additional insightsin facilitatingthe design of aconcisepolicy direction interms of regional investmentsand interventionsby forming specificpriorities andtargetsfor rural developmentthat can acceleratepovertyreduction. 26 Table 2.12: UnderstandingRural EmploymentChoices, (ProbitEstimations) Agriculture Non-agricultural Wage Farmer Wage Self-employed Householddemographics Household size -0.0105*** 0.0079*** -0.0001 0.0007 Number o f Children <6 0.0049 0.0051 -0.0185*** 0.0077*** Number o f Children 7-16 0.0068** 0.0001 -0.0106*** 0.0047*** Number o f Adults >60 -0.004 -0.0007 -0.0078 0.0086** Householdheadcharacteristics Female household head(yes=l) -0.0427*** -0.1135*** 0.1429*** 0.0054 Household head age (years) 0.0006* -0.0003 -0.0004 -0.0001 Household head literate (yes=l) -0.0650*** -0.0113 0.0789*** -0.0024 Householdassets and other characteristics Ownhouse (yes=l) -0.0137 0.1129*** -0.0520*** -0.0313*** Number o f rooms -0.0322*** 0.0276*** -0.0086*** 0.0062*** Rainfedland tjeribs) -0.0038*** 0.0028*** -0.0009 0.0011*** Irrigatedland tjeribs) -0.0001* 0.00001 -0.0001 0.0001*** Number o f cows owned -0.0008 0.0012 0.0001 0.00001 Number o f poultry owned 0.00001 0.0012** -0.0023*** 0.0001 Access to electricity (yes=l) -0.0528*** -0.0242** 0.0226** 0.0398*** Community and district characteristics Community population (#) 0.0001 0.00001*** -0.0001*** 0.00001 Agro-ecological zone Both irrigate and rainfed 0.0043 -0.0108 0.0131 -0.0008 Rainfed only 0.0248** 0.0182 -0.0618*** 0.0207*** Grazing only 0.0209 -0.0824*** 0.0157 0.0424*** Market incommunity -0.0049 -0.0383*** 0.0315* 0.0165* Food for work program incommunity -0.0047 -0.0577*** 0.0809*** -0.0120* Cash for work program incommunity -0.0270* 0.0202 -0.0165 0.0243** Food aid program incommunity -0.002 -0.0146 -0.0092 0.0187* Poppy produced indistrict 0.0117 -0.001 -0.0225** 0.0089* Observations 18987 18987 18987 18987 Log likelihood -9836 -10577 -11832 -5459 Fit(adjusted correct predictions count) 0.76 0.72 0.64 0.91 Dependent variables: Non-farm employment choice injob designated by column heading. Comparison household demographic category is adults 16-60. Comparison agro-ecologicalzone is "irrigated". Comparisonregion i s Central. Marginal effects reported. Significant levels:* = 90%, ** = 95%, *** =99% 27 111. A PROFILE OFPOVERTYAND VULNERABILITY INRURAL AFGHANISTAN 3.1 The previous chapter described the way in which people in rural Afghanistan make a living. A picture emerges o f poor people, often struggling to stay alive, but at the same time seeking opportunity and making gains wherever possible. This chapter, by way o f complementing what has already been presented, details a profile o f poverty and vulnerability inrural Afghanistan. Itpresents a snapshot intime -the summer o f 2003 when the NRVA was carried out -o f the situation o f 11,200 households living throughout the country.48 Poverty 3.2 A standard approach to identifying characteristics associated with poverty is to divide the sample o f households surveyed into "quintiles" - i.e., five groups each comprising 20% o f the sample - based on a selected indicator o f welfare. Here we use total food expenditure per capita as the welfare measure. Data from the NRVA suggest that close to half o f the population in the sample o f 11,200 households had food consumption expenditure (inthe summer of 2003) that fell below the consumption threshold, that is the cost ofpurchasing 2100 calorieslday o f the typical (to rural Afghanistan) basket of food items consumed by the poorest household^.^^ Since the survey was carried out at the time o f a bumper harvest, when domestic production o f wheat was at an all-time high, and incomes were buoyant, it would be expected that the year-round average percentage o f households inthis situation would be higher. Table 3.1: RuralPovertybyRegion: NRVA sample YOof sample `YOof people Measurementof population fallingbelow the how far people per capitafood fall belowthe expenditure threshold(P2)* threshold Overall sample 100 48 7 Regions: North-east 11 33 5 South-east 12 39 4 south 19 42 4 Central 15 43 5 East 14 47 4 West 10 62 11 North 16 66 12 West-central 3 71 17 * A larger number suggests more "deep" poverty. See Appendix I1for details. Source: NRVA 2003. 3.3 Poverty i s widespread throughout rural Afghanistan, but there appear some variations across geographical areas. On the basis o f total food expenditure per capita and actual patterns o f calorie consumption (quantity of calories and dietary patterns5'), the west, north-west and eastern parts o f the country seem poorer, while the south and the northeast seem relatively better off. (Table 3.1 and Figure 3.l).51 Notwithstanding the limitations o f the data, similar patterns o f variance at the regional level are fkrther reproduced at the provincial level, district level, village level and within villages, underscoringthe heterogeneous nature o frural livelihoods (discussed inChapter 2) and the variance inwelfare outcomes over both short and longer distances.52 48Although a large sample and with sufficient numbers o f observations to make inference into the situation o f rural households, the NRVA should not in any way be considered statistically representative o f all rural Afghanistan. 49 An individual is classified as poor ifhis annual food consumption is below the poverty line, defined as the annual cost o f consuming 2100 calories/day o f the typical basket o f food items among the poorest. For more details see Appendix 11. See Appendix 111for details on calorie consumption and Figure3.3. for dietary patterns. These rates are only suggestive since the data are not representative at the regional level (nor at the provincial level). j2 See, for example, "Three villages inAlingar, Lagham: A case study o f rural livelihoods", AREU, Kabul, November 2003 28 Figure 3.1: Rural Poverty, by Regionss3 Regional p o v e r i y rates 0 < 3 5 3 5 . 4 5 B 4 5 . 5 5 5 5 . 6 5 > 6 5 Source: NRVA 2003. 3.4 When asked about their ability to meet their food needs over the last 12 months prior to the survey, about a third o f the rural sample o f households claimed they "usually" had problems meeting those needs, while another third "sometimes" had food security problems (Figure 3.2). Rather predictably, poor households were more likely to have problems meeting food requirements compared to the non-poor or better o f f food consumption quintiles. These results are consistent with findings o n food security obtained from the other modules o f the NRVA and do indicate that rural food security remains a critical ~ o n c e m . ~ ~ The Correlates of Poverty 3.5 Poverty is multidimensional, involving a complex interplay between l o w assets (physical, financial and human), years o f insecurity and drought, indebtedness, poor infrastructure and public services, traditional roles and other constraints. As such, exploring its causes and correlates requires an analytical approach that examines these many dimensions both individually and simultaneously (to capture the interactions). This section summarizes the patterns emerging when exploring the correlates between poverty quintiles, as measured by food consumption, and a number o f household demographic, social and regional characteristics (Table 3.2). The main objective o f this exercise i s to capture some o f the most important attributes that seem to be related with higher or lower poverty. One important caveat is that this analysis does not capture the causes o f poverty per se but rather offers an understanding o f the variation infood (expenditure) consumption due to various factors. 53The poverty estimates are not statistically representative of the regions. As such, they can only be used for qualitative comparisons. 54MRRDand WFP. (2004) Report on Findings from the 2003 National Risk and Vulnerability Assessment inRural Afghanistan. 29 Figure 3.2: Rural Food Security "Problems SatisfyingFood Needs Duringthe Last 12 months?" 60% 50% +-r 30% 20% 10% 0% Q1 (low) Q2 Q3 Q4 Q5 (high) 1 Rarely ESometimes Often1 Source: NRVA 2003. 3.6 Household size and composition. The average household in the NRVA sample comprises 7.5 persons, comprising four children, and three to four adults. While there is little variation in household size across consumption quintiles, household composition does differ: the better off households have few children and more adults, suggesting a possible link between income generation capacity and dependency, although in the Afghan context children are a source o f income as well as dependents. 3.7 Female-headed households. There is a strong relationship between the sex o f the household head and household welfare: female-headed households, defined as such by the households themselves, are more concentrated inthe lowest consumption q~intilesGiven the constraints that women face inruralAfghanistan, especially vis-& . ~ ~ vis working outside the home, ownership o f assets and the low levels o f literacy (only 10% o f female household heads are literate), it is not surprising to find female-headed households amongst the poorest. The proportion o f female household heads among the poorest quintile (13%) is more than three times higher than that inthe better off quintile (4%). 3.8 EducationAiteracy and employment of household head. Human capital and non-agricultural employment are bothpositively correlated with higher food consumption. Illiteracy is high inrural Afghanistan, and only about 30 percent o f household heads can read and write: among women it i s only 10 percent. Levels of education are very low, on average only 1-2 years o f schooling. Nonetheless, household heads who can read and write have a lower incidence o f consumptionpoverty. This is partly mediatedthrough the labor market and access to non-agricultural employment, and may also reflect better farming practices. With regard to the labor market, while the poorest are more likely to work as agricultural laborers (on other peoples land) or in other daily wage employment (off-farm), the better off are farming their own (irrigated) lands, working in trade or other enterprises and for the public sector. Migration abroad, usually more associated with the educated, is also positively correlated with higher food consumption, whereas intra-rural migration is more closely aligned with the poor. 55 Itshouldbenoted inthe Afghan context, that the de facto number of female householdheads in the NRVA samplemayhave been significantlyunder reportedas aresult o f socialstigma associatedwith separationfromthe husbandandthe fact that male relativesmayhaveofficially taken over role as householdhead, in the absence of ahusband. Also, the relativewelfare of family sub unitswithin apolygamoushousehold can vary dependingon householddynamics. 30 Table 3.2: Various HouseholdCharacteristicsbyConsumptionQuintiles All Q1 Q2 43 44 Q5 Households (poorest) (better off) Household Demographics Household size 7.4 7.6 7.6 7.6 7.3 7.5 Numberofchildren under 6 years 1.3 1.2 1.2 1.1 1.o 1.2 Numberofchildren 7-16 years 2.8 2.9 2.9 2.9 2.6 2.8 Numbero felderly >60 years 0.3 0.3 0.3 0.3 0.2 0.3 Householdswith disabled (%) 22 13 12 12 9 14 Household Head Female householdhead ("A) 13 9 8 5 4 8 Household head literate ("h) 18 23 29 30 33 27 Hhheadworks inagriculture (%) 65 63 62 63 58 62 Hhheadworks innon-ag. employment 35 37 38 37 42 38 Household assets and activities Landownership: some land ("ho f hh) 60 74 79 82 85 76 Among those owning land- rainfed 28 25 24 22 26 25 irrigated 47 61 66 73 76 65 Livestock (# o f cows) 0.4 0.7 0.6 0.8 1.1 0.7 Grows poppy ("ho f households) 3 3 5 5 8 5 Resides inpoppy growing area (%) 10 7 9 8 14 10 Incidence o f migration ("3 o fhholds) 15 17 20 18 16 17 Distributiono f migration - Rural-mal 15 16 14 9 10 13 Rural-urban 14 11 17 18 6 13 Abroad 71 73 69 73 87 74 Source:NRVA 2003. 3.9 Households with disabled members. A high percentage o f rural households have disabled family members, and are among the poorest.56About 14percent o f rural households have disabled family members. These households are generally more likely to have lower levels o f food consumption, and some households exhibit very low levels o f consumption, suggesting extreme vulnerability. These findings imply that special attention is needed to understand the circumstances o f these "vulnerable" groups and that policies and programs should be appropriately targeted. 3.10 Assets. The role o fphysical and productive assets is critical. Physical assets can facilitate mobility out o f poverty byprovidingincome generating opportunities. Indeed, the findings do suggest that owning productive assets like land or animals are inversely related with poverty and food consumption (Table 3.2). One quarter o f the households in the NRVA sample were landless, and some o f these households are extremely vulnerable. Among those with land, ownership o f irrigated land has a more positive relationship with higher food consumption that those deriving their livelihood from rainfed land, possibly indicating both the role o f inherited capital inreproducing poverty status and the higher and more stable income potential o f irrigation and access to water. After water, access to livestock and tools for farming i s the next most important consideration, and poor households report a high incidence o fproduction constraints arising from lack o flivestock and tools. 56Bothphysicalandmentaldisabilitiesare includedhere. 31 3.11 Opium poppy. Households associated with poppy production have significant lower poverty rates. The issue o f opium and poppy production is o fparticular importance for the Afghanistan economy ingeneral and for the rural population in particular. Overall, only 5 percent o f rural households in this survey reported growing poppy. The possibility that this indicates a significant under-reporting by households due to the illicit nature o f this commodity cannot be ruled out. Nevertheless, while also correlating with land ownership, the poverty rates would appear to confirm that households producing poppy are less likely to be poor. Inparticular, only 33 percent among households reporting production o fpoppy are poor, compared with 49 percent among those who are not involved in poppy. Inaddition, households residing incommunities where poppy is produced are also less likely to be poor. This i s most likely due to the labor intensive nature o f poppy cultivation, involving large numbers o f workers, especially duringthe harvest, as well as the general effect of opiumon the local economy. 3.12 Access to Basic Services and Utilities. The supply o fbasic services and utilities is extremely low inrural Afghanistan. The overall low coverage, as indicated by the NRVA data, reflects the historical context and decades- long conflicts in the country. For example, less than 16 percent o f the rural households have access to electricity (Table 3.3). In addition, a third o f the rural household reported have no sanitation services available to them (although it is highly probable that even for those reporting services, the quality o f that service may be unsatisfactory) and almost all households reported using predominantly traditional technologies for cooking (99%) and lighting (86%).Access to clean drinking water is also low. For example, more than half o f the rural households reported that they obtain their drinking water from open water sources such as rivers, dams or water pools (Table 3.3). In addition, the other half o f the rural households gets their water from wells or hand-pumps within the community. The NRVA data did not allow for sufficient detail to objectively assess the safety and quality o f all water sources and based on existing surveys, it is highly likely that the majority o f households do not regularly access safe drinking water for cons~mption.~' 3.13 While the absolute level o f access to these services i s low, poorer rural households have even lower access than the non-poor. For example, a third o f the households inthe highest quintile have access to electricity, compared to 11percent among households inthe lowest quintile (Table 3.3). Access to clean water is slightlyimproved among the higher consumption quintiles, but the overall similarity confirms the extent to which all households in rural Afghanistan may face significant healthrisks. 3.14 Public facilities and services within rural communities are scarce. Less than half o f the rural communities inthe NRVA have aprimary school within the community and only 13percenthave a secondary school (Table 3.4). Equally low is the existence o f public transportation (34 percent) and health facilities (9 percent) within rural communities. Interestingly, access to public facilities such as schools and health facilities is low for all households irrespective their welfare level (Table 3.4). A number o f community level characteristics suggest that household location matters when it comes to access to services. For example, households residing in larger communities are more likely to have higher food consumption expenditures. Larger community populations are typically in close proximity to larger urban areas and as such to a larger set o f income opportunities and access to services. 3.15 Education. School enrollment rates among children are low. Only half o f the children o f schooling age attend primary school. The difference between boys and girls is large: while more than 70 percent o f the boys are enrolled in primary school, only a third o f girls o f primary school age go to school. Furthermore, children from poorer households have lower enrollment rates. These patterns are even more pronounced for secondary school where the (net) enrolment rate i s 7 percent (2 percent for girls). Interestingly there appears to be no significant differences ingender disparity between poor andnonpoor households. 57 Accordingto UNICEF's 2003 MICS, only 31% o frural householdshave access to drinking water from apumpor protected spring. 32 Table 3.3: Access to Basic Infrastructureby FoodConsumptionQuintiles, Rural Afghanistan2003 Q1 42 43 44 Q5 All (Lowest) (Highest) household S Access to electricity (%) 11 13 15 15 27 16 Traditional lighting source (%) 90 89 87 85 78 86 Traditional cooking source (%) 100 99 99 99 98 99 Source o f drinkingwater (%with open 50 source) 51 50 49 48 44 Access to adequate toilet facility (%) 5 6 8 8 9 7 No sanitation facilities (%) 36 28 28 24 20 28 Source: NRVA 2003. Table3.4: Access to Basic Facilities and Services, Rural Afghanistan2003 Q1 Q2 43 Q4 Q5 All (Lowest) (Highest) household S Facilityor servicewithin community Public transportation 30 37 39 34 31 34 Primary school 50 51 48 44 46 48 Secondary school 11 15 13 12 12 13 Health facility 9 9 10 9 10 9 Source: NRVA 2003. Table 3.5: NetEnrollment,Rural Afghanistan 2003 Primary(ages 6-12) Secondary (ages 12-18) Male Female All rural Male Female All rural All rural 70 32 51 11 2 7 Poor 67 30 49 15 2 8 Non-poor 73 34 54 21 4 12 3.16 The low school attendance rates can be attributed to both supply and demand factors. While only three percent o f the rural households report that the direct costs associated with attending school is the main cause for not sending their childrento school, 22 percent report that lack o f schools is the mainreason for non-attendance. Lack of books, absent or poor quality teaching, and poor school facilities are additional factors contributing to low enrollment. But other reasons, such as "family reasons" (43 percent) or work (17 percent) indicate a complex pattem o f reasons for non-enrollment in school. Across the consumption quintiles, reasons for not sending children to primary school are similar. With regard to secondary school, households in lower consumption quintiles are more likely to report lack o f books and teachers as the main schooling problems, while household in the higher quintiles report that inadequate facilities as well as supplies are the bigger problems. 33 3.17 Health. Despite recent achievements, Afghanistan still ranks poorly on all health indicators. While data for most health indicators are not available, the existing estimates are poor: infant mortality o f 165, and under five mortality of 257 per thousand live births, maternal mortality o f 1600 per 100,000 live births, and immunization coverage (Diphtheria, Tetanus, Pertussis - DTP3) less than 30 percent. The rate o f chronic malnutrition (moderate and severe stunting) is estimated to be in the order o f 50 percent, reflecting a combination o f poor caring practice, micronutrient deficiency and food insecurity.58 (Table 4.3 inChapter 4.) Most o f the burden o f disease results from infectious causes, particularly among children where diarrhea is widespread (largely due to limited access to safe drinking water), acute respiratory infections, and vaccine preventable illnesses (such as measles). Among adults, tuberculosis is widespread, particularly among women. Inaddition, almost half o f the existing health facilities do not have female staff, which implies that women are very unlikely to use those facilities. 3.18 A number o fnew insights related to health and heath services emerge from the NRVA data. The incidence o f illnesses such as fever or malaria is relatively high- more than 40 percent o f the rural households reported at least a member who suffered with fever in the monthprior to the survey (Table 3.6). Inaddition, more than a third o f the households had a member who suffered with diarrhea. Interestingly, the incidence of these health problems i s similar across the consumption quintiles, reflecting, for example, the finding that restricted access to protected water points applies fairly evenly across all households. However, there are notable differences in the care sought when sick: two-thirds o f people inthe lowest consumption quintile sought no medical attention and those that did went to a traditional healer. At the other end o f the spectrum, family members from better o f f households sought care from a private doctor as well as the traditional healer. Expense and access to health facilities are the differentiating factors inbehaviors o f sick people seeking care from different consumption quintiles. Overall access to health services is very low. Only nine percent o f the communities in the NRVA have immediate access to a health facility, and over half o f the rural households that needed medical assistance in the month before the survey did not seek medical assistance (Table 3.6). Among those who did seek medical assistance (47 percent), the majority visited a traditional healer (13 percent o f cases) or a private doctor (12 percent o f cases). Reasons for not seeking medical assistance include cost (37 percent), facilities not available (30 percent), and distance to a medical person (24 percent). Table 3.6: HealthStatus and FacilityAccess by Consumption Quintiles, RuralAfghanistan2003 Q1 Q2 Q3 4 4 Q5 All (Lowest) (Highest) Households Households with at least one member with: Feverhalaria last month e?) 37 42 44 45 43 42 Diarrhealast month(%) 31 28 31 28 30 29 Healthfacility used mostfrequently last month, if needed None 67 61 52 40 40 53 Traditionalhealer 11 10 18 11 13 13 Private doctor 6 7 8 22 21 12 Basic health centre 6 10 9 10 11 9 Other 10 12 13 17 15 13 Reason not used any healthfacility last month, if needed Too expensive 44 35 34 34 33 37 Too far 22 24 22 29 27 24 Facilities not available 25 29 35 30 33 30 Other 9 12 10 h R " 4 I Source: NRVA 2003. 58Securing Afghanistan's Future, Health and Nutrition Technical Annex, January 2004. 34 Nutrition and Malnutrition 3.19 The incidence o f malnutrition, especially chronic malnutrition, in rural Afghanistan is thought to be high. Unfortunately, the unavailability of nationwide anthropometric information hinders an in depth and systematic analysis o f the topic. Most malnutrition occurs during pregnancy and the first 2 years o f life. Typically, there are a number o f factors associated with malnutrition: (i) household level access to food; (ii) type o f food consumed the and variety in the diet; (iii) access to health services and facilities; and (iv) home carelweaning and feeding practices. All three factors are present inrural Afghanistan. 3.20 An examination o f the household level dietary patterns reveals a number o f interesting insights, and suggests micronutrient deficiencies inmany people. For the total NRVA sample, typically 70 percent o f all calories come from cereals and tubers (mostly wheat). When the data are analyzed by group, there is greater dietary diversity among the upper consumption quintile, while poor households typically have little dietary diversity. In particular, more than 75 percent o f the daily calories for the lowest quintile are derived from cereals and roots compared with about half for the highest quintile (Figure 3.3). Overall, the rural diet inAfghanistan is particularly narrow, with the average household consuming less than 5-7 distinct food sources, with wheat being the dominant food item. The data also confirm the significance o f regional variations infood consumption patterns, with the west o f Afghanistan shown to correspond to the worst indicators for dietary diversity. Figure3.3: Daily CaloricDistributionPer FoodGroup, By ConsumptionQuintile 100% 80% 60% 40% 20% 0% Source: NRVA 2003. Determinantsof Rural Poverty 3.21 Insummary, anddrawing ona multivariate analysis (as compared to the bivariate analysis presentedabove: Table 3.7), whereby different factors are simultaneously compared with the poverty indicator in an effort to distill their separate effects, the main determinants o f rural poverty include: Location significant geographical differences in rural poverty, with higher poverty in the West and - lower inthe North-East; Numberof children -large households with more children tend to bepoorer; Gender of household head -Female headedhouseholds tend to bepoorer; Disability -households with a disabled member tend to bepoorer; 0 Migration-households who have migrateddomestically (including IDPs) tend to be poorer, but those who have returned from abroad are relatively better ofc 35 Remittances external remittancesplay an important role inalleviating poverty; - Human Capital education levels are closely associated with non-agricultural employment, and with - lower levels o fpoverty; Publicsector employment -employment inthe public sector, inparticular, leads to lesspoverty; Land -land ownership has a positive impact on welfare, especially irrigated land; Landlessness share-cropping and working as an agricultural laborer is positively associated with - poverty; Poppy -the cultivation o f opiumpoppy leads to higher levels o f income; and 0 Infrastructure and services -access to basic services is not a strong determinant o fpoverty: everyone is equally badly served. Are Things Getting Better? 