Report No. 32378-XK Kosovo Poverty Assessment Promoting Opportunity, Security, and Participation for All June 16, 2005 Poverty Reduction and Economic Management Unit Europe and Central Asia Region Document of the World Bank Table of Contents EXECUTnTE SUMMARY ......................................................................................... i 1 INTRODUCTION . ................................................................................................ 1 2. THE CHALLENGE OF POVERTY MONITORING ..................................... 3 A. THE SOCIAL STATISTICAL INFRASTRUCTURE...................................................... B. INSTITUTIONAL ANALYSIS....11 3 ARRANGEMENTS CAPACITY FORPOVERTY AND 3. THE MANY DIMENSIONS OF POVERTY .................................................. 14 A. LIMITED INCOME 14 B. OPPORTUNITIES..................................................................... LOW EDUCATIONAND HEALTH 25 C. D. POVERTYASA DEFICITINEMPOWERMENT POVERTYASA LACK SECURITY ................................................................... CAPABILITIES .................................................. OF 30 ....................................................... E. MULTIPLE DEPRIVATIONS POVERTY LINKAGESTHELIFECYCLE........44 AND IN 49 4. A FRAGILE SOCLAL COHESION ................................................................. 53 A. INCOME INEQUALITY 53 B. ........................................................................................ SPATIAL C. THEETHNIC DIMENSION................................................................................... DISPARITIES........................................................................................ 55 D. 58 GENDEREQUALITY ........................................................................................... E. 60 DISPARITIES ACROSSGENERATIONS................................................................. 65 5. THE PRO-POORPOTENTIAL OF PUBLIC SERVICE D E L M R Y .......67 A. SUFFICIENCY ..................................................................................................... 67 B. EFFECTIVE~SS ................................................................................................. C. 71 EQUITY ............................................................................................................. 75 6. POLICY DIRECTIONS .................................................................................... 78 A. PROMOTINGINCOME B. OPPORTUNITIES ............................................................... 78 .................................................... 79 C. RAISINGHEALTH AND EDUCATIONABILITIES D. REDUCING VULNERABILITY .............................................................................. 80 E. PROMOTINGEMPOWERMENT ............................................................................. 81 SUPPORTINGANEVIDENCE-BASED DIALOGUEONPOVERTY ........................... 82 SELECTED REFERENCES .................................................................................... 84 ANNEX: DESIRABLE AND MEASURABLE LIST OF POVERTY INDICATORS ............................................................................................................ 86 ListofTables Table 1: Summary o f Poverty Diagnosis and Policy Directions ................................. xii Table 3: Opportunities and Challenges to Measure the MDGs inKosovo..................10 Table 2: Comparison o f Sample Distributions from LSMS. LFS. andHBS.................6 Table 4: Key Macroeconomic Indicators..................................................................... Table 6: Extreme Poverty Rates by Selected Socio-economic Groups ...................... Table 5 Poverty Estimates Derived from the 2000 LSMS and 2002 HBS..................15 17 20 Table 8: Labour Status and Sources of Income........................................................... Table 7: Income Structure by Poverty Status and Location (percent) ......................... 23 24 25 Table 10: Selected Key Educational Indicators inKosovo ......................................... Table 9: Kosovo and the MillenniumDevelopment Goals ......................................... 26 Table 11: Infant Mortality Rate (IMR) and Life Expectancy (LE) inKosovo as Compared to Regional and EUData.................................................................... 27 Table 12: Selected Matemal and Child Health Outcomes inEurope and Central Asia Table 13: Simple IncomePoverty Projections............................................................. .............................................................................................................................. 28 31 Table 14: Main Labor Market Indicators..................................................................... 32 Table 15: Households Expenditureson Health, 2000.................................................. 34 Table 16: Blood Lead Level (BBLinpglL) inPregnant Women and Children, 199235 Table 17: Geographical Distributionof MainEnvironmental Hot Spots inKosovo ..36 Table 19: Number of Reported Cases andForms of Domestic Violence.................... Table 18: Access to Basic Infrastructure Services (percent o f households) ................38 44 Table 20: Indicators of PoliticalRights and Civil Liberties inSouth-East Europe, 2004...................................................................................................................... 45 Table 21: Opinions on the Biggest Problems Facedby Kosovo (percent o f respondents) ......................................................................................................... 45 Table 22: Level o f Satisfaction with Public Institutions (percentage of respondents) 47 Table 23: Opinions about the Presence of Corruption inVarious Public Services (percent o f respondents)....................................................................................... 48 Table 24: Access to Information and Communication Technologies inKosovo around ..48 Table 25: The Multiple Deprivationof the Extreme Income Poor.............................. 2002.................................................................................................................... Table 26: InequalityIndices......................................................................................... 50 54 Table 27: Projected Extreme Poverty Rates with different GrowtMnequality Paths ............................................................................................................... Table 28: SelectedSocio-economic Indicators by Regions......................................... (percent) 54 55 Table 29: Selected Socio-economic Indicators by Type o f Settlements (percent) ......57 Table 30: Selected Key socio-economic Indicators by Ethnicity (percent) ................-59 Table 32: Selected Socio-economic Indicatorsby Gender.......................................... Table 31: Ranking o f Opinions on the Biggest Problems inKosovo, by Ethnicity ....60 Table 33: Selected Socio-economic Indicators by Age Categories (percent) ..............64 Table 34: Public Social Spending inKosovo and SelectedCountries in2003 ...........65 Table 35: Poverty Reduction Impact o f Social Transfers............................................ 68 74 Table 36: Poverty Incidence of Public Spendingon Basic Education, 2002 (percent o f households) .......................................................................................................... 76 Table 37: Social Assistance and Individuals Characteristics o f Recipients Among the Poor and Extreme Poor (percent ofhouseholds) ................................................. 77 Table 38: Multidimensional Objective Poverty Indicators Within the Life cycle Approach.............................................................................................................. 86 Table 39: Non-age Specific Quantitative MultidimensionalPoverty Indicators........-92 Table 40: Poverty Indicators Within the Life Cycle: Opportunities............................ Table 41: Poverty Indicators Within the Life Cycle: Security ............................ 96 Table 42: Poverty Indicators Within the Life Cycle: Empowerment ...........................; ......95 96 ListofFigures Figure 1: GDP Per Capita inSelected Economies o f South-east Europe in2002 (in current US$) ......................................................................................................... 17 Figure2: Dependency o fHouseholdsonNaturalResources (percent o fhouseholds)3 1 Figure3: SelectedIndicators of HousingConditions (percentage of households) ......40 Figure 4: Number o f ReportedCrimes ........................................................................ 42 Figure 5: School Enrolmentby Level and Poverty Status in2003 (percentage o f Figure6: Primary and Secondary SchoolEnrollmentby Gender andGrades, 2002 ..62 children) ............................................................................................................... 51 Figure 7: Female-Male RatiosinLabor Market Indicators, 1989-2002...................... Figure 8: Social Assistance Recipients, 2001-2003..................................................... 63 70 Figure9: Targeting Effectiveness; Share of Pre-transfer Poor, Very Poor andNon- poor Receiving Social Assistance Benefits (percent o f households) ...................73 Figure 10: Percentage of Teachers inPrimary and Secondary Schools with Inadequate QualificationRespectively by Poverty at the Municipality Level ....................... 76 ListofBoxes Box 1: The Three Standard Foster-Greer-Thorbecke (1984) Aggregate Poverty 11 12 Box 3: Construction of the Absolute Poverty Lines .................................................... Box 2: Improving the HBS for Poverty Analysis ........................................................ Measures .............................................................................................................. 18 Box 4: Voices o f Vulnerable Groups inKosovo ........................................................ 22 Box 5: EnvironmentalDegradationinKosovo ............................................................ 33 Box 6: ReconcilingEconomic Activity with EnvironmentalProtection: The Trepca Box 7: People Talk about the Poor State of Infrastructure Services............................ Complex............................................................................................................... 37 39 Box 8: Youth Voices About Their Lack o f Influence inDecision Making.................46 Acknowledgements This report was preparedby a team led by Alexandre Kolev (Task Team Leader and main author), and comprising Betty Hanan, Giorgio Tamburlini (Health), Cem Mete (Education), Janet Owens (Gender and Poverty Monitoring), Tiziana Tamborrini (Environment), Hjalte Sederlof (Social Protection), Sasun Tsirunyan (Income Poverty), Paolo Verme (Multidimensional Poverty and Life Cycle), Selma Kapo, Vigan Behlui (Qualitative Poverty Assessment), ShpendAhmeti, NandShani andNeilBush(Poverty Dialogue). The team would like to thank Asad Alam, Cheryl Gray, Tracey Marie Lane, Pierella Paci, Gloria L a Cava, Peter Darvas, Michael Stanley, and Ruslan Yemtsov who provided excellent comments, suggestions and guidance on the poverty report. The team would also like to thank Viviana Mangiaterra and the Children and Youth Unit, as well as the Italian and Gender Trust Funds for the support it has received. The peer reviewers Jeni KlugmanandKaspar Richterprovidedinvaluablesupport andguidanceto the team. The team graciously acknowledges its indebtedness to the United Nations Development Program (UNDP) and the Statistical Office o f Kosovo (SOK) teams for their close collaboration and sharing o f data with the World Bank team. The team would also like to thank officials from the Provisional Institutions o f Self-Government (PISG), inparticular M. Behxhet Haliti and Ms. Flaka Surroi from the UnitedNations Interim Administration Mission in Kosovo (UNMIK), and colleagues fiom the United Nations Children's Fund (UNICEF), United Nations Population Fund (UNFPA), UK,Department for International Development (DFID), the Swedish International Development Cooperation Agency (SIDA), the German Agency for Technical Cooperation, the World Health Organization (WHO), and the UnitedNations HighCommission for Refugees (UNHCR). The team would also like to thank Orsalia Kalantzopoulos, Nancy Cooke, and Robert Jauncey for their guidance, as well as Sidi Boubacar and Kanthan Shankar for their invaluable support andadvice throughout the process o f writing and disseminating this report. Specific comments on the background papers commissioned for the report were receivedfrom Government anddonors during a first workshop organized inJune 2004jointly with UNDP. A first discussion draft o f the report was presented to the Cabinet, and subsequently discussed with the donor community, inJanuary 2005. Feedbackreceivedwere incorporated in a revised discussion draft that was further discussedwith civil society during a series o f consultations organizedby UNDP inApril 2005. The production o f this report has also greatly benefited from the logistical support providedby Valery Ciancio, Alison Panton andJudy Wiltshire. EXECUTIVESUMMARY 1. Thepopulation of Kosovo has suffered substantially over thepast decadefrom deep economic, social andpolitical turmoil. Duringthe 199Os, Kosovars were subject to the shock of transition, the neglect and exclusion o f the Albanian community during the Milosevic era, and the ensuing armed conflict in 1999. Existing studies indicate that by the end of the decade, output had more than halved, income had collapsed, less than half o f the population was employed, andhalfwas inpoverty. 2. The reconstruction efforts since the end of the conflict have succeeded in putting theKosovo economy on a growthpath, but littleprogress was made with the clarification of Kosovo's political status. The province has benefited greatly from the restoration o f peace and stability, and frommassive inflows of donor assistance.There is still some uncertainty on economic data, but these suggest that progress with reconstruction, macroeconomic stability and the establishment o fpublic institutions have ledto a recovery of economic activity since 2000. GDP growth reached 21 percent in2001`and has since averaged 4.3 percent. GDP per capita increasedfrom about US$400 in 1995 to US$790 in2003. Gross National Disposable Income (GNI) per capita was even higher at about US$975 in 2002 and US$1170 in 2003, reflectingthe importance o fremittances. 3. Despite considerable progress with reconstruction, the challenge of poverty reduction in theyears to come remains enormous. The recent growth performance has been drivenby a post-conflict boom fmancedby official development aid flows, but it has not led to a significant creation o fjobs and it i s unlikely to be sustainable without a clarification of KOSOVO'S political status and the maintenance o f peace and security for all. While recent growth and moderate consumption inequality have most likely led to a reduction inpoverty between 2000 and 2003, still about 37 percent o f the population was estimated to live in poverty according to the 2002/03 Household Budget Survey (HBS) data. A fragile social cohesion, poor environmental conditions, and deficient industrial and basic infrastructures could further undermine the sustainability o f economic growth and its impact on poverty reduction. There are additional concerns that the pro-poor potential o f public social service deliverycouldbejeopardized by weak govemance, alow level o f administrative capacity and inadequate resources. Last, but not least, the lack o f a comprehensive and integrated social statistics system in Kosovo could limit the ability o f Government to formulate pro-poor policies andplan hture investments. 4. The aim of this report is to contribute to the on-going policy dialogue on poverty reduction in Kosovo and to support the formulation and prioritization of well-informed public policies. Specifically, the report has been prepared as an input toward the gradual building o f a donor-supported, home-grown, comprehensive poverty reduction strategy. It i s articulated around the following key questions: (i) what extent does the social statistic To system inKosovo allow policy makers to form an accurate picture of the poverty situation and hence to formulate well-informed pro-poor policies? (ii) is the true extent and What natureo fpoverty? (iii) are the poorest groups? (iv) How cohesive i s the Kosovo society? Who (iv) I s public social service delivery sufficient, effective and equitable? This report has been produced as a result o f working collaboratively with SOK and UNDP. In addition, nine background papers on various aspects o f poverty were commissioned and their fmdings have been incorporatedselectively inthe mainreport'. The Challenge of Poverty Monitoring 5. Substantial progress has been made by SOK in the collection of household level data on consumption and income, but the lack of a recent population census remains a major impediment. SOK i s conducting an increasing number o f surveys that are meant to be representative at national level and for urbdrural and the Serbhon-Serbs sub populations. The current fielding o f the HBS represents a substantial progress with the collection o f household level data on consumption and income necessary for monitoring income poverty. However, Kosovo lacks a recent population census to assess the accuracy o f existingdata and improve the quality o f sampling designs. Without a population census that provides up-to- date population estimates, not only is it almost impossible to assess whether the estimates generated from various surveys are representativeat national level and for sub groups, but it is also very difficult to improve the quality o f sampling designs. 6. The current household survey system does not allow the comparison of critical income and non-income poverty estimatesacrosssurveys and over timefor theperiod 2000- 2003. The investigation conducted in this report shows that there are some methodological differences between the Living Standards Measurement Survey (LSMS) and the H B S in the collection o f data on consumption and income which make the comparison of incomepoverty estimates problematic. Previous World Bank poverty estimates were based on the 2000 LSMS. The new estimates reported in this report are based on the 2002/03 HBS. The 2000 LSMS and the 2002/03 H B S are not entirely comparable, mainly due to differences in consumption modules and the use o f survey-specific povertylines. Only after future rounds o f the HBS have been undertaken and analyzed, can one make more definite statements about the trends inpoverty and extreme poverty. There are other inconsistencies in the household survey system that make the comparison of non-income indicators of poverty such as educational achievements andhealth outcomesacross surveys andover time most difficult. 7. insufficient. Critical gaps in data collection further undermine Kosovo's capacity to measure The data necessary to measure non-income dimensions of poverty is still the multidimensional aspects o fpoverty, including several indicators necessaryto monitor the MillenniumDevelopment Goals (MDGs), andto analyze the poverty linkages and the causes ofpoverty. 8. Access to SOK data remains a problem. The Statistics Office maintains a comprehensive and up-to-date website of the office's most recent publications and statistics generated from its survey system. Unfortunately, only published data i s available for public use. The statistics office enforces a strict policy on data access. No data is ostensibly allowed to leave the premises o f the office. As a result, much o f the data i s under-utilized and the monitoring and analysis o f poverty i s made more difficult. Data access policies vary across the region. In eastem Europe and central Asia, countries with an open access to survey data include Albania, Bosnia and Herzegovina, Estonia, Georgia, Hungary, the Kyrgyz Republic, andMoldova. Consistent with intemational standards, the sharingof household level data can improve poverty monitoring program capacity and lead to efficiency gains in poverty analysis. The list of the backgroundpapers i s providedinthe reference section. .. 11 TheManyDimensionsofPoverty 9. Income poverty appears widespread but is relatively shallow. Four years after the end o f the conflict, income poverty remains widespread. Calculations based on the HBS shows that in 2002 about 37 percent o f the population in Kosovo was living below the poverty line o f Euro 1.42 per adult equivalent per day and 15.2 percent were below the extreme (food) poverty line o f Euro 0.93 per day. However, the size o f the poverty gap (11.4 percent) and the severity of poverty index (4.9 percent) indicate that poverty was not extremely deep. For all income groups, income from work was the main source o f income (59.9 percent) followed by private remittances from abroad (15.2 percent). For the extreme poor, state transfers constituted the third largest source o f income (18.4 percent). Income diversification and own-produced food were important coping mechanisms. 10. Income poverty is found to affect disproportionately children, the elderly,female- headed households, the disabled, non-Serb ethnic minorities, the unemployed, and precarious job holders. The poverty profile has remained practically unchanged since 2000 and is consistent with other recent quantitative poverty profiles. Poorer living conditions among the elderly, female headed households, families with children, the disabled, and the unemployed is also supported by qualitative evidence. There is no accurate information on internally displaced persons (IDPs), but available evidence from small scale surveys points to their extremely difficult living conditions. the Kosovo economy togenerate sufficient viablejobsfor its large youth population. So far, 11. Joblessness is an important cause of incomepoverty that reflects the difficulty of the growth performance in Kosovo has not been driven by a revival of the formal private sector leading to more sustainable formal job creation. There i s still a lack of consistency in labor market data, but data from the LFS show persistently highILO unemployment rates in the post-war period: 41.2 percent in2001,47.2 percent in2002, and 44.4 percent in2003. An overall low labor demand, relative to the large youth labor force, is a major problem in KOSOVO,that is strongly associated with income poverty. While many existing formal one jobs are found in the service sector, and are closely reliant on the presence o f a large expatriate community, informal employment is very high with half o f workers with no - written contracts. Informal jobholding seems to be widespread in construction and in agriculture, andjob opportunities inthe industrialsector and inagro-processing have been so far very limited. Prospects for hture reduction in poverty will crucially depend on the removal o f barriers to a more labor intensive growth, specifically: (i) difficulty to attract the FDInecessaryto revitalize the industrial sector inthe absenceo f solutions to address political risks and to overcome delays with the privatizationprocess; and (ii) intense competition to the domestic producers inthe agricultural sector induced by subsidized agricultural imports. 12. Educational and health outcomes are still low, but there are signs of improvement. Besides the lack o f income, education and health capabilities are among the primary dimensions o f individual well-being. The report finds that the overall level o f completed education is relatively low inKosovo and that illiteracy i s still significant. In2002, half o f the adult population had only achieved primary education or less and 6 percent were illiterate, largely reflecting the legacy o f the past and its adverseimpact on older cohorts. Yet, since the end o f the conflict, considerable progress was made. Progress toward the achievement of universal primary school enrolment has been substantial, with gross primaryenrolment rates up to 95.4 percent in2003. The illiteracyrate was also reducedto less than0.5 percent among children and youth. Yet, more needs to be done, especially on pre-school and secondary school. Few children in Kosovo attend pre-school institutions and enrolment drops substantially at secondary level, with a gross enrolment rate o f only 75.2 percent in2003. As regards the health status ofthe population, there is still considerable uncertainty on estimates, but available evidence points to verypoor health outcomes which are also among the worst in south-east Europe. Health around birth and over the first year o f life is a major problem, as ... 111 indicated by current infant mortality rates ranging from 18 to 49 per 1,000 (4 to 10 times higher than the EUaverage). Inadequate nutritionis a particularproblemthat affects a fairly largenumber of children. The factors behindthe high maternal and infantmortality rates are fairly well identified and largely preventable. There are also signs o f improvements inthese indicators. Quantitative information on the health status later on in the life stages i s very scarce. Yet, available data suggest that tuberculosis and disability are major problems. The incidence of tuberculosis has declined substantially in recent years but remained 5 times higher than in westem European countries. And between 5 to 7.5 percent of the overall population have some kind o f disability. Qualitative data further indicate that mental health problems are widespread, especially for youngpeople. 13, Exposure to health risks is widespread, largely resulting from environmental pollution. Environmental pollution and contamination is widespread and represents a serious hazard to health in Kosovo. The major causes o f environmental health risks are due to: (i) outdated mining practices and industrial infiastructure that ignored environmental impacts; (ii) housing conditions andquality ofbasic infrastructure services, and(iii) poor weak environment management systems. Contamination of soil and foodstuff by heavy metal and lead is probably one o f the most serious problems in areas where the mines and industrial facilities are located. Inenvironmental hot spots, the lead intake o f people eating crops has been calculated to be more than three times higher than the WHO and FA0 recommended maximum weekly intake and about 15 times greater than the standardsfor EUcountries. The first studies conducted in the early 1990s also found some evidence o f lead poisoning in children from northem Kosovo, with blood lead concentration at birthinMitrovica two times above the intemationally accepted level. The on-going planned large scale testing program will probably c o n f m the earlier results. Water contamination is also a major environmental risk,Outdoor airpollutionis mostproblematic around outdated industrialinfrastructureand in Pristina due to polluting and unregulated traffic, Despite the decline o f industrial activity and the shut downofthe mostpollutingindustrialfacilities, Kosovo remains one of the highestair polluters inSouth-East Europe, with 5.5 tones of COzemissions compared, for instance, with 4.06 tones in Croatia and 3.8 tones in Romania. Indoor air pollution could affect a vast majority of households. Other important health risks are the threat o f a rapid HIViAIDS epidemic andthe prevalence o f riskybehaviors among youth. 14. A large share of thepopulation couldfall into incomepoverty as the result of a still unstablepolitical situation and the related risk of economicslow down. Besides the political instability and the related risks o f economic slow down that can affect the entire population, the report finds that a combination o f other and often inter-related factors explain a fairly large level of income insecurity and vulnerability to income poverty in Kosovo. These include: (i) environmental degradation which canjeopardize the livelihoods o f those who rely on natural resources, especially inrural areas; (ii)a highrisk o f unemployment andprecarious employment; (iii)constraints to internal mobility that can limit the capacity to mitigate income shocks; (iv) unexpected private health-care expenses that appear catastrophic for some households; and (v) low coverage o f the formal social safety net. Yet, households' capacity to self-insure against temporary income shocks is improved by access to family- based networks and private transfers, the reliance on multiple sources o f income, and the ownership of durables that could serve as collateral to access credits or be sold to compensate for temporary loss of income or to cover unexpected expenses such as health care. 15. The lack of personal safety is an important dimension of poverty that is not solely related to inter-ethnic tensions. The Standards for Kosovo stipulate that "all people in Kosovo are able to travel, work, and live in safety and without threat o f fear o f attack, harassment or intimidation .. .". Physical safety remains problematic in Kosovo, both at a national and household level. In the first place, there are elevated environmental risks that constitute a national security issue. Kosovo belongs to a major seismic area but its housing stock does not seem to comply with the security codes, puttinga large share o f the population iv at risk. Risks of landslides are high as the result of the high deforestation rate and are an important issue inthe Bardhand Mirashminingareas. Flooding can be a potential threat for households located near environmental hot spots. Unexploded mines used during the war remain an important safety issue. Inter-ethnic tensions are also an important potential source o f conflict, and o fparticular concern to internally displaced people. While the growing feeling o f insecurity observed between 2002 and 2003 i s not supported by crime statistics and only a few reported crimes in 2002 and 2003 were ethnically motivated, the March 2004 riots indicate that inter-ethnic violence remains a major issue. Finally the tiafficking inwomen and girls and domestic violence against women emerge as other important sources o f physical insecurity. 16. The empowerment of people in Kosovo is weak Empowerment refers to the capability of all people, regardless o f their income status, ethnicity, race, religion, or gender, to participate in,negotiate with, influence, and hold accountable institutions that affect their lives. Empowerment is an important dimension o f well-being, one that i s also well reflected inthe "Standards for Kosovo" under headings I(the functioning ofdemocratic institutions) and I1 (the rule of law). .The findings of this report show that there have been positive developments but several challenges remain. On the one hand, UNMIK has established a legal framework that promotes a democratic society and the rule o f law in Kosovo, gender equity is promoted in formal state institutions and has enabled a greater representation of women inpublic life, and participation and representation o f ethnic minorities is guaranteed by the Constitutional Framework. On the other hand, achievements in political and civil freedom i s mixed, the unresolved legal status o f Kosovo generates a general sense o f powerlessness and voicelessness, youths' ability to influence the institutions that affect their lives seems also very limited, many Internally Displaced Persons (IDPs) feel voiceless, women's influence in decision making remains at stake, and the participation for all in the political life is challenged by the risk o f non-participation o f Serbs in local elections. Moreover, the move toward more transparency inpublic management remains hinderedby the lack o f good administrative data. Corruption is considered a problem of medium to high importance depending on the sector, andit tends to affect disproportionately the poor. Access to internet remains also limited. The level of knowledge o f youth on reproductive health, HIV/AIDS,other sexually transmittedinfections(STIs) and safe sexpractices is stillvery low as i s the level o f environmental awareness, and indicate that access to information on these issuesremainproblematic. 17. The experience of poverty is more devastating at early ages. Around birth and during early childhood income poverty and environmental risks are major causalfactors for severe and sometimes irreversible health consequences. During adolescence, when schooling is no longer compulsory, income poverty has also a strong negative influence on school enrolment. Later on in the life cycle, low health outcomes contribute to income poverty through their adverse impact on employment. During adulthood, low education achievements become a major cause o f income poverty, mainly through lower wages and higherunemployment. Often, the different aspects ofpovertyreinforce eachother. InKOSOVO, the vast majority o f the extreme income poor are also poor inthe dimension o f education. The overlap between the income andhealth dimensions o fpoverty i s more difficult to capture with existing data. A nonnegligible share o f the extreme poor also live inpoor housing conditions. A large majority of the extreme poor are infrastructure poor - but infrastructure poverty is widespread andaffects all income groups almost equally. A Fragile Social Cohesion 18. Social cohesion is fragile and this may undermine the sustainability of economic growth and reduce the impact of growth on poverty reduction. While overall consumption V inequality is moderate, disparities in economic and social well-being by location, ethnicity, gender andage are important and maybe important sources o f discontent andpoliticalunrest. 19. Consumption inequality is moderate and shall thusfacilitate the transformation of average income growth into poverty reduction. Measures o f income or consumption inequality provide information on the extent o f disparities inmaterial well-being prevailing across households or individuals, and are thus key indicators o f social cohesion. Inequality measures are also important to forecast the impact o f economic growth onpoverty reduction. Income inequality is high in Kosovo relative to other neighboring countries, with a Gini coefficient o f 0.49, but consumption inequality i s moderate relative to neighboring countries, as reflected by a Gini coefficient o f 0.30. The simulations conducted inthis report also show that moderate consumption inequality is a positive development for KOSOVO, as an increase in inequality can reduce and even compromise the pro-poor potential o f economic growth. 20. Spatial socio-economicdisparitiesare important andjustifi a geographicalfocus in the poverty reduction strategy. Kosovo has pockets o f poverty in regions where local unemployment i s often high. Income poverty is lower inrural areas than in secondary cities, but this is mostly due to the consumption of own produced food and does not reflect better economic opportunities inrural areas. Employment opportunities inrural areas are indeedas bad as insecondary cities and worse than in Prishtina. Minorities are also indisadvantage due to restrictions in freedom o f movement, giving them limited access to markets where they could sell their products. The regional distribution of infrastructure services is relatively flat. Housing conditions seem to vary quite substantially from one region to another and are substantially worse insecondarycities, compared to Prishtina, but not as bad as inrural areas. Inrural areas, distance to secondary schools is a major barrier to education, and physical access to medical centers is also more difficult. This maybe one reason why secondary school completion rates are substantially worse in rural areas. In 2002, the secondary school completion rate was only 46 percent inrural areas, as opposed to 64 percent in secondary cities and 77 percent inPrishtina. 21, Socio-economic disparitiesbetweenAlbanians and Serbs are modest, but non-Serbs ethnic minorities face a substantial disadvantage. Consumption inequality i s largely explained by disparities within communities andnot by inequality between Albanians, Serbs and other minority groups predominantly composed of Slav Muslims and Roma. Moreover, socio-economic disparities between Albanians and Serbs are modest, but the maintenance o f a parallel system for the provision o f social services to Serbian minorities may undermine the development o f a unified social protection system and the perception o f social cohesion. Regardless o f actual circumstances, there also are areas o f agreement and divergence in the perceptions o f Albanians, Serbs and other minority about the main problems inKosovo. For all ethnic groups, there is a broad consensus that unemployment and uncertainty about the final status o f Kosovo are among the biggest problems in Kosovo. However, a sense o f insecurity is one o f the top problems reportedby the Serbs, and also a non-negligible concern among other minority groups while it i s not seen as a major problem by the Albanians. Further investigations shows that non-Serbs ethnic minority groups face a serious disadvantagenot only inthe dimension o f incomebut also with respect to education. 22. Gender inequity remains widespreadand interacts closely with ethnicity, indicating that the status of women in Kosovo is still at stake. The report shows that during childhood and adolescence, girls' education becomes at risk at secondary level among Albanians and non-Serbs ethnic minority. During adulthood, women face a substantial and growing employment disadvantage, but women disadvantage in terms o f lower wages is less pronounced. The combination o f limited schooling, early marriage, and entrenched attitudes among family and community about the fdfillment o f traditional roles help explain women labor market disadvantage. The analysis also shows that the higher incidence of poverty among female-headed households is mostly due to gender differences in educational vi attainment and sources o f income, with female-heads beingless educated and more reliant on pension. 23. Large pools of unemployed and idle youth is a worrisome fact in Kosovo and a major potential cause of social instability. Income opportunities are substantially worse among youth, with a youth unemployment rate of nearly 75 percent according to the LFS, which is more than two times higher than the adult unemployment rate. A major issue for youth is the difficulty to find decent work and to be occupied, which adds to their lack o f participation and voice inKosovo society and may contribute to increase the overall level o f ethnic and social tensions. There are also fewer income opportunities for the elderly, whose main concern is the substantialloss o f income incurredwhen leaving employment andjoining the pension system. The fact that pensioners are among the poorest, yet not entirely destitute, relates to the existence o f close family ties which guaranteea minimumsupport. The set-up o f a new pension system has also been instrumental in ensuring a minimum income for the elderly. The 2002/03 HJ3S data can serve as abenchmark to monitor the impact o f the pension system on household welfare. There is also a large literacy gap between youth and older people which hels the disparities in life styles and aspirations across generations: in 2002, while illiteracy affected less than 0.5 percent o f individuals o f ages 24 or less, it affected more than48 percentofindividualsof ages 65 andabove. The Pro-poor Potentialof Public Service Delivery 24. Insuf$cient public spending and low coverage are issues of medium to high importance depending on the sector. Compared to regional average, overall public spending on education is reasonable (6.1 percent o f GDP) but public spending on health (4.5 percent) and social protection (5.8 percent) is low. Within education, a relatively highproportion o f public spending goes to primary education and very little on preschool. The downward trend indonors' financial assistanceto education andhealth sectors also raises concems about the sustainability o f pro-poor public spending inthe social sector. Interms o f coverage o f social services, the overall provision o f public health care services seem to be reasonable, but the provision o f schools at pre-primary and secondary level appears to be inadequate, especially given the very young age o f the Kosovo population. Finally, while pension coverage seems to be high according to recent administrative data, social assistance delivery is very limited in terms o f both overall public spending, amount o f benefits, and coverage o f the extreme poor. While developing a more comprehensive social assistance scheme is unlikely to be fiscally sustainable in the current context, there are ways o f making the system work better for the poor. 25. Besides suf$ciency, the pro-poor potential of public social services may be compromisedby a low effectiveness of service delivery.The ongoinghealth reformprocess in Kosovo has achieved important results but the low quality o f services remains a major impediment for an effective delivery o f health care. The lack o f quality o f health care affects several crucial areas, especially antenatal care, and reflects, among other factors, the inadequacy and inefficiency instaffing allocation. Inthe education sector, concerns about the quality o f schooling result from the presence of teachers with inadequate qualification. Finally, the social assistance scheme appears moderately effective, with little errors o f inclusion but large errors o f exclusion that reflect the cash limit. Overall, the delivery o f public social transfers has a large impact on poverty reduction, but the main incidence is provided bythe basic pensionwhile the impact o f social assistanceis more modest. 