Report No. 23390-KOS Kosovo Poverty Assessment (In Two Volumes) Volume 1: Main Report December 20, 2001 Poverty Reduction and Economic Management Unit Europe and Central Asia Region Document of the World Bank CURRENCY EQUIVALENTS (Exchange Rate Effective December 20, 2001) Currency Unit = DEM DEM I = US$0.4494 US$1 = 2.2254 FISCAL YEAR January 1 - -- December 31 ABBREVIATIONS AND ACRONYMS ADL : Activities of Daily Living AOR : Area of Responsibility CSW : Center for Social Welfare DES : Department of Education and Science DLA : Department of Local Administration DM : German Marks FPL : Food Poverty Line FRY : Federal Republic of Yugoslavia GDP : Gross Domestic Product HIF : Health Insurance Fund IDP Internally Displaced Population IMF : International Monetary Fund IOM : International Organization for Migration JAMA : Joumal of American Medical Association JIAS : Joint Interim Administrative Structure LSMS : Living Standard Measurement Survey MOH : Ministry of Health MTS : Mother Theresa Society NGOs : Non-Governmental Organizations OECD : Organization for Economic Cooperation and Development OSCE : Organization for Security and Cooperation in Europe OZ : Elementary Units PEC : Per-Adult Equivalent Consumption PL : Poverty Line SFRY : Socialist Federal Republic of Yugoslavia SIZ : Self-managing Community of Interest UNFPA United Nations Population Fund UNMIK : United Nations Interim Mission in Kosovo WFP : World Food Program WHO : World Health Organization Vice President: Johannes F. Linn, ECAVP Country Director: Christiaan J. Poortman, ECCO4 Sector Managers: Maureen Lewis (ECSHD); Kyle Peters, Daniela Gressani (ECSPE) Task Team Leader: Pierella Paci, ECSHD KOSOVO Poverty Assessment Table of Contents Acknowledgements . . ............................... vi Executive Summary .................................... vii The data and the poverty indicators used ......................................... vii The profile of income poverty in Kosovo ........................................ viii Main sources of income ................................................. ix Poverty and inequality in educational attainments ................................................x Health status and access to health care ........................................... xi Conclusions and recommendations ................................................. xii PERMBLEDHJE E EKZEKUTIVE ...................................... xiv IZVRSNI REZIME ..................................... xxii 1. Introduction ........................................................ I A. The Background ..................................................1 Poverty in Kosovo in 2000 ......... .........................................3 SSoi ..................................................3 What are the Characteristics of Poverty in Kosovo? .................................................3 B. Providing a Benchmark for Future Monitoring .............................................. . 5 C. The Data ..................................................6 D. The Purpose of the Study and the Structure of the Report .................................. 6 2. Poverty as Lack of Opportunities ...... ....... ....... 8 A. Defining Poverty ..................................................8 Choosing a Welfare Indicator ..........8.......................................8 Setting a Poverty Line and Choosing Poverty Indicators ............................................8 B. The Extent of Poverty in Post-Conflict Kosovo ..................................................9 Headcount ................................................. 10 The Role of Food Aid in Containing Poverty ................................................. 11 C. Who are the Extremely Poor ................................................. 12 What are the Characteristics of the Extremely Poor and How Do They Differ From Those of the Poor? ..................................................................................................... 14 D. Identifying the Correlates of Extreme Poverty ..................... 15 Standard Correlates of Poverty ..................... 15 Conflict-Related Correlates ..................... 18 E. Conclusions and Policy Implications ..................... 19 iii 3. Income Sources and social assistance . ............................... 23 A. How Much Income is Generated? .............................................................. 23 B. What are the Main Sources of Income . ............................................................. 23 C. Income Related Inequalities in Sources . ............................................................. 24 D. Individual/Household Income from Private Sources in the Last 12 Months .... 26 Employment and Occupation Structure .............................................................. 26 Household Employment Structure and Poverty ....................................................... 27 Why are Earnings so Low in Kosovo? ... 28 Correlates of Labor Market Outcomes ... 28 E. Inter-household Remittances and Other Transfers ... 29 F. Social Assistance ... 29 G. Assistance from International Donors ... 32 H. Household Income and Consumption Compared ... 32 I. Conclusions and Policy Implications ... 33 4. Poverty and Inequality in Education Attainments .......... 35 A. Educational Status of the Population of Kosovo ................................................ 35 B. Income Related Inequalities in Educational Outcomes ...................................... 37 C. Analyzing Expenditure ......................................................... 38 Public Expenditure on Education ......................................................... 38 Household Expenditure on Education ........................................................ . 40 D. Identifying Causal Factors . 41 E. Conclusions . 41 Ensure that Poor Families Can Afford Schooling . 42 Promote the Education of Specific Groups . 44 5. The Health Status and Access to Health Care ................. 45 A. Health Status of the Population ..................................................... 45 Correlates of Poor Health Outcome ..................................................... 47 Income Related Inequalities in Health Outcome ..................................................... 48 B. Health Care Provision and Utilization ..................................................... 49 Patterns of Health Care Utilization ............................................ ........... ... 49 Barriers to Access ... 50 C. Health Care Financing ... 52 Private Expenditure in Health ... 52 D. Policy Levers and Policy Design ... 53 Developing a Blue Print for Improving Health Outcomes ... 54 6. Towards Buildinga Strategy for Poverty Alleviation ..... 55 A. The Poor Versus the Extremely Poor ............................................................ 55 B. Improving the Odds for the Poor: Broad-based Growth Strategies and Structural Reforms ............................................................ 56 iv C. Towards a Strategy to Reach the Extreme Poor .................................................. 57 The Extremely Poor are Deprived in More Than One Dimension ............................ 58 Fiscal Sustainability ..................................................... 59 Local Priorities .................................................... 59 D. Designing Interventions to Alleviate Extreme Income Poverty .......................... 59 E. Improving Educational Outcomes .................................................... 61 Ensure that Poor Families Can Afford Schooling .................................................... 61 Promote the Education of Specific Groups .................................................... 61 F. Improving Health Status and Access to Health Care .......................................... 62 Developing a Blue Print for Improving Health Outcomes ........................................ 63 G. Towards a Comprehensive Strategy for Poverty Reduction .............................. 63 References ............ ......................... 65 v ACKNOWLEDGEMENTS This document was prepared by Pierella Paci (HDNVP), Task Leader, Edmundo Murrugarra, James Stevens, Esperanza Lasagabaster, William Tracy (ECSHD), Marina Wes (ECSPE), Taies Nezam (ECSSD), Benedicte De la Briere (formally DECRG), Hjalte Sederlof, Soumya Mitra, and Arjan Gjonca (consultants). Ian Conachy produced the manuscript. The report is based on the data from the 2000 Living Standard Measurement Survey of Kosovo. The survey was a joint venture of the United Nations Development Programme (UNDP) and the Food and Agriculture Organization (FAO) and the World Bank. Financial support was provided by the Dutch Government, the UK Department for Intemational Development (DFID) and the Intemational Organization for Migration (IOM). It was carried out for the benefit of this report by the Statistical Office of Kosovo and is publicly available at http://www.worldbank.org/Ilsms/guide/select.html. Valerie Evans, Field Manager, and Armando Levinson (consultant) were invaluable to the survey execution. The report benefited from discussions with Chrik Poortman, Country Director, intemational and local partners, as well as with participants to the preparation and dissemination workshops in both Pristina and Gracanica. It was extensively discussed with the members of the Joint Interim Administrative Structure. Detailed comments were provided by Andras Horvai, Country Officer, Giuseppe Zampaglione, the World Bank Representative in Kosovo. The peer reviewers, Jesko Hentschel (AFTMI) and Philip O'Keefe (ECSHD) and Harold Aldemnan (RDV), provided feedback and advice. Maureen Lewis, Sector Manager (ECSHD), Kyle Peters, and Daniela Gressani, Sector Managers (ECSPE) provided overall guidance to the team. vi EXECUTIVE SUMMARY 1. It is generally recognized that the events of the last decade have claimed a heavy toll on the long-term welfare of the population of Kosovo. As Kosovo goes beyond the emergency relief phase, taking into consideration the situation of those that the conflict has left behind is vital to the effectiveness of any strategy for economic development. The present study is designed to inform the current policy debate in the area of poverty alleviation and social service delivery and to feed into the strategies of the Joint Interim Administrative Structure (JIAS), the World Bank and other donors in these areas. This report identifies the characteristics of the poorest and most vulnerable groups in Kosovo, discusses key issues that affect the poor, and suggests a set of policy levers which are likely to be most effective in improving their welfare. It is particularly valuable in identifying priorities and suggesting ways in which the existing resources can be better targeted. 2. Poverty is defined in a multidimensional way that extends beyond low levels of income. Thus the report covers a wide range of issues including consumption, income, education, health and social protection. THE DATA AND THE POVERTY INDICATORS USED 3. This report predominantly is based on data from the Living Standard Measurement Survey (LSMS).' This household survey carried out between September and December 2000 - covers 2,880 households and is statistically representative of both the Albanian and Serb population, but not of the other ethnic groups. 4. A consumption based measurement of welfare is used to analyze income poverty. The focus of this volume is on absolute poverty and the extent of extreme and non-extreme poverty is estimated with reference to a food poverty line (FPL) and a Complete poverty line (PL) respectively. The definition of these two poverty thresholds and their relative values are given in Box 1. Box 1: Measures of Income Poverty 1. Poverty Lines * Food Poverty Line (FPL). cost of minimum food basket to provide 2,100 calories per adult (DM 1.8539 per adult, per day). * Complete Poverty Line (PL): allows for the households need to consume goods other than non food (DM 3.499 per adult per day). 2. Definition of Poverty * 'Extreme poor' if household consumption < FPL. * 'Poor' if household consumption < PL. 5. In the absence of indicators of learning outcomes, educational outcomes are defined in terms of illiteracy rates, educational attainment, and enrolment rates. Finally the health dimension of poverty is captured by a combination of indicators of (i) health status and (ii) access to health care. For more details on the LSMS see http://www.worldbank.org/lsms/guide/select.html vii THE PROFILE OF INCOME POVERTY IN Kosovo 6. The extent of poverty and extreme poverty. The findings of the report on income poverty are summarized in Box 2. The LSMS data suggest that in late 2000 poverty was widespread, affecting over half of the population. However poverty was relatively shallow: an average annual contribution of DM 400 per person would have been sufficient to bring a poor household out of poverty. 7. The conclusion of pervasive but not deep poverty is reinforced by the finding that a comparatively low 12 percent of the population lived in extreme poverty - i.e., had consumption levels below the FPL. The extreme poverty gap2 is also relatively small at only 2.5 percent of the food poverty line. In other words, an extremely poor individual would have needed an average of DM 144 per year to achieve the minimum food requirement. Thus, assuming perfect targeting, the cost of a social assistance program designed to eliminate extreme poverty would have been around 34 millions DM, which corresponds to around 1.2 percent of GDP according to the IMF estimate of GDP per capita . With a more realistic assumption of a 25 percent leakage, the cost of social assistance would rise to just over 45 millions DM, around 1.7 percent of GDP.4 Box 2: Incidence of Income Poverty Extreme Poverty Overall Poverty Headcount Gap Headcount Gap Total 11.9 2.5 50.3 15.7 Rural 11.6 2.7 52.0 16.1 Urban 12.5 2.4 47.5 15.1 Source: World Bank estimates from LSMS 8. Focusing on the extremely poor. These results suggest a significant clustering of households just below the PL. One implication of this conclusion is that the poor/non- poor distinction is somewhat arbitrary in the context of post-conflict Kosovo. It suggests that, given the limited budgetary resources, an effective poverty reduction strategy should concentrate on the clearly identifiable and distinct group of the extremely poor, leaving the issue of consumption and income inequalities to a separate growth strategy. 9. Identifying the characteristics of the extremely poor. Overall the 'typical' household in extreme poverty is of Albanian ethnicity, it has 7 members and a high dependency ratio, due to a large number of children and the presence of at least one elderly person. It lives in a rural area, has been displaced at least once during the conflict and is now back to its original place of residence. The household head is around 50 years of age, with a low level of education, works in agriculture but has less than one hectare of land and no machinery. 2 The extreme poverty gap measures the consumption deficit of the extremely poor compared to the poor, or the gap between actual consumption and the food poverty line. The poverty gap, on the other hand, measures the consumption deficit of the poor, i.e., the gap between actual consumption and the poverty line. Both measures are defined as percentage of the reference poverty line. 3 The IMF estimates GDP per capita to be around US$805. This study uses DM as currency unit to avoid problems of establishing the purchasing power parity (PPP) of the DM in Kosovo. 4 Simulations on proxy-means tests for targeting using data from several Latin American countries shows leakage rates ranging between 22 to 34 percent. See Grosh, M. (1994) viii 10. The prevalence of extremely poor households in rural areas reflects the fact that Kosovo is predominantly a rural society, rather than the higher poverty risk of rural households. Indeed, the incidence of extreme poverty is slightly higher in urban areas than in rural ones and extremely poor households in urban areas differ from their rural counterparts in a number of ways. Most notably, in urban areas education levels tend to be higher and the heads of extremely poor household have been out of work and have not looked for a job for more than 12 months. 11. The typical Serb household in extreme poverty shares with the Albanian the rural residence, the low level of education and the agricultural activity. However it is generally very small, headed by an elderly and with no children. 12. Identifying the Correlates of Extreme Poverty. The strongest correlates of extreme poverty include: * Dependency ratios of 0.5 or higher; * In rural areas, no or small (less than 1 hectare) land holdings; * In urban areas, labor market status - being unemployed or out of the labor force - and being disable; and - Precarious and/or damaged housing, and lack of sewage and piped water. 13. In addition to these correlates that apply to the overall population, there are other smaller groups characterized by a very high risk of extreme poverty. These include the Internally Displaced Population (IDPs) and 'other ethnic groups' - especially Romas, that often display other characteristics associated with high risks of extreme poverty.5 Finally, households with a disabled head (as evidenced by holding a disability card) also present significantly higher risks of extreme poverty especially if Serbs and/or living in urban areas. MAIN SOURCES OF INCOME 14. Despite the destruction due to the conflict in the year prior to the survey, wages and earnings from business activities remain the main source of income for the average household. Household transfers from relatives form the second most important source of income for the Albanian household. Nearly 50 percent of them receive remittances in cash or in kind from relatives abroad and to a much smaller extent inside Kosovo. Cash remittances average DM 4,983 a year with an additional DM 1,473 receive on average as in-kind transfers. By way of contrast, only 5 percent of Serb households receive transfers of this type and the average amount received is less than a third that of the Albanians. 15. Over 60 percent of the population report receiving food aid, winter fuel and clothes in the 12 months prior to the survey from a variety of donors and the percentages are higher in rural areas. For the average Albanian households the food aid is worth a total of DM 382 while the average Serb household received only DM 176, partly as a reflection of its smaller size. 16. Over 20 percent of Albanian households receive additional non-cash foreign aid in the form of building repair. The average value of this assistance is DM 5,825 per household. 5 It is however important to remember that the survey used for the analysis is not statistically representative of the 'other ethnic groups'. The findings on these groups therefore, need to be taken with caution. ix 17. Inequalities in income sources. The relative importance of different sources of income varies at different points of the income distribution. Unsurprisingly, the role of work related income is highest for the top income decile while the contribution from public transfers declines considerably as we move from the poorest to richest households. More surprising is the inverted U-shaped pattern of the share of receipts from both household transfers and international donors. These sources contribute significantly to income in the poorest two deciles. Their role then declines only to increase again for households in the top four income deciles. This implies that the absolute value of these sources increases with income. 18. Nevertheless, household transfers are the more redistributive source of income reducing income inequality - as expressed by the Gini coefficient - by 17 percent from 0.76 for eaming only - to 0.63.6 International aid has a smaller redistributive effect for both ethnic groups while public assistance predominantly affects the Serbs. This reflects the fact that compared to the Albanians, Serb households receive much less in household transfers, and international assistance is very much concentrated in the bottom income groups. For this ethnic group public transfers constitute over 50 percent of the income for households in income deciles 2-5. POVERTY AND INEQUALITY IN EDUCATIONAL ATTAINMENTS 19. In the area of educational attainment the most striking result of the report is the large inequalities existing across income, ethnic groups and gender. While net enrolment rates in primary school are on average high (97 percent), only 76 percent of the children under 14 from ethnic groups other than the Serbs and Albanians - 'other ethnic groups' - are enrolled in school. The situation is particularly worrisome in urban areas where children from these ethnic groups in the bottom consumption decile have net enrolment rates in primary education as low as 15 percent. Their low enrolment rates coupled with their concentration on the lowest consumption decile (28 percent) explains the very worrying result that as many as 20 percent of children from urban households in the lowest consumption decile do not enroll in primary school. 20. Inequality in secondary school enrolment is even more dramatic. Only just over half of the Albanian girls aged 15-18 are enrolled in school in 2000, compared to nearly three quarters of the Albanian boys. By way of contrast the average net enrolment rates for the Serbs are nearly 80 percent, with a slight gender bias in favor of girls. Net enrolment rates in secondary education for the 'other ethnic groups' drop to less than 55 percent on average and to less than 40 percent for girls. Youngsters from the poorest decile in rural areas are less than half as likely to enroll in secondary education than their counterparts in the top decile. Although less striking, the difference in urban areas is significant, with only 62 percent of the poorest children enrolling as opposed to nearly 90 percent of those in the top decile. 21. The LSMS results point to economic factors as the main correlates of non- enrolment. For the entire population, the private cost of education is identified as the main reason for non enrolment (37 percent), with an additional 6.5 percent pointing to work as 6 The Gini coefficient is the most commonly used measure of income inequality. Its value ranges from 0 - complete equality of income - to I - complete inequalitity, ie one person has all the income, all others have none. Graphically the Gini coefficient represents the area between the Lorenz curve and the diagonal, ie the line of perfect equality. x the main factor. The LSMS findings show that nearly all households who had children enrolled in the academic year prior to the survey (99 percent) made considerable out-of- pocket payments. On average 6 percent of these payments cover fees and another 12 percent cover tuition costs. Total books and equipment account for Il percent, while 'in- kind' and 'other expenses' add to over 25 percent of private expenditure. Although the amount paid differs across the consumption distribution, private expenditure for education - especially primary - are high also at the lower end. Geographical and ethnic differences are also apparent, with expenditure being higher in urban areas particularly for the Serbs. Safety is also identified as an important barrier to enrolment. Other reasons related to access to schools are significant especially in rural areas, with distance to school cited by 5.4 percent of the sampled population. HEALTH STATUS AND ACCESS TO HEALTH CARE 22. The limited available data on health status suggest that Kosovo ranks lowest in Europe on virtually every health indicator. Infant mortality, maternal health and the decline in immunization coverage are areas of particular concern. However very little is known on the causes of such poor performance and the existing data do not allow any existing inequalities in these indicators to be analyzed. The LSMS information on three indicators of adult health status shows strong ethnic and gender-related inequalities. The Serb sample reports systematically worse health indicators than the Albanians. While part of the difference reflects the different age structure of the two populations, differences persist also within each age group.7 A systematic gender bias is also noticeable. Following a pattern common to many other countries, the percentage of women reporting poor health outcomes is greater than men's with the exception being the 'disability' indicator. 23. Income related inequalities also emerge and their pattern is one common to many countries. Ill health appears to be concentrated among the worse off households and this is particularly true for the self-assessed health indicator. However the degree of inequity is surprisingly low, possibly reflecting the recent conflict in Kosovo. 24. As in the rest of the region, health care appears to be hospital- and specialist-based and utilization patterns and rates are similar across ethnic groups and gender, once differences in health status are taken into account. Excluding minor ailments for which self treatment was considered sufficient, the most common barrier to access to health care is the cost of the service. An average of 28 percent of those that reported having had a medical problem for which they did not seek treatment, did so because they could not afford health care. 25. The role of affordability as a barrier to entry comes as no surprise when one considers the major out-of-pocket payment the population of Kosovo appears to face. Over 95 percent of Albanians paid something for the services received, independently of whether these are provided in public or private facilities. The average payment is well in excess of the expected amount of the co-payment and the most expensive item of 7 To some extent the difference in health status across ethnic groups may reflect the tendency for the healthier Serbs to leave the country, leaving behind the less economically active. However data on this issue is not available. xi expenditure is pharmaceuticals followed by 'gifts'. Finally the Serbs appear to have easier access to free health care and tend to face lower costs. CONCLUSIONS AND RECOMMENDATIONS 26. This report suggests that, at the end of the year 2000, 50.5 percent of the population of Kosovo lived in poverty and over 12 percent lived in extreme poverty. It is clear that transfer programs and the provision of social services alone cannot eliminate poverty of this magnitude. Moreover, the difference between the extremely poor and poor households, as it emerges from the analysis above, suggests a two-prong strategy for poverty reduction in Kosovo today. 27. Improving the odds for the poor. The households who in 2000 were poor-but not extremely poor-have a stock of assets and a level of financial and human capital that suggests their poverty status to be responsive to labor-intensive, economy-wide growth. They are predominately self-employed fanners with some assets and comparatively larger land holdings. When living in urban areas they are relatively young, have low dependency rates and have a head with a good education level but who is currently unemployed or out-of-the-labor force. 28. A major component of the poverty reduction strategy must therefore focus on the renewal of broad-based economic growth driven by key structural reforms and institution building to support the transition to a market economy. The development of an adequate and sustainable rural strategy designed to increase productivity in agriculture and develop job opportunities in related sectors is an important element of this growth strategy. Experience elsewhere shows that a major source of new jobs would lie in the small and medium enterprise (SME) sector. The expansion of small and medium scale enterprises, in both service activities and light industry, is of crucial importance to promoting efficient labor-intensive growth. Yet for this to happen, Kosovo needs to further develop and implement rules and market mechanisms, both legal and institutional, under which the private sector can succeed. 29. Levers to alleviate extreme poverty. Even under highly optimistic assumptions, it is difficult to envisage the economic growth and development of social services in the next few years to be sufficient to improve significantly the odds for the hardcore of individuals that represent the extremely poor. For them, specific anti-poverty interventions are required to avoid the creation of an underclass marginal to the economy and to the social structure and unable to access the social services. 