3.22 Qualitative perceptions o f welfare change indicate that households do perceive an improvement over time. While the static nature o f the NRVA and lack o f earlier data limit the ability to quantify how welfare has changed over time, a question on "how things compare with last year" reveals that welfare in rural Afghanistan is perceived to have improved in the 12 month period, summer 2002-2003. Almost 40 percent o f the sample perceived welfare improvements while for about 25 percent o f the sample the situation had deteriorated. Not surprisingly, non-poor households were more likely to perceive welfare improvements over the last year with around 30% o f the poorest quintile reportingimprovements, as against around45% o fthe better offquintile. (Figure 3.4). Figure3.4: Perceptionsof change "How do things comparewith lastyear?" 50% - 40% 30% 20% 10% 0% Q1 (low) Q2 Q3 Q4 Q5 (high) Eil Worse Same 0 Better Source: NRVA 2003. Shocksand Vulnerabilities 3.23 As discussed earlier inthis report, households living inrural Afghanistan have hadto weather many shocks over the years. The poverty profile o f rural households (detailed above) is to a large extent the outcome o f the interplay between assets, opportunities (which facilitate improvements in welfare) and shocks (which can result in {greater} poverty). According to the NRVA, during the year preceding the survey almost 80 percent o f households inthe sample had been affected by at least one covariate (or aggregate) shock, and 23 percent o f households had been affected by an idiosyncratic (localized) shock. Some households were affected by more than one shock - such "bunching" o f shocks is thought to be associated with severe poverty.59By far the most widespread shock reported was drought, affecting more that half o f the rural population in the sample.60 Next in importance came farming shocks (loss o f crops and animals), natural disasters, epidemics, and food price increases (Figure 3.5). Surprising the incidence o f violence as a shock was small (less than 5 percent of households feeling its effect). With regard to idiosyncratic shocks, loss o f employment, ill-health and death in the family were the most common. (Figure 3.6). 59 Tesliuc (2001) 6o By the time the NRVA survey was carried out, the worst o fthe drought had passed and indeed the snow-melt and rains o f the 2003 agricultural season were very good. It i s most likely that people responded to this question in a less well-defined timeframe than the past year, remembering the bad time of recent years. 36 Table 3.7: Poverty Correlates: Log of Food Consumption per Capita Coefficient Household demographics Household size -0.05*** Number of Children <6 (adults 17-60excluded category) -0.03*** Number o f Children 7-16 -0.02*** Number o f Adults >60 -0.20* Members with physical disabilities (yes=l) -0.06*** Household head characteristics Female householdhead(yes=l) -0.06*** Household head age (years) 0.00 Household head literate (yes=l) 0.07*** Household head works inagriculture -0.05*** Household head works in(public sector excluded) Private sector -0.12*** Self employed -0.01 Household assets and other characteristics Own house (yes=l) 0.03** Number o frooms 0.07*** Owns rainfed land(yes=l) 0.07*** Owns irrigated land (yes=l) 0.14*** Number o f cows owned 0.005*** Number o fpoultry owned 0.001*** Produces poppy 0.10*** Access to infrastructure Access to electricity (yes=l) 0.06*** Access to sanitary service (yes=l) 0.05*** Community characteristics Community population (#) 0.001** Agro-ecological zone (irrigated only excluded): Bothirrigate andrainfed -0.001 Rainfed only -0.068*** Grazing only -0.129*** Market incommunity -.03 Constant 8.53*** Obs: 7779 : from the overall sample, a number of households lack complete data and are omitted from these estimates. R2:0.25 Dependent Variable: Logofper capita food consumption Also includes regional dummies (not reported) *significant at 10%; ** significant 5%; *** at significant at 1%. 37 Figure 3.5: Incidence of Covariate Shocks (% of Household Reporting) I Drought Natural Food Violence shocks price increase Figure 3.6: Incidence of Idiosyncratic Shocks (% of Households Reporting) 15 10 5 0 Serious health Eqloymntshock Deathinfarnily Violence shock Source: NRVA2003 38 3.24 The incidence o f shocks across food consumption quintiles shows some variation such as the greater importance o f farming shocks in the higher quintiles, where farming is more prevalent. For the poorer groups, epidemics - leading to ill-health and death, food price increases, and loss o f employment stand out. But overall, the patterns o f shocks are quite similar, with the large covariate shocks dominating. Table 3.8: Covariate Shocks by Food Consumption Quintiles All Q1(Low) Q2 Q3 4 4 Q5(High) Households Incidence of covariate shocks (%) Drought 51 55 56 59 54 55 Farming shocks 38 45 47 50 50 46 Natural shocks 37 38 36 38 43 38 Epidemics 36 29 27 28 23 29 Foodprice increase 29 25 27 28 20 26 Farmgate price decrease 6 4 5 6 6 5 Violence 3 4 4 5 6 5 Highinflux o freturnees 4 4 4 3 3 4 Households affected with: 0 covariate shocks 20 22 22 20 23 21 1covariate shocks 19 18 19 20 19 19 2 covariate shocks 29 24 23 21 24 24 >2 covariate shocks 32 36 36 39 37 36 Farming shocks: crop andpest diseases, reduced availability o f grazing areas; Natural shocks: earthquakes, landslides, flooding, hailstorms; Epidemics: unusually high level o f human diseases. Source:NRVA 2003. Table 3.9: Idiosyncratic Shocks by Food Consumption Quintiles All Q1(Low) 4 2 4 3 Q4 Q5(High) Households Incidence of idiosyncratic shocks (%) Employment shock 13 12 11 10 8 11 Serious health shock o f working member 16 12 13 12 9 12 Death infamily 10 9 9 6 5 8 Violence 2 2 1 2 2 2 Households affected with: 0 idiosyncratic shocks 71 76 77 78 84 77 1idiosyncratic shocks 18 14 14 14 10 14 2 idiosyncratic shocks 9 8 7 7 5 7 >2 idiosyncratic shocks 2 2 2 1 1 2 Emplovment shock: loss o femployment for a household member, reduced salary, bankruptcy o f family business. Violence (violence or theft). Source:NRVA 2003. The Impact of Shocks on Welfare 3.25 The impact o f shocks on welfare is extensive. In fact, very few o f the households that report that they were affected by large covariate or idiosyncratic shocks indicate a negligible impact on incomes or assets. For example, 39 almost 60 percent o f the rural households affected by drought report an income loss as a result o f the shock. In addition, 21 percent indicate that they lost assets while another 16 percent reports that they lost both. Only seven percent reported no losses due to the shock. Similar patterns emerge from other covariate and idiosyncratic shocks. For example, only six percent o f the households affected by various farming shocks (pests, livestock diseases) report that it didnot affect their incomes or assets. Similarly, more than 95 percent o f households affected by idiosyncratic shocks like a loss o f employment or a serious illness o f a breadwinner report that their incomes and assets were negatively affected. There is only a small variation inthe incidence o f reported income and asset losses due to the various shocks across consumption quintiles. As a general pattern, households in the upper consumption quintiles reported a higher incidence o f asset losses, reflecting perhaps the fact that these households have a higher asset base and as such they can depend on it more for coping. Households in the lower consumption quintiles, by comparison, reported a higher incidence o f income loss. Still, these patterns only reflect the incidence but not the magnitude o f income and asset losses and as such they offer only partial insights about the potential quantitative impact o f these shocks on household welfare. 3.26 One way to partially understand the magnitude o f the impact o f the various shocks on household welfare is to explore whether households have managed to recover from their reported losses. The patterns suggest that shocks in rural Afghanistan have long run impacts, which implies that few households have the ability to adequately manage risks. For example almost 70 percent o f the rural households who were affected by drought have not recovered the income and asset losses (Table 3.10). In addition, another 29 percent has only partially recovered while more than half o f the rural households affected by other covariate shocks like farming or other natural shocks report that they have not recovered the income and asset losses. Further, two thirds o f rural households that were affectedby employment shocks have not recovered their losses (Table 3.11). 3.27 Despite the already high level o f non-recovery for all rural households, poorer household are even less likely to recover from the various shocks. For example, the probability o f a rural household in the lowest food consumption quintile to partially recover from covariate shocks like drought or farming was significantly lower compared to non-poor households (Table 3.10). In addition, 80 percent o f the households in the lowest quintile report that they have not recovered the losses from employment shocks, compared to 60 percent among those inthe highest quintile (Table 3.11). Similarly, 76 percent o f the households inthe lowest quintile have not recuperated the income and asset losses from serious health shocks as opposed to only 39 percent for those in the highest quintile. While the information on recovery from various shocks is a subjective concept for each household and individual, these observed patterns do suggest that shocks have large impacts on welfare on everyone affected and that the poor are more vulnerable and less capable to effectively recover. RiskManagement Instruments: CopingWith Shock 3.28 Faced with a multitude o f risks and shocks, rural households in Afghanistan rely mainly on informal and private risk management instruments. The main coping mechanisms when a shock hits are reductions in expenditures, including food consumption, and sale o f assets. (See Tables 3.10 and 3.11). For example, almost 40 percent o f the rural households affected by drought reduced food consumption as the main coping mechanism.6' In addition, another 25 percent decreased overall expenditure and another 17 percent sold assets (such as livestock). A smaller percentage (9%) relied on informal coping instruments (such as borrowing from family and friends). Similar coping patterns emerge for other covariate and idiosyncratic shocks, such as farming shocks, employment losses or health shocks (Tables 3.10 and 3.11). 6' Itis important to note that among richer households, a reduction infood consumption will not necessarily have the same impact as for those inpoorer households. 40 Table3.10: CopingMechanismsandRecovery fromSelected Covariate Shocks, by FoodConsumptionQuintiles Drought Farmingshocks Q1 Q2 43 Q4 Q5 Q1 42 43 Q4 Q5 (Low) (High) (Low) (High) Coping mechanism(YO) Decrease infood consumption 36 38 41 36 39 31 32 35 33 34 Decrease expenditures 21 25 24 28 25 23 22 21 22 24 Use informal networks 14 13 10 5 5 14 15 11 6 6 Formalborrowing 2 2 1 1 1 1 2 1 1 2 Asset sales 13 14 16 20 22 17 19 23 28 24 Other 7 3 3 3 2 7 3 3 4 2 None 7 4 5 6 6 7 6 6 7 8 Recoveredfrom covariate shock (Yo) Not at all 76 73 68 64 57 69 69 62 57 49 Partially 21 24 29 34 38 26 28 34 40 44 Yes 3 3 3 2 5 5 3 4 3 7 Source: NRVA 2003. Table 3.11: CopingMechanismsand Recovery from SelectedIdiosyncraticShocks, by FoodConsumptionQuintiles Employmentshocks Health shocks Q1 Q2 43 44 Q5 Q1 Q2 43 Q4 Q5 (Low) (High) (Low) (High) Coping mechanism(YO) Decrease infood consumption 32 41 53 49 43 23 30 26 29 30 Decrease expenditures 22 20 15 17 24 20 16 16 12 20 Use informalnetworks 17 19 10 10 4 24 26 18 17 8 Formalborrowing 3 1 1 1 1 4 1 2 2 2 Asset sales 10 15 16 21 22 15 24 32 33 36 Other 1 3 3 6 3 3 13 3 5 6 1 None 3 1 0 0 4 1 1 1 0 3 Recoveredfromcovariateshock ("/.I Not at all 81 81 73 78 60 76 75 65 50 39 Partially 16 17 26 21 35 22 23 33 50 59 Yes 3 2 1 1 5 2 2 2 0 2 Source: NRVA 2003. 3.29 By contrast, very few households report using formal or public instruments to cope with shocks. In particular, no more than three percent o f rural households report that they borrow money from formal lenders to cope with the various shocks (Tables 3.10 and 3.11). Similarly, few households report that the existence o f public programs (such as public works programs) has helped them to cope with various shocks. In fact, the incidence o f participation ina number o f safety net programs that aim to target households at "risk" is not significantly different between households impacted by a shock than those not impacted (Table 3.12). As the analysis detailed below shows, safety net programs are not well targeted to those most in need in terms o f welfare measures nor to those affectedby shocks. 41 Table 3.12: Program Participation and Shocks Incidence Program participant (%) N o Yes Covariate shocks Drought 54 56 Farming shocks 48 44 Natural disaster 39 38 Epidemics 27 30 Foodprice increase 22 30 Idiosyncratic shocks Employment shock 9 13 Serious health shock o f working member 10 15 Death infamily 7 9 Violence 1 2 Programparticipant refers to a household that participates in any o fthe following three programs: food for work, cash for work, relief food. Similar results emerge in looking at these pattems by each programseparately. Source: NRVA 2003. 3.30 The majority o f the main risk coping strategies employed by rural households have long-run adverse implications. Inparticular, hannful coping mechanisms such as the decrease in food consumption have important implications both in the short term in the form o f lower productivity, but also in the long run via health and nutritional consequences among children. While reduced consumption o f food is a strategy employed by poor and better offhouseholds, for the better off this may involve mixingbarley flour with wheat flour and decreasing overall dietary diversity, for poorer households reduced consumption can meanjust one meal a day, reliance on bread and tea and consumption o f wild `famine' foods. Similarly, while asset sales may be an appropriate smoothing mechanism for households that are relatively rich in assets, inthe context o f rural Afghanistan, selling assets may have huge implications for the ability o f households to generate income in the future, and as such may accentuate poverty and poverty traps. Finally, while coping mechanisms such as child labor, indentured labor and "sales" o f daughters are less frequent in the NRVA sample (combined in the category "other"), their presence and therefore utilization among some households, albeit small, reveals the destitution and absolute lack o fbetter riskmanagement instruments inrural Afghanistan. 3.3 1 Finally, there are distinct differences in the coping mechanisms between poor and non-poor households. Specifically, while poorer households are more likely to use informal networks to cope with covariate and idiosyncratic shocks, less poor households are more likely to use asset sales to cope with shocks. As expected, poorer households have fewer assets and as such their capacity to use them to cope with shocks is limited. Instead, in addition to food consumption reductions, they have to rely on support from family and friends inthe form o f informal borrowing or transfers. This i s frequently reported by rural households as a potentially damaging disinvestment in social capital and as such it can incur a significant burdenfor household members (and may give rise to increasedpersonal and psychological insecurity). The Performance of Some Formal Safety Nets 3.32 A number o f safety net programs are inexistence inrural Afghanistan, and range from humanitarianrelief feeding to asset creating cash-for-work programs. This section complements the previous analysis by evaluating how a few o f these programs are performing in terms o f their ability to target those households and individuals in "need". The aim is not to provide a full and formal evaluation (the data to do that are not available) but instead to provide initial insights as to how these programs are performing and how they might become more effective. With the challenge o f resource scarcity and a policy commitment to harness the entrepreneurial and risk management capacities o f poor people rather than displace them, it is o f crucial importance that policies and programs involving direct transfers inthe form o f social welfare and safety net mechanisms are efficiently and effectively targeted. 42 3.33 The NRVA collected information on participation in three programs during the year summer 2002 - summer 2003. They are (1) food for work (FFW); (ii) for work (CFW); and (iii) food aid (RFA). The first cash relief two are public works schemes, designed with the dual objectives o f creating rural infrastructure and providing short- term employment for some o f the poor and vulnerable to supplement household incomes and food consumption. Relief food aid was administered mainly by the World Food Program and its objective is to reach extremely poor and food insecure households. Typically FFW and relief feeding programs are offered in the same community: the former designed for the able bodied, and the latter for those unable to work on labor-based projects. Further details on the programs can be found inBox 3.1. 3.34 The participation rates in these programs among the NRVA rural sample varies, but the data indicate that up to 47 percent o f the m a l population may have participated inat least one o f these programs. By far the most widespread program is the FFW, where more than a third o f the rural households in the sample have participated (Table 3.13). In addition almost 20 percent o f the households had participated inthe CFW and another 13 percent had received RFA. (Three percent o f rural households had participated in a few other public and private programs that unfortunately were not identified.) Table 3.13: Program Incidence2002/2003, Rural Afghanistan Householdsparticipating(YO) Food for work 33 Cash for work 19 Relieffood 13 Other programs (PublicLJ"G0) 3 Source: NRVA 2003. 3.35 Who Participates In These Programs? Inprinciple, these programs are supposed to target households that are either food insecure (RFA, FFW) or that need additional income and are prepared to work for a l o w wage (FFW, CFW). As such, evaluating these programs should be onthe basis o fmeeting these objectives. 3.36 From the three programs, only Relief Food Aid is slightly progressive in targeting the extreme poor households (Ql); two perform as ifthey were entitlement programs, i.e., befitting households inall welfare the other categories relatively equally. (Table 3.14). Households inthe poorer consumption quintile do participate in all three programs, but only inthe relief feeding program do they receive a larger share o f the program than other households. It is striking that households in the "better-off' category (QS), that i s the top 20 percent o f households, also participate in all three programs. While it is possible that they were less well o f f at the time o f participating in the program (the data refer to the year preceding the survey) and as such were more eligible beneficiaries, this seems quite unlikely. The concentration coefficient which captures whether program participation is progressive or regressive indicates the following: for the first two programs it is near 0, implyingthat it is welfare neutral while for RFA it is (-0.18) suggesting its moreprogressive nature.62 62The concentration coefficient ranges from +Ito -1, with positive values indicating a regressive distribution, negative values a progressive distribution, and a value o f zero a perfectly equitable distribution. The formula for calculating the concentration coefficient is: where Nis the total number o f groupings o fthe welfare variable used (e.g., 5 food consumption quintiles) and x, is the coverage rates among group i. 43 Box 3.1: Some safety net programs Labor Intensive Public Works Programs (cash-based). These interventions, which have been operating since 2002, are generally designed to address the following objectives: improve employment inrural areas at a minimumwage, as a safety net to as many people and inas short a time as may be feasible. rehabilitate in a sustainable way Afghanistan's basic infrastructure (rural roads and irrigation channels). prevent a humanitarian catastrophe by ensuring winter access to remote villages. The largest single LIPW initiative, the NEEP, was fully operational by the time the NRVA was carried out inthe summer o f 2003. The program was designated for 13 provinces but rapidly spread throughout the country for reasons o f "wide visibility". Within provinces, the criteria for selection o fproject sites included: areas o f intense IDPs and returnee activity; drought affected areas, especially where out-migration was visible; areas severely affected by war; where existing NGO programs (often food-for-work) could be scaled up quickly; areas where few alternative employment opportunities existed; kuchi areas; and areas sufficiently stable and secure to be able to undertake works. Participation inthe program was self-selection, predicated on the notion that at such a low wage ($2/day) only the poor and needy would want to participate (defined as refugees, IDPs, disabled, women heading households, nomads, demobilized ex-combatants and poppy-ban affected households). As time progressed however, at $2/day, labor became scarce during the summer months, and wage rates have been increased. Some barriers to the involvement o f the poor and vulnerable have been identified and include distance to project site, share-cropping and other arrangements bindinglabor to landlords, andtraditional value vis-a-vis women's work. Food for worWfood for asset creation programs. These programs are similar inprinciple to the cash-for- work programs, except that payment is made in food commodities (wheat, flour, cooking oil). FFW is targeted to areas o f acute and very high food insecurity areas where households meet less than 50 percent o f caloric requirements. Wages are set below casual labor rates so that FFW is largely self-targeting and does not compete with other work. FFW activities are targeted through community participation, allowing villagers, including women, to agree on projects to be supported and determine who should participate. Food for food programs have been a feature o f life in rural Afghanistan for many years, and have been supported by the World FoodProgram. Relief food aid, under the umbrella o f the WFP, consists o f a number o f programs designed to assist food- insecure households to meet basic food requirements, with particular attention to the most vulnerable (returning refugees, IDPs, malnourishedchildren, TB patients). Major programs inoperation in2002/2003: Returning refugees: 150kg o f wheat per family, provide at the place o f return, with follow-on support for 3 months on a self-targeting basis; IDPs: support for both returning to their homes (150kg. o f wheat) as well as those remaining in camps; Vulnerable rural households: households living in FFW communities but unable to participate- such as widows, orphans, households headed by elderly or disabled person, and households without income earners. Rations for 5 months; Assistance to TB patients and their families, mainly women. Foodrations for eight months; and School-feeding programs for girls and teachers. 44 Table 3.14: ProgramParticipation andDistributionby Poverty Groups (%) Foodfor work Cashfor work Relieffood Q1(low) 24 22 33 42 22 21 21 4 3 20 22 18 4 4 18 18 14 45 16 17 14 Total 100 100 100 Concentration coefficient -0.08 -0.05 -0.18 3.37 Using different targeting eligibility criteria, for example food insecurity, whether the households had unemployed household members, categorical targets, one can further evaluate whether these programs reach "eligible" households. For each o f the programs and each of the eligibility criteria a variety o f indicators such as errors o f inclusion (e.g., non-eligible households that participate in the programs) or errors o f exclusion (e.g., eligible households that do not participate in the program) are calculated. The results are presented in Table 3.15. None o f the three programs are particularly effective inreaching the "eligible" population using any o f these criteria. 3.38 In conclusion, to fully evaluate these programs, additional information not collected in the NRVA is needed. For example, additional data on the level o f transfers is important to further assess whether the limited resources that these programs have are targeted to the most needy and eligible. Also, one ideally should know the overall cost associated with delivering a unit o f assistance to programbeneficiaries. The patterns above, however, suggest that the programs are not proving very effective inreaching the poorest and most vulnerable. Clearly steps need to be taken in order to improve some o f the targeting and selection criteria that may exist in these programs, and to help restructure and redesign some o f the objectives and goals o f the various existing and potentially new programs. The detailed evaluation o f one cash-for-work program (inHazarajat: see Box 3.2) suggests some pointers for more successful labor-intensive public works programs, inter alia: Seasonality-more activities inthe winter months, with before winter and end winter payments to coincide with most needytimes o f the year; Types of activities- activities which the extreme poor andvulnerable canhelp with (a particular challenge with regardto the inclusion o f different groups o f women and people with disabilities) - and location o f activities, in-or near to the village; Durationof employment- long enough to provide a significant cash injectionandprovide a meaningful altemative to other coping strategies; Create quality and durable infrastructure -based on community prioritiesandhamessing community participation; Modeof payment-preference for cash-for-work over food-for-work projects as muchas possible, but with some seasonal variation; and Support traditionaldistribution mechanisms-support communities to usetransfer instruments for less-able bodied households. 