26. Although there have been some successes, challenges remain with ensuring an equitable delivery of public social services in Kosovo. In the health sector, equity in service delivery may be compromised by management inefficiency and corruption that cause undue out-of-pocket expenditures and disproportionately hit the poorest. In the education sector, vii while a prioripublic funds allocated to basic education seems to equally benefitpoor andnon- poor children, there seem to be large disparities between the poor andnon-poor inthe quality o f teaching. With respect to access to social assistance, Serb households and households with more than four children appear disadvantagedfor reasons that remainto be explained. Directions for Public Policy 27. To reduce poverty, national policymakers would need to think of developing a multi-pronged strategy reflecting the various dimensions of poverty and the needfor more robust evidence on the nature and the changes of poverty over time. This strategy cmld be articulated around the following five main goals: (i) promoting income opportunities; (ii) raising education and health capabilities; (iii) reducing vulnerability to income and non- income poverty; (iv) empowering disadvantaged groups; and (v) promoting an evidence- based dialogue on poverty. 28. To be realistic, this multi-pronged strategy would need to take into account the context of tight fiscal constraints in which public programs need to be developed. A summary diagnosis of the poverty situation along with policy options is provided inTable 1 as a first input into the on-going building of a broader Kosovo Development Strategy. As such a strategy i s developed, it would be necessary to ensure that it i s properly costed and affordable. This would require proper prioritization and an appropriate sequencing o f the proposedreforms and interventions. 29. Growth-enhancing and efficiency reforms should be at the center of this poverty reduction strategy. This would make the State more effective and provide additional fiscal spaces for public interventions. 30. Theprospectsfor growth and poverty reduction will crucially depend not just on domestic policies but also on decisions made by the international community. Sound domestic policies are crucial for sustaining growth and attacking poverty, but it is equally important to acknowledge that in the short-term the prospects for growth and poverty reduction inKosovo will still be heavily influenced by the level o f donor assistance and the clarification by the international community o f Kosovo's political status. An important concem is that delays in the settlement o f KOSOVO'S status combined with a rapid decline in donor assistance i s likely to have severe adverse social and economic repercussions on the population o f Kosovo. numerous poverty challenges ahead. This support is likely to become more effective if 31. Kosovo still needs the support of the international community to address the embedded within the framework of a Kosovo Development Plan. Given the level o f uncertainty regarding donor assistance, i s it important for national policy makersto undertake the costing of pro-poor public interventions, establish priorities, and identify which interventions will be fmanced from own resources and can be fiscally sustainable, and which ones would needdonor assistance.This costing andprioritization exercise could be developed within the framework o faKosovo DevelopmentPlanas initially developedby ESPIG. (i) Promoting income opportunities 32. Kosovo faces an important poverty challenge in the coming years, requiring a strong commitment by national policymakers to implement a policy program that sustains and accelerates broad-based growth. A major conclusion of this report is that growth with equity is essential for poverty reduction. Prospects for sustained and broad-based growth within the scope of national policymakers will largely depend upon: (i) strengthening the policy environment conducive to private sector led-growth driven by increasing exports and ... Vlll inwardinvestment; and (ii) enhancing social cohesion and ensuringpeace and stability inthe Province. 33. Future poverty reduction will depend crucially notjust on the level of growth but on the extent to which growth will be labor intensive. To maximize the impact on poverty reduction, the links between growth and job creation need to be enhanced. This would require: (i)improving the overall investment climate to attract the much needed investments to revitalize the industry; and (ii) working out fairer trade arrangements and developing advisory services to help small farmers face the intense competition due to largely subsidized agriculture imports. 34. Specific measures need to be undertaken to reduce the relatively large labor market disadvantage faced by youth. To start with, pilot school-to-work and youth entrepreneurship programs could be developed, their implementation evaluated, and scaled- up, ifsuccessful. 35. But theprospectsfor growth andpoverty reduction will crucially depend notjust on domesticpolicies but also on the decisions made by the international community. Sound domestic policies are crucial for sustaining growth and attacking poverty, but it is equally important to acknowledge that in the short-term the prospects for growth and poverty reduction inKosovo will still be heavily influenced by the level of donor assistance and the clarification by the international community o f Kosovo's political status. An important concem is that delays in the settlement o f KOSOVO~S status combined with a rapid decline in donor assistance i s likely to have severe adverse social and economic repercussions on the populationof Kosovo. 36. Kosovo still needs the support of the international community to address the numerous poverty challenges ahead. This support is likely to become more effective if embedded within the framework of a Kosovo Development Plan. Given the level o f uncertainty regardingdonor assistance, i s it important for national policy makersto undertake the costing of key pro-poor public interventions, establish priorities, and identify which interventions will be financed from own resources and can be fiscally sustainable, and which ones would needdonor assistance. This costing andprioritization exercise could be developed within the framework o f aKosovo Development Planas initiatedbyESPIG. (ii) Raising education and health capabilities 37. Reconciling a prudent macroeconomic andfiscal management with theprotection of essential investmentsin basic services and human capital is an important challenge that needs to be addressed. The scope to raise health and education abilities depends very much on the protection o f essential productive investments in human capital as well as in basic infrastructure. Kosovo still has important unmet basic needs inthe areas o f education, health, and social infiastructure services that will have to be properly estimated andbalanced against fiscal consideration. 38. Inefficiency and inequity in theprovision of education and health services deserve specific public interventions. Efficiency gains can be achieved ina number of areas, leading to savings that can be reinvested to improve the overall functioning and effectiveness o f the education and health sectors. At the same time, the low quality o f primary care and the presence of a large number o f teachers with inappropriate qualification deserve particular attention. Further efforts are requiredto reduce undue out-of-pocket health expenditures that are resulting from corruption andinefficiencyindrugprescription. ix 39. To break the vicious cycle of poverty, policy makers need to address the most pressing inequality in education by location, gender, and ethnicity. Improving health outcomes around birth and during thefirst years of life is also essential. The most urgent objectives are to: (i)raise school enrolment at secondary level in rural areas; (ii) improve health outcomes around birth; and (iii) reduce the large inequities in secondary school enrollment rates acrossgender andethnicity. (iii) Reducingvulnerability 40. High environmental health and safety risks need to be reduced through the promotion of a more responsiblegrowth and better environmental management. Prospects for sustainable development and a more environmentally responsible growth would require: (i) inthepowersector;(ii) reforms continuous investments to rehabilitate basic infrastructure services; (iv) implementing sustainable principles in forestry and agriculture activities; and (iii)enforcing environmentalsafeguards. 41, The eradication of extremepoverty should become the major objective of the social safety net. The findings of this report shows that the elimination o f extreme poverty at no additional costs is possible through improvement inthe targetingefficiency and equity o f the nascent social assistance system. Policymakers should rethink the objective o f the public social assistance system, aiming it at the extremely poor, while leaving the situation o f the "non-extreme" poor to be tackled through family-based safety nets and the overall growth strategy. 42. Continuous efforts are needed to support conflict prevention. Policies supporting conflict prevention are usuallymore effective iftargeted on children andyouth. Keymeasures include: (i)promoting tolerance in school curricula; and (ii) involving unemployed and idle youth from different ethnic background inlocal development projects. (iv) Empowering disadvantaged groups 43. Promoting women andyouth influence in both economic, social andpolitical life is essential. Immediate attention should be devoted on a better representation o f women in leadership government positions. Youth and women servicing NGOs should be supported by policymakers- especially inthe areas o fpolitical rights, health counseling and business skills development. Youth organizations should be heavily involved in the building o f youth policies. 44. There is an urgent need to improve the level of knowledge of youth on HIV prevention and to raise awarenesson environmental issues. Influencingthe public debate to reverse the social stigma associatedwith HIV/AIDSi s also essential. (v) Supporting an evidence-based dialogue onpoverty 45. Improving the statistical infrastructure is essential to establishing a poverty dialogue based on robust evidence. In the first place, this require conducting a population census. Important knowledge gaps on key socio-economic outcomes and their distribution across different population groups need to be filled-in within SOK's current survey program. Further attention needs to be paid on ensuring the comparability of income and non-income poverty estimates over time. 46. The monitoring of critical socio-economic outcomes requires unfettered access to household data among government ministries and thepolicy community. Affording access X to non-identifiable household-level data is essential to promoting transparency in poverty analysis, facilitating data analysis, andimprovingthe production o f statistical data. Conclusion 47. Substantialprogress has been made with the reconstruction efforts, yet poverty in Kosovo is still widespread and has many dimensions. This report has shown that the lack o f opportunities, security and participation indecision-making are major aspects o f poverty in Kosovo. Income poverty is large and is driven primarily by the difficulty of the Kosovo economy to generate sufficient viable jobs for its large youth population. Health and education outcomes are still low, and mirror the legacy o f the past as well as the flaws that remain in the sufficiency, effectiveness and equity o f social service delivery. There i s also a large sense of insecurity -both inthe dimension o f income, health, andpersonal safety - that reflects a combination o f factors, including the unresolved political status o f Kosovo, environmental depletion, and violence. Finally, there are large disparities in economic and social well being by location, gender, age, and ethnicity which contribute to a fragile social cohesion and may ultimatelyundermine the sustainability o f economic growth and the impact of growth onpoverty reduction. 48. Looking forward, the building of a home-grown and donor-supported comprehensivepoverty reduction strategv embedded in the Kosovo Development Plan is particularly important. This would help address the numerous challenges ahead in a more coordinated way, and also underpin the implementation o f key development agendas in KOSOVO, including the "Standards for KOSOVO",MDGs, and the development priorities the outlined by the Kosovar government. xi Table 1:Summary o f Poverty Diagnosis and Policy Directions Overall M a i n poverty-related issues Policy directions Coals D In2002, about 37% ofthepopulationwas livingbelowthepovertylineof pc%8 Euro 1.42/ adult equivalent per day. based growth, and increases the contribution of the private D The populationliving below the extreme (food) poverty line of Euro 0.93 sector to overall growth. 3er equivalent adult per day was estimated at 15%. D Joblessness is widespread andcould affect nearly half o fthe labor force. I Women and youth face substantially worse labor market outcomes. D Insome areas, unemployment tends to be higheramong ethnic minority as0 I goups. Make growth more labor-friendly by improving the investment D Income poverty is disproportionately high among non-Serbs ethnic climate and supporting agriculture competitiveness. ninority groups. %I Enhance social cohesion by strengthening the unification o f the social protection system and reducing the most pressing inequality issues by gender, ethnicity andage. Facilitate access to credit. D Only about 4% of children o f ages 3-5 attend primary school. Reconcile prudent macroeconomic and fiscal management with I Only about 73% of children of ages 7-15 complete primary education and the protectionof essential investments inhuman capital. rimary enrolment. Improve the efficiency and equity of public health expenditures D School enrolment drops substantially at secondary level, with a gross by rationalizing the use of drugs, addressing staffing issues, :nrollment rate around 75 percent, and is worse among girls and in rural private sector regulation, andfighting conuption. mas. Health estimates are imprecise, but the overall health status of the population is poor and among the worse in the region. Infant mortality rate couldbe between 18-44/1,000 in2003. Improve the quality and accessby the poor to education. D Health aroundbirth andover the first year o f life is a major problem. I A large share of the population could fall into poverty as the result ofthe Reduce environmental risks through continuous efforts on the risk o f economic slow down relatedwith the unstable politicalsituation. rehabilitation o fbasic infrastructure services andthe The risk o f becoming unemployed for mid-age workers is high with the enforcement ofenvi"!ntal safeguards, and ensurethat prospectso f further restructuring andprivatisation. economic growth is environmentally responsible. D Privatehealth expenses can be catastrophic for some households. D The coverage ofthe extreme poor bythe formal safety net is too low. Adjust the social assistance system as to improve the coverage Environmental pollution and contamination are widespread andrepresents Of the extreme poor. fairly large number.of households to healthrisks. highest risk ofbecoming unemployed. Despite a low incidence, the threat of a rapid HIV/AIDS epidemic is high. Consider the introduction of mitigationmechanism for Kosovo belongs to a major seismic area but its housing stock does not individuals that are likelyto lose their jobs as the result o f seem to comply with the security codes. enterprise restructuring. e Interethnic tensions are high. Support conflict prevention and implementmeasures to fight the D The trafficking inwomen and girls is on the rise. traffickinginwomen and girls anddomestic violence. Domestic violence against women is a serious safety issue. e The unresolved legal status o f Kosovo generates a general sense of Improve the level ofyouth howledge onreproductive health, powerlessnessandvoicelessness. HIV/AIDS, other STIs, and safe sex practice, and facilitate Youth's ability to influence the institutions that affect their lives seems access to supportive services. very limited. Women's influence indecision makingremains at stake. Increase awarenesson environmental issues. e Participation o f ethnic minorities in political life i s challengedby the risk ofnon-participation of Serbs inlocal elections. Corruption is considered a problem o f medium to high importance . depending on the sector. Support the participation of women and youth indecision The level of knowledge of youth on HIV/AIDS and safe sex practices i s making. very low. Environmentalawareness tends to be very limited. The lack o f a recent population census precludes an assessment of the Maintain an active dialogue on poverty within govemment. representative nature o fpoverty estimates. e The household survey system does not allow the comparison of critical Contribute to the mainstreaming o f poverty issues into income andnon-income poverty estimates for the period 2000-2003. government policy. The data necessary to measure non-income dimensions o f poverty i s Improve the social statistical infrastructure and conduct a insufficient. populationcensus. e Access to SOK data remains a problem. Provide accessto SOK data and create a data user group. xii 1. INTRODUCTION 1.1 Kosovo is presently under a UnitedNations' (UN) interimadministration (UNMIK), pursuant to UNsecurity Council Resolution 1244 (1999) of June 10, 1999', supported by a NATO-led Kosovo Force (KFOR), and with increasingly autonomous ProvisionalInstitutions of Self-Government (PISG). Its population o f about 2 millions inhabitants is among the youngest in Europe, with half the population under 25. The population is predominantly Albanian. Other ethnic groups represent less than 12 percent o f the population and are composed predominantly o f Serbs, Slav Muslims, and Roma. Kosovo is also the poorest economy inthe Balkans, with a GDPper capita of approximately US$790 in2003. 1.2 The population o f Kosovo has suffered substantially over the past decade from deep economic, social andpolitical turmoil related with the transition shock inthe early 199Os, the neglect and exclusion o f the Albanian community under the Milosevic era inthe 199Os, and the ensuing armed conflict in 1999. Existing studies indicate that by the end o f the decade, output had more than halved, income had collapsed, less than half o f the population was employed, and halfwas inpoverty (World Bank, 2001). 1.3 Since the end o f the conflict in 1999, the province has benefited greatly from the restoration o f peace and stability, and from massive inflows of donor assistance. The lack o f good and comparable statistics precludes an accurate assessment o f the economic and social changes that took place inKosovo since 2000, but recent studies indicate that progress with reconstruction, macroeconomic stability and the establishment of public institutions have led to a recovery o f economic activity and a revival in the labor market,(World Bank, 2003; World Bank, 2004a). There is also some indication that Kosovo's economic progress may have contributed to an overall reduction in income poverty compared to the immediate post- conflict period. 1.4 The challenge o f poverty reduction inthe years to come remains enormous, however. The recent growth performance has been driven by a post-conflict boom financed by official development aid flows, but it has not led to a significant creation o fjobs and is unlikely to be sustainable without a clarification o f Kosovo's political status and the maintenance o f peace and security for all. A fragile social cohesion, poor environmental conditions, and deficient industrial and basic infrastructures could further undermine altogether the sustainability o f economic growth and its impact on poverty reduction. There are additional concerns that the large pro-poor potential o f public social delivery could be jeopardized by weak governance, low level o f administrative capacity andinadequateresources. Last, but not least, the lack o f a comprehensive and integrated social statistics system in Kosovo could limit the ability o f Government to formulate pro-poor policies andplanfuture investments. 1.5 Today in KOSOVO, building o f a donor-supported and home-grown poverty the reduction strategy (PRS) embedded within the establishment o f a Kosovo Development Plan (KDP) appears particularlyimportant. Not only it could help address the numerous challenges ahead in a more coordinated way, but it could facilitate the harmonization of donors' assistance necessaryfor the implementation o fkey development agendas inKosovo. First, the Government o f Kosovo has agreed to an action plan for implementing the "Standards for 'Hereafter referredto as Kosovo. Kosovo" - a set o f short-term requirements endorsedby the UN Sechty Council that must be met for a final settlement o f Kosovo's politicalstatus - andwhich include the rightfor all to participate hlly and safely in economic, political and social life. The need to meet this. condition requires a close monitoring o f key socio-economic indicators disaggregated by gender, ethnicity and other relevantdimensions and calls for a holistic approach. Second, the formulation o f a PRS would resonate closely with the four broad development priorities outlined by the Kosovar government, all o f which are closely related with the goal o fpoverty alleviation3. Finally, the lay-out o f a strategy to reduce poverty in existing development plans could contribute to advance regional and international development agendas set-out by the European Union (EU) Stability and Association process and the Millennium declaration and to speed-up the progress toward the achievement o f the EU standards and the Millennium Development Goals (MDGs). 1.6 The main objective o f this report is to contribute to the on-going policy dialogue on poverty reduction in Kosovo and to support the formulation and prioritization o f well- informed public interventions. It does not attempt to provide a comprehensive strategy for poverty reduction. Rather it tries to be selective by addressing the following questions: (i) To what extent does the social statistic systeminKosovo allows policy makers to get an accurate pictureofthe poverty situation andto formulate well-informed pro-poor policies? (ii) is What the extent and aspect ofpoverty? (iii) are the poorest groups? (iv) How cohesive is the Who Kosovo society? (iv) I s public social delivery sufficient, effective andequitable? 1.7 The report draws on nine background papers commissioned for this poverty study, based on the analysis o f quantitative and qualitative data. The major sources o f quantitative data include data collected by the Statistical Office of Kosovo (SOK), namely the Household Budget Survey (HBS), the Labor Force Survey (LFS), the Living Standards Measurement Survey (LSMS), and the Demographic and Health Survey (DHS). The report has also benefited from access to the UNDP Human Development Indicators Survey (HDIS), the UNDP and Riinvest Kosovo Early Warning Reports (KEWR), the UNDP Kosovo Mosaic Survey (KMS), the UNICEF Population Services International Knowledge Attitudes and Practices (PSI-KAP) survey, and the U N F P M O M Socio Demographic and Reproductive Survey (SDS). The quantitative analysis was further complemented with qualitative data collected for the study through focus groups andpersonal interviews. 1.8 The key findings o f the backgroundpapers have been summarized in this report and articulated around cross-sectoral issues under the following structure: Section 2 begins with an assessmento f SOK's social statistics system. Section 3 brings together recent evidence on the multi-dimensional nature o f poverty, encompassing opportunities, security, and empowerment, and provides a profile o f the income poor. An investigation o f the level of social cohesion is then discussed in Section 4. The sufficiency, effectiveness and equity of public social delivery are discussed in Section 5. The last section concludes by presenting some directions for public policy that could be integrated in a holistic strategy for poverty reduction inKosovo as part o f the establishment o f a Kosovo Development Plan. The four broad development priorities endorsedby the Government o?ICosovoare: (i) fostering economic development and growth, and increasing employment, through creatingthe conditions for a market-basedeconomy; (ii) improvingthe living standardo f vulnerable groups, and enhancing the quality of and access to education and health; (iii) ensuring an efficient and transparent public administration; and (iv) providing equal opportunities for all citizens. 2 2. THE CHALLENGEOF POVERTYMONITORING 2.1 Attacking poverty requires in the first place the set-up o f a public social statistics program that can help take stock o f the poverty situation and monitor the impact o f growth and public policies onpoverty reduction. Bothincome andnon-income dimensions o fpoverty need to be monitored. Non-income indicators, including data collected to measure trends in health and education outcomes across key population groups may follow those of income related dimensions o fwelfare, butthey maybe less responsive to changes inoverall economic growth or to reductions inincomepoverty. Social statistics should address identified needsby users within Kosovo, but at a minimumshould cover the indicators necessary to monitor the Standards for Kosovo, EU standards, and those used to measure progress towards the MillenniumDevelopment Goals (MDGs). Additionally, social statistics that are derivedfrom household surveys may provide a validity check of data collected from various administrative registries and databases. 2.2 This chapter documents the efforts madeby SOK with the support o f the international community toward the establishment o f a poverty monitoring system in Kosovo. It also discusses some o f the remaining challenges, with a view o f providing an opportunity to improve comparability o f data across various surveys, increase data collection inareas where critical indicators are missing, and improve the overall efficiency in data collection and poverty analysis. The results show that substantialprogress has beenmade with the collection of household level data on consumption and income. Yet, the collection o f data necessary to measure non-income dimensions of poverty, in particular health, educational and environmental outcomes, remains particularly weak. The comparison o f critical estimates across surveys and over time i s also problematic. Another area o f concern i s the absence o f a recent census. As a result, sampling remains a difficult exercise in Kosovo and makes an assessment o f the representativeness o f the various surveys almost impossible. The fmdings o f the report also point to the difficulty o f sustaining a knowledgeable cadre of personnel for poverty monitoring within SOK and the importance o f continuous technical assistance to produce key statistics on economic and social welfare. Finally, data sharing and interactions between data producers and users are found to be extremely limited and further complicate the monitoringo fpoverty inKosovo. A. THESOCIAL STATISTICAL INFRASTRUCTURE 2.3 Summary, An increasingnumber of household surveys i s being collected inKosovo with the objective o f monitoring key social and economic indicators. This section provides a review o f KOSOVO~S social statistical infrastructure, focusing on the household surveys collected by SOK. The review does not cover administrative data sources. The Kosovo household survey program, the representativenesso f the surveys and the comparability o f the data are first discussed. An assessment o f the adequacy o f data collection to measure key poverty outcomes and their distribution across the population, inparticular those necessaryto measurethe progress towards the achievement o f the MDGs and the Standards for Kosovo, is then provided. The fmdings show that substantial progress has been made by SOK in the collection o f household level data on consumption and income, but the lack of a recent population census remains a major impediment to assess the accuracy o f existing data and improve the quality of sampling designs. Moreover, the current household survey systemdoes 3 not allow the comparison of critical income andnon-income poverty estimates across surveys 'and over time, and the collection of data necessary to measure non-income dimensions of poverty i s still insufficient. The KosovoHouseholdSurvey Program 2.4 SOK has been collecting an increasing number of household surveys that are meant to be representative at national level and for urbadrural and the serbshon-Serbs sub population. SOK is currently fielding a Household Budget Survey (HBS), a Demographic and Health Survey (DHS), a Labor Force Surveys (LFS), and an Agriculture surveys (AS). SOK had also participated in2000, along with the World Bank, UNDP, FAO, and IOM, in the implementation o f a Living Standards and Measurement Survey (LSMS) which was used for the 2001 World Bank poverty assessment for Kosovo. The present poverty study relies extensively - but not exclusively - on three SOK household surveys, namely the HBS, the LFS and the DHS. 2.5 The HBS i s an on-going survey and contains extensive and detailed information on household income and expenditures over a one year period. The first round was conducted between June 2002 and May 2003 and was used for this report to construct a welfare aggregate basedon the standardWorld Bank consumption approach to welfare measurement. The household budget survey contains other useful information on educational attainment, labor market status of household heads, and housing conditions. The LFS provide hrther information on economic opportunities and allows the construction and the monitoring o f basic labor market indicators such as labor force participation, unemployment and wages since 2001. The 2001 and 2002 LFS have been extensively analyzed in the 2003 World Bank Labor Market Study and also used for this report. The latest DHS was collected in July 2003 and contains mostly information on school attendance, educational attainment, and reproductive health. 2.6 The sampling designfor the household surveys is identical across the HBS, LFS, and the DHS but slightly different from the LSMS. The sampling design is based on the two-stage design used for generating the LSMS sample. The first stage involves the random selection of enumeration areas and the second stage involves the random selection o f households. The selection of primary sampling units i s conducted once for all three samples (HBS, LFS, DHS). The enumeration areas are selected within ruralhrban and Serbhon-Serb strata for each o f the regions. Inthe rural strata, enumeration areas are delineated villages. In the urban strata, enumeration areas are urban settlements represented by well-delimited parcels o f land. These enumeration areas have been further enumerated and refined since samplingwas conducted for the LSMS. 2.7 The second-stage sampling ensures that different households are covered for each o f the three surveys. Both stages use sampling without replacement: the first stage to increase the efficiency (reduce the variance) o f the sample; the second stage to reduce the response burden of the twelve households selected from each primary sampling unit and for each survey. Given that greater variability inthe Kosovo population is assumed to exist across communities, rather than within communities, the sample for each o f the various household surveys should, intheory, be comparable. Representativeness and Comparability 2.8 In order to develop a comprehensive and representative profile of poverty, it is important that estimates generated from the various household surveys be representative of the Kosovo population andconsidered comparable across surveys and over time. This implies exploring the following dimensions: (i) the overall representativeness o f the surveys; (ii) the 4 comparability o f the population coverage across surveys; and (iii) comparability o f the the indicators across surveys. 2.9 The lack of a recent population census makes sampling particularly difficult in Kosovo and raises the issue of representativeness. The sampling design of the SOK household surveys follows the basic premise o f the LSMS design but the methodology and enumeration o f the primary sampling units has subsequentlybeen refmed with consultations from Statistics Sweden. These consultations reviewed and concurred with the basic elements o f the LSMS sample design, but concluded that physicalenumeration o f the primary sampling unitswas requiredto update the samplingframe. 2.10 While these revisions may have improved the potential representativeness of household-level data collected within the household survey program, they may no longer reflect more recent changes inKosovo's population. Without up-to-date population estimates, it is unlikely that any further changes ineither the design or size of the sample will affect the efficiency o f household survey estimates. Further enhancements or modifications to the householdsurvey program cannot serve as a reliable substitute for a population census. 2.11 Thepopulation coverageseems to be consistentacross surveys in the dimensionsof age and gender, but there are some disparities with respect to ethnicity and ruravurban. Data collected from the various SOK surveys are intended to estimate outcomes from the general population. Since these surveys share the same sampling frame, they should be comparable inprinciple.However, giventhe migration dynamics o f Kosovo's population, it is important to understand whether population coverage is similar across the various survey samples. It is estimated that approximately 800,000 persons have returned to Kosovo since the end o f the conflict. These migration flows could ultimately affect the characteristics and profile o fthe population. Thus, without comparing population aggregates across surveys, it i s unclear whether poverty estimates generated from these surveys can represent any real trend. 2.12 The data in Table 2 represent the distribution o f the sample populations with respect to key demographic variables o f the LSMS, LFS, and HBS at the individual level of detail. The sample design o f each o f the household surveys incorporates urbanandrural location and Serb and non-Serb ethnicity inthe selection o f the primary samplingunits and the respective samples should thus be representative o f these elements. Non-Serb minority groups refer mostly to the Slav Muslims and the Roma4. As mentioned above, the sample design o f the household surveys are similar, but the methodology o f the LSMS was reviewed after its implementation and revised slightly before implementation o f the H B S and LFS surveys. Table 2 indicates that the sample distributions across age and gender remain relatively constant across the three survey populations, but ethnicity and location vary more significantly acrossthe surveys. 2.13 To allow strict comparison of critical estimates across surveys, SOK would need to take into account the differences in variance arising from stratified design effects. The three surveys are based on a similar sampling design inwhich a primary sampling unit o f a village or cluster is drawn from a stratified population o f rural and urban localities, and Serb and non-Serb ethnicities across the five regions in the first stage and 12 households are selected fi-om each Primary Sampling Unit (PSU) inthe second stage. The stratified design affects the variance o f survey estimates and should thus be incorporated into the variance estimation. Without accounting for differences in variance arising from stratified design effects, it is difficult to compare critical estimates across the population. In particular, although ethnicity and locationvary more significantly across the surveys (Table 2) it can not be confirmed whether there is a statistical significance without standard errors that are ~ Inthe HBSquestionnaire, other non-Serbethnicityrefers to the Turkish, Bosnian,Montenegro, AshkalianandRoma. 5 corrected for the sample design effect. Differences in the distribution o f these variables between the LSMS implementationand that o f the other household surveys may be explained by the revised enumeration in the primary sampling units. It is also plausible that the d e f ~ t i o n so f urban and rural locations could have been modified after the revised enumeration. 2.14 Various computer software packages used to process and manipulate survey data have been devised to account for these design effects. STATA is one such package that incorporates information on strata, primary sampling units and survey weights - all o f which are necessary for capturing the effect o f the stratified design. STATA is used widely across the World Bank to generate statistics from household survey data and estimate both income and non-income elements o f poverty. Poverty estimates from the LSMS survey data were calculated using STATA software. 2.15 On the other hand, SPSS, a widely used statistical software package outside o f the B e , accounts only for sample weights and therefore does not accurately adjust for the design effect when calculating sample variances. This means that estimates should not be compared without adjusting for standard errors when stratified design effects are in places5 Given that poverty is estimated currently by SOK fiom the H B S with SPSS software - and that the H B S data are not accessible for processing outside SOK the adjusted standarderrors - are not available for the HBS estimates and thus the LSMS estimates and H B S estimates cannot be strictly compared. 2.16 Population aggregates cannot be strictly compared across the DHS, LFS, and H B S for the same reason, but it i s more likely that these survey estimates are comparable because the first stage o f sampling is the same for each o f these survey populations. Data collection for each o f the surveys occurred more recently and the elapsed time between data collection is relatively short. Changes inthe population occurring from changes inmigration are less likely to have affected differences inthe sample populations between these household surveys. Table 2: Comparisonof Sample Distributionsfrom LSMS,LFS, and HBS YOof Population by Groups LSMS LFS H B S (2000) (2002) (2002) Age 0-15 33.91 32.26 32.35 16-30 28.04 29.09 28.84 31-45 17.1 18.67 18.04 46-60 12.12 11.95 12.07 61-75 7.18 6.46 7.21 76-90 1.59 1.46 1.46 90+ 0.08 0.11 0.04 Gender Male 49.23 50.69 50.9 Female 50.77 49.3 1 49.1 Ethnicity Albanian 85.27 88.1 86.73 Serbian 10.35 6.85 7.89 Others 4.4 5.05 5.38 Location Urban 34.82 37.46 54.83 Rural 64.8 62.54 45.17 A true statistical comparisoni s basedon a t-test that uses standarderrors corrected for the strata designeffect. 