30. The results of this report suggest that the design of these interventions must reflect the existing large degree of overlap between different dimensions of poverty. Poor educational attainment and - to a lesser extent - poor health status are strong correlates of extreme income poverty. At the same time, being extremely poor is a significant barrier to enrolment in education and to health care utilization. 31. Addressing extreme income poverty. The finding of this study point to a number of possible policy levers to address the issue of extreme income poverty. Amongst those are: xii * Improving the short-term prospects for non-agricultural employment via labor intensive economic growth and a selective system of community-based public works focusing on the extremely poor; * Increase access to land for extremely poor households involved in agriculture; * Boosting the consumption level of the most needy by a sustainable social assistance system; and * Developing an adequate system of income related family allowances. Each of these options are discussed and assessed in the report. 32. Improving educational outcomes. The findings of this report point to the need to: * ensure that poor families can afford schooling; and * promote enrolment of girls and some minority groups in secondary education and, increasingly, in primary. 33. High costs of education appear to be the major barrier to enrolment. This reflects the high direct parental contributions paid across the income distribution. The report presents recent developments in this area and suggests some policy levers worth investigating. 34. A common intervention in low-income countries is to distribute family assistance programs throughout the school system, making them conditional on continued school attendance. This option is investigated in the report as a possible way of better targeting social assistance and breaking the vicious cycle between income poverty and poor education attainment. 35. Dealing with poor health outcomes and barriers to health care utilization. Affordability appears to be the main barrier to access in health care as well. In this context is it essential to * legalize and contain under-the-table payments; * establish realistic levels of health service output; and * collect more reliable disaggregated information on health care utilization and health status. 36. The need for future monitoring. Finally it is important to remember that this study represents a snap-shot of the living conditions of the population of Kosovo a year after the conflict. As such it closely reflects the unusual environment in which it was carried out. Given the lack of reliable information on trends, it is difficult to establish how far the results of this study are a reflection of the peculiarities of the year of the survey. Only future studies will be able to separate the transient finding from the long-term structural factors. In the context of medium-term monitoring, this study will represent an important benchmark against which the economic development of Kosovo and any future changes in the welfare of its population can be assessed and evaluated. xiii PERMBLEDHJE E EKZEKUTIVE 1. Eshte nje fakt qe pranohet gjeresisht se ngjarjet e dhjetevjecarit te fundit kane shkaktuar nje tendosje afatgjate ne mireqenien e nje numri te madh Kosovaresh. Tashme qe Kosova po kapercen fazen emergjente, duhet patur paraysh se gjendja qe la pas konflikti eshte e mjaft e rendesishme per konceptin e ,faredolloj strategjie te zhvillimit ekonomik. Ky studim eshte konceptuar par te ndihmuar debatin aktual te politikave ne lidhje me zbutjen e varferise dhe levrimin e sherbimeve sociale me strategjite e Struktures se Perbashket te Perkohshme Administrative (JIAS), Bankes Boterore, dhe donatoreve te tjere ne kete aspekt.. Ky raport percakton karakteristikat e me te varferve dhe grupeve me te prekshme prej varferise ne Kosove, analizon ,eshtjet kryesore qe prekin te varferit, si edhe sugjeron nje set politikash qe mund te jene efektive ne permiresimin e mircqenies. 2. Varferia perkufizohet ne menyre shume dimensionale, duke kaluar pertej nivelit te ulet te te ardhurave. Raporti permbledh nje game te gjere ceshtjesh, ku perfshihen konsumi, te ardhurat, arsimi, shendetesia dhe mbrojtja sociale. TE DHENAT DHE TREGUESIT E VARERISE QE JANE PERDORUR 3. Raporti bazohet ne te dhenat e nxjerra nga Anketa e Nivelit te Jeteses (LSMS)8. Anketa eshte nje vrojtim statistikor i kryer mbi bazen e familjes, i realizuar gjate periudhes shtator - dhjetor 2000, e cila perfshin 2,880 familje . 4. Eshte perdorur metoda e parl1ogaritjes se mireqenjes ne baze te konsumit. Vellimi I perqendrohet te varfria absolute, nderkohe qe shtrirja e varferise ekstreme dhe jo ekstreme eshte percaktuar duke ju referuar kufirit te varferise ushqimore, (FPL) dhe te kufirit te varferise se plote (PL). Perkufizimet e ketyre dy kufijve te varferise dhe vleres se tyre relative jepen nd Kutine 1. Kutia 1: Matja e Varferise ne lidhje me te ardhurat 1. Kufijte e varferise * Kufiri i varferise ushqimore (FPL): kostoja e nevojshme par te siguruar shporten minimale ushqimore 2100 kalori per nje person te rritur (DM 1.8539 per person te rritur ne dite) * Kufiri i Varferise se plote (PL): familjet mund te konsumojne edhe te mira te pervec ushqimit (DM 3,499 per person te rritur ne dite) Perkufizimi i Varferise: * "Varfdri e ekstreme": kur konsumi i familjes eshte < FPL * "Varferi": ne rast se konsumi i familjes eshte < PL 8 Per hollesi td metejshme per LSMS shih shtesen I td Volumit II. xiv 5. Ne mungese te treguesve mbi rezultatet e mesimdh&nies treguesit e arsimit jane percaktuar mbi bazen e perqindjes se analfabetizmit, arritjeve arsimore dhe parqindjes se rregjistrimit ne shkolle. Se fundi, pesha qe ze shendeti ne varferi eshte percaktuar nepermjet nje gershetimi treguesish te (i) gjendjes shendetsore dhe (ii)mundesise se manjes se sherbimeve mjekesore . PROFILI I VARFERISE NE KOSOVE SIPAS TE ARDHURAVE 6. Shtrirja e varferise dhe varferise ekstreme. Perfundimet ne te cilat ka arritur raporti lidhur me mat en e varferise sipas nivelit te te ardhurave paraqiten ne Kutine 2. Te dhenat e LSMS-se tregojne se aty rreth fundit te vitit 2000 varferia ishte shume e perhapur, dhe prekte me shume se gjysmen e popullsise. Megjithate, thellesia e varferise ishte relativisht e vogel: nje kontribut mesatar vjetor prej 400 markash gjermane per person, do te ishte i mjaftueshem par te nxjerre nje familje te varfer nga niveli i varferise.Konkluzioni i arritur se varferia eshte mjaft e perhapur por jo e thelle perforcohet edhe nga perfundimi qe nje perqindje relativisht e ulet (12 %) jetonte ne varferi ekstreme -qe do te thote se kishte nje nivel konsumi nen FPL. Hendeku i varferise ekstreme eshte gjithashtu relativisht i vogel vetem 2.5 % i kufirit te varferise ushqimore. Kutia 2 Shperndarja e Varferise sipas te ardhurave Rregjistrimi i Rregjistrimi i Hendeku i Hendeku i varferise ekstreme pergjithshem i varferise ekstreme varferise varferise Totali 11.9 50.3 15.7 15.7 Rurale 11.6 52.0 16.1 16.1 Urbane 12.5 47.5 15.1 15.1 7. Me fjale te tjera nje individ ne varferi ekstreme do te kishte nevoje per nje shume prej 144 DM markash gjermane per te permbushur kerkesat minimale per ushqim..Keshtu, nqs do te kishim nje percaktim perfekt te perfituesve , kostoja e nje programi perkrahjeje sociale i hartuar per te eleminuar varferine ekstreme do te arrinte rreth 34 million marka gjermane, qe korrespondon me rreth 1.3 % te PBB sipas vleresime te bera nga FMN mbi PBB per fryme ( 750 US $).Po te benin nje supozim me realist ku humbjet do te arrinin 25 %, kostojae perkrahjes sociale do te arrinte mbi 45 million marka gjermane, rreth 1.7 % e PBB 8. Varferia Ekstreme. Keto rezultate flasin per ekzistencen e nje grupi familjesh qe ndodhen nen kufirin e varferise se plote. Nje implikim i ketij rezultati eshte se dallimi i varfer dhe jo i varfer ka qene disi arbitrar ne kontekstin e Kosoves se pasluftes. Duke patur parasysh burimet e kufizuara buxhetore ,studimi sygjeron se nje strategji efektive per reduktimin e varferise duhet te fokusohet ne percaktimin e qarte te grupeve lehtesisht te dallueshme te varferise ekstreme, duke ia lene ceshtjen e pabarazise se te ardhurave dhe konsumit nje strategjie zhvillimi te vegant&. 9. Percaktimi i vecorive te varferise ekstreme. Pergjithesisht, familja 'tipike' ne varferi ekstreme perbehet prej 7 antaresh, me nje koeficient varesie te larte, gje qe eshte rrjedhoje e numrit te madh te femijeve dhe pranise ne gjirin e familjes te pakten e nje personi te moshuar. Kjo lloj familje banon ne zonat rurale, e cila si pasoje e konfliktit eshte zhvendosur te pakten nje here dhe tani eshte rikthyer ne vendin e saj te banimit. Kryefamiljari eshte rreth 50 vjeg, me nivel te ulet arsimor, punon ne bujqesi, ka me pak se nje hektar toke dhe nuk zoteron asnje makineri bujqesore. xv 10. Mbizoterimi i familjeve ne varferi ekstreme ne zonat rurale pasqyron me teper faktin qe Kosova eshte nje shoqeri rurale, sesa prezencen enje risku te larte te varferise ne familjet rurale. Permetej shkale e varferise ekstreme eshte me e larte ne zonat urbane se ne ato rurale dhe familjet ne varferi ekstreme ne zonat urbane ndryshojne shume prej homologeve te tyre te zonave rurale. Ne shumicen e rasteve niveli i arsimit ne zonat urbane eshte i larte dhe kryefamiljaret e familjeve ne varferi ekstreme kane qene te paune dhe nuk kane kerkuar pune per me shume se nje vit. 11. Familja tipike serbe ne varferi ekstreme ka te perbashket me homologen e saj shqiptare, residencen rurale nivelin e ulet arsimor dhe veprimtarine bujqesore. Sidoqofte, pergjithesisht ajo eshte shume me e vogel ne perbeje, me nje kryefamiljar ta moshuar dhe pa femije. 12. Percaktimi i elementeve qe ndikojne ne varferine ekstreme . Nder elementet qe ndikojne me shume mbi varferine ekstreme rradhiten: * Koeficint i larte i varesise 0.5 (ose me i larte) * Ne zonat rurale, pronesia e vogel mbi token (me pak se nje hektar), ose asnje lloj pronesie mbi te. * Ne zonat urbane, gjendja e tregut te punes - te papune ose jashte forces se punes - ose te paafte per pune. a Shtepi ne gjendje te keqe ose, dhe te demtur, mungese kanalizimesh te ujrave te zeza dhe uji te pijshem. 13. Pervec ketyre elementeve te cilet ndeshen ne gjithe popullsine ne pergjithesi, ekzistojne edhe disa grupe me te vogla, te cilet karakterizohen nga nje risk me i larte i varferise ekstreme.Ketu perfshihet Popullsia e Zhvendosur Brenda Territorit (IDPs) dhe "grupe te tjera etnike" - ve,anerisht romet, te cilet shpesh shfaqin edhe karakteristika te tjera qe lidhen me nje risk me te larte per te rene ne nje varferi ekstreme . Se fundi, familjet te cilat kane nje kryefamiljar te papafte per pund (te evidentuar nepermjet nje karte paaftesie) paraqesin nje risk me te larte per te rene ne varferi ekstrem , sidomos kur ato jane serbe dhe/ose jetojne ne zonat qytetare. BURIMET E KRYESORE TE TE ARDHURAVE 14. Pavaresisht nga shkaterrimet qa ishin rrjedhoje e konfliktit te nje viti pararealizimit te vrojtimit, rrogat dhe te ardhurat nga veprimtaria e biznesit vazhduan te mbeteshin mesatarisht burimi kryesor i te ardhurave per familjet. Dergesat ne te holla nga ana e te afermeve per familjaret e tyre perbejne burimin e dyte me te rendesishem te te ardhurave.Perafersisht rreth 50 % e tyre marrin dergesa ne te holla ose ne natyre nga te aferm te tyre jashte vendit dhe keto dergesa vijne duke rene ne zonat e thella te Kosoves.Dergesat ne te holla arrijne mesatarisht ne 4,983 marka gjermane ne vit shume te ciles i shtohen edhe 1,473 marka gjermane te marra si transferta ne natyre. Ne krahasim me kete vetem 5 % e familjeve serbe marrin transferta te kesaj natyre m nje vlere mesatare me pak se 1/3 e asaj qe marrin familjet shqiptare. 15. Rreth 60 % e popullsise kane deklaruar se kane marre ndihma ushqimore, veshje dhe lende djegese per ngrohje ne dimer, nga donatore te ndryshem, nje vit me pare se kryerja e ketij vrojtimi dhe se perqindja e tyre ishte me e larte ne zonat rurale. Per nje famije mesatare shqiptare vlera e ndihmes ushqimore ishte 382 DM marka gjermane, xvi ndersa per nje familje mesatare serbe kjo shume ishte 176 DM marka gjermane, pjeserisht ky ndryshim eshte o si rezultat i madhesise me te vogel te familjes shqiptare. 16. Me teper se 20% e familjeve shqiptare kane marre ndihme shtese per ndertime dhe riparime. Mesatarisht kjo ndihme arriti deri ne 5825 DM par familje. 17. Pabarazia e burimeve ne lidhje me te ardhuarat. Rendesia relative burimeve te ndryshme te te ardhuarave varion ne pika te ndryshme te procesit te shperndarjes se te ardhurave.Sic edh epritej roli i te ardhurave nga puna eshte me i larti perdhjeteshen me te ardhura me te larta, ndersa kontributi nga transfertat publike ulet ndjeshem linearisht me kalimin nga familjet me te varfra ne ato me te pasura.Cka bie me shume ne sy eshte kurba ne forme U e kthyer mbrapsh e kontributit si nga transfertat familjare dhe nga donatoret nderkombetare, Keto burime kane nje rol te rendesishem per te ardhurat ne 2 familjet me te varfera te dhjeteshes , roli i tyre me pas bie per tu ngritur perseri ne 4 familjet me te ardhura me te larta te dhjeteshes..Kjo do te thote qe vlera apsolute e ketyre dy Ilojeve te te ardhurave rritet linearisht me rritjen e te ardhurave. 18. Megjithate transfertat familjare jane burimi me rishpemdares i te ardhurave duke reduktuar pabarazine e te ardhurave-sic shprehet edhe nga vlera 17 e koeficientit G1NI.. Ndihma nderkombetare ka nje efekt me te vogel rishperndares per te dy grupet etnike nderkohe qe asistenca publike ndikon kryesisht mbi serbet. Kjo reflekton faktin se krahasuar me shqiptaret , familjet serbe terheqin shume me pak dergesa familjare dhe asistenca nderkombetare eshte me shume e perqendruar ne grupet me te ardhura me te ulta.Per kete grup etnik transfertat publike perbejne mbi 50 % te te ardhurave te familjeveqe bejne pjese ne intervalin 2-5 te grupit dhjetesh. VARFERIA DHE PABARAZIA NE FUSHEN ARSIMORE 19. Rezultati qe te terheq me shume vemendjen ne lidhje me treguesit e arsimit eshte ai qe nxjerr ne pah pabarazite e medha qe ekzistojne ne nivelin e te ardhurave, grupeve etnike dhe perkatesise gjinore. Ndersa perqindjet e rregjistrimeve neto ne shkollat fillore jane mesatarisht te larta (97%), vetem 76% e femijeve nen moshen 14 vjec, qe vijne nga grupe etnike te tjera jo nga Serbet dhe Shqiptaret - "grupet e tjera etnike" - jane te rregjistruar ne shkolla.Situata paraqitet me e keqe vecanerisht ne zonat urbane ku femijet qe i perkasin ketyre grupeve etnike gjenden kryesisht ne nivelin me te ulet te konsumit te grupit dhjetesh (28 %) dhe perqindja neto e regjistrimit ne shkolle eshte me e ulet se 15 %. Pra gershetimi i ketyre dy faktoreve shpjegon edhe rezultatet me shqetesuese se 20 % e femijeve nga zonat urbane kane nivelin me te ulet te konsumit dhe nuk ndjekin shkollen fillore. 20. Ne arsimin e mesem , pabarazia eshte edhe me e madhe. Gjate vitit 2000, vetem pak me shume se gjysma e vajzave shqiptare te moshes nga 15 deri ne 18 vjec ishin te rregjistruara ne shkolle, krahasuar me 3/4 e djemve shqiptar te regjistruar Ne krahasim me kete perqindja mesatare neto e regjistrimit ne shkolle per Serbet ishte afersisht 80 % me nje prirje te lehte gjinore drejt femrave. Perqindjet neto ta rregjistrimit ne arsimin e mesem per "grupet e tjera etnike" kane rene mesatarisht nd me pak se 55% dhe ne me pak se 40 % par vajzat. Me pak se gjysma e te rinjve me te varfer te dhjeteshesne zonat fshatare kishte te ngjare qe te rregjistrohej ne arsimin e mesam ne krahasim me rinjte me te pasur dhjetesh. Ndonese me pak i dukshem, ndryshimi ne zonat urbane eshte gjithesesi xvii mjaft domethenes, pasi vetem 62% e femijeve me te varfer regjistrohen ne arsimin e mesem, perkundrejt 90% te femijeve te pasur te dhejteshes. 21. Sipas rezultateve te LSMS-se arsyeja kryesore per mos-rregjistrim ne shkolle eshte faktori ekonomik. Ne nivel popullate , kostoja private e arsimimit eshte identifikuar si arsyeja kryesore per mos-rregjistrim ne shkolle (37%), ndersa 6,5% te tjere nxjerrin si shkak se duhet te punojne. Rezultatet e LSMS-se tregojne se pothuajse te gjitha familjet, te cilat kane femije te rregjistruar ne shkolle gjate vitit te fundit akademik (99%) kane bere shpenzime nga xhepi i tyre ne menyre suplementare. Ne 6% te rasteve, keto shpenzime kane mbuluar pagesat shkollore, 12% e tyre shkonin per koston e mesimdhenies, 11 % e tyre shkonin per shpenzime per libra shkollore dhe pajisje, nderkohe qe shpenzimet ne "natyre" dhe "te tera" arrinin deri ne 25%. Ndonese shumat e paguara ndryshonin perkundrejt zerave te tjere te shpemdarjes se konsumit, shpenzimet private per arsimim - vecanerisht te nivelit fillor- ishin gjithashtu te larta edhe per masen e varfer te popullsise. Ndryshimet jane te dukshme edhe nga pikpamja gjeografike dhe etnike; shpenzimet jane me te larta ne zonat urbane, vecanerisht per serbet, per te cilet keto shpenzime jane thuajse dy here me te larta se sa shpenzimet mesatare te familjeve. Nje pengese serioze per rregjistrim ne shkolla eshte percaktuar gjithashtu edhe siguria publike. Ne zonat urbane jane percaktuar edhe disa pengesa te tjera , per ndjeken e shkollave, , mes te cilave rradhitet edhe largesia e vendbanimit nga shkolla, gje qe permendet si arsye nga 5,4% e individeve te anketuar. GJENDJA SHENDETSORE DHE MUNDESIA E MARRJES SE NDIHMES SHENDETSORE 22. Te dhenat e pakta qe disponohen lidhur me gjendjen shendetesore, tregojne se Kosova gjendet ne nivelin me te ulet ne Europe, ne pothuajse te gjithe treguesit shendetesore. Vdekshmeria foshnjore, shendeti i nenes dhe renia e nivelit te vaksinimit dhe imunizimit, shfaqen si probleme mjaft shqetesuese. Anketimi i Serbeve tregon sistematikisht tregues shendetsor me te perkeqesur ne krahsim me shqiptaret .Ndersa nje pjese e ketyre ndryshimeve reflektojne ndryshimin ne strukturen e moshes midis ketyre dy popullsive, ndryshime shfaqen gjithashtu ne menyre te perseritur edhe per te njejtat grup mosha. Kjo ndoshta eshte pasqyrim i seleksionimit qe ka ndodhur, i largimit te Seberve te shendetshem nga Kosova.Gjithashtu edhe prirja gjinore shfaqet dukshem.Duke ndjekur edhe shembullin e vendeve te tjera, perqindja e grave me tregues shendetsor te perkeqesuar eshte me e madhe se ajo e burrave me perjashtim te treguesit "i paafte" 23. Pabarazia ne varesi te te ardhurave shfaqet pothuaj njelloj si ne shume vende te tjera .Treguesit e shendetit shfaqen me te perkeqesuar ne familjet me gjendje ekonomike me te keqe dhe kjo eshte vecanerisht e vertete ne rastin e treguesit te vetevlesimit te shendetit.Shkalla e barazise eshte cuditerisht e ulet dhe ndoshta reflekton konfliktin e fundit ne Kosove. 24. Sikurse edhe ne pjesen tjeter te rajonit, kujdesi shendetesor eshte kryesisht i llojit spitalor dhe atij te specializuar, nderkohe qe modelet dhe shkalla e perdorimit te tyre jane te ngjashme ne te gjitha grupet etnike dhe gjinite kur studiohen dallimet ne gjendjen shendetesore. Duke perjashtuar semundjet e vogla, per te cilat eshte konsideruar i mjaftueshem vettrajtimi, pengesa me e zakonshme per te marre perkujdesjen shendetesore duket se ka qene kostoja e shebimit perkates. Mesatarisht 28% e personave te anketuar, te cilet kane patur nje problem shendetesor, por qe nuk kane kerkuar trajtim, eshte pergjigjur xviii se ata kane vepruar keshtu per shkak se nuk mund te perballonin shpenzimet per kujdes mjekesor. 25. Pamundesia per te perballuar shpenzimet mjekesore, qe konsiderohet si nje nga pengesat per kujdes shendetsor, nuk duhet te na habise po te mbajme parasysh shpenzimet e medha suplementare qe kosovaret duhet te bejne per kete sherbim nga xhepi i tyre. Mbi 95% e shqiptareve kane paguar per sherbimet e marra, pavarsisht nga fakti qe ato sherbime sigurohen nga sistemi publik apo privat mjekesor. Pagesa mesatare ishte shume me e larte se sa shuma e parashikuar e bashkepageses, nderkohe qe zeri me i shtrenjte ne teresine e ketyre shpenzimeve ishte blerja e ilaceve, i pasuar nga "dhuratat". Se fundi, duket se serbet nuk kane pasur me shume vestiresi per te siguruar ndihme mjekesore me te lire dhe jane perballur me kosto me te ulta ne kete drejtim KONKLUZIONE DHE REKOMANDIME 26. Ky raport tregon se, ne fund te vitit 2000, 50,5% e popullsise se Kosoves ka jetuar ne varferi dhe se mbi 12 % e popullsise ka jetuar ne varferi ekstreme. Wshte e qarte se nje varferie e kesaj madhesie nuk mund te eliminohet vetem permes programesh transfertash dhe marrjes se masave per sherbimet sociale. Per me teper, ndryshimi ndermjet familjeve te varfra dhe estremisht te v'irfr, sikunder rezulton edhe nga analiza e mesiperme, nxjerr ne pah domosdoshmerine e miratimit te nje strategjie te dyfishte per reduktimin e varferise tani ne Kosove. 27. Permiresimi i mundesive per te varferit. Familjet qe ne vitin 2000 ishin te varfera - por ekstremisht te varfera - zoterojne nje lloj pasurie si edhe nje fare kapitali financiar dhe njerezor, gje qe tregon se gjendja e tyre e varferise mund te reagoje shpejt ndaj nje rritjeje te pergjithme ekonomike te mbeshtetur intensivisht mbi punen.Ne pjesen me te madhe keta individe jane fermere te vetpunesuar, me nje p pronesi toke relativisht te madhe. Kur ata jetojne ne zonat qytetare, ata jane relativisht te rinj, kane nje koeficient varesie relativisht te ulet dhe kane nje kryefamiljar me nje nivel te mire arsimor, por qe, aktualisht eshte i papune ose jashte forces se punes. 28. Nje komponent i rendesishem i strategjise se reduktimit te varferise, duhet te jete rritja ekonomike me baze te gjere, qe do te sigurohet nga zbatimi i reformave kyqe strukturore dhe ngritjes se institucioneve te afta per te mbeshtetur kalimin ne ekonomine e tregut. Zhvillimi i nje strategjie te pershtatshme e te qendrueshme rurale, e konceptuar per te rritur rendimentin ne bujqesi dhe per te krijuar mundesi punesimi ne sektoret perkates duhet te perbeje nje nga elementet me te rendesishem te kesaj strategjija rritjeje. Pervojat e vendeve te tjera kane treguar gjithashtu se nje burim i rendesishem per krijimin e vendeve te reja te punes eshte sektori i ndermareve te vogla dhe te mesme (SME). Zgjerimi dhe perhapja e ndermarrjeve te vogla dhe te mesme, si ne sektorin e sherbimeve, ashtu edhe ne ate te industrise se lehte, kane nje rendesi vendimtare per promovimin e nje zhvillimi te bazuar mbi punen intensive. Por qe te realizohet nje gje e tille, Kosova ka nevoje te zhvilloje dhe te zbatoje me tej rregullat dhe mekanizmat e tregut, si ne planin ligjor ashtu edhe ne ate institucional, pasi sektori privat mund te kete sukses vetem ne kete menyre. 29. Mekanizmat per zbutjen e varferise ekstreme. Edhe ne rastin e supozimeve me optimiste, eshte e veshtire qe ne vitet e ardhshme te parashikohet nje rritje ekonomike dhe nje zhvillim i sherbimeve sociale, i afte te permiresoje mundesite per personat dhe grupet xix e individeve qe gjendjen ne varferi ekstreme.. PerteVetem keshtu do te evitohet evituar krijimi i nje nenklase te marxhinalizuar per ekonomine dhe strukturen shoqerore e cila do te jete e paafte per te marre sherbimet sociale. 30 Rezultatet e ketij raporti tregojne se konceptimi i ketyre nderhyrjeve duhet te pasqyroje edhe shkallen e larte te mbivendosjes qe konstatohet nga dimensionet e ndryshme te varferise. Niveli i ulet arsimor dhe gjendja e keqe shendetesore jane elemente shume te rendesishem qe ndikojne mbi te ardhurat dhe qe te cojne ne kufijte e nje varferie ekstreme. Ne te njejten kohe, varferia ekstreme perben nje pengese te madhe per rregjistrimin ne sistemin arsimor si edhe per marrjen e kujdesit shendetesor. 31. Trajtimi i varferise estreme ne lidhje me te ardhurat. Perfundimet e ketij raporti kane nxjerre ne pah nje sere mekanizmash te mundshem per te trajtuar ceshtjen e varferise ekstreme ne lidhje me te ardhurat Nder to perfshihen: * Rritja e te ardhurave nga bujqesia duke rritur siperfaqen e tokes ne perdorim nga familjet ne varferi ekstreme; * Permiresimi ne punesimin afatshkurter per ata persona qe gjenden ne varesi ekstreme ne zonat urbane, nepermjet nje sistemi perzgjedhjeje komunitet-pune publike; - Rritja e nivelit te konsumit per me te varferit nepermjet nje sistemi asistence sociale te qendrueshme; * Zhvillimit te nje sistemi pagesash duke marre ne konsiderate te ardhurat e familj eve; Ne raport analizohen dhe vleresohen secili prej ketyre opsioneve. 32. Permiresimi i treguesve te arsimit. Perfundimet e ketij raporti nxorren si te domosdoshme: * T'ju sigurohet familjeve te varfra mundesia e perballimit te shkollimit * Nxitja e regjistrimit ne arsimin e mesem te vajzave dhe e disa grupeve te pakicave dhe rritje e regjistrimeve ne arsimin fillor. 33. Kostoja e larte per arsimin eshte deklaruar qe te jete pengesa kryesore per regjistrimin ne shkolle. Kjo pasqyron pagesat e larta direkte te paguara nga prinderit qe ndikojne ne shperndarjen e te ardhurave. Raporti paraqet zhvillimet e fundit ne kete fushe dhe sygjeron disa mekanizma te politikave qe duhen studiuar me tej. 34.. Nje nddrhyrje e zakonshme ne vendet me nivel te ulet te ardhurash eshte zbatimi i programeve se asistences familjare ne te gjithe sistemin shkollor, duke e kushtezuar ate me frekuentimin e rregullt dhe te vazhdueshem te shkolles. Ky opsion eshte studiuar ne raport si nje mekanizem i mundshem per percaktimin e objektivave te qarta per vajzat dhe grupet e tjera etnike. 35. Trajtimi i treguesve te perkeqesuar shendetesore dhe i pengesave per marrjen e ndihmes shendetsore. Duket se pamundesia per te paguar perben gjithashtu pengesen kryesore per te marre ndihmen e nevojshme mjekesore.Ne kete konteks eshte e domosdoshme qe te: * Legalizohen dhe frenohen pagesat nen dore; * Percaktohen nivele realiste te sherbimit shendetsor, dhe xx * Mbledhja e te dhenave me te besueshme dhe me te hollesishme per gjendjen shendetsore dhe shkallen e marrjes se saj. 36. Nevoja per monitorim ne te ardhmen. Se fundi eshte e rendesishme qe te kujtojme se ky studim paraqet nje veshtrim te shpejte mbi nivelin e jeteses se popullsise se Kosoves vetem nje vit mbas konfliktit. Si i tille ai reflekton dukshem edhe situaten e pazakonte ne te cilen ai eshte kryer .Duke patur parasysh mungesen e te dhenave per tendencat ne pergjithesi, eshte e veshtire per te percaktuar se deri ne cfare mase rezultatet e ketij studimi ndikohen nga viti i kryerjes se tij.Vetem studimet e ardhshme do te bejne te mundur ndarjen e perfundimeve tranzitore nga faktoret strukturor afatgjate. Ne konteksin e monitorimit afatmesem ky studim paraqet nje objektiv te rendesishem nepermjet te cilit mund te vleresohet zhvillimi i ekonomise se Kosoves dhe cdo ndryshim i ardhshem ne mireqenien e popullsise. xxi IZVRSNI REZIME 1. bpste je poznato da su dogadaji poslednje dekade ostavili teske posledice na dugorocnu socijalnu zastitu velikog broja Kosovara. Obzirom da Kosovo upravo izlazi iz faze hitne humanitarne pomoci, od vitalnog znacaja za buducu strategiju ekonomskog razvoja jeste ozbiljna briga o onima koji su ostali na Kosovu nakon konflikta. Ovaj izvestaj je uraden kao informacija za razgovor o tekucoj politici u oblasti smanjenja siromastva i pruzanja socijalnih usluga, kao i o strategijama razvoja Zdruzenih privremenih administrativnih struktura (JIAS), Svetske banke i drugih donatora u ovim oblastima. Izvestaj je posebno znacajan pri identifikaciji prioriteta kao i za sugerisanje budu6ih nacina a radi boljeg iskorisenja postojecih resursa. Ovaj izvestaj identifikuje karakteristike najsiromasnijih i najosetljivijih grupa na Kosovu, pokrece kljucna pitanja koja pogadaju siromasne i daje predlog skupa politickih poluga koje de verovatno biti najefikasnije za poboljsanje njihovog socijalnog stanja. 2. Siromastvo je definisano na visedimanzionalan nacin, koji se proteze izvan oblasti niskog prihoda. Na ovaj nacin izvestaj pokriva sirok dijapazon pitanja, ukljucuju6i potrosnju, prihod, obrazovanje, zdravstvenu i socijalnu zastitu. Korisceni podaci kao indikatori siron1astva 3. Ovaj izvestaj se prvenstveno koristi podacima iz Pregleda ispitivanja 2ivotnog standarda (LSMS)'. Ovo je bilo ispitivanje onih domacinstava koja predstavljaju statisticki reprezentativan uzorak, a sprovedeno je u periodu od septembra do decembra 2000 godine i obuhvatalo je 2.880 doma6instava. Tabela 1: Merenje siroma*tva usled dohotka Pravci siromastva: Pravac siromaA§tva ishrane (FPL) - cena minimuma potrosacke korpe za isbranu, koja obezbedjuje 2100 kalorija po odraslom (DM 1.8539 na dan, po odraslom) All, domacinstva moraju trositi i druge proizvode osim hrane: Koliko? Dodatnih 47% * Pravac siroma.wtva (PL): DM 3.499 na dan, po odraslom = 1.47 x 1.8539 2. Definicija siromastva: "Ekstremno siromasgtvo". ukoliko je potrosnja domacinstva < FPL "Siroma§tvo": ukoliko je potrosnja domacinstva < PL I Za dodatne informacije o LSMS videti dodatak I u Delu II xxii 4. Merenje socijalnog stanja na osnovu potrosnje korisceno je za analizu siromastva usled dohotka. U Delu I, paenja je usmerena na apsolutno siromastvo, a obim esktremnog i neekstremnog siromastva je procenjen u odnosu na pravac siromastva ishrane (FPL) i praavc potpunog siromastva (PL). Definicija ove dve granice siromastva kao i njihova relativna vrednost dati su u Tabeli 1. 