45 Table 3.15: Program Targeting Indicators under Various Eligibility Criteria (YO) Assumed targeting criteria Inclusion error: Exclusion error: YOof non-eligible group % of eligible group receiving not receiving Poverty Food for work 45 63 Cash for work 47 79 Relief food 38 83 Food intake gap Food for work 81 70 Cash for work 79 85 Relieffood 77 85 Unemployed in household Food for work 88 67 Cash for work 89 75 Relieffood 85 83 Cash for Work Program IDPheturnees 80 80 Households with disabled person 89 85 Female headed household 92 81 Source: NRVA 2003. 46 Box 3.2: Oxfam's Cash for Work inHazarajat: LessonsLearned for Labor Intensive Public Works Background: The Oxfam cash for work program was funded by ECHO in a project site where Oxfam hadbeenpresent since 1991. The intervention was focused on livelihood protection over the winter period when employment opportunities are reduced and to ensure the prevention o f damaging coping mechanisms. According to Oxfam, cash based transfers were chosen: as a result o f the fact that food based transfers were frequently re-distributed and were often not perceived as a development activity by communities; for asset building; and, to minimize the potential for adverse impact on agricultural activities. Targeting and Activities: Targeting mechanisms were facilitated through the introduction o f Oxfam developed criteria but managed by communities to the extent possible. Where greater investment in shurdcommunity participation had taken place, targeting problems were minimalized. Communities had clear ideas regarding the identification o f most vulnerable households including homeless; land less; small landowners; widows and people with disabilities. Cash for work activities included: construction o f water reservoirs and river protection walls, protection against erosion, safe guarding irrigation channels, tree planting, vegetative propagation, fodder collection and planting, etc. A women's embroidery project component was also introduced and Oxfam acknowledged that the identification o f acceptable work for Hazara women was a challenge, given the inappropriateness o f extemal manual labor. Traditional community safety nets instruments were reinforced through the contribution by participants o f cash transfers to the non-able bodied. Payment: Halfo fthe cashpayments were distributedbefore winter and a second distributed, after winter, which coincided with the communities' greatest needs for cash. Wages levels were around $2lday for unskilled, and $4 for skilled, and interestingly, although community representatives argued that other agencies were paying higher rates, local govemment supported the principle o f greater self targeting through lower payments. Monitoring confirmed that participating households accessed 70 days work and around 90% o f wages were spent on food, much o f it stocks for the winter season. Despite the existence of high debts, money was not spent onrepayment. Impact: The household impact o f the intervention included increased purchasing power; prevention o f migration and distress asset sales; and, creation o f new community assets through the infra-structure creation. Interestingly, communities appeared most interested inthe overall and longer term impact o f the infrastructure on livelihoods (particularly when these were located within the local vicinity) rather than the more immediate impact o f the cash receipts on household well-being. Cash-for-work appeared to have focused communities on raising the quality o f outputs, design, materials, etc. Another impact o f employment generation, requiring further research, was the decreased acceptance o f poorer households with regard to exploitative sharecropping arrangements. Market traders were particularly supportive o f cashinjection and markets appeared to respond well to increased purchasing power. 47 IV. URBANPOVERTYAND VULNERABILITY Introduction 4.1 Urban poverty is as much a multi-dimensional phenomenon as rural poverty. In common with their rural counterparts, the urban poor suffer from various deprivations, e.g., lack o f access to employment; adequate housing and services; health, education and personal security. However, it is also useful to consider some o f the general characteristics that distinguish urban poverty (as highlighted in B o x 4.1). What i s clear is that to understand the deprivations that poor people face and to address them effectively, the impact o f local context and external forces must be grasped. An understanding o f poverty and vulnerability must recognize the differences between rural and urban populations. While there are common characteristics to both rural and urban poverty, generalizations that overlook the diversity o f locations must be avoided. Box 4.1:UnderstandingUrbanLivelihoodsinAfghanistan For rural Afghans, cities can present tremendous opportunities for social and economic development. As evidenced by the desire o f many retuming refugees to resettle in urban areas, urban migration i s often the most rational choice, given the improved prospects to secure livelihoods through paid employment. However, in spite o f the positive opportunities that urban areas can offer, it should not be automatically assumed that urban vulnerability is less acute in Afghanistan. Given the underlying factors o f ineffective govemance and collapse o f services, a significant proportion of the urban poor can find themselves living in as bad or worse living conditions than the rural poor, with no access to water or sanitation and with high risk o f accidents (e.g., fires) and personal violence. For many people, urban Afghanistan provides a highlythreatening and risky environment. Urban centers are characterized by the dense concentration o f people and by the corresponding need for complex delivery systems to meet their resource and service needs (food, water, shelter, transport, water, etc.). Urban households frequently have a higher number o f dependents to able bodied adults as compared to rural areas. The competition for space means that housing and land are at a premium. In addition, the near total monetization o f the urban economy (as opposed to partial in the rural economy) means that people's livelihood strategies are dominated by the need for cash to meet expenditures. At the same time there is an inherent instability o f wage income opportunities in the informal employment markets o f cities. As a result o f insecure livelihoods, a high transience rate tends to exist in urban areas as opposed to rural, which can have a negative impact on the formation o f social capital and informal networks. 4.2 However, with exception o f a handful o f largely qualitative studies, initiatives involving the systematic assessment and analysis o f urban poverty and vulnerability at the household level throughout Afghanistan have not yet been undertaken. Accelerated socio-economic change has taken place in recent years, due to rural-urban migration and retuming refugees and IDPs. It will therefore be crucial to ensure that better understanding is developed. It should be noted that NRVA data were not collected inurban areas although the data do highlight some key aspects o f rural - urban linkage, as seen for example, in the significance o f regional urban labor markets for under-employed rural household members. 4.3 The purpose o f this chapter, which i s based primarily on secondary sources, is to provide an overview o f what is known about urban poverty and vulnerability in Afghanistan, drawing together findings from a variety o f thematic surveys and assessments. Some o f these with a specifically urban focus (most often Kabul), look at vulnerability through the eyes o f specific groups such as women or children. Other pieces o f work represent preliminary analysis based on focus group discussions and drawing extensively o n valuable case study material to explore the parameters and spatial mapping o f vulnerability more generally. Additional findings are taken from the National Multiple Indicator Cluster Survey, incorporating specific urban based data.63 While at this stage it is not possible to draw definitive conclusions regarding the relative nature o f poverty and vulnerability inurban areas with the situation in rural areas, nevertheless, some general conclusions can be reached regarding broad similarities and differences. 63 "Moving Beyond2 Decades of War: Progressof Provinces", MICS 2003 - a surveyundertakenby UNICEF inconjunction with the CSO. 48 Background 4.4 The current urbanpopulation o f Afghanistan i s estimated to be around 6.4 million, which constitutes about 30 percent o f the total p ~ p u l a t i o nIt~is~estimated that this figure will have doubled by 2015 (at a growth rate twice . the average growth rate o f the country). Over 70 percent o f the urban population are concentrated in six cities. The return o fInternally Displaced People (IDPs) andrefugees has intensified since cessation o f hostilities in2002. Municipality Population Population ReturneedCity Total Population (Est. 1999) (Est. 2002) (Est. 2002) (Est. 2002) Kabul 1,780,600 2,559,900 128,434* 2,688,334 Kandahar 294,500 308,900 2,720 311,620 Heart 231,600 243,000 657 243,657+ Mazar-I-Sharif 170,600 179,000 1 179,001 Jalalabad 89,000 93,400 50,788 144,188** Total 2,566,300 3,384,200 182,600 3,566,800 4.5 Thousands o f returnees have chosen Kabul as their residence in order to benefit from the facilities and social services that the capital i s expected to offer and to escapethe potential conflict arisingfrom land andproperty disputes with the new `owners' o f their former homes. Owing to the influx o f returning refugees over the past two years, the population of Kabul alone has increased from 1.7 million in 2000 to over 2.5 million (possibly over 3 million) in2003. 4.6 The crisis arising from such rapid growth is particularly evident in the shortage o f housing in the capital city. I t is believed that over 75,000 damaged houses need to be rebuilt in Kabul alone. Almost half o f Kabul's residents are living under the threat o f relocation, inpart due to the disruption o f the administration during decades o f protracted conflict. The majority o f new urban residents are landless returnees and IDPs who have settled in unplanned and hence unserviced portions o f the city, such as the surrounding hillsides. Most urban Afghans have little or no access to conventional services. For example, only 20% o fKabul residents have access to pipedwater. In other major provincial cities levels o f access to water are comparable, and in over 20 smaller provincial capitals no hnctioning piped water system exists at all.65 Towns and cities continue to function within the limited mandate accorded to district municipalities under the 1964 Constitution and who lack the resources and skills to effectively manage the urban environment. 4.7 Kabul has suffered tremendously from the negative impact o f protracted conflict and the collapse o f systems for urban management and investment. As a result o f its location, residents o f Kabul have to cope with the additional challenges associated with the winter season. However, while the majority o f recent studies undertaken relate to Kabul, a few have sought to incorporate an initial assessment o f the situation in other cities.66 From this work it i s apparent that while urban areas share multiple constraints, there are often significant variables relating to local context that needto be considered. 4.8 The city o fJalalabad inthe easternNangarhar province experiences wide seasonal fluctuation inpopulation numbers as a result of in flows from Kabul (during winter) and ongoing migratory outflows (largely short term), both to surrounding rural areas (agricultural labor opportunities) and to Pakistan, as a result o f its strategic location near the border. The local economy in Jalalabad is closely integrated with that o f Pakistan. There is a significant proportion o fIDPs living either inthe city or incamps near by. By contrast, the city o f Herat inwestern Afghanistan TISA (2004) SecuringAfghanistan'sFuture. Ibid. 66 Schutte(2004) UrbanVulnerability inAfghanistan:Case Studiesfrom Three Cities 49 i s situated in close proximity to Iran and Turkmenistan. Due to the connectivity with neighboring countries' markets and the resulting unregulated customs duties from border traffic, the city is relatively developed as compared with other urban areas inthe country. Nevertheless, a wide range o f income variation exists and for the urban poor access to secure shelter, regular incomes and basic services may be as problematic as is the case with other cities. Dimensionsof UrbanPovertyandVulnerability 4.9 Whle the end result o fpoverty may be similar for rural and urban households inAfghanistan (insecurity o f employment, poor access to education, food insecurity, illhealth, increased mortality rates, etc.) the underlying factors can vary significantly (Table 4.2). Table 4.2: UnderstandingDistinctiveAspects ofUrbanPovertyand VulnerabilityinAfghanistan Dimensionof Povertyand DistinctiveAspectsofUrbanPovertyand Vulnerability Vulnerability Income 0 Greater dependence on cash and market access for consumption needs Inherent instability o f employment opportunities derived from informal sector Education Insufficient service provisioninrapidly growing cities 0 Inability to afford school expenses Personal safety I security risks deters school attendance Health Pollutiono f air and water Injuryanddeathsarisingfromtraffic Unsafe working conditions (informal sector) Overcrowded andunhygienic living conditions Tenure Land and housing oflen not affordable in authorized areas - poor occupy land illegally and construct houses without official permission Personal Drug/ alcohol abuse; domestic violence; crime Family breakdown and reduced support for children - vulnerable urban populations frequently lack extended family ties, which can provide safety nets. Financial Lack of access to credidsafety nets from social and informal networks. 4.10 The nationwide 2003 Multiple Indicator Cluster Survey (MICS)67 data provide some crucial insights necessary for undertaking analysis o f poverty and vulnerability in urban Afghanistan as compared with both rural Afghanistan and urban populations in other countries. Table 4.3 provides an overview o f core indicators from this survey that are o fbroad relevance for this chapter. `'CSOAJNICEF (2003) 50 Table 4.3: SelectedIndicators for Poverty and Vulnerability in Urban and RuralAfghanistan ndicator Sational 121 97 115 183 142 172 5 reporting incidence inpast 15 days) 30 30 30 nuedfeeding givenduring diarrhea 54 51 53 19 19 19 27 32 28 73 66 71 90 74 85 not taking place inhealth facility 96 68 89 dwomen (<49 years) who have not heardo fa method to delay 78 55 72 94 79 90 2 3 3 3 4 3 6 6 6 hildren 12-23 months that have received DPT 3 immunization 23 48 30 hildren 12-23 months that have received Polio 3+ immunization 43 71 51 73 84 76 55 73 60 31 61 40 59 87 67 77 60 67 literacy rate: Male ("?o f pop. aged 15 and over) 64 40 57 literacy rate: Female (% o fpop. aged 15 and over) 92 72 86 61 81 67 30 63 40 Source:Afghanistan:Multiple IndicatorCluster Survey 2003, UNICEF andAfg mistanTrans malAuthorit! :so. 4.11 The MICS data point to significant levels o f poverty and vulnerability inboth urban and rural Afghanistan. For several indicators, mainly relating to accessing basic services such as delivery o f mother in health facility, literacy and school enrollment (particularly for girls), access to safe drinking water and sanitation facilities, the situation is slightly improved in urban areas. While the infant and child mortality rates are around twenty percent lower in urban areas than in rural areas, this differential is not perhaps as significant as would be expected in the urban context. However, for both urban and rural areas diarrhea prevalence, the proportion o f disabled children and female headed households, the incidence o f child labor, orphans and fertility rates, no significant variation exists. This may suggest that in both urban and rural contexts, similar factors underlie poor human development and vulnerability, or that different factors leadultimately to similar outcomes. 4.12 Living Conditions and Service Provision. Internationally, many poor urban people are excluded from services on grounds of cost, discriminatory or inadequate administrative and legal practices, and failure o f urban 51 managers to keep up with growth o f cities - especially on the urban fringe. And yet in crowded urban areas, residents may be more dependent on the provision o f private and public services for water, sanitation, fuel, waste disposal, etc. As a result o f over-crowded unsanitary conditions, infant mortality rates are frequently higher among the urban poor than for their rural counterparts.68 Box 4.2: Afghanistan's Urban Water Supply and Sanitation Crisis Conflict, lack o f investment and maintenance and drought have resulted in a poor piped water supply service limited to about 20% o f the population o f Kabul and 13 other major urban centers. The remainder relies either on wells fitted with hand pumps or informal sources such as karez and streams. Various donors and NGOs have recently developed small piped schemes in smaller urban centers. The vast majority o f the urban population relies on pit latrines for sanitation. This situation, combined with the intermittent availability o f existing piped services, inadequate water treatment and the contamination o f groundwater has resulted in a precarious sanitary situation. I Source: TISA (2004), The UrbanWater Supply and SanitationSector. 4.13 For decades the majority o f urban residents in Afghanistan have relied for their water supply upon a combination o f surface water flowing through open ditches and stored in open reservoirs as well as ground water from hand dug wells. As cities have grown with no planning and design, many settlements exist inpoorly drained areas prone to flooding and subsidence with stagnant and contaminated pools constituting a severe health hazard (see Box 4.2). Former systems for the management o f city wide solid waste collection have long fallen into disrepair so that waste piles up instreets and creates a further risk for people. Most o fthe population relies on surface latrines, which frequently overflow onto streets andpavements exacerbating unsanitary condition^.^' 4.14 Protracted conflict inAfghanistan has devastated infrastructure and deprived the country o fnew investment that would raise services to pre-war levels. Poor health and social tension and instability are among the most serious consequences o f service delivery deficiencies. Diarrhoeal disease is a key cause o f under five mortality and severe indoor air pollution resulting from use o f traditional heating appliances affects over 80% o f vulnerable households unable to afford alternatives. Access to food, fuel and water can involve three hours o f time a day. Table 4.4: Urban Service Delivery Deficiencies City/town Access to safe Percentage Number of houses Water destroyed Roads destroyed Kabul 29 60 63,000 Kandahar 20 25 3,700 Mazar-I-Sherif 40 13 3,000 Jalalabad 50 50 3,000 Heart 85 46 3,000 Kunduz 0 23 2,000 4.15 A recent surveyof 1,927 people inKabul found that 47 percent o fprimary household income earners were illiterate. Illiteracy rates per age group and by sex are shown in Table 4.5. The results appear to show a positive correlation o f illiteracy and age. World ResourcesInstitute(1997) The UrbanEnvironment. 69 EuropeanCommission(2002) EC RapidReactionMechanismAssessment Mission: Urbanrecoveryin Afghanistan. 52 MaleIlliteracyYO FemaleIlliteracy YO Under 18years 18 35 18-25 years 31 51 25-35 years 39 64 35 -45 years 38 69 45-55 years 47 83 Over 55 years 67 88 UrbanAsset Vulnerability 4.16 Understanding the asset vulnerability o f differentsocio-economic groups can provide key insights into their relative exposure to various risks. Inthe context o furban Afghanistan (as to a great extent inrural Afghanistan) cash is essential for the purchase o f food, land, shelter, energy, transport, water from vendors and sanitation from pay on use facilities. 4.17 Often people's key assets are their land, house, and capacity to sell labor as well as the ability to participate in social networks. The cost of living for essential non-food expenditure may be relatively higher inurban areas (water and sanitation user fees, fuel charges, school uniforms, higher cost o f foods, etc.), implying for example, that comparing wage levels inrural and urban areas may not always provide an accurate insight into purchasing power levels. 4.18 Housing and land. Safe and secure housing provide a range of benefits including collateral for securing credit, a potential source o f income (rental) and home-based enterprises, and a means to reduce personal insecurity - particularly for those most at risk such as women and children. However, the cost o f conventional housing is frequently beyond the reach o f the majority of the poor. Often rooms are rented at exploitative rates and the poor state of accommodation contributes to vulnerability. A recent survey involving 3,600 household inKabul found that 56% o f households live in conditions where approximately 6 people share one room; 14% o f households live in partially destroyed houses; in 68% o f houses there are no windows or missing windows; 51% households have access to electricity but 98% o f households either have no access or cannot afford to use electricity for cooking and heatingpurposes. 4.19 For the past 23 years most land transactions in Kabul were unrecorded and many former landowners left their landunattended, to find upon retum that it hadbeen sold (many times) intheir absence. Over 75,000 destroyed houses require reconstruction (140,000 overall in the six largest urban centers o f Afghanistan). Almost half the residents of Kabul are under the threat o f eviction or relocation.'' As a result, poor households inKabul often have to move from place to place insearch o f living quarters. 4.20 Income and Employment. The growth ofurban centers attracting surplus labor from surrounding areas i s not always matched by corresponding public investment. This in tum can lead to restricted levels o f private investment and therefore future expansion o f employment opportunities. Many poor people work in the informal sector and with relatively insecure income flows. This often pushes people into debt in the absence o f efficient financial services. Inthe context o f Kabul, levels o f income are reported to fluctuate significantly. Income insecurity (as a result o f the dependence o f many households on informal wage labor) is a significant factor underlying poverty. 70 ACF (2004) 53 Box 4.3: The Challengeof UrbanLandManagementinAfghanistan "The thirdfloor o f the Kabul municipality office is a scene o f absolute chaos. There is hardly any room to move or breathe as hundreds o f desperate Afghans cram the comdors, all shuffling and jostling in their efforts to reach the property office and discover if - at long last - they have been allocated a precious piece of~and".~' Rapid urban growth constitutes a major challenge inAfghanistan, compounded by the massive influx of returningrefugees and displaced people. The urban challenge is all the more daunting as a result o f the backlog o f land and housing needs caused by a combination o f years on turmoil and a rigid master planning system, a deteriorated housing stock, and pressure to make more land available for housing. A mismatch between land and housing prices and income levels, outdated and incomplete land registration system, and pressure to resolve disputed property claims are adding to the complex nexus o f urbanissues facing the country. Source: Wiley (2004) `Urban Land Managementin Afghanistan:The Case of Kabul'. 4.21 Social Organization. Social capital (informal networks) is very differentinurban areas o f Afghanistan as compared with rural areas. The urban context can offer greater opportunities for social organization, but can also serve to alienate specific groups o f poor people (non-home owners, returnees, certain ethnicities, etc.) Initial research in the Afghan context suggests the relative fragmentation o f social networks in urban areas due to accelerated change inrecent years, regular shifts o f accommodation among non-home-owners and many ethnicities, frequently leading to limited social support for insecure households living with extreme poverty. UrbanRisks and Shocks 4.22 The NRVA data confmed the significance o f covariate shocks such as drought, pests and other natural shocks for a high proportion o f the rural population. Similar data are not available for the urban population in Afghanistan, but it mightbe expected that such shocks would generally be felt more indirectly, for example through price fluctuations, reduced seasonal farm based employment for urban household members, etc. Other potential covariate shocks affecting segments o f the urban community might include epidemics, violence and a sudden influx o f refugees into a particular neighborhood. The annual onset o f the winter season, while not an unexpected shock, cannevertheless constitute a serious riskperiod(with seasonalbunching o f shocks) for the extreme poor. 4.23 Idiosyncratic shocks such as loss o f employment opportunities or reduced income, the illhealth or death of an eaming family member or domestic violence can also be expected to impact on urban households, although due to lack o f data, the identification and quantification o f shocks cannot be completed. For many households, idiosyncratic shocks ultimately represent the final human outcome o f the general risks and recurrent shocks associated with urban poverty: unreliable and erratic labor markets; hazardous living environments; insecurity of tenure and shelter; lack o f health care; food insecurity; pressure on intra and inter household relations and overall a sense o f powerlessness and depression. Perhaps the complexities o f urban vulnerability most clearly highlight the challenges associated with simple causal models based on shock (cause) and impact (effect), as shocks lead to outcomes which themselves come to constitute shocks with additional impacts (increased household food insecurity, greater reliance on child labor, etc.). Who are the Vulnerable? 