6 2.17 Differences between the LSMS and the HBS in the measurement of the werfare indicator used to estimate income poverty makes poverty comparison even more problematic. In both the LSMS and the H B S a consumption measure of welfare i s used to estimate income poverty6.A per adult equivalent measure o f total consumption is constructed for each householdandused to compare households accordingto minimumcaloric and daily expenditurerequirements. These basic requirements constitute the Kosovo food and non-food poverty lines. However, differences between the LSMS and the HBS inthe level o f details for the number o f food items and the time period for recording the expenditures make the poverty figures hardly comparable. 2.18 The LSMS-derived incidence o f poverty estimate is indeed based on a consumption aggregate that contains components o f consumption that were excluded in the HBS. The LSMS collects information on consumption for 46 various food items obtained through purchases, gifts, and home-based production over a 30-day period. The non-food items include the rental value o f housing, services such as basic utilities, health and education, and the imputedrental value o f durable goods. This last category i s omitted inthe HBS. Similar to the LSMS format, the H B S prompts respondents with a long list o f durable goods, but it does not ask for a purchase date (age o f the good) or the estimated purchase price at acquisition. Thus, imputingthe rentalvalue o f durable goods i s impossible with the HBS. 2.19 Inaddition, contrary to the LSMS, the HBS uses a diary format to collect recurrent food and service expenditures over a one-day period. The use o f different formats and recall periods between the two surveys is likely to cause differences in omitted and included expenditures as well as potential categories. Additionally, the LSMS was administered in a few calendar months while the HBS is administered over the entire calendar year. 2.20 The household survey program does not provide consistent information on key socio-economic indicators such as educational achievement, health related expenditures, and informal employment. The measurement o f key indicators, including educational achievement, health related expenditures, and informal employment varies across surveys. As for education, the LSMS includes a detailed education module which encompasses 33 questions on current enrollment, past enrollment, absences, expenditures, distance and mode o f transportation to school. In contrast, the education component inthe HBS, LFS, and the DHS surveys has been reduced to between one to four questions only. The parsimonious modules do not reveal much about educational achievement or potential constraints for staying in school. Moreover, differences in the wording o f the education questions across surveys would leadto different estimates on enrollment status. 2.21 With respect to health, the LSMS asks detailed questions about health expenditures for public and private in-patient and out-patient services. Within these services, costs are further disaggregated by prescription, transportation, fee, and g;lfts or voluntary payments. Frequency o f visits i s also recorded for each type o f service. Health expenditures reported by HBS respondents are captured in the one-day recall diary and they are not captured at all in the DHS survey because respondents are not asked about health visits. The DHS survey focuses on contraceptive knowledge, practices, preferences for number and gender o f children, and live birth histories. Differences in the formats used to elicit information on health expenditures betweenthe LSMS andthe H B S will cause differentestimates. The HBS format i s likely to result inan under-reporting o f expenditures, especially voluntary payments and related transportation costs. Householdconsumption is preferredto household income as a measure of welfare because households are likely to smooth consumption over time while income sources may vary seasonally. 7 2.22 There are also large differences in the measurement o f informal employment. Definitions applied in the HJ3S and LFS to estimate informal sector activity appear to be inconsistent with I L O definitions. Inparticular, ILO size restrictions for delineating formal and informal sector business activity are not used inKosovo. Inaddition, detailed questions on transport ownership and construction activity could presumably pick up activity in areas where self employment is fiequent but they are not includedinthe LFS survey. The LSMS takes a more meticulous approach to eliciting information on any and all income generation activities than either the LFS or the HBS. Additionally, the timing o f the LFS survey may have led to under-reporting o f self-employed agricultural work. Differences intimingbetween the LSMS and the LFS could be partially responsible for large differences in the unemployment estimate between the two surveyperiods. GapsinDataRequirements 2.23 Critical gaps in data collection undermine Kosovo's capacity to measure the multidimensional aspects of poverty, the poverty linkages and the causes of poverty. The limited coverage o f education, health, employment and environmental indicators in the surveys i s especially problematicbecause critical socio-economic issues are present ineach o f the related sectors, andbecause it undermines the calculation o f key indicators to monitor the MDGs andother desirable indicators. While Table 3 below provides an assessment of the gap in data requirement to measure the MDGS, a more comprehensive list of desirable and measured indicators is presented inthe appendix inTables 38-42. 2.24 Available questions on education included in the household surveys make it most dij-$cult to monitor primary completion rates and secondary school enrollment. School enrollment data compiled by administrative records indicate that school attendance varies by ethnicity and gender. Yet, administrative data can not be usedto calculate enrollment rates o f girls and boys as the denominator is not collected insuch data. Additionally, administrative education data provides no information on the source o f constraints to access and participation. While education enrollment and school attendance rates are usually obtained fiom household surveys, the current household survey system in Kosovo does not allow monitoringsecondary school enrollment. 2.25 The educational attainment questions inthe HBS andthe LFS use indeedonly general categories o f highest level completed, currently attending primary school, did not complete primary school, no school, and literacy, and make it impossible to construct estimates on primarycompletion rate and current secondary school enrollment. These gross categories do not enable tracking o f school attendance by grade neither, while this could be helphl in understanding the critical years or periods when parents determine whether to extend educational investments in their children. From a policy perspective, its important to understand how education investment behavior within the household may be influenced by either government education policy, including prices established for tuition or learning materials, or more specifically by characteristics o f local schools. 2.26 The lack o f community level data in current household surveys covering aspects o f the curriculum, teacher-student ratios, teacher absenteeism, and distance to the communities serviced h t h e r prevent obtaining important insights into the determinants o f school enrollment andattendancedecisions made by households. 2.27 Collection of data that are necessaryfor monitoring trends in healthfacility access, usage, and outcomes are missing within SOK household survey system. Although some small-scale surveys have been conducted to estimate underlyinghealth trends, the scope and sample size o f these surveys have produced unreliable estimates in critical health indicators andthe absenceof such estimates from SOKhousehold survey datahas ledto a largerange in 8 published estimates. Additionally, it is assumed that births and deaths occurring outside of hospitals have been historically under-reported and administrative registries are unreliable sources for the generation o f such estimates. Given the unreliability o f public registries, survey data i s essential for estimating important trends and for checking reliability o f these registries. 2.28 Imprecise estimates from small scale surveys further show that Kosovo exhibits high infant and maternal mortality rates - some o f the highest rates reported in Europe. This suggests that critical gaps exist in the coverage, access, and quality o f prenatal health care. Yet, detailedanalysis within thehealthsector on access, use, satisfaction andquality ofhealth care that could explain the observed outcomes cannot be undertaken within the current household statistics system. 2.29 Currently the HBS contains some cursory questions on health expenditures, but no information on health status. Moreover, DHS surveys are conducted infrequently in Kosovo and do not provide all the necessaryinformation. A typical DHS is implementedwith a large sample size and covers a range o f health topics. Data collection on a large sample size yields the possibility to estimate trends on rare events such as infant and matemal mortality. Unfortunately, the latest Kosovo DHS conducted in 2003 incorporated neither o f these aspects. Key information on health status, health service quality, and direct and indirect costs associatedwith obtaining health care and that usually contributes to low utilization or lack of affordability, are missing. The sample size was apparently too small for producing reliable estimates on infant mortality and the range o f topics were restricted to contraception prevalence and family composition preferences. While contraception use i s important from a gender perspective, as it may help to explain the sex birthratio (the male-female gap i s larger thanthe global average), andto understandriskybehaviors, such as unsafe sex practices, it is not sufficient for keepingabreastof important health indicators. 2.30 The lack o f community level data on health services in SOK surveys further prevent the understanding o f health outcomes, access to health services, and consumer satisfaction with the quality andcoverageo fhealthservice delivery. 2.3 1 Thepresent available information on employment, unemployment, and laborforce participation across surveys does not allow generating reliable trends. The dynamics underlying unemployment and employment trends across gender, ethnicity, and locality are critical for understanding the critical linkages between labor market activity and equitable economic growth prospects. Inparticular, labor market participation in rural areas and in general among women is critically low. Yet it i s difficult with the present available information on employment, unemployment, and labor force participation across the LSMS, the HBS and the LFS to generate reliable trends. Further information on labor market activity and the standardization of the timing when surveys are administered is necessary to understandthese dynamics. Informal sector employment activity may thrive inKosovo as it is assumed across the Balkans but it i s universally under measured. 2.32 There are limited data to monitor the environmental MDG indicators and track the health-environmental linkages. Some o f the environmental h4DG indicators can only be measuredthrough community level data and are currently unavailable through SOK survey system. Land use issues and carbon dioxide emissions are includedas MDG indicators which must be measured at the community level by those possessing technical knowledge on measurement. Moreover, relevant environmental issues may be dependent on the location. Landusage is more likely to be a pressingrural issue, whereas carbon dioxide emissions are more likely to be an urbanproblem. Additionally, the linkages between health outcomes and environmental conditions such as inadequate and untreated sewage, the burningo f wood for cooking fuel inside homes, waste disposal, sanitationpractices, andthe sources o f water used bythe householdcannotbemonitoreddue to the absence o fahealthmodule inthe HBS. 9 Table3: Opportunities and Challengesto Measurethe MDGsinKosovo Goal 1: Eradicateextreme poverty and hunger Indicators: NationalPoverty headcount ratio; Poverty gap ratio Share o f poorest quintile innational consumption Prevalence of underweightchildren Proportion of populationbelow minimum level of dietary energy consumption Source andcurrent availability: The poverty headcount and ratio are currently estimated from the 2000 LSMS and the 2002 HBS. Due to comnarabilitvissues a trend cannoi be esiimated but with 2003 H B S Dove& trends can be estimated between 2002 and2003. The prevallnce o f underweightchildren canbe estimated from the UNICEF k I C iurvey. Goal 2: Achieve universal primary education Indicators: Net enrollment inprimary school . Proportion o f pup& start&g grade 1who reach grade 5 Primary completionrate Literacy rate o f 15 -24 year olds Source andcurrent availability: Data on enrollment is collected within the Ministry of Educationmonitoring system, but it is not reportedby age which is required to estimate net enrollment. Because ofthe absence o f current census data, it's difficult to estimate total populationo f school-age children andthus calculate enrollment rates. The current household surveys collect information on education activity for members o frespondent households but neededdetail is lacking. Goal 3: Promote gender equality and empower women Indicators: Ratio o f girls to boys inprimary, secondary, andtertiary education Ratio of iterate women io men-15-24 ye& old Share o f women inwage employment in the non-agriculturalsector Proportion o f seats held by women innational parliament Source and current availability: Enrollment ratios can be calculated using the Ministry o f Educationdata. The education questions in the HBS don't capture information on current secondary or tertiary school enrollment. Literacy ratios have been computed from the DHS data, but the age range is too broad for this indicator. Female employment rates are calculated from the Labor Force Survey. However, some discrepancy exists over the estimated rates from the survey data because informal sector activity is underestimated from the 2002 survey. Goal 4: Reduce child mortality Indicators: Under-five mortalityrate Infant mortality rate Proportion o f 1 year-old children immunized against measles Source and current availability: The sample size ofthe 2003 DHS is too small to estimate reliable mortality rates for either infants or children. Immunization data are currently not collected inthe DHS survey. Provided accurate andup-to-date birthregistry, measles immunizationrates of 1 year-old children could be estimated with public healthdata. Goal 5: Improve maternal health Indicators: Matemalmortalityratio Proportiono fbirths attendedby skilledhealthpersonnel Source andcurrent availability: Matemalmortalityrates canbe estimatedwith up-to-date birthand deathregistries, however, deaths resulting from matemal causes are often misclassified. Birthregistries may overestimate the number of births attendedby skilled personnel ifbirths that are not attendedby skilledpersonnel are less likely to be reported. Household surveys are sometimes usedto estimate matemalmortality but often don't produce reliable, current results. The 2003 DHS sample is too smallto estimate such Condomuse rate o fthe contraceptive prevalence rate Condom use rate at last high-risk sex Percentageo f 15-24 year old populationwith comprehensive knowledge o f HIV/AIDS Contraceptive prevalence rate Number o f children orphaned by HN/AIDS Prevalence anddeath rates associatedwith tuberculosis Proportionof tuberculosis cases detected and cured under directly observed treatment short course Source and current availability: Contraceptive prevalence rate can be monitored fiom the 2003 DHS. Prevalenceof HIV and TB should be monitoredthroughthe public health information system because clinical diagnosis is required. Goal 7: Ensure environmental sustainability Indicators: Proportiono f land area coveredby forest Land area protectedto maintainbiologicaldiversity GDP per unit o f energy use Carbondioxide emissions (per capita) Proportiono fpopulationusing solid fuels Proportiono f population with sustainable access to an improvedwater source Proportiono f people with access to improved sanitation Proportiono f people withaccess to secure tenure Source andcurrent availability: Access to improved water source, sanitation, and securetenure can be monitored throughthe HBS. Other indicators must bemonitoredat the community levelthrough the Ministry o fthe Environment andSocial Plannine and the - - MinistryofAgriculture andForestry. Goal 8: Develop a global partnership for development Indicators: Unemployment rate o f 15-24 year olds Source andcurrent availability: The unemployment rate of 15-24 year olds can be estimated fiom the available Labor Force Survey 10 B. INSTITUTIONAL ARRANGEMENTSAND CAPACITY FOR POVERTY ANALYSIS 2.33 Summaiy. The previous section has shown that substantial progress has been achieved by SOK toward the implementation o f surveys that are meant to be nationally representative. Yet, some important challengesremained interms o f designingrepresentative samples, undertakingpoverty comparison, and measuring non-income dimensions o f poverty. Inthis section, the discussion onpoverty monitoring is extended beyondthe social statistical infrastructure and includes an assessment of the level o f SOK's capacity for poverty analysis andthe extentto which institutionalarrangementsare conducive to anefficient use of existing poverty data. The discussion shows that sustaining a knowledgeable cadre o f personnel for poverty monitoring is a challenge. Limited data sharing and interactions between data producers and users are also important impediments for analyzingpoverty. 2.34 SOK has embraced the standard World Bank approach to income poverty measurement. The World Bank has beenproviding technical assistance to SOK to produce income poverty estimates based on SOK households surveys on consumption and income. Past poverty analysis undertakenjointly by SOK and the Bank had relied on the 2000 LSMS and were published inthe 2001 World Bank Kosovo Poverty Assessment. New estimates of income poverty have been constructed using the 2002/03 HBS and are presented in the following chapter. 2.35 While measuring income poverty i s not exempt from problems, SOK with the assistancefrom the World Bank has followed the standard World Bank methodology usedfor computing estimates o f consumptionpoverty. These estimates are based on consumption data collected from the HBS and adjusted as neededto account for differences across households infamily compositionandfor differences inconsumer pricesbetweenurbanandruralareas. Two absolute poverty lines referring to extreme poverty (the food poverty line) and poverty (the complete poverty line) have been constructed using the structure o f expenditures o f the HBS7. Three commonly used indicators of poverty are subsequently reported in this report: thepoverty headcount index, the poverty gap andthe severityofpoverty (Box 1). Box 1:The Three Standard Foster-Greer-Thorbecke (1984) Aggregate Poverty Measures ~ The headcount index (PO) i s a measureof the prevalence of poverty. It denotes the percentage of households who are poor-as defined by the poverty line-as a proportion of total population. This measure is insensitive to the distributionof the poorbelow the poverty line. The poverty gap index (Pl) is ameasureof the depthof poverty, andit denotes the gapbetweenthe observed consumptionlevels of poor households and the poverty line. Assuming perfect targeting of resources (transfers), thispoverty gapindexindicatesthe total amountneededto bringall householdsinpovertyup tothe povertyline. The poverty severity index (P2) measures the degree of inequality in distribution below the poverty line, giving greater weight to householdsat the bottomofthe consumptiondistribution. wheren= totalpopulation, q =number of those with consumptionyi less thanthe poverty line z. 2.36 Sustaining a knowledgeable cadre of personnel for poverty monitoring is a challenge. Officials from SOK have emphasized that capacity for data analysis among staff is limited if non-existent. The most recent poverty profile was intended to be produced as a collaborative effort between SOK staff and an intemational consultant fmanced by the World 'For more details about the construction of the poverty estimates, see Tsirunyan, 2004. 11 Bank. The consultant allocated many hours to training staff on the basics of data cleaning, preparation, and analysis. Unfortunately, many of the staff who benefited from the time o f the consultant either left the department or SOK altogether. Thus, although individuals have benefited from technical support, SOK has not been able to develop and sustain a knowledgeable cadre o fpersonnel for poverty monitoring and analysis. Low salaries for local staff in the statistical office are indeed a major barrier for attracting and retaining qualified staff, While ongoing training is critical inthe short-term to maintain SOK's capacity, systemic incentivesfor qualified staff retention can only be addressedin the context of a longer-term programfor civil service reforms. 2.37 SOK remains in great needs of technical assistance in the short-term. Despite obvious difficulties to raise SOK's capacity inpoverty analysis through technical assistance only and without a more holistic approach, the continuation o f Bank technical assistance remains critical in the short-term. SOK has been requesting additional support for the next round of consumption data analysis, This will allow cleaning and analyzing the new rounds o f the Household Budget Survey (Hl3S) that i s being collected and that will permit the comparison o f poverty over time. Additional support to SOK may also be needed so as to support the improvement o fthe H B S for poverty analysis (Box 2). Box 2: Improvingthe HBS for Poverty Analysis One option to improve the H B S for poverty monitoring would be to introduce a core H B S survey with rotating modules, This may offer a holistic approach to monitoring multi-dimensional aspects of poverty that is inherently more sustainable and preferable to anLSMS-type option. The incorporation o f rotating modules inthe H B S with core survey questions to estimate the various non-income dimensions o f poverty may require various sample sizes. Therefore, the sample size should be adjusted to reflect the indicators being measured within the rotating modules. Inparticular, the measurement of non-frequent events, such as child and matemal mortality, indicates that the sample size must b e increased for more precise measurement. Particular emphasis within rotating modules in the H B S should include the coverage o f critical socio-economic issues and areas where knowledge is lacking. Specifically, SOK may wish to develop rotating modules to: (i) improve the existing education module (enrollment, school attendance, related expenditures, school quality); (ii) introduce a health module (health outcomes, access to, and use of health services); (iii) improve the questionnaire on the possessiono f durables to enable the imputation o fthe rental value o f durable goods (age or purchase date of the durable good, the purchase price at the time of acquisition); and (iii) the incidence and prevalence of monitor gender-based violence. The particular rotating modules in the H B S would also need to be complemented with community-level questionnaires to measure the quality and coverage of community service delivery (school and health facilities) and some o fthe environmental h4DGindicators (land use, carbon dioxide). IP 2.38 Limited data sharing and interactions betweendataproducers and users complicate the underlying analysis necessaryfor poverty monitoring. The Statistics Office maintains a comprehensive and up-to-date website o f the office's most recent publications and statistics generated primarily fiom the HBS, LFS, and DHS (www.sok-kosovo.org). In fact the office posts statistics and publications on the website without delay. Draft chapters based on the analysis ofthe DHS data were posted on an `as ready' basis. 2.39 Unfortunately, data sharing does not go beyond the SOK website. The statistics office enforces a strict policy on data access. No data is ostensibly allowed to leave the premises o f the office. SOK enforces the rule so that it can obtain training on data manipulation and analysis for its SOK staff from the interested data user. As discussed earlier, however, this practice has not been very effective to sustain SOK's capacity for poverty monitoring. The parsimonious attitude toward sharing o f household survey data has also complicated the underlying analysis necessary for this poverty report. Unfortunately the lack of data sharingi s endemic across the various ministries. These entities also do not share data and often do not 12 releaseit to SOK. Such attitudes amongthe government implythat most data is not used.The various ministries produce data solely for their own needs. Much o f the data, therefore, is under-utilized. 2.40 Data access policies vary across the region. In Eastern Europe and Central Asia, countries with an open access to survey data include Albania, Bosnia and Herzegovina, Estonia, Georgia, Hungary, the Kyrgyz Republic, and Moldova. The World Bank fully supports access to individual records o f household surveys that have been stripped of identificationnumbers and information. The sharing o f such data can improve the capacity for poverty monitoring, lead to efficiency gains in poverty analysis, and contribute to greater policy coherencewithin the Government. 13 3. THE MANYDIMENSIONS OF POVERTY 3.1 Attacking poverty further requires a thorough understandingo f the multidimensional nature and causes o f poverty that will help establish a comprehensive poverty reduction strategy in Kosovo. It i s now widely recognized that poverty encompasses the lack o f opportunities, security and empowerment. The importance o f assessing the various dimensions ofpoverty early on inthe life cycle is also gaining increasingrecognition with the better understandingthat risks are not homogeneously distributedalong the life cycle and are typically higher in the early stages of life, with important, long-term and sometimes irreversible consequenceson later stages o f life. 3.2 This chapter starts by bringing together recent evidence on income and non-income dimensions of poverty and documents how different aspects o f poverty interact andreinforce each other. It also takes a detailed look at relevant dimensions o f child and youth poverty with the objective of informingpolicy makers on age-specific risks and opportunities, and on the "right" timing and nature o f interventions. The findings point to the lack of opportunities, security and participation in decision making as the major aspects o f poverty in Kosovo. Incomepoverty appears widespread butrelatively shallow. There are no accurate estimates o f the evolution of poverty over time, but tentative interpretations suggest that income poverty has most likely declined in the immediate post-war period. Non-income poverty i s also widespread and takes the form o f low achievements in education and health and a large vulnerability to poverty - including high income and physical insecurity and widespread environmental healthrisks. Voicelessness andpowerlessnessare also important dimensions o f poverty resulting from deficits in the functioning and accountability o f state institutions, and the patriarchal structure o f private institutions. Often - but not always - these many dimensions o fpoverty seemto interact andreinforce each-other. A. LIMITEDINCOME OPPORTUNITIES 3.3 Summary. Identifyingpoverty with a lack o f economic opportunities, as measuredby low levels of consumption or income, has a long tradition. The previous chapter has shown that while it is not exempt fi-omproblems, there is a standard World Bank methodology used for computing estimates of income or consumption poverty. This sections focuses on the income dimension o f poverty in KOSOVO, on consumption data collected through based household surveys and on the definitiono f two nationalpoverty lines for poverty and extreme poverty. This sections starts by presenting the macroeconomic background. It then provides several indicators o f aggregate welfare and income poverty and examines in more details the profile of the income poor and the sources o f income. Finally, it presents a tentative assessment of the impact of recent economic performance on poverty. This sections finds that the reconstruction efforts have succeededto put Kosovo on a growth path since 2000. Income growth with moderate inequality most likely led to a reduction in poverty, but the lack o f accurate data preclude an accurate assessment o f the evolution o f poverty between 2000 and 2003. Four year after the end o f the conflict, income poverty remained widespread but relatively shallow. Income poverty was found to affect disproportionately children, the elderly, female-headedhouseholds, the disabled, non-Serbs ethnic minorities, the unemployed andprecariousjobholders. 14 T h e Macroeconomic Background 3.4 The 1990s have been characterized by a deterioration of all economic indicators. The macroeconomic situation inheritedin1999hadbeen damagedbypoor economic policies, broken external trade and financial links, international sanctions, a lack o f investment, and ethnic conflict. Data for the 1990s are limited but it is estimated that during this period all economic indicators deteriorated. Output fell by more than 50 percent inthe early 1990s and by another 20 percent following the conflict at the endof the decade. Inflationwas very high throughout the 1990s as aresult o f monetary financingof fiscal andquasi-fiscal deficits. 3.5 Since the end of the conflict in 1999, the reconstruction efforts have succeeded in putting the Kosovo economy on a growth path. Estimates of macroeconomic data for the post-conflict period need to be treated with great care. Due to the weak statistical infrastructure inKosovo, economic data are subject to frequent revisions. The latest estimates compiled by the IMF and the Kosovar authorities in2003, and reported inthe 2004 Kosovo Economic Memorandum (World Bank, 2004), indicate that there has been a recovery o f economic activity andpositive growth since 2000 (Table 4). GDP growth reached 21 percent in2001 andhassince averaged4.3 percent. GDPpercapitaincreasedfrom aboutUS$400in 1995 to US$790 in2003. The recent growth performance was largely drivenbyhigh levels o f public and private investments reflecting massive inflows o f donor assistance after the cessation o f hostilities in 1999 and the shift toward liberal market policies. Gross National Disposable Income (GNI) per capita was evenhigher at about US$975 in2002 and US$1170 in 2003, reflecting the importance of remittances. The Euro became the de facto local currency and price stability was achieved, with zero inflation in 2003 compared to double digits inflation rates in 2000. Current account deficits were reduced thanks to improved budgetary performance and rising private savings. Government revenues increased remarkably, from about 17percent o f GDP in2000 to 45 percent in2003, and ledto abudget surplus averaging 3.5 percent o f GDP during 2001-2003. Table 4: Key MacroeconomicIndicators - _p__ Year P a 2000 2001 2002 2003 2004 National accounts Real GDP growth 21.2 3.9 4.7 6.5 Investment(% GDP) 35.6 40.0 40.7 38.0 45.3 Gross domestic savings (% GDP) -198.6 -114.7 -99.4 -85.2 -71.3 Government balance (%GDP) Revenue 17.2 28.0 38.9 44.7 45.4 Expenditure 31.5 24.4 33.5 42.5 50.8 Balance (excl. grants) -14.3 3.6 5.4 2.2 -5.5 Donor grants 21.5 7.3 3.0 2.2 0.0 External accounts (YOGDP) Exports o f goods 0.5 3.0 7.4 8.2 21.7 Imports o f goods 156.4 113.7 108.3 99.2 117.4 Workers' remittances 73.7 52.0 42.9 43.2 42.5 Current account before grants -174.6 -110.2 -96.5 -76.9 -69.8 Savings and investment(% GDP) National savings 44.9 38.5 14.8 7.6 7.1 Investment 66.9 57.7 48.3 40.5 45.3 Monetary aggregates(% GDP) Broad money 99.6 78.5 70.9 75.4 74.5 Bankcredit to private sector 0.4 2.2 6.8 16.1 19.8 Prices CPI inflation 11.7 3.6 0.0 0.0 Source: World Bank Kosovo Economic Memorandum(2004). Note: a 2000 to 2003 data are estimates _. . , and 2004 data are projections; excluding donor-financed spending; excluding donor-related imports. 15 3.6 The business environment and the investment climate have improved considerably, but remain adversely affected by political risks, delays in reforms, and asymmetry in trade agreements. With the shift toward liberal market policies since the establishment of the United Nations Interim Administration Mission in Kosovo (UNMLK) in 1999, and the Provisional Institutions of Self Government (PISG) in 2001, the climate for trade and investment has also improved quite substantially. The former trade regime was replaced by a simple regime with no quantitative barriers and a single 10 percent tariff rate. Findings from the 2003 Investment Climate Assessment concluded that a sound legal framework for a free market economy has largelybeenput inplace, and that the business environment inKosovo was generally seen by domestic small and medium f i i s as more favorable than in many countries in the region. However, major constraints in the business environment and investment climate included: (i) the unreliability o f power supply which i s seen as a major barrier for doingbusiness and a consequenceofthe lack o freform inthe power sector; (ii) an unfair and informal competition, inparticular the absence o f reciprocal treatment under trade agreements within the region which gave competitors an unfair advantage; (iii) highpolitical uncertainty; and (iv) delays with the privatization process. 3.7 The recent growth performance has not been labor intensive growth. Despite an increase inprivate sector activity since 1999, the recent growth performance inKosovo has not translated into a significant creation o f decentjobs. This, combined with a large youthful labor force, has translated into high unemployment, especially among youth. While most o f existing fonnal jobs are found inthe service sector, and are closely reliant on the presence o f a large expatriate community, job opportunities inthe industrial and agricultural sectors have beenvery limited (World Bank, 2003d). The lack of a strong link between economic growth and job creation has been attributed to: (i) the difficulty to attract FDInecessary to revitalize the industrial andminingsector; and (ii) the intense competition to domestic producers in the agricultural sector inducedby agricultural - and largely subsidized imports. - 3.8 Sustaining the recent growth performance remains a serious challenge. Despite substantial progress, Kosovo faces the challenge o f sustaining robust growth in the coming years. The recent growthperformance was driven by a post-conflict boom financedby official aid and is unlikely to be sustainable. Growth rates have already declined from 21.2 percent in 2001 to a yearly average o f 5 percent during 2002-2004, reflecting the overall reduction in donor assistance, as well as Kosovo's difficulty to establish a sound business environment and attractive investment climate in the absence of a more predictable political and legal environment. Since 2000, donor grants have fallen by 70 percent and will most likely continue to decline. Prospects for income growth, as identified in the KEM, will largely depend on a number o f related factors: (i)the maintenance of peace and security; (ii) the speedo fresolving KOSOVO'Sstatus; and (iii) implementation o f a set o f reforms that legal the promote private sector led growth. The latter includes reforms in the power and mining sectors, and the support o f agro-processing and commercial farming. The Extentof IncomePoverty 3.9 Kosovo is thepoorest economy in the Balkans. With a GDP per capita o f US790 in 2003, Kosovo i s also one o f the poorest areas in south-east Europe and the poorest economy in the Balkans. In 2002, GDP per capita was indeed US$ 1,537 in Albania, US$1,924 in Serbia and Montenegro, US$1,860 inMacedonia, and US$1,362 inBosnia and Herzegovina (Figure 1). 16 Figure 1:GDP Per Capita in Selected Economies of South-eastEuropein2002 (incurrentUS$) 12,000 1 11,202 10,000 8,000 6,000 4,000 2,000 0 Source:World Development Indicators. Note: GDP data for Kosovo are basedonpreliminary estimates andrefer to 2003. 3.10 Absolute income poverty in 2002 appears widespread but is relatively shallow. World Bank estimates of poverty are presented in Table 5 according to different poverty indicators. The construction o f the absolute poverty lines in summarized inBox 3'. The 2002 H B S data suggest that, four years after the end o f the conflict, poverty remains widespread in Kosovo. In2002, about 37 percent o f the population o f Kosovo was living below the poverty line of Euro 1.42 per adult equivalent per day. However, the size o f the poverty gap (the average distance of the poor fi-om the poverty line) and the severity of poverty (the weighted sum of poverty gaps) indicate that poverty was not extremely deep. It was estimated that among the poor, the average incremental consumption amount necessary to escape from poverty was only 11percent o f the poverty line (Table 5). The conclusion o f large but not deep poverty i s reinforced by a relatively flat distribution o f consumption, with an estimated Gini coefficient of 30 percent. Extremepoverty remains nonethelessan importantproblem in KOSOVO,still about 15percent ofthepopulation livingbelow the extreme (food) poverty with line o f Euro 0.93 per equivalent adult per day. Table 5 Poverty EstimatesDerived from the 2000 LSMS and 2002 HBS - Data Year Overall ' Food Poverty Poverty Severity of Extreme (food) source poverty poverty Headcount Gap poverty poverty line line (YO) (YO) (YO) headcount (YO) HBS 2002 1.42 /day O0.93Iday 37.0 11.4 4.9 15.2 Source: 2002 HBS T h e Profile of the Extreme Poor 3.11 Designing an effective poverty reduction strategy requires a good understanding of which socio-economicgroups in society suffer the mostfrom incomepoverty and how large they are. Table 6 presents the extreme poverty indices (extreme poverty headcount index) * At the time of writingthis report, no international povertylines were constructedbecauseofthe lack of PPP indicesfor Kosovo, andhencethe comparison of poverty in Kosovo with other countries difficult was not attempted. 17 across socio-economic groups and their share intotal extreme poverty (the extreme poverty share), using per-adult equivalent measuresg. While the results presented inTable 6 are based on`univariate correlates and use a per adult equivalent scale, most o f them tend to be relatively robust to the choice o f different equivalent scales and the use o f multivariate analysislo. 3.12 Extreme poverty is disproportionately high among children and the elderly. In KOSOVO,incidence of extreme poverty is the highest among children (17.3 percent among the pre-school children and 16.9 percent among children o f age 6-14) and the elderly (17.1 percent), and the lowest among working age adults (13.3 percent). Given the young age o f the Kosovo population, chldren o f ages 0-14 also constitute a large share o f the extreme poor (34.5 percent). 