5. U nedostatku indikatora rezultata ucenja, obrazovni rezultati su definisani u smislu stopa nepismenosti, obrazovnih dostignu6a i stopa upisa. Konacno, zdravstvena dimenzija siromastva sagledana je u okviru kombinacije indikatora i) zdravstvenog statusa i ii) dostupnost zdravstvene zastite. Profil siromastva usled prihoda na Kosovu 6. Stepen siromastva. Rezultati i nalazi izvestaja o siromastvu usled dohotka sumirani su na Tabeli 2. Podaci iz ankete LSMS ukazuju da je siromastvo bilo rasprostranjeno u drugoj polovini 2000. godine, pogadaju6i vise od polovine stanovnistva. Ipak, stepen siromastva je relativno mali: Prosecan doprinos od DM 400 godisnje po osobi izvukao bi siromasno domacinstvo iz siromastva. Radi zadovoljenja minimalnih potreba za hranom, ekstremno siromasnom pojedincu bilo bi potrebno 144 DM godisnje. Dakle, uz pretpostavku savrsenog planiranja, program socijalne pomoci izraden radi eliminisanja ekstremnog siromastva kostao bi oko 34 miliona DM, sto, prema procenama MMF za GDP (bruto nacionalni dohodak) po glavi stanovnika odgovara 1.2 % GDP. Sa relisticnim pretpostavkama o 50 % odlivanja troskova socijalne pomoci, doci 6e do povecanja na oko 2 % GDP. Tabela 2: Podaci o siromastvu usled dohotka Ekstremno Ukupno Jaz Tezina siromastvo siromastvo siromagtva siromastva Ukupno 11.9 50.3 2.5 15.7 Ruralni 11.6 52.0 2.7 16.1 Urbani 12.5 47.5 2.4 15.1 7. Zakljucak sveobuhvatnog, ali ne dubokog siromastva pojacan je nalazom da je srazmemo daleko nizih 12 procenata stanovnistva zivi u ekstremnom siromastvu - tj. da su nivoi potrosnje bili ispod nivoa FPL. Ovi rezultati ukazuju na znacajno grupisanje domacinstava neposredno ispod linije siromastva. Implikacija ovog rezultata jeste da je razlika izmedu siromasnog/ne-siromasnog unekoliko proizvoljna u kontekstu Kosova posle sukoba, i ukazuje da u svojim naporima za smanjenje siromastva, ogranicena budzetska sredstva kao i efikasna strategija za smanjenje siromastva treba da se koncentrisu na posebnu grupu koja jasno mo2e da se identifikuje kao grupa ekstremno siromasnih, ostavljajuci pitanje nejednakosti u potroTnji i prihodu za posebnu strategiju rasta. 8. Identifikovanje karakteristika ekstremno siromasnih. U principu, "tipicno" domacinstvo koje je ekstremno siromasno sastoji se od 7 clanova visokog stepena zavisnosti, zbog velikog broja dece i prisustva bar jedne starije osobe. Doma6instvo zivi u seoskoj xxiii sredini, raseljeno je bar jednom tokom sukoba i sada se vratilo na svoje prvobitno mesto stanovanja. Glava porodice ima oko 50 godina, nizak nivo obrazovanja, radi u poljoprivredi ali ima manje od jednog hektara zemlje i nikakvu masineriju. 9. Ipak, preovladavanje ekstremno siromasnih domacinstava u seoskim sredinama odraz je cinjenice da je Kosovo u najvecoj meri ruralno drustvo, te da je pojava ekstremnog siromastva neznatno visa u gradskim sredinama nego u seoskim. Ova ekstremno siromasna domacinstva razlikuju se od seoskih po nivou obrazovanja i statusu zaposlenosti glave porodice. U gradskim sredinama nivo obrazovanja je visi, a glave ekstremno siromasnih domacinstava ne rade i ne traze posao duze od 12 meseci. 10. Tipicno srpsko domacinstvo koje je ekstremno siromasno deli sa albanskim stanovanje na selu, nizak nivo obrazovanja i poljoprivrednu aktivnost. Medutim ovo domacinstvo je, u principu, malo, glava porodice je starija osoba i nema dece. 11. Identifikovanje korelacija ekstremnog siromastva. Najsnaznija korelacija ekstremnog siromastva obuhvata: - Stope zavisnosti u visini - 0,5 ili vise - U seoskim sredinama, bez ili sa malo (manje od 1 hektara) zemlje. - U gradskim sredinama, status na trzistu rada - nezaposlen ili van radne snage - i invalid. - Nesigumo i/ili stanovanje u oste6enim kudama, i nedostatak kanalizacij i vodovoda. 12. Pored ovih korelacija koje se odnose na celokupno stanovnistvo, postoje druge manje grupe koje se karakterisu veoma visokim rizikom ekstremnog siromastva. One obuhvataju intemo raseljena lica (IDP) i "druge etnicke grupe" - posebno Rome, koje cesto iskazuju druge karakteristike koje su povezane sa visokim rizicima ekstremnog siromastva. Na kraju, doma6instva u kojima je glava porodice invalid (kao dokaz nosilac je invalidske karte) takode predstavljaju znacajno vise rizike ekstremnog siromastva, posebno ako su Srbi i/ili zive u gradskim sredinama. Izvori prihoda siromasnih 13. I pored razaranja zbog sukoba u godini pre ispitivanja, plate i zarade iz poslovnih aktivnosti su ostale glavni izvor prihoda za prosecno doma6instvo. Transferi od rodaka cinili su drugi najvazniji izvor prihoda, gde je skoro 50% albanskih domacinstava u seoskim sredinama primalo doznake u gotovini ili naturi, u poredenju sa samo oko 5% srpskih domacinstava. 14. U proseku, domadinstva na Kosovu su dobijala doznake u iznosu od DM 5506 od rodaka izvan Kosova, pri cemu su albanska domadinstva primala u proseku DM 5662 u poredenju sa DM 1475 koliko su primala srpska domacinstva. 15. U toku 12 meseci pre pocetka istrazivanja, preko 60% stanovnistva dobijalo je pomoc u hrani a procenti su bili i visi u ruralnim oblastima. Albanska domadinstva su, u poslednjih xxiv 12 meseci, dobila pomo6 u hrani vrednu 382 DM, a dok su za isti period srpska domacinstva dobila samo 176 DM u hrani. Preko 20% albanskih doma6instava primilo je i dodatnu pomoc u vidu gradjevinskog materijala, zimskog ogreva, odece, itd. U proseku, ova pomoc iznosi 5825 DM po domacinstvu. Siromastvo i nejednakost u stecenom obrazovanju 16. Najupadljiviji rezultat izvestaj u oblasti postignuca u obrazovanju su velike razlike koje postoje u prihodima, po etnickim grupama i polu. Dok je neto stopa upisa u osnovne skole u proseku visoka (97%), samo 76% dece mlade od 14 godina iz drugih etnickih grupa osim Srba i Albanaca - "druge etnicke grupe" - je upisano u skolu, a razlika prema polu je u okviru ove grupe velika - samo 69 procenata upisanih devojcica naspram 89 procenata upisanih decaka. Pored toga, cak 20 procenata dece iz desetice doma6instava sa najnizom potrosnjom u gradskim sredinama nije upisano u osnovnu skolu. 17. Nejednakost kod upisa u srednje skole je jog dramaticnija. Samo nesto preko polovine albanskih devojaka izmedu 15 i 18 godinaje bilo upisano u skolu u 2000. godini. Neto stope upisa na srednje obrazovanje za "druge etnicke grupe" palo je ispod 55 procenata u proseku, a na manje od 40 procenata za devojke. Sto se tide mladih iz najsiromasnije desetice iz seoske sredine verovatno je bilo da ce manje od polovine da se upise na srednji nivo obrazovanja u odnosu na one iz najbogatije desetice. Mada manje upadljiva, znacajna je i razlika u gradskim sredinama, sa samo 62 procenata upisane najsiromasnije dece nasuprot skoro 90 procenata one iz najbogatije desetice. 18. Nalazi istrazivanja LSMS ukazuju na cinjenicu da su ekonomski faktori glavno obrazlozenje za neupisivanje u skolu. Za celokupno stanovnistvo, privatni troskovi obrazovanja bili su identifikovani kao glavni razlog za neupisivanje (37 procenata), sa dodatnih 6,5 procenata koji ukazuju na rad kao glavni faktor. Nalazi LSMS pokazuju da su skoro sva domacinstva koja imaju decu koja su upisana u poslednjoj skolskoj godini (99 procenata) davala znacajne doprinose >>iz d2epa<. U 6% slucajeva ova placanja su pokrivala upisne troskove, u dodatnih 12% slucajeva pokrivala su skolarinu, dok su davanja »>u naturio i »ostali trogokvio iznosili i preko 25%. Mada se pla6eni iznos razlikovao po distribuciji potrosnje, privatni troskovi za obrazovanje - posebno osnovno - bili su visoki i u delu sa manjom potrosnjom. 19. Geografske i etnicke razlike takode su ocigledne, sa troskovima koji su visi u gradskim sredinama, posebno za Srbe koji se suocavaju sa gotovo dvostrukim srednjim troskovima domacinstva. Bezbednost je takode identifikovana kao vazna prepreka za upis. Ostali razlozi koji se odnose na pristup gkoli znacajni su posebno u seoskim sredinama, sa tim daje 5,4 procenta stanovnistva iz uzorka navelo udaljenost skole. xxv Zdravstveni status i pristup zdravstvenoj zastiti 20. Ogranikeni raspolozivi podaci o zdravstvenom statusu ukazuju na to da se Kosovo nalazi na najnizem mestu u Evropi, prakticno po svim indikatorima zdravlja. Smrtnost novorodencadi, zdravlje majke i pad broja vakcinisanih su oblasti koje posebno brinu. Medutim, veoma malo se zna o uzrocima tako loseg stanja, a postojeci podaci ne dopustaju bilo kakve nejednakosti kod ovih indikatora. Informacije iz ankete LSMS o tri indikatora zdravstvenog statusa odraslih pokazuju velike nejednakosti vezane za etnicku pripadnost i pol, ali ima malo podatka o nejednakosti koja se odnosi na prihod. 21. LSMS sadrzi detaljne informacije o koriscenju zdravstvene zastite u toku cetiri nedelje, i u toku 12 meseci pre sprovodjenja ankete. Kao i u ostalom delu ovog regiona, zdravstvena zastita je na osnovu bolnicke ili specijalisticke zastite, a uzorci i stope koriscenja su slicni u etnickoj grupi i po polu, posto se uzmu u obzir razlike u zdravstvenom statusu. 22. Iskljucuju6i manja oboljenja za koja se smatra da je samolecenje dovoljno, najcesca prepreka pristupu zdravstvenoj zastiti izgleda da su troskovi usluga. U proseku, 28 procenata onih koji su prijavili da su imali zdravstveni problem za koji nisu trazili lecenje, razlog za to je bio jer nisu mogli sebi da priuste zdravstvenu zastitu. 23. Uloga dostupnosti kao prepreka za dolazak ne iznenaduje kada se uzme u obzir da je to najvece gotovinsko placanje sa kojim se suocavaju na Kosovu. Preko 95 procenata Alabanaca je platilo dobijene usluge, bez obzira da li su one date u drzavnim ili privatnim objektima. Prosecno placanje bilo je daleko vece od ocekivanih iznosa doplate, a najskuplja pozicija troskova bili su lekovi praceni "poklonima". Na kraju, izgleda da su Srbi imali laksi pristup besplatnoj zdravstvenoj zastiti i izgleda da se suocavaju sa nizim troskovima. Zaklju&ci i preporuke 24. Ovaj izvestaj pokazuje da je, na kraju 2000. godine, 50,5% stanovnistva Kosova zivelo u siromatvu. Jasno je da se siromastvo ovog obima ne moze eliminisati samo transferom programa i pruianjem socijalnih usluga. Stavise, razlika u karakteristikama ekstremno siromasnih i siromasnih domacinstava, kako se vidi iz gomje analize, ukazuje na dvosmemu strategiju za smanjenje siromastva na Kosovu danas. 25. Poboljsanje izgleda za siromasne. Domacinstva koja su u 2000. godini bila siromasna - ali nisu bila ekstremno siromasna - imaju zalihu sredstava i nivo finansijskog i ljudskog kapitala koji ukazuju da njihov status siromasnog brzo reaguje na sveukupni radno- intenzivni rast privrede. Ona su prvenstveno samostalni poljoprivrednici sa nekim sredstvima i relativno velikim zemljisnim posedima. Kada zive u gradskim sredinama relativno su mladi, stopa zavisnosti im je niska i imaju dobar obrazovni nivo, ali su trenutno nezaposleni ili se ne nalaze na trzistu radne snage. 26. Glavna komponenta strategije smanjenja siromastva zbog toga mora da se usmeri na obnavljanje ekonomskog razvoja na sirokoj osnovi, pokrenutog kljucnim struktumim reformama i institutcionalnom izgradnjom za podrsku tranzicije na trzisnu ekonomiju. xxvi Razvoj adekvatne i odrzive strategije za seoske sredine projektovane za razvoj produktivnosti u poljoprivredi i razvoj radnih mesta u sektorima koji su sa njom povezani je vazan elemenat ove razvojne strategije. Iskustva sa drugih strana su takode pokazala da ce glavni izvor za nova radna mesta biti sektor malih i srednjih preduze6a (SME). Ekspanzija malih i srednjih preduzeca, kako u uslu2nim delatnostima tako i u lakoj industriji, jeste od presudne vaznosti za unapredenje efikasnog radno intenzivnog razvoja. Ali, da bi se to dogodilo, Kosovo mora dalje da razvija i primenjuje zakone i trzisne mehanizme, kako pravne tako i institucionalne, pod kojima ce privatni sektor uspeti. 27. Mere za ublazavanje ekstremnog siromastva. Cak i prema izrazito optimistickim pretpostavkama, tesko je zamisliti da de ekonomski razvoj i razvoj socijalnih sluzbi u narednih nekoliko godina znacajno poboljsati izglede "ekstremno" siromasnih. Za ovu neresivu grupu pojedinaca, potrebne su specijalne intervencije protiv siromastva da bi se izbeglo da postanu jedna niza klasa pojedinaca, marginalna za ekonomsku i drustvenu strukturu. 28. Rezultati ovog izvestaja ukazuju da projekat za ove intervencije mora da bude odraz postojeceg velikog stepena preklapanja raznih dimenzija siromastva. Slaba obrazovna postignuca i log zdravstveni status imaju veliku korelaciju sa ekstremnim siromastvom prihoda. Istovremeno, ekstremno siromastvo postaje znacajna prepreka za upis na skolovanje i koriscenje zdravstvene zastite. 29. Pitanje siromastva usled dohotka. Strategija za poboljsanje prihoda i nivoa potroTnje ekstremno siromasnih moze da koristi jedna ili vise od slededih politickih mera: * Pomaganje nivoa potrosnje onih koji za tim imaju najvecu potrebu odrzivim sistemom socijalne pomo6i; * Poboljsanje perspektiva za zaposljavanje izvan poljoprivrede i pogodnosti tih pojedinaca preko selektivnig sistema javnih radova sa ucescem zajednice; i * Porodicni dodaci vezani za dohodak. * Svaka od ovih opcijaje razmatrana i procenjenja u ovom izvestaju. 30. Poboljsanje nivoa obrazovanja. Nalazi ovog izvestaja ukazuju da su visoki troskovi obrazovanja glavne prepreke za odlazak na skolovanje. Na Kosovu je obrazovanje skoro u potpunosti obezbedeno drzavnim sistemom. Direktni roditeljski doprinosi skolama nisu zvanicno obavezni, mada ih ima i gotovinski troskovi domacinstava za obrazovanje izgleda da se protezu na velik deo prihoda prilikom njegove raspodele. UNMIK je tokom poslednje dve godine uspeo da stavi skolama na besplatno raspolaganje neke udzbenike i materijale za najbitnije predmete, ali ovo izgleda nisu ni bili glavni izvori javne potrosnje. Bez obzira na ovo, UNMIK treba dobro da razmotri usmeravanje svojih ogranicenih sredstava na ona davanja siromasnim doma6instvima koja 6e rezultirati povecanim upisom njihovih clanova u osnovne, a posebno srednje skole, gde su gotovinska pla6anja visa. 31. Uobicajena intervencija u zemljama sa niskim prihodom je da se raspodeljuju programi pomoci za porodice kroz skolski sistem, tako da budu uslovljeni neprekidnim pohadanjem skole. To je svakako opcija koju treba dalje ispitati u vezi sa porodicnim dodatkom koji je gore razmatran. Moze se, na primer, razmisljati o pausalnoj stopi dodatka xxvii za svako dete koje pohada skolu, koja se placa svakom domacinstvu koje ima vise od troje dece ili o necem slicnom. 32. Ranije analize ukazuju da je pristup obrazovanju za devojke ogranicen, posebno u seoskim sredinama, kao i za "druge etnicke grupe" sirom Kosova. S obzirom da je ranije na Kosovu bio postignut skoro potpun upis na osnovno obrazovanje i na jaku tradiciju obaveznog osnovnog obrazovanja, povratak na pun upis u osnovno obrazovanje na Kosovu ne mora da trazi skupe ili slozene intervencije. Medutim, porast upisa devojaka i "drugih etnicTkih grupa" u srednji nivo obrazovanja moze da zahteva usmereno i ciljno investiranje. Potreba za ovom vrstom intervencije treba da se adekvatno proceni. 33. Bavljenje losim zdravstvenim rezultatima i preprekama za koriscenje zdravstvene zastite. Izgleda da je finansijska dostupnost takode glavna prepreka pristupu zdravstvenoj zastiti. U ovom kontekstu od sustinske je vaznosti da se legalizuju i zaustave placanja ispod ruke i da se ustanove realni nivoi zdravstvene sluzbe. Ovo je odlucujuci korak u smanjenju korupcije u sistemu, kao i u osiguravanju finansiranja kako visoko prioritetne intervencije tako i subvencionisanja onih koji nisu u mogu6nosti da sebi priuste potrebnu negu. 34. Potrebne su pouzdanije rasclanjene informacije identifikovanje teku6ih slabosti sistema. Kao deo ovih prava potrebno je da se regulisu realni finansijski nivoi sa proporcionalnom zastitom. Ovo moze da se odnosi na usluge koje drzavni sektor mora da prestane da pruza, i1i da znaci proveru pristupa jednakog osiguranja u zdravstvenoj zastiti. Budu6i pristup mozda 6e morati da napravi kompromis za smanjenje prevelikih kapaciteta kako fizicke infrstrukture tako i u ljudstvu. 35. Na kraju, takode je vazno uloziti napore da bi se obezbedilo vakcinisanje u drzavnom zdravstvu, adekvatan pristup cistoj vodi i smanjenje ekoloskih rizika po zdravlje. xxviii 1. INTRODUCTION A. THE BACKGROUND 1.1 Kosovo was the poorest province of the Socialist Federal Republic of Yugoslavia (SFRY). In 1988, output per head in Kosovo was 28 percent of the SFRY average. Immediately after the Second World War, Kosovo was predominantly an agricultural economy, and agriculture continues to be a very important sector in Kosovo today. Nevertheless, during the 1970s-80s, Kosovo's development became a priority for the SFRY government, and deployment of significant resources from the national investment fund led to some degree of industrialization and the development of mining as a major sector of economic activity. By 1988 industry and mining comprised nearly half of the total national product, with agriculture down to just over 20 percent.9 Earnings, however, remained low at 67 percent of the average for the whole Yugoslavia, 74 percent of the average for Serbia and 44 percent that of Slovenia.10 Furthermore, health care statistics- with the exception of life expectancy-were among the lowest in Europe. During the same period, access to Albanian language education increased significantly. As a result, enrolment doubled between 1968 and 1978, and illiteracy dropped to below 10 percent. Kosovo also operated its own social protection system under the general umbrella of the SFRY legislation. This included pensions and unemployment benefits, means-tested child allowances and cash transfers for households with no resources. 1.2 The suspension of the autonomy of Kosovo within the republic of Serbia, in 1989, marked the beginning of a downward trend in the living standards of the ethnic population of Kosovo as a whole and of the Albanians in particular. In the first half of the 1990s, GDP contracted by 50 percent, to less than US$400 per head. The decline was particularly evident in industry and mining, so that by 1995 the structure of output had reverted to that of the early 1970s.11 The drop in employment that followed affected predominantly the Albanian population who were expelled from employment in the public sector. Ethnic Albanians responded by developing a large informal economy. 1.3 The response of the ethnic Albanians to the restriction on the use of the Albanian language at school was along similar lines. In 1991-92, the majority of ethnic Albanian students left the formal educational system. This marked the beginning of the overcrowded and under-funded parallel system of education for the Albanians, which ran 9 World Bank (2001) 0 World Bank (2000) The World Bank (2001). ibid 1 along side the traditional Serb system for a decade. Similarly, the centralization in Belgrade of the financing of health care and the dismissal of Albanians from senior positions in the health sector resulted in the Albanians creating a 'parallel' system of health care delivery. This was based on a combination of: * a private health care system run almost exclusively by Albanians and meeting the needs of those who could afford to pay the fees charged; and, * a parallel charitable health care system organized through the Mother Theresa Society and designed to provide basic health services free of charge to those members of the Albanian community who were unable to afford private medical care and were unable-or unwilling-to use the official system. 1.4 Finally, the centralization of the social security system in Belgrade in 1989, introduced a systematic discrimination against Albanians in terms of eligibility and access to the system. Thus as from 1990, for many ethnic Albanians the main safety net has been the extended family network, the remittances from the Albanian community outside Kosovo, and the services of many humanitarian organizations, such as the Mother Teresa Society (MTS). 1.5 Ethnic violence escalated in 1998 and 1999, putting increasing pressure on the fragile economy and social fabric, and leading to displacement. More than 800,000 ethnic Albanians were displaced to Macedonia, Albania and other foreign countries. Another 500,000 were displaced within Kosovo, and 9,000-10,00012 were killed or are missing. At the end of the crisis, as the majority of the Albanian refugees returned to Kosovo, more than 100,000 Serbs and Romas fled.'3 1.6 The conflict ended in June 1999 with UN Resolution 1244, which accorded Kosovo 'substantial autonomy and meaningful self-administration,' while recognizing the 'sovereignty and territorial integrity of the Federal Republic of Yugoslavia.' The resolution also established the United Nations Interim Administration Mission in Kosovo (UNMIK) as a transitional government with four areas of responsibility: (i) humanitarian affairs; (ii) interim civil administration; (iii) institution building; and (iv) reconstruction. Six months later, UNMIK signed an agreement on a Joint Interim Administrative Structure (JIAS) with a number of local representatives. This structure currently administers Kosovo. During this period local and international actors have been attempting to establish a sustainable system of government and lay the ground work for the next stage of economic development. 1.7 The consequences of the 1999 conflict on the living standards of the population were severe. Industrial output collapsed, agriculture activities came to a halt, and livestock was decimated. Currently the economy appears to be on the way to recovery. Thanks to massive donor assistance, domestic savings and inflows of diaspora funds, the 12 The lower figure is a UNFPA/IOM (2000) estimate; the higher figure is a UTNHCR (1999) estimate. 13 UINHCR (1999) estimate. 2 entrepreneurial spirit of the 1980s is reviving, repairs to housing and infrastructure are under way, and agricultural activities have begun again. 1.8 However, the conflict has claimed its toll on the long-term welfare of a large proportion of the population. As Kosovo goes beyond the emergency relief phase, taking into consideration the situation of those that the conflict has left behind, is vital to the effectiveness of any strategy for economic development. JIAS recognizes the difficulties of promoting sustainable growth and providing adequate social protection. Alleviating widespread poverty and addressing existing inequalities is an essential part of its development strategy for Kosovo. Poverty in Kosovo in 2000 1.9 This report provides a profile of poverty in Kosovo as defined in a multidimensional way, focusing on lack of opportunity and lack of capabilities (Box 1.1). It then identifies the correlates of poverty and assesses the key issues that affect the poor. Box 1.1: A multi-dimensional Definition of Poverty The World Development Report 2000 defines poverty as a multidimensional concept extending beyond low levels of income. This report focuses on three dimensions of poverty: * Lack of opportunities: Low levels of consumption/income in relation to an agreed poverty line. This is generally associated with the level and distribution of physical assets (such as land), human capital and social assets; and market opportunities which determine the returns to these assets. * Low capability: Low human development indicators, such as education and health both at aggregate level and among a particular socio-economic group. * Vulnerability and coping strategies: exposure to risk and income shocks which may arise at the national, local, household and individual level. The availability and ease of access to different coping strategies is important in this context. Source: World Bank estimatesfrom LSMS. What are the Characteristics of Poverty in Kosovo? 1.10 Three factors suggest that poverty in Kosovo may have unique characteristics: * The events of the last decade, including the role of the 1999 conflict; * The marked ethnic differences in the demographic structure of the population; and * The unique role played by the family networks and migration in the socio- economic structure of the ethnic Albanian population. 1.11 The legacy of the last decade has left a society which is ethnically segregated along geographical and economic lines. The Serb population is concentrated in the Northern region and in small, predominantly rural, enclaves in the South from which it is prevented from traveling by security restrictions. The private economy is 3 overwhelmingly in the hands of the Albanian population, prevented for a decade from working in what used to represent the main employment source, the public sector. Infrastructure is very poor-especially in Albanian areas-after years of neglect followed by months of conflict. The 1999 conflict took a heavy toll on Kosovo. At the end of the conflict, physical and human damage were widespread. The conflict affected most severely housing, agriculture, and telecommunications. 1.12 The demographic structure of the population-as it emerges from the LSMS sample. The size and composition of Kosovo's population by ethnicity and location of residence - as it emerges from the LSMS sample - is given in Table 1.1. Kosovo households are large - with over 6 members on average - and extend beyond the single family unit. Larger households are more common among Albanians (6.76 members per household) than the Serbs (3.93 members per household) and other ethnicities (5.36 members per household). As expected, household size is higher in rural than in urban areas. Kosovo's population is young. The average age of the population is 27.3 years and 50 percent of the population is below the age of 22.5 years. This is mainly due to the high fertility of ethnic Albanians which is the highest in Europe at present, with a total fertility rate of over 2.5 children per woman14. Table 1.2 shows the distribution of Kosovo's population, by ethnicity, according to the LSMS findings. It is important to note that the picture emerging reflects the large population movements occurred during and after the conflict. More details on this are given below. Table 1.1. Household Size in Kosovo Rural Urban Total Albanian 7.57 5.93 6.76 Serb 4.22 3.64 3.93 Other 5.13 5.56 5.36 Total 6.90 5.54 6.22 Source: World Bank estimates from LSMS Table 1.2. Population by Ethnic Group (in percentage) Albanian Serb Slav Muslim Roma Turk Others Total 88.1 7.0 1.9 1.7 1 0.3 100 Source: World Bank estimates from LSMS 1.13 Family networks and migration. The Albanian population of Kosovo is characterized by a household structure-often referred to as Zadruga by Balkan scholars-which consists of extended and multiple households, which often involve several generations of the same family living together and functioning as a single economic unit. The entire social and economic structure for the Albanian population evolves around strong kinship relations and large family households. Migration is an important part of this social structure with large households selecting one or more members-usually male-for migration as a strategy for income diversification. This strategy became particularly common during the past decade and the recent conflict when job opportunities for Albanians in Kosovo were increasingly limited. 14 UNFPA and IOM (2000) 4 1.14 Internally displaced population, refugees and fertility. One of the major consequences of the recent conflict was the large fraction of the population that were internally displaced or migrated abroad as refugees. The Kosovo LSMS evidenced different patterns across ethnic groups, as it is evident from Table 1.3. The Albanians suffered significant displacement during the conflict. However, a year later the largest proportion of displaced population was Serb. Table 1.3 Place of Residence Before The Conflict (% ofpopulation) Albanian Serb Other Total Same place 89.8 91.4 96.2 90.2 Other Municipality 4.1 7.5 0.2 4.1 Abroad (returnees) 2.9 0.6 1.0 2.6 Newbom 3.3 0.6 2.6 3.1 Total 100 100 100 100 Excluding returnees and newborns: Same place 95.7 92.5 99.8 95.6 Other Municipality 4.3 7.5 0.2 4.4 Source: World Bank estimatesfrom LSMS. 1.15 In addition, the LSMS shows that the largest fraction of returning emigrants and of recent births were Albanian.'5 These categories combined account for more than 6 percent of the ethnic Albanian population. The Serbs on the other hand had the lowest incidence of returnees and fertility rates, reflecting the out-migration from the region and the age structure of those still residing there. The LSMS was not able to capture the extent of out-migration since the survey was not implemented outside Kosovo. Available evidence suggests that between 100,000-150,000 Serbs left Kosovo in recent years, which would explain the low share of Serbian ethnicity in the 100,000. 