4.24 Urban vulnerability can manifest in many ways. This makes the application o f broad categories such as widows, people with disabilities, retuming refugees, children, etc., highly problematic, given the significant degree o f heterogeneity within such categories. Nevertheless, certain classifications may allow for correlation with a higher probability o f risk exposure. Drawing on existing studies, the following three factors emerge as relevant indicators to gauge household vulnerability: (i) location o f residential area (inso far as it conditions employment opportunities, services, access to shelter, environmental and sanitary conditions, etc., and also as often reflected in levels and sources o f income); (ii)social capital (informal networks); and, (iii) dwelling is owned or rented. whether 71Abdel Nasser Saeed(ARR No. 20,2002), quotedin Wiley (2004) 54 4.25 On the basis o f these studies, it is apparent that the characteristics o f urban vulnerability can be quite different to that experienced by rural communities. The analysis indicates that while poverty expressed by household food insecurity may appear to be relatively less significant inKabul for example, health risks appear to constitute a more critical dimension o f vulnerability, for both poor and better o f f households. In the context o f urban Afghanistan, the main factors determining structural vulnerability include access to health services, safe and regular water supply and sanitation and acceptable housing. At the same time, a number o f factors can be identified as determining the inherent vulnerability o f socio-economic groups. These include being a woman o f childbearing age, lacking regular income and being an insecure tenant. 4.26 Cumulative deprivation leads to higher levels o f vulnerability. Different groups (gender, age, ethnicity, etc.) can be more vulnerable than others. For example, exacerbated personal insecurity is a central dimension o f poverty for many urban women and children. Children, the disabled and the elderly may be highly vulnerable to ill health and dependent o n the household for su port In an urban context, the exposure to risks such as disease and accidents may be higher than in rural areas.' Family problems can result in homeless and destitute individuals (street children, the mentally ill,the elderly and abandoned women). 4.27 Thousands o f IDPs and returning refugees have chosen to reside inurban centers rather than return to their original homes, despite the fact that the overwhelming majority came from rural areas.73Thus an important portion o f the population living in cities are still rural in their mindset, cultural practice and social behavior. However, a recent vulnerability assessment undertaken by ACF in Kabul (2004, see B o x 4.4) established that on the basis o f community perspectives and experiences, the returnee status o f a family does not appear to be a significant determining factor of vulnerability in itself, although the rapid influx o f households may constitute a significant shock for existing residents and the most disadvantaged among the new arrivals. Additional studies have also concluded that returnees are not necessarily the most vulnerable fringes o f the population (Rodey, 2003 and Grace, 2003). Many returnees amve with assets, existing social networks and have at least one able-bodied male adult in the household. 4.28 Income/Food Insecurity. The relatively dynamic urban cash economies of present day cities in Afghanistan are dominated by the informal sector. Employment opportunities appear to be greater in Kabul than in other cities, particularly in the construction sector. However, while petty trading and small business expand, waged income is often unstable and there canbe intense competition for resources. Clearly the majority o f the population in urban areas are dependent upon the purchase o f food to meet the daily requirements o f household members. Of 526 households surveyed recently in Kabul (ACF, 2004), 53 percent a f f i e d that they were able to access a larger amount of food than during the same time last year. However, at the same time, food clearly continues to be a primaryconcern (ranking third after income insecurity and tenure insecurity). When asked to give the main reasons for taking loans, food security was cited as the primary purpose (followed by health expenditure). Household interviews confirmed that decreasing consumption o f food (particularly eggs and meat) is a common phenomenon, as is the purchase o f food on credit. Kitchen gardens provide an important component o f food availability for many households, but tend to be less prevalent for poorer groups for whom overcrowded residential areas do not allow for such ~pportunities.~~ An assessment o f households inpoorer districts o f Kabul found that 13% owned livestock to supplement their diet. However, for many households in these areas a typical meal was reported to consist only o f bread and tea. Eighty-five percent o f women reported eating fruit less than once a week, 49% less than once a month. Ninety-seven percent o f respondents consume dairy products less than once a month. The daily expenditure on food for 62% families is between $1 - $2/day, and for 36% families, less than $ l / d a ~ . ~ ' ''The concentration of people inareas where the provisiono f water, sanitation, garbage collection and health care is inadequate, creates conditions where infectious and parasitic disease thrive and spread. At the same time, highpopulation densities can potentially provide opportunities for improved health of the population. 73 Wiley (2004) 74 ''ACF (2004) CARITAS Germany (2003) 55 Box 4.4: KabulVulnerabilityMapping A recent exercise aiming to better understandrisks andvulnerability inKabul, based on amapping o f different livelihood zoneshighly vulnerable neighborhoods to ensure sensitivity to specific communities' needs, generated the following findings: 0 Neighborhoods that are effectively excluded from even the limited provision o f existing municipality services are frequently the most vulnerable. Eleven highly vulnerable neighborhoods have been identified (gozars), which usually host the highest population density and represent up to 60% o f the total population o f Kabul. The most vulnerable o f these are the hillside communities andmost recent settlements indry river beds. 0 Respondents inthe assessmentexercise consistently expressed their frustration and sense o f powerlessness with regard to the functioning of the health system in the city, particularly with regard to regular referral to private practices and difficulty to access health services in emergency situation such as childbirth. Maternal health and child delivery is a priority concern o f residents. 0 The recent influx o f population is reported to have put a significant stress on an already fragile water and sanitation situation. Inmost of the neighborhoods, garbage and night soil collections are either over stretched or non-existent. 0 Water consumption and access, particularly among the hillside communities, raise serious concerns. 0 Sixty percent o f households spend over 25% o f their income on rent and this drain on purchasing power frequently results inthe inability to obtain sufficient fuel for the winter. 0 Households living in rented accommodation had a very high turnover having moved on average 1.58 times inthe last year. 0 Incertainareasofthe city, morethanhalfallhouseholdsliveina 0 34% o f the main income earners in households rely on irregul householdhighlyvulnerable to s 0 Insecurity o ftenure and fear o fe exacerbates the stress of daily life. 0 Nevertheless, more than half o f the sampled population reported having more to eat now than a year ago (19% said they were eating less). Source: ACF AfghanistanKabul Vulnerability Ma~ping.'~ 4.29 Uncertainty and irregularity o f income are expressed as a major concern both for daily laborers and civil servants alike. Results from a survey o f over 500 households indicate an overwhelming reliance on daily waged labor throughout the city. One third o f all primary income earners are unskilled waged laborers. Construction workers interviewed stated that they currently earned 120Af. plus two meals per day but that they would be happy to take a cut inpay levels inreturn for the security o fregular income. Civil servants interviewed stated that they had not beenpaid for four months. Factors limiting livelihood opportunities and levels o f disposable income can include the following: high number o f dependents (or all dependents) in household; no literate members o f household; few employable skills, no assets (productive or saleable); limited social networks; indebtedness; no savings; poor urban en~ironment.~~ Inone survey among the Kabul population, 67% o f households reported earning a daily income of between US$1-$2, with a 50% reduction duringthe winter season.78 4.30 Winter. The winter season in Afghanistan results in heightened vulnerability for many people in urban areas, mainly as a result o f exposure to cold temperatures; less disposable income due to reduced employment opportunities and higher expenditure requirements (fuel, clothes, blankets; shelter; food, etc.). However the majority 76The study was basedon apurposive samplingo f 9 representativeneighborhoods(based on CSOIAIMS mapping)within the city of Kabul, sharingsimilar characteristicsin terms of income, access to infrastructure,services, housing, etc. Fifty household interviewswere carriedout in eachneighborhood, together with severalgroupdiscussions.Fieldwork was carriedout during the final quarter of 2003. 77Ibid. 78Ibid. 56 o f those at heightened risk during winter are also those who remain vulnerable throughout the year. Winter vulnerability i s therefore largely a matter o f poverty coupled with poor housing conditions. As shown inTable 4.6, classification o f the urban population in Kabul, based on residentiallretumee status over time, indicates that those who have resided inKabul on a permanent basis share a similar level o fvulnerability with recent returnees. Table 4.6: Factors affectingKabulHouseholds' Susceptibility to Threats Associatedwith Winter Households who Householdswho Householdswho Refugeereturnee havenever left have livedin have arrivedin households Kabul Kabul for one Kabulless than a year or more year ago from another province Living in damaged 78% 68% 67% 73% shelter Burning garbage 6% 28% 8% 13% for fuel Owning no heater 0% 16% 42% 20% Risk o f eviction 38% 38% 17% 53% due to squatting or renting from non- family Owning no 87% 84% 92% 90% productive assets Owning no other 81% 69% 67% 67% valuable assets 4.3 1 Child Malnutrition. Results o f the anthropometric surveys undertaken inKabul from 1995-2003 (Table 4.7) show high levels o f chronic malnutrition (stunting) for children under five and significant seasonal variations in acute malnutrition (wasting). The results point to a marked improvement in nutritional status since 2002, although the absolute figures remain high. The fact that acute malnutrition peaks have occurred during the summer o f 2002 may be correlatedto the marked increase inthe incidence o f diarrhoeal disease at this time o f year.79While it is not possible to compare these figures with those for rural areas due to the lack o f reliable and representative anthropometric data nationwide, a comparison with robust surveys in, for example, Badghis Province'' suggests that the incidence o f chronic malnutrition in Kabul is fairly similar, while the incidence o f acute malnutrition may be higher inKabul than that among rural households. Additional data on estimated average household income inKabul has also been shown to correlate closely with the incidence o f households with diarrhoea inyoung children. (ACF, 2004). 4.32 Vulnerability of Women. The findings from a recently undertaken assessment o f women's vulnerability inKabulhighlight some dimensions ofthe risks, which many women have to manage." The assessmentfound that: 16% o f all households surveyed were female headed; 22% o f the women had been refugees and that 91% o f the women were illiterate. The most significant challenges reported were meeting daily food and (during the winter) fuel requirements; 49% o f all women identifiedfood insecurity as the predominant experience o f poverty. 79 This causal relationshipi s suggestedby figureson diarrheacases in 2002 and 2003 gatheredat ACF supplementary feeding centers, with increaseanddecreaseofdiarrhea cases being slightly earlier thanthat ofmalnutrition. (ACF, 2004) UNICEF (2002). CARITAS Germany(2003) The Assessment involved the use of structuredquestionnaireswith 3,600 women from 10 o f the poorest districts ofKabul.The purposeof the exercisewas to enhancethe understandingofshocks, vulnerability andrisk managementstrategieso fpoorwomen as articulatedby the women themselves. 57 Table 4.7: Results of Anthropometric Surveys in Kabul (1995 - 2003) I Time of Kabul Survey Global Acute Malnutrition (% Global Chronic Malnutrition inZscore) (% inZscore) November 1995 6.2 -- May 1996 6.7 -- December 1996 5.1 55.2 June 1997 6.8 52.1 December 1997 7.5 63.O February 1999 8.7 61.3 February 2000 2.8 53.7 October 2000 8.0 57.5 March2001 3.8 62.6 August 2002 11.7 47.9 INovember August 2003 8.1 34.4 2003 44.1 Source: UNICEF (2003) 4.33 The material poverty o f Afghan women is often compounded by cultural restrictions on the roles they can play, most obviously apparent in the high degree o f immobility faced by many. Women without an able-bodied husband or male household head are structurally the most vulnerable with low levels o f income and the highest share o f long term illness. Fifty-eight percent o f divorced or widowed women reported that a male family member had taken over household responsibility. Inmany cases women are dependent on the approval o f men (even when this is not their husband) for financial resources, food, shelter, heating, clothing, medical expenses, etc. Itis c o m o n for only male household members to purchase household requirements from the market. Women report that men often prevent dependents from seeking medical care, either for themselves or their children. They are often socially and politically invisible. Widows and divorced women are frequently subject to social exclusion, isolation and harassment making it very difficult to secure sufficient income and resources to maintain minimum household welfare. 4.34 Maternal Mortality. The tragedy o f maternal mortality in Kabul highlights, inparticular, one dimension of women's vulnerability. I t has been estimated that 1out o f every 6 women o f childbearing age inKabul dies due to complications related to pregnancy and childbirth.82The cost o f private midwife services (reportedly around US$20) practically excludes poor households from ensuring some provision o f qualified assistance for child delivery. 4.35 Mental Health Problems. Mental illhealth is also emerging as a serious issue for a significant proportion of the urban population in Afghanistan (Hunte 2003, Grace 2003). This can be due to a combination o f factors including precarious and overcrowded living conditions, income insecurity, domestic violence, conflict related trauma and health worries (particularly reproductive issues for women). After more than two decades o f war, WHO estimates that 95% o f Afghans have been affected psychologically and that at least 20% suffer from mental health problems such as anxiety, depression, and other symptoms o f post-traumatic stress disorder? War victims, people with disabilities and women in general appear to be most at risk to mental health problemss3In a survey o f Kabul women undertaken by International Medical Corps (IMC 2003), it was concluded that 98% met the standard diagnostic criteria for post-traumatic stress syndrome, major depression or severe anxiety. Related to the combination o f poor mental health and socio-economic insecurity, is the phenomenon o f rising drug addiction in Afghanistan (Box 4.5). '*UNICEF(2002) Furthermore,only 25% of newbomswhose motherhas died, are likelyto survive due to malnutrition during the first year of life. '3Schutte (2004) 58 Box 4.5: RisingDrugAddiction inAfghanistan A rising trend o f drug addition is reported, both among Afghans inside Afghanistan and those living in the refugee communities in neighboring Iran and Pakistan. The prospects for treatment are bleak given the acute shortage o f facilities for basic healthcare, let alone the complicated packages required for drug-abuse victims. According to UNODC, nearly 40% o f drug users within Afghanistan begantheir habit as refugees, before they returned, thereby highlightingthe significance o f drug abuse prevention programs for refigees before they return. UNODC report that despite a ban on drug use and its criminalization by the Afghan Government, drug use is o n the increase inpart due to easier access to opium, coupled with the highprevalence o f mental illness related to war-trauma. In July 2003, the first ever UNODC assessment o f drugs usage in Kabul (based o n interviews with 200 users and 100 specialists), showed that heroin opium, alcohol and pharmaceutical drugs are being used by thousands o f people across the city, with a significant proportion o f women and returnees among the users. The report estimated a minimum 60,000 drug addicts out o f Kabul's estimated 3 million populations, while acknowledging that actual numbers were bound to be much higher. It is worth noting that there are estimated to be nearly one million drug addicts in Iran and half a million inPakistan. Untilrecently the consumption of hard drugs in Afghanistan was not widely apparent with taking drugs associated with social stigma, but this situation may well be changing fast. Urgent measures may be neededto curb this trend. Source:UNODC (2003). 4.36 PersonalInsecurity of Children. The particular experience o f vulnerability associated with being a child was explored in a study o f children in Risks and shocks identified by children and their families included the following: political repression, war, displacement, poverty, family loss and separation, family tensions, physical illness, danger from the physical environment, heavy and exploitative work and peer relationships. Perceived shocks therefore include economic, environmental, political, relational and personal. The negative consequences o f these threats affect children's overall social and emotional wellbeing, including their social development, morality, behavior, mental health and the opportunities they are able to access. When discussing the impact of conflict o n their lives children identified the most damaging consequences as physical destruction, lost opportunities for education, the threat o f displacement and the distress o f witnessing fighting. 4.37 It was clear from children's discussions that poverty exacerbates other dimensions of vulnerability. For example, poverty was reported as increasing tensions in the home, forcing children to work o n the street, pressuring parents to pulltheir children out o f school, marrying daughters at an early age, etc. Children expressed various ways that poverty increases the negative effect o n emotional and social relationships. Families saw their economic situation as a key factor in their ability to cope with difficulties. The significance o f personal relationships for borrowing in cash and kind is clear, although children stated that it was better if assistance came from within the family. Many reported that one o f the things they disliked doing most was being sent by parents to borrow from neighbors. 4.38 One o f the greatest fears expressed by children and parents was about the risks faced by young people when outside the home and the threat o f disappearance or kidnapping. There is a widely held belief that children are commonly the victims o f kidnapping for sale or organ extraction. Adolescent girls talked about the pervasive fear o f insecurity. Parents also feared the possibility o f sexual abuse of children. Research suggests that despite parents concerns primarily for their daughters, sexual abuse is statistically more o f a risk for boys than girls inAfghanistan. 4.39 For girls, the impact o f the Taliban era was particularly acute, given the restrictions on their movement and attending school. Many girls talked o f the disappointment and frustration they felt every morning as they watched their brothers go to school while they were left behind. Girls recognized that there were elements o f protection inthe way that their families restricted them and took special note o f where they went and whom they met. However, at 84 In2002 SC/US, with support fromUNICEF, conductedasignificant consultationwith childrenandtheir familiesliving in Kabul. The aim ofthe research, usingprimarily qualitativemethods, was to listen to children's views andto access their experiencesof daily life inorder to guidethe development of support programs for war affectedchildreninAfghanistan.More than 600 peopleparticipatedinthe groupdiscussions. 59 the same time they talked about frustration over the severe restrictions placed on their movements and opportunities and the high probability o f suffering depression and anxiety as a result o f confinement within the home. Girls also mentioned special disappointment if they had been in school for sometime but were then stopped from going because their parents felt like they were getting too old to be out inpublic. Girls regrettedthe loss o f opportunities for education and the hope for their futures that education gave them. Adolescent girls were also worried about their ownpersonal security. Copingwith Risks and Shocks 4.40 The results o f household interviews and focus group discussion^^^ in Afghan cities have highlighted the following coping strategies as routinely employed to guard against various dimensions o f economic insecurity: purchase o f food on credit; borrowing money from informal networks; child labor; sharing accommodation; renting out rooms in own houses; reducing food quality and diversity; women secretly saving portions o f daily 'housekeeping' money received from male householdhead; sale o fjewelry; sale o f assets such as carpets. 4.41 Informal credit was reported as the most frequent strategy used to cope with shocks. According to the data gathered, debt burdens frequently reach 38,000 Afs. (Le., about $760). However, a common theme emerging from qualitative assessments o f urban society inAfghanistan, particularly Kabul, relates to the perception that there has been a weakening or breakdown of support networks and informal safety nets among poor communities as a result o f decades o f difficulties endured by residents. For many households, it is not possible to keep up with traditional social activities involving hospitality and gift giving. However, at the same time, a poor household living in one or two rooms only may face no choice but to accept long-term guests, especially relatives, who arrive in Kabul. For recently arriving poor households, there is a reduced probability o f being able to access informal credit from shopkeepersthat are so often requiredto make basic food purchases and meet daily needs. 4.42 As a result o fthe pervasive lack of secure employment inurbancenters, householdmembers may generally engage in multiple entrepreneurial or waged productive activities, and combine intermittent daily labor with petty trade. However, female household members face more restrictions with regardto the employment Opportunities they can access, leading either to household dependence on fewer wage earners, or inthe absence o f an able-bodied adult male, to extreme poverty and destitution. Such is the pressure on ensuring the social/cultural standing o f young unmarried women and girls (with home confinement an important indicator o f human capital), that they are rarely involvedinincome generation outside the home. Three household case studies are presented inTable 4.8 to provide a better illustration o f the multiple but insecure and low value opportunities harnessed by poor people in urban communities o f Afghanistan to live from day to day. Income from a variety o f sources is pooled within the household and can help to smoothnecessaryconsumption, particularly for food, health and fuel related expenditure. 4.43 As noted in The Children of Kabu?', poor households inthe capital city frequently depend on the financial contribution o f their children to meet household expenditure. Child labor is also important for the collection o f alternative fire materials from refuse. Young boys from the age o f 10 can be highly active, undertaking petty trade (e.g., street vendor), bicycle repair, etc., often managing to combine work with school attendance. Older children (of both sexes) were found to be highly aware o f their potential ability to bring additional income into the households. Children inKabul were familiar with the experience and consequences o f poverty. For many, the impact o f poverty was described as not being able to afford medicine or treatment when a member o f the family falls ill.Many o f the children interviewed described how they experienced their own or family members' sickness as an economic shock -anevent that canpotentiallyplacecripplingfinancial pressuresonthe household. The anticipationofbecoming sick and not being able to afford treatment was expressed as a key worry by children. Many children inKabul were able to supplement household income through carpet weaving, a skills that many have learned from Pakistan (a significant number o f boys have been sent to Pakistan to leam carpet weaving skills). Children reported that the mainproblemwith carpet weaving was the impact on health. 