3.13 The presence o f children inthe households is also a strong correlate of poverty. What i s interesting i s that the positive association between children and poverty i s true whatever equivalent scales are used, but the incidence of children on poverty is stronger when per capita rather than per adult equivalent measures are used. With a per adult equivalent scale, the extreme poverty headcount rises from 10.7 percent among families with no children to almost 17percent for families with three or more children. Box 3: Constructionof the Absolute PovertyLines I I Two absolute poverty lines were constructed, jointly with SOK, usingthe 2002/03 HBS data: a food or extreme poverty line, estimated at 0.93 /day; and (ii)a complete poverty line,taking into account the need for non-food consumption, and estimated at l.4l/day. The estimation o f the poverty line followed the same method utilized earlier by the world Bank and SOK for calculating the poverty lines usingthe 2000 LSMS.These absolute poverty lines are based on the cost of basic needs approach. A minimal daily calorie intake of 2100 calories/person/day is adopted. The cost of the food basket providing this nutritional level is then estimated, using consumption habits o f the third-to-fifth deciles o f the population and the price information contained in the HBS. This leads to the "food poverty line", and i s treated as athreshold for identifying extreme poverty. Clearly, non-food basic needs should be accounted for as well. The cost of non-food basid needs is estimated on the basis of calculating the food share for households with per capita food consumption in the neighborhood o f the "food poverty line" (between the food poverty line and 1.2 times the food poverty line). The calculated food share is 65.9 percent. The complete poverty line i s then calculated as the sum o f the food component (the food poverty line) and non-food component, where the food poverty line corresponds to 65.9 percent of the complete poverty line. 3.14 Households with disabled members have a higher incidence of extreme poverty. Like in many other regions in the world, in Kosovo the presence o f disabled members is associated with higher poverty rates. The extreme poverty headcount increases from 14.3 percent among households with no disabled members to 17.9 percent among households with one or more disabled members. 3.15 Female-headed households are much poorer than male-headed households. While female-headed household accounts for only 4.7 percent o f all households inKosovo, their risk o f extreme poverty (28.2 percent) is much higher than for male-headed households (14.6 percent). Female-headed households also represent a small, but non-negligible share of the Per-adult equivalent measures are usedto account for differences across households in family composition and the economies o f scale. See Tsirunyan (2004) for more details. lo Multivariate analysis provides more precise estimates as to the role o f selected socio-economic characteristics on poverty, taking into account the impact o f other factors that are also associated with poverty (see Verme, 2004). 18 extreme poor (8.7 percent). The reasons for a higher incidence o f extreme poverty among female-headed households are discussed in Chapter 4 and explained by gender differences in households composition, education andlabor market status. 3.16 Ethnic groups other than Albanians and Serbsface a strong disadvantage. Interms o f ethnic groups, extreme poverty is significantly higher among groups other than Albanian and Serbs with a combined extreme poverty incidence for these groups o f 31 percent. While the HBS does not allow for a more detailed classification by ethnicity, other ethnic groups refer altogether to the Roma, Slav Muslims and other minority groups. Given the population structure, Albanians constitute the majority o f the extreme poor (86.1 percent) but the risk o f being poor among 'other' ethnic groups is more than two times higher than among the Albanians or the Serbs. Serbs show extreme poverty indexes slightly below that o fAlbanians. 3.17 Multivariate analysis confirms a significant disadvantage among other ethnic groups, whose probability o f being extreme poor is higher by 14.6percent. Chapter 4 provides M e r insight as to the reasons for higher incidence o f extreme income poverty among non-Serbs ethnic minorities by looking at the disparities in educational and employment outcomes by ethnicity. 19 Table 6: Extreme PovertyRates by Selected Socio-economicGroups Extreme Poverty Index YOof the extreme poor Risk Au 15.2 100.0 100.0 Age of individuals 0-5 17.3 13.0 113.8 6-14 16.9 21.5 111.1 15-24 15.8 22.4 108.3 25-64 13.3 36.2 87.9 65+ 17.1 6.9 106.1 Gender of head Female-headed 28.2 8.7 185.5 Male headed 14.6 91.2 96.0 Presenceof children( 4 5 y.0) 0 10.7 16.6 70.4 1 13.6 16.1 89.5 2 16.4 17.0 107.8 3 or more 16.9 50.3 111.1 Presenceof disabledmembers 0 14.3 74.5 94.1 1 or more 17.9 25.5 117.8 Educationof householdhead None: can't readwrite 20.4 8.1 137.4 None but can readwrite 17.9 3 120.4 Attending primary school 15.6 21.9 107.6 Uncompleted primary school andnot attending 19.5 5.3 131.8 Primary 17.6 41.4 119 Secondary 9.6 18 65 Vocational 6.9 1 46.9 University or higher 7.2 1.2 48.6 Labor status of householdhead Employers 9.6 1.2 63.3 Employees 9.6 19.0 63.2 Subsistence farmers 16.2 12.6 106.2 Per diemworkers 21.4 11.0 140.9 Other self-employed 10.0 1.8 65.7 Retireadisabled 16.2 34.1 106.8 Unemployed 22.8 15.6 150.2 Housekeepers 28.1 3.5 184.8 Others 12.1 0.8 79.6 Ethnicgroup of householdhead Albanian 14.6 86.1 96 Serbian 13.9 5.8 91.3 Other 31 8 203.1 Location Prishtina 7.7 3.1 50.2 Other Urban 19.1 26.6 125.4 Rural 14.8 70.2 96.9 Region Gjakova 13.8 8.4 90.6 Gjilani 10 5.5 65.9 Mitrovica 24.2 24.2 158.9 Peja 22.6 15 148.2 Prizreni 15 13.9 98.7 Pristina 7.3 16.3 47.8 Ferizaji 28.9 16.6 189.6 Source:HBS, 2002. Note: The risk index i s computed as the ratio o fthe poverty rate among a particular group to the averagepovertyrate inthepopulation. 3.18 Extremepoverty affects disproportionately lower educated headed households. The education level o f the household head is a strong correlate o f extreme poverty, with secondary education being the main threshold. While the incidence of extreme poverty is above average for all households whose heads have completed only primary school or less (ranging from 20 15.6 percent to 20.4 percent), it falls substantially for those with secondary education and more (from 7.2 percent to 9.6 percent). The vast majority o f the extreme poor is made of households whose heads have only completed primary education o f less (79.8 percent), reflectingthe overall low educational standards o fthe population. 3.19 Joblessness is a strong correlate of extreme poverty, and so are precarious jobs. Using the labor market classification reported inthe HBS, one can see that housekeepersand unemployedheads are the groups with the highest extreme poverty indices (28.2 percent and 22.9 percent respectively). Everywhere in the region unemployment i s a strong correlate of poverty, and it is particularly pronounced in countries with weak unemployment benefit schemes. In Kosovo, the fact that there is no targeted unemployed assistance do not exists contributes to accentuate the adverse impact o f unemployment on poverty. The incidence o f extreme poverty i s also very high among per diem workers (21.5 percent) who do not enjoy the same working conditions as regular employees. 3.20 Extreme poverty is the highest in seconday cities and exceeds rural poverty. The extreme poverty headcount is the highest in urban areas outside the capital Prishtina (19.1 percent). Extreme poverty rates are somewhat smaller in rural areas (14.8 percent) but still two times higher thaninthe capital city (7.7 percent). While the incidence o f extreme poverty i s higher in secondary cities, the vast majority o f the extreme poor live in rural areas (70.2 percent) as do most o f the population in Kosovo. The finding that the poorest are found in other urbanareas remains unchangedeven when other correlates o fpoverty are controlled for, andis robust to the use ofnationalversus urbdruralpricedeflators. 3.21 The region of Ferizaji has the highest extremepoverty rate, followed by Mitrovica and Peja. Regions with the highest poverty indexes are Ferizaji (28.9 percent), followed by Mitrovica (24.2 percent) and Peja (22.6 percent), and about half o f all the extreme poor are located inthese regions. These fmding are robust to the use o f conditional probabilities and point to a strong regional dimension o f poverty in Kosovo despite the relative small size o f the province. 3.22 Thepoverty profile has remainedpractically unchanged since 2000. Comparing the findings ofthe 2002 HBSwith the previous World Bank analysis of income poverty based on the 2000 LSMS suggests that the profile of poverty has remainedrelatively stable over time. In 2000, the incidence of extreme poverty was already higher inurban areas, among non- Serbs ethnic minorities, among families with numerous children, and among the unemployed. However, the situation o f female-headed households seems to have deteriorated in relative terms: in2000 the ratio o f the female-headed to male-headed extreme poverty rate was only 1.1,as opposedto 1.9 two years later. 3.23 The 2002 profile of incomepoverty that emergesfrom the HBS is also consistent with other recent quantitative poverty profile. A World Bank labor market study (2003) investigates the determinants of income poverty, using the 2002 LFS, a relative concept o f poverty and income rather than consumption as a measure o f welfare. This study finds that income poverty for the unemployed is much higher than for the average population, and that being underemployed or employed inprecarious work and unemployment duration increases the probability o fbeingpoor. Among the employed, the povertyrisk is lower for employment inthepublicsector andhigherforprecariousworkerswhere salariesarelikelytobelower. 3.24 Worse living conditions among the elderly, female headed households, families with children, the disabled, and the unemployed is also supported by qualitative evidence. The qualitative poverty study has identified the elderly, female headed households, the disabled, and the unemployed as the most vulnerable groups in society and thus echoes the quantitative findings (see Box 4). 21 3.25 Manyolder respondentsinthe qualitativestudy mentioned the low levelo fpension as a source of poverty (in2002 the average pension was about 6 times lower than the average wage) and some even report that they do not receive a pension at all. As for single mothers, they usually talk about their difficulties to combine work and domestic duties, and complain about the adverse consequences o f traditional views that value women only as mothers and housewives. For the disabled and the unemployed, the problem is often their incapacity to work and/or to find jobs, and the fact that they have to rely mostly on social aidbut that it is not enough to make ends meet. The situation describedby many families with a large number of children is somewhat different but not better, as many report the most difficulties to meet the increased income andnon-income needs associatedwith their parentresponsibilities. 3.26 What is also interesting i s that respondents have mixed views about the extent o f poverty in rural andurban areas. Often, it is arguedthat life i s easier inthe cities and that the urban population has more opportunities. On the other hand, it is also recognised that villagers can more easily fulfiltheir basic needs as they can rely on home produced food. ' Box 4: Voices of Vulnerable GroupsinKosovo Society should do more to help the elderly. (A.D., 67 year old male, Albanian, Ferizaj) IhavealwaysbeenmorefragilethanotherwomenweresinceIhadnohusbandandIwasahidsomeone would say something bad about me. (H.B.,70 year old female, Albanian, Mitrovica) The state should care more about single mothers. Icannot go to forest to get some fire wood; Imust pay someone to bringit to me and to cut it. (Leposavic) IhaveneverworkedbutnowIevencan'tbecauseIundenventasurgeryfortwotimes.(N.N.,47yearold female, Bosnian, Peja) First thingw e do when we get our Social Benefit is to secure the flour and salt, and also the cleaning products for the children and clothing; the rest is neglected. Ican never manage to buy some new clothes for my children. (P.P:, 40 year old female, Bosnian, Peja) Ithinkthatlifeinthevillageisbetterbecauseyoucanhaveacowandtogetmilk,sourmilkanditiseasierto provide food. However, it is easier to find a job in;own than it is invillages.(k.M., 46 year old female, Albanian, Gjilan) Life is easier in village, if you have a land and are able to plow it. But if you have children and want to educate them it is easier to do this intowns, because there are no travel expenses. It i s easier to findjob in towns. (M.H-,43 year old female, Albanian, Gjilan) When the war started inKosova, the Serbs displaced us from the village and destroyed all of our wealth that we owned init. Our house was burned down to the base and so was our store with all its equipment inside. My husband's family found shelter at a nearby village while Iand my family came to live with my own parents in Gjakova.(H.M., 26 year old female, Albanian, Gjakova) Nowadays, my family has worse living conditions than previously. It is harder for us to live inthe city since we don't have any choice o f work, while when we used to live in the village, we could at least plow the land and grow cattle.(Z.R., 57 year oldmale, Albanian, Gjilan). 3.27 Qualitative evidencefurther points to a high incidence of poverty among internally displacedpersons -an issue that is not well captured in recent quantitative surveys. The qualitative study, andother small scale surveys, finds that internally displaced persons (IDPs) remain among the most disadvantaged households in Kosovo. One group o f IDPs is composed o f ethnic Albanians who lived inrural areas before the war and had been forced to look for shelter incities after the conflict started. While agriculture and cattle breedingused to provide them with decent income before the war, most o f them are uneducated and under- qualified and can not easily find a job in the cities (Box 4). They usually have to turn for assistance to the state social services and humanitarian organizations. Another specific problem of this group i s that they have lost their pre-war homes andare still facing difficulties 22 infmdinghousingaccommodationsincities. They oftenhave to rentrooms inother's houses. Another group o f IDPs include ethnic minorities displaced after the war, andwaiting for their re-integration in Kosovo. According to a 2004 survey conducted by the IDP information Center on IDPs in the North of KOSOVO, the biggest problem faced by respondents was accommodation, followed by the return of property, lack o f employment andlegal issues. And the vast majority o f these IDPs were willing to return to their place o f origin if security and freedom o f movements were guaranteed. Sources of Income 3.28 For all incomegroups, incomefrom work is the main source of incomefollowed by private remittancesfrom abroad. Inpoor countries, income from work usually represents a very large share o f total income. Labor is usually and understandably the only valuable asset that poor people have. InKOSOVO, because o f very highunemployment and low employment, income from work i s less important, although it remains the first source o f income estimated at almost 60 percent for all households and 42 percent among the extreme poor (Table 7). A disproportionately high level o f income derive from cash remittances from relatives and friends, mostly from those living abroad (15.2 percent), and this is true for both the extreme poor (15.5 percent), the poor (14.5 percent) andthe non-poor (15.4 percent). Table 7: Income Structureby Poverty Status and Location(percent) - - AU Households Very-poor Poor Non-poor Urban Rural Prishtina Cashwages and salaries net o ftax 59.9 41.7 55.9 61.2 66.5 54 87.2 Wages inkind 3.1 3.5 3.4 2.9 0.8 4.2 0 Rent, dividends, interest 0.8 0.2 0.5 0.9 1.9 0 3.9 Social welfare benefits 1.7 9 4.5 0.7 2.1 1.6 1.1 Pensions 6.2 9.4 7.1 5.8 4.5 1.3 2 Cash remittances from Kosovo 2 3.7 2 2 3.1 1.6 2.6 Cash remittances from abroad 15.2 15.5 14.5 15.4 11.4 18.2 2.4 Game o f chanceilottery 0.1 1.1 0.3 0 0 0.1 0 Other income or transfer 11.2 15.9 11.8 11 9.7 13 0.9 Total 100 100 100 100 100 100 100 Source: HBS 2002. Note: Labor status categories refer to the head o f the household only o f ages 16-65. 3.29 For the extremepoor, state transfers constitute the third largest source of income. For the extreme poor, a very highproportion o f income is constituted by transfers from the state through pensions and social assistance (18.4 percent). The combination of these two types o ftransfers comes close to the share of income from private transfers from relatives and friends living abroad or in Kosovo. This finding is consistent with the 2003 World Bank Kosovo labour market study, which fmds that about 25 percent o f households reported to receive cash or non-cash assistance from friends and relatives living abroad, about 10 percent reported receiving cash or in-kind assistance from friends and relatives living in Kosovo and 26 percent declared to have borrowed money from friends and relatives (which is evidently notperceived as assistancebyall respondents). 3.30 Income diversijication and own produced food are important coping mechanisms. Available evidence from the HBS data also indicate that most households in KOSOVO, regardlesso f the labor market status o f the household head, rely on several sources o f income (Table 8)' '.The only households which basically rely only on one source o f income are those headedby an employee, but these are also the least poor andprobably those inthe lowest-risk activity (see Table 4). Consumption data from the HBS hrther indicate that own produced food constitute an important source of subsistence. The imputed value of home food production is estimated at about 14.3 percent of total household consumption on average, and ~~ Labour status is defined as the main activity performed over the previous 12 months which evidently captures what respondents regard as the main activity. The question on sources o f income instead may actually include income that was generated by the different types o f activities performed inthe course o f the year. 23 16.3 percent among the first consumption quintile. It i s thus evident that for the poor, relying on multiple sources o f income and home production o f food are important coping mechanisms. Table 8: LabourStatusand Sources of Income Sources of Income Labour Status Pub.Wages Priv.wages Farming Per diem HHbus. Pensions Remit. Other Tot. Employers 14.5 51.0 0.0 1.5 29.5 0.0 1.9 1.7 100 Employees 72.7 21.7 1.9 0.1 1.4 0.1 1.4 0.7 100 Subsistencefanners 6.9 8.5 73.2 2.0 1.1 1.o 6.3 1.1 100 Per diemworkers 2.9 8.9 2.7 75.1 1.3 0.5 4.4 4.3 100 Otherself-employed 2.6 35.0 0.0 0.0 55.1 0.0 0.0 7.4 100 Retireddisabled 30.9 15.3 8.6 10.4 3.5 16.2 9.9 5.4 100 Unemployed 22.2 14.7 3.7 6.8 4.5 4.4 17.6 26.2 100 Housekeepers 30.4 13.0 5.3 15.7 3.1 1.5 11.8 19.3 100 Others 15.4 28.4 3.6 3.6 2.4 3.6 2.4 40.8 100 Total 35.9 18.0 10.5 10.9 5.5 , 5.5 6.9 6.9 100 Source: HBS 2002 Economic Growth and Poverty Trends 3.31 Has the growth peiformance of 2000-2003 been pro-poor? Since the end o f the 1999 ethnic conflict, the province has benefited greatly from the restoration o f peace and stability, and from massive inflows o f donor assistance.'h this context there are good reasons to believe that income poverty may have declined. However, while the growth-poverty linkages can be simulated from survey data, which helps illustrate the importance o f both growth and equity for poverty reduction (see Section C inChapter 3 and Section A inChapter 4), the actual links cannot be empirically tracked duringthe period 2000-2003 for the reasons discussed below. 3.32 The lack of comparable poverty data precludes an accurate assessment of the changes inpoverty that tookplace between 2000 and 2003. A number o f surveys have been undertaken in Kosovo by different institutions to derive poverty figures, but so far none o f these estimates are strictly comparable. The previous poverty estimates were based on the 2000 LSMS and reported inthe 2001 World Bank Poverty Assessment (World Bank, 2001). Available World Bank estimates point to an overall poverty rate o f 50 percent in 2000 based on the LSMS and 37 percent in 2002 based on the HBS. Correspondingly, the poverty gap, severity o f poverty and extreme poverty were estimated respectively at 15.7 percent, 6.8 percent and 12 percent in the LSMS, and 11.4 percent, 4.9 percent, and 15.2 percent in the HBS. As discussed inChapter 2, however, the 2000 LSMS andthe 2002 HBS are not entirely comparable due to differences inconsumptionmodules andthe use o f survey-specific poverty lines. Only after future rounds o f the HBS have been undertaken and analyzed, can one make more defmite statements about the trends inpoverty andextreme poverty. 3.33 At best one can infer that overallpoverty hasprobably declined. The steep decline shown by the overall poverty estimates would suggest that poverty has probably declined between 2000 and 2002. The difference is indeed sufficiently large to assume that even accounting for the statistical errors associated with the poverty estimates, one would still b e able to conclude that overall poverty has declined. 3.34 But no inference can be made about the evolution of atremepoverty. The poverty gap, the severity o f the poverty index and the extreme poverty headcount are rather close between the two years and do not show a consistent trend across the three measures. It is 24 possible that changes inthese measures reflect the sensitivity o f these estimates to the use o f different welfare aggregate and slightly different poverty lines, rather than real changes. Given the comparability problems listed above, there are not sufficient elements to argue for a positive or negative trend inextreme poverty. B. LOW EDUCATION AND HEALTH CAPABILITIES 3.35 Summaiy. Besides the lack o f income, education and health capabilities are among the primary dimensions o f individual well-being, and ones that are well reflected in the Millennium Development Goals (Table 9). This section complements the previous discussion by providing an assessment of poverty that goes beyond low levels o f consumption to encompass low achievements in education, health, and environment. The findings show that important progress were made in education achievements since 2000, but the level o f completed education was still relatively low inKOSOVO, compared to the region. As regard health, there is still considerable uncertainty on estimates, but available evidence points to very poor healthoutcomes which also the worst inthe Balkans. Table9: Kosovo and the MillenniumDevelopment Goals MDGs Current Situation inKosovo Projections by 2015 1. Halve between 1990 and 2015 In2002, about 37% of the population With an annual average growth the proportion of people in extreme was living below the pove$-line of rate of 2% and no change in poverty and who suffer form Euro 1.42/ adult equivalent per day. inequality, per capita is estimated hunger The population living below the that extreme poverty may be extreme (food) poverty line of Euro reducedto less than 8 percent 0.93 per equivalent adult per day was estimated at 15%. 2. Achieve universal completion of Only about 73% of children of ages 7- Data from the 2003 HDIS primary educationby 2015 for both 15 completedprimaryschoolin2001. indicate that while overall rates boys and girls In 2001 and 2002, the ratio of girls to are high, rates for ethnic groups boys inprimarywas 0.92. other than Albanian or Serb are much lower. There is a possibility that the target will be met. 3. Eliminate gender disparities in School enrolmentdrops substantially at Gender disparities are greater primary and secondary education secondary level, especiallyamong girls after primary school, especiallyin by 2005 andfor all levels by 2015 and in m a l areas. The ratio of girls to rural areas where 29% of girls boys in secondary was estimated at compared to 61% of boys finish 0.72 in2001 and0.79 in2002. secondary school. The target is unlikely to be met. 4. Reduce under five mortality Under-five mortality rates are not Although no data on U5MR is rates by two-thirds between 1990 available. WHONNICEF estimated available, Kosovo suffers from and2015 IMR at 35/1000 for both 2000 and one of the highest IMRs in ECA 2001. and is therefore unlikely to meet the MDGtarget. 5. Reduce matemal mortality rate Data on MMis conflicting. According With the limited data available, it by threequarters between 1990 to UNFPA MMR was 153 per 100,000 is hard to tell if the MDG will be and 2015 while MOH dataindicates 12. met. 6. Halt and reverse by 2015 the HIV infection rate appears low. Since Based on available data, Kosovo spread of HIV/AIDS, malaria, and 1986, only 60 HIViAIDS cases have appears to be on track to meet other diseases beenreported. TB notification rate was AIDS target. 78 per 100,000 in2001. 7. Ensure environmental In 2001, the proportion of population Much of the water supply in sustainability and access to and with access to improved water source Kosovo is contaminatedand there improved water source was estimated at 68 percent of the are not waste water treatment population (not taking into account systems with only 28 percent of water &om wells). homes connected to the sewage system. 25 EducationalOutcomes 3.36 The overall level of completed education is relatively low. The education system in Kosovo has undergone substantial shocks during the 1990s. Discriminatory practices against the Albanians encouraged this group to pull out o f formal education and to create a parallel educational system. Later, the war disruptedformal education. As a result, the overall level o f completed education appears today low inKosovo, compared to the general achievements of other neighbouringeconomies. According to the 2002 HBS, over halfo f the population inage 26-65 has primary education or less andonly 5.2 percent have ahigher education degree. 3.37 Illiteracy is still significant in Kosovo although declining. The level o f illiteracy increases very rapidly with age. According to the HBS, while the level o f illiteracy is about 0.5 percent among the population younger than 26, it increases to almost 49 percent among the population o f ages 65 and older. Yet, although estimates are not entirely comparable across surveys, it seems that illiteracy is declining, with a rate o f 6.5 percent in 2000 (based on the Demographic and Socio-Economic and Reproductive survey (SDS) conducted by UNFPA and IOM), 5.9 percent in2002 (based on the HBS), and 5.8 percent in 2003 (based on the HDIS). 3.38 Few children in Kosovo attendpre-school institutions. While pre-school education is usually not compulsory and i s designed to prepare children for school, recent evidence shows that preschool programs improve the prospects for physical and cognitive development thereafter. In Kosovo, the pre-school educational system is quasi inexistent and leaves early child care in terms o f health and educational provision almost entirely on parents. Before the war, pre-school education used to be divided into kindergartens (age 3-5) and pre-primary education (age 5-7). Yet, only a minority o f childrenhad access to pre-school institutions and these were located close to the factories to serve workers' children. 3.39 The war and the transition process dismantled the system. With the recent reform of the educational system, pre-primary education hasbeenabolished and absorbedinto a formal pre-school cycle that covers children from the age o f 3 to 6. The most recent data on pre- school enrolment rates .are for 2001 and shows that at that time, there were only 7,343 children in Kosovo (about 4 percent of children in age 0-5) in pre-school education spread over 34 institutions (Table 10). Most of these institutions were in the Prishtina region and virtually all inurban areas. Table 10: Selected Key EducationalIndicators in Kosovo 2000 2001 2002 2003 Adult illiteracy rate(agel5+) 6.5' 5.ga 5.8' % o f adults (age 26-64) with primary education or less 53.3 a 43.7d Pre-school enrollment rates (age 3-6) 4.0e Primary enrolment rate (age 7-15) 91.7' 90.5 a 95.4d Rate o f children completing primary education 73e Ratio o fgirls to boys inprimary 0.92' 0.92' General secondary enrolment rate (age 15-19) 5 9 3 75.2 Ratio o fgirls to boys insecondary 0.72' 0.79' Girls reaching grade 9 58.1' Higher education enrolments 17.3' Ratio of girls to boys intertiary - 0.79 ' 0.82 ' Source: 'HBS; bLFS, ILO unem loyment rate for adulthood refers to age 15-64; 'Demographic and Socio- economic reproductive survey; HumanDevelopment Indicators Survey; eSOWMESTAJNICEFGender review; a fSOK;,gper thousand live births, early neonatalmortality rate (first week) from UNICEF Survey on prenatal care; DHS; 'World Bank estimates; 'RIINVEST; kUNICEF estimates; 'Kosovo Obstetrician and Gynaecology 26 Association; %omen 15 to 49 years attendedby skilled health personnelat delivery, based on UNICEF-MNSS; 'at 1year, based onNIP= estimates; OLSMS. Note:HBS-based illiteracy rate for 2002 refers to individualsof age 7 and above. 3.40 School enrolment is high at primay level but enrolment drops substantially at secondary level Kosovo enjoys high enrolment rates in basic education. According to the HDIS, 95.4 percent o f children o f ages 7-15 were currently enrolled inprimary education in 2003. Although surveys are not strictly comparable, this seems to be an improvement over the recent years. In2002, according to the HBS, the enrolment rate inprimaryeducation was 90.5 percent, and about 91.7 percent in 2000 according to the SDS (Table 10). From an international perspective, these enrollment rates are high: the average primary school enrollment rate (net) is 79.5 percent in low income countries, about 91 percent in middle income countries, and93 percent inupper middle income countries. 3.41 However, the share o f children enrolled insecondaryeducation drops significantly. In 2003, according to the HDIS, the proportion o f children o f ages 16-19 currently enrolled in secondary education was only 75.2 percent. This i s much better than in 2000, where the secondary enrolment rate was estimated at 59.5 percent by the SDS, and reflects the reconstruction efforts, but it nonetheless indicates that improving access to secondary education remains an important challenge in Kosovo. Problems in service delivery include insufficient space and classrooms, with schools working on a 2 to 4 shifts per day. Health Outcomes 3.42 Although there is still considerable uncertainty on estimates, the overall health status of the population in Kosovo is poor and probably among the worst in south-east Europe. Traditionally, infant mortality rate (IMR) and life expectancy (LE) for males and females are the indicators that best describe the overall health status o f a population. In KOSOVO, are major differences in the currently available estimates for both these there indicators. For LE, SOK, based on the 2003 DHS provides an estimate o f 67-71 years for females and 63 to 67 for males, which is significantly lower than available estimates for neighbour countries. Current IMR estimates vary from 18 to 49 per 1,000, which would be the highest in SEE and from 4 to 10 times (depending on the estimate) higher than the EU average (Table 11). The comparison o f these health indicators across countries i s not exempt of problem, however. These are based on different data sources and the differences may reflect the disparities indata collection and surveys, rather than true differences. Table 11: Infant Mortality Rate (IMR) and Life Expectancy(LE) inKosovo as Comparedto Regionaland EUData - Countries IMR LEatbirth LEat birth Males females Kosovo (2003) 18-49 63-67 67-71 Serbia &Montenegro (2001) 13.3 69.7 74.8 FYROM(2001) 11.8 68.9 74.9 Bosnia& Herzegovina(2001) 7.6 69.3 76.4 Albania (2001) 11.6 66.3 73.2 EU(2000) ' 4.5 75 81 Source: WHO, Atlas of Health in Europe, 2003 and SOK. 3.43 Health around birth and over thejirst year of life is a major problem. There are only rough estimates o f perinatal mortality, infant mortality, under-five mortality, maternal mortality, and o f the incidence o f vaccine preventable disease, but available figures indicate that the health situation around birth and over the first year o f life is particularly bad in KOSOVO,amongtheworst inEurope (Table 12). and 27 3.44 Inadequate nutrition is a particular problem that affects a fairly large numher of children. According to the micronutrient survey carried out by UNICEF in2002, low height for age is present in 10percent o f children 6-59 months, low weight for height in4 percent o f children, and a highweight for height in8 percent. Sixteen percent o f children have mild and moderate anemia and 36 percent Vitamin A deficiency. Low urinary iodine is present in 50 percent o f school age children and iodine deficiency is severe in 14 percent o f them. Breastfeeding (BF) i s common practice in Kosovo, but for a short duration and with a low proportion of infants being exclusively breastfed. This is an important issue since formula feeding is expensive, cow's milk is nutritionally inadequate, while exclusive breastfeeding for the first six months provides the greatesthealthbenefits at the lowest cost. 3.45 The factors behind the high maternal and infant mortality rates are fairly well identifled. Weak prenatal care, along with illegal abortions, are important contributors to maternal mortality and morbidity including permanent reproductive health problems. Abortion is still a common fertility control method inKosovo and there is anecdotal evidence that the private sector performs illegal late abortions after ten weeks. Causes o f death within the neonatalperiodinclude prematurity (59 percento f all childdeaths), infections (9 percent), asphyxia and birth injuries (32 percent). Beyond the prenatal period, infections and particularly respiratory infections are the first cause o fhospitaladmission anddeath. Table 12: SelectedMaternal and Child HealthOutcomes inEurope and CentralAsia - P RegiodCountry Infant Under- Maternal Child TBC DPT Polio mortality five mortality birthcare immunisation immunisation immunisation rate mortality ratio rate rate rate rate (4) (1) (2) (3) (5) (6) (7) 1999 2001 2001 2001 2001 2001 2001 Kosovo -35 35-40 21* 95 98.5* 97** 91** Czech Republic 4.6 5 3.3 99.9 98.5 99.2 97.2 H w a r y 8.4 9.4 5.2 99.6 100 99.9 99.9 Poland 8.9 9 3.5 99.8 94.9 98.2 97.7 Slovakia 8.3 8.2 15.6 99.4 88.7 Slovenia 4.5 4.7 Estonia 9.5 10.9 7.9 99.7 99.3 75.5 65.5 Latvia 11.3 13.7 25.4 99 96.5 93.9 94 Lithuania 8.6 10.8 12.7 100 99.3 94.7 91.5 Bulgaria 14.6 15.1 19.1 98.9 97.8 93.6 94.3 Romania 18.6 21.9 34 96.5 99.6 96.5 96.7 Albania 12.3 16.9 Bosnia-Herzegovina 7.6 8.7 9.5 99.9 99 91 91 Croatia 7.7 9.2 2.4 97 94 94 FYRMacedonia 14.9 12.9 14.8 97.6 96.8 90.6 91.5 Serbia andMontenegro , 13.6 15.3 6.9 Belarus 11.5 11.6 14.2 99.9 99.3 99.2 99.2 Moldova 18.2 20.3 43.9 99 99.6 97.1 98.1 Russia 16.9 18.3 36.5 99.3 96.6 95.8 96.7 Ukraine 12.8 14.4 18.1 99.9 98.1 99.1 99 Armenia 15.4 18.8 21.8 96.4 94.5 96.8 Azerbaijan 16.5 24.8 25.4 99.6 98.5 98.2 95.8 Georgia 17.5 13.3 66.8 96.7 96.4 89.3 82.8 Kazakhstan 20.7 25.1 50 98.5 Kyrgyzstan 22.7 29.5 43.8 98.7 98.8 98.9 98.8 Turkmenistan 25.4 36.9 9 98.2 98.7 99.3 Uzbekistan 20.2 33.5 97.9 98.5 99.4 Source: WHO, Unicef Kosovo and UnicefTransMONEE database.. (*) 2002 (**) 2003. (1) Probability o f dying betweenbirth andone year o fageper 1,000livebirths.(2) Probabilityo fdyingbetweenbirthand five year o f age per 1,000 live births. (3) Annual number o f deaths from pregnancy related causes per 100,000 livebirths. (4) Proportion ofbirths attended by skilled personnel. (5) Percentage o f one year old children immunised against tuberculosis. (6) Percentage o f one year old children immunised against diphtheria, pertussis and tetanus. (7) Percentageo f one year old children immunised against polio. 28 3.46 The causes of death within the neonatal period are largely preventable and there are signs of improvements.A substantial proportion o f death within the neonatal period can be avoided with appropriate interventions in the prenatal and immediately postnatal period. Data fiom the 2000-2002 WHO/"ICEF Survey on perinatal care show indeed a decreasein the early neonatal mortalityrate (ENMR) from 14.8 per 1,000 live births in2000 to 12.6 per 1,000 live births in2002. This i s largely due to interventions that have taken place since the end o f the conflict to improve quality o f care through training and better facilities and equipment. 3.47 Quantitative information on the health status later on in the life stages is very scarce. Yet, available data suggest that tuberculosis and disability are major problems. There are no available quantitative data that could inform on the health status o f older children, youth, adults and elderly people in Kosovo, nor systematic data on chronic and infectious diseases and risk factors. Available data on the incidence o f HIV/AIDS indicate that while Kosovo remain a low level epidemic region, tuberculosis and disability remain major public healthproblems inKosovo. 3.48 Although the incidence o f tuberculosis (TB) has decreased in the aftermath o f the conflict (fiom 128/100,000 in2000 to 53/100,000 in2003), it is still far above the average of western European countries (10/100,000), and similar to the highlevels inthe former Soviet Union. Moreover, the majority o fnew caes occur at a relatively young age. 3.49 The recording o f HIV/AIDS in Kosovo started in 1986. Since then, 60 HIV/AIDS patients have been recorded, but the real number o f individuals infected by the virus is likely to be higher. The majority o f the patients have been males between 30 and 39 years o f age. By the end of 2002, twenty-two patients had died. Inthe absence of a systematic STI-HIV surveillance system, the number o f HIV infected persons is not known. Government statistics report 1095 registered STIs cases duringthe period 1990-2000. 3.50 Disability, as a consequence o f injury and chronic illness, is also a major problem, due to the consequences o f the conflict and o f past neglect o f prevention, care and rehabilitation. Available estimates show that between 5 percent to 7.5 percent o f the overall population have some kind o f disability, and up to 50 percent o f the population over 65 suffers from limitingillness and disability. Many disabled children are still institutionalized. 3.5 1 Qualitative data further indicate that mental health problems are widespread, especially for young people. There i s no official data on the prevalence o f mental health problems. Immediately after the conflict, the focus was on those traumatized during the conflict. It was estimated that 62 percent o f those surveyed had been close to death, 49 percent had been victims o f torture or abuse, 42 percent had been separated fiom family members, 26 percent had experienced the murder o f a family member or fiiend, 10 percent had been imprisoned, and 4 percent had been sexually abused. It is safe to assume that the consequences o f conflict are still largely present in KOSOVO, not only in terms o f post- traumatic stress disorder (PTSD), but also because o f the impact o f the loss o f parents, sons, relatives for many o f which a mourninghealing process could not take place because o f the manywho disappearedandfor whom there still is no information. 3.52 Qualitative data also indicate that mental health i s a serious problem for young people inKosovo, due to a combination o f factors: (i) the psychological sequel of the conflict; (ii) falloftraditionalwaysofcontainingand/ormaskingpsychiatricproblems; (ii) the the high risk for social exclusion (see Section D); (iii) the general frustration about non- and solution o f status quo. But there are no official data on the prevalence o f mental health problems among young people inKosovo. 