1.16 Finally, the LSMS suggests displacement among "other ethnic groups" to be low, but this rate could also be negatively affected by out-migration to neighboring countries. B. PROVIDING A BENCHMARK FOR FUTURE MONITORING 1.17 This study represents a snap-shot of the living conditions of the population of Kosovo a year after the conflict. As such it closely reflects the unusual environment in which it was carried out, a period in which socio-economic infrastructure was on the way to recovery but remained torn by a decade of civil unrest. Economic production was at a minimum. Agriculture, the main source of income for most of the population of Kosovo in the past, was still hindered by the presence of landmines and the displacement of the population. Private economic activities were restarting and large numbers of Albanians were coming back from abroad after spells of migration of different lengths. International relief efforts were still intense but declining. Given the lack of reliable 15 Returning emigrants are defined as individuals currently living in Kosovo but living abroad before the conflict. 5 information on trends, it is difficult to establish how far the results of this study are a reflection of the peculiarities of the year of the survey. Only future studies will be able to separate the transient finding from the long-term structural factors. 1.18 The main purpose of this report is to inform the JIAS transitional government in its difficult task of bringing the economy of Kosovo beyond the emergency phase and in providing a picture of poverty and opportunity in Kosovo at the close of the conflict. Because of its timing it is also expected to act as an important benchmark for monitoring future trends on poverty and living conditions. C. THE DATA 1.19 This report uses data primarily from the Living Standard Measurement Survey (LSMS). This is a household survey carried out between September and December 2000, and covering 2,880 households.16 Great effort was made to create a representative sample of the population of Kosovo. In the absence of a census, up-to-date listings of households were created using different methodologies for rural and urban areas. The details of the methodology used are provided in the LSMS documentation. The sample was then drawn to be representative at the level of the existing areas of responsibility (AORs) and locations (urban and rural).17 It is also designed to be representative of the two main ethnic groups (Albanian and Serbian) but not of other ethnicities, which in the report are referred to as 'other ethnic groups'. This category comprises a variety of peoples, often very different in their characteristics, ranging from Muslim Slav (including Bosniac and Gorani) to Roma, and including Turk, Croat and Montenegrin. D. THE PURPOSE OF THE STUDY AND THE STRUCTURE OF THE REPORT 1.20 The present study on poverty is designed to provide inputs into the current policy debate and into the development strategy of the JIAS, the World Bank and other donors. Its focus is on poverty diagnostics and to provide options for a poverty reduction strategy for Kosovo as it enters a new phase of development. For this purpose, the report distinguishes the hard core of extreme poor from the rest of the poor. The latter have enough human and physical capital to benefit from the recovery of the economy and the economic growth that is expected in the next years. The welfare of the former, however, continue to be at high risk in this new stage of development. The report identifies priorities and suggests ways in which the existing resources can be better targeted and can be used more effectively to improve the welfare of this particularly vulnerable group. 1.21 The rest of this volume is organized as follows. Chapter 2-which focuses on poverty as lack of opportunities-provides a profile of income poverty and describes the characteristics of the income poor. Chapter 3 outlines the sources of livelihood of the poor. Chapter 4 describes poverty as lack of education and identifies its correlates and 16 For more details on the LSMS see http://www.worldbank.org/Isms/guide/select.htm1. 17 Kosovo is divided into five areas, called UNMIK and NATO Areas of Responsibility (AOR), roughly equivalent to the former regions: Southeast (United States of America), East, including Pristina (United Kingdom), North (France), West (Italy), and South (Germany). 6 characteristics. Chapter 5 looks at the health status of the Kosovo population. The final chapter offers recommendations for policies that could improve the stake for the most vulnerable groups and suggestions on a possible poverty reduction strategy for the future of Kosovo. In addition a number of background papers is collected in Volume II. 7 2. POVERTY AS LACK OF OPPORTUNITIES A. DEFINING POVERTY Choosing a Welfare Indicator 2.1 In a post-conflict environment, the measurement of welfare faces significant difficulties. Following a well-established procedure, this report employs a consumption- based measure. Such a measure has the advantage of reflecting the long run socio- economic condition of households better than income-based indicators. However, in a period of extensive reconstruction, like the year 2000 in Kosovo, the value of consumption aggregate may be inflated by the need to replace basic assets and depleted capital. This would lead to unusually high levels of expenditure on non-food items and a generalized over-estimation of household welfare. Thus the resulting estimates of the incidence of poverty would underestimate its real value. 2.2 In the Kosovo region, the structure, size, and composition of households vary considerably across ethnic groups and geographical location. In order to take these differences into account, when comparing welfare across households, per-adult equivalent consumption (PEC) is used. This is derived using a Kosovo specific equivalent scale,'8 which is used to define the equivalent size of households on the basis of the structure and number of individuals in a household. PEC consumption is then obtained by dividing household consumption by its equivalent size. The household level PEC are then price-adjusted to allow for intra-regional variations in prices. Setting a Poverty Line and Choosing Poverty Indicators 2.3 The focus of this study is on absolute poverty.19 By using the calorie intake information from the Survey, this study estimates the food poverty line (FPL) as the cost of the food basket required to provide 2,100 calories per day (per adult). For the sake of comparison with other studies, it is worth mentioning that this is approximately equivalent to US$0.813 per adult per day.20 Households below the FPL are then considered to be in extreme poverty. Households, however, do not spend only on food, and a poverty measure should provide some room for expenditure on non food items. A Complete poverty line or simply poverty line (PL), it estimated and households with a 18 For more details see Chapter 1 of Volume 2. 19 In accordance with common practice Volume 2 uses a combination of relative and absolute poverty measures to describe poverty (Chapter 1). 20 For more details on the estimation procedure and the values of the basket, see Chapter 1 of Volume 2. 8 daily consumption below such PL are considered to be in poverty. For cross-study comparisons it is again interesting to note that the estimated PL is about US$1.50, which is very close to the median PEC for Kosovo. The FPL and the PL can also be expressed in terms of household consumption.21 For the modal household structure of four adults and two children the minimum food consumption is about DM 200 per month, and the total household budget --including clothing, utilities and other expenditures - amounts to DM 370.22 2.4 Given the chosen poverty line, this study presents three standard poverty indicators to capture different dimensions of the problem. The measures used are (i) poverty incidence, (ii) depth, and (iii) severity. Poverty Incidence is the head count ratio that indicates the fraction of the population in poverty (or extreme poverty). Poverty depth measures the consumption deficit or the gap between actual consumption and the poverty line audits also referred to as the 'poverty gap'. This is interpreted as the aggregate deficit of the poor relative to the poverty line. T Severity of poverty measures the poverty gap giving more importance to those households that are farther from the poverty line. B. THE EXTENT OF POVERTY IN POST-CONFLICT Kosovo 2.5 The LSMS data suggest that, one year after the end of the conflict, poverty was widespread in Kosovo.25 It is evident from Table 2.1 that in Fall 2000, about half of the population of Kosovo had consumption levels below the PL (column 2). Although international comparisons are difficult due to the country specific poverty line used, this is a very high incidence of poverty compared to the rest of the region. Similar orders of magnitude are only found in Tajikistan and Moldova. However, the value of the poverty gap suggests that the depth of poverty was relatively small. An average annual contribution of DM 400 per person would have brought a poor household out of poverty on average. The conclusion of pervasive but not deep poverty is reinforced by the finding that only 12 percent of the population was in extreme poverty (column 1). An extremely poor individual would have needed an average of DM 144 per year to achieve the minimum food requirement.26 21 Since the FL and PL are expressed in terms of adult-equivalent consumption, the household consumption needs to account for economies of scale and different costs of children compared to adults. 22 The estimation of monthly budgets on food and other expenditures for other household structures can be found in Chapter 1 of Volume 2. 23 For more details see Chapter I of Volume 2. 24 For more details see Chapter 1 of Volume 2. 25 Volume 1 focuses on the absolute measures of poverty described above. For details of the relative measures refer to Chapter 1 of Volume 2. 26 The potential cost of filling this gap with social assistance under different hypothesis on the efficiency of the target is discussed in Chapter 2. 9 Table 2.1. Kosovo Poverty Indicators (poverty measures in percentages, standard errors in parenthesis) Extreme Overall Extreme Overall Poverty Severity of Poverty Poverty p Severty Headcount Headcount Gap Poverty (1) (2) (3) (4) Total 11.9 50.3 15.7 6.8 (0.9) (1.5) (0.7) (0.4) Location Rural 11.6 52.0 16.1 6.9 (1.1) (2.0) (0.8) (0.4) Urban 12.5 47.5 15.1 6.7 (1.7) (2.3) (1.2) (0.8) Regions (AOR) 1. North-East 12.4 46.5 15.2 6.7 (2.1) (3.3) (1.6) (1.0) 2. South-West 15.9 59.5 20.0 9.3 (2.2) (3.8) (1.7) (1.0) 3. West 6.8 42.5 11.6 4.6 (1.4) (2.7) (1.1) (0.5) 4. North 12.2 58.4 17.5 7.2 (1.6) (2.5) (1.1) (0.6) 5. South-East 11.2 47.6 14.3 6.2 (1.9) (3.2) (1.4) (0.7) Survey samples Albanian 12.0 49.7 15.6 6.8 (1.0) (1.6) (0.7) (0.4) Serb 11.0 58.6 17.5 7.3 (1.9) (2.9) (1.3) (0.8) Source: World Bank estimates from LSMS 2.6 This result suggests a significant clustering of households just below the poverty line. This is evident from Figure 2.1 that shows the density of the logarithm of PEC and the monthly food and poverty lines (in logs: 4.02, 4.65). The high concentration of households between the two vertical lines is evident. 2.7 The picture that emerges is one of a relatively small hardcore of extremely poor households and a low percentage of very rich households, with a large concentration of households on either side of the poverty line. This is confirmed by the very low level of inequality, as measured by an estimated Gini coefficient of 0.29.27 The implication of this result is that the poor/non-poor distinction is somewhat arbitrary in the Kosovo context and suggest that in their poverty reduction efforts policy makers should concentrate on the clearly identifiable and distinct group of the extremely poor, leaving the issue of inequality of consumption and income to be addressed by the growth strategy. 27 See footnote 5 above for details on the Gini coefficient. 10 Figure 2.1: Kosovo per adult-equivalent consumption distribution. .098- 0 0 0 2.05 6.75 ln(PEC) Source: World Bank estimates from LSMS. The Role of Food Aid in Containing Poverty 2.8 In the post-conflict year 2000 the Kosovo economy was operating at minimum capacity, employment opportunities were limited, agricultural production decimated by a year of conflict and a sizeable proportion of the population was still displaced. In order to mitigate the potential effects of this situation on the welfare of the population, a number of internationals donors provided large quantities of emergency food aid. 2.9 According to the LSMS data the average value of food aid distributed was about DM 20 per month per household, and varied between DM 32 for the poorest population quintile to DM 11 in the richest group. Thus food from aid represented almost 20 percent of total food consumption for households in the poorest quintile, versus less than 2 percent among the richest. 2.10 The poorest quintile received about 31 percent of the total food aid, compared to the richest quintile that captured only 10 percent. This evidence is suggestive of a targeting mechanism that is reaching the poorest but still showing signs of leakages. This is confirmed by comparing poverty measures with and without food aid (Table 2.2). The incidence of extreme poverty would have been almost 15 percent, as compared to 12 percent without food aid; and overall poverty incidence would have been 52 percent instead of 50.3 percent. Thus food aid has been well targeted, in so far as it has reduced extreme poverty more than overall poverty, however its impact has been modest. 11 Table 2.2: Impact of Food Aid on Poverty and Inequality Poverty Percentage indicators indicators poverty indicators without food aid reduction (percent) (percent) Extreme poverty (< Food Poverty Line) Incidence 11.9 14.6 23 Extreme Poverty Gap 2.5 3.9 36 Extreme poverty severity 0.8 1.5 47 Overall poverty (including extreme) (< Food Poverty Line) Incidence 50.3 52.2 3 Poverty Gap 15.7 17.8 12 Poverty severity 6.8 8.4 19 Inequality indicators Gini coefficient 0.2845 0.2992 0.0146 Theil entropy measure 0.1336 0.1476 0.0140 Theil mean log deviation 0.1346 0.1531 0.0185 Source: World Bank estimates from LSMS. C. WHO ARE THE EXTREMELY POOR 2.11 Before the conflict, the MTS defined the most needy members of the Albanian population to be the disabled, elderly, orphans, landless and unemployed.28 Immediately after the conflict, conflict-related damages were the most important indicators of poverty. These included destroyed housing, and destruction of harvest, irrigation and other related infrastructure and livestock. In addition loss of family members and conflict-related trauma were claimed to be strongly associated with increased vulnerability as they were believed to undermine coping strategies.29 Six months after the conflict, the poorest households remained those directly affected by the conflict (e.g., displacement, destroyed housing) plus those lacking resources (without salary, pension, remittances, savings, or property), with many dependents (orphans, elderly, disabled), and geographically and/or economically isolated.30 28 World Food Program. (2000) 29 ibid. 30 ibid 12 2.12 About one year after the conflict the qualitative poverty assessment conducted by the Inter-Agency Sub-Group on Poverty reported the local perception of poverty as given in Box 2.1.31 Box 2.1: Local Perceptions of Poverty The poor have irregular or no income. They do not have savings. They have few skills and little education. They have little or no land. They often do not have livestock. They eat poor quality food. They depend on handouts from neighbors and humanitarian organizations. They live in old mud or stone houses with one or two rooms that lack basic amenities. They dress very poorly and cannot even afford to wash their clothes. Most of their children do not receive more than primary education. They can not afford to buy medicine when they are sick. The poor have no means to recover from the losses incurred during the conflict and rebuild their lives without assistance. Source: Kosovo: A Qualitative Poverty Assessment. 2.13 The reference to poor quality of food in the Qualitative Poverty Assessment suggests that the category under consideration is closely related to the definition of 'extreme poverty' used in this report. However, the portrait of the 'typical' extremely poor household as it emerges from the LSMS reflects only partially the local perceptions captured in the Qualitative Poverty Assessment. 2.14 The extremely poor are predominantly Albanian but the incidence of extreme poverty is the same across Serbs and Albanians. As shown in Table 2.1 above, the Kosovo LSMS shows the population of Kosovo in 2000 to be predominantly Albanian (88 percent). Given that the Albanians make up such a large share of the overall population, it is not surprising that they also represent the majority of the poor and the extremely poor. However, the percentage of individuals in extreme poverty is very similar across the two main ethnic groups while it appears to be much higher for the Roma population (60 percent) and the Muslim Slav and Gerami (23 percent). 2 While the Roma represent only 1.7 percent of the population of Kosovo, they make up 8.4 percent of the extremely poor. For the Gerami and Muslim Slavs the situation is only slightly better: they represent 1.6 percent of the population, but 3 percent of the extremely poor. 2.15 The majority of the extremely poor live in rural areas but the incidence of extreme poverty is higher in urban areas. As more than 60 percent of the Kosovo population lives in rural areas, it is not surprising that the majority of the poor and extremely poor households live in rural areas. As shown in Table 2.1, the proportion of the population in extreme poverty is slightly higher in urban areas. This could be due to a combination of factors. First, rural households might be able to consume from farm production, either from crops or livestock, avoiding extreme poverty. Second, the impact of the conflict on urban economic activities (and labor markets) may have hit some urban households particularly hard. More importantly, currently internally displaced households and returnees are more likely to live in urban centers, where they make up more than 11 31 Oxford Policy Management (2001). 32 In deriving conclusions on the 'other ethnic groups' however, it is important to remember that the data are not statistically representative for these ethnic groups. 13 percent of the urban population and face poor living conditions and consumption restrictions.33 What are the Characteristics of the Extremely Poor and How Do They Differ From Those of the Poor? 2.16 The rural poor differ from those in urban areas. Box 2.2 begins by summarizing the characteristics of the typical extremely poor household in Kosovo in 2000 and highlighting how they differ from poor households. Identifying these characteristics is important to the design of policy measures which will have the largest impact on poverty. Box 2.2. Stylized Facts on The Extreme Poor and 'Typical' Poor Household Extremely Poor Poor (not extremely) * Rural Albanian households with 7 * 8 members (3 adults) members at least half of which are <15 or >60. * Household head around 50 years of age * Household head has secondary with low level of education. education Overall * Formerly Internally Displaced (IDP). * Higher asset value than extreme * Live in damaged dwellings. poor. * Have < I hectare of land and no machinery. * Use latrines. Same characteristics as 'overall' except that: * Smaller households with 1 child or * Head has higher education attainment without children Overall and Urban * Head is not participating in the labor market * Better sewage (flushing toilet) * Rural households concentrated in North * Equally likely in urban and rural and Southeast regions. areas. * 3-member households with elderly * Larger and younger households Only head and no children. with 2 children. Serb * Head has low level of education and * Head has secondary education. works in agriculture. * Did not move during the conflict. Source: World Bank estimates from LSMS. 33 According to the LSMS about 66 percent of the currently internally displaced population and 73 percent of the returning population live in urban areas. 14 2.17 Overall the 'typical' household in extreme poverty has 7 members with a high dependency ratio, due to a large number of children and the presence of at least one elderly person, it lives in a rural area, has been displaced at least once during the conflict and is now back to its original place of residence. The household head is around 50 years of age, with a low level of education, works in agriculture but has less than one hectare of land and no machinery. The households in poverty - not including the extremely poor - however, are larger, have a better educated head and more assets. 2.18 Poor Serbs differ from the Albanians. The typical Serb household in extreme poverty shares with the Albanian the rural residence, the relatively low level of education and the agricultural activity. However, it is very small, headed by an elderly person and with no children. The poor Serb household, however, is equally likely to live in rural as in urban areas, has a larger household and two dependents under 18. D. IDENTIFYING THE CORRELATES OF EXTREME POVERTY 2.19 Table 2.1 above identifies ethnic origin and the urban/rural dimension as important correlates of poverty. Using multivariate analysis, a number of additional variables that are strongly correlated with poverty have been identified. These are represented in Box 2.3 and can be distinguished into two groups: * The traditional correlates of poverty, i.e., household composition, education and health status of the household head, poor employment opportunity and size of land used; and 3 Those factors that are more linked to the post-conflict situation from which Kosovo is emerging: displacement, destruction of housing, infrastructure and disabilities. Standard Correlates of Poverty34 2.20 The presence of vulnerable groups such as children under 15, the elderly and female heads is larger among the extremely poor. The link is particularly strong in rural areas and for specific ethnic groups, suggesting a higher vulnerability in those areas. 2.21 Fraction of elderly in the household. Among the Albanians almost all the poor and extremely poor (93 percent) live in comparatively young households where the elderly represent less than 25 percent of the members. Among the Serbs, by contrast, 35 percent of the extremely poor live in households where more than 25 percent are elderly. Incidence of extreme poverty among elderly dominated households is particularly high in rural areas, resulting in higher relative risks of extreme poverty, particularly among the Serbs. 34 For details of the numbers given and relevant tables please see Chapter 1 Volume II. 15 2.22 Fraction of children under 15. Among the Albanians most of the poor and extremely poor (72 percent) live in households where more than a quarter of the members are children. Poverty and extreme poverty incidence increases with the fraction of children in the household (more clearly among the Albanians), but in rural areas the effect becomes significant only when the children make up more than half of the household members. Box 2.3. Factors Affecting Risk of Extreme Povert3 5 Increase risk Decrease risk Overall Dependency ratio Non participation in labor force Secondary education Agricultural worker Current Internally Displaced (IDP) Self-employment Precarious dwelling and stones/mud materials Latrine or absence of sewage Land and agricultural assets Urban Dependency ratio Household head holding a disability card Self-employment Non participation in labor force Current IDP. Not moved during conflict Precarious dwelling or damaged dwelling Latrine Serb Older head with primary education Household head holding a disability card Younger heads Non participation in labor force Little machinery and livestock Latrine Lack of piped water (private or public) Source: World Bank estimatesfrom LSMS. 2.23 Female headed households (FHH). On average, the incidence of extreme poverty among female headed households is slightly larger (14 percent) than other households (12 percent), but this difference is negligible when considering poverty, rather than extreme poverty. However, significant differences are observed between urban and rural areas. While the incidence of extreme poverty among female headed households is considerably higher in rural areas, in urban areas the difference is very small. This suggests an increased vulnerability of females in rural areas that jeopardize consumption opportunities, which could be associated with the absence of prime age adult male members.36 Moreover, the higher incidence of extreme poverty among female headed households seems to be due predominantly to their demographic structure and 35 Extreme poverty risks are factors that significantly increase the probability of being in extreme poverty. These are not simply relative risks but actually conditional relative risks, that is resulting from extreme poverty regressions. The control household is one living in an intact dwelling made of cement and bricks, flushing toilet, centralized water, with a salaried head that never moved during the conflict. 36 This is particularly corroborated by the qualitative study that points to the death or disappearance of a husband, father or other male relative as a source of vulnerability in rural areas. See Oxford Policy Management (2001). 16 other characteristics, such as low employment status and education levels. When these factors are taken into account the disadvantage of female headship seems to disappear.37 2.24 Education of the household head. The education of the household head is a strong correlate of poverty. The incidence of poverty among households whose head is not educated is more than 60 percent higher than the average and is more than twice that of households with a head who attended secondary school. 2.25 Labor market characteristics of the poor. Most of the extremely poor are non- participants in the labor market38 (41 percent) - a category which is unusually large in the labor market of post-conflict Kosovo - or unemployed. Moreover a very high proportion of these without work fall into extreme poverty. In urban areas the non participants account for 52 percent of the extreme poor. Lack of employment opportunities, therefore, appears to be a major cause of poverty. 2.26 Land.39 About 32 percent of the rural population live in households with less than 1 hectare of land holdings. The incidence of extreme poverty is highest among this group of farmers (19 percent), while landowners with more than 2 hectares of land have a very small incidence of extreme poverty. 2.27 Agricultural machinery and other assets. Individuals living in households with no agricultural machinery have the highest relative incidence of extreme poverty (29 percent more than other categories). Moreover, almost 69 percent of the extreme poor have no agricultural machinery, which appears to represent a major obstacle for increasing production and productivity in rural areas. While the lack of assets, especially for small farms, may have predated the conflict, the extensive destruction of farm buildings, agricultural machinery and animals during the conflict may have worsened the situation.40 Most of the destruction took place in western municipalities which is corroborated by the large fraction of population with no assets in the Western AOR (63 percent) compared to 49 percent in the Northeast or Southeast. 