85 Hunte and Schutte (2004) 86 SC USRJNICEF (2003) 60 Table 4.8: Urbancase studies to highlightthe copingeconomy ofUrban Afghanistan HouseholdStatus HouseholdComposition HouseholdEconomy Returnedrefugeehousehold Mother 55 years Son regularly seeks work as from Iran Son 22 years unmarried daily laborer. Eams around $2 (considered head o f household) /day but work is intermittent. Daughter, 25 years, widow, five H e also sells corn on the cob children inseason. Elderly mother crochets washcloths which are sold by son for $0.10 each Returnedrefugeehousehold I Mother 42 years, widow, head 13 year old son is the main fromPakistan o f household earner -he does masonry Two sons (13 and 10 years) work and eams $1.20Iday Daughter 12 years 10 year old son works ina bicycle repair shop after school and earns around $0.2Olday both sons sell drinking water inthebazaar and canmake $1.60Iday IDP household e Extendedfamily o f 3 25-year-old son is main generations with 11members earner. H e has secure e Patriarch 62 years employment as money e Matiarch55 years collector onprivatebus. H e e Daughter inlaw 25 years, earns $2/day, 6 days a week widow, 3 children Patriarch searches for daily e Son 25 years, with wife and labor and can make $1-2/day. daughter e Daughter 18 years unmamed -Son e 13 years unmarried Box 4.6: ChildLabor inUrban Afghanistan Child labor is a common form o f coping with shortage and irregularities o f income generating opportunities, often resulting in the erosion o f education and welfare levels. In some cases children may be the sole breadwinners o f the family. Young boys are frequently expected to find poorly paid employment, engage in petty trade activities, collect garbage for fuel, and to beg. Such activities, while necessary, ultimately incur disinvestment in human capital, thereby reducing the range o f future options for children and their families. However, a significant proportion o f children appear to manage both school attendance and work. Jalal is 38 and lives in a squatter settlement on the outskirts o f Herat with his wife and 5 children. He i s a construction worker but unable to find work much o f the time. H i s three sons (14, 11 and 7) work as vendors (the eldest) and for the local blacksmith (youngest). His eldest daughter (16) weaves carpets but is suffering from very poor health as a result. The youngest daughter (9) assists her mother with domestic work. Inthe past Jalil migrated to Iran in search o f work but had a very bad experience there. He has decided that his eldest daughter must be mamed to reduce household expenditure and assist the family through receipt o f the bride price. I Source:QuotedinSchutte (2004). 61 4.44 Nevertheless, the strong desire for children to be educated appears as part o f a household's long-term risk management strategy and education is recognized as an investment inthe future. Ina recent assessment o f women's vulnerability, while over 90 percent of all women in some o f the poorest districts o f Kabul were illiterate, and despite pervasive poverty, over 60 percent o f all children were attending school. 4.45 With insecurity o f health, income and food, resources acquired through social relations are o f critical importance, often as a last resort. In urban Afghanistan, it has been noted that as a consequence o f widespread uprooting and poverty, both inter- and intra-household relations are put under high pressure, which may eventually erode their potential value as safety nets. In focus group discussions, women were found to frequently report domestic violence against them and their chlldren, with aggression often accompanied by drug abuse o f husbands and elder sons. Heightened reliance on extended family networks is greatly valued, although often constrained by spatial dispersal o f members and the fact that relatives may be experiencing similar shocks at the household level. 4.46 Several social protection interventions are underway in Kabul including cash for work programs implemented by a number o f NGOs andpublic bread distribution (WFP UrbanBakeries Project). However, in-depth interviews with several poor households in Kabul have highlighted people's perceptions that without good relations to local government representatives (wakil) it can be extremely difficult to secure selection to access the benefits o f interventions such as participation in an NGO project or obtaining a card for food aid distribution. The residents o f Kabul involved in recent interviews have also expressed confusion and distrust regarding ways to capture much talked about intemational assistance. Accusations o f corruption at all levels o f intemational and national institutions are common. When assistance does amve from various sources, frequently the mechanisms for public information and transparent targeting are unclear. Variants, Sources andDeterminantsofUrbanVulnerability 4.47 Drawing on Moser's8' asset framework, Schutte (2004) and others, and given existing data constraints, it is possible to distinguish four significant variants o f vulnerability in urban Afghanistan as highlighted in Table 4.9: (i)vulnerability to income failure; (ii)vulnerability to food insecurity; (iii)vulnerability to bad health; and, (iv) vulnerability to social exclusion and disempowerment. These variants may complicate simplistic classification o f vulnerable groups incities and towns such as people with disabilities, widows, returnees, people livinginareas with relatively worse services etc despite the correlation between certain groups and exposure to risk. The linkage between these variants should also be highlighted, for example, the death o f an adult male may result inthe creation of a female-headed household characterized by increasedvulnerability to loss o f income and hence food insecurity. *'Moser (1998): The Asset Vulnerability Framework: ReassessingUrbanPoverty Reduction Strategies. 62 Table 4.9: Variants, Sources and Determinants of Urban Vulnerability inAfghanistan Income Failure Food Insecurity Poor Health Social Exclusion and (mental and Disempowerment physical) Source o f Vulnerability Exposure to an Decline o f Poor living and Powerlessness and unreliable and entitlements working conditions marginalization erratic labor market Determinant o f Quality of health Market Quality o f housing Quality of intra- and Vulnerability fluctuations inter-household Levels o f Education Confinement relations Income levels (restricted mobility) Household Resources to invest in Composition Education levels Overcrowding maintenance o f social capital Diversification of Access to clean Access to clean income sources water and waterisanitation Social affiliation and sanitation membership in various Income and education social networks. levels Education levels Access to health care Those who may be Peoplewith Households with Working children Macro: People most exposed to risks disabilities a high unconnectedto broader . (to be confirmed with dependency rate Women social structures improved data) Elderly Households with People living inpoor Micro: The `invisible': Widows and less diversified housingl such as people with female-headed income sources environments disabilities, women, and households the chronically sick. Women with People with disabilities disabilities, especially women with disabilities War victims 63 V. POLICYDIRECTIONS: TOWARDSA NATIONAL STRATEGY FOR POVERTYREDUCTION Introduction 5.1 The previous chapters of this report have shown that a large number of Afghanistan's people are poor, some are extremely poor, suffering from hunger, frequent illness, and homelessness. Others are vulnerable to poverty, and the frequent occurrence o f drought, political, economic and security instability, conflict, and natural disasters pose shocks that can easily send additional people into poverty. Inaddition, the failing state for more than two decadeshas left a country absent o f infraskucture, basic services and the core institutions o f a modem economy (such as banks) further exposing the people o f Afghanistan to poverty. Divesting or mortgaging assets - physical (land, livestock, housing), human (children) and financial - to help smooth consumption or to simply survive has been the order o f the day. This strategy not only has major implications for short term productivity, but it can have equally detrimental effects on sustainable growth and poverty reduction in the longer run. In addition, household strategies such as risk-aversion cropping patterns or income maximization in the short-run are not necessarily optimal for longer term growth andpoverty reduction. 5.2 Yet the people o f Afghanistan are not on the brink o f starvation, nor have they been inrecent history, due mainly to widespread personal and private initiative, and community support mechanisms. Afghans are resourceful, resilient, creative, opportunity-seeking, and entrepreneurial people (as witnessed by the high incidence o f labor migration, entrepreneurial activity wherever they are located, trading networks, and remittances). Their achievements in the face o f adversity are noteworthy. Given this spirit o f survival and independence, significant opportunities lie ahead as political stability retums, the economy begins to grow, and the institutions of state begin to function. Poverty and vulnerability in Afghanistan, while extensive, have been managed principally by people themselves, with little support provided from the outside. While sometimes resulting in detrimental coping arrangements, the way forward is nonetheless to build on the strengths o f private initiative and avoid displacing it with government interventions. 5.3 Sustained, equitable economic growth i s the essential path for poverty reduction - for in the absence o f growth, there will be little to share among the poor. Although poor at present, there are many people inAfghanistan who canparticipate ina growing economy ifgiven the endowments and opportunities - education and skills, health, jobs, business opportunities and credit, land and water. Even some o f the disabled, certain categories o f women, intemally displaced people - which include some o f the nomadic Kuchi - can, if equipped with assets and capabilities, enjoy a level o f participation in economic activity. Government has a key role to play in laying the foundations for sustained economic growth, equipping and assisting its citizens to participate and benefit from growth as the main avenue for poverty reduction, andputting inplace a minimumsafety net for those who genuinely cannot help themselves. The latter, however, needs to build carefully on community solidarity and mutual support mechanisms, which have proventhe backbone o f Afghan society for centuries. 5.4 This chapter outlines a strategy for poverty reduction. The focus is largely on the short-term, when the country will continue to face stability issues and the government will be laying the ground for a sustained economic recovery. With this in mind, the chapter discusses the elements o f broad-based growth, highlighting those areas o f the government's development strategy that are most critical to poverty reduction and growth. The chapter underscores the importance o f equipping the poor with assets and capacities to participate in growth, and singles out some elements o f the public investment program that are key to this. The chapter concludes with a review o f the poorest and most vulnerable who are presently falling outside government's social protection remit, and suggests some ways to assist thesepeople. The EssentialCharacteristicsof the Poor- Re-visited 5.5 To re-cap from previous chapters, the rural population can be divided into three group - the extreme poor, the moderately poor (or those at risk of becoming very poor), and the better off. Using this grouping, the essential characteristics o f the ruralpopulation are presented inTable 5.1 below. 64 Table 5.1: The Essential Characteristics of Three Groups of Rural Households The moderately poor The extreme or those at risk o f The better off poor becoming poorer Share inNRVA 20% 50-60% 20-30% sample Estimated rural 3.5 million 10.5 million 3.5 million population (approx.) Adequate calories No Yes Yes Problems meeting Yes Sometimes Rarely food needs year round Share o f cereals indiet 70% 60% 50% Households headedby 13% 6-7% 4% women Literate household 18% 24% 32% head Incidence o f disability 22% 12% 9% ~~ Land ownership Highincidence o f Some landlessness (20- Mostly landowners of irrigated landlessness 25%), some with land land; engage sharecroppers (40%); some (including irrigated) sharecropping Employment pattems Farmandnon- Waged labor as well as Self-employed farming & farm unskilled self-employment (farm business, and regular wage labor, daily waged and non farm) based emdovment Migration Some, mainly Most prevalent, Some, predominantly abroad short-term domestic and abroad domestic Access to electricity 11% 14% 25% Use o fhealthprovider Occasionally, Occasionally, most often Often (but not always), most when sick most likely traditional healer and likely to use private doctor but traditional healer, health center but private also traditional healer rarely health doctor for some center or private doctor Coping strategies - Reduce food Reduce food intake, opt Reduce quality o f diet, reduce what households do intake, opt for for cheaper calories. other expenditures. Sale o f when a "shock" occurs cheaper calories; Mortgageisale o f assets; assets, poppy cultivation tumto informal informal networks for networks for food credit; migration, poppy and credit; cultivation migration; poppy laboring ource: NRVA, 2003 5.6 Although there are similarities across the three categories, there are also some important differences. Of particular noteis access to assets - land, water, education, skiis and finance. Land, especially irrigated land is more concentrated, and an increasingly larger number o f rural households are landless. Access to grazing lands, fundamental to the livelihoods o f the nomadic kuchi, is also becoming more scarce. On- and off-farm employment, including job-related migration, is closely associated with levels o f education as well as social networks, both more effectively developed among the better off. Further, there is a close association between land ownership and non- farm business development, reflecting the absence o f capital markets and the reliance on own-capital for investment. There are also important gender dimensions across poverty groups, most significantly a concentration o f female headedhouseholds inthe "extreme poor" category. 65 5.7 Inthe urban areas, although less inknown about their situation, the poor are likewise dependent on low wage, irregular laboring jobs inthe informal sector. In the absence o f adult males, children often become the main bread winners for the family. Housing is a particular problem, and over-crowding is a frequent situation. The threat o f eviction or relocation is a common problem. Rent and fuel costs consume a large part o f incomes, leaving many urban households food insecure for all or part o f the year. Access to potable water i s an increasing problem, and statistics on morbidity and mortality are little better in the urban areas than in the rural areas. The poor face additional challenges by living in a high risk environment: accidents, personal violence, the kidnapping o f children and crowded unsanitary living conditions. 5.8 As to the location o f the poor, insights from the NRVA and other information sourcessssuggest that there may be some regional concentrations, such as a higher incidence o f poverty in the west and north-west o f the country, and less poverty in the north-east. But in all likelihood, there is heterogeneity in household welfare both within communities and across communities throughout the country. This poses a particular challenge for any form o f program targeting in the rural areas. (See Box 5.4 on targeting.) By contrast, in urban areas, location o f residential area (inso far as it conditions employment opportunities, services, access to shelter, environmental and sanitary conditions, etc.) i s considered an important factor is determining where the poor and vulnerable can be found. Stability o f income source, social capital, and status o f housing (owned or rented) are other proxy indicators of urbanpovertyhlnerability. Economic Growth: The Essential Path for Poverty Reduction" 5.9 The essential path for poverty reduction in Afghanistan is sustained, equitable economic growth. In the absence o f growth, and the conditions necessary to facilitate it, the scope for poverty reduction i s severely constrained. Thus, the key challenge for Afghanistan over the medium-termis to put in place the institutions, policies and services that will generate sustained, broad-based economic growth. As noted in the report, Securing Afghanistan's Future (SAF, 2004), in addition to the level, the quality o f economic growth will be critical: "the quality o f growth is very important. What is needed is growth with improving social indicators (reflecting investments inhuman capital) and without a significant deterioration inincome distribution. This ensures continuing broad consensus around the policies and decisions needed to create the enabling environment for economic growth. In the Afghan context, growth needs to be labor-intensive7 sustainable macro-economically and financially, environment-friendly and conducive to social development. Growth also needs to be reasonably well-balanced ethnically and regionally inorder to avoid exacerbating political tensions among different groups and region^."^' On this basis, growth will hlly support poverty reduction. 5.10 Afghanistan has areas o f good potential for growth, ranging from agricultural production and processing to mining, construction, trade and other services. Exploiting Afghanistan's geographical position as a "land-bridge" between Central and South Asia can also be a source o f growth, through trade, transit and joint development o f resources. For these and other growth potentials to be realized, however, the Afghan economy will need to move beyond its present, largely "informal" character. The informal economy has been a coping mechanism for Afghan's survival during conflict, flexibly responding to shocks and short-term opportunities, and has shown considerable dynamism. On the other hand, the informal economy is the outcome o f conflict-related insecurity, short time horizons, erosion of rule o f law, and lack o f public services which have influenced entrepreneurs' decisions toward staying informal, relatively small and not making longer-term investments. Realistically it will take many years for the bulk o f the economy to become formalized, but a core dynamic formal sector should be supported as a matter o f priority. This canbuildon and complement the legitimate activities o f the informal sector. 5.11 The government's SAF report puts forward an ambitious goal of 9% annual non-opium GDP growth over the next 12 years. Ifrealized, this will have important positive implications for overall poverty reduction. However, Afghanistan faces some major constraints in sustaining such a growth rate, which will have to be led by a strong, competitive private sector. Key challenges include: (i)improved security and rule o f law; (ii) a pro-private sector regulatory framework, free o f corruption; (iii) an effective, accountable state; (iv) provision o f essential support See, for example, AREU (2003) "Three villages in Alingar, Laghman: A Case Study ofRural Livelihoods". 89 This sectiondraws uponthe World Bank's EconomicReport, "Afghanistan: State Building, SustainingGrowth and Reducing Poverty". September2004. 90SAF (2004), page 7. 66 services such as infrastructure (power, roads, serviced land, water), finance, insurance, business support services, agriculture extension and marketing, and quality assurance for export; and (v) development o f human capital skills, both technicaliprofessionalimanagerial for business leadership and a literate, skilled workforce. The most important and immediate steps for the Government are summarized inthe next few paragraphs. 5.12 Security, rule of law, and state- building. The first and most fundamental requirement for sustained, broad-based economic growth is to build security, an appropriate legal framework and functioning judicial system, effective law enforcement mechanisms, and a capable and effective State. This in tum will encourage investment, and provide an environment more conducive to business development. In the medium term, the stability, security, and prosperity o f Afghanistan will also depend on the ability o f the State to raise the resources needed to fund the provision o f necessary public services. 5.13 Macroeconomic policy framework. A stable macroeconomic framework is very important for growth. The Government's intention to sustain recent achievements in fiscal, monetary, and exchange rate management and deepen budget and public administration reforms is critical. More generally, the policy framework will need to be responsive to potential shocks affecting the economy. 5.14 Provision of infrastructure and services. Another key element o f a strong growthstrategy and especially important for poverty reduction is effective delivery o f infrastructure and services, including, business services, micro-finance and research and development, especially in agriculture. These all have significant extemalities in terms of increasing productivity and raising incomes. 5.15 Enabling Regulatory Framework and Structural Reforms. A fourth element of the growth strategy seeks to improve the enabling environment to set clear rules for entrepreneurs and not require them to go through endless steps in the bureaucracy to comply with these rules. Regulatory and structural reforms are needed in a number o f areas: product and financial markets, tax, and customs. 5.16 Agricultural Development. Agriculture, half o f the Afghan economy, is critical for future growth, poverty reduction, and export development. Maintaining robust agricultural growth requires adequate investments and a sound enabling environment, as with any business. But agriculture faces additional business risks: the weather, especially lack o f water; pests and diseases affecting crops and livestock; and price volatility. (See figure 3.5 in chapter 3.)Key priorities to promote agricultural growth include: Irrigation: the Government's emphasis on rehabilitation o f surface irrigation facilities and subsequently major new investments i s appropriate given the significant variation in rainfall, the scope for more intensive cultivation and the need for water conservation. Attention to operation and maintenance o f irrigation systems warrants similar attention to the creation o f new infrastructure. Land tenure: pattems o f land ownership and land use are highly complex in Afghanistan, and can be controversial to resolve. At the same time, the absence o f a legal framework and institutions to deal with matters o f land tenure result in insecurity and are a potential constraint to investment. Government will need to proceed carefully in this difficult area, involving local communities in the search for workable solutions.; Research and extension: dissemination o f improved agricultural techniques and management practices, suitably adapted to Afghanistan, will be essential to realize the full benefits o f irrigation and other investments. Both public and private sectors have a role to play, working closely with communities, and focusing on marketing as well as production.; 67 Marketing: Investments in market facilities, cold storage, infrastructure (roads, electricity, telecommunications etc.), quality support and market management will all be important. Rural Credit: Building on micro-finance provided through MISFA,91 larger levels o f credit will be needed for agriculture and agriculture-related investments. These are important medium-term goals, and are best seenas part o f the development o f a broader financial systemandrural finance market. 5.17 The Government has already made a very good start by maintaining macroeconomic stability and initiating some important structural reforms, as well as capacity building inthe public administration. These very successful and important initiatives need to be complemented by capacity building at all levels o f govemment and delivery o f key infiastructural and social services, by easing the regulatory burdenon the private sector and reducing the scope for corruption, and most urgently by improving security and strengthening the rule o f law. If successful, these policies will pave the way for wholesale poverty reduction throughout much o f the country and for a large share o f the population. Equippingthe Poor with Assets and Capacities for Growth: Public Investment Priorities 5.18 Sustained over some years, economic growth can increase opportunities for most Afghan people, including poor people in both rural and urban areas. Its impact inthis regard will depend on how well equipped poor people are to participate in growth, and to what extent growth helps the poor to build their assets - physical, financial, human and social. Having the ability to accumulate physical capital such as (irrigated) land, livestock and small businesses, having access to (micro) finance and business support services, as well as finding employment opportunities will be fundamental to poverty reduction for people who are presently poor. Improvements in opportunities for poor people will also require building their human capital - education, skills and health - which is especially important for people who are dependent on their labor for a livelihood - such as the landless. Asset buildingon the part o f the poor needs to be supported by a range of well-targeted, cost-effective programs. While the overall trust o f the Government's development strategy provides a solid foundation for poverty reduction, greater emphasis will be needed in some areas to ensure that the poor are equipped to contribute and participate in a growing economy. 