29 c. POVERTYAS A LACK SECURITY OF 3.53 Summary. As articulated by the people themselves in KOSOVO,experience of the poverty is also a sense o f vulnerability to a reduction in different dimensions o f their well- being.This sectionprovides an assessment ofmajor risks occurringat national, community or household level inKosovo and that could adversely affect individuals inthree dimensions o f well-being: income, health and personal safety. The analysis shows that a large share o f the population could fall into income poverty or remain into poverty as the result o f a still unstable political situation and the related risk o f economic slow-down. Exposure to health risks i s widespread, largely resulting from environmental pollution. Finally, the lack of personal safety appears as an important dimension o f poverty that is not solely related with inter-ethnic tension. Income Insecurity 3.54 A fairly large level of income insecurity and vulnerability to income poverty in Kosovo i s the results o f a combination o f inter-related factors, includingthe risks o f economic slow down and political instability that can affect the entire population, the environment degradation that canjeopardize the means o f livelihood o f the many households who rely on natural resources, an overall high risk o f unemployment and precarious employment, the existence o f constraints to internal mobility that can reduce the capacity to mitigate income shocks, the occurrence o f unexpected private health expenses that can be catastrophic for some households, and the low coverage o f the current formal safety net. Households' capacity to self-insure against temporary income shocks is however quite important, mostly thanks to a large access to family-based networks and private transfers, the reliance on multiple sources o f income, and the ownership o f durables that could serve as collateral to access credits or be sold to compensate for temporary loss o f income or to cover unexpected expenses such as health care. 3.55 Sustaining growth is expected to be much harder over the comingfew years. This contributes to a high risk of remaining orfalling into incomepoverty. As discussed earlier, the recent growth performance in Kosovo has been driven by a post-conflict boom fmanced by official aid and is unlikely to be sustainable. Prospects for income growth, as identifiedin the Kosovo Economic Memorandum, will largely depend on the maintenance o f peace and security - an issue that remains a challenge as illustrated by the March 2004 riots - as well as on the speed o f resolving KOSOVO'S legal status and implementing a set o f reforms that promote private sector ledgrowth. Simulations have been conducted to examine the impact of policy reforms and status resolution on growth. The results show that compared to a "base- case" scenario which assumes rapid reform and status resolution by 2005, the delay o f both reforms and status resolution may reduce growth rates from an average o f 7.2 percent to 2.3 percent for the period 2005-2007, and from 5 percent to 2.2 percent in2008 to 2010. 3.56 Additional simulations conducted for this report show that holding income inequality constant, lower growth rates would contribute to a less rapid decline inpoverty (Table 13). In the "low-case" scenario with an average real per capita consumption growth o f only 2 percent per year during the next 5 years, the extreme poverty rate would decline only from 15.2 percent to 11.5 percent. Inother words, without sustained growth rates, the risk o f remaining or enteringinto income poverty would remainhigh. 30 - - Table 13: SimpleIncomePovertyProjections - Annual Poverty Rate Consumptiongrowth 5 years 10 years -3% 22.2 30.6 -2% 19.6 25.5 -1% 17.6 19.6 0% 15.3 15.3 1Yo 13.3 11.5 2% 11.5 8.1 3% 10.2 6.2 Source: HBS, 2002. 3.57 Environmental depletion is a serious problem that couldjeopardize the livelihood and income of many householdsthat rely on natural resources,particularly in rural areas. InKOSOVO,biodiversity andthe naturalenvironment is animportant source o f sustainable the economic opportunities for a large part o f the population and it needs to be preserved. As shown in Figure 2, a fairly large number o f households live o f natural resources and are vulnerable to environmental risks. This is especially true inruralareas, where accordingto the H B S 90 percent o f households have land, 55 percent have livestock, and 15 percent live of their own food products. Other estimates from the Kosovo Forest Agency show that about 10 percent o f the overall population, and 20 percent o f the rural population, relies on forestry production for its living. O f major concern is the very poor state o f the environmental conditions inKosovo (Box 5), which constitutes, among other risks, a considerable threat for the income and livelihood of the large share o f the population which depends on natural resources for their living (e.g. farm andgrazingland, fish, forest production, wild food). Figure 2: Dependency of Householdson Natural Resources (percent ofhouseholds) #All URural 100, 89.6 90 80 70 60 50 40 30 20 20 10 0 Have land Have livestock Rely on subsistencefarming Rely on forestryproduction for living Source: HBS, 2003. Data on forestry are from the Kosovo Forest Agency. 3.58 Widespread unemployment and deficit of decent job opportunities contribute to large income insecurity. The evidence discussed earlier shows that unemployment and precarious job holding are strong correlates o f poverty. In KOSOVO, formal employment is very scarce and unemployment i s very high. As a result, the risk that households or individuals will fall into poverty is very high. Available data show that, as in other former Yugoslav republics, unemployment in Kosovo was already high during the pre-transition period with an estimated rate o f 36 percent in 1990. This rate climbed throughout the 1990sto an estimated 68 percent in 1999 before the war. Since the end o f the conflict, and despite the recent growth performance and an increase inprivate sector activity, the situation inthe labor markethasnot significantly improved andremainsworrisome (Table 14). 31 3.59 In 2002 Kosovo had the lowest labour force participation rate, the highest unemployment rate (registered and non registered) and the lowest educational level o f all former Yugoslav republics (World Bank 2003). There are also some concems about the large share o f the informal economy. Data from the 2002 LFS shows that although 90 percent o f workers were permanent workers, less than half had actually signed a regular contract and about a quarter were estimated to be inprecarious working positions. For those who work and have a regular job, wages are satisfactory if compared with other countries intransition with anaveragemonthly wage estimatedfromthe 2002 HBS at 249 Euro. 3.60 However, manyo f existing formaljobs are still reliant on intemational aid andpublic provision and are mostly found in services related in some form to the intemational community. In 2002, according to the LFS, general government employed 6 percent o f the workforce, education 8 percent, and health 4 percent. The share o f Kosovo's employment in manufacturing was very small relative to neighboring countries (8 percent o f total employment), and a large share of Kosovo's employment was still engaged inagriculture (39 percent), mostly in some kind o f subsistence farming. Apart from agriculture, trade and construction constituted the most important sectors o f employment (12 and 8 percent respectively), largely inthe informalsector. 3.61 While Kosovo has probably one o f the most flexible labor market in the world, a combination o f factors is often mentioned that explain the lack o f a substantial reduction in unemployment: (i)pattern of growth that has not been labor intensive; (ii) retum o f a a the quite largenumber o fKosovar asylum seekers fiom western European countries, whose status in these countries is no longer sustainable; (iii) negative income effect on the local the individual and/or household labor supply related with the large level of workers remittances from abroad; and(iv) a growing populationofworking age. - Table 14: Main Labor Market Indicators 1989 2000 2001 2002 Employmentrate (YO) 22.2 40.9 19.6 21.8 Unemploymentrate (%) 36.3 12.1 41.2 47.2 Labor force participationrate (%) 34.9 46.5 33.4 41.3 Source: World Bank (2003), Kosovo Labor Market Study 3.62 Barriers to internal mobility are important and may reduce households' capacity to cope with income shocks. Intemal mobility and access to transportation is an essential element that allows households to cope with income shocks through better access to employment, to markets, to family networks, to health services or to social assistance centres. InKOSOVO, however, the evidence points to the existence o f barriers to mobility that i s not only related with the fragile security situation, but also with the deficiencies o f transport services. 3.63 Poor transport infrastructure services is a not a novel phenomenon in Kosovo. In former Yugoslavia, the province was the least serviced interms o f transport. Chronic under- finding o f the transport systems during the 1990s and the war have further deteriorated the transport network. The 2003 World Bank transport review estimated that the road network consists o f 3,800 kilometres including 623 km o f main paved roads and 1300 km o f secondary roads o f which 920 kmare paved. Intercitypublic transport is mainly providedby Kosovo Trans a cooperative owned by the government (51 percent) and the employees (49 percent). About 80 percent o f the fleet and almost all service equipment were destroyed during the conflict but thanks to the post-war effort the fleet was back to about 200 buses operating in 2003. Local urban transport is provided by a mix o f municipality and private companies while the railway network i s constituted by a single non electrified track o f about 330 kmwhich i s inneedo fupgrade anddoes not currently contribute significantly to the local 32 economy. Over 60 percent of fteight traffic is represented by KFOR activities and only 40 percent by commercial activities. Passenger traffic i s limited to a twice a day service that covers the length o f the network. 3.64 Private health expenses among households who incur such costs are high in Kosovo. For some of these households, they could be catastrophic and lead to income paver@. According to estimates drawn form the HBS, in2002, private spending on health, in the form of out-of-pocket expenses for private services and pharmaceuticals, official co- payments and informal payments to professional staff in the public system, amounted to 7.5 Euro/month and represented about 1.75 percent o f GDP. This is lower than the previous estimate based on the LSMS and which gives a figure o f 3.58 percent o f GDP in20001*.The average household health expenditure in 2002 (7.05 Euros /month) appears relatively low if compared with, for example, the average expenses for coffee or tea (7.8 Euros) and footwear (10 Euros). Box 5: EnvironmentalDegradationinKosovo metals, such as lead and zinc, and may not be used even for industrial needs, before prior water treatment. The main sources of contamination are the industrial effluents discharged in the main river system. Another serious damage on rivers comes from human-generated waste and untreated wastewater. As Kosovo i s located in the upper watersheds o f four rivers, polluted rivers in Kosovo are also responsible for a typical transboundary environmental issue that involves the entire region since these rivers provide water also to downstreamneighboring countries. Less than one third of forested area can be considered healthy and productive. Kosovo benefits from a significant forest area that covers 42.1 percent of total land. Out of the forested area, private forest ownership cover 33.7 percent. Available studies show that out of the total forest area, approximately 17 percent i s in a degraded status, and almost 1/3 is subject to uncontrolled harvesting. Land degradation at thepower plants and lignite mines is common. Tailing dumps and old mines that are no longer in operation deteriorate a large area o f land that can no longer be used for agriculture activities. Non-reclaimed outside dumps and old mines that are no longer in operation further represent a potential source of contamination of the territory in the event of a natural disaster (flood, earthquake). Biodiversity conservation is at risk Kosovo is exceptionally rich in plant and tree species considering its relatively small area. Kosovo is also very diverse from a geological perspective with volcanic, metamorphic, and sedimentary rocks o f varying ages and origins present. However, the ecosystem in Kosovo has been under stress due to a long period o f mismanagement and neglect. National protected areas, as a share of total land area, are rather small in Kosovo (4.3 percent) compared with regional and world average (7 percent for ECA and 11.7 percent worldwide) and slightly below the European required threshold (5 percent). This leaves a large proportion o f the natural dicinal plants, fire, and the deposit o f sand 3.65 However, these are averages and include a majority o f households that do not incur any health expenditures inthe period considered, either because they did not seek or did not needhealth services, or because they did not pay for the services they received. Considering only households that incurred healthexpensesprovide a different story, as illustrated inTable '*The reasons for this major difference inthe estimates o f the private health expenditure may be due to: i)different survey tools, methods and interpretation; ii)a change in private expenditure patterns with respect to the immediate post-conflict period (private procurement o f drugs was more prevalent, the public system was less able to provide free drugs, prices were higher and consumption too may have been higher due to consequences of the conflict); iii)the GDP per capita has increased at an estimated rate o f approximately 5 percent over the last 3 years; or to various combination of the above factors. 33 15. What emerges from this table is that: (i) average health expenditures among those who the incur such expenses i s much higher than among the total population; (ii) average health the expenditures, and the proportion of householdsthat incur health expenses, is, not surprisingly, a direct function o f the total consumption; (iii) expenditures, as proportion o f the these average total consumption, are also higher in the lowest quintile; and (iv) a small but significant proportion o f households (8 percent in the poorest quintile and 4 percent in the richest), incur substantial expenses - i.e. greater than 10 percent o f the total monthly consumption - thatmay, particularlyfor the poorest, be catastrophic. 3.66 Participation in the formal safety net in Kosovo does not provide an adequate cushion against adverse income shocks. The social protection system in Kosovo i s still modest. The main social cash transfers are pensions and social assistance. There i s neither unemployment insurance schemes nor child benefits. An analysis o f the effectiveness o f the formal safety net in Kosovo is discussed inChapter 5. It shows that although social transfers have an important poverty alleviating effect among households who receive benefits, these are not sufficient to lift them out o f poverty. Moreover, the coverage o f the poor and the extremely poor is very low. According to the H B S data, in 2002, only 21 percent o f the extremely poor (pre-social transfers) were receiving social assistance benefits. This indicate that the vast majority of households in Kosovo cannot rely on the formal safety net as a sufficient protection scheme against income poverty. Table 15: Households Expenditures on Health, 2000 1'' -uintile 5ithq u i n ndecile Average household health expenditures among all 2.38 3.41 11.74 14.13 households(Euro) % of householdsthat incurredcosts 27.8 28.4 48.0 49.2 Average health expenditures among households that 8.5 13.3 20.7 38 incurred costs (Euro) Average expenditure in proportion to average total 8.4 3.0 household consumption for those who incurred costs (%I % of households whose monthly health expenditure 8.0 8.0 4.0 2.6 was more than 10%of their total expenditure Source: HBS. 3.67 Households' capacity to self-insure against temporary income shocks is not negligible, however, due to the ownership of physical assets, income diversifcation, and reliance on family-based safety-nets, including remittances from abroad. In Kosovo, possessiono f physical assets is relatively important, and so i s land ownership inrural areas. According to the HBS, in2002, 94 percent o f the population had a television, 47 percent had a video player, and 52 percent owned a car. Inrural areas, nearly 92 percent owned land and 55.5 percent had some livestock. An investigation o f the sources o f incomes further points to the existence o f a large private and family-based safety net, including remittances from abroad, and the reliance on multiple sources o f income, including among the poor. This indicates that households' capacity to hedge against temporary income shocks either by selling their assets and/or by relying on other sources o f income, including private transfers from relatives and friends, is not negligible. Exposure to Health Risks 3.68 Besides a fairly large level o f income insecurity, the population inKosovo tends also to have a highexposure to health risks. The major health risks are the environmental degrade 34 and pollution resulting from: (i) mining practices and industrial infrastructure that outdated ignored environmental impacts; (ii)poor housing conditions and quality o f basic infrastructure services, and (iii)weak environment management systems. Other important health risks are the threat o f a rapid HIV/AIDS epidemic and the prevalence o f risky behaviors. 3.69 Environmentalpollution and contamination is widespread and represents a serious hazard to health. Environmental studies indicate a rather bleak environmental situation in Kosovo. While the extent o f environmental problem and its incidence on health is still not well monitored in Kosovo, the few available studies show that environmental contamination i s a serious healthissue. 3.70 Contamination of soil andfoodstuff is probably one of the most seriousproblems. Contamination o f soil and foodstuff by heavy-metal and lead is a serious problem in some areas o f Kosovo where the mines sites and industrial facilities are located. Inenvironmental hot spots areas, lead concentration in soil exceeds the tolerable level. In particular, environmental samplings realized in these areas show that spinach contained 20-30 times more lead than accepted levels and potatoes 30 times more. The lead intake of people eating crops in theses areas has been calculated to be more than three times higher than the WHO andFA0recommended maximumweekly intake andabout 15 times greater thanthe standard for EUcountries. 3.71 There are quite alarming data on leadpoisoning in childrenfrom northern Kosovo. Unacceptably high levels o f calcium, nickel, arsenic and mercury have been found insoil and foodstuffs innorthern Kosovo and have contaminated the population. Although smelting and mining activities have ceased since July 2004 human exposure from the historical, industrial environmental contamination in the area continues. The first studies, conducted inthe early -1990s (Table the 16) showed that inMitrovica, blood lead concentrations at birthwere 20.8 pg/dl well above intemationally accepted (but now questioned as too high) level o f 10 pg/dl and much higher than the level in Pristina - and capable o f serious damage to the developing Central Nervous System (CNS). In fact, a government study conducted in the late 1980s found that 4 year old children in Mitrovica had significantly lower values for intelligence, verbal expression and perception, compared with children in Pristina. A planned large-scale testingprogram is on going andwill probably c o n f mthe earlier findings. Preliminary results from a health risk assessment o f aged 2-3 years, currently being completed by WHO in Mitrovica show Blood Lead levels appear to be associated with level of soil contamination where they reside. These preliminaryresults analyzed by WHO, show that children from the Roma community inNorth Mitrovica are the most exposed groups. Ingeneral, Roma, Ashkali and Egyptians tend to be disproportionately exposed to insecure living conditions, including lead contamination. Table 16: BloodLeadLevel (BBL inpgL) inPregnant Women and Children, 1992 MitrovicaN=208 PristinaN=l84 MeanBBLat midpregnancy 19.9 5.6 MeanBBLumbilical Cord 22.2 5.5 Mean BBL child at birth 20.8 4.9 MeanBBLchild at 2 years 35.4 8.5 Source: MoH, 2003 3.72 Water contamination is also a major environmental risk. InKOSOVO,majorities the o f rivers appear polluted by chemicals and bacteriological polluters. River pollution firther contribute to the pollution o f land through the irrigation system. There are fewer and more imprecise estimates on the level o f contamination o f drinking water, but major concerns are with the quality o f water from wells in rural areas. A report from 1989 by the Institute for 35 Public Health shows that 74 percent o f wells in rural areas had unacceptable levels o f bacterial contamination, More recent and limited rural water assessments by sanitary and water inspectors c o n f m these fmdings. There is also evidence that an important proportion of infantdeaths causedby diarrhea is relatedto water contamination. 3.73 Outdoor air pollution is most problematic around outdated industrial infrastructure, while indoor air pollution could affect a vast majority of households. Despite the fact that outdoor air pollution seem to have declined inthe last 15 years due to the decline of industrial activity and direct emissions to the air during the 1990s and the decision o f UNMIK to shut down some o f the more polluting industrial facilities soon after the war, Kosovo remains one o f the highest air polluter in South East Europe. Available estimates point to 5.5 tones o f C02 emissions in KOSOVO, compared with the 4.14 tones in FYR o f Macedonia, 4.06 tones in Croatia, 4.06 tones in Serbia and Montenegro, 3.86 in Bosnia-Herzegovina and 3.85 inRomania". Air quality also seem to be a serious problem around miningand industrial cites, which are.distributedquite evenly over the territory (Table 17). An environment samplingconducted by the French KFOR inMitrovica and surrounding villages in November 1999 and June 2000 shows that lead concentration inthe atmosphere exceededthe EUstandard. Table 17: GeographicalDistributionof MainEnvironmentalHot SpotsinKosovo Regions Main industries Polluting factors Environmental damages Giakova `Metaliku': metal factory; chrome mine (Giakove); . _ . Wastewater from Metaliku ContaminationofKrena Constructionmaterial (&can); Reported contamination from war river; PlasticFactory (Rahovec); mines Sandand gravel extraction from Drini I Bardheriver (Rahovec.) Gjilan Magnesiumand nickelmine; Solid particles and heavy metals in `Celiku-Metal' company (construction materials in industrialwastewater. metal); IBGfactory producingnickelsadmiumbatteries; RadiatorFactory,; Buildingmaterialfactories; ,VINEXmetal factory (viti). Trepca complex; Lead and general mine Contamination of Ibar Cobaltmine inVushtrri; contamination, river. Factoryof Plastic inSkanderaj; Chemical wastewater `Chrystal': crystal glassware production; Largezinc depositinZvecan. Peja Beer factory; Industrialwastewater; Contamination of Drin Leather factory; Chemical contamination from seed river. Seed factory; factory;, Textile factories; Unexplodedmines from war. Electro-motorsfacilities; Contruction Industry; Mine-bauxite quarries (Kline). Prinen `Dratex' textile manufacturing industry; Industrial wastewater, Contamination of Heavyindustrialtruck andmetal industryin Suhareka; Fertilizers Bistrica and White Drin Mining and Fertilizers used in the vinehards rivers. (Suhareke); Prishtina Obilic industrial complex; Dust pollution, Pollution of Sitnica Trepca complex: Badovac, Kishnica and Ajvalija Toxic gas emissions: NOx, S02, River; minesandConcentrator (Granica); CH4, C02, CO, phenol, Contamination of Floatation of Lead and Zinc (Kishnica), Smelter of Heavy metal inwastewater drinking water supply `Feronickel' (Gllocovc), not functioning, for Prishtina. Leadand otherminerals' mines (Novoberde) Ferizaj ConstructionIndustry Possiblechemical contamination Pipe factory Lignitedeposit ~ 13 IEA statistics, OECD `Emissionsfrom fuel combustion 2002' 36 Source:adapted from UNMIK,Health and Environment, 2003 and Premiere Urgence's report. Note:most of the industrial facilities listedinthe table are not working at the moment. 3.74 Indoor air pollution is also believedto be a seriousproblem given the widespreaduse o f dirty fuels, and could be responsible for the high proportion o f infant deaths caused by respiratory diseases. 3.75 A major cause of environmental contamination in Kosovo is due to outdated mining practices and poor industrial infrastructure. In the past, a very old and environmental unfi-iendly industrial system used to discharge waste and wastewater into the rivers and soil, without any treatment, contaminating surface and groundwater. Soon after the conflict, partly due to environmentalconcem, a largepart o f the industrial activities, including some of the high-polluted facilities in the Trepca complex in Mitrovica in 2000, was shut down, but some continue to operate, like the thermal Power Plants in Obiliq. While inmany cases pollution is the result o f accumulated industrial waste originated from the past activity, it is also resulting from outdated industries that continue to operate. Even if not all the industrial facilities are working at the moment, most of them, mines in particular, are responsible for wastewater and leakage from past pollution which represents a serious threat for groundwater, soil andriver contamination (see Box 6). 3.76 TrafBc is a growing source of airpollution in Kosovo. Infact, the number of cars is rapidly increasing and most o f them are not in line with the new European standards. Most cars are old or imported as second-handcars from Germany and Switzerland. They are fuelled predominantly with benzene without catalytic converter. Inaddition, road infrastructure is a problem, causing regular traffic jams that increasethe amount of gas emissions inthe form of COZ, CO, NO, anddust. 3.77 A source of indoor air pollution and unhygienic lifestyles is attributed to the unreliable power supply and the lack of adequate heating and water supply systems. Available data from the LSMS, H B S and HDIS reported in Table 18 all point to rather low connections rates to andpoor quality o f basic infrastructure services in KOSOVO, compared to other transition economies (World Bank, 2004b) 3.78 Although electricity coverage is high in KOSOVO, with 99 percent o f households covered by electricity, power supply remains very unreliable and i s a major source o f dissatisfaction among households andbusinesses. Regular power cuts have produced the need for wood buming stoves for cooking, and is a source o f huge wood cutting and indoor air pollution. Box 6: ReconcilingEconomicActivitywith EnvironmentalProtection: The Trepca Complex I One of the main sources o f soil contamination in Kosovo comes from the mine sites that are spread all over the territory. One main complex is the Trepca industrial park located inMitrovica, but with mines also inother regions. The Trepca Miningand Metallurgical complex i s one ofthe biggest of its kind in Europe. After the World War 11, Trepca developed as a significant engine of growth for the former Yugoslavia. Trepca' units were located all throughout Kosovo, Serbia and Monte Negro and they are currently split up between the Serbian Northpart of Mitrovicaand the Southem units in South o f Prishtina andNovoborde, under the Albanians control. As all the other Socially Owned Enterprises, Trepca is waiting for its ownership status to be defined. KTA has been given a mandate to privatize socially owned enterprises. However, technical, environmental, and social legacy is delaying the process. Operations were interruptedby in 2000 when KFOR closed the operation partly for environmental reasons. Between 2000 and 2003 there has been strong donor support to rehabilitate the mines with the intention to resume the mining operations later this year, andexport leaazinc concentrate. 37 Trepca management has plans to re-start a limited mine production at the end of this year conditionally to the respect of UNMIK rules and EuropeaniInternational standards. This plan i s consideredcritical to the potentialfuturebusiness activities andthe privatizationprocess. 3.79 Access to the district heating i s also very low. Overall, 1.9 percent of households report district heating connection and a large majority of households have to rely on unsafe andfuels like wood (95 percent) for heating, which further contributes to indoor air pollution. 3.80 The same Table 18 shows that while the proportiono fhouseholds with runningwater has increased since 2000, in 2002, still only 54 percent o f households reported to be connected to the central water system, and 31 percent were relying on wells as the main source o f water. Discontinuity o f water supply is also a common problem in Kosovo, resulting from the frequent leakage due to the poor condition and illegal connections, and from often-long power cuts and the fluctuating voltage that affect the water supply pumping. Household's misuse o f water, for irrigationfor instance, and lack o f awarenessand incentives to repair leakages, are also reasons o f shortage inwater supply. Irregular water supply i s also a serious health issue, given its importance for hygienic lifestyles, and an important source of dissatisfaction among households (see Box 7). Table 18: Access to Basic Infrastructure Services (percent of households) - LSMS(2000) H B S (2002) HDIS (2003) Dwelling electrified 99.2 99.3 98.7 Main source of water is central pipeline 54.1 72.6 Main source of water is own pipeline 9.4 Running water inside the dwelling 51.4 63.5 Mainsource of water i s wells 31.0 Access to drinkingwater 72.6 Use of district heating 1.9 Use of wood stow for heating 94.8 Ownin-house flush toilet 49.0 61.4 Ownlatrine outside the house 38.3 Garbage collected by truck 10.0 Disposal o f street container 27.5 Lackingwaste disposal (%of households dumping, 62.8 62.5 burning,burying and other) 3.81 Poor sanitationfacilities and the lack of municipal waste management systems are other factors that contribute to water contamination and increase health risks. Sanitation facilities are particularly poor in Kosovo. According to the HBS, only 61.4 percent o f households report having their own in-house flush toilet and 38.3 percent rely on outside house latrine (Table 18). 38 Box 7: People Talk about the Poor State of Infrastructure Services The lack of drinking water makes our life difficult, as we have to travel far to get it. (J.D.,20 year oldmale,Albanian, Ferizaj) We also have a lot of problems with the water. It happens that there i s water for two or three hours duringthe night and then it gets cut off and we don't have any water for the whole day. (T.M, 46 year oldmale, Albanian, Podujeva) Idon'tknowwhoistoblameforrestrictioninelectricity.Thisiscausingalotofdamage.TV sets, stoves, washing machines and radios keep burninginpeoples' houses. Citizens go to complainto Electrokosova every day. (A.J.,60 year oldmale, Albanian, Gjakova) We needto fix the sewage system becauseit is not connectedto the main canal. Ihave a duct nearbywhich smells. A diseasemay spreadfrom it.(A.H.,38 year oldmale, Albanian, Podujeva) There i s no sewage and nobody comes to remove garbage. Therefore the lack of sewage system and having no removal service of the garbage it shows a great risk for any epidemics, especially it is very risky for childrenandthere can be many insects. (M.R., 30 year old male, Roma, 3.82 There i s no wastewater treatment in Kosovo. Sewage system i s practically not existent in m a l area and covers only a low percentage of urban area. Only 28 percent o f homes are connected to a sewage system and only 7 percent in rural area. Wastewater is usually disposed o f in open channels, which contaminate surface and groundwater. Most o f towns and many villages have piped sewer systems that usually empty untreated water into streams or rivers. Septic systems are still uncommon. Sewerage system without treatment i s normally found in cities and larger towns, less commonly in few villages. In the lack of a proper sewage system, most villages report open channels. 3.83 The lack o f waste collection is also a major issue in Kosovo and an additional source o f water contamination. Although a lot has been done with the help o f donor agencies and international organizations, equipment and transport are still insufficient for a proper collection and disposal o f all the municipalwaste. Only 10 percent o f families report having their waste collected by truck and only 27.5 percent declare to have a street container. Another 55 percent report dumping their garbage. Recycling infrastructure and services are still not available, apart from few initiatives. Industrialandmedical waste remain also a major environmental issue, although the situation for medical waste has improved quite substantially inrecent years. 3.84 Poor housing conditions further expose a fairly large number of households to unsafe and insecure living conditions. While most houses inKosovo appear to be well-built, with no major damages, and provide sufficient space per capita, there is still a fairly large number o fhouseholds with inadequate housing conditions and exposed to unsafe and insecure living conditions. Available indicators of housing conditions based on the HBS are reported inFigure 3. In 2003, about 10percent ofhouseholds hadhouses made of mud/stones, and25 percent lived in overcrowded houses. In crowded houses, when ventilation facilities are not available, the use o f dirty fuels for cooking and heating represent a dangerous source o f indoor pollution. 3.85 The adverse impact o f the war on housesremains also visible. While UNDP estimates that 130,000 houses were damaged during the war and that approximately 60,000 had been reconstructed by 2002, there was still a fairly large share o fhouseholds reporting either major 39 damaged houses (9 percent) or minor damages (18 percent) mostly because o f the conflict. These figures may have also significantly worsened in some areas due to the March 2004 clashes. 3.86 Finally, while more than 98 percent o f households owned their dwelling, qualitative evidence indicates that tenure security i s actually a problem for many famiIies, as the legal status o f ownership was often not clearly defmed. Today in KOSOVO, many houses are built without regular authorization and with illegal connections to most public services, such as electricity and water. This was also true during the boom in construction activity that took place after the war. The high building activity conducted without the respect o f building codes, and the related risk o f sewerage flooding due to the poor status o f wastewater canalization andtreatment system, also increases the risk of flooding. Figure3: SelectedIndicators of HousingConditions(percentage ofhouseholds) 25 24.8 20 15 10 5 0 Walls madeof Major damages Minor damages Does not own More than 3 mdlstone principledwelling persons per room Source: HBS, 2002. 3.87 Kosovo meets all the conditions for a rapid spread of the HIV/AIDS epidemic. According to the UNAIDS classification, Kosovo i s a low level epidemic region, but there are all the ingredients for a rapid spread o f HN/AIDS. An increase, even o f alarming proportions, might occur for several reasons: persistent stigmatization of AIDS, large young population, high unemployment rate together with a growing N drug use, a growing sex market, highexternal mobility o f Kosovars anda large internationalcommunity. 3.88 Youth risky behaviors are important risk factors affecting youth health status. Health-related lifestyles and behaviours among youth are rapidly changing but changes not always occurs in the right direction. Youth may be better connected to the globalized world but the disappearanceof traditional values, the lack of opportunities for higher education and employment, idleness, and the lack o f information on health issues are conducive to risky behaviours suchas the practice o funsafe sex, substance abuse, prostitution, andviolence. 3.89 Child exploitation and worstforms of labor can be an important health riskfor a small number of children. Little is known on the incidence o f child labor inKosovo and on the situation of child workers. According to data from the LSMS, the percentage o f children inthe age group 10-14 engagedinsome kindofwork was about4.4 percent in2000. There is also some concern that some children may be working inhazardous conditions. Ina survey on child workers conducted by UNICEF in 2003, 8 percent o f children declared that they suffered frombadhealth as a consequence o f their working conditions. 40 PhysicalInsecurity 3.90 The Standards for Kosovo stipulate that "all people in Kosovo are able to travel, work, and live insafety and without threat o f fear o f attack, harassment or intimidation(...)". Physical safety remains nonetheless problematic inKOSOVO,a national level, as a result both of elevated environmental risks, at community level, due to inter-ethnic tensions, and at household level, with a large prevalence o f domestic violence and the rise in human trafficking. 3.91 Kosovo belongs to a major seismic area but its housing stock does not seem to comply with the security codes andputs a large share of thepopulation at risk. A major public safety issue inKosovo is that while Kosovo belongs to a major seismic area and is ratedrelatively highinterms o f risky seismological factors, security codes o f construction are hardly respected. At the moment, enforcement o f these security codes by municipal authorities does not seem to be a priority - overlooking the fact that in Skopje, only a few miles from the border, an earthquake in 1963 completely destroyed the city and killed many people. 3.92 Risks of landslidesandflooding are also high and arepotential threatsfor several households located near environmental hot spots. Landslides can be a serious security problem, particularly for households living close to the mines, where tectonic disturbances in the form o f fractures affect the coal seam, and where the risk o f landslides increases tremendously in the case o f heavy rainfalls. The high deforestation rate, especially inthe last years with the extensive use o f wood fuel, further contributes to the instability o f slopes. Floodingi s also aproblem that hasbeen aggravated bythe boominillegal construction. 3.93 Unexplodedmines used during the war remains a potential important safety issue. The presence of unexploded mines from war is reported in certain areas and is a serious source of landdegradation and security-concem for ruralpeople. 3.94 Inter-ethnic tensionsare apotential source of conflict and create an overallfeeling of insecurity among communities. This has direct implications for the freedom o f movement and equal access to services andincome opportunities, particularly for IDPs. The March 2004 riots in Kosovo have drawn again the attention of the international community to the still fragile security situation inKosovo. The next chapter explores the potential sources o f social instability and shows that the tensions between the Kosovo Albanians and Serbs cannot be explained by major economic and social disparities, and thus appear mostly politically and historically driven. 3.95 Yet, inter-ethnic tensions remainvisible inKosovo andgenerate an overall feeling of insecurity, especially among the Serbs. According to the KEWRs, the percentage of respondents who declared to feel safe or very safe on the streets declined betweenNovember 2002 andNovember 2003 from 56.1 percent to 53.5 percent for the Albanians and from 39.9 percent to 12.4 percent for the Serbs. The same data also show a decline in the number o f respondentswho declared to feel safe or very safe at home: from 68.8 percent to 66.5 percent among the Albanians and from 73.9 percent to 23.2 percent for the Serbs. Other data from the 2003 K M S would point to an even more negative perception o f security. Lack o f general and personal security is ranked among the top five problems by all ethnic groups inthis survey and particularlybythe Serbs which rankedthis problem second after freedom o fmovement. 