2.28 Livestock. Rural households tend to diversify activities as a means to reduce the exposure to adverse shocks (pests, floods, prices, etc). In Kosovo, almost all rural households have some animals on their farm.41 On the basis of self-reported values, about 26 percent of the rural population lives in households with livestock worth less than DM 1,000. Most of the population of Kosovo (47 percent) lives in households that have livestock worth between DM 1,000 and 2,000. The distribution of livestock value seems more uniform than that of land or machinery, reflecting the use of livestock as a savings 37 This is evident from the results of the regression analysis in Chapter 1 Volume 2. 3S These are defined as individuals without a job who have not looked for one in the last 12 months. 39 Household land holdings include land owned, rented and borrowed. 40 World Bank (2000) 41 Only 9 households out of 1,440 reported no animals in their farm. 17 instrument and potential direct consumption during an adverse shock.42 Extreme poverty risk shows a large increase for households with less than DM 1,000 in livestock, while the composition of the poor and extreme poor offer no clear difference across livestock holders. 2.29 Living conditions Extreme poverty is twice as common among households living in accommodations with walls made of stones and mud. Having no toilets and/or using latrines is also a strong correlate of poverty. Conflict-Related Correlates 2.30 Previous qualitative analysis has identified conflict-related damage as an important determinant of poverty. Here three conflict related dimensions are identified: displacement, housing destruction and disabilities. 2.31 Displacement. According to the LSMS, about 70 percent of Kosovo's population was displaced during the 1999 conflict (March 24 to June 9), with about 36 percent displaced inside Kosovo and about 64 percent displaced outside Kosovo, mostly in Albania, FYR Macedonia, and Montenegro. The majority of this population returned to the areas from where they had fled within a few months of the end of the conflict. Perhaps because of this relatively short displacement period, the incidence of poverty is not higher among this group. However, 4 percent43 of the total population, and 8 percent of all Serbs, were still displaced within Kosovo -'Internally Displaced Population' (IDP) - at the time of the survey and an additional 3 percent was represented by individuals who were outside Kosovo before the conflict and have returned to Kosovo after the conflict: the 'diaspora returnees' or 'returning migrants'. The currently IDP and those returning emigrants are twice as likely to be extremely poor than the rest of the population. This is due to a combination of poor access to land and the lack of non-agricultural employment opportunities. 2.32 Housing. On average households living in precarious dwelling (tents, barracks and other) or damaged dwellings are also twice as likely than the rest of the population to be extremely poor with a 24 percent incidence rate. In urban areas the link between precarious and damaged accommodation and extreme poverty is even stronger with nearly half of the extremely poor living in this type of accommodation. 2.33 Disabilities. The LSMS includes information on two measures of disability: (i) the official Disability Card and (ii) self-reported functional disability that measures the 42 In a separate study it would be important to examnine the scale and compositions of livestock holding by socioeconomic status, since poorer households may tend to have small livestock (chicken, sheep) while richer households may have larger animals (cows). This not only reflects the available resources allocated to savings but also the liquidity constraints faced by each type of household. 43 This figure only refers to those living in households headed by a displaced person. In addition, it does not include those living in collective centers or camps. As one would expect, the qualitative assessment found living conditions in collective centers and camps to be particularly difficult, with many of them overcrowded and lacking basic necessities. 18 difficulty to perform some activities. Unfortunately neither of these definitions allows the distinction between conflict-related and pre-existing disabilities to be drawn. On average just over 6.7 percent of the population lives in a household whose head holds a disability card but more than 14 percent of the heads-23 percent of the Serbs-are classified as disabled according to the functional criteria. 2.34 The statistically representative results of the LSMS confirm the earlier conclusions of the qualitative study: disabilities, however measured, are strong correlates of extreme poverty. E. CONCLUSIONS AND POLICY IMPLICATIONS 2.35 The difference in the characteristics of the extremely poor and the poor households, as highlighted above, suggests an important conclusion. The households who are currently poor-but not extremely poor-have a stock of assets and a level of financial and human capital that will allow them to take full advantage of future growth opportunities. If living in rural areas, they are self-employed farmers with some assets and comparatively larger land holdings. In urban areas they exhibit low dependency rates, are relatively young and have a head with good education levels but who is currently unemployed or out-of-the-labor force. It is therefore justifiable to think of them as 'post- conflict' poor that have been brought into poverty by the economic decline of the last decade and/or by the recent conflict. The extremely poor, on the other hand, have characteristics that make them much more unlikely to benefit from economic growth. They tend to live in households with high dependency ratios and a poorly educated head. If in agriculture, they have little land and no machinery and, if Serbs, they are predominantly elderly. These findings point towards the need for a two-prong strategy for poverty reduction in Kosovo today. 2.36 Improving the odds for the poor. For the non-extremely poor the key to higher welfare is broad based economic growth. Since the majority of these individuals live in rural areas and are predominately farmers, it is important to prepare a sustainable rural strategy which builds on the relatively high level of education of these individuals and their access to land. The main objective of this strategy should be to increase productivity in agriculture and develop job opportunities in related sectors. The findings of this study also show that the non-extremely poor in urban areas are well-equipped to benefit from any increase in job opportunities in the non-agriculture labor market which growth can bring along. Another essential component is the return to some form of socio-economic stability. This is needed to improve the opportunities of the Serbs in general and those living in the enclaves in particular. The current fragmentation of society and market opportunities along ethnic and geographical lines is likely to prove a considerable constraint on the development of a modern agricultural sector. 2.37 Alleviating extreme poverty. By way of contrast, even high labor-intensive growth is unlikely to be sufficient to deal with extreme poverty. For this hardcore of individuals specific anti-poverty interventions are required. The effectiveness and efficiency of these policy levers can be greatly improved by their being informed by the findings on the characteristics of the extremely poor and the correlates of extreme 19 poverty as they emerge from Box 2.2 and 2.3 above. Amongst others, these identify the following correlates of extreme poverty in Kosovo: * High dependency ratio; * Being IDP's; * No or small (less than 1 hectare) land holdings; * Shortage of machinery and equipment in agriculture; and * Lack of employment in urban areas,- being unemployed or out of the labor force. Of these only the last three represent factors of possible intervention in a medium term poverty reducing strategy. 2.38 Small land holdings in rural area is a strong correlate of extreme poverty. In addition to improved opportunities for non agricultural employment, extremely poor rural households would benefit from improved access to and a more efficient use of the available land. Most of the agricultural land is under private ownership but about 12 percent of the agricultural land is under the regime of collectively owned farms. Ancillary evidence showed that in these cooperatives only 25 percent of the land was cultivated. Although just a small fraction of the total arable land in Kosovo the unused land could represent a significant improve in the economic condition of the rural households with excessively small holdings. . A policy lever in rural areas could entail improving the conditions for land use by clearly defining ownership and property rights. In addition improving access to credit and to input markets remains a priority. 2.39 Another serious restriction to production in rural areas is the lack of machinery and tools for agriculture. Among those in extreme poverty in rural areas, almost 70 percent have no machinery and the average value of the capital stock (DM 630) is less than half that of the poor (DM 1,270). 2.40 The large fraction of non participants and unemployed among the extremely poor might be attributed to the effects of the recent conflict and lack of development of a proper business environment. However, in a labor market with high unemployment, the poor human capital endowments of the extremely poor suggest that they will be slow at benefiting from any labor intensive growth. In circumstances of labor market shocks and high unemployment, public works programs could provide a useful short term device to improve the employment among the most vulnerable groups. They could also support the rehabilitation of socio-economic infrastructure in Kosovo. 2.41 Simulating the poverty reducing impact of alternative policies. Using regression analysis it is possible to simulate the potential effect of interventions in the three areas described above.4 The results are reported in Table 2.3. These policies 44 The simulation is based on the regression of per adult-equivalent consumption on a number of household characteristics. Those characteristics include demographic composition, labor force participation and asset holdings such as land, machinery and livestock. 20 represent very different interventions and involve very different costs, so the table also includes gross estimates of their costs and cost-effectiveness measures. These simulations only provide gross estimates on poverty reduction effects based on the observed correlations. The first row reports the simulated effect of increasing land use to a minimum of 2 hectares for all households in rural areas. This corresponds to an increase of about 1 hectare per average household in extreme poverty. About 21,000 rural households are in extreme poverty. The policy would therefore involve a total of 21,000 hectares of land, which is around the amount of land currently left uncultivated by collective farms.45 Of course this is not more than a theoretical simulation as in practice providing access to this land is not straightforward. Issues of ownership and property rights are important constraints and so are geographical considerations, such as the fact that . 2.42 The second row reports the simulated effects of increasing the capital stock available to extremely poor households working in agriculture to the average value of DM 2,255. This represents an overall increase in the capital stock of DM 1,855 The total cost of the policy is about DM 39.4 million. 2.43 The third and final row simulates the poverty alleviation effect of an 8 percent increase in salaried employment of the individuals currently non -participating in the labor force. This hypothesis is consistent with the 10 percent growth estimated by the IMF. The increase in salaried employment is shared equally across rural and urban areas. This policy is more difficult to cost. If the increase was to come exclusively from public works type programs, the wage bill would be around DM 1.2 under the assumption of perfect targeting. Table 2.3. Simulated Poverty Reducton Effects of Alternative Policies Extreme Percentage Estimated Cost per poverty reduction cost percentage headcount point (Percent) (Percent) (DM million) (DM million) 1. Status quo 11.9 2. Increased access and use of land 9.0 24.4 21.26 7.3 3. Agricultural stock reconstruction 11.8 0.8 39.38 48.7 4. Employment growth 10.6 10.9 1.22 0.9 Total 7.6 36.1 22.83 5.3 2.44 The simulation predicts the combined effect of these policies to be a reduction in extreme poverty of 4.4 percentage points (36.7 percent). However, significant differences exist in the effectiveness of the policies analyzed and their cost. Increasing non- 45 The cost of recovering land for agricultural use is estimated as a fixed percentage of its value. The report uses a 3 percent of the value as the amount needed to reuse such land. 21 agriculture employment is the most cost-effective measure. The policy reduces the incidence of extreme poverty by more than 10 percent at a costs of about DM 0.9 million for each one percentage reduction each. Increasing land use has the largest effect but the cost of such policy is relatively high (7.3 millions for percentage point). Moreover, the feasibility of increasing access to land depends on the geographical distribution of the poor and the available land and their access to inputs, finance and crop markets. Finally policies intended to increase machinery in rural areas have a very small effect and are very costly. 22 3. INCOME SOURCES AND SOCIAL ASSISTANCE 3.1 The IMF estimates the GDP of Kosovo at about $805 per capita (equivalent to around DM1,600), a level at the low end of the range of transition economies in Central and Eastern Europe.46 Because of the high level of remittances, and some humanitarian aid, 1IMF estimates show that national income per capita is in the range of $1150 (equivalent to around DM 2150). Since 1999, unemployment has hovered around 50 percent. Given the dislocations in the economy, remittances from overseas, and aid and social assistance programs administered by JIAS and other international agencies have become an increasing component of income. Although growth rates are expected to be significant overt the medium term, unemployment rates could grow further and remittances are likely to decline as those who fled to other countries before and during the conflict return. Analyzing the current composition of income is therefore important to both identify potential sources of income vulnerability and to provide a baseline against which future developments can be monitored or evaluated. A. How MUCH INCOME IS GENERATED? 3.2 The LSMS results suggest the average per capita income of households in Kosovo in 2000 to be DM 1322. Income is received in either cash or kind but the in-kind share very rare and on average they represents a very small part of total income. 3.3 Given the significant difference in the structure of the household between Albanians and Serbs per adult equivalent household income is a better indicator of household welfare than the adult per capita value.47 Average per adult equivalent income in Albanian households is DM 1313, more than double that of Serb households. With a Gini coefficient of 0.58, per adult equivalent income is unequally distributed across households. B. WHAT ARE THE MAIN SOURCES OF INCOME 3.4 The LSMS captures Kosovo at a critical time, after a decade of ethnic discrimination and conflict. The economy was operating at minimum capacity, earnings were limited and households were relying heavily on aid from international donors and 46 IMF (2000) 47 See Appendix Al of Chapter I in this volume for more details on the role of equivalence scales. See Chapter 6 of this volume for more details 23 remittances from abroad. The rural economy was slowly beginning to revive with a spurt in agriculture, although income from land was negligible. 3.5 In urban areas, formal employment was limited after the conflict. Some employment was available through relief and reconstruction programs48. At the same time, there was evidence of a reviving economy with the re-establishment of household enterprises and other businesses. As in the case of agriculture, some household enterprises reported either no income or an outflow of resources, presumably due to the setting up costs of restarting a business after the conflict. Finally across the region high levels of emigration to Western Europe translated into remittances from family members overseas.49 Income transfers from donors within and outside Kosovo also played an essential role in supporting income levels. Figure 3.1: Per equivalent household income by source in male and female headed households 100% 2 80%- a 60% - _ Misc. income o_ Public transfers 40% -lntni. assistance 20% U Household transfers EL 0 Labor eamings 0% Male headed Female headed households households Source: Kosovo LSMS and World Bank calculations. 3.6 Figure 3.1 shows that despite the destruction due to the conflict in the year prior to the survey, wages and eamings from business activities remain the main source of income for the average household. This is particularly evident in male headed households, where the share of income from wages and self employment is roughly 50 percent. Transfers from relatives form the second most important source of income, and are particularly prominent in female headed households and in rural areas. Such transfers are received from relatives residing both within and outside Kosovo. The combination of eamings and household transfers represents more than 80 percent of the income of the average households. The rest is represented predominantly by international assistance. C. INCOME RELATED INEQUALITIES IN SOURCES 3.7 The relative importance of different sources of income varies at different points of the income distribution (Figure 3.2). 48 World Bank (2000) 49 For more details on the role of migration as a coping strategy see Chapter 6 of this volume. 24 Figure 3.2: Per equivalent household income by income decile (All households) 1 00% ? 80% U M Misc, income -~~~~~~~~~~~ U~~~~~~~~~ Public transfers co 3.8 Unsurprisi%gly, the role of eaniings highest for thetop lint,om assistance 40% h Household transfers 0 Labor earnings Q. 20% 0% 1 2 3 4 5 6 7 8 9 10 Income Deciles Source. World Bank estimates from LSMS. 3.8 Unsurprisingly, the role of earings highest for the top income decile while the contribution from public transfers declines considerably as we move from the poorest to richest. More surprising is the inverted U-shaped patter of the share of receipts from both household transfers and international donors. These sources contribute significantly to income in the poorest two deciles. Their role then declines only to increase again for households in the top four income deciles. This implies that the absolute value of these sources increases with income. 3.9 Compared to the Albanians, Serb households receive much less in household transfers, and international assistance is very much concentrated in the bottom income groups. Households in income deciles 2-5 receive over 50 percent of the income from public transfers. 3.10 Table 3.1 shows the Gini coefficients for total household income and from its major components. Unsurprisingly, earnings are the most unequal source of income with a Gini coefficient of 0.76. Household transfers reduce inequality substantially, bringing the Gini coefficient down to 0.63 (a 17 percent reduction). Intemational aid has a smaller redistributive effect for both ethnic groups while public assistance predominantly affects the Serbs. Table 3.1: Inequality in Household Income by Source ofIncome (Gini coefficients) All Serbs Urban Rural areas areas Per equiv. income - earnings 0.76 0.85 0.72 0.78 plus household transfers 0.63 0.81 0.62 0.63 plus international aid 0.60 0.76 0.61 0.59 plus public assistance 0.58 0.56 0.58 0.57 Source: World Bank estimates from LSMS. 25 3.11 Figure 3.3 visualizes the results of table 3.1. The Lorenz curve for earnings is very steep showing a large degree of inequality. Household transfers reduce inequality substantively and shift the Lorenz curve up at low levels of income. The redistributive effect of aid is smaller but significant while public assistance appears to have almost no effect on income inequality for the total population. It has however, an important effect for the Serbs.50 Figure 3.3: Per equivalent household income by source and income decile 2 1.0 0.9 -- D 0.8 - Diagonal o U 0.7 N a ig a D0.6 Erig co2 0.5 ---- plus household 0 0.4 transfers > 0.3 - -plus international aid E 0.21 - plus pub. assistance O 0 1 2 3 4 5 6 7 8 9 10 Cumulatve share of income Source: World Bank estimates from LSMS. D. INDIVIDUAL/HOUSEHOLD INCOME FROM PRIVATE SOURCES IN THE LAST 12 MONTHS 3.12 Households generate income in a variety of ways. The focus here is on three potential sources. - Individual labor income; * Household transfers; and * Income from miscellaneous sources. Employment and Occupation Structure 3.13 Only 55 percent of individuals of working age were actively employed in the year proceeding the survey (Table 3.2).5S While 6.6 percent of individuals were unemployed, as many as 38 percent did not have a job but had not been looking for one in the last 12 months.52 This report refers to them as being 'out-of-the-labor-force'. As expected the percentage of those not actively looking for a job is considerably lower among males (columns 4,5,6). However, the 20 percent average rate suggests a significant 'discouraged worker' effect as individuals stop looking for employment in a very depressed labor 50 See Chapter 2 Volume II. 51 Working age here refers to individuals aged 21 to 65. 52 Unemployed are defined as individuals who do not currently have a job but have actively been looking for one in the last 12 months. Individuals who have not been looking for a job are classified as 'out-of-the- force-force'. 26 market. The apparent difference in unemployment rates between the Albanians and the Serbs may reflect more differences of access to unemployment benefits and discouragement towards job search than differences in labor market conditions per se. After a decade of discrimination in employment opportunities and eligibility for unemployment benefits, it is not surprising that the Albanians feel 'discouraged' in their job search and therefore stop looking before finding a job. Table 3.2: Sector ofmain empment all persons aged 21-65 All Male Albanian Serbs Total Albanian Serbs Total Waged workers 24.9 24.5 24.8 39.9 27.5 38.8 Agriculture 19.2 36.9 20.5 21.0 41.4 22.8 Non agricultural self employment 10.7 1.5 9.8 13.2 2.5 12.3 Unemployed 5.6 14.1 6.6 6.5 13.6 7.2 Not in labor force 39.6 22.9 38.4 19.3 15.0 19.0 Total 100 100 100 100.0 100.0 100.0 Source: World Bank estimates from LSMS. 3.14 Among the Serbs, agriculture is the most common source of employment, and non-agricultural self employment is minimal. In contrast Albanians are most likely to be employed in paid employment followed by agriculture. Their self-employment rates are high reflecting a decade of private entrepreneurship. Household Employment Structure and Poverty 3.15 The employment status of its members has important implications for the level of welfare of the household and for its vulnerability. Households without working members are the poorest, and the risk of poverty declines with the proportion of embers receiving a regular wage. On average 9 percent of the households have no earners and this proportion is higher for the Serbs. Over a quarter of Albanian households had 5 or more earners and over half of them 3 or more. 3.16 The composition of employment varied across income deciles, as shown in Figure 3.4. As expected, the poorest households are characterized by a large average proportion of agricultural workers. As households become richer, the proportion of workers in the formal sector or self employed increases. The poorest and richest households had a larger share of self employed persons as compared to households in the interim income deciles. This reflects the heterogeneity of self-employment and household enterprise in Kosovo. 27 Figure 3.4: Household composition of employment by income decile 100% - t 80%- 60% -OSelf emp. worker X 0% EAgri worker o U Wage earner 5D 40%- co 20% 0% 1 2 3 4 5 6 7 8 9 10 Income Deciles Source: World Bank estimates from LSMS. Why are Earnings so Low in Kosovo? 3.17 The Kosovo LSMS shows that earnings from wages and self employment are very low in 1999-2000, the year preceding the survey.53 This is due to the effect of three main factors: * The post-conflict period is very unusual as it is characterized by low economic activity, and low employment in both urban and rural areas. Formal employment was most affected; * Even when formal employment opportunities were available, earnings were low; and * As the economy was just beginning to recover from the conflict, individuals were beginning to invest in household enterprises and agriculture but were yet to generate earnings. As a result, returns to self employment and agriculture were low or negative. Correlates of Labor Market Outcomes 3.18 The analysis in this report shows that education is the most important correlate of employment status.54 Both the level of education and its specialization are important determinants of employment outcomes. In addition poor health and urban residence have strong negative impacts on employment. 3.19 Estimates of the determinants of earnings fail to show strong correlates. This may be a reflection of the special characteristics which were in operation in 2000 in the Kosovo labor market. The combination of both transition and post-conflict elements 53 It is important to note that information on income from agriculture is not included in the calculation of work related income. 54 See also Chapter 2 in Volume II 28 appear to have weakened the traditional link between human capital and earnings in a country where many of the professionals are still paid only a nominal stipend. E. INTER-HOUSEHOLD REMITTANCES AND OTHER TRANSFERS 3.20 The second major source of income is intra-household transfers. The majority of these transfers went to Albanian households and almost 50 percent of Albanian households in rural areas received them. Among the Albanians, transfers were further targeted to female headed households reaching 62 percent of them as compared to 42 percent of households headed by men. The majority of these transfers came from outside Kosovo, Albania or countries of the former SFRY and their average value was DM 4,983 per year. 3.21 Remittances and transfers to households were also received in the form of food and other goods. However, the occurrence of this type of transfers is less frequent. The mean value of in-kind remittances for the few households receiving them was DM 1473. Transfers from outside Kosovo were DM 1638, double the amount received by households from within Kosovo. 3.22 By way of contrast, only 5 percent of Serb households received transfers from relatives and the average amount received was less than one third of that of the Albanians. F. SOCIAL ASSISTANCE 3.23 Over 8 per cent of households had at least one member eligible to receive benefits from a public agency. For the Albanian population these were primarily social assistance payments from the Center for Social Welfare (CSW). The Serbs also received pensions, unemployment and other benefits from Belgrade. 3.24 Following the Emergency Financial Assistance Program, the new social assistance scheme was introduced in November 2000 with payments to the so-called 'Category One' beneficiaries: families without resources where no one was capable of work. Benefits varied from DM 65 per month for a single person up to a maximum of DM 120 per month for a family of five or more. 3.25 In November 2000 the eligibility for social assistance was extended to 'Category Two' households, with a member who was capable of work but unable to find a job. Because take-up of Category One was higher than expected, eligibility for this second phase was subject to stringent criteria designed to target the most vulnerable. However, at the time the program was designed statistically representative information on vulnerable groups in the post-conflict environment was not available. In addition, the institutional structure to determine eligibility and administer the overall program was minimal. Thus the program had to be simple to administer and the categories easily identifiable. To be classified as vulnerable, a household should fulfill the following conditions: 29 * No member had received payment for any work in the month preceding the application; * Every member capable of working has registered with the Office of Employment; * No resources and in addition the households should have not receiving assistance from other family members working within or outside Kosovo; * Owned - or had access to - less than one-half (0.5 HA) hectare of land; * At least one child under the age of five, or full-time care to an orphan under the age of 15; * No motorized vehicle, including tractors and automobiles; and * Been on the list for distribution of food aid in September 2000. 