5.19 The Government's development strategy is embodied inthe twelve national development programs o f the National Development Framework,92 organized around three "Pillars" o f development: (i) and social capital, human (ii)physical infrastructure and natural resources, and (iii) sector development, including security and rule of private law and public administration reform and economic management. (See Box 5.1). All o f these are important foundations for poverty reduction. However, a number o f elements o f the government's investment program are most critical for the poor. 5.20 Public investment programs at the core of poverty reduction: Education and training: access to quality schooling for girls and boys is o f the highest priority for poverty reduction in the longer-run, as one o f the drivers o f growth. Inthe more immediate future literacy training for women, skills development and employment/self-employment services for a wide range o f people can play large dividends. Experimenting with different delivery models for schools and skills development training should be pursued. Health and nutrition: improvements inmorbidity and mortality will come from a variety o f interventions ranging from better matemal and child healthcare, a more varied diet, clean drinking water and improved sanitation, and education. Inthe short-run the emphasis should be on potable water and provision o f basic health services. As drought continues to plague much o f the country, the shortageiabsence o f drinking water in the rural areas has become a major concern, and warrants topmost attention o f the government. The provision of drinking water and modem sanitation schemes are drivers for a reduction in infant and child mortality, improvements inhealth more generally, learning inschool and worker productivity; as with education, provision o f health services can come from a variety o f different providers. Livelihoods and Social Protection: this consists o f a number o f govemment programs, operating through a variety o f parallel interventions: community level asset creation, short-term employment generation, '*Micro-financesupport facility inAfghanistan(MISFA). See Table 5.2. The Government is currently updatingthe NDF, which was preparedinApril 2002, under the umbrellaofan InterimNational DevelopmentStrategy. 68 micro-finance, and participation and inclusion in local decision making. Core elements o f this program include National Emergency Employment Program, National Solidarity Program, rural community water and sanitation schemes, area based development projects and micro-finance. All these programs are o f the highest important. (Table 5.2) Natural resource management: investment inagriculture for the settled population and rangelands for the livestock economy (important for the kuchi and women) will have important implications for the rural poor with some assets as well as the landless who work as agricultural laborers; provision o f rural finance, water management, rural infrastructure, tool and seeds, research and extension services are all important in the longer-run. Urban management:urban services, especially water and sanitation, solid water disposal, and housing are o fthe highest priority for the urbanpoor. 5.21 With these programs, the Government o fAfghanistan hasput inplace many o fthe elements o f good policy aimed at equipping the poor with assets and capacities that will support their active participation in a growing economy. Many o f these programs have the dual attribute o f being good for both growth and poverty reduction. The immediate challenge is to ensure adequate funding and effective implementation o f these programs. There is a danger that the government i s taking on too much at once, and the launch o f new programs - for example, those associated with poppy eradication - could result in limited fiscal and administrative capacity being spread over too many areas. A t this stage, data are inadequate to ascertain which o f the national investment programs are most cost- effective as poverty reduction instruments - the on-going Public Expenditure Review should begin to shed light on this. Inaddition, data onprogramparticipation, best gleaned fiomhouseholdsurveys, can assistwith monitoringand evaluation and provide insights for improved policy design. Inso far as choices have to be made as to "where" to implement programs, inthe absence o f a poverty map (see Box 5.4), usingavailable information as much as possible to focus on the poorer districts inthe poorer provinces, andpoor urban communities should take priority. Continued investment in collecting data and building knowledge about livelihood systems, the multiplicity o f risks, the heterogeneity o f households and the differences that exist between localities will be central to ensuring optimum returns to investments inpublic sector programs. The Social Protection Program 5.22 Some o f the commendable features o f government's livelihoods and social protection policies and programs include favoring productive safety netsg3such as labor intensive public works and other employment creation activities over direct transfer programs, strengthening livelihoods through education, skills development and micro-finance, providing health services, drinking water and sanitation, and working with and through communities. These directions are largely in line with emerging good practice fiom post-conflict countries. (see Box 5.2) 93Inthe longer-run conditional transfers (cash or in kind) to individual households and/or communities predicated on sustained investment in children's health and schooling might be an additional productive safety net instrument to consider. 69 Table 5.2: Three Successful National Programs and Poverty Reduction The NEEP program has generated around 10 million labor days through the establishment o f a nationwide labor intensive public works program for the construction o f rural infrastructure. This includes rural roads, bridges, irrigation schemes, water supplies, schools, clinics and shelters. In addition to these cash-for-work activities, a number o f parallel food-for- work schemes have been supported by WFP and implemented by NGOs. One o f the broader challenges facing NEEP at this time i s to clarify which o f the dual objectives o f (i) rural infrastructure development, and (ii) emergency employment generation should predominate, in so far as there may be tension between the two. If it is the former, that i s primarily an asset creation program, further consideration o f quality and durability o f the infrastructure created is essential, as well as the interface o f this program with other rural development programs. The NSP, Afghanistan's flagship intervention for community empowerment, provides resources through block grants for community managed small-scale reconstruction and development inrural areas. It also makes provision for grants to some o f the poorest individuals within the community. Its modus operandi promotes more participatory and inclusive decision making and governance at the village level. The target is to cover the country's approximately 20,000 villages over a four year period. As o f 2004, 5,000 village-level, community-development councils had been elected, and 2,350 sub-projects hadbeen submitted for approval. The establishment o f MISFA, as an umbrella agency to provide oversight, technical support and fimding to NGOs has proven a success. Since its inception in2003, MISFA has supported 12 micro-finance institutions and more than 80,000 clients in 17 provinces. Around 90% o f clients are women, with a repayment rate o f 98%. Given the absence o f financial services in both rural and urban areas, MISFA and the implementing partners have a very critical role to play inpoverty reduction. MISFA can also make an important contribution to promoting alternative livelihoods to opium, both as a source o f finance for alternative activities as well as a vehicle to break the viscous cycle o fopium creditlopiumdebt. 5.23 Two major challenges face the design o f social protection. First, how best to support both short-term emergency measures to deal with the immediate needs o f poor and vulnerable people as well as put in place the foundations for a longer-term social protection system, compatible with an economy undergoing reconstruction and development. Fortunately some programs, such as public works programs and the National Solidarity Program can serve both short and longer-term goals. Inother ways, these time periods require a different set o f social protection policies and priorities, and desirable programs such as a more comprehensive social safety net and institutions supporting flexible labor markets will have to wait till fiscal and administrative capacities are strengthened. In the meantime, govement should encourage humanitarian efforts to continue to provide some form o f a safety net. 5.24 The second challenge refers to how to prioritize among the many competing immediate needs for support: traditional poor and vulnerable groups (some o f the kuchi, rural landless) versus new conflict-related vulnerable groups (such as demobilized soldiers, war disabled, war widows and orphans, displaced people) versus new vulnerabilities (such as drought related, or where opium poppy i s being eradicated). There is no easy answer to this 70 dilemma, and it will be a matter o f balancing fiscal and administrative capacities with political priorities. However, as discussed above, although poor at present, many Afghans can and want to participate in a growing economy - even some o f the disabled, certain categories o f women, internally displaced people (which include some o f the nomadic Kuchi) - can, if equippedwith assets and capabilities, enjoy a level o f participation in economic activity. Moreover, spendingfunds on education and skills, health, jobs, business opportunities and credit today will generate returns tomorrow, and contribute to a growing national economy. Favoring social protection interventions that provide both "protection" inthe short runand underpinequitable growth inthe longer runare "win win" policies. Box 5.2: Emerging Lessons for Social Protection from Post-Conflict Countries As a number o f countries aroundthe world emerge from conflict, some lessons on good practice canbe identified. Many o f these lessons are o f direct relevance to Afghanistan. t Conflict leads to new groups of vulnerable people with a particular set o f problems: displaced, demobilized, and disabled persons; child soldiers, orphans and widows; t Tension can arise between providing scarce resources to the poorest/most vulnerable and new vulnerable people; t The destruction of data collection infrastructure, and the absence o f reliable data makes the formation o fpolicy and the monitoringo fprograms very difficult; t The largely informal nature of the economy, with few formal systems challenges service delivery; t Implementation of safety net and other public sector programs can be difficult where conflict continues, or insecurity to service delivery agents i s o f concem; t Butspacefor boldreformexists-anew erawith new optimism. Priorities for reform include: Labor market - the creation o f an environment that is attractive to employment creation and to equip people with the right skills and information to enter jobs. Demobilized soldiers, displaced and disabled persons can face discrimination - public works programs can help ease these barriers inthe short run until longer-term training and labor market interventions take hold. Post-conflict boom in reconstruction canbe a huge boost for employment; Social assistance - need to focus on the most vulnerable, avoid entitlements. Keep it simple because o f the shallow administrative capacities and limited funds. Categorical targeting can be a good method. Gradually introduce public social assistance schemes to replace humanitarian aid, but NGOs can continue to play a criticalrole inservice delivery; Social services - social workers should be trained to provide special support services, counseling and rehabilitation to vulnerable groups such as the disabled, former child soldiers, and women most affected by the conflict; Pensions - war puts pension systems under additional pressure as the number o f contributorstvalue o f contributions fall and new demands arise - such as the early retirement o f military personnel. To keep the pension system working and to make payments on time - an important aspect o f the credibility o f a government - keep it as simple and as transparent as possible. Pensions could best be o f a flat rate nature and calculated on a pay-as-you-go basis inthe short run, before raising individual pension levels and introducing differentiation inthe longer run; Social funds - there i s a good fit between the conditions o f a post-conflict country and the characteristics o f social funds (such as NSP). They are flexible and can be adapted to different circumstances and priorities, and can serve both emergency and longer-term development needs. They can provide much needed services and infrastructure at the community level as well as built social capital. These public sector interventions should build on those aspects o f informal and private coping mechanisms that are legal andnot detrimental to people's current and future welfare. Source: "Emerging from Ethnic Conflict: Challenges for Social Protection Design in Transition Countries". World Bank Social Protection Discussion Paper Series, December 2002. 71 5.25 Although designed to address the needs o f poor and vulnerable people, thus far core elements o f the social protectionprogram have proven more effective inreaching the broad population as compared to the poorest: see for example the information on participation in public works programs, presented in Table 3.14 in Chapter 3, which suggests a "welfare neutral" distribution o f parti~ipation.~~Given the importance o f getting programs up and running, the political imperative to show activity throughout the 34 provinces, the administrative challenges and the absence o f data to guide resource allocation decisions, these outcomes are not unsurprising. Moreover, many o f the existing programs, by nature o f their design, are not well equipped to reach the very poorest and most vulnerable - i.e., extremely disabled people, certain categories o f female headed households, displaced, nomadic kuchi. These categories o f people carry physical, spatial or social attributes that may prevent their participation in mainstream economic activities, especially activities requiring a laboring effort outside o f the home (disabled and women). Little can be done to change those characteristics, but at the margin the programs do have the potential to be more effective inreaching out to some o f the poorest and most vulnerable. Indeed operational manuals for programs such as NSP and NEEP make provision for greater outreach. For example, inaddition to the investments ininfrastructure, NSP block grants can be used for (i) self-help savings and credit schemes for women and disabled, and (ii) productive asset transfers for vulnerable women and disabled. T o make this effective, there is a need to develop clear methodologies for how this assistance can be best targeted andprovided Box 5.3: Key Design Features of a Good Public Works Program To realize the full potential o f a public works program as a poverty-reducing and risk-coping instrument the following are important design features: The wage rate should be set at a level which i s no higher than the prevailing low-season market wage for unskilled manual labor inthe setting inwhich the scheme is introduced. Restrictions on eligibility should be avoided; the fact that one wants work at this wage rate should ideally be the only requirement for eligibility. Ifrationingisrequired(because demandfor work exceedsthebudget available at the wage set) then the program should be targeted to poor areas. However, flexibility should be allowed infuture budget allocations across areas, to reflect differences indemand for the scheme. The labor intensity (share o f wage bill intotal cost) should be as highas possible. The level o f labor intensity will depend on the relative importance attached to immediate income gains versus (income and other) gains to the poor from the assets created. This will vary from setting to setting. The projects should be targeted to poor areas, and try to assure that the assets created are o f maximum value to poor people inthose areas. Any exceptions inwhich the assets largely benefit the non-poor- should require co-financing from the beneficiaries, and this money should go back into the budget o f the scheme. Public works should be synchronized to the timingo f agricultural slack seasons. Inorder to encourage female participation, the appropriate formo fwages is important - for example, women can benefit from piece rates or task-based wages; sometimes wage inthe form o f food has attracted more women to work sites. Also, provision o f childcare or preschool services can improve participation by women. Transaction costs to the poor are kept low-one important means to accomplish this is through locating project sites close to villages. It is also necessary to ensure appropriate mediation o fNGOs for protecting the rights o f the poor vis-a-vis program managers. The program should include an asset maintenance component. Source: K. Subbarao (2001) "Systemic Shocks and SocialProtection:RoleandEffectivenessof Public Works Programs". World BankAfrica RegionHumanDevelopment Working Paper Series. drawing on best practice lessons from other projects. Some pointers for improved effectiveness of labor-intensive public works programs (NEEP) can be found inB o x 5.3 (also B o x 3.2 in Chapter 3), and include consideration o f the types o f activities, their location, the wage rate and mode o f payment. The particular circumstances o f 94 Itis widely acknowledgedthat communitycapture of programresourcescan seem to be "welfare neutral" at first glance, but when evaluatedinbroadercontext, it may in fact be moreprogressive. 72 Afghanistan (such as the topography and the climate; the role o f women), however, do render some o f these pointers to be less realistic than inother countries. A Safety Net for the Poorest? 5.26 One o f the outstanding, more immediate - and most difficult - challenges for Afghanistan's social protection strategy lies ineffectively reaching the poorest andmost vulnerable people who are not able to participate inasset-creating, income-generating, labor-based programs described above?' yet who are least able to weather any further shocks. For example, when a winter such as the winter of 2004105 occurs, their very survival is threatened. The first questionto tackle is "who are the poorest and most vulnerable?". 5.27 The extreme rural poor, detailed as the poorest quintile in Table 5.1 are not an homogenous group. On average, however, this group exhibits a higher incidence o f landlessness and rain-fed small-scale farming (as compared to irrigated farming), less livestock ownership, less literacy, more female headed households and households with disabled members than inthe other welfare groups. Some o f these households may enter into share- cropping arrangements with land owners, with advances in cash andor kind. These can be highly exploitative arrangements under certain conditions. The extreme poor might actually reside inthe homes o f the land-owners as a way to enhance security of food and employment. The primary asset o f the extreme poor is labor and wage employment (farm and non-farm) i s central to their livelihood strategies. Insecure and low value daily laboringjobs (frequently paid in a combination o f cash and meals) tend to predominate as a result o f the relatively low levels o f humancapital within these households. The children of extreme poor households are less likely to attend school and often make an important income contribution. Households without members able to engage in gainful employment may have little choice but to depend on the charitable support o f relatives and informal social networks. Extremely poor households are highly vulnerable to sickness and premature deaths, have neither the means nor the ability to seek medical care when sick, face volatile employment opportunities and are potentially at risk from fluctuating food prices. 5.28 Inrecent times Afghanistan policy makers have designed programs for certain categories ofpeople inthe belief that they are among the poorest and most vulnerable. These include people with physical disabilities, female- headed households, the nomadic kuchi, IDPs and returning refugees, and highly disadvantaged children. The NRVA collected a variety of information related to a number o f such groups, and these data are presented in Table 5.3. These data allow a more detailed understanding o f the nature o f the poverty o f these groups, as well as a validation (or not as the case may be) o ftheir socio-economic positionrelative to other poor people. 5.29 Insummary, the datashow the following: Returned refugees and IDPs have a slightly higher incidence o f poverty compared to the overall NRVA sample, but all other indicators for this group are similar to the rural average. Infact, they are slightly better off interms o f land ownership, access to irrigated land, and literacy o f household heads. Data on refigees and IDPs who have returned to urban locations rather than to their original rural homes (the majority) further c o n f i i that these households are no worse o f f than others. Many retumees arrive with assets, including education and skills, existing social networks, and able-bodied men, and are well positioned to engage ineconomic activity andbenefit from a growing economy. Households with disabledpeople are visibly among the poorest along a number o f dimensions: they have higher poverty rates and food gaps than average, and less land ownership and access to irrigated land. There is also a higher incidence o f female-headed households in this category. However, by no means are all households with disabled people poor. Depending upon the severity o f the disability, how this impacts on both the individual as well as the household as a larger entity, the activities o f non-disabled household members, and existing household assets including social capital, there is considerable variation inhouseholdwell-being. Female-headed households have the highest incidence o f poverty o f all vulnerable categories. Although they also have a h g h calorie intake gap, the gap is less than some other categories, suggesting that perhaps these households do better than some inmeeting basic food requirement. This 95These includedisabledpeoplewith the moredebilitatingdisabilities;poor, marginalizedfemale-headedhouseholds; and destitute anddisplacedfamilies, manyof whom are Kuchi; drugaddicts; orphansand street children. 73 may well reflect targeting o f female-headed households in humanitarian programs as well as intra- community charitable arrangements. Female-headed households also have a higher than average presence o f disabled members. They are more likely to be landless, lack access to irrigated land, and have lower ownership o f other assets, such as livestock. Female-headed households tend to have restricted mobility outside the domestic compound, and restricted formal entitlements/control with regard to productive assets. Table 5.3: Selected Characteristics among "Vulnerable" Households households (%) Incidence o f disability 14 100 19 16 14 (%I Literate household head 27 20 20 27 24 (%) 81 74 56 0 76 Irrigatedland (%) ~~~~ 74 57 43 0 65 Livestock (# COWS) 0.7 0.8 0.5 0.3 0.7 5.30 The conclusion to draw from these insights is that there are indeed people in Afghanistan who can be recognized as very poor and vulnerable, and indeed the Government's use o f targeting by characteristic (categorical targeting) has, for most part, been appropriate given its ease o f administration. (See Box 5.4 o n targeting). However, it should be acknowledged that not all people in the categories - disabled, female-headed households, IDPs and refbgees (and Kuchi - not discussed here) - are equally poor, and that other people not included inthese categories, are among the poorest. This highlights the broad-brushed nature o f this method o f targeting. 74 Box 5.4: Targeting Targeting is the process o f identifying who is eligible to receive support, and devising a scheme to ensure that the intended support reaches its "target". This is an important consideration for government policy in poverty reduction(or provision o f any other publicly funded program) ifthe objective is to maximize the resource transfer to the poor for any given level o f expenditure. But targeting itself incurs costs - data and information costs, administrative costs (ensuring that the right people get the intended resources), as well as incentive effects at the level o f recipients. Behavior o f non-recipients and broader political support for narrowly targeted programs are also important considerations. Thus the choice o f targeting method is a complex interplay between administrative, fiscal and political considerations. There are five widely used targeting techniques for safety net programs: + Means tests: an assessment of the income o f the individualihousehold. Most accurate method where capacities to do it well exist (a high level o f literacy, reliable documents and extensive administrative capacity). A variant is proxy means testing, whereby more easily identifiable assets (housing, ownership + of durable goods) are usedas a proxy for income. Indicator-based, or categorical targeting - for example, demographic characteristic (number o f children), gender, disability, ethnicity, land ownership. An easy to administer approach but not very + accurate. Self-targeting - this is situation whereby participants decide for themselves whether or not they + participate inthe program (there is often a cost to participation, such as inpublic works program). Community-based - uses some form o f community representative, individual or committee to decide + who benefits from the program. Geographical - the program provides support to the residents o f a particular area. An important require for this approach is a good map which charts where there is a concentration o f intended beneficiaries. Source: Adapted from "Targeting Programs:Conceptsand Evidence". Subbarao, K. (2004) 5.31 In considering a more "finely tuned" safety net for the poorest and the most vulnerable inAfghanistan, there are a number o frealities that cannot be ignored: a Extreme poverty and vulnerability in Afghanistan are not well understood: their precise nature, the numbers o f people affected, their essential conditions and location are yet to be fully documented - this report has only begun to scratch the surface o f a very complex situation. a In addition to economic dimensions o f poverty, social and political conditions afflicting the most marginalized and disempowered communities, households and individuals in Afghanistan need to be addressed. a Institutional and fiscal capacities are extremely limited, thereby constraining the scope and affordability o f (new) programs. M u c ho f the implementation capacity lies with NGOs. a The existing cash transfer scheme for war-related disabled persons and the families o f martyrs, awarded irrespective o f poverty status, may have establisheda precedence for entitlement programs. In addition, some interest groups have established powerful lobbies, often fully backed by special interest ministries (e.g., disabled, martyrs, kuchi, women's affairs). N e w vulnerable groups, though arguably not among the poorest - such as retiring civil servants and displaced public sector workers - are also making demands on the national budget. a Hitherto, many o f the programs designed to reach the poorest and most vulnerable have been supported by UNagencies and implementedby NGOs - these typically exist outside the domain o f the national development program and the nationalbudget, and often provide in-kindsupport. 