3.96 Yet, the growing feeling of insecurity observed between 2002 and 2003 is not supported by crime statistics. What is remarkable, however, is that perceptions o f insecurity as measured in public opinion polls do not seem to be closely related to crime statistics. Available crime data indicate that the number o f reported crimes has generally declined between 2000 and 2002. As shown inFigure 4, all major crime statistics consistently declined 41 over the period with the exceptions o f grievous assaults and arson. However, a press release from UNMIK police in January 2003 reported that "The increase in reported Grievous Assaults is believed to represent a greater willingnesson thepart of the victims to report such crimes instead of indulging in personal revenge. The number of Arsons increased but the severity of the crimes decreased." Figure4: Numberof ReportedCrimes m o o 0 .2001 m o o 2 700 590 600 500 400 300 200 100 0 3.97 Moreover, only few reported crimes in 2002 and 2003 were ethnically motivated. The January 2003 UNMIK police report argued that "Most Arsons were motivated by personal disputes between individuals (rarely resulting ffom ethnic conflict) and typically involved damage to non-residential property such as wood piles and hay stacks.'' The 2002 UN Civilian Police Report states that "Of the 68 murders recorded in 2002, 60 involved Albanianvictims, six of the deceasedwere Serbian and two were o f other ethnicity. Inmost murder cases the ethnicity o f the suspect was the same as that of the victim (...) Crime generally occurs within ethnic groups and i s motivated by personal reasons not involving ethnicity." 3.98 The OSCE-UNHCR Tenth Assessment o f the Situation o f Etlmic Minorities in Kosovo covering the period May-December 2002 would c o n f m that the crime situation was improving during the year. The report found an improved situation in relation to crimes committed against ethnic minorities and UNHCR found that where aggressions occurred these were often related to properties and therefore economically motivated crimes14. Such improvements encouragedKFOR to reduce the level o f protection and security measures and this factor, rather than growing inter-ethnic confrontations, may help to explain the increased insecurity felt by the Serbs between2002 and 2003. 3.99 The trafficking of human beings is another source of physical insecurity for women and girls in Kosovo. Recent studies by UNICEF and the Intemational Organisation for Migration (IOM) identify Kosovo as a place o f origin, destination, and intemal trafficking inwomen and girls, mostly for sexual exploitation, and induced by a large male expatriate population. The routes into Kosovo are through Macedonia and Serbia. Points of destination beyondKosovo are through Montenegro into Albania, andthen to ItalyandWesternEurope. 42 3.100 The multi-national police force working in Kosovo estimates that at least 90 percent o f foreign women working as prostitutes are victims o f trafficking. Other estimates from IOM, based upon interviews conducted between February 2000 and August 2001, show that women and girls are trafficked from the Republic o f Moldova (60 percent), Romania (19.5 percent), Ukraine (10 percent) and Bulgaria (7 percent) followed by infrequent cases from Albania and Russia. The majority o f the victims were recruited with promises o f jobs and about 9 percent were kidnapped. About half o f the recruiters were women and half were unknown to the victim prior to recruitment. The conditions of the Kosovo sex industry are poor and compared to slavery. Less than 15 percent o f the interviewed women reportedbeing paid and a little over half were sexually abused. 3.101 Recent investigations also show that Kosovo residents have become targets. Some o f the risk factors o f being trafficked include age, educational attainment o f the victim, poverty o f the victim's household, and the prevalence o f household violence. It has been estimated that about 85 percent of the Kosovar victims have achieved at best a primary education. Absent job skills and running from desperate circumstances, many victims are lured into traffickingwith the promise of employment that can help them improve their lives. 3.102 Domestic violence against women is a serious safety issue. Violence against women by knownmen is expressedbothphysically and psychologically. The occurrence o fviolence within the family may include battering, the sexual abuse of female children, dowry related violence, and marital rapeby either a spouse or non-spouse. It is perceived that the incidence o f violence may have increased since the conflict and the breakdown o f traditional family structures. 3.103 According to a study done by UNIFEM in 2000 on violence against women in KOSOVO, which intends to be a representative sample, 23 percent out of the 213 and respondents had experienced either psychological or physical violence by a known man during 1999 or 2000. The respondents also indicated that follow up services for victims o f violence were neededbut largely absent. O f all women who disclosed experiencing violence, about 54 percent o f urban women and 7 percent o f rural women had not talked with anyone about the violence. Social marginalization, lack o f employment opportunities, lack o f willingness among victims to seek redress through the courts were cited by the study as possible reasons for the persistenceo fviolence. 3.104 The reservation by victims to report occurrences o f domestic violence may result because the family in Kosovo is considered a private sphere and closed to public criticism. Victims may also perceive limitedaccess andability to seek redress within thejudicial system in Kosovo. Intheory property law entitles female spouses to an equal share of the family assets inthe event of a divorce. Inreality, women often relinquish their claims to the family o f the spouse andtheir childrenbecome the property o fher spouse's family as well. 3.105 Other data from the 2003 annual report by the Center for Protection o f Women and Children (CPWC) indicate that the incidence o f domestic violence i s a serious issue, but the credibility o f such data remains at stake (Table 19). The information is based on information collected from women who have sought assistance,against violence from the Center but it is not always clear whether calls asking for information about domestic violence are not registered as cases. The forms o f violence reported by victims are often multiple. Inabout 90 percent o f the cases, psychological violence i s experienced. Physical violence occurred in approximately 60 percent o f the cases and sexual violence occurred inabout one-third o f all cases. These cases include domestic violence, trafficking, forced prostitution and incest. Out o f the total victims who report an incident, a little over half are victims o f domestic violence. The vast majority o f female adult victims are victimized by their husbands. Another 45 percent o f women are subjected to violence by family members other than a spouse. 43 Approximately 82 percent of the victims reside in rural areas and about 40 percent o f the women are betweenthe ages o f 30 and40 years. Table 19: Numberof ReportedCasesand Forms ofDomesticViolence Year Incest Partner Parent Other Homicide/Suicide 2000 13 1378 151 344 2001 20 434 49 1922 2002 10 771 564 135 5 2003 15 754 246 1228 25 Source: CPWC, 2003. D. POVERTYAS A DEFICITINEMPOWERMENT 3.106 Summary. Empowerment refers to the capability of all people, regardless o f their income status, ethnicity, race, religion, or gender, to participate in, negotiate with, influence, and hold accountable institutions that affect their lives. Empowerment is an important dimension of well-being, one that is also well reflected inthe "Standards for KOSOVO" under headings I(the functioning o f democratic institutions) and I1(the rule o f law). This section provides a tentative assessment o f the level o f empowerment inKosovo with a focus on three dimensions: (i)the scope for participation in, and influence over, the decision making process; (ii) the extent to which state and civil institutions are responsive to the citizens; and (ii) scopeforaccessingreliablesourcesofinformation.Thefindingsshowthattherehas the been positive development but several challenges remain ahead. One the one hand, UNMIK has established a legal framework that promotes a democratic society and the rule o f law, gender equity is promoted in formal state institutions, andparticipation and representation o f ethnic minorities is guaranteed by the Constitutional Framework. On the other hand, achievement in political and civil freedom i s mixed, the unresolved legal status o f Kosovo generates a general sense o f powerlessness, youth's and women's participation in society remains at stake, and participation for all in political life is challenged by the risk on non- participation o f Serbs in local election. Corruption and low level o f information are also problems insome areas. ParticipationinPublicLife andDecision Making 3.107 UNMIKhas establisheda legalframework thatpromotes a democratic society and the rule of law in Kosovo. Followingthe end o f the conflict inJune 1999,the UnitedNations Security Council Resolution 1244 (UNSCR 1244) placed Kosovo under temporary UN administration (UNMIK),and with the objective o fproviding"substantial autonomy and self- government" to the people o f Kosovo while reaffirming the sovereignty o f the FRY over the territory o f Kosovo. Since the end o f the conflict, a lot has been done by the international community to put inplace democratic andeffective institutions. 3.108 A Joint Interim Administrative Structure was set up in February 2000 in order to share the administrative responsibility with the people o f Kosovo and the administration o f municipalities was handedover to locally elected representatives after the municipal elections o f October 2000. The adoption o f a new Constitutional Framework in May 2001 that guarantees political rights and civil liberties to the citizens, the first election for a Kosovo- wide assembly in November 2001 and the first Presidential election in March 2002 were further major steps toward self-govemment andgreater humanfreedom. 3.109 Achievements in political and civil freedom are mixed. Kosovo is rated as "partly free" in 2004 by Freedom House. Despite incredible efforts by the international community to put quickly in place democratic and effective institutions and bring peace and political stability, the results so far have been mixed. Independent evaluation by Freedom House, an 44 international non-governmental organization which monitors the state o f political rights and civil liberties all over the world, has rated Kosovo as "Partly Free" in2004. This ratingplaces Kosovo as a mixed performer in terms o f civil and political freedom, relative to other countries inthe region (Table 20). Table 20: Indicators of PoliticalRightsand Civil Liberties inSouth-EastEurope,2004 Score Rating Politicalrights Civilliberties Averagescore Albania 3 3 3 Partly Free Bosnia-Herzegovina 4 4 4 Partly Free Bulgaria 1 2 1.5 Free Croatia 2 2 2 Free Kosovo 5 5 5 Partly Free Moldova 3 4 3.5 Partly Free Romania 2 2 2 Free Serbia and Montenegro 3 2 2.5 Free Slovenia 1 1 1 Free Source: Freedom House. Note: Free refers to an average score o f 1to 2.5; Partly Free, from 3 to 5.5; andNot Free, from 5.5 to 7. 3.1 10 The unresolved legal status of Kosovo generates a general sense of powerlessness and voicelessness. Little progress has been made toward the resolution o f Kosovo final political status. Discussions about the final status are forecasted for 2005 at the earliest. But as o f today, the citizens in Kosovo have been living under temporary UN administration for already five years, and important decisions affecting their lives have remained out o f their control. The lack of clarity about KOSOVO'Sstatus has been identifiedas a major source legal of economic insecurity. It i s also responsible for a growing feeling o f frustration and exclusion among the population. According to public opinionpolls, the uncertainty about the final status o f Kosovo became the first preoccupation o f the population in 2003, before unemployment and poverty (Table 21). In the qualitative poverty study, youths provide a candid assessment of the status-quo and express their resentment with the impossibility for the Kosovar sport teams to berepresentedinintemational competitions. Table 21: Opinionsonthe Biggest ProblemsFacedby Kosovo (percent ofrespondents) November 2002 November 2003 In&astructure(roads) 3.6 0.5 Power supply . I 15.6 6.8 Urbanproblems 0.9 0.5 Environment 1.o 0.1 Poverty 11.2 16.2 Prices 2.7 1.4 Healthcareservices 0.7 0.3 Public andpersonalsecurity 2.6 3.3 Education 1.o 0.4 Interethnic relations 1.9 1.7 Unemployment 21.6 28.4 Uncertaintyabout the final status ofKosovo 19.6 31.7 Social problems 2.3 1.2 The fate ofthe missing 7.9 4.6 Organizedcrime and corruption 6.2 2.1 Unsocialphenomena(prostitution, drugs) 1.2 0.3 Somethingelse 0.2 0.4 Total 100.0 100.0 Source: Bank staff calculations based on the KEWR. 3.111 Youths' abilig to influence the institutions that affect their lives seems very limited. The qualitative poverty study points to a very limited influence o f youth on the institutions that affect their lives, whether in the family, at school, or inthe community (Box 8). Inthe family, important decisions affecting youths are often taken without any consultations with 45 them. The blame is usually on the persistenceof traditions, especially inrural areas, that put old males at the center o f decision making processes. There is also a general feeling among youth that many local institutions that dealwith youth issues are not responsive to their needs. Often, these are led by old people, and youth have no representations. Participation insocial life and cultural activities is also very much influencedby nepotism - what matters is who's daughter or son's you are, rather thanwho you are. 3.112 Many IDPs feel voiceless. Although there are municipal community services for displaced persons in KOSOVO, a recent survey conducted by the IDP Information Center on 200 IDPs fiom Northern Kosovo shows that many displaced persons from Northern Kosovo, about 83 percent o f respondents, have not heardabout existing local supporting services. And an equally largepercentage o f survey respondents, 73 percent, shared the feeling that no one was representing their interests. Box 8: Youth Voices About Their Lack of InfluenceinDecisionMaking Ithinkthatforamaleyoungpersonitiseasiertotakepartonthefamilydecisionmakingprocess thanfor a female youngperson. (Peja) The fact that the youth are not part of the decision making process is illustratedbest by the other fact that youth do not have the right to decide entirely by themselves where to spend their vacation Ithinkthat 80percentofthedecisionsaretakenbyparentsratherthanbyyoungpersonsin decisions that regard mostly the young person himselfierself. (Peja) Generally, the decisions are made byfamily, andnot by youth. (Rahovec) Youth do not go very often in concerts because even when concerts are organized they do not satisfy the needs of the young people, as they are organized by people who do not know what the youth Youth associations are much needed as they organize different activities for youth, such as computer training, acting courses, etc. (Peja) e from the young people themselves, but 3.113 Gender equity is promoted in formal state institutions and has enabled a greater representation of women inpublic lye. An infrastructure. also exists at levels o f both elected and non-elected govemment that promote gender equity into the national and local arena. Women have achieved increased representation in the Parliament and municipal assembly through the promotion of a quota system that reserves 30 percent o f the seats for female candidates in closed party lists. In the last Kosovo Assembly, 33 o f the 120 seats (27.5 percent) were occupied by women. Within govemment, various committees exist to mainstream gender into government policy. Such committees include: (i) Inter-ministerial an committee that exists at the Prime Ministerial level to meet on gender issues; (ii) a Gender- equality Commission o f the Kosovo Assembly that supports the mainstreaming o f gender .issues in the legislature; (iii) Gender affair office within the UN provisional Government a that facilitates the deployment o f municipal gender affairs officers to each municipality; (iv) a Gender Section in the Ombudsperson Office; and (v) a Gender task force within the Statistics Office to address gender statistics. 3,114 Despite increased representation in public lye, women's influence in decision making remains at stake. While there has been some progress in terms o f women's representation inpublic society, their ability to influence decision making remains limited. At the moment, women in Kosovo do not hold leadership positions within the 14 parties 46 represented inparliament, nor do they hold any leadership positions inparliamentary groups. O f the 18 committees in the last Parliament, women chaired only 7 o f them. Women occupy 28 percent o f all municipal assembly seats. There i s one female minister out of 10 ministers, one female permanent secretary out o f 9 permanentsecretaries, and only 2 females out o f 30 chief executive officers in municipalities. Women account for 25 percent o f judges, 17 percent o fprosecutors and 15percent o f the Kosovo Police Service (KPS). 3.1 15 Participation and representation of ethnic minorities is guaranteed by the Constitutional Framework -yet it is challenged by the risk of non-participation of Serbs in local elections. A system o f quotas is inplace to make sure that all communities and ethnic groups can be represented in elections. According to Constitutional Framework, 10 out o f 20 seats reserved for minorities belong to the Serbs. The self-government is m e r held accountable to ensure an equal access to justice and to make basic public services, such as health care, utilities and education, available to all communities. Yet, the riskofboycottingby Serbs of local elections comes out regularly and contributes to the overall political uncertainty. Accountability 3.1 16 The move toward more transparency in public management remains hindered by the lack of good administrative data. Assessing the extent to which state institutions in Kosovo are responsive to the citizens is a difficult task. At the first place, the lack of good administrative statistics, with the exception o f public fmances and monetary statistics, contribute to a lack o f transparency and information about the quality of the services provided bythe public administration. 3.117 Public opinion polls show a net decline in satisfaction with state institutions, but this is not necessarily related with the level of government effectiveness. Public opinion polls datainKosovo provide interesting insight as to the level o f satisfaction of the population with various public institutions. According to the Kosovo Early WarningReports (KEWRs), the percentage of people unsatisfied or very unsatisfied with the political situation increased from 31.8 percent inNovember 2002 to 47.9 percent inNovember 2003 (Table 22). While all major institutions in Kosovo lost trust during this period, the decline was more dramatic for the UN institutions indicating a growing resentment toward the foreign community. These satisfaction data need to be treated with great care, however. Kosovo has a rather complex governance structure, with the responsibility for different public services shared between PISG, UNMIK and KFOR. In this context, it is evident that the population may have difficulty inhaving a correct perception of who i s responsible for what and who is to blame whenproblems arise. Thus, the levelo f satisfaction with public institutions is not necessarily a good indicator o f government effectiveness. Rather, it i s a good indicator of people's perceptions, which canhelp understand and forecast people's attitude toward the state. Table 22: Level of Satisfactionwith Public Institutions (percentage ofrespondents) ___p Nov-02 Nov-03 Percentagechange UNMIK 63.8 28.4 -35.4 SRSG 73.1 43.1 -30 Government 74.1 68.5 -5.6 Parliament 76.7 65.3 -11.4 Local government 52.0 48.3 -3.7 Source:KEWRs No. 1-5 3.118 Corruption is considered a problem of medium to high importance depending on the sector, and it tends to affect disproportionately thepoor. A major indicator o f the extent to which public institutions are responsive to the citizens is the existence o f corruption. In KOSOVO, according to survey data from the KEWR, a large majority o f the population reports 47 the presence o f corruption among professionals in customs, in health and in the local administration (Table 23). Corruption is seen as particularly widespread incustoms and inthe health sectors. The perception o f widespread corruption among health workers is further confirmed by the qualitative poverty study and quantitative data on private heath expenses, which indicate that an important reason for incurringsignificant health expenses when health care is supposedto be provided for free is because o f informal "grey" payments (see Chapter 4). Corruption is a regressive tax which hurts the poor the most. It is income forgone by the public sector and it discourages the most vulnerable groups in society to benefit from using public services. Table 23: Opinions about the Presence of CorruptioninVarious PublicServices (percent of respondents) Health Education Customs Localadministration Quite present 28.5 23.3 41.5 36.2 Very much present 27.4 16.9 38.4 19.6 Quite or very much present 55.9 40.2 79.9 55.8 Source: Bank staff estimates basedon the KEWR#2, September-December 2002. Access to Information 3.119 Access to internet remains limited. ICTs are an important tool for empowering people because they provide access to information that can create earnings opportunities and improve access to basic services. ICTs also give citizens a voice to demand government support and reform. While in KOSOVO, nearly everyone has a televisions (94 percent), a vast majority has a radio (74 percent), and half the population has a mobile phone (51 percent), few seem to have access to a fmed phone: in 2003, a total o f 101,059 telephones were connected, compared with over 315,000 mobile telephones in service. Access to the internet at home remains also extremely low, with only 6 percent o f households having access to a computer with a telephone line (Table 24). The availability and use o f internet remains also quite limited among young people. According to the 2003 PSI-KAP survey of youth between the ages o f 15 and25, only 45 percent o f youth are accessinginternet, mostly through internet cafks. Table 24: Access to Informationand CommunicationTechnologies inKosovo around 2002 Listof items Percent of households Television 94 Video Player 47 Radio 71 Satellite dish 65 Have access to a computer 10 Have access to a computer with a telephone line 6 Mobile phone 51 Source:Hl3S (2002), HDIS (2003) for accessto computers. 3.120 The level of knowledge ofyouth on reproductivehealth, HIV/AIDS) other STIs and safe sex practices is still very low. The Government and many NGOs have put many efforts intrying to increasethe knowledge about reproductive health, HIV/AIDSandother sexually transmitted infections (STIs) among youth. Precise estimates about youth knowledge on reproductive health and HIV/AlDS prevention are still lacking, but early pregnancy and other available evidence suggests that the level of knowledge o f youth remains very low and that youth still experience difficulties to access quality information about safe sex practices. According to the 2003 PSI-KAP survey, although a majority of youth in Kosovo have heard 48 about HIV/AIDS (as many as 85 percent o f respondents inthe 18-24 age group), 23 percent did not know if there was HIV/AIDS inKosovo, and 16 percent didnot know that HIV risk could be reduced by usingcondoms. The same survey also showed that only 31 percent knew where to be tested for HIV, only 2 percent had been tested for HIV, and only about half o f sexually active youth were using condoms always or most o f the time. A low level of awareness on reproductive health and STIs ingeneral, and HIVprevention inparticular, also emerged fkom the 2003 SOK-UNFPA DHS. According to this survey, the Contraceptive Prevalence Rate (CPR) among women o f age 15-19 and 20-24 was only 1.7 and 12.3 percent respectively, and 0.8 and 3.2 percent for the modern CPR. And only 64.1 percent and 68.1 percent o f young women o f age 15-19 and 20-24 who have heard o f AIDS cited the use o f condom as a way o f avoiding HIV infection. 3.121 Additional evidence from the qualitative poverty study also shows that access to information about safe sex practices seems to be limited. Many respondents inthis study did not know much about the details o f preventive measures and unsafe sex was common practice. The main sources o f information on reproductive health were obtained from the media and older friends, and more rarely from school. In most cases, the issue was taboo among parents. 3.122 Environmental awareness tends to be very limited. According to data from the KEWR, people in Kosovo show a very low concern for the environment (Table 21). This contrasts with the importance o f the environmental risks discussed earlier and indicates that environmental education and awareness is still very limited. The lack of awareness and informationon environmentalissues i s a matter o f concern as it may limitthe scope to reverse the environmental degrade and preventpeople to deal with contamination of basic goods and undertakepreventive measures. E. MULTIPLE DEPRIVATIONS POVERTYLINKAGESTHELIFECYCLE AND IN 3.123 Summary. The previous sections have shown that poverty in Kosovo goes well- beyond income poverty to encompass low achievements in health and education, a large vulnerability to poverty, and a sense o f powerlessness and voicelessness. Yet, assessing poverty inits many dimensions does not necessarily mean that the extent o f poverty would be higher than ifmeasuredinone dimension only. Often, the different aspectsofpoverty tend to interact and reinforce each other inways that can exacerbate the deprivation that poor people face. This section does not intendto explore all the different poverty inter-linkages. Instead, it focuses on the multipledeprivations experienced by the income poor andtries to identify how various dimensions o f poverty interact differently depending on the life stages. The analysis shows that the vast majority o f the extreme poor are also poor inthe dimension o f education and infrastructure. There are no data on health and income status. Around birth and during early childhood, income poverty and environmental risks are major causal factors for severe health consequences. During adolescence, when school i s no longer compulsory, income poverty has a strong negative influence on school enrollment. Later on, low health outcomes contribute to income poverty through their adverse impact on employment. Duringadulthood, low education achievements become a major cause o f income poverty, mainly through lower wages andhigher unemployment. TheMultipleDeprivationsofthe ExtremeIncomePoor 3.124 I n Kosovo, the vast majority of the extreme income poor are also poor in the dimension of education. As shown inthe previous section, low education achievements and poor health status are strong correlates o f poverty. Additional data from the HBS reported in Table 25 shows a large overlap between extreme income poverty and education poverty. In 2002, while about 53 percent o f household heads were poor in the dimension o f education 49 (i.e. with primary education or less), the proportion o f extremely poor household heads who were also poor inthe dimensiono feducation stood at 80 percent. 3.125 The overlap between the income and health dimensions ofpoverty i3 more dijjicult to capture with existing data. The only tangible health outcome that can be disaggregatedby income poverty status i s disability. Only few o f the extreme income poor are disabled, indicating a small overlap, but the riskof income poverty associatedwith disability is high. 3.126 A non negligible share of the extremepoor also live in poor housing conditions. Poor housing conditions are usually captured through several indicators discussed earlier, each representing one aspect o f the problem. The same Table 25 shows that among the extreme poor, the overlap with housing poverty would be in the range from 12 percent (houses with major damages) to 35 percent (overcrowded house). This is also higher than the overall incidence o fpoor housing conditions inthe whole population. 3.127 A large majority of the extreme poor are infrastructure poor - but infrastructure poverty is widespread and affects all income groups almost equally. Almost half of the extreme poor do not have access to a central pipeline and one third has to rely on wells. Nearly all extreme poor have to use wood stow for heating, and about 3 out o f 5 extreme poor households lack waste disposal and in-house flush toilet. However, what i s remarkable i s that apart ffom sanitation, infrastructure poverty i s widespread and affects also the non-income poor. - Table 25: The Multiple Deprivation of the ExtremeIncomePoor - ___= Non-incomedimensions of poverty YOofnon-income poor householdsinthe Ratioof extreme below dimensions among: poor to all All population(a) ExtremePoor (b) @/a) Low educationcapabilities(of householdhead) Less thanprimary 22.1 38.3 1.7 Primary 30.9 41.4 1.3 Poor healthstatus (of householdhead) Disabled 15.8 16.3 1.03 Exposureto poor housing and livingconditions Househas major damages 8.6 12.0 1.4 House has minor damages 18.5 26.7 1.4 Walls made ofmudlstones 9.9 15.2 1.5 Morethan3 personsper room 24.8 35.3 1.4 Dwellingi s not electrified 0.7 1.4 2.0 No centralpipelineas main source of water 45.9 50.8 1.1 Mainsource ofwater is wells 31.0 35.9 1.2 Householddoes not use districtheating 98.1 100.0 1.o Use of wood stow for heating 94.8 97.3 1.o Doesnot ownin-houseflushtoilet 38.6 57.1 1.5 Own latrineoutsidethe house 38.3 55.8 1.o 1.5 Lacking waste disposal 62.5 61.4 Source: HBS, 2002 50 The ChangingPovertyLinkagesinthe LifeCycle 3.128 Around birth and during early childhood incomepoverty and environmental risks are major causalfactors for severe and sometimes irreversible health consequences. There is a major information gap on health outcomes across income groups in Kosovo, but a vast literature shows that the great majority o f the adversehealth outcomes (mortality, morbidity and disability) are more frequent in the poorest population. The lst 5* quintile infant to mortality rate typically ranges fiom 1.2 to over 2, usually increasing with the level of inequality. In KOSOVO, tentative extrapolations would give a risk o f death inthe first year o f life 1.5 times higher among households in the poorest income quintile, compared to the richest quintile. 3.129 Income poverty is an important cause of poor child health outcomes mostly through its negative effects on nutrition, exposure to poor water and sanitation and to indoor air pollution, and access to health care services. InKOSOVO, the impact o f income poverty on nutrition is not well established, but there is a good match between the proportion o f children livingbelow the extreme (food) poverty line (about 17percent) andthe proportiono f children showing low height for age (10 percent) andhavingiron deficiency anemia (16 percent) - two of the best indicators o f child chronic under nutrition. Moreover, the H B S data show that the composition of food intake in the lowest consumption quintile is skewed towards bread, cereals and potatoes. Exposure to poor water and sanitation and to indoor air pollution has been describedpreviously as greater inthe poorest households. Finally, the impact o fpoverty on access to health care is documented inthe LSMS, which shows that the cost o f the service was the most common problem, and one that hurt the poor the most. This issue further came out as amajor probleminthe recent qualitative poverty study. 3.130 Available evidence also shows that environmental contamination in Kosovo is responsible for lead poisoning among children and pregnant women, with well known negative repercussionson child's health (see Section C). 3.131 During adolescence, when schooling is no longer compulsory, income poverty has also a strong negative influence on school enrolment. Available data from the HDIS show that in 2003, while there was almost no significant difference in the enrolment rate between poor and non-poor children at basic level, the disparities increased substantially at secondary level and became enormous at tertiary level (Figure 5). Econometric results o f the determinants of primary and secondary school enrolment further indicate that household poverty status starts to affect significantly schooling outcomes during adolescence, when school is no longer compulsory (after the age 15). Figure5: School Enrolmentby Leveland PovertyStatus in2003 (percentage of children) "'1. 1w -" 96.2 96.0 80 60 40 20 0 Basic(ages7to 15) Semndary (ages16to 19) Tertiary(ages20 to 24) Source:HDIS.Note: enrolment rates refer to children currently enrolled or who have completed. 51 3.132 Later on in the life cycle, low health outcomes contribute to income poverty through their adverse impact on employment. There is very little analysis on the linkages between health outcomes and labor market outcomes in Kosovo, but available data point to a strong association between poor health or disability status and joblessness that may explain the higher incidence o f extreme poverty observed in the H B S among households with disabled members. In the LSMS, indeed, while the youth unemployment rate was 25.2 percent on average, it was 30.2 percent for the disabled. Moreover, while the share o f youth not ineducation andnot inemployment was 46 percent an average, it was 57.8 percent among the disabled. 3.133 During adulthood, low education achievements become a major cause of income poverty, mainly through lower wages and higher unemployment. As discussed earlier, extreme poverty rates are much higher among low educated adults. The large overlap observed between the income and the education dimensions of poverty during adulthood reflects the adverse impact that low education achievements have on wages and employment. An insight into the return of education on wages is provided by the 2003 World Bank labor market study. Based on multivariate analysis, the report finds that returns to education on wages are significant inboth the private and public sectors although less important than is usually found in other economies. The same report shows substantially lower employment rates and higher unemployment rates among individuals with elementary education or less, as compared with those with more education. The impact of education on employment and unemployment is also statistically significant. 52 4. A FRAGILE SOCIAL COHESION 4.1 Social cohesion may be defined as the capacity o f a society to manage potential causes o f social instability, such as excessive disparities in economic and social welfare, or misperceptions about these inequities resulting from imperfect information or the manipulation of information by various interest groups. Enhancing social cohesion is important in itself, but also for sustaining economic growth, which can be threatened by discontent and political unrest if actual or perceived disparities within society are too wide, and for making economic growth more conducive to poverty reduction. In Kosovo, such an agenda hrther underpinsthe challenge o f implementing the "Standards for Kosovo" endorsed bythe UNSecurity Council, which reaffirm the needsto build an inclusive society as one of the requirements for a final political settlement. An analysis o f social cohesion i s thus o f central interest in contemporary KOSOVO, and it complements well the macroeconomic risks assessment performed inthe World BankKosovo Economic Memorandum. 4.2 This chapter does not explore all sources o f social instability in Kosovo. Rather, it focuses on the most pressing actual and perceived socio-economic inequalities, and tries to understand how location, ethnicity, gender and age may shape these inequalities and become a source o f social instability. While the previous chapter has raised some concerns about the ability for all inKosovo to participate indecision making and access reliable information, the results inthis chapter further point to the difficulty for some groups to participate effectively ineconomic andsocial life. Location and gender are important factors, as shownby the large socio-economic disparities across regions and between Pristina, other cities and rural areas, and by the strong disadvantage of Albanian women in terms of health, education and labor market outcomes. There are also large disparities across the life cycle, with youth and the elderly facing a worse economic position, although for different reasons. A major issue for youth is the difficulty to find a decent work, which adds to their lack o f participation and voice inKosovo society. For the elderly, the main concem is related with the substantial loss of income incurred when leaving employment and joining the pension system. Finally, a potential source o f social instability and inter-ethnic tension i s found in the continuous existence of parallel structures for the provision o f social services in Serb enclaves, while socio-economic disparities between the Albanians and the Serbs tend to be modest. This confiis that tensions between the two communities are not socially nor economically driven. These tensions are nonetheless an important factor o f social instability, one that contributes to a sense o f insecurity and a lack o f freedom o f movement among the Serb minority. At the same time, the evidence points to a strong disadvantage o f non-Serbs minority groups in terms o f income opportunities and education outcomes. A. INCOMEINEQUALITY 4.3 Summary. Inequality is a broader concept than poverty in that it is defined over the entire population, not only for the population below a certain poverty line. Measures o f income or consumption inequality provide information on the extent o f disparities inmaterial well-being prevailing across households or individuals, and are thus key indicators o f social cohesion. Inequality measures are also important to forecast the impact o f economic growth 53 onpoverty reduction. This section startsby on overview o fthe extent of inequality inKosovo. It then provides some simulations to quantify the impact of average consumption growth on poverty reduction under different inequality scenarios. The fmdings show that consumption inequality is moderate and shall thus facilitate the transformation of average income growth intopovertyreduction. 4.4 Income inequality is high in Kosovo but consumption inequality is moderate. Inequality is measured on total household consumption and income, using three commonly used indicators: the Gini coefficient, the Mean log deviation measure (GE(O)), and the so- called Theil's T index (GE(1)). The values are reported inTable 26 and show that income inequality i s high in KOSOVO, with a Gini coefficient at 0.49. Consumption inequality is instead much lower, with a Gini coefficient o f 0.30. The difference between income and consumption inequality is not surprising. Consumptioninequality is usually a better proxy for the distribution o f material well-being than income becausehouseholds usually smooth their consumption over time while income is subject to underreporting andusually shows a higher variability. - Table 26: InequalityIndices - Total consumption P Total income Coefficient of variation 0.63 1.43 Ginicoefficient 0.30 0.49 Theil meanlog deviation measure 0.15 Sources: HBS 2002. Note: All figures are calculated for households' income and consumption andbased on aper adult equivalent. 