3.26 Important targeting deficiencies emerge.55 On the one hand only 16.5 per cent of the extremely poor declare to have received social assistance, on the other more than 80 per cent of program resources are granted to households who are not extremely poor and 35 percent of them to the not poor. 3.27 To a large extent, errors of inclusions reflect the fact that eligibility for Category I was determined exclusively on the basis of documentation presented by the applicant. The use of a limited number of easily observable proxy indicators for eligibility for Category II should have represented an improvement but it came to late for the benefit to be reflected in the LSMS results. Moreover, the choice of eligibility criteria for Category II appears to result in rather inefficient targeting. The criteria used by the CSW - and presented on Table 3.3 - include demographic characteristics, assets ownership, participation in food aid programs and labor force variables.56 For each criterion the table provides: D The incidence of extreme poverty for households with that characteristic - (column 2); * The increased risk of being extremely poor for households holding that characteristic as compared to the average household - ie, relative risk - (column 3); * The fraction of all households satisfying that criterion (column 4); and * The fraction of extremely poor households with the characteristic in question (column 5). 3.28 The results in the table suggest three strong proxies for extreme poverty: * Being a rural household with less than half hectares of land; * Not having a car or tractor and; and * Not having any working adults in the household. For more details see Chapter 3 of Vol II. 56 The LSMS labor market data refers to employment in (i) the last seven days and (ii) the last twelve months. In addition there is not information on being registered as available to work. The only information available is on actual job search - ie, having looked for a job in last seven days and twelve months -. 30 3.29 Using these criteria as entitlement categories minimizes the probability of excluding the extremely poor from the benefit (exclusion error). However, column 4 shows that nearly 50 per cent of households do not own a vehicle. This implies that ownership of a motorized vehicle is an inefficient targeting criterion since it applies to a large part of the households, as it has large potential for leakage. On the other hand the proportion of rural households having less that half hectare of land is only 23 percent as compared to 40 percent amongst the extremely poor. This makes this variable a relatively efficient targeting instrument. The equivalent of land ownership in the urban context is employment status. Over the whole population, 20 percent of the households have had nobody working in the last seven days. However the proportion rises to over 40 percent for the extremely poor households. Thus this indicator has the potential of minimizing leakage. 3.30 By contrast, Table 3.3 suggests that the presence of a child under 6 is both a poor and an inefficient proxy for extreme poverty. The incidence of extreme poverty amongst households with young children is only slightly higher than for the other households and the extremely poor are only slightly more likely to have young children than the rest of the population. Table 3.3 Incidence of Extreme Poverty in Different Category II Categories Incidence Fraction Fraction of of Extreme Relative of Households that satisfy each condition ovErty Re lati on the extreme Poverty risk (%/) population po % (%)0 (% poor (%) Possession of less than 0.5 HA'7 21.41 74.30 22.99 40.07 Not Having a Car or Tractor 19.71 73.96 49.00 85.24 Household has at least one child age four or less 12.66 11.74 43.93 49.06 No adulte8 member of the household has worked 22.81 101.32 20.26 40.79 in the last 7 days At least one adult member of the household has 27.18 19.16 25.14 29.95 looked for work in the last 7 days59 Households that meet all of the above criteria 42.10 277.2 1.00 3.78 All population 11.9 Source: Kosovo LSMS and World Bank calculations. 3.31 It is also important to point out that the combination of the criterion above is an effective way of identifying the extremely poor since the small group of households that satisfies all the criterion is nearly three times more likely to be extremely poor than those 57 Excludes Urban Households 58 Aged 15+yrs. 59 All households where at least one adult has worked in the last seven days are excluded from these caluculations. 31 who do not. However, only a very small percentage of the extremely poor (3.8%) satisfy all the conditions described above and would therefore qualify for Category II assistance. In other words Category II benefits suffer from a very large exclusion error. 3.32 The LSMS data also suggests a large degree of exclusion error in Category I benefits that only apply for 13 percent of the extremely poor. This implies that over 80 percent of extremely poor households have characteristics that make them ineligible for either type of benefits. 3.33 Finally it is interesting to note that the potential leakage from the current system is high. Of the households that qualify for Category II benefits nearly 80 percent are not extremely poor, with 34 percent being non poor. The targeting is only marginally better for Category I, which uses as eligibility criteria characteristics held predominantly (58 percent) by the non extremely poor and in over 20 percent of the cases by the non poor. G. ASSISTANCE FROM INTERNATIONAL DONORS 3.34 More than 63 percent of Albanian households and 58 percent of Serb households reported receiving food aid in the 12 months prior to the survey.60 The majority of these receipts were to households in rural areas. While MTS was the major donor, particularly for donations to Albanian households, other international agencies such as the Red Cross, the Orthodox Church and the World Food Program made donations to Serb households. There are also significant differences in amount of food aid received probably as a reflection of different household size. While Albanian households received on average DM 382 worth of food in the last 12 months, Serb households received only DM 176 worth of food aid. 3.35 Over 20 percent of Albanian households received additional non-cash assistance aid in the form of building repair. The average value of this assistance was DM 5825 per household. H. HOUSEHOLD INCOME AND CONSUMPTION COMPARED 3.36 Figure 3.5 compares the distribution of monthly income and consumption across consumption deciles. For the most part, we find that levels of consumption are higher than income levels. This is not contrary to expectations in a post-conflict economy where households consumed unusually high levels of stored food and other accumulated items to compensate for the reduced earning capacity. Spearman rank correlation coefficient of 0.35 between monthly consumption and income shows that income and consumption expenditure were not very highly correlated in the year before the survey. 60 Data on food aid receipts are taken from the consumption module of the Kosovo LSMS because the income modules provide information only for the last 6 months before the survey. 32 Figure 3.5: Average monthly income and consumption by consumption decile 300- p~200- 0 E 100 *U-Perequiv consunpton 0 1 2 3 4 5 6 7 8 9 10 Consumption deciles Source: World Bank estimates from LSMS. I. CONCLUSIONS AND POLICY IMPLICATIONS 3.37 In the post-conflict year 2000 the Kosovo economy was operating at minimum capacity and in achieving their consumption level, households were relying heavily on aid and remittances from abroad. This was particularly true for poor and extremely poor households. 3.38 As Kosovo moves out of the stage of post-conflict crisis into a more sustainable stage of development, the level of aid is declining fast and many predict the value of the remittances to fall progressively over the next few years, as many of those living abroad return home and those that remain loosen their links with the family. Thus households are to become increasing reliant on their own resources to finance their consumption. 3.39 In this context the household capacity to generate income becomes an important indicator of the household future resources and its economic vulnerability. The findings of this study suggest the existence of an important difference between the poor and the extremely poor in this context. The poor households have characteristics that make them likely to benefit from the predicted growth of the years to come so as to be increasingly less reliant on external sources and more able to finance their consumption out of labor income. The characteristics of the extremely poor, on the other hand, suggest that this group of households in the medium-term will continue to rely on external funding for a large part of their consumption. 3.40 As foreign aid declines and remittances become more uncertain, the importance of social assistance in reducing extreme poverty increases. The social assistance program currently operating was designed and launched during a short timeframe to quickly reach some of the most vulnerable groups within society. It has the advantage of being simple to administer and to meet the tight budget requirements. However this study shows that it suffers from important errors both of inclusion and exclusion. According to the LSMS data, more than 80 per cent of program resources are granted to households who are not extremely poor and 35 percent of them to the not poor. Moreover, only a very small 33 percentage of the extremely poor (3.8%) satisfy all the conditions required to qualify for Category II assistance. 3.41 In the light of the increasing importance social assistance will be acquiring in this phase of development, a careful revision of the qualifying criteria for inclusion in the scheme is required. Moreover, administration of the program will be key to meeting social assistance objectives and budget constraints. 3.42 Finally a well-targeted social assistance program, with around 25 percent leakage will cost around 1.7 percent of GDP. This is calculated from the findings of Chapter 2 that an extremely poor individual would need an average DM144 per year to achieve the minimum food requirement. Thus, assuming perfect targeting, the cost of a social assistance program designed to eliminate extreme poverty would have been around 34 millions DM, which corresponds to around 1.2 percent of GDP according to the IMF estimates of GDP per capita (IJS$ 805). With a more realistic assumption of a 25 percent leakage, the cost of social assistance would rise to just over 45 millions DM, around 1.7 percent of GDP. Simulations on proxy-means tests for targeting using data from several Latin American countries shows leakage rates ranging between 22 to 34 percent.61 61 For more detail see Grosh (1994) 34 4. POVERTY AND INEQUALITY IN EDUCATION ATTAINMENTS 4.1 Earlier parts of this report suggest that in Kosovo, as in many other countries, low educational attainment is a strong correlate of extreme poverty. Thus children born into poor households appear to be trapped in a vicious cycle, in which the low incomes of their families translate into poor educational outcomes. This, in turn, has a direct negative impact on the future welfare of their own families. As such, improving education outcomes is an important component of the strategy to eradicate poverty in the long-run. Moreover, education is an important dimension of poverty and has important effects on health. A. EDUCATIONAL STATUS OF THE POPULATION OF Kosovo 4.2 In the absence of indicators of learning outcomes, this report defines educational outcomes in terms of illiteracy rates, educational attainment, and enrolment rates. 4.3 Illiteracy rates. Unsurprisingly, illiteracy appears to be concentrated among the oldest age cohort (over 55 years of age) and falls to below 5 percent for the population aged 45 and under. 62 These trends are representative of both Albanian and Serb groups. However, a sharp increase in illiteracy among the younger cohorts of the 'other ethnic groups' has been experienced over the last decade. Having dropped to single digits for the cohorts aged 26-45, illiteracy among these ethnic groups has risen again very sharply, reaching 22 percent for the cohort aged 16-17 and about 15 percent of the cohort aged 18- 25. 4.4 Educational attainments. Large inequalities in education attainments also emerge among the literate.63 On average, the population of Kosovo have about 10 years of education with almost a two year gap between urban and rural areas. Differences among ethnic groups follow the same pattern as for illiteracy. While the Serb-Albanian difference is small-around 0.5 of a year-the 'other ethnic groups' receive on average two years less education. In addition, gender differences in attainment are large, with gaps for women of more than two years among Albanians in rural areas. Finally, of concern is the decline in educational attainment. Among the non Serbs, the cohort aged 16-25 on average has one year less education than their older counterparts (26-45), possibly reflecting the difficulties in access to and attendance at school during the last decade. 62 Literacy is defined as the ability to read, even with difficulty. 63 Educational attaimnent is defined as the highest grade achieved for adults who have left school. 35 Figure 4.1: Educational Attainment a: RURAL 0. co c?0 6 0.4 0.2 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 years of education -l--- AIbanians Boys Albanians Girls Serbs Boys - Serbs Girls b: URBAN 1 0.8 cy0.6 to 0 E 0.4 0 0. 0.2 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 years of education -4- Albanians Boys --_ Albanians Girls .- Serbs Boys - Serbs Girls Source: World Bank estimates from LSMS. 4.5 Figure 4.1 presents the educational profile of men and women of different ethnic groups in urban areas (Figure 4.1 a) and rural areas (Figure 4. lb). This shows clearly that drop-outs occur overwhelming between primary and secondary school. However, gender and ethnic differences are striking, especially in rural areas. 4.6 Enrolment. In the year 2000 enrolment was nearly universal-around 97 percent-for children of primary school age (6-14). However, the disaggregation in Table 4.1 raises some important concerns. Only 76 percent of the children under 14 from the 'other ethnic groups' were enrolled in school and their net enrolment rates dropped to just 36 50 percent in urban areas. Gender difference within this group were large, with only 69 percent of girls enrolled versus 85 percent of boys. Table 4.1. Net enrolment rates by age groups and gender/ethnicity Male Female Total Albanian 6-14 98.2 96.8 97.5 15-18 74.2 56.2 65.3 Total 6-18 91.0 84.0 87.6 Serb 6-14 100 99.1 99.5 15-18 86.9 92.9 89.9 Total 6-18 95.2 96.9 96.1 Other 6-14 85.4 69.3 76.8 15-18 54.4 40.3 46.4 Total 6-18 77.4 61.1 68.5 Source: World Bank estimates from LSMS. 4.7 Inequality in secondary school enrolment was even more dramatic. Gender inequality in secondary education was of particular concern among Albanians: last year, only just over half of the Albanian girls aged 15-18 were enrolled in school. In addition, net enrolment rates in secondary education for the 'other ethnic groups' dropped to less than 55 percent. Among these groups, girls were particularly disadvantaged having enrolment rates of less than 40 percent. This suggests that the legacy of the last decade on the educational status of the 'other ethnic groups', which is noticeable for the young adults, is continuing to claim its toll on the younger cohorts. B. INCOME RELATED INEQUALITIES IN EDUCATIONAL OUTCOMES 4.8 Although enrolment in primary education is almost universal, large income- related inequalities exist. For example, in urban areas, as many as 20 percent of children from households in the lowest consumption decile were not enrolled in primary school. The great majority of these are children from other ethnic groups - 66 percent. Extremely poor children aged 6-14 from these ethnic groups and living in urban areas have a net enrollment rate below 15 percent. 4.9 Even larger income-related inequalities emerge in net enrolment rates into secondary education. As Table 4.2 shows, youngsters from the poorest decile in rural areas are less than half as likely to enroll in secondary education than their counterparts in the top decile. Although less striking, the difference in urban areas is significant, with only 62 percent of the poorest children enrolling as opposed to nearly 90 percent of those in the top decile.64 Again the ethnic composition of different deciles here plays a major role. While only 4.6 percent of 15-18 years old are from the 'other ethnic groups', these children represent nearly 14 percent of those in the bottom consumption decile. 64 See Chapter 4 Volume II for details. 37 4.10 Table 4.2 shows also a strong correlation between household poverty and educational attainment among the population over 20. An adult in the top expenditure decile has on average 2.5 years more education than one in the bottom expenditure group. One possible explanation for the correlation between poverty and educational attainment is provided by the lower employment opportunities open to individuals with lower educational attainment as described in Chapter 3 above. Table 4.2: Educational Enrollment and Attainment Enrolment rates in secondary Highest grade achieved Expenditure education decile Cohort 15-18 Cohort 20 and above Rural Urban Rural Urban 1 35.9 62.1 8 9.6 2 53.6 53.5 8.5 9.7 3 49.0 73.9 8.8 9.9 4 60.5 78.6 9.3 10.6 5 59.1 84.3 9.1 10.4 6 63.6 80.9 9.4 10.8 7 61.0 79.1 9.1 11.2 8 61.4 84.2 9.3 11.1 9 72.0 88.3 9.3 11.6 10 75.8 89.2 10 12.1 Total 59.3 78.1 9.1 10.8 Source: World Bank estimates from LSMS. C. ANALYZING EXPENDITURE 4.11 During the last academic year, spending for education was financed through a combination of public financing-largely contributed by the international community- and private out-of-pocket contributions. The relative importance of the two components varied across different levels of education. The LSMS data indicate that in the year 2000 household cash contributions were just over 43 percent of total expenditure for primary education, but may have exceeded 60 percent for both secondary and tertiary education.65 Public Expenditure on Education 4.12 Table 4.3 shows that public expenditure favors the better off. This is due to a number of factors including: * a strong inverse correlation between poverty and utilization of education services, particularly in secondary education; and * the unequal distribution of public expenditures across municipalities. 65 It is worth mentioning that this expenditure includes the cost of travel and food as well as fees, books, etc. 38 4.13 With a concentration index of -0.024, expenditure on primary education is slightly pro-poor due to: * the concentration of primary school children in the relatively poor households; * the slightly lower enrolment rates of those children; and * the little variation in unit costs across municipalities. 4.14 By contrast a concentration index of 0.12 shows expenditures in secondary education to be pro-rich. This reflects higher enrolment rates among better off children as well as larger differences in expenditure per pupil (that reflect differences in pupil- teacher ratio) across municipalities. The pro-rich character increases further when higher education is considered, with the concentration index reaching 0.37. Table 4.3: Shares ofpublic expenditure in education by consumption decile and type of school (percent) Level of Share of expenditure received by each consumption decile Total (DM) Quasi Gini education 1 2 3 4 5 6 7 8 9 10 Primary .6 10.1 10.9 10.9 10.5 10.7 9.9 8.9 9.5 8.9 71,318 - 0.024 Secondary 5.6 7.9 7.7 10.4 11.2 11.2 8.3 10.2 13.9 13.7 22,592 0.120 Higher .5 3.4 3.6 6.4 6.9 10.1 12.4 13.4 15.3 26.2 12,322 0.367 Total .9 8.9 9.4 10.3 10.2 10.7 9.8 9.7 11.1 111.9 106,232 0.052 Source: World Bank estimates from LSMS. 4.15 The inequalities in public expenditure at different levels of education are summarized in Figure 4.2. The concentration curve for primary education lies above the diagonal showing its slight pro-poor effect. The position of the concentration curve for expenditure on secondary and higher education shows them to be pro-rich. Figure 4.2: Public expenditures in education favors the rich 100 51 80 */ 70 70 ~~~~ 60 ~~~~~~~~~Diagonal 50* ____________ ~~~~Primary ----Secondary 40 - , 30 1 -~~~~~~~~~~~~~~~----igher 30~~- 20 - . 10 0 0 1 2 3 4 5 6 7 8 9 10 Cumulatie share of consumption Source: World Bank estimates from LSMS. 39 Household Expenditure on Education 4.16 The LSMS data shows that nearly all households who had children enrolled in the last academic year (99 percent) paid some out-of-pocket contribution to school expenses.66 On average 6 percent of private expenditure covered fees and another additional 12 percent covered tuition costs. Total books and equipment accounted for 11 percent, while 'in-kind' and 'other expenses' accounted for over 25 percent. Moreover, the typical-i.e., median-expenditure per child was reported to be considerable, averaging nearly DM 200 for primary education, just over DM 400 for secondary education and about DM 280 for higher education. 4.17 Figure 4.3 shows that private expenditure for education increases with income in a progressive manner. This pattern is in accordance with expectations and with the experience of other countries. What is somewhat surprising is that private expenditure is comparatively high even for households in the bottom deciles. Table 4.4 suggests that nearly 5 percent of the total private expenditure in education comes from the bottom income decile. Considering the high level of private expenditure in education this implies that the worse-off in Kosovo spent a large percentage of their income on education with important negative implications for their welfare. Figure 4.3: Private expenditures E 100 90 90 80 - > 70 Cnupi Diagonal 'C 60 Pa 0.7 Primary Seco-dar 1.7 ----5.2 6.0 8.1 8.6 8.6 14.8 14.5 28.7 421,60Secondary 40 - - Total 3.1 4.5 5.0 309 7.4 11.3 10.8 -11.5 13.8 24.6 1,444,647gher - ~ ~ ~ - - > 20 , 6 s 0 0 1 2 3 4 5 6 7 8 9 10 Cumulatiwe share of cons umption Source: World Bank estimatesfirom LSMS. Table 4.4: Shares of private expenditure in education by consumption decile and type of school Level of Consumption decile Iotal (DM) Education 1 2 3 4 5 6 7 8 9 10 Primary 4.7 6.6 7.0 11.1 9.5 13.6 12.4 8.8 13.1 13.3 668,654 Secondary 1.7 3.9 5.2 6.0 8.1 8.6 8.6 14.8 14.5 28.7 421,606 Higher 1.7 1.2 1.2 4.1 2.3 10.4 10.3 12.7 14.4 41.7 354,387 Total 3.1 4.5 5.0 7.9 7.4 11.3 10.8 11.5 13.8 24.6 1,444,647 Source: World Bank estimates from LSMS. 66 See previous footnote. 40 D. IDENTIFYING CAUSAL FACTORS 4.18 The LSMS identifies the most important reasons for not enrolling in school. The results point to economic factors as the main rationale for non-enrolment. 7 4.19 Out-of-pocket payments as the main barrier to entry. Over the entire population, the private cost of education was identified as the main reason for non enrolment (37 percent) with an additional 6.5 percent pointing to the need to work as the main factor. This is not surprising when considering the magnitude of the out-of-pocket payment faced by the average household. The findings show that nearly all households who had children enrolled in the last academic year (99 percent) paid considerable out- of-pocket contribution to school expenses. Geographical and ethnic differences are also apparent, with expenditure being higher in urban areas particularly for the Serbs who face nearly twice the median household expenditure. 4.20 Other barriers to entry. Safety was also identified as an important barrier to entry especially by the 'other ethnic groups'. Other reasons related to access to schools are significant especially in rural areas, with distance to school cited by 5.4 percent of the sampled population.68 4.21 Identifying the correlates of enrolment in secondary school. Box 4.1 identifies the main correlates of secondary education enrolment, for non-Serb girls, the group at higher risk of non-enrolment. Box 4.1: Determinants of enrolment in secondary education for non -Serb girls. The probability of enrolment increases with: * Educational level of father and mother; * Residence in urban area; * Household income. The probability of enrolment decreases with: * Household size; * Belonging to other ethnic groups'. E. CONCLUSIONS 4.22 Education policy makers in Kosovo need to focus on these findings to improve educational attainment and break the vicious cycle that links low educational outcomes and poverty: Primary school enrolment is very low amongst the 'other ethnic groups' especially in urban areas; 67 These factors, however, appear to be less important for the Serbs. 68 See Chapter 4 of Volume II for more detailed information. 41 * Enrolment in secondary education is low and significant inequalities emerge - with the poor, girls, and 'other ethnic groups' emerging as the most disadvantaged groups; * In rural areas, only about a third of the 15-18 year olds from the poorest decile are enrolled in secondary education, as opposed to over two thirds of those from the most affluent decile. The results are only slightly less unequal in urban areas. Over 60 percent of the poorest youth in the 15-18 age group are enrolled in secondary education, versus over 90 percent of the same age group from the most affluent decile; and * The main barrier to access seems to be the costs of education. This reflects the large cash and in-kind contributions paid by households of all income groups. The second most common cause of dropping out of school is lack of security. This is particularly important for the 'other ethnic groups'. 4.23 These findings show the importance of: * Ensuring that poor families can afford schooling; and * Promoting enrolment in secondary education, and increasingly in primary, of girls and some minority groups. Ensure that Poor Families Can Afford Schooling 4.24 In Kosovo education is provided almost entirely by the public system. Direct parental contributions to schools are not formally compulsory, although they are pervasive (99 percent of the households) and are considerably large across the income distribution. On average 6 percent of private expenditures on education covers fees and another 12 percent covers tuition costs. Total books and equipment accounted for 11 percent of household expenditure on education, while 'in-kind' and 'other expenses' accounted for over 25 per cent. JIAS has succeeded over the past two years in making some textbooks and education materials available to schools free of charge for core subjects, but many key texts and materials currently required by the system are often only available on the informal market or sometimes not available at all. Donor funding is now being used to extend the availability of books-and to streamline the curriculum to reduce the number of required texts. JIAS should strongly consider targeting its limited funding on these inputs to poor households to increase their enrolments in primary and especially in secondary education, where out-of-pocket contributions are highest. At the same time the charging of fees in the public sector should be prosecuted and the development of a private fee-paying system should be regulated and encouraged. 42 4.25 A common intervention in low income countries is to distribute family assistance programs through the school system, making them conditional on continued school attendance. This option - generally referred to as 'conditional transfer programs' - should be considered as an alternative to the current unconditional social assistance and as such it is discussed in chapter 6 which follows. A description of the international experiences with this type of programs is presented in Box 4.2. Box 4.2: Conditional Transfer Programs: experiences and requirements Shaping social assistance as a conditional household transfers to address low coverage of social services among children is growing in Bank client countries. These programs provide a cash transfer in exchange for childrens' attendance at school or their receiving health check- ups thereby initiating a "price" for social assistance in return for a social benefit. Two examples of these type of transfer are the PROGRESA program in Mexico and the Bolsa Escolha in Brazil; a recent similar program is under implication in Turkey. Both LAC programs use the same mechanism: household level cash transfers conditional on the verification of school attendance by the children in Brazil and Mexico and regular health care screening in Mexico. Both have been shown to have been successful in improving the targeting of social assistance, and in increasing educational and health outcomes among children. These programs, however, have some requirements. Most importantly, they intensively use the existing network of public social services, such as health care centers and schools. Coverage in communities without these networks or for groups with limited access to those services is much more difficult to guarantee. In the Kosovo context one such group could be the Roma communities, which appear to be the most needy and are also largely marginalized. Thus more imaginative ways of reaching these groups needs to be identified. Second, in order to gain large impacts on the beneficiary population, some programs are targeted to pre- selected groups, such as literate populations. While this pre-selection exploit community level spillovers, it also excludes others from participating. Thus the challenge is to identify the groups to target and design programs accordingly. 4.26 Making education more affordable will go a long way towards improving educational participation but is unlikely to eliminate the barriers to access which are currently faced by girls, particularly in rural areas, and by the children from 'other ethnics groups'. Given past achievements of nearly complete primary enrolment in Kosovo and the strong tradition of compulsory primary education, it is expected that a return to universal primary enrolment in Kosovo may not require highly costly or elaborate interventions. However, given the continued ethnic tensions, security both inside and outside the school may continue to represent an important constraint on enrolment. Improving security outside the school is an issue that goes far beyond the responsibility of the educational system and is, for obvious reasons, a high priority for JIAS and the entire international community. Until some form of multi-ethnic society is achieved, in troubled areas, the focus should be put on working with local police, UNMIK police, and KFOR to ensure safe transit to school and a safer environment inside the school. Particular attention should be given to the very vulnerable position of the 'other ethnicities'. 43 Promote the Education of Specific Groups 4.27 Increasing the enrolment of girls and 'other ethnic groups' in secondary education may require focused and targeted interventions. How to best target these groups is a challenging question. Research from other countries shows that the direct costs of schooling pose more of a barrier for these vulnerable groups. Although evidence for Kosovo on this subject is not available, targeted interventions to reduce or eliminate direct costs are likely to have positive effects on the enrolment of these groups. Indirect or opportunity costs are also traditionally important deterninants of educational attainment in low income countries, but are thought to be somewhat less significant in Kosovo. A potential policy lever in this context is to target these vulnerable groups through an attendance conditional benefit of the type described above. Alternatively, outreach programs to encourage parents to send their children to secondary school should complement other efforts. 44 5. THE HEALTH STATUS AND ACCESS TO HEALTH CARE 5.1 Kosovo's health sector has crumbled under the combined effects of neglect by Belgrade, ethnic disagreements, and chaos after the break up of the former Yugoslavia. The legacy of these difficulties has left a population with inadequate health coverage. This, combined with the difficulties of the past decade, has resulted in deteriorated health status, limited access to care for some populations, and a system that is heavily financed through out-of- pocket payments. The concern is that the negative effect of this status quo has fallen disproportionately on the most vulnerable groups in society. However, the data available are limited and as such do not allow a full analysis of this claim. A. HEALTH STATUS OF THE POPULATION 5.2 Information on health status indicators in post-conflict Kosovo is limited. However, available data suggest that, even before the crisis, Kosovo had the poorest statistics in Europe on virtually every health indicator. The findings of a national survey conducted by the International Organization for Migration (IOM) and the UN Population Fund (UNFPA) between November 1999 and February 2000 confirm that Kosovo's relative position has persisted. 5.3 Kosovo reports the highest infant mortality in Europe at 45 deaths per 1,000 live births with about 36,000 births per year. A high 40 percent of in-hospital mortality was due to infant deaths, many of them in the neonatal period. Kosovo also exhibits one of the lowest percentages of ante-natal care in pregnant women, and in births attended by a professional. In 1999, 15 percent of pregnant women did not see a health care worker and 20 percent gave birth at home without professional help. Another area of concern is the reported decline in immunization coverage. However no information on income-related inequalities in health outcomes is available.69 5.4 Recent epidemiological studies commissioned by UNMIK offer a health profile highlighting health problems and incidence of illness and injury. Despite no recent outbreaks, tuberculosis and Hepatitis A are endemic in Kosovo. Chronic adult morbidity is concentrated in cardiovascular, renal and lung diseases. The trauma associated with the recent conflict places special needs on the health system 2.5 percent of the population is 69 In designing a manageable LSMS questionnaire a separate 'Female Questionnaire' was not included as information of the type described above had just been collected by IOM and UNFPA. 45 estimated to have some form of physical disability. At the same time, the incidence of psycho-social problems are of considerable concern. A recent survey indicates that 62 percent have been close to death, 49 percent have been victims of torture or abuse, and 42 percent have been separated from their families (JAMA, 2000 284-569-77).. This requires both rebuilding the system but also ensuring that appropriate and adequate care are accessible. 5.5 In addition, there is widespread concern that the conflict may have worsened the already precarious situation of safe drinking water and sewage disposal, and concern is heightened with endemic, waste-borne Hepatitis A. The problem is at least partly linked to chemical contamination in industrial and mining areas. 5.6 Evidence on perceptions of health status is also instructive. The LSMS collects individual level information on three indicators of adult health status. The first is a 'self- assessed' health indicator that takes five possible values. The second is a 'yes/no' variable referring to the loss of at least one day of primary daily activities. The third is an indicator defined by a small set of variables assessing the existence of a disability. The results-summarized in Table 5. 1-suggest that the health status of the population varies significantly across ethnic groups, gender and age groups. 'Limiting illness' refer to the percentage of the sampled population that reported at least one day of primary daily activities missed due to poor health in the past four weeks. The 'bad health' category includes the percentage of surveyed people who rated their health status as 'poor' or 'very poor' in the past four weeks. 'Disability' refers to the sampled population who reported difficulties in performing activities of daily living (ADL) in the past year.70 Table 5.1: Percentage ofpopulation reporting poor health (past four weeks, unless otherwise indicated) Limiting illness Bad health Disability * Albanian Serbs Albanian Serbs Albanian Serbs Male 16-24 7.4 2.2 1.9 5.0 1.2 0.0 25-49 10.9 15.3 4.7 17.6 3.1 4.4 50-69 26.7 32.0 21.1 52.0 18.1 22.6 70 and over 46.3 70.0 53.9 80.5 54.1 60.2 Total male 14.6 22.1 9.3 29.9 6.1 5.4 Female 16-24 8.0 8.3 0.9 8.3 0.3 1.4 25-49 12.7 13.2 5.3 22.6 3.7 3.1 50-69 28.5 49.7 28.4 67.4 24.6 32.2 70 and over 49.8 65.6 59.7 70.7 60.4 61.8 Total female 16.4 28.8 11.5 38.3 4.3 4.2 Total (M,F) 15.5 25.4 10.4 34.1 8.9 15.6 * past 12 months. Source: World Bank estimatesfrom LSMS. 70 These activities include the ability to dress without help, stand up from a sitting position, and use the bathroom without help. 46 5.7 Table 5.1 shows clear ethnic differences in the reported health status. The differential persists across the different indicators of health status, with Serbs reporting systematically worse health outcomes than Albanians. The percentage of Serbs rating their health status as 'poor' and 'very poor' is over three times that of Albanians. Similarly, the percentage of Serbs reporting limiting illnesses and disability is almost twice as large as that of the Albanians'. While the difference in the average values reflects the different age structure of the two populations, differences persist also within each age group. This may reflect a selection process that have resulted in the healthiest Serbs leaving Kosovo. 5.8 A systematic gender bias is also noticeable: the percentage of women reporting poor health outcomes is typically greater than men's with the exception being the 'disability' indicator. Differences across age groups are also interesting, especially when considering ethnicity and gender differentials. As would be expected, the percentage of people reporting poor health outcomes increases with age, however, the pace of such increases varies according to gender and ethnicity, with Serbs and women having fewer complaints. Correlates of Poor Health Outcome 5.9 A primary set of health status determinants relates to access to water and sanitation and refuse collection. A clear ethnic difference in the pattern of causality emerges.71 For Albanians, trucked-in water is by far the preferable means of water distribution from a health stand point and public taps appear to have a negative effect on their health status. On the other hand, Serbs drinking trucked-in water have a higher incidence of poor health. Similarly, the best option for garbage disposal for the Albanians is having street containers but for the Serbs it is collection by trucks. Consistent results across ethnic groups are shown only on the link between sanitary systems and health status. Health status of both Albanians and Serbs are negatively affected by the use of latrines-especially when shared-and flush toilets provide a preferable solution. 5.10 These results provide an indication of the general public health environment confronting the population of Kosovo, as well as the thorny institutional and ethnic issues that undermine effectiveness in these areas of health prevention. However, the details are still blurred. Why Albanians and Serbs fare so differently under alternative water supplies and garbage collection arrangements is not at all clear. The geographic divisions of ethnic neighborhoods and possible differences in the quality of tap water across the region obviously contribute to the discrepancies in experience. However, institutional factors, such as public sector management, infrastructure maintenance and service reliability, are also likely to play an important role. Further analysis is inhibited by the lack of information. 71 See Table 7 of Chapter 5 in Volume II. 47 5.11 Box 5.1 highlights the other main causal factors behind poor health. Box 5.1: Determinants ofpoor health outcome * Apart from the differences highlighted in the text in the role of water and sanitation, the correlates are siniilar for different ethnic groups but gender differences exists. • The probability of reporting 'bad health' increases in age and declines with o Individual education; and o Household size. * For women living in urban areas reduces the incidence of 'poor health' while living in a female headed household increases. * The correlates of the incidence of limiting illness in the last four weeks are similar to those for bad health but the coefficient are less significant and more of an ethnic difference emerges. Source: World Bank estimates with LSMS data. For details see Chapter 5 Vol 11. Income Related Inequalities in Health Outcome 5.12 Figure 5.1 presents the incidence of poor health by expenditure deciles for two of the indicators available from the LSMS. The pattern that emerges is one common to many countries. Ill health appears to be concentrated among the worse off households and this is particularly true for the self-assessed health indicator. However the degree of inequity is surprising low. It is also interesting to note that no clear pattern exists in the distribution of disabilities across income deciles, possibly reflecting the recent conflict in Kosovo. Figure 5.1: Health staus by income 0.9 4 *~0.8 1 _0.7 -J ~0.6 -_ _ __ _ a + diagonal ' 0.5 -J- SAH poor 0.4ulativeLimiting Illness 0.4 - ~ ~ ~ ~ ~ 4 0.2 0.1 0 0 1 2 3 4 5 6 7 8 9 10 Cumulative percentage of consumption Source: World Bank estimates from LSMS. 48 B. HEALTH CARE PROVISION AND UTILIZATION 5.13 The health care system is a patchwork of the previous parallel systems that have survived the conflict and continue to provide services. As in the rest of the region, health care is hospital- and specialist-based. Currently, there are six hospitals in Kosovo with 4,856 beds, and 280 primary care facilities. The average length of stay in the hospital is 10.6 days and bed occupancy is 68 percent. Health sector employment financed by the Kosovo Consolidated Budget is estimated at around for 13,000 in 2001, well in excess of the 10,582 forecast in the budget. In comparison with the much wealthier OECD countries, these estimates reflect an over-supply of beds, low hospital utilization, and an over-supply of staff. Primary care is provided through a network of clinics defined as health houses and 'ambulanta'. Major towns have a single health house, while 'ambulanta' offer less extensive outpatient care but are more widely available. A private sector is beginning to emerge, concentrating in ambulatory care, diagnostic services, pharmaceuticals and dental care.72 5.14 Table 5.2 shows the utilization rates by type of health care facility and ethnic group for the sampled population according to gender and self-reported health status. In addition, the average number of visits and duration of inpatient stays are also indicated. Table 5.2: Utilization of Health Care Service (past four weeks, unless otherwise indicated) Ambulanta and Private Doctor Private Nurse Other Facility Hospital* Health Houses Albanians Serbs Albanians Serbs Albanians Serbs Albanians Serbs Albanians Serbs Utilization Rate (percent) Good health status 10.9 8.6 2.5 0.8 1.0 0.1 2.0 1.2 4.5 2.7 Poor health status 50.8 47.3 12.5 1.2 5.5 0.2 13.8 8.1 15.8 17.3 Men 12.9 17.8 2.8 0.3 0.9 0.1 2.1 3.0 3.7 5.8 Women 16.9 24.3 3.6 1.3 1.7 0.3 3.2 3.1 5.5 8.3 Avg. no. of visits 2.0 3.2 1.7 3.5 4.8 3.0 1.4 2.2 1.2 2.2 Avg. days of stay - - - - - - - - 14.2 30.3 * past 12 months Source: World Bank estimatesfrom LSMS. Patterns of Health Care Utilization 5.15 Table 5.2 shows clearly that 'ambulanta and health houses' are by far the most commonly utilized health care facilities by all ethnic groups and across genders. The patterns of utilization are also similar across ethnic groups, although Albanians show a higher tendency to use private facilities than the Serbs. This is clearly a legacy of the past and a reflection of developments over the last decade. 72 Institute of Health Sector Development (2001) report to the Department of Health and Social Welfare of UNMIK, Kosovo, Education and Health project, health component B-03/00, status report no 1. 49 5.16 On average, the Serbs appear to make greater use of the health system. This may be due to a combination of their age structure and comparatively poorer health status. However, when the utilization patterns for healthy and less healthy individuals are compared, it is the Albanians who have relatively higher utilization rates. 5.17 Hospitals are highly utilized with 16 percent of Albanians and 17 percent of Serbs with poor health status and over 5 and 8 percent of Albanian and Serb women, respectively, having at least one hospital visit in the past year. The average number of visits and the length of stay of Albanians are almost half those for the Serb samples. However, much of this difference is explainable by the comparatively worse health status of the Serbs. The high use of hospital care across both income groups probably reflects historical reliance on the specialist-based and hospital-centered model of care that prevailed in the SFRY. This is coupled with a lack of trust in the developing primary health care sector. As a result, many hospitals are providing both primary care services as well as inpatient care. Despite this, as discussed above, the average length of stay in the hospital is long by international standards and the bed occupancy rate comparatively low, suggesting extensive excess capacity in the hospital sector. 5.18 Finally, hospital care is provided almost exclusively by the public system, with private humanitarian hospitals providing a mere 2 percent of the care for the Albanian and 4 percent of that of the Serbs. Barriers to Access 5.19 On average around 9 percent of the adult population reported having had a medical problem in the four weeks prior to the survey but have not sought medical care. Excluding minor ailments for which self treatment was considered sufficient, the most common barrier to health care access is the cost of the service. An average 28 percent of those that reported having had a medical problem for which they did not seek treatment, did so because they could not afford health care (Table 5.3). This percentage increases to nearly one third for the prime age cohort (age 25-49). 5.20 A detailed analysis of income related inequalities in the extent - and type - of barrier to access to the health service is prevented by the limited number of people who where in need of, but did not seek, health care in the four weeks prior to the survey. These individuals total 1,600 of which 250 were extremely poor. However, the finding suggest the proportion not seeking medical care for a declared health problem to be slightly higher for the bottom consumption decile (13 percent) but the extent of income- related inequality is limited. Unsurprisingly financial consideration seem to have been more binding for the worse-off. Over 40 percent of the individuals in the bottom two consumption deciles who did not seek medical care when in need, were prevented from doing so by financial consideration. The corresponding figure for the top two consumption decile was around 10 percent. 50 Table 5.3: Main Barriers to Access: Reasonsfor not Seeking Treatment Minor Minor Too Unsafe to Poor Too Age ailment, ailment,no Other Total self aiment far travel service expensOve tetdtreatment treated 15 < 55.2 14.1 5.3 0.9 0.8 21.1 2.7 20.9 16-24 53.3 16.6 2.2 1.1 0.0 26.6 0.3 11.4 25-49 44.6 12.8 5.7 1.2 1.0 30.1 4.6 28.5 50-69 41.5 11.0 9.2 2.3 0.5 31.1 4.3 28.1 70 + 39.3 8.0 14.9 2.8 0.5 28.6 6.0 11.1 Average 46.3 12.5 7.2 1.6 0.7 27.9 3.8 100.0 Source: World Bank estimates from LSMS. 5.21 The role of affordability as a barrier to entry comes as no surprise when considering the considerable out-of-pocket payments. Over 95 percent of Albanians pay for the services received, independently of whether these are provided in public or private facilities.73 Moreover, the average payment is in excess of the expected amount of the co-payment. For example - given the prescribed contribution of DM 1-2 for primary care visit, and the average number of visits to primary care units over the last four weeks, which equals 2 - the average cost per individual should be around DM 4. However, the average individual appears to pay close to DM 6 for general expenses and an added DM 11.3 in what are described as 'gifts'. It is not possible to distinguish between gratitude and required gratuity, something that qualitative research can provide, and there is extensive experience in the region on these issues.74 The Serbs appear to have easier access to free health care and tend to face lower costs. It is also interesting to note that by far the most expensive item of expenditure is pharmaceuticals, followed for Albanian patients, by gifts, and general expenses for Serb patients. 5.22 Another important barrier to access, especially in rural areas, is the distance of the facility. Safety, however, does not appear to pose a major impediment. Not surprisingly, the cohort of people most affected by the distance and safety factors are senior citizens, especially those over 70. Lastly, values associated with the perception of quality of the service provided are very low but it is not clear whether medical quality is compromised or whether patients are frustrated with access to and treatment within the system. However, this perception combined with the cost barTier suggest inherent problems in the operation and accountability in the health system. 73 See Chapter 4, Volume II for more details. 74 Lewis (2000) 51 C. HEALTH CARE FINANCING 5.23 In 2000 the financing of health care in Kosovo continued to have the characteristics of a mixed system where out-of-pocket contributions were considerable and public financing came predominantly from donors, with minor contribution from the general tax system. The element of social insurance which was characteristics of the pre- conflict system was lost and no private insurance has filled the gap.75 Private Expenditure in Health 5.24 Private expenditure on health is of three types: * payment for services provided by the private sector, mainly doctors, dentists, diagnostics and drugs; * a co-payment introduced in the summer of 2000 for pharmaceuticals and treatment received in the public sector; and * unofficial compensations to medical staff operating in the public sector. 5.25 Little is known about the relative importance of these components in total. Indeed until the LSMS data were available little was know about the magnitude of the private contributions to the health care system. Only the Kosovar Albanian Health Survey Report gave some estimates of private expenditure on health for the Albanian population in 1999. 76 5.26 The findings of the LSMS suggest that over 95 percent of Albanians pay for the services received, independently of whether these are provided in public or private facilities77. Moreover, the average payment is in excess of the expected amount of the co-payment 5.27 Figure 5.2 shows the distribution of out-of-pocket health expenditure across income deciles. Unsurprisingly, the better-off appear to spend more on health care than households at the bottom end of the distribution. What is noticeable, however, is the very low degree of progressivity of expenditure. This suggests that even the worst-off face large out-of-pocket contributions. Indeed the expenditure for pharmaceuticals, which constitutes a large proportion of total expenditure, is very close to being proportional to income. Given the high costs that the poor face, it is not surprising that financial considerations are described as the most common barrier to access for those who, in the four weeks before the survey, experienced a health problem that would have required medical attention. 75 The public allocation for health care is described in Chapter 5 Volume II. 76 International Rescue Committee, Institute for Public Health of Kosovo, WHO and US Centers for Disease Control (1999) 77 See Chapter 5 Volume II for details. 78 For more details on the co-payment system see Chapter 5 Volume II. 52 Figure 5.2: Private Expenditure on Health by Income Decile 0.9 e 0.8 u0.7- 0.6 ----diagonal @ 0.5 ----- _//primary care -&- Pharmaceuticals 0.4 -Bt- Hospital care S0.3- 2 if 0.2- 0.1 0 1 2 3 4 5 6 7 8 9 10 Cumulative ahare of consumption Source: World Bank estimates from LSMS. D. POLIcy LEVERS AND POLIcy DESIGN 5.28 Both health status and the state of the health care system are in need of upgrading. The results of the past decade have taken a toll on the health status of the population, and the health care system has provided minimal support due to a combination of historical treatment by Kosovo in SFRY and the chaotic system of the recent past. The neglect of the health system has led to a multi-funded health system and comparatively poor health indicators. Of particular concern are: - the high infant mortality * the low percentages of ante-natal care in pregnant women, and the small proportion of births attended by a professional; * the reported decline in immunization coverage 5.29 While lack of data prevent an accurate analysis of income-related inequalities in health outcomes, the expectation is that the income poor carry much of the burden of poor health. This is confirmed by the analysis of self-reported illness amongst adults, which shows poor health to be somewhat more common at the lower part of the consumption distribution. 5.30 The poor health outcomes appear to be related to the quality of health care received rather than to barriers to access. The percentage of the population that reports having had a medical problem in the four weeks prior to the survey without seeking medical care is comparatively small and does not vary substantially across the income distribution. This suggests the need for upgrading - rather than expanding - services in all their aspects: infrastructure, human resources, equipment and management. 53 5.31 Excluding minor ailments for which self-treatment was considered sufficient, the most common barrier to health care access was the cost of the service. An average of 28 percent of those that reported having had a medical problem for which they did not seek treatment, did so because they could not afford health care. The financial consideration was particularly binding for the worse-off individuals. Over 40 percent of the individuals in the bottom two consumption deciles who did not seek medical treatment when in need, were prevented from doing so by its cost. Developing a Blue Print for Improving Health Outcomes 5.32 The findings of this study suggest that, without radical efforts to restructure the health system neither the poor nor the general population will be adequately served. The following priorities are recommended: * Produce a blueprint for prioritizing across health investments given limited resources; * Determine efficiency in provision and produce a plan to downsize where possible to concentrate on priority services; * Align realistic financing levels with commensurate care, upgrading services in all their aspects and increase efficiency of health care provision; D Formalize under-the-table payments via a system of means-tested co-payments and establishing realistic levels of health service output; * Focusing on child and maternal health; and * Undertake efforts in public health to ensure immunization coverage, adequate access to clean water, reduced environmental health risks. 5.33 Adopting some of these measures may result in services that the public must withdraw from providing, or means testing access to ensure equity in health care. Access may need to be compromised given the current costly overcapacity of both physical and human infrastructure. Formalizing and legalizing under-the table-payments could help to reduce the corruption in the system, and ensure that the public sector finance high priority interventions as well as subsidizing those unable to afford needed care. It is also designed to reduce the existing financial barriers to access. Indeed, the high incidence of out-of- pocket payments commonly leads to under-utilization by those least able to pay and a system low on accountability.79 79 See Lewis (2000) 54 6. TOWARDS BUILDING A STRATEGY FOR POVERTY ALLEVIATION 6.1 At the time the LSMS was conducted the economy was starting to recover after a decade of operating at minimum capacity and the aftermaths of the recent conflict. Individuals were just beginning to invest in agriculture and in other small enterprises but earnings were still negligible. Formal employment in other sectors remained limited and paid very low wages. This resulted in: * widespread poverty, with over 50 percent of the population having inadequate consumption levels; and * a large reliance on remittances from abroad and aid from international donors. A. THE POOR VERSUS THE EXTREMELY POOR. 6.2 A close look at the distribution of consumption reveals a significant cluster of households around the poverty line, with a small hardcore of extremely poor households and an equally small number of very rich ones. The implication of this conclusion is that in the Kosovo context, the distinction between extremely poor and poor household is more meaningful than the rather arbitrary split between poor and non-poor. This conclusion is confirmed by the clear difference in the characteristics of the extremely poor and the poor households. 6.3 The households who were poor in 2000-but not extremely poor-have a stock of assets and a level of financial and human capital that allows them to take full advantage of any future growth opportunities. In rural areas, they are self-employed farmers with some assets and comparatively large holdings. In urban areas, they exhibit low dependency rates, are relatively young and have a head with a good education level, but who is currently unemployed or out-of-the-labor force. It is therefore justifiable to think of them as 'post-conflict' poor that have been brought into poverty by the economic decline of the last decade and/or by the recent conflict. The characteristics of the extremely poor, on the other hand, are worrisome. These 'hardcore' poor exhibit a number of characteristics which make them unlikely to benefit fully from any future economic growth. They tend to live in households with high dependency ratios and a poorly educated head. If in agriculture, they have little land and no machinery and, if Serbs, they are predominantly elderly. 6.4 The extremely poor are also comparatively worse off in terms of other dimensions of poverty. Their sources of income are more volatile than the rest of the population, as 55 they depend almost entirely on a combination of household transfers and aid from international donors. They have a higher incidence of poor health status and disabilities. Their children are twice less likely to enter secondary school than children from households with an average level of consumption. By contrast the 'poor'-but not extremely poor-are much closer to the rest of the population in all these dimensions. 6.5 These findings suggest the need for a two-prong strategy for poverty reduction in Kosovo today. For the poor, broad-based economic growth and improved access to good quality social services is the key to poverty alleviation. This strategy will also increase the welfare of the many individuals whose consumption levels fall just above the PL. However, economic growth by itself is unlikely to be sufficient to bring the extreme poor out of poverty. Neither will their capabilities increase just as a result of a general improvement in access to services. For this hardcore of individuals, specific anti-poverty interventions are required. B. IMPROVING THE ODDS FOR THE POOR: BROAD-BASED GROWTH STRATEGIES AND STRUCTURAL REFORMS 6.6 This report suggests that, at the end of 2000, just over half of the population of Kosovo lived in poverty. It is clear that poverty incidence of this magnitude cannot be eliminated by transfer programs and the provision of social services alone. Moreover, the analysis shows that these individuals share a number of characteristics that indicate their poverty status will be responsive to economy wide growth. A major component of the poverty reduction strategy must therefore be a focus on the renewal of economic growth that is broad-based and driven by the key structural reforms and institution building which are necessary to support Kosovo in the new stage of development. 6.7 JIAS acted early to lay the fundamental conditions for macroeconomic stability by permitting the use of the deutsche mark in all transactions without disturbing the legal status of the dinar. This placed the onus of macroeconomic policy on fiscal policy. The basic framework for credible fiscal management has been established, including budget formulation and implementation procedures, continuing reform in the tax structure, and the integration of locals in technical and policy tasks. Local revenue mobilization for the budget has risen impressively over the past two years, from one-third of the total budget in the final four months of 1999 to two-thirds in 2001, with the balance financed by donors. For sustained macroeconomic stability, and in the run-up to the provincial elections scheduled for November 2001, it will be of the utmost importance to focus efforts on institutional capacity building, to support and strengthen the quality of the civil service. 6.8 To create the conditions for durable growth in Kosovo, it is essential that fiscal discipline be maintained. This will require (i) responsible budget management; (ii) a fturther widening of the tax base through tax policy reform and strengthened administration; and (iii) the further development of fiscal institutions ensuring the efficiency and probity of public spending, especially at the local level. 56 6.9 The diagnostics presented in Chapter 2 suggest that the majority of the poor live in rural areas and are predominately farmers. Within the context of long-term macroeconomic stability and growth, long-term improvements in the welfare of these individuals, and reductions in poverty head-count are conditional on the development of an adequate and sustainable rural strategy designed to increase productivity in agriculture, improve access to markets and develop job opportunities in related sectors. Given the importance of household transfers as a source of income, migration to urban areas and abroad and repatriation of earnings is an important component of this strategy. 6.10 The findings of this study also show that the non-extremely poor in urban areas are well-equipped to benefit from any increase in job opportunities in the non-agriculture labor market, which growth will bring. 6.11 Given the important role played by self-employment for the Albanian community and on the basis of the experience shown elsewhere, the major source for new jobs will lie in the small and medium enterprise (SME) sector. Thus expansion of small and medium scale enterprises, in both service activities and light industry, will be of crucial importance to promoting labor-intensive growth. Yet for this to happen, Kosovo needs to further develop and implement rules and market mechanisms, both legal and institutional, under which the private sector can succeed. In essence, there are two core sets of private sector development policies that need to be pursued. 6.12 First, Kosovo needs to establish a business environment where there is respect for the rule of law, property rights and claims are recognized, contracts can be enforced, information and services in the marketplace are available, and investors can have access to capital. A pressing challenge at the moment is to strengthen institutional capacity, and in particular the capacity of the judicial system. 6.13 Second, Kosovo needs to expand the private sector's composition by transferring potentially viable public enterprises to private ownership or divest productive assets from the public sector into private hands. This process is similar to what is required in agriculture to increase access to land. JIAS is currently in the process of setting up the legal framework required for the is process to begin in manufacturing. C. TOWARDS A STRATEGY TO REACH THE EXTREME POOR 6.14 Even the high growth which is predicted for Kosovo in the next few years is unlikely to improve significantly the odds for the extremely poor. For this hardcore group of individuals, specific anti-poverty interventions are required to avoid their becoming an underclass of individuals marginal to the economy and the social structure. The effectiveness and efficiency of these policy levers can be improved and guided by the reports findings on the characteristics and the correlates of extreme poverty. Box 6.1 summaries the major correlates of extreme income poverty and other dimensions of poverty based on the analysis of the multivariate analyses in Volume II. The emphasis is on identifying the most vulnerable groups according to different definitions of poverty. 57 Box 6. 1: Correlates of extreme poverty Dimension of Highest incidence/lower outcome among poverty Consumption * Only primary education or less * Non participation in labor force/unemployment * Agricultural workers * Currently IDP * Have less than 1 hectare of land and no machinery * Disabled * Latrine or absence of sewage * Precarious dwelling or stones/mud materials Education * Rural population; * Ethnic minorities; * Girls; * Income poor; * Father with low educational attainment; Children from large households. Health Health status: * Serbian; * Women; * Individual with low education; * Small households (especially for women). Health care * No significant differences emerge when controlling utilization: for age and health status but * Albanians make comparatively more use of private facilities with higher out-of-pocket payments. * Poor face somewhat higher barriers to access Source: World Bank estimates from LSMS. The Extremely Poor are Deprived in More Than One Dimension 6.15 The results above are instructive in a variety of ways. The first is the degree of overlap between different dimensions of poverty. A comparatively poor educational attainment and - to a lesser extent - poor health status are strong correlates of income poverty. Being extremely poor is a significant impediment to enrolment in secondary education and limits access to health care. The key to alleviating extreme poverty therefore begins with identifying the characteristics of the individuals that score relatively badly in more than one dimension of poverty. 6.16 Among the traditional correlates of poverty, particularly important are: * being unemployed-or out-of the-labor force -if living in urban areas, or working in agriculture if in rural areas; and * living in a household with a high dependency ratio. 58 6.17 It is also worth noting that the incidence of extreme poverty is slightly higher in urban areas than among rural communities and that gender represents the single most important correlate of poor educational attainment. Post-conflict correlates are also important. 6.18 Having identified the characteristics of the individuals who are most likely to fall into extreme income poverty and human capital deprivation, the challenge is to design a strategy that satisfies the following constraints: * it is fiscally sustainable and; * reflects local priorities. Fiscal Sustainability. 6.19 The total recurrent budget for Kosovo in 2001 is DM 500 million, with DM 95 million allocated to social protection. Roughly two thirds of the budget is expected to be collected in customs duties and sales taxes. The remainder is being funded by contributions from international donors. Donor funding for recurrent expenditures is expected to disappear altogether over the next two to three years. As the economy grows and the tax base widens, tax revenues should rise. The tentative revenue forecast for 2002 is in the range of DM 550 million, increasing to over DM 600 million by 2004. Much of this rise will, however, substitute for donor funding. 6.20 JIAS has been under-spending on non-wage budget expenditures. To promote equitable development, this type of expenditure should be increased in a targeted and coordinated fashion. In the context of the 2002 budget cycle, JIAS should take the opportunity to evaluate spending priorities. Local Priorities. 6.21 Discussions with local stakeholders has also identified priority groups and programs for social protection. Priorities concern war veterans, civilian victims of the recent conflict, the elderly, and the handicapped. Of this only the last two categories appear to be high poverty risk categories on the basis of the analysis above. D. DESIGNING INTERVENTIONS TO ALLEVIATE EXTREME INCOME POVERTY 6.22 The findings of this study point to a nurnber of possible policy levers to address the issue of extreme in come poverty. Amongst those are: * Improving the short-term prospects for non-agricultural employment of the extremely poor via a selective system of community-based public works; * Increase access to land for extremely poor households involved in agriculture; and, * Improving the targeting of the current social assistance system and increasing its coverage of vulnerable groups. Within this context a transfer program conditional on children attending school should be evaluated. 59 6.23 Public works. The findings of this study point to underemployment in agriculture and lack of employment opportunities in the non-agricultural sector as main correlates of extreme poverty. Category II of the current social assistance scheme has sought to help this group by providing benefits for poor households with unemployed members actively looking for a job.80 However, prolonged cash assistance and lengthy periods of inactivity reduce both job search incentives and readiness for work. In circumstances of labor market shocks and high unemployment, public works programs offer a useful alternative to cash transfers. At the same time, public works schemes can be designed to support the rehabilitation of socio-economic infrastructure in Kosovo. In designing the schemes, care would have to be taken so that they do not simply substitute for existing jobs, and that wages are sufficiently low to self-select the poor. It is also important to recognize that Kosovo has just been through a massive reconstruction phase which has been more extensive in some areas than in others - e.g., roads versus school rehabilitation. Any public works scheme should therefore be carefully selected with a view to reflecting local needs and minimizing the disincentives for gainful employment. 6.24 Increase land use. Small land holdings in rural area is a strong correlate of extreme poverty. Most of the agricultural land is under private ownership but about 12 percent of the agricultural land is under the regime of collectively owned farms. Ancillary evidence showed that in some cooperatives only 25 percent of the land was cultivated. The unused land represents a small fraction of the total arable land in Kosovo but could represent a significant change for those households involved in its cultivation. This report simulated the effect of granting extremely poor rural household access to this land so as to increase their usage of land to 2 hectares. This policy would reduce the incidence of extreme poverty by around 25 percent bringing the head count down to 9 percent. However it is a relative expensive policy and faces potential problems such as pre-existing property rights, the quality of the land and the geographical distribution of the unused land and the extremely poor households. 6.25 Social assistance. The current social assistance scheme has kept expenditures under control, and it has targeted categories that display high poverty risk, but, with a relatively high degree of targeting errors-both of exclusion and inclusion. On the one hand only 16.5 per cent of the extremely poor declare to have received social assistance, on the other more than 80 per cent of program resources are granted to households who are not extremely poor and 35 percent of them to the not poor. 6.26 To a large extent, errors of inclusions reflect the fact that eligibility for Category I was determined exclusively on the basis of documentation presented by the applicant. The use of a limited number of easily observable proxy indicators for eligibility for Category II represents a step in the right direction. However, the criteria used to target the benefit appear to be too stringent without eliminating leakage. According to the analysis in chapter 3 only 13 percent of the extremely poor households meet the eligibility requirements for Category I benefits and only a further 4 percent have the characteristics s See Chapter 3 of this volume for more details. 60 required for Category II eligibility. Chapter 3 suggests conditions that could be released without further increasing the already very high potential leakage of the system. 6.27 Family allowance. Finally the findings of this report suggest a strong correlation between dependency ratio and the incidence of extreme poverty. This points to family allowances as an effective lever in poverty alleviation which will reduce the vicious circle of intergenerational poverty. However, with limited public resources, a flat-rate universal benefit of this type is unlikely to be fiscally sustainable, even in the medium term. The number of young children in the household can then be used as a proxy for poverty, following the finding of this report of a strong correlation between these two variables. The suggestion is that the transfer should be conditional on the children attending school to break the poverty, lack of education cycle which appears to exist in Kosovo as in many other countries. E. IMPROVING EDUCATIONAL OUTCOMES 6.28 The findings of this report suggest that a poverty alleviation strategy in the area of education needs to focus on two essential components: * Ensuring that poor families can afford schooling; and * Promoting enrolment in secondary education, and increasingly in primary, of girls and some minority groups. Ensure that Poor Families Can Afford Schooling 6.29 In Kosovo education is provided almost entirely by the public system. Direct parental contributions to schools are not formally compulsory, although they are pervasive (99 percent of the households) and are considerably large across the income distribution. On average 6 percent of household expenditure on education covers fees and another 12 percent tuition costs. Total books and equipment accounted for 11 per cent, while 'in-kind' and 'other expenses' account for over 25 per cent. 6.30 JIAS should strongly consider the possibility of: * Targeting its limited funding on the provision of textbooks and basic equipment for the very poor; * Outlawing, auditing and overseeing institutions and individuals that collect 'under-the-table' fees and other contributions in the public education system; * Improving the efficiency in the use of public resources and reduce provision costs; * Encouraging the development of a private market for education, once the adequate regulatory framework has been set up; and * Adapting the social assistance system to make benefits conditional on children with many young siblings attending school up to the age of 18. Promote the Education of Specific Groups 6.31 Making education more affordable will go a long way towards improving enrolment and attainment but is unlikely to eliminate the barriers to access which are 61 currently faced by girls, particularly in rural areas, and by children from 'other ethnics groups'. Security both inside and outside the school appears to represent an important constraint on school enrolment for both these groups. This is an issue that goes far beyond the responsibility of the educational system and is, for obvious reasons, a high priority for JIAS and the entire international community. 6.32 In addition research from other countries shows that the direct costs of schooling pose more of a barrier for girls than for boys. Although evidence for Kosovo on this subject is not available, targeted interventions to reduce or eliminate direct and indirect costs should be evaluated. Indirect or opportunity costs are also traditionally important determinants of girls' educational attainment in low income countries, but are thought to be somewhat less significant in Kosovo. The 'conditional transfer program' suggested above will alleviate the financial barriers to entry. 6.33 How to improve the enrolments of 'other ethnic groups' poses an additional challenge. Before a strategy in this area is formulated, a more detailed analysis of what represents the main barriers to entry for these groups is required. F. IMPROVING HEALTH STATUS AND ACCESS TO HEALTH CARE 6.34 Both health status and the state of the health care system are in need of upgrading. The results of the past decade have taken a toll on the health status of the population, and the health care system has provided minimal support due to a combination of historical treatment by Kosovo in SFRY and the chaotic system of the recent past. The neglect of the health system has led to a multi-funded health system and comparatively poor health indicators. Of particular concern are: * the high infant mortality; * the low percentages of ante-natal care in pregnant women, and births attended by a professional; and * the reported decline in immunization coverage. 6.35 While lack of data prevents an accurate analysis of income-related inequalities in health outcomes, the expectation is that the income poor carry much of the burden of poor health. This is confirrned by the analysis of self-reported illness amongst adults which shows poor health to be somewhat more common at the lower end of the consumption distribution. 6.36 The poor health outcomes appear to be related to the quality of health care received rather than to barriers to access. The percentage of the population that reports having had a medical problem in the four weeks prior to the survey but not having sought medical care is comparatively small and does not vary substantially across the income distribution. This suggests the need for upgrading - rather than expanding - services in all their aspects: infrastructure, human resources, equipment and management. 62 6.37 Finally the main constraint to health care utilization appears to be represented by the high out-of-pocket contributions paid to obtain access to health care. An average of 28 percent of those that reported having had a medical problem for which they did not seek treatment, did so because they could not afford health care. The financial consideration is particularly binding for the worse-off individuals. Over 40 percent of the individuals in the bottom two consumption deciles who did not seek medical care when in need, were prevented from doing so by the cost of the treatment. 6.38 The role of affordability as a barrier to entry comes as no surprise considering the high out-of-pocket payments the population of Kosovo appears to face. Over 95 percent of Albanians pay for the services received, independently of whether these are provided in public or private facilities and the payments are large across the consumption distribution. Developing a Blue Print for Improving Health Outcomes 6.39 The findings of this study suggest that, without radical efforts to restructure the health system neither the poor nor the general population will be adequately served. The following priorities are recommended: * Establish realistic levels of health service output commensurate with public funding levels; * Define parameters of a health reform strategy that aligns resources and services; * Formalize under-the-table payments via a system of means-tested co-payments; * Undertake efforts in public health to ensure immunization coverage, adequate access to clean water, reduced environmental health risks; and * Focusing on child and maternal health. 6.40 Adopting some of these measures may result in services that the public must withdraw from providing, or means testing access to ensure equity in health care. Future access may need to be compromised to reduce overcapacity of both physical and human infrastructure. Formalizing and legalizing under-the table-payments, and reducing excess capacity could help to reduce the corruption in the system, and ensure that the public sector is financing both high priority interventions as well as subsidizing those unable to afford needed care. It should also be designed to reduce the existing financial barriers to access. Indeed, the high incidence of out-of-pocket payments commonly leads to under- utilization by those least able to pay81, and a system low on accountability. G. TOWARDS A COMPREHENSIVE STRATEGY FOR POVERTY REDUCTION 6.41 The findings of this report show that extreme poverty in Kosovo is a complex and multifaceted condition which span across different dimensions. As such in order to achieve durable progress in reducing poverty, a comprehensive development strategy which is inclusive and pro-poor is needed. The basis for this strategy is macroeconomic stability and sound economic development. The use of the deutsche mark as currency 81 Lewis (2000) 63 places the onus of macroeconomic stability on fiscal policies. It is therefore essential that any poverty reducing package is designed to be fiscally sustainable. 6.42 A realistic strategy will also need to acknowledge the limited local institutional capacities and the weak accountability mechanisms, which will negatively affect service delivery and the effective implementation and management of public policies, including those on poverty. The strategy is therefore to proceed on two parallel lines. In the short run the objective is to design policies that are simple to administer and implement and can be easily monitored. In the longer term the aim is to built the institutional capacity and put in place stronger accountability mechanisms. These are essential conditions for private sector development, adequate public expenditure management, and equitable delivery of essential public services, all of which are critical to broad-based sustainable growth and poverty reduction. 6.43 A successful strategy also requires regular evaluation of the effects of policies used and monitoring of the socio-economic conditions of the population. This is particularly important in the Kosovo context where both the economy and the society are in a very fluid state. However, the policy makers at the moment lack the information and data needed to develop and monitor policy, and the capacity to elaborate that information in order to strengthen the strategic orientation of its poverty reduction agenda. The LSMS represents a snap-shot of the living conditions of the population of Kosovo a year after the conflict. This is the information the report is based on. As such it closely represent a benchmark against which the economic development of Kosovo and any future changes in the welfare of its population can be assessed and evaluated. However future policy making will require regular updating of this information and it is essential to develop local capacity to collect, analyze and disseminate reliable, useful and timely statistical and other data on trends in household consumption, income and other key socio-economic indicators, as well as on other key aspects of economic and social activity. 6.44 Finally it is important to combine the strengthening of the monitoring mechanisms with the increased capacity to generate and analyzed data into a culture of using this information to inform the development and implementation of new policy initiatives through improved monitoring of the impact and effectiveness of Government programs. 64 REFERENCES Grosh, M. (1994) Administering Targeted Social Programs in Latin America. From Platitudes to Practice. World Bank Regional and Sectoral Studies. Washington, D.C. IMF. 2000. Kosovo - Macroeconomic Issues and Fiscal Sustainability. Institute of Health Sector Development. 2001. "Report to the Department of Health and Social Welfare of UNMIK, Kosovo, Education and Health project, health component B-03/00, status report no 1." UN Population Fund (UNFPA) and the International Organization for Migration (IOM). National Survey, November 1999 and February 2000 International Rescue Committee, Institute for Public Health of Kosovo, WHO and US Centers for Disease Control. 1999. "Kosovar Albanian Health Survey Report." Oxford Policy Management with Care International UK. 2001. "Kosovo: A Qualitative Poverty Assessment, Data Analysis Report." World Bank. 2000. Conflict and Change in Kosovo: Impact on Institutions and Society, draft. Piere]la Paci Q:\Kosovo\pa\draft PA\Kosovo PA Dec 5, 2001-revised.doc December 4,2001 1:45PM 65