5.32 Currently under consideration in Afghanistan is a National Vulnerability Program (NW)which aspires to provide support to some o f the poorest and most vulnerable people. As its design moves forward, here are guiding principles for the shodmedium term: 75 Focus o n data collection and further analysis o f extreme poverty and vulnerability before launching any new programs. There is an unresolved tension between special interest groups and a poverty- basedassessment o fthose most inneed. Review the design and effectiveness o f on-going programs under implementation by NGOs, and assess the potential for scaling-up. 0 Try and influence the design and implementation o f existing programs to ensure that they are as inclusive as possible. Modify the existing cash transfer program for war-related disability and martyrs families as a matter ofpriority to ensure consistency intarget groups and modalities. Only as a large resort launch new programs. Programdesign shouldbe grounded inthe very limited administrative capacity and flexible indemands o n the national budget. 5.33 With regard to the launching o f any new direct (targeted) support to the extreme poor and most vulnerable members o f society, questions remain as to how best to target these individuals and their families. Self- targeting, categorical targeting and working through community "solidarity" channels remain the preferred mechanisms, although the extent to which community networks can be relied upon - if at all in urban areas - to effectively provide for the poorest and most vulnerable remains an open question. This could be an agenda item for the NVP as it develops a monitoring and evaluation capacity. More generally, continuing research o n the nature of poverty and vulnerability, and linking findings to enhanced targeting o f programs will be key to increasing the efficiency and effectiveness o f limitedresources ina very difficult delivery climate. 5.34 While there i s a public policy gap vis-a-vis a safety net for the poorest and most vulnerable, which the NVP i s tasked to address, there are a number o f schemes runby NGOs, and supported by UNand bilateral agencies which have provided humanitarian support over the years and helped prevent mass destitution and hunger. For example, food, cash and housing have been provided to returning refugees; relief food aid has been provided to IDPs, female headed households and disabled persons alongside food-for-work programs. One o f the key challenges for the government and the aid community is to increasingly bring humanitarian assistance into the framework of the government's development program, and ensure that it is managed as an efficient and effective element of the broader national development strategy. Food aid i s o fparticular importance to the country's safety net.96Inthe short term food-based safety net programs can make an important contribution to the welfare o f the poor if properly managed. These ideas are presentedinB o x 5.5. 96 While a resourcethat has played a critical humanitarian role inAfghan livelihoods for many years and which has the potential to shift into a developmental mode, food aid also has its own vulnerabilities: its impact on local food production, moral hazard allowing more farmers to shift into the cultivation o f opium poppy knowingthat if all else fails food aid will arrive, the huge implementation capabilities o fNGOs tied up in distributing food, and the counterpart cashhudget requirements. 76 Box 5.5: food-Based Safety Nets: Gettingthe most our of Food Aid Economic theory suggests that cash transfers are the preferred means o f assistance to address poverty since they are the most "pure" form. In the ideal world the right amount o f cash would be provided to those persons and individuals falling below the poverty line with accurate provision such that income (cash) poverty would be eliminated. There are, however, a number o f real world problems that prevent such an approach, among which the lack o f fiscal means and administrative capacity (to identify the poor and administer the programs) are paramount. Over time, as these binding constraints lessen, this increases the scope for using more cash transfers as the core o f the social safety net. International experience suggests that "productive" safety nets, that i s public works programs which create physical infrastructure and conditional cashtransfers, which support humancapital development are to be preferred over non-conditional transfers. Inthe short term, food-based safety net programs can make an important contribution to the welfare of the poor: food has a direct impact on welfare, as well as an indirect impact as it frees up income for other uses. The fact that women often control food-related transfers may explain the widely documented evidence that food transfers lead to a greater increase in households' net food consumption as compared to an equivalent cash transfer. Food-based programs can be the direct provisiono f food (supplementary feeding programs, food for work programs), or through cash-like instrumentstied to the provision o f food (food stamps, food coupons). Food-based programs can be more effective at reaching the poor than cash programs, especially ifthe food is inthe form o f inferior, less-preferred foods which are more heavily consumed by the poor (though not nutritionally inferior), Food-based programs can be self-targeting, in so far as only those most in need will be interested in the programs. The direct provision o f food is most valuable when food supply is disrupted, either because food is not available in the markets or access to markets is particularly difficult (such as during winter months). When there is evidence that markets are beginning to work, the provision o f food stampslfood coupons can be an important development (or indeedcash instruments not tied to food) so long as the increase indemand does not pushupprices. Implementation of food stampslfood coupons requires a reliable system for printing and distributing the stampslcoupons, and good banking system to assure that retailers can redeem the stamps for cash. As with supplementary feeding, which is usually given inconjunction with health care and schooling, food stampslcoupons are oftenprovided as an incentive for participation in a welfare program. Only in emergency situations do unconditional food-based transfers make sense, where they canbe the only resource a household receives and the difference between life and death. Thus food-based transfer programs can serve a variety of purposes, ranging from income support, guaranteed minimum level of consumption, improved participation in socially beneficial programs, and nutritionallhealth improvements. Costs vary widely depending on the size o f the transfer, the size and distribution o f the target group, and logistics, but can be very substantial. Cost-effectiveness, often hard to assess, is nonetheless an important consideration when food-based transfers are being considered - both across different food-based programs and vis- &vis non-food programs. A final matter is political support: it is often easier to gain political support for transfers in the form o f food as compared to cash-butby the same token it is often hardto dismantle such programs. Source:Food-BasedSafetyNetsandRelatedPrograms:World Bank SocialProtectionDiscussionPaper Series, September2002. Conclusion 5.35 Muchofthis chapter has discussed an optimistic scenario inwhich sustained economic growthprovides a "rising tide" which will carry the bulk o f the Afghan people to greater prosperity so long as the complementary investments inenabling capacities, services and infrastructure are made. This coupled with a very carefully designed safety net for the poorest, left inlarge part to NGOs, self-targeting and community based arrangements, maximizing the utility o f the food aid resource and drawing very selectively on Afghan budget resources, can make for a comprehensive poverty reduction strategy. But what ifthere is a major downtum inthe economy (perhaps the result of a successful opium poppy eradication program), or other covariate shocks - such another major drought, which impact negatively on lives and livelihoods occur; or more idiosyncratic shocks resulting injob loses or ill-health o f a 77 family member continue to throw people deeper into poverty. Afghanistan cannot prevent most o f these shocks; it cannot insure against all such risks nor can it afford to provide transfer payments to compensate those affected when shocks hit. On the other hand, the Government would want to intervene and try and avoid the worst forms o fprivate coping arrangements that have prevailed in the past when adversity struck. Finding the right balance between supporting the upside - growth and prosperity - yet being preparedfor the down side is an important consideration. The social protection strategy needs to be sufficiently flexible and adept so that it canprevent poor households from adopting coping arrangements which have negative longer-run consequences (liquidation o f assets, withdrawal o f children from school, child labor and early marriage o f daughters). 5.36 But let's not forget that Afghans are strong people and for decades Afghan society has coped with adversity. Afghans have coped with frequent drought, with civil war and conflict, with pests and disease. Strong community organizations, extended families and mutual support obligations together with entrepreneurship and opportunism have kept people alive. As Government is able to extend the scope and reach of its social protection system, it is critical that itbuilds on these strengths o f Afghan society. 5.38 A final word about the formulation of a national development (poverty reduction) strategy. A good strategy is not only technically and economically sound - it needs to be fully "owned" by the poor as well as by other stakeholders in the society. In this regard, the role o f civil society groups, including the private sector, to identify key policy, institutional, and infrastructure constraints, as well as to suggest ways forward, will be vital to establishing a poverty reduction strategy that clearly identifies the problems that it seeks to address and builds ownership across the society. The role o f both national and international NGOs in this process will be very important for enhancing ownership. Inthis context, the Government is appropriately planning to engage inextensive consultations with different segments o f society as it develops its (Interim) National Development Strategy, a process that is targeted to reach a first stage o f completion with the preparationo f an I-NDS later in2005. 78 Appendix I: and Sampling Issuesg7 Data Since taking office, the Transitional Government o f Afghanistan has articulated a clear strategy for the country's reconstruction and development, and through the National Development Framework, a number o f programs and projects have been put in place. The Government committed to ensure that the benefits o f reconstruction and development are felt by the entire population, particularly the poor and vulnerable. However, to- date, lack o f data as well as institutional and capacity constraints seriously hinderedempirical analysis to understand the complex issuesrelatedto reconstruction, growthandpovertyreduction. T o address these limitations, a data collection effort led b y the Ministry of Rural Rehabilitation and Development and supported by a number o f the international development agencies was implemented during the summer o f 2003.'' The data were collected at a time when the country was experiencing an unusually bumper harvest as a result o f the end o f three years o f drought. The National Risk and Vulnerability Assessment (NRVA) 2003 data collected a variety o f community- and household-related information aiming at identifying areas for policy intervention and welfare monitoring. Despite various limitations and constraints, the data offer one o f the first comprehensive insights on rural welfare in Afghanistan. In fact, the NRVA 2003 is the first dataset from Afghanistan that measures consumption poverty and explores various issues related to the many dimensions o f rural poverty and welfare. The National Risk and Vulnerability Assessment (NRVA) 2003 was the result of a multi-agency collaboration. It was ledby the Ministry o f Rural Rehabilitation andDevelopment and implementedjointly withthe WFP and analytical contributions from the World Bank, FA0 and other agencies. Its main objectives were to: (i) support the government's efforts to derive information on rural poverty, vulnerability and risk management and fill the large information gaps that exist; (ii) improve the design of social protection programs and other policy outputs on a systematic basis; and (iii) provide longer-term partnering and collaboration on poverty and risk analysis and enhance institutional development and capacity building thus allowing for ownership and sustainability o f the analysis and results. Afghanistan's last census was implemented in 1979. While the Central Statistical Office (CSO) i s currently conducting a pre-census exercise in Afghanistan, the immediate lack o f credible population census data have hindered the ability to design an instrument that was based on a random population-based sample framework that would allow estimating national or sub-national representative statistics. Instead, the sample design was implemented intwo stages. The community selection phase was done based on agro-ecological zoning and estimated land areas per zone. Specifically, during the first stage, all known communities within each district were classified into 1o f 5 agro- ecological zones (irrigated, mainly irrigated, rainfed, mainly rainfed, grazing). This was done with the aim o f capturing the geographical diversity and heterogeneity in rural Afghanistan. Using the estimated land area for each agro-ecological zone per district, a sample o f communities was drawn to represent the relative land size o f eachzone within the district. Unfortunately, inthe absence o f a census that lists all existing communities and population within a given district or geographical zone (Le., agro-ecological zone) implies that even inthe event that the populationo f each zone can reasonably be estimated, the ability to construct ex-post sampling weights to adjust for the selection o f communities is not possible. Inaddition, selecting communities strictly on the rule above was not always possible (for a variety o f reasons including security and accessibility), rendering community selection into a non-random set o fcommunities. During the second phase (household selection), all households within a community were a-priori classified in one o f three wealth groups using a self-selection perception subjective methodology developed by the WFP. Typically, all households would be present in the shura meeting and as such a ``listing'' o f all households within each wealth group could be formed. Then, households would then randomly be selected from each wealth group to be interviewed in more detail (household questionnaire). Since the NRVA collected information on the community and wealth-group sub-populations, weights to correct for the probability o f selection for a given household in the community is possible. This eliminates any potential biases from the oversampling o f poor households using the wealth group approach discussed above. As such, all NRVA sample estimates calculated can ''SeealsoHttp://www.afghanistangov.org/nrva 98The WFP, FAO, UNICEF, Dfid, AREU andthe World Bank. 79 (and should) be correctedto adjust for the intra-community selection probabilities for each household. Nonetheless, the inability to correct for the first stage in the absence of population figures and a rigorous sample frame implies that results from the NRVA are not necessarily representative o f all rural Afghanistan and any interpretations beyond the sample estimates should be treated with caution. Still, thls is the first reliable household data in Afghanistan that includes information on a number o f socioeconomic issues inrural Afghanistan from all the provinces. In addition, out of the more than 300 districts, data were collected from all but 16 districts. As such, despite its limitations, the information from the NRVA2003 is bothunique andtimely. The NRVA survey collected information from various levels of analysis: districts, communities, wealth groups and households. The data collection was implementedbetween July and September 2003 during a year that markedthe end o f a multi-year drought. The NRVA initially aimed at collecting information from about 12,000 households. Security and constraints in completing some o f the modules related to constructing the consumption poverty indicator resulted ina decrease inthe final number o f households surveyed and analyzed. The final household data have information from 11,227 households inrural Afghanistan, and capture information from 1800rural communities, consisting o fmore than 170,000 households and more than 1,280,000 individuals (about 10 percent o f the estimated rural population). 80 Appendix 11:Measuring Poverty Conceptual Issues - Assessing poverty relies on some measure of welfare. Since well-being cannot be measured directly, consumption expenditures are typically used as an indirect monetary measure o f welfare. Ideally, measuring consumption poverty is done by collecting household level information on food -bothpurchased and non-purchased -andnon-foodexpenditures suchasclothingandhousing(seeBoxAII.1).Thisallowsanestimateoftwomeasures of consumption poverty: (i) extreme poverty; and (ii) poverty. Extreme poverty refers to those households general whose annual total consumption per capita (both food and non-food) is below the extreme poverty line, defined as the annual cost o f purchasing a typical (to the country) bundle o f food items that meet a minimumlevel o f caloric requirements (e.g., the cost of 2100 calories per person per day). Similarly, general poverty refers to those households whose annual total consumption per capita (both food and non-food) is below the general poverty line, defined as the cost o f purchasing a typicalbundleo f both food and non-food i t e m . m.1: aswewelfare . 1 -, I) k? - Y 1 I , .I J 2 1 Consumption expenditure is used instead o f income for several reasons First, consumption is considered a better indicator of standards o f living since it fluctuates less than income during a month or year. When incomes change (e.g., in different seasons), individuals tend to use their savings (in cash and kind) to smooth consumption throughout the year. Second, consumption data tend to be more accurate than information on individuals' incomes. Intemational experience has shown that respondents tend to provide more accurate information on consumption than income. The latter is often underestimated or difficult to measure due to informal or in-kind income (or income from illicit activities such as poppy income). Finally, using consumption as a measure o f welfare has the advantage that poverty lines can be derived from the same data andnot from other information sources. Consumption also has several advantages over other welfare measures, such as indicators o f basic needs (such as access to water, electricity, and schooling; malnutrition) or caloric intake (food gap). While consumption is an objective measure o f welfare, indicators o f basic needs are based on various subjective definitions, including the level at which such needs would be "satisfied" and the respective weights assigned to their components. Moreover, indicators o fbasic needs are not responsive to short-term changes, since they mainly reflect public investments. As such, they are less useful for monitoring changes in economic conditions. While the NRVA 2003 collected the data necessaryto construct a measure o f annual food consumption, no information on non-food expenditures was collected inthe household module. As such, the poverty indicators and poverty lines that can be constructed using the NRVA 2003 household data are limited to measuring annual food consumption and constructing a poverty line analogous to the extreme poverty line discussed above, thus somewhat limiting the analyses that can be done. For example, analyses on distributional issues based purely on food consumption offers few insights about the overall distribution o f both food and non-food expenditures. Still, one o f the strengths o f the methodology described inthis section to construct poverty estimates with the NRVA is that the poverty estimates derived can in fact be interpreted as general poverty rates as defined above. In addition, much analysis can focus on the general patterns o f socio-economic issues using the monetary measure o f food consumptionexpenditures, which initselfis highly correlated with poverty. 81 Box AU.2: Calculating aGeneral Poverty Rate and Future Poverty Estimates Comparisons One concern with using food consumption alone to measure welfare is the issue o f comparing the NRVA 2003 poverty rates with hture poverty rates derived from more comprehensive surveys. This is crucial as monitoring poverty over time is an important element for understanding progress and welfare changes. While calculating the number o f poor people that consume less in food expenditures than the cost o f 2100 calories per day is possible using the existing data, the ability to compare this to a poverty rate constructed infuture surveys is notpossible. Assuming that better data will be collected on both food and non-food expenditures in the future, the comparison o f changes inpoverty usingthe poverty indicator from the NRVA 2003 is not possible due to the non-randomness o f the sample (at the community level). That is, since the sample is not representative of rural Afghanistan and since the ability to correct for this is limited, there is no way to compare this "baseline" poverty rate with a future one derivedfrom a sample that does represent rural Afghanistan. Nonetheless, the NRVA poverty rate can be interpretedas a "general poverty rate" as defined above. This is based on two assumptions: (i) share o f food to total expenditures decreases with wealth (Engel's law); the and (ii)defining the general poverty line as the annual cost o f 2100calories (using the same assumptions as inthe NRVA2003), plusthe cost for the typical non-food bundle amongthose households near the extreme (cost of food-only) poverty line. Using these assumptions, it can be theoretically shown that the poverty headcount rates calculated using the NRVA 2003 will be identical to those that could have been calculated ifdataontotalconsumption(bothfoodandnon-foodexpenditures) wasavailable andusingageneral poverty line as defined above. As such, the povertyrate calculated usingthe NRVA 2003 is interpreted as general rural poverty rate. Still, even if hture data do collect information on both food and non-food consumption expenditures, the poverty estimate cannot be used to monitor poverty changes through time due to the non-randomness o f the sample. At the best, they canbeusedqualitatively. See Lanjouw and Lanjouw (2001) for a complete exposition. MeasuringPoverty using the NRVA 2003 Measuring poverty requires the following steps: (i) calculate the annual value o f food consumption per capita; (ii)define a food poverty line; and (iii)calculate the number o f households and individuals whose consumption is below the poverty line. These steps are discussed below. Annual value of food consumed per capita The NRVA collected information on the quantities of food consumed in the household during the last 7 days." This is usedto obtain an annual value of quantity consumed for each food item(inkilos). Since the annual food consumption estimates above are at the household level, a final step is required to calculate population (individual) based consumption measure. As such, a share o f the total food consumption must be allocated to each household member. Per capita consumption is used inthis case, that is, the total value o f food consumption o f the householddividedby the number o f household members. While there are several other ways o f allocating household consumption to the different members, taking into account different requirements, economies o f scale, and the presence of public services inthe household, per capita consumption was used due to its transparency and since no credible information on intra-household or age-specific food allocation and requirements for Afghanistan was available. 99This includesbothpurchasedandnon-purchasedfood items. 82 The cost of livingis not uniformthroughout the country. As such, the value o f food consumption can be adjusted to account for regional variation in prices. To that extent, price indices were constructed for each province using the information collected on prices at the district level using the following approach: (i) using the food consumption information from the household module, the "rural average consumption in lulos" is calculated for each food item; (ii) average prices for each food item are then calculated for each province using the district- level prices reported inthe district survey. With these prices, the purchase cost o fthe "rural average consumption in kilos" in each province can then be calculated; and (iii) the cost o f the "national average consumption in finally, kilos" is normalized using Kabul as the base province. This produces the food price geographical index, which can be usedto deflate the annual food consumptionthus accounting for regional price variation. Constructingthe rural poverty threshold The rural poverty threshold represents the annual cost o f the minimum daily caloric requirement recommended for rural Afghanistan (see Box AIL3 for international poverty lines). Inthe absence o f information on caloric requirements for Afghanistan, it was decided that the rural poverty line should measure the cost o f consuming 2100 calories per day to be consistent with a number of studies that have used the same threshold for the last decade or so (WFP, FAO). The rural poverty threshold was calculated as follows: (i) the ranking based on total annual food using consumption per capita above, households at the 2"d and 3`d quintiles o f food expenditures were selected; (ii) on the basis o f the food consumption pattems o f these households (20" -60" percentiles), caloric shares supplied by each food item in total food consumption was calculated;'00 (iii) the cost o f the food basket required for satisfying the minimum caloric requirement of 2100 calories per day was then calculated; and (iv) finally, an adjustment to account for "wasted food" or "calorie availability" was applied for each food itemseparately."' As such, the cost o f the food basket required for reaching 2,100 calories per day was adjusted accordingly. This i s the cost o f the minimumcalorierequirements, inother words, the value ofthe ruralpovertythreshold. ForAfghanistan in2003, the poverty threshold was calculated as Af. 4621 per capitdyear, which corresponds to $0.30/capita/day (also Table AII.1). Calculatingthe number of individualsbelow the rural poverty threshold Finally, based on the above methodology, individuals were ranked from the lowest to the highest level o f annual per-capita food consumption. An individual is classified as poor ifthe per capita food consumption is less than the ruralpoverty threshold. loo Using food caloric composition data from FA0 and USDA. lo'Also see Subramanian and Deaton (1996) 83 BoxAII.3: International versusnationalbasedpoverty lines National poverty lines can be set using local social and economic conditions as well as local perceptions o f poverty. As a result, the standard o f livingbelow which a personis deemedto be living inpoverty tends to be differentfor each country. The most common approach to setting national poverty lines is the "cost o f basic needs" approach, which uses nutritional or caloric requirements. This approach first identifies a minimally acceptable diet based on local knowledge o f the foods consumed. Inlow-income countries, this minimumdiet is usually dominatedby cheap sources of calories, with little other nutritionalcontent. Then, the cost o f obtaining this minimum diet at current market prices is calculated. T h i s cost defines the "extreme" poverty line. Adding the cost for non-food necessities to this extreme poverty line results to the estimated cost o f obtaining a minimally adequatediet plus other necessities, the national poverty line. International comparisons o f poverty entail both conceptual and practical problems. Different countries have different definitions o f poverty, and consistent comparisons between countries can be difficult. Local poverty lines tend to have higher purchasing power in rich countries, where more generous standards are used than in poor countries. Poverty measures based on an international poverty line attempt to resolve these comparability issues. The commonly used $1 a day standard, measured usually in 1985 international prices and adjusted to local currency using purchasing power parities (PPP) is a typical poverty measure that is used to compare poverty across countries. PPP exchange rates are used because they take into account the local prices o f goods and services not traded internationally. Still, PPP rates were designed not for making international poverty comparisons but for comparing aggregates from national accounts. As a result, there i s no certainty that an international poverty line measures the same degree o f need or deprivation across countries. Nonetheless, inthe case o fNRVA 2003, since data were only collected inrural Afghanistan and only food consumption expenditures exist (as opposed to both food and non-food expenditures), the estimate o f poverty rates using international poverty lines would require additional assumptions on the distribution o f food and non-food consumption (since international comparisons are based on national, full consumption expenditures that include both food and non-food items). As credible information to calculate these is not available, for the purposes o f this study, international comparisons cannot credibly be made and should be avoided. Sensitivity Analysis Inorder to assessthe robustness and sensitivity ofthe povertythreshold calculatedabove, povertyrates were recalculated for different values o f the poverty line so as to evaluate how sensitive the results are to the specific choice o fpoverty line (Table AII.2). For example, a 5 percent increase o f the value o f the poverty threshold raises the poverty line fiom Af. 4621 (the baseline) to Af. 4852. This is also the equivalent o f changing the caloric requirement for the typical rural Afghani from 2100 calories to 2205 per day. Based on the new poverty threshold, the new poverty rate would be 51.4% (compared with the baseline 48.1%). This represents a 3.3 percentage point difference with the baseline poverty rate. Inaddition, the impliedelasticity o f poverty line changes to the poverty rate is 1.37. The last column in Table AIL2 lists all implied elasticities for different levels o f poverty line (or caloric requirement), increases or decreases, and suggests that there are no extreme changes inpoverty rates due to changes inpoverty lines. 84 Table AII.l: Calculation of Food consumption Patterns for Rural Poverty Line ITEM KcallKg Afghanisper Annual Daily consumption inKcal Kcalcontribution to Annual comUInpti0 Per 2,100/dayJperson cost of n 2,1 OOJday Kg KgJ'yeadH Household Person % H ~ Almonds 2470 118 48 1.3 9 1.2 0.05 1 16 Animal fat 9020 95 11 0.6 15 2.0 0.08 2 6 Apple 490 25 50 15.4 21 2.8 0.10 2 40 Apricot 470 18 39 6.0 8 1.0 0.04 1 12 Apricots 2960 44 15 0.6 5 0.6 0.02 1 3 Barley 3270 8 2 9.8 88 11.7 0.45 9 8 Beans 3500 27 8 50.3 483 64.3 2.44 51 144 Beef 1240 99 80 23.0 78 10.4 0.40 8 243 Broccoli 200 7 37 2.9 2 0.2 0.01 0 2 Butter 6930 121 18 1.7 32 4.3 0.16 3 22 Cabbage 160 9 57 1.5 1 0.1 0.00 0 1 Carrots 370 7 18 4.0 4 0.5 0.02 0 3 Cauliflower 150 10 65 4.2 2 0.2 0.01 0 4 Cheese 3100 83 27 2.5 21 2.8 0.11 2 22 Chick peas 3570 26 7 15.5 152 20.3 0.77 16 44 Chicken 1270 92 73 12.1 42 5.6 0.21 4 119 Coriander 190 11 59 1.3 1 0.1 0.00 0 2 Cucumber 170 5 28 15.2 7 0.9 0.04 1 8 Dogh 383 12 31 178.0 187 24.9 0.95 20 222 Driedmeat 6295 10 2 1.7 29 3.9 0.15 3 2 DriedTomato 2590 51 20 3.8 27 3.6 0.14 3 20 Driedvegetable 2386 55 23 4.9 32 4.3 0.16 3 Egg 1420 68 48 12.0 47 6.2 0.24 5 87 Eggplant 330 10 30 65.9 60 7.9 0.30 6 68 FishCarp 460 90 196 0.9 1 0.1 0.01 0 8 Ghee 8730 115 13 4.2 100 13.3 0.51 11 51 Goat 1570 141 90 2.7 12 1.6 0.06 1 40 Grapes 670 22 3 39.4 72 9.6 0.37 8 91 Honey 3120 49 48 0.9 7 1.0 0.04 1 14 Hot pepper 290 25 85 11.5 9 1.2 0.05 1 30 Krut(Yogurt) 4842 62 13 11.2 149 19.9 0.75 16 75 Leek 440 8 18 5.9 7 1.0 0.04 1 5 Lentil 3540 8 5 14.8 143 19.1 0.72 15 28 Liver -beef, 1440 02 71 1.o 4 0.5 0.02 0 11 goat Maize 3630 7 2 55.3 550 73.3 2.78 58 0 Melon 270 7 27 127.1 94 12.5 0.48 10 97 Cantaloupe Milk 855 16 19 70.7 166 22.1 0.84 18 120 Mint 240 11 46 3.6 2 0.3 0.01 0 4 Mulberry 3330 30 9 2.9 27 3.5 0.13 3 9 Mulberry 820 19 23 7.3 16 2.2 0.08 2 14 Mung 3400 19 6 19.3 180 24.0 0.91 9 40 Muttonllamb 2355 147 63 3.6 24 3.1 0.12 2 57 Okra h i t 390 13 34 65.7 70 9.4 0.36 7 92 Onion 420 8 19 92.6 107 14.2 0.54 1 80 Orange/citrus 330 20 60 0.4 0 0.0 0.00 0 1 Peach 460 21 46 2.2 3 0.4 0.01 0 5 Pistachio 3330 196 59 0.1 1 0.1 0.00 0 1 Plum 460 16 35 2.0 3 0.3 0.01 0 3 85 Pomegranate 430 25 57 1.1 1 0.2 0.01 0 3 Potato 750 9 12 119.5 246 32.8 1.24 26 110 Radish 280 7 27 2.5 2 0.3 0.01 0 2 Raisins 2930 44 15 1.4 11 1.5 0.06 1 7 Rice 3630 23 6 147.1 1463 195.0 7.40 155 363 Spinach 250 7 29 3.9 3 0.4 0.01 0 3 Sugar - brown 3860 24 6 13.0 138 18.4 0.7C 15 33 Sugar - refined 3860 22 6 63.0 666 88.8 3.3; 71 148 Sweet potato 730 8 11 4.1 8 1.1 0.04 1 4 Tomato 180 10 57 92.3 45 6.1 0.22 5 101 Turnip 230 5- 20 _. 3.8 2 0.3 0.01 0 2 Vegetable oil 8840 45 5 84.2 2040 272.0 10.32 217 404 walnuts 2770 54 20 1.3 10 1.3 0.05 1 7 Wheat 3570 10 3 1202.1 11757 1567.7 59.49 1249 1255 Wildleaves 190 6 29 27.6 14 1.9 0.07 2 16 Yogurt-drained 1530 18 12 62.0 260 34.6 1.31 28 118 TOTAL 2635 100 2100 4621 Source: NRVA 2003. Calorie sources: FA0 and USDA. Table AII.2: Rural poverty line sensitivity Poverty linechange Poverty line Equivalent catoricNew Poverty Poverty change Ifmplied as a YOoforiginal inAfghanis poverty line Yo ' , Elasticity * % 1 I > -10% 4201 1187 40.4 7.7 16.0% 1.60 -9% 4239 1319 41.1 7.0 14.6% 1.62 -8% 4278 1450 41.9 6.2 12.9% 1.61 -7% 4318 1576 42.8 5.3 11.0% .57 -6% 4359 1694 43.6 4.5 9.4% .56 -5% 4401 1802 44.3 3.8 7.9% .58 -4% 4443 1897 45.1 3.0 6.2% .56 -3% 4486 1976 45.8 2.3 4.8% .59 -2% 4530 2037 46.6 1.5 3.1% .56 -1% V 4575 2079 47.4 0.7 1.5% .46 0% - Original 4621 2100 48.1 0.0 0.0% 1% 4667 2121 48.8 -0.7 -1.5% -1.46 2% 4713 2142 49.6 -1.5 -3.1% -1.56 3yo 4759 2163 50.2 -2.1 -4.4% -1.46 4% 4805 2184 50.9 -2.8 -5.8% -1.46 5% 4852 2205 51.4 -3.3 -6.9% -1.37 6% 4898 2226 52.0 -3.9 -8.1% -1.35 7% 4944 2247 52.8 -4.7 -9.8% -1.40 8Yo 4990 2268 53.4 -5.3 -11.0% -1.38 9% 5037 2289 54.1 -6.0 -12.5% -1.39 10% 5083 2310 55.1 -7.0 -14.6% -1.46 C rce: NRVA 2003. 86 PovertyMeasures The poverty indices used inthis study are three special cases o f additively separable measures developed byFoster, Greer andThorbecke (FGT)."' The general poverty measure is: [Eq. 11 where: c i = estimated consumption of the ithperson ina populationof size n Z =the poverty lineq =number o fpersons whose consumption ci isbelow poverty line Z and; a =is a non-negative parameterthat reflects the measure's aversionto poverty Depending on the value o f a, a number o f possible poverty indicators can be constructed (all o f the indices below are expressed as a percentage). Headcount index - PI, (a = 0). This is the Headcount ratio and, as can be seen fIom Equation 1, it is simply qln or the proportion o f the population below the poverty line. As such, the headcount index provides information on the incidence o f poverty. It provides little information about the distribution o f people below the poverty line. Poverty Gap PI (a = 1). T o determine the depth o f poverty, a second version o f the FGT poverty - measure, called the Poverty Gap index i s used. The index is the aggregate poverty deficit o f the poor relative to the poverty line. It is the ratio between the amount o f money required to bringthe level o f consumption o f the poor at the poverty line (i.e., perfect targeting) and the cost o f the poverty line given to everyone irrespective o f their poverty status (imperfect targeting). Poverty Severity -P2 (a = 2). This measure captures the severity o f poverty and demonstrates how poverty is distributed among the poor. It gives more weight to poorer households or individuals (e.g., with higher levels o f poverty) so that it takes into account the distribution o f consumption (or other welfare indicator) among the poor. As such, a higher P2 suggests that the poor are more heterogeneous as a group, with larger differences between the extremely poor to those with lower levels ofp ~ v e r t y . " ~ lo2 Foster, Greer and Thorbecke (1984). Also see Ravallion (1992) 87 Appendix 111:Calorie Gap Analysis The NRVA collected information on the quantities o f food consumed inthe household during a seven day period preceding the interview. This information was collected on a "self-reporting" basis, a methodology far less accurate than that often used in nutrition surveys where the quantity o f food is actually weighed prior to consumption. Once aggregated to calculate annual food (calorie) consumption, and then divided by the number o f household members eating the food, an estimated per capita calorie consumption number can be arrived at. Using the threshold o f 2100 calorieslday, the percentage of the sample population consuming less than this threshold can then be estimated (Table A3.1 below). In the overall sample, 21.5% of the people were consuming less than 2100 calorieslday. There is, however, considerable variation across the country, with people inthe west, north, south and central areas all having higher calorie gaps. The north-east by comparisonhas a lower calorie gap figure. Table AIII.l: Calorie Gap Estimates % of sample Calorie Gap population Rate* All sample 100 21.5 ByRegion North-east 11 13 South-east 12 15 East 14 16 Central 15 24 South 19 24 West 10 25 North 16 26 West-central 3 37 *Source: NRVApeople consumingless than 2100 calories/day. 2003 Percentageof Figure AIII.l: Food ConsumptionDistribution (lorenz curves), B y Region Source:NRVA 2003. Food consumption in rural Afghanistan is relatively unequally distributed when viewed across welfare quintiles. Around one third o f the total food (calorie) consumption is consumed by the highest quintile o f the rural 88 pop~lation.`~~ compares with 12 percent o f total food consumption consumed by the poorest quintile (Table This A3.2). This implies that the top quintile consumes almost two and a half times the food consumed by the lowest quintile. In terms o f food expenditure, the top quintile spends more than four times on food what is spent by the lowest income quintile, reflecting greater dietary diversity and consumption o f more expensive food items (such as meat, dairy and oil). The overall gini coefficient - a measure o f inequality - for food expenditures distribution in rural Afghanistan is 29 percent, and as highas 32 percent inthree regions (north, north-east and west-central)(figure A3. 1).'05 Table AIII.2: Distributionof CaloriesacrossWelfare Groups Caloriesper cost f 0 Distribution of Personper Distribution of purchasing food day Calories(%) caloriesAfdp.a. expenditure(%) Quintiles 41-Poorest 20% 1708 12 2292 9 0 2 2270 16 3580 13 4 3 2650 19 4762 18 44 3106 22 6237 23 45 - TOP20% 4263 31 10151 38 Overall 2800 100 5403* 100 Source: NRVA 2003. * Thisis equivalentto $1 10 at the exchangerate of US$I = 49 Afghanis. Understanding whether the observed distribution o f calories and the associated food gaps are due to supply or demand problem is o f particular interest. That is to say, are people going hungry because food is simply not available (i.e., a supply problem) or because they have neither the means to grow their own food andor purchase it inthe market. First, there are a large number ofpeople who are landless - about a quarter of the households inthe NRVA sample. There is a higher incidence o f (food consumption) poverty among landless households than those having some land, and average calories intakes for this category o f households is among the lowest. These households also have little livestock, another source o f food. Although some may sharecrop, the core livelihood is agriculture and non-agriculture labor. These households are, in all likelihood, net purchasers o f food. Other net purchasers o f food would be households whose land holding is inadequate (in terms o f quantity or quality) to produce sufficient food to satisfy household subsistence needs, and households who have a diversified income portfolio and prefer to acquire food inthe market. A number of observations seem to suggest that the supply o f food in rural Afghanistan, and its local availability is not the major issue (though the winter season does present logistical problems and location specific shocks, such as disease, cannot be ruled out). Inaddition to growing domestic production, data suggest that private imports o f wheat (flour) into Afghanistan have beensubstantial inrecent year, with large numbers o ftraders who are highly responsive to market forces bringing grain from Pakistan, Kazakhstanand other countries inthe region. Food aid has also entered Afghanistan in large quantities, though evidence suggests that this has not increased net food availability, rather substituted for private trade.lo6 The overall availability o f food is borne out by the NRVA data which show that, based on the caloric consumption estimates, the average availability is around 2800 calorieslday (Table A3.2). To the extent that 2100 calories per person per day i s a reasonable caloric threshold for rural Afghanistan, interms o f satisfying the minimumdaily requirement o f a person, this supports the idea that there are enough calories inrural Afghanistan to supply the required caloric intake, ifonly they were more evenly distributed. Low supply o f food in a particular area (village, district) would suggest that a significantly high number o f households would exhibit food intake gaps. Infact, the opposite appears true: food intakes gaps are generally low in any one area, suggesting a dispersion o f food insecure households throughout the sampled areas. While these NRVA based-results are not representative for the whole country and thus cannot be taken as conclusive, the body o f IO4Theremaybe aproblemwith the calorie consumptionnumbersfor the top andbottomquintiles, where mergedhouseholds might be sharingfood. Thus the "tails" ofthe distribution should betreated with great caution. lo'The gini coefficient is a measure of inequalityor distributional differencewithin aparticularpopulation. It varies between 0 and 1, with higher inequalitybeinggini values closer to 1. IO6See "The impact of food aid on wheat markets in Afghanistan, 2000-03". World Bank mimeo (2004) 89 evidence weighs in favor o f demand side (insufficient purchasing power) rather than food supply issues as the root cause o f household food insecurity. Figure AIII.2: Food Supply -Distribution of District-level Caloric Intake Gaps Districtswith low food gaps Districtswith large food gaps Source:NRVA 2003. 90 REFERENCES ACF, (2004). Kabul Vulnerability Mapping. Actionaid, (2003). What Poor Afghans Think. ADB, (2003). Proposed National InstitutionalArrangements for Emergency and Disaster Management inAfghanistan. A F S U N A M UnitWFP, (2002). Pastoralist Vulnerability Study. AREU, (2003). Three Villages inAlingar, Laghman: A Case Study ofRuralLivelihoods. AREU, (2003). UrbanWinter Vulnerability: The KabulExperience. AREU, (2004). Some Notesonthe Livelihoods o fthe UrbanPoor inKabul, Afghanistan. AREU, (2004). UrbanLandManagement inAfghanistan: The Case ofKabul. AREU, (2004). GenderRoles inAgriculture. Asian Development Bank, (2002). Afghanistan Recovery and Reconstruction: Preliminary Needs Assessment. Agriculture and Water Sub-Sectors. Barfield, N.(2004). Nomadic Pastoralists inAfghanistan: reconstruction o f the Pastoral Economy. CARITAS Germany, (2003). Women inKabul: A Needs Assessment, Chaudhuri, S. (2000). EmpiricalMethods for Assessing Household Vulnerability to Poverty, Columbia University, Draft. Coke, A. (2004). Wheat SeedandAgriculture Programming: Its Potential to Impact on Livelihoods. Deininger, K. and L. Squire (1998). New Ways o f Looking at Old Issues, Journal o f Development Economics, 57: 259-87. Environment Advisory Group for Afghanistan, (2003). Strategy Paper for the Integration o f Environmental Management and Recovery inthe NationalDevelopment Planfor Afghanistan. European Commission, (2002). Report o f the EC Rapid Reaction Mechanism Assessment Mission: Afghanistan Urban Recovery. FAO, (2003). Contribution to FoodAid Policy Development for Afghanistan. FAO, (2003). National Livestock Census, Interim Report. Favre, R. (2003). Grazing LandEncroachment. Foster, J., J Greer and E. Thorbecke (1984). A class of decomposable poverty Measures, Econometrica, 52: 761- 765. Ghobadi, N., Koettl, J., and Vakis, R., (2005) "Moving out o f Poverty: Migration Insights from Rural Afghanistan". Draft paper, World Bank. Government of Afghanistan, (2004). Securing Afghanistan's Future: Accomplishments and the Strategic Path Forward. Report prepared for the International Conference March 2004. 91 Grace, J. (2004). Gender Roles inAgriculture: Case Studies o f Five Villages in Northern Afghanistan. Groupe URD, (2002). The Global Study onparticipation o fAffected Populations in Humanitarian Action: Afghanistan Monograph. GTZ, (2004). Case Study Afghanistan. HealthNet International, (2003). Mental HealthinBasic HealthCare: Experiences and Results of a programinNangarhar Province. Heitzmann, K., Canagarajah, R., and Siegel, P (2002) The Sourceso f Vulnerability: A Rationale and Guideline for the Assessment o f Risks and Risk Responses. Social Protection Discussion Paper No. 0218, Washington D.C, The World Bank. Holzmann, R. and S. Jsrgensen. 2000. Social RiskManagement: A New Conceptual Framework for Social Protection and Beyond, Social Protection Discussion PaperNo. 0006, Washington D.C.: The World Bank. Hunte,(2004). SomeNotes onthe Livelihoods ofthe UrbanPoor inKabul, Afghanistan. AREU Case Study Series. Lanjouw, P. and Lanjouw, J. (2001). How to compare apples and oranges: poverty measurement based on different definitions o f consumption. Review o fIncome and Wealth, (47) vl. Lautze, S. et a1(2002). Qaht-e-Pool ?A Cash Famine? FoodInsecurity inAfghanistan 1999-2002. Maletta, H.(2003), Winters o fDiscontent: SeasonalChanges inwages and FoodPrices inAfghanistan. Maletta, H. (2003). Women at Work gender, wealth, wages and employment inrural Afghanistan. Mansfield, D. (2004). Coping Strategies, Accumulated Wealth and Shfting Markets: The Story o f Opium Poppy CulviationinBadakshan 2000-2003. Mansfield, D. (2004). Diversity and Dilemma: Understanding Rural Livelihoods and Addressing the Causes o f OpiumPoppy Cultivation inNagarhar and Laghman, Eastem Afghanistan. A Report for the Project on Alternative Livelihoods. Molla, D. (2003). FoodAid, Wheat Prices and Poppy CultivationinAfghanistan Moser, C. (1998). The Asset Vulnerability Framework; ReassessingUrbanPoverty Reduction Strategies. MRRDNVFP, (2003). FoodAid, Wheat PricesandPoppyCultivationinAfghanistan. National Surveillance System, (2003). Summary Findings for Southern Kohistan District, Fall 2003. Pain, A. (2004). Understanding Village Institutions: case Studies on Water Management from Faryab and Saripul. Pain, A and Lautze, S. (2002). Addressing Livelihoods inAfghanistan AREU Issues Paper Series. Pinney, A. National Riskand Vulnerability Assessment 2003: A Stakeholder-Generated Methodology. Afghanistan Research and EvaluationUnitWorking Paper Series. January 2004. Ravallion, M. (1992). Poverty Comparisons: A Guide to Concepts and Methods. World Bank LSMS Working Paper No. 88. 92 Schutte, P. (2004). UrbanVulnerability inAfghanistan: Case Studies from Three Cities. Sobrado, C. (2001). Measuring poverty using the ENCOVI 2000, inGuatemala Poverty Assessment. Report No. 24221-GU Stigter, E. (2005). TransnationalNetworks and Migration from Herat to Iran. AREUCase Study Subbarao, K. (2001). Systemic Shocks and Social Protection: Role and Effectiveness o f Public Works Programs. World Bank Africa RegionHumanDevelopment Working Paper Series. Subbarao, K.(2004). Targeting Programs: Concepts and Evidence (mimeo). Subramanian, S. and Deaton, A. (1996). The Demandfor Food and Calories. The Joumal o f Political Economy, v.104No. 1. Strategic Monitoring Unit (SMU), (2001). Badakhshan. Teas, M., (2002). Building an Approach to UrbanVulnerability Analysis inAfghanistan. Tesliuc, E. and K.Lindert (2004). Risk and Vulnerability inGuatemala: A Quantitative and Qualitative Assessment. Social ProtectionDiscussion Paper No. 404, Washington D.C., The World Bank. TISA, (2004). Securing Afghanistan's Future. Turton, D.et al, (2002). Taking Refugees for a Ride? The Politics o f Refugee Returnto Afghanistan. UNDP (2004): Afghanistan: National HumanDevelopment Report "Security with a Human Face: Challenges and Responsibilities". UNEP/ATA, (2003). Post Conflict EnvironmentalAssessment inAfghanistan. UNICEF/SC-US, (2003). The Children o fKabul: Discussions with Afghan Families. UNICEFLJS (2002), Centers for Disease Control and Prevention (CDC), (2002). Nutrition and HealthSurvey, Badghis Province, Afghanistan. UNICEF/CSO (2004). Moving Beyond 2 Decades o f War: Progress o f Provinces", MICS 2003 - a survey undertaken by UNICEF inconjunction with the CSO. UNODC, (1999). Access to Labour: The Role o f Opium in the Livelihood Strategies o f Itinerant Harvesters Working inHelmand Province, Afghanistan. UNODC, (2000). The Role o fWomen inOpiumPoppy Cultivation inAfghanistan. UNODCICND, (2003). Afghanistan Farmers intentions' Survey 2003/4. Vakis, R (2003) The Afghanistan National Riskand Vulnerability Assessment 2003: A short commentary, The World Bank. WCLRF, (2003). BAD: A Painful Sedative. Wily, L.(2003). LandRights inCrisis: RestoringTenure Security inAfghanistan. Wily, L.(2004). Landrelations inBamyan Province. 93 Wiley, L.(2004). UrbanLandManagement inAfghanistan: The Case of Kabul. World Bank (2002). Emerging from Ethnic Conflict: Challenges for Social Protection Design in Transition Countries". World Bank Social Protection DiscussionPaper Series, December 2002. World Bank (2002). FoodBased Safety Nets andRelatedPrograms. World Bank Social Protection Discussion Paper Series, September 2002. World Bank, (2003). Emergency Urban reconstruction Project: Project Information Document (PID). World Bank (2004). Afghanistan - State Building, Sustaining Growth andReducing Poverty. World Bank (2005). Towards a National FoodPolicy Framework for Afghanistan. Draft report.