4.5 An increase in inequality may reduce and even compromise thepro-poor potential of economic growth. To better understand the combined role o f growth and inequality on poverty reduction, a number o f simulations were conducted that provides possible extreme poverty rates under different consumption inequality and growth scenarios. The results are presentedinTable 27. Inthe case of moderate annual consumption growth o f about 2 percent inthenext5 years, whichcorrespondsto the low-case growthscenariospelledoutinthe 2004 World Bank Kosovo Economic Memorandum, extreme poverty may increase or decrease depending on the evolution o f inequality: with a 1 percent annual inequality increase, the extreme poverty rate would decline only slightly from its 2002 level o f 15.3 percent to 13.5 percent; with a 2 percent inequality increase, the poverty reductionpotential o f growth would be already offset with a extreme poverty rate as high as 16 percent; and with a 3 percent annual inequality increase, extreme poverty would actually increase to almost 18 percent". Table 27: ProjectedExtremePoverty Rateswith differentGrowthDnequality Paths (percent) Extreme Poverty Ratein 5 Years with differentgrowth andinequality paths Annualchange in inequality -3% -2% -1% 0% +I% +3% +2% -3% 16.5 18.7 20.0 22.2 24.7 26.2 27.4 f l5For details on the calculations, see Tsirunyan, (2004). 54 B. SPATIAL DISPARITIES 4.6 Summary. Besides the overall distribution o f income, the monitoring o f spatial disparities in income and non-income dimensions o f well-being i s essential to inform policy makers on the extent of territorial imbalances and to support the formulation of a more balanced development strategy, makingboth sectoral policies which have a spatial impact and regional policy more coherent. This section explores the extent o f spatial disparities in economic opportunities and other dimensions o f well-being inKosovo. The results show that spatial socio-economic disparities are important and justify a geographical focus in the poverty reduction strategy. 4.7 Kosovo has pockets of poverty in areas where local unemployment is often high. Table 28 presents selectedsocio-economic indicators byregions based on the 2002 HBS data. It shows that there are large regionaldisparities inbothincome poverty andunemploymentin Kosovo. In2002, the regional extreme poverty rates rangedbetween 7.3 percent inPrishtina and 28.9 percent in Ferizaj. As for unemployment among household heads, the range was between 6.4 percent in Peja and 23.7 percent inFerizaj. The same Table 25 also shows that the regions with the highest prevalence o f poverty (Ferizaj and Mitrovica) were also the regions with the highest unemployment rates. What i s probably more surprising i s that high poverty in Peja (22.6 percent) does not seem to be supported by a disproportionately high level o funemployment (only 6.4). But inPeja, the level o f income inequality was the highest, and may well explain a high incidence of poverty despite a relatively lower incidence o f unemployment. - - Table 28: Selected Socio-economic Indicators by Regions - Giakova Gjilani Mitrovica Peja Prizreni Prishtina Ferizaj Poverty and inequality Average daily household consumption per 1.87 1.96 1.71 1.98 1.81 2.06 1S O equivalent adult (Euros) Extremepoverty rates e?) 13.8 10.0 24.2 22.6 15.0 7.3 28.9 Ginicoef. 0.15 0.16 0.19 0.23 0.13 0.11 0.16 Contribution o fbetween group inequality to 2.8 overall inequality (mean log deviation) Labor market status of household heads (YO) Unemployment rate 17.5 16.1 23.7 6.4 7.9 15.9 23.7 Employment rate 54.7 55.7 50.9 67.9 58.9 56.3 45.7 LFPrate 66.3 66.4 66.7 72.6 64.0 66.9 59.9 Housing and living conditions (YO) House has major damages 8.6 18.6 13.5 10.2 6.2 4.1 10.6 House has minor damages 31.4 29.7 29.3 21.0 9.2 9.2 25.4 Walls made o fmud/stones 8.3 26.1 6.0 7.3 15.2 4.4 20.2 More than 3 persons per room 18.4 21.1 28.3 40.9 19.3 22.4 29.3 Dwelling is not electrified 1.6 0.2 0.7 0.5 1.5 0.0 2.9 N o central pipeline as main source ofwater 42.8 70.3 36.2 54.2 34.1 42.8 64 Main source o f water is wells 17.9 40.5 24.4 39.2 24.0 30.2 52.9 Household does not use district heating 98.3 99.3 96.7 97.3 96.4 98.2 99 Use o fwood stow for heating 96.1 96.9 95.1 97.0 95.7 92.1 98.5 Does not own in-house flush toilet 35.1 49 39 32.4 44.6 32.6 55.1 Own latrine outside the house 35.1 49.0 38.9 29.5 44.1 32.6 55.1 Lacking waste disposal 71.5 64 51.8 --. / L 59.8 65.7 50.3 Source: HBS, 2002. Weighted data. Note: Labor market status based on household heads of ages 16-65. The HBS does not allow estimating labor market indicators according to strict ILO definitions and may not be strictly comparablewith indicators from the LFS. 4.8 The regional distribution of infrastructure services is relatively flat, but housing conditions seem to vary quite substantially from one region to another. In fact, the 55 percentage o f households with seriously damaged houses in 2002 (before the March 2004 riots), was as low as 4 percent inthe Prishtina area, and as highas 18.6percent inGjilani.The quality o f housing materials varies also quite a lot across regions, with 26.1 percent o f households' houses built in mud and stones in Gjilani, compared to only 6 percent in Prishtina. 4.9 Income poverty is lower in rural areas than in secondary cities, mostly due to the consumption of own produced food. The distribution o f income poverty (as measured by consumption) by types o f settlements is shown in Table 29, along with other key socio- economic indicators. In the dimension o f income, the incidence o f extreme poverty is the lowest inthe main city Pristina (7.7 percent), but is higher in secondary cities (19.2 percent) than in rural areas (14.8 percent). However, this reflects the importance of subsistence fanning for family consumption in rural areas, compared to secondary cities, rather than better economic opportunities. Infact, the contribution o f own produced food to household consumption - which is includedinthe welfare aggregate usedto derive the poverty figures - was muchhigher inruralareas (19 percent) than inurbanareas (less than 2 percent). 4.10 Consumption inequality across regions and types of settlements is not very large. The decomposition o f inequality into within and between groups inequality shows indeedthat most o f inequality i s explained by within groups, indicating that consumption inequality does not vary substantially betweenregions andtype o f settlements (Tables 28 and29). 4.11 Employment opportunities in rural areas are as bad as in Secondary cities. Unemployment rates inrural areas are lower than in secondary cities, but this is because o f a large hidden unemployment inthe form o f subsistence farming and it does not reflect better employment opportunities inrural areas than in secondary cities. The employment rates are indeed identical in rural and secondary cities, and attachment to the labor market is even lower inrural areas than in secondary cities (Table 29). InPristina, unemployment is as high as in other cities, but there are more job opportunities, as indicated by a higher local employment rate. 4.12 Housing and living conditions are substantially worse in secondary cities, compared to Prishtina, but not as bad as in rural areas. All indicators o fhousing and living conditions reportedinTable 29 point to a significant disadvantage in secondary cities, relative to Pristina. This i s especially true for water and waste services, sanitation, and quality o f housing. The housing andliving conditions inrural areas remain nonetheless the worst. While the differences in the quality o f housing between rural areas and secondary cities are not so marked, there are large disparities in terms o f access to water and waste services, and sanitation. 4.13 Distance to secondary schools is a major barrier to education in rural areas. Physical access toprimary schools and medical centers is also more difJicult in rural areas. Overall, Kosovo exhibits a fairly good distribution o f community services; few households need transport to reach the closest health or primary education facility. However, this is not always the case for ruralhouseholds. According to the 2003 HDIS,basic services coverage in tenns o f medical facilities and primary and secondary schools is better in urban areas than in rural areas. Inrural areas, over 33 percent o f households have to cover over 3 kmto reach the closest medical centre, compared with 4 percent in urban areas. As for the availability o f schools, the rurayurban difference is not so highfor primary schools but becomes substantial for secondary schools. Indeed, while only 7.1 percent o f ruralhouseholds have to cover more than 3 km to reach a primary school, compared to 0.6 percent inPristina and2.4 percent in other urban areas, more than 71.4 percent o f households have to cover more than 5 km to reach a secondary school inrural areas, compared to 16.6 percent inPristina and 10.3 percent inotherurbanareas. 56 4.14 Illiteracy and secondary school completion rates are worse in rural areas. The evidence shows indeed that compared to urban areas, inrural areas adult illiteracy is higher and children exhibit much lower secondary completion rates. However, primary enrollment rates are not significantly different inrural and urban areas. Later on, the rural disadvantagein secondary education i s reflected in a large gap between rural and urban inthe proportion of adults with secondaryor more education. Table 29: Selected Socio-economic Indicatorsby Type of Settlements(percent) Urban Rural Secondarv cities Prishtina Poverty andinequality Average daily householdconsumption per 1.94 1.74 1.84 equivalent adult (Euros) Extremepovertyrates 14.8 19.2 1.7 Ginicoef. 0.15 0.17 0.10 Contribution ofbetweengroup inequality to overall 0.6 inequality (meanlog deviationmeasure) Labor marketstatus of householdheads (YO) Unemploymentrate 13.4 21.4 21.0 Employment rate 55.2 55.2 64.5 LFP rate 63.8 70.2 81.6 Educationaloutcomes %of individuals26+ withprimary or less 63.7 46.6 43.2 % of adults26+ with secondary 29.5 40.4 43.5 %of adults 26+ with vocational 3.0 6.9 4.4 Yoof adults 26+ withuniversity or higher 3.8 6.1 8.9 Illiteracy (cannot read nor write among IS+) 11.7 8.2 8.7 Primary school enrolment rate (7-15 yrs) 90.2 90.6 93.7 Secondaryschool completionrate (19-25) 45.9 64.0 76.8 Distanceto primaryschools (YOof households) < lkm 54.3 67.4 63.3 1-3 km 38.6 30.2 36.1 3-5 km 5.9 2.1 0.6 >5km 1.2 0.3 0.0 Distanceto secondary schools(% of households) < lkm 11.0 42.4 27.3 1-3 km 17.6 47.3 56.1 3-5 km 21.4 8.9 16.6 >5km 50.0 1.4 0.0 Distanceto medicalcentre(YOof households) < lkm 33 58 1-3 km 34 38 >3 km 33 4 Housingandlivingconditions (% of households) House has major damages 10.0 7.2 1.1 House has minor damages 20.4 18.5 2.0 Walls made ofmuastones 11.2 8.8 2.1 More than 3 persons per room 27.5 21.0 14.0 Dwelling is not electrified 0.8 0.7 0.0 No central pipeline as main sourceofwater 64.5 8.4 0.8 Mainsourceofwater iswells 43.5 6.2 0.0 Household does not use district heating 100.0 99.0 79.2 Use of wood stow for heating 98.9 91.3 71.0 Doesnot own in-house flushtoilet 51.9 14.1 0.0 Ownlatrine outsidethe house 51.4 13.9 0.0 Lackingwaste disposal 88.1 11.0 0.8 Source: HBS (2002), weighted data. Data on the distance to medical centre andprimary school are based onthe HDIS(2003). Note: the HBS does not allow estimating labor market indicators according to strict ILO definitions and may not be strictly comparable with indicators from the LFS. , 57 C. THEETHNICDIMENSION 4.15 Summary. Social stability can be achieved with population coming from very differentbackgrounds interms o fethnicity, religionor race. The European integration process i s one example, with the assimilation in one supra national territory o f populations that less than fifty years ago were still fighting against each other. What matters is to set-up institutions and policies that prevent ethnicity to become a factor o f division, such as excessive socio-economic disparities along ethnic lines, or misperceptions about these disparities, that may result from imperfect information andor the manipulation o f information by various interest groups to promote specific political and economic interests. This section provides an assessment of the extent to which disparities in actual and perceived social and economic well-being across ethnic groups are a potential source o f social instability in Kosovo. The analysis fmds that socio-economic disparities between Albanians and Serbs are modest and hence does not appear as a factor that can fuel the tensions between the two communities. At the same time, non-Serbs ethnic minorities face a substantial disadvantage. 4.16 Most of overall consumption inequality is explained by disparities within communities and not by inequality between Albanians, Serbs and other minority groups. Table 30 presents selected socio-economic indicators broken down by ethnicity. What emerges from this table is that there is very little difference in consumption inequality between Albanians, Serbs, and other minority groups in Kosovo. The decomposition o f inequality into within and between groups inequality shows that the vast majority o f overall consumption inequality is explained by intra-ethnic inequality (99.5 percent) and very little by inter-ethnic inequality (0.5 percent). 4.17 Socio-economic disparities between Albanians and Serbs are modest. Indeed, average consumption and the incidence o f consumption poverty are almost identical for the Albanians and the Serbs (Table 28). Inthe dimensions o f literacy and education, Serbs tend to exhibit much better outcomes, with the highest secondary enrolment rates and the lowest illiteracy rate. In terms o f housing and living conditions, Serbs report a higher incidence o f damaged houses and houses made of mud, worse sanitation, and a lower access to waste disposal, while Albanians have a lower access to water services and live in much crowded houses. As regards labor market status and wages, the econometric analysis performed inthe 2003 World Bank Labor Market study points to a higher probability o f being unemployed among the Serbs in rural areas (8 percent higher than for the Albanians) but not in urban areas, and a wage disadvantage among the Serbs but inthe private sector only, where Serbs earned, other things equal, 36 percent less than Albanians. In contrast, in relation to employment probabilities, there were not statistically significant. At the same time, Serbian females tend to enjoy muchbetter employment prospects thanAlbanians females (Section D). 4.18 Yet, the maintenance of parallel and discriminatory structuresfor theprovision of social services to Serbian minorities is a source of social instability and plays against the inclusion of Serbs into Kosovo society. The collection of quantitative information on parallel structures is extremely difficult, but qualitative evidence, including from interviews with officials from the OSCE, indicates that the maintenance o fparallel structures for the provision of servicesto Serbian communities, inparticular pension, remains an important barrier for the unification o f the social protection system. The Kosovo social protection system i s not exempt of problems (Chapter 5), but tremendous efforts have been made to provide pensions to all individuals of age 65 and above, regardless of their ethnic background. The fact that several Serbian pensioners may receive two pensions has probably been instrumental in keeping their poverty rates fairly low, but it may undermine the level o f perceived social cohesion. 58 Table 30: Selected Key socio-economic Indicators b y Ethnicity (percent) Type of settlement Albanians Serbs Other Povertyandinequality Average daily householdconsumptionper equ. adult 1.9 2.0 1.6 Extreme povertyrates 14.6 13.9 31.0 Average net monthly earnings 198.7 183.5 202.4 Gini coefficient 0.30 0.29 0.31 Contribution of betweengroup inequality to overall 0.5 inequality (Theil mean logdeviationmeasure) Compositionof populationandlabor force status Shareinpopulation 88.1 6.8 5.0 Shareinemployment 87.8 8.8 3.5 Shareinunemployment 86.0 10.1 3.9 Educationoutcomes Illiteracy (cannot reador write among 15+) 7.8 1.8 6.3 Primary enrolment rates (7-15) 96.6 95.2 86.6 Secondary enrolment rates (16-19) 78.3 96.3 41.7 Housing and living conditions House has major damages 6.6 26.5 7.2 House has minor damages ? ? ? Walls made of mudlstones 7.9 22.7 19.2 Morethan 3 persons per room 27.0 11.1 13.4 Dwelling is not electrified 0.8 0.0 2.5 No central pipeline as mainsource of water 48.7 36.8 12.7 Mainsource of water is wells 32.6 26.0 11.0 Use o f wood stow for heating 94.5 96.6 96.1 Doesnot own in-house flush toilet 37.4 47.2 43.3 Ownlatrine outsidethe house 37.0 46.8 42.8 Lackingwaste disposal 33.4 56.3 29.5 Source: HBS, 2002, weighted data, except employment and wage data based onthe 2002 LFS, and primary andsecondary enrolment rates based on the 2003 HDIS. Note: the H B S does not allow estimatinglabor market indicators according to strict ILO definitions and maynot be strictly comparable with indicators from the LFS. 4.19 Other non-Serbs ethnic minority groups face a serious disadvantage in the dimensions of income and education. While there are no major socio-economic inequities between the Albanians and the Serbs, the same Table 30 shows a clear disadvantage among non-Serbs ethnic minority groups inthe dimensions o f income and education. In2002, among other non-Serbs ethnic groups, the risk o f extreme poverty was more than two times higher than that of the Albanians and the Serbs. Moreover, in 2003, according to the HDIS, secondary enrolment rates among non-Serb minority groups were two times below the rates of Albanians and Serbs. Multivariate analysis based on the HBS further shows that Serbian heads are 30 percent more llkely to have dependents who have completed secondary school thenAlbanian householdheads, while other ethnic minority groups are 19 percent less likely to have dependents with secondary school completion than Albanian households. The World Bank Labor Marketstudy also indicates that the probability of unemployment was higher by 6 percentagepoints for non-Serbian minorities, compared to the Albanians, inurban areas, but that there were no significant differences with respect to unemployment. 4.20 Regardless of actual circumstances,there are areas of agreement and divergencein theperceptions of Albanians, Serbs and other minority about the mainproblems in Kosovo. Besides the true level o f socio-economic disparities in a society, another indicator o f social cohesion is the extent to which subjective well-being differ across groups. People's perceptions on well-being may differ from actual facts, andthis i s becausehumanperceptions are based on information that is often imperfect and that i s easily manipulated by interest 59 groups. To what extent people's perceptions in Kosovo match the overall distribution of economic and social welfare discussed above? How far people belonging to different communities share common opinions about the main problems in Kosovo? Table 31 sheds some light on these issues by presentingthe opinions o f Albanians, Serbs, and other minority groups on a range o f issues, as captured bythe EWRin2002 and2003. 4.21 For all ethnic groups, there is a broad consensus that unemployment and uncertainty about thefinal status of Kosovo are among the biggestproblemsin Kosovo. In 2002, out o f the 5 biggest problems perceived by Albanians, 3 were also among the 5 biggest problems cited by the Serbs and 4 were among the 5 main problems reported by other non- Serbs minority groups. For all groups, unemployment anduncertainty about the final status o f Kosovo were among the five biggest problems, and these both inthe 2002 and2003 survey. 4.22 A sense of insecurity is one of the topproblems reported by the Serbs, and also a non-negligibleconcern among other minority groups, but it is not seen as a major problem by the Albanians. Albanians and Serbs were instead very different inrelation to the weight attributed to public and personal security and inter-ethnic relations. While Albanians ranked these issues as very low, for the Serbs, insecurity was the third major problem in 2002 and became the first problem in2003. Inter-ethnic relations were also ranked as the second major problem by Serbs in2003. What is remarkable is that this strong feeling o f insecurity was not matched with crime data (see Chapter 3), but they obviously reflected a high risk o f inter- ethnic clashes, as well as the incidents experienced by Serbs in transports, which have contributed to limit their freedom of movement and their internal mobility. Among other minority groups, the sense of insecurity was not as large as among the Serbs, but it was on the rise (from the eighth biggestproblemin2002 to the sixth in2003). Table 31: Rankingof Opinionsonthe Biggest Problems inKOSOVO, - - by Ethnicity November 2002 November 2003 Albanians Serbs Others Albanians Serbs Others Unemployment 1 4 1 2 5 1 Uncertainty about the final 2 2 4 1 4 3 status ofKosovo Power supply 3 9 2 4 4 Poverty 4 6 3 3 I 2 The fate of the missing 5 5 8 5 3 I Organized crime and 6 I I 6 I 6 corruption Idfastructure (roads) 7 13 6 9 Prices 8 8 8 I 5 Social problems 9 13 6 8 8 I Public and personal 10 3 8 11 1 6 security Unsocial phenomena 11 10 9 13 9 9 (prostitution, drugs) Environment 12 12 14 Education 12 11 10 Urbanproblems 13 11 11 10 10 Healthcare services 14 13 10 12 8 Interethnic relations 15 1 5 14 2 I Something else 16 14 12 14 6 Source: KEWR. Note: "-''means no answer. D. GENDER EQUALITY 4.23 Summay. The UN Millennium Task Force and the World Bank have been developing comprehensive frameworks for monitoring progress towards gender equality in society to demonstrate where firther emphasis i s needed in data collection. The Task Force 60 and the Bank have proposed that gender equality be conceptualized along three dimensions: capability (including education, health, and nutrition outcomes), opportunity (access to productive assets, income, and employment), and agency (the ability to make choices that alter outcomes). Human abilities raised through the capabilities dimension may be a prerequisite to achieving increased access to economic opportunities and the ability to utilize themand engage effectively insociety. While the previous chapter has raised concems about women's ability to participate effectively indecision making inKosovo, this sections takes a further look at the extent o f gender disparities inincome and employment opportunities and in health and education capabilities. To the extent possible, it also tries to identify some of the underlyingcauses. The findings show that gender inequity remains widespread, especially in education and employment outcomes, andinteracts closely with ethnicity. 4.24 The higher incidence ofpoverty amongfemale-headed householdsis mostly due to gender differences in sources of income and educational attainment. The previous chapter has shown that gender o f headship has a differential impact on poverty and that female households face an elevated risk o f being poor. According to the HBS, about 28 percent o f female-headed households are extreme poor, compared to 15 percent among male-headed households. Why are female-headed households more likely to be poorer than male-headed households? 4.25 Available data from the H B S shows that on average, female-headed households support a larger ratio o f elderly individuals to income earners compared with male-headed households. Female-heads are also older than male-heads. Thus, female-headed households are more likely to rely upon pension income as the main source o f household income .(22 percent) and less llkely to rely upon labor income (40 percent) than male-headed households (among which only 14percent rely on pension as the main source o f household income and 58 percent rely on labor income). Given the low level o f pension, compared to wages, the combined characteristics o f higher dependency on pension income from other members and old age help explain the higher incidence o fpoverty among female headed households. 4.26 Another reason i s the fact that while education o f the household head is negatively associatedwith poverty, the educational attainment o f female heads i s concentrated within the lowest levels. According to the HBS, over 80 percent o f female heads have obtained at most ,primaryeducation, compared with 42 percent for male heads. 4.27 Girls' education becomes at risk at secondary level among Albanians and non- Serbs ethnic minority. Compulsory schooling in Kosovo covers both primary education, grades 1-5, and lower secondary education, grades 6-9. Duringgrade 9, students are provided with academic counseling and take a comprehensive exam before continuing with non- compulsory upper secondary education, grades 10-12. Secondary education offers the choice o f a vocational track or a general education track which prepares students for post-secondary education. Despite these enrollment requirements, 2001 school enrollment data compiled by the management team o f the World Bank funded education participation project indicate that school attendancevaries by ethnicity and gender. 4.28 Overall, the share o f girls' enrollment lags behindboys, especially at secondary level (Table 32). This is equally true for Albanians girls, who constitute the large majority o f girls, and for girls fromnon-Serbs minority groups butnot for Serbian girls. Inaddition, Albanian - girls are likely to drop out o f school earlier than boys. Female enrollment suffers the largest decline between the 8& and 9& grades when it decreases precipitously by 28 percent (Figure 6). However, when females survive this critical year, they are no less llkely than males to complete secondary school. Although welcome by the education community, the recent education reform that expanded compulsory education to grade 9 has not led to significant changesinschooling outcomes -especially among girls. 61 4.29 To further explore the determinants behind education outcomes within the household, a multivariate analysis was conducted to identify the net contributions o f individual, household, and community characteristics on primary school enrollment and secondary school completion. Location, ethnicity andincome appear to interact with gender andbecome significant causes o f disparities at secondary level. The probit results o f the determinants of primary school enrollment for household dependentsbetween the ages o f 6 and 15 years of age indicate indeed that the gender and relationship o f the dependent to the household head weakly affects primary school enrollment and that ethnicity plays no role on primary. However, for secondary school enrollment, the probit estimates show that both biological daughters o f the household and other female dependents are 12 percent and 23 percent less likelyto have completed secondary school comparedto the householdhead's son. In addition, ethnicity becomes an important factor. Serbian heads are 30 percent more likely to have dependents who have completed secondary school then Albanian household heads, while other ethnic minority groups are 19 percent less likely to have dependents with secondary school completion thanAlbanianhouseholds. Figure6: Primaryand SecondarySchoolEnrollmentby Gender and Grades, 2002 +Girls -m- Boys 30000 20000 15000 10000 5000 0, 1 2 3 4 5 6 7 8 9 10 11 12 Grade Source: Ministry of Education, Science andTechnology, World BankEMISproject. 4.30 Women face a substantial and growing employment disadvantage. Access to employment is an important indicator o f economic opportunity and gender equality both within families and the larger community. The contribution o f women's labor income to family budgets may equip women with opportunities to influence expenditures inways that are more beneficial to children compared to spending patterns influenced exclusively by male-controlled income. Female employment opportunities outside o f the home may change the relative return to investments ingirls and at least influence household allocation inamore gender-neutral way. Investments in girls and economic opportunities for women may also improve the efficiency allocation of resourceswithin the economy as well. 4.31 Figure 7 shows the evolution o f gender disparities in basic labor market indicators between 1989 and2002, usingdata from administrative sources for 1989, the LSMS for 2000 andthe LFS for 2001 and2002. The level andthe trends inthe female-male employment and unemployment ratios highlight that conditions for females have been historically worse than for menand that the women disadvantage inthe labor market has increased over time. 62 4.32 Besides a relative and growing disadvantage o f women compared to men, the LFS also show that for both men and women, the situation in the labor market has deteriorated in absolute terms. Some of the variation in the absolute employment trends may have resulted from differences insurvey methodologies implemented across the period. However, estimates covered by the last two years are based on similar methodologies and reinforce the direction o f the negative trends inwomen employment andunemployment relative to men. The gender assessment conducted for this study further shows that the share of female employment in Kosovo is relatively l o w compared to neighboring post-transition countries16. Figure 7: Female-MaleRatiosinLabor MarketIndicators, 1989-2002 -+-Femle-MaleLabor Force ParticipationRate -m- Female-Male ErrploymentRatio --tFemale-MaleUnemploymentRate 2 - 1.8 - 1.6 - 1.4 - 1.2 - 1 - 0.4 4 It 0.2 m I 1989 2000 2001 2002 Source: Adapted from Owens (2003) 4.33 The overall trends in female employment are influenced heavily by the Kosovo Albanian and urban patterns as Albanian women workers and urban female employment accounted respectively for approximately 82 percent and 62 percent o f national female employment. Yet, the gender employment gap varies quiet substantially across ethnic groups and between rural and urban areas. In general, urban residence and Serbian ethnicity are positively associated with female employment and tend to reduce the male-female employment gap. 4.34 Women disadvantagein terms of lower wages is lesspronounced. According to the LFS, on average, female workers earn approximately 89 percent o f the earnings o f male workers (Table 32). However, compared with other countries the female-male wage gap seems relatively small. Evidence o f women's limited pay gap is further highlighted in the World Bank Kosovo Labor Market Study (2003). The authors noted that the male-female wage gap was substantial in the private, but small inthe public sector. The earnings function results showed that once other observable differences are accounted for, the male-female gap in the private sector amounted to 8 percent, and was insignificant inthe public sector. The lack o f a gender wage gap in the public sector most likely reflected the presence o f a public sector wage scale. 4.35 The combination of limited schooling, early marriage, and entrenched attitudes amongfamily and community about thefulfillment of traditional roles help aplain women labor market disadvantage. Considering the persistence o f the trend suggesting that female l6 See Owens (2004). 63 employment opportunities have decreased, it is important to distinguish the underlying determinants associatedwith the downward spiral of bothmen's and women's labor market activity fiom those associatedwith the accelerateddecline inwomen's employment. 4.36 One reason for the female labor market disadvantage is the difficulty o f women to keep pace with men on earning academic credentials that translate to valuable workplace skills. While education i s a strong predictor o f labor force attachment among females, among the Albanian and the non-Serbs ethnic minority groups, fewer females than males complete secondary schools. Serbian women, incontrast, are no more likely to leave school earlier than men and exhibit a less unfavorable employment situation. Other factors that negatively influence women's labor force attachment are marriage, non-Serbian minority, young age, the number o f children under the age o f 7 years, and rural area residence. On the other hand, besides education, Serbian ethnicity increases the probability o f female labor force participation. Early withdrawal from school and early marriage among females are also more likely to occur in rural areas, where the influence of traditional norms are more strongly embedded inhousehold decisions, and contribute to the greater labor market disadvantage of women inruralareas. Table 32: Selected Socio-economicIndicators by Gender Men Women Poverty Extremepovertyrates by sex ofheadship(%) 14.6 28.2 Monthly average wages (Euros) 202.9 180.2 Employmentopportunities(YO) Unemployment rate 36.9 68.9 Employmentratio 35.1 8.4 LFP rate 55.5 26.9 Individualeducationoutcomes (%) Illiteracy (cannot reador write among7+) 2.8 9.1 Primaryenrolmentrates (7-15) 95.7 96.6 Secondary enrolment rates (16-19) 84.4 71.8 Highestlevelo f educationof HHhead(YO) Les than primary 14.5 47.2 Primary 27.6 32.5 Secondarylvocational 49.1 18.4 University 8.7 1.2 Healthoutcomes Life expectancy(years) 63-67 67-71 Source: HBS, 2002, weighteddata, LFS 2002 for employment andwage data, and 2003 HDIS for primary andsecondary enrolmentrates. 4.37 Kosovo has a remarkably smallfemale-male gap in life expectancylargely that may reflect thepoor quality of antenatal care and theprevalence of abortion. Life expectancy is an important indicator in the quality o f life and the relative size o f the male-female gap is evidence o f a gender differential in life quality. Under equality o f social and economic conditions, women are expected to outlive men. When the gap between female and male life expectancies is small, women may be vulnerable to maternal mortality, gender-based violence, sex-selected infanticides or other social risks. In contrast, large mortality gaps suggest that men may be susceptible to behaviors or events that elevate male mortality, including binge drhkmg, violence, and accidents. 64 4.38 Compared with neighboring countries, Kosovo's female-male gap o f 4 years is small. Life expectancy for men and women is estimated between 63-67 years and 67-71 years, respectively (Table 32), and suggests that factors are in place that undermines women's potential quality of health. A high maternal mortality, which is the highest among Europe, may contribute to the small size o f the gap, andcould reflect the poor quality o f antenatal care andthe prevalence ofat riskabortion. E. DISPARITIES ACROSS GENERATIONS 4.39 Summary. The previous chapter has shown that consumption poverty was disproportionately concentrated among children and the elderly. This section extends the discussion on age-relatedinequities inthree areas: (i) literacy, by looking at the prevalence o f illiteracy by age groups; (ii) employment, by bringing updated evidence on youth relative labor market disadvantage; and (iii) income, by documenting the relative loss in income incurredby pensionersonce they leave the labor force. The fmdings show that a large literacy gap between youth and the elderly fuels the disparities in life style and aspirations across generations and may be a source of misunderstanding. Income opportunities are substantially worse among youth and the elderly. O f major concern, and a major potential cause o f social instability, are the largepools o f unemployed and idleyouth. 4.40 There is a large literacy gap between youth and older people which fuels the disparities in life styles and aspirations acrossgenerations. Illiteracyis a major dimension o f poverty that prevents individuals to access essential information and to understand and adapt to a rapidly changing world. InKosovo, the evidence shows that while illiteracyi s a marginal problem among youth, it increases rapidly with age and becomes a major issue among the elderly (Table 33). This is evidently related with the disruption o f formal education experienced by older cohorts o f Albanians during the Milosevic era and the war. Such a strong divide in literacy between the youth and the elderly in Kosovo society is a potential source o f inter-generational tension as it reduces the ability o f older people to understand the life-styles and aspirations o f youth. The qualitative study points indeed to a large sense o f misunderstandingbetweenyouth and older family members. Table 33: SelectedSocio-economic Indicatorsby Age Categories (percent) 0-14 15-24 25-64 65+ Incomewell-being Extremepovertyrates 17.0 15.8 13.3 17.1 Employmentopportunities Unemployment rate 74.3 33.1 Employment rate 9.3 27.6 Educationoutcomes Illiteracy (cannot read or write among 15+) 0.4 0.5 4.6 48.9 Source: HBS 2002, weighted data, and LFS 2002 for the employment data. 4.41 Youth access to decent work remains a major challenge. Access to employment is evidently a cardinal point inunderstandingincome but also non-income dimensions o f well- being. In most cases, the lack o f decent work determines whether people are able to enjoy a decent standard of living. It is also a factor usually correlated with psychological instability, depression, marriage instability, crime, social exclusion and eventually social instability when this extends to large portion of the population. Having a decent job thus provides a sense of social inclusionthat justifies existence itselfto some extent. 65 4.42 While unemployment is a serious problem in Kosovo, the evidence also indicates a strong disadvantage o f youth relative to adults (Table 31). In2002, according to LFS data, the incidence of unemployment was more than 2 times higher for youth (of ages 15-24) than for adults (of ages 25-64). 4.43 Pensioners incur a substantial income loss when leaving the labor force as the result of smallerpensions relative to wages. InKosovo, reachingthe pension age andleaving the labor force is clearly associated with a strong decline in income. In 2003. The basic pension was set at 35 Euros, which is less than one fifth o f the average wage. 4.44 Thefact thatpensioners are among thepoorest, yet not entirely destitute, relates to the existence of closefamily ties which guarantee a minimum support. A still imperfect pension system inKosovo creates a large financial dependency o f pensioners on other family members. In fact, only 2.8 percent o f the elderly live alone and their well-being is, to some extent, guaranteed by a household structure and values among society which are rather inclusive of the elderly. 66 5. THE PRO-POORPOTENTIAL OF PUBLIC SERVICE DELIVERY 5.1 The delivery o f public services inhealth, education and social protection has a large potential on poverty alleviation. Yet, the realization o f the expected positive poverty impact depends altogether on the sufficiency o f public provision, the effectiveness o f public interventions, and the extent to which delivery i s equitable..In Kosovo, given the decline in international donor assistance, the fact that the vast bulk o fpublic social spending depends on general budget resources, and that there i s still a rather weak administrative capacity and fragile social cohesion, a critical challenge for poverty reduction is to protect the financing o f essential services inhealth, education and social protection, while enhancing the effectiveness andequity o fpublic social delivery. 5.2 The aim of this chapter is to assess the pro-poor potential o f public social service delivery by looking at the sufficiency, effectiveness, and equity o f public provision inhealth, education and social protection. The findings shows that while public spending on education compares well with other regions, it is relatively low on health and social protection, and there are important gaps, in view o f the population needs, in the availability o f secondary schools and the provision o f social assistance. Moreover, in the health and education sectors, the potential welfare gain associated with the current provision of services appears constrained by the poor quality o f services as evidenced bypoor health outcomes around birth and the presence of a large number of teachers with inappropriate level o f qualification. Inthe area o f social protection, the delivery of public social transfers is found to have a fairly large impact on poverty reduction, but the main incidence is provided by the basic pension while the poverty impact of social assistance is more modest, probably as the result o f large errors o f exclusion and small amounts of benefits. Finally, the delivery o f public social services in Kosovo i s not exempt from equity concerns. Inhealth, phenomena o f both inefficiency and corruption create undue out-of pocket payment which disproportionately hit the poorest. In education, there is some evidence that the poorest children may be disproportionately affected by the poor quality of teachmg. With respect to social assistance, Serb households and households with more than four children appear disadvantaged for reasons that remain to be explained. A. SUFFICIENCY 5.3 Summary.While the forthcoming World Bank Public Expenditure Reviewwill take a close look at the issue o f public spending from the point o f view o f fiscal sustainability and spending efficiency, this section focuses on the sufficiency o f public social services from a welfare perspective. Assessing the sufficiency o f public social delivery i s a rather complex task. It requires looking at both supply-side factors including the overall amounts allocated - to public social services, the administrative capacity to spend these amounts, and the efficient use o f the spending - and demand-sidefactors - such as population's needs. This section starts by reviewingthe amounts o f public spending inhealth, education and social protection, as compared to other countries. It then turns to a tentative assessment o f the extent to which the aggregate demand for public social services is satisfied. It shows that compared to regional average, overall public spending on education i s reasonable but public spendingon health and 67 social protection is low. Within education, a relatively high proportion o f public spending goes to primary education and very little on pre-school. The decline in donors' financial assistance on education andhealth raises concems about the sustainability o f public spending inthe social sector. Interms ofcoverage, the overallprovision ofpublic healthcare services seems reasonable, but the provision o f schools at pre-primaryand secondary level appears inadequate. Finally, while pension coverage seems to be high, yet not universal among the eligible population, social assistance delivery is very limited. PublicSocial Spending 5.4 Compared to regional average, overallpublic spending on education is reasonable butpublic spending on health and socialprotection is low. Table 34 provides the amount of public spending on health, education and social protection in 2003 as a percentage o f GDP and as a percentage of total public expenditures in Kosovo and selected countries. In2003, KOSOVO'S spending on education was 6.1 percent of GDP and 15.6 percent of total public public spending, and compared favorably to other countries inthe region. However, both the relative and absolute amount of public resources allocated to health and social protection appeared to be low compared to regional average. These amounted to respectively only 4.5 percent o f GDP and 11.5 percent o f all public spending for health, and 5.8 percent of GDP and 16.8 percent o f all public expenditures for social protection. Within social protection, public spending was allocated mostly to pension payments (3.3 percent o f GDP) and little to social assistance (2.5 percent o fGDP). 5.5 Within education, a relatively high proportion of public spending goes to primay education and very little on preschool. Looking at the 2002 distribution of public funds across primary, secondary and tertiary levels shows that primary education - where the social rates o f returns are the highest -receives the highest allocation (60.8 percent o f all public spending), while secondary education and universityihigher-educationreceiveda much smaller share (18.1 percent and 13.3 percent respectively). Spending on preschool education was even smaller (only 4.4 percent o ftotal public educational spending). - Table34: PublicSocial Spendingin Kosovo and Selected Countriesin 2003 - Kosovo Bosnia & Serbia & Slovenia Herzegovina Croatia Montenegro As a percent of GDP Social security & welfare 5.8 16.1 18.3 13.1 17.0 Health 4.5 6.2 6.8 5.9 5.5 Education 6.1 5.6 4.4 5.9 5.5 As a percentage of totalpublic expenditures Socialsecurity & welfare 16.8 31.3 36.8 34.2 40.7 Health 11.5 12.1 13.7 15.4 13.2 Education 15.6 10.9 8.9 15.4 13.2 Note: The values for Kosovo are budgeted amounts, but for all others the amounts are actual spending. Source: Kosovo authorities and forthcoming South Eastern Europe Regional Study on Public Expenditure Policies (World Bank) 5.6 The downward trend in donors'$nancial assistance on education and health raises concerns about the sustainability of pro-poor public spending in the social sector. Data on donor assistanceon education and healthfrom the PISG are onlyjointly available and indicate a steady decline, from 59.9 million euros in 2000 to 20.4 million euros by 2003. In 2003, about 13 percent o f total donor assistance focused on health and education sectors. While 68 public spending on education andhealth play a keyrole inpoverty alleviation, the reduction indonors' financial assistanceinthese sectorsraises concems about the sustainability ofpro- poor public spending. Sustainability depends on the overall capacity o f the public fmance to raise taxes, the capacity - so far limited-to enforce co-payments for health, on donor support, but also on the capacity to improve the efficiency of current spending. A forthcomingpublic expenditure review bythe World Bank will investigate these issues inmore details. Coverage of Public Social Services 5.7 The overallprovision ofpublic health care services seems to be reasonable.At the aggregate, the present situation o f the Kosovo health system is characterized by a quantitatively reasonable provision. This largely reflects the efforts with the restoring o f infrastructure (most health facilities have been refbrbished and re-equipped) and staffing o f facilities. At the end o f 2003, the public health system inKosovo comprised 6 mainhospitals, 365 primary health care facilities, 10 tuberculosis dispensaries, 5 health institutions and 7 community mental health centers. The hospital network consisted o f approximately 5000 inpatients beds or 2.6 bedsper 1000 population, which is well below the EUaverage o f 4, but sufficient to meet the needs taking into account that the age structure o f the population is a major determinant o f the needs o f hospital beds and that the population o f Kosovo is very young. 5.8 Theprovision of schools at pre-primary and secondary level appears inadequate. Availability o f primary schools i s not an issue for a majority o f the Kosovo population. According to the HDI, about 60 percent o f households have a primary school less than 1km away, and more than 95 percent have a primary school less than 3 km away. Availability o f preschool and secondary schools, on the other hand, emerges as an important issue. As regards pre-school education, accordingto data from SOK, inAugust 2001 there were only 34 pre-school institutions staffed with only 402 teachers. Most o f these institutions were in the Prishtina region and virtually all in urban areas. As for secondary schools, data from the H D I S show that in2003 only about 24 percent o f households have a secondary school within lkm and 28 percent have a secondary school located at 5km or more. As discussed earlier, there are also some important disparities in the availability o f education facilities between rural and urbanareas. 5.9 Pension coverageseems to be high, yet not universal among the eligiblepopulation. The basic pension is a Govemment-sponsored non-contributory universal scheme. It is conceived as a flat rate monthly benefit available to all residents o f Kosovo, as well as Kosovar refugees living outside the territory, age 65 and over, irrespective o f prior contributions or work history. Inthat sense, it is an old-age assistance program. It is pay-as- you-go andfinanced from the Kosovo CentralBudget (KCB). The scheme took effect inmid- 2002. The basic pension is set to equal the extreme poverty line (the food basket) and adjusted annually to reflect increases inthe cost o f living. For 2003, this level was 35 euros; for 2004 it is 40 euros. Because the basic pension i s tied to the cost o f the food basket and indexed to the CPI, the cost o f the basic pension system i s likely to increase at a rate lower than the overall growth rate o fthe economy, thw maintaining fiscal sustainability inthe longrun. 5.10 Using administrative data on recipients, some 103,000 pensioners, or 94 percent of the estimated 65-and-over population, are receiving pensions. A coverage of less than 100 percent is surprising in a situation where a pension i s available to anyone age 65 or over. It may reflect that some elderly cannot provide proof o f age, and some do not know how, or bother, to apply for the benefit. What i s a priori also surprising is the large discrepancy on pension coverage between administrative and survey data. In the 2002 HBS, only less than half o f eligible individuals (individuals o f age 65 and over) were receiving a pension. But this 69 i s probably because some delays took place in the distribution o f pensions after the scheme became effective in2002 which are pickedupinthe 2002 survey data. 5.11 Social assistancedelivery is very limited. The primary objective of a social assistance program should be to provide a safety net for the poorest in society. The benefit can be provided in cash or in kind, or it can involve public works schemes and other income- generating programs that produce assets. It can be means-tested or targeted at specific categories o f poor people. The choice o f social assistance program i s determined by local conditions-culturaland socio-economic factors, domestic income andresourcesavailable for social assistance; and the role o f civil society, i.e. the strength o f informal safety nets (family, kinship, and community support structures). An important consideration indesigning a social assistance program is that it does not unnecessarily encroach on already existing and well- functioning informal social safety netmechanisms insociety. 5.12 The Kosovo social assistance scheme was introduced inthe year 2000 and revised in December 2003. It was, and remains, a cash limited categorical scheme that targets poor households, while aiming at fiscal sustainability and administrative simplicity with easily identifiable categories. The scheme as it was duringthe 2002 H B S targeted two categories of beneficiaries: (i) Category One, composed o f families without resources where no one is capable o f work, or expected to make themselves available for work (single mothers, children); and (ii) Category Two, embracing families with at least one childunder 5, or caring for an orphan under 15; other members capable o f working should be registered as unemployed; no member can own a motorized vehicle or more that half a hectare o f land; no resources are being provided by relatives outside the family. Category Two has subsequently beenrevised with the introduction o f a new SocialAssistance Law inOctober 2003, involving the introduction o f means and asset testing for eligibility and calibration o f the benefitI7. 5.13 In 2003, a monthly average of only 51,000 families received social assistance that totaled 32.3 million euros, or 7 percent o f state budgetary expenditures, and 2.5 percent of GDP. This is low in comparison to the estimated 277,000 families living in extreme poverty in2002, anddrivenmostlybyfiscal limitation. Administrative data further indicate that the numbero fbeneficiaries hasbeendecreasing, drivenby a fall inthe numberof Category One- eligiblehouseholds (Table 8). 60000 , 50000 55000 45000 40000 * Figure 8: Social Assistance Recipients, 2001-2003 +Category One -i-Category Two +-Total recipients 35000 30000 25000 2000.0 'I 2001 2002 2003 Source: Ministryof Labor and Social Welfare l7 For further details onthe social protectionsysteminKOSOVO,Sederlof, 2004. see 70 5.14 The decline in Category One beneficiaries reflects improvements in the administration o f the program, as more stringent procedures for the verification o f eligibility has reduced errors o f inclusion inthe systemand identified mis-classification of families on to Category One when in reality they are eligible for Category Two. The increase in Category Two is inpart explained by this reclassification, as well as a deteriorating economic situation inKosovo that has resulted inrising unemploymentand the resultant increaseinapplicants. The Ministry o f Labor and Social Welfare believes that the rate o f increase, especially o f Category Two beneficiaries, would have been higher - possibly resulting in an overall increase inbeneficiaries -without a more stringent verification process. B. EFFECTIVENESS 5.15 Summap Besides sufficiency, the pro-poor potential o f public social services depends on the effectiveness o f service delivery, as indicated by the level of gain in well- being or reduction in poverty produced by these services. This section does not provide a rigorous quantitative impact assessment o f public social delivery in Kosovo, which anyway would be impossible given data limitation. Instead, it reviews some o f the most important impediments to an effective delivery o f public social services, focusing on the quality o f health care and education, andthe targeting efficiency andpoverty impact o f social transfers. The analysis shows that in the health sector, the on-going health reform process in Kosovo has achieved important results. Yet, the low quality of services remains a major impediment for an effective delivery o f health care. Inthe education sector, concerns about the quality o f schooling result from the presence o f teachers with inappropriate level o f qualification. Finally, the social assistance scheme appears moderately effective, with little errors o f inclusion but large errors o f exclusion. Overall, the delivery of public social transfers has a large impact on poverty reduction, but the main incidence i s provided by the basic pension while the impact o f social assistance is more modest, as the result of large errors o f exclusion andsmall amount ofbenefits. Quality of Health Care and Education 5.16 The ongoing health reform process in Kosovo has achieved important results but the low quality of services remains a major impediment for an effective delivery of health care. Little informationis available on the extent to which the public health systeminKosovo is able to deliver effective health interventions that produce significant health gains. Yet, there are some concerns, driven by quantitative and qualitative evidence, that despite substantial progress inrecent years with the reform process, the effectiveness o f the health care system remains underminedby the low quality o fhealth care services18. 5.17 Successes and difficulties toward greater effectiveness o f the health care system have bothcharacterized the reformprocess. Successeshave included: i) policy development andre- organization of services, with emphasis given on primary health care and gradual integration of vertical structures in the health care system; ii)infrastructure, with most health facilities refurbished and re-equipped; iii)training, through the development and implementation o f new curricula for family doctors, nurses and health care managers; and iv) the establishment o f a Health Information System (HIS), which is now inplace, although yet in a development stage. 5.18 The main difficulties have regarded: (i) lack of adequate managerial capacity and the tools, particularly at peripheral level; (ii) lack o f regulation and control o f the private the sector, includingthe private use o f public health facilities; (iii)the lack o f quality assurance strategies and systems; and (iv) the lack o f capacity to address emerging health issues, such as ** SeeTamburlini, 2004. 71 adolescent andyouth health. MoHis aware o f the problems and of the need o furgentpolicies to address them, starting with the rapidgrowth of a largelyunregulated private sector. 5.19 The lack of quality of health care affects several crucial areas, especially antenatal care. InKosovo, the lack o f quality, i.e. the failure of delivering the effective interventions, appears as an important problem, especially in areas where health interventions can make a major difference with respect to important health outcomes. A relevant example is given by perinatal care and more generally by reproductive health. For instance, an antenatal care survey conducted by UNICEF in 2003, involving 382 respondents, 66 percent rural and 34 percent urban, 69 percent client o f private services and 31percent o f public services, showed that accesduse was satisfactory but that quality as measuredby the proportion of visits that - includedthe procedures considered essential for good quality -was poor both in clients o f private andpublic services. 5.20 As far as delivery care and neonatal care is concerned, the situation seems to have improved with respect to the recent past. There is widespread concem, however, that in the absence o f a rapid establishment and enforcement of regulatory policies, there will be an increased shift to unregulated private practice with consequent possible deterioration of the standardso f perinatal care. 5.21 The quality o f child and youth health services deservesalso hrther attention. A recent study in 30 primaryhealth facilities showed that despite the great number o f healthworkers trained in the use o f effective diagnostic and treatment protocols for the common childhood illnesses, the quality o f care is still far from acceptable. There i s no information on the quality o f youth health services whose provision, anyway, has not gone beyond pilot experiences. 5.22 One reason for the low the quality of health services is the inadequacy and inefficiency in staffing allocation. Government employment inhealth amounts to 0.6 percent of the population, below the EUaverage o f 0.8 percent andthe OECD average o f 0.7 percent. While the number o f the administrative staff may be excessive, with a ratio o f support staff to professional staff (doctors andnurses) o f 0.37 in2003, the average ratio o f staff (professional and support) per hospital bed was only 1.15, which is low compared to the UK, the Scandinavian countries and the Netherlands, were the ratio may be as highas 2.7 to 5.0. The ratio of doctors (139/100,000) andnurses (361/100,000) to the population is very low relative to Westem Europe (respectively 346/100,000 and 777/100,000) and similar only to that o f Albania (139/100,000 and 391/100,000). The concentration o f doctors (40 percent) andnurses (40 percent) inhospitals i s below most European countries. It is clear, however, that there are too many doctors inrelation to nurses. The relationo fnurses to doctors averages 2.6 compare with 2.2 inWestem Europe. Butthe figures are not an indication o f overabundance o f doctors but rather of a shortage o f nurses, which may contribute to alter the overall quality of the health services providedto the population. 5.23 I n the education sector, concerns about the quality of schooling result predominantly from the presence of teachers with inadequate level of qualification. The major issue with the quality o f schooling in Kosovo, and thus with the effective delivery o f educational services to the population, is the quality o f teachers, probably more so than the quantity o f teachers. The pupil-teacher ratio, which indicates the average number of pupils servedby a teacher, andis one indicator for the quality o f schooling, compares favorably with other regions: 20 inprimaryschool, and 19 in secondary school. For comparison, the average pupil-teacher ratio in 2000 was 40 in low income countries; 22 in middle income countries and21 inupper middleincome countries. 5.24 In contrast, another proxy for the quality of schooling which raises concem is the large presence o f teachers with inadequate level o f qualification, a byproduct of the long- 72 lasting unofficial "parallel" system o f education. According to the HDIS, in 2003, 22.2 percent of all teachers had inadequate level o f qualification in primary schools, and 28.1 percent in secondary schools. The presence o f teachers with inappropriate qualification help explain the relatively low pupil-teacher ratio in Kosovo, since many teachers experience difficulties to teach multiple subjects in early levels of schooling. But there are also other reasonsrelated with local circumstances, such as the existence o f small rural primary schools that are instrumental to keep high enrollment rates in primary education and the needs o f ethnic communities. TargetingEfficiencyandPovertyImpactof PublicSocialAssistance 5.25 The social assistance scheme appears moderately effective, with little errors of inclusion but large errors of exclusion that reflect the cash limit. Using data from the 2002 HBS, Figure 9 looks at the effectiveness o f social assistance (i.e. excluding pensions) as measured by the coverage o f pre-social transfer poor, extremely poor and non-poor populations. The higher the coverage o f the poor andvery poor andthe lower the coverage o f the non-poor, the more effective are the social assistance benefits in reaching the needy population. In examining the findings, it should be noted that social assistance programs are supposedto provide benefits only to the "pre-social transfers" poor. 5.26 The data in Figure 9 suggests that social assistance benefits are at best only moderately effective inreaching the pre-transfer poor and extremely poor population, i.e. the coverage o f the poor and very poor is low: only some 22 percent o f the extremely poor declare to have received social assistance, and some 16percent o f the poor. At the same time, coverage o f the non-poor is even lower - only 4 percent o f the non-poor received social assistance over the period under examination. This would tendto indicate that targeting errors of inclusion are modest, while low coverage (higherrors o f exclusion) point to the categorical constraints that are imposed on the scheme as a function o f cost constraints. Figure 9: Targeting Effectiveness: Share of Pre-transfer Poor, Very Poor and Non-poor Receiving SocialAssistance Benefits(percent of households) 25 1 21.5 Poor Very poor Non-poor Source: H B S 200212003 5.27 Errors of exclusion reflect the effects of the cash limit. Proxy indicators for poor families were inpart chosen in order to limit the number o f eligible families to the available envelope, thus explicitly generating errors o f exclusion. Moreover, at the time the program was designed, statistically representative information on vulnerable groups was not available. Therefore, the scheme did not make a distinction between the poor and the extremely poor, nor did it attempt to raise well-being to some pre-defined level, i.e. no notional poverty line was set. Now the situation i s different, and ongoing household surveys do provide a basis for 73 identifying the poor and the extremely poor, and allow more informed decisions to be made about poverty targeting. The recently introduced revisions to eligibility do reflect the increased understanding o f the profile o f the poor and needy in Kosovo, that is discussed in Chapter 3 (Table 4). 5.28 To a largeextent, errors of inclusionmay reflect the fact that eligibility is determined mainly on the basis o f documentation presented by the applicant. It may also reflect the environment - it is particularly difficult for a benefit officer to establish need in an environment where there is a large informal sector and extensive family- and hnship-based safety nets. 5.29 Overall, the delivery of public social transfers has a large impact on poverty reduction, but the main incidence is provided by the basic pension while the impact of social assistance is more modest, as the result of large errors of exclusion and small amount of benejits. How significant are public transfers (social assistance and pensions) in alleviatingpoverty? To answer this question, apoverty impact analysis is conducted usingthe HBS. The poverty situation o f households is compared with and without transfers: A sensitivity analysis is included, based on a household's ability to compensate for the loss, or absence, o f social transfers at the 25, 50 and 75 percent level o f the value o f the transfer (Table 35). 5.30 The empirical evidence from the 2002 HBS suggests that social transfers are an important policy instrument for poverty reduction in Kosovo. If social transfers were eliminated, incidence o f poverty and extreme poverty would increase by 5.7 and 6.1 percentage points, respectively. Interestingly, the main influence on poverty incidence is provided by the basic pension: withdrawing the pension raises poverty and extreme poverty by4.4 and4.6 percentagepoints respectively. The impact o f social assistanceis more modest -anincreaseof1.4and1.7percentagepoints, respectively, forpovertyandextremepoverty. This may mainly be a reflection o f the relatively large size o f the pension as compared to the social assistance benefit - an elderly, pension-eligible couple would have been drawing 70 euros per month in2003, which is five euros more than a five-person householdwould have receivedinsocial assistance. 5.3 1 The elimination o f social transfers would not only push non-poor into poverty, but it would also increase depth and severity o f poverty. At the zero and 25 percent compensation levels, the poverty gap and severity o f poverty are significantly higher as a result o f the elimination o f social transfers. This indicate that social transfers may have an important poverty alleviating effect onhouseholds, even ifthey may not be sufficient to lift them out o f poverty. Again, the impact o f the basic pension is more marked than that o f the social assistancebenefit. Table 35: PovertyReductionImpactof SocialTransfers Percentage point changes in pre-transfer poverty indices` Social transfersb Pensions Social Assistance Poverty Extreme Poverty Severity Poverty Extreme PovertySeverity Poverty Extreme Poverty Severity poverty Gap poverty Gap poverty Gap 75 percent 2.7 1.4 1.2 0.8 2.0 1.0 0.8 0.6 0.7 0.4 0.4 0.3 50 percent 3.9 3.5 2.7 2.5 2.9 2.6 2.0 1.9 1.0 0.8 0.7 0.6 25 percent 5.0 5.1 4.4 5.2 3.8 3.8 3.3 4.3 1.3 1.3 1.1 0.9 Opercent 5.7 6.1 6.2 9.2 4.4 4.6 4.7 7.9 1.4 1.7 1.5 1.3 Source: Kosovo HBS 2002/2003. Note: "Pre-transfer rates" show the poverty incidence assuming different levels of compensation by households if a social transfer i s withdrawn. "Social transfers" show the combinedeffect of pensions and social assistance on poverty incidence. 74 C. EQUITY 5.32 Summay. While the effectiveness of public social delivery refers to the actual welfare gain associated with the intervention, an equitable delivery o f public services is generally defined as the capacity for all eligible beneficiaries to be provided with the same quantity andquality ofpublic services, regardlesso f their characteristics. Equitymatters inits ownrightfrom the view point o fpublic ethics, and also becausetogether with sufficiency and effectiveness, it is instrumental inmakingsocial servicespro-poor. This section reviews some o f the successes and challenges with ensuring o f an equitable delivery o f public social services in Kosovo. In the health sector, the major issue i s the existence o f undue out-of- pocket payment that disproportionately hit the poorest. In the education sector, public funds seem to equally benefit the poor and the non-poor, but there could be large disparities in the quality o f schooling by income level. With respect to social assistance, extreme poor Serb households and extreme poor households with more than four children appear disadvantaged for reasonsthat remainto be explained. 5.33 In the health sector, equity in service deliverymay be compromisedby management inefficiency and corruption that cause undue out-of-pocket expenditures and disproportionatelyhit the poorest. The health system in Kosovo i s publicly financed and i s meant to provide a universal coverage with a generous exemption policy to the co-payment principle.As a result, there should not be a significant proportion o f the population reporting excessive costs as a mainbarrier to access. However, the benefits o f the system, especially for the poorest, are compromised by phenomena o f mismanagement and corruption that cause undueout-of-pocket payments even whenhealthcare is supposedto beprovidedfor free. 5.34 A major concern is the unnecessaryprescriptiono f drugs that are not included inthe list of drugs that are provided for free, and the fact that essential drugs supposed to be for free are often not available in the Health Centre or Hospital pharmacy and have to be bought. According to the HBS, the cost o f drugs account for 65 percent o f the private expenditure, but a high proportion o f these out-of-pocket payments are for drugs that are supposed to be provided for free, or that are not essential and therefore shouldnot beprescribed. 5.35 Informal payments are also widespread and hit the poorest patients worst. In 2000, according to the LSMS, patients reported paying more for the use o f public health facilities thanwouldbe expectedifthe only chargesby the facilities were the officially sanctioned co- payments. The difference was due to unofficial under-the-table payments to the providers. This situation does not seem to have changed much over the last few years. Inthe 2003 EWR, 53 percent o fthe Albanian respondents said that they hadhadto makeunofficialpayments for health services. The comparatively high level o f finance passing through the unregulated private and 'informal' sectors does not support the implementation o f a rational health strategy and underminesefforts to achieve sustainability and improve service delivery. Private sector regulationis thus a priority to avoid further leaking o f resources from the public sector and to protect consumers from poor quality anddouble standards. 5.36 I n the education sector, while a priori public funds allocated to basic education seems to equally benefit poor and non-poor children, there seems to be large disparities between thepoor and non-poor in the quality of teaching. Table 36 presents the incidence o f public spending on basic education in 2002 based on the HBS. For comparison, it also presents incidence of public spending based on the 2000 LSMS and reported in the 2001 Kosovo Poverty Assessment. As already mentioned, it i s not possible to deduce enrollment rates for secondary and higher education with the 2002 HBS, and thus the incidence o f spending figures can be conducted only for basic education. The results show that a fairly equal distribution of public funds allocated to basic education across poor and non-poor 75 households which i s largely due to relatively highenrollment rates inbasic education even for children coming from poor households. Yet, it i s impossible to know from these aggregate data whether schools attendedbypoor andnon-poor children receive similar amount o fpublic fundsper child. Table 36: Poverty IncidenceofPublic Spending on BasicEducation,2002 (percent of households) - q__ HouseholdConsumptionQuintiles 1 2 3 4 5 (poorest) (richest) 2002 (HBS) Basic education(9 years) 18.0 17.4 20.9 22.3 21.3 2000 (LSMS) Basic education(9 years) 19.7 21.8 21.2 18.8 18.4 Secondary 13.5 18.1 22.4 18.5 27.6 Higher 16.8 19.7 20.9 19.5 23.0 5.37 Equity concerns with the quality of education emerge from the HDIS data that show that not only there are a large numberso f teachers whose level o f qualification ininadequate inKOSOVO,that there are adisproportionately largepercentageofunder-qualifiedteachers but inmunicipalities where the incidence ofpoverty is larger(Figure 10). Thus, these disparities between poorer and richer municipalities would mean that de facto a higher share of poor children is at risk o f being trained by less-qualified teachers and may receive lower-quality education services. Figure 10: Percentageof Teachers in Primary and Secondary Schools with Inadequate Qualificationby Poverty at the MunicipalityLevel 75 75 ** * + e5 65 * * * 55 s! te645 * * 35 L 35 25 * * * , 15 2: 5 5 10 15 20 25 30 35 40 45 15 20 25 30 35 40 45 50 55 Percentageof unquallfied primaryschool teachers Percentageof unqualified sacondaryschodteachers Source: 2003 HDIS andthe EducationinKosovo2000/01, Kosovo EducationCenter. 5.38 With respect to access to social assistance, Serb households and households with more than four children appear disadvantagedfor reasons that remain to be explained. Targeting effectiveness can also be examined from an equity perspective. Table 37 looks at the distribution o f social assistance among poor and very poor households by ethnicity, gender ofhousehold head, andnumber o f children ina household - characteristics that usually show strong correlation with poverty. 76 5.39 From the point of view o f ethnicity, non-Albanian and non-Serb "other" ethnic groups display higher poverty incidence, and the share o f such poor and extremely poor families receiving a social assistance benefit exceeds the national average. Likewise, the share o f poor and extremely poor Albanian families receiving social assistance benefits is above average. The Serbs, on the other hand, appear to be particularly disadvantaged, with only 5 percent of poor Serbs and 4 percent of extremely poor Serbs receiving a social assistance benefit. It would be important to examine more closely the reasons for this remarkable difference intargeting o f social assistance. Table 37: SocialAssistanceandIndividualsCharacteristicsofRecipientsAmong the Poor and ExtremePoor (percent of households) Share of Poor ReceivingSA Share of Extremely Poor Rec. SA All 15.7 21.5 Albanian 16.3 22.4 Serb 5.0 3.9 Other 16.4 25.4 Female-headed households 20.2 24.1 Householdswith no children 10.9 18.9 1-2children 16.7 24.6 3-4 children 20.5 26.2 5 andmore children 10.2 11.9 Source: HBS2002/2003 5.40 The social assistance benefit appears to be relatively well targeted at poor and extremely poor female-headed households. The share o f poor and extremely poor female- headed households receiving social assistance is 4.5 and 2.6 percentage points higher than the national average, respectively. 5.41 A higher than average share o f poor and extremely poor households with up to four children receive social assistance benefits, and targeting improves as the number of children inthe household increases. However, when the number of children exceeds four, the share of poor and extremely poor families receiving social assistance benefits drops significantly below the national average. Again, it would be useful to examine why targeting declines with regard to families with over four children. 77 6. POLICY DIRECTIONS 6.1 The previous chapters have sought to contribute to the on-going poverty dialogue in Kosovo. The statistical and institutional capacity for poverty monitoring and analysis was reviewed in Chapter 2. An assessment o f the multidimensional nature o f poverty, the profile o f the income poor, and the poverty linkages were conducted in Chapter 3. An analysis o f social cohesion was provided inChapter 4. The pro-poor potential o f public service delivery was reviewed in Chapter 5. The fmdings show that the lack o f opportunities, security and participation in decision-making are major aspects o f poverty inKosovo. Income poverty is found to disproportionately affect children, the elderly, female-headed households, the disabled, non-Serb ethnic minorities, the unemployed, and precarious jobholders. Besides income poverty, there are also large disparities ineconomic and social wellbeingby location, gender, age, and ethnicity. Finally, flaws in the provision of public social services are identified in the area of sufficiency, effectiveness and equity, and may contribute to the reduction o f the pro-poor potential o fpublic service delivery. 6.2 The aim of this final chapter is to turn some o f the previous fmdings into options for public actions. This chapter does not attempt to provide a comprehensive strategy for poverty reduction. Rather, it tries to be selective by supporting the formulation and prioritization o f selectedpolicy options with the understanding that these would still have to be decided by the Kosovars themselves, and consistently articulated with a broader Kosovo Development Plan, behind which donor assistance will be aligned. A major conclusion of this chapter is that policy makers would need to think o f developing a multi-pronged strategy reflecting the various dimensions o f poverty and the need for more robust evidence on the nature and the changes o f poverty over time. To be realistic, this multi-pronged strategy would need to take into account the context o f tight fiscal constraints in which public programs need to be developed. As sucha strategy i s developed, it would be necessary to ensure that it i s properly costed and affordable. This would require proper prioritization and an appropriate sequencing o f the proposed reforms and interventions. Growth-enhancing and efficiency reforms should be at the center of this poverty reduction strategy. This would make the State more effective and provide additional fiscal space for public interventions. Finally, it is important to recognize that in Kosovo, the prospects for growth and poverty reduction will crucially depend not just on domestic policies but also on decisions made by the intemational community. Kosovo still needs the support o f the intemational community to address the numerous poverty challenges ahead. This support i s likely to become more effective if embeddedwithin the framework o fa KosovoDevelopmentPlan. A. PROMOTINGINCOME OPPORTUNITIES 6.3 Kosovo faces an important poverty challenge in the coming years, requiring a strong commitment by national policymakers to implement a policy program that sustains and accelerates broad-basedgrowth. International evidence indicates that domestic policies are crucial for sustaining growth and generating income opportunities. These include prudent macroeconomic and fiscal management, policies that allow a stable and predictable environment for private sector activity, and measuresthat contribute to social stability. 6.4 Working toward the clarification of Kosovo's political status is crucial for sustaining growth and generating income opportunities. The 2004 World Bank Kosovo 78 Economic Memorandum indicates that the recent growth performance has been driven by a post-conflict boom financed by official development aids flows and it i s unlikely to be sustainable without a clarification o f Kosovo's political status and the maintenance o f peace and security. Considering that the resolution of Kosovo's political status is pendingupon the respect o f a set o f "Standards for Kosovo", the priority for policymakers should be to implement measures that will speed up the progress toward the achievement o f these standards. Measures should include implementing a policy program that supports the establishment o f political and institutional stability and the rule o f law, and that promotes a favorable environment for private sector led growth, as spelledout inthe 2004 KEM. 6.5 Making growth more labor friendly is equally important. The quality o f growth matters. The links between economic growth and the creation o f decent jobs need to be significantly enhanced. So far, Kosovo has experienced what is often called a "jobless" recovery - growth with little job creation despite a very unregulated labor market and low - payroll taxes. International evidence suggests that facilitating competition in product and fmancial markets, supporting the development of small and medium size enterprises, and supporting exports may altogether play an important role in generating employment. In KOSOVO, a sound legal framework for free-market has already beenput inplace, the business environment for domestic firms is favorable, and the Kosovo Micro Credit Bank plays an important role incapitalizing small businesses. Yet, prospects for future reduction inpoverty will crucially depend on the implementation o f reforms and policies that will make Kosovo more attractive to FDIandwill promote agricultural competitiveness. 6.6 Enhancing social cohesion is essential to sustain economic growth and make growth pro-poor. This report further illustrates through micro-simulations that growth with equity matters for poverty reduction. Large socio-economic inequality across location, gender, ethnicity, and age, along with the existence o f parallel systems for the provision o f social services to Serb enclaves needto be addressedby policy makers as they can contribute to discontent and political unrest and ultimately threaten the sustainability o f growth. Excessive socio-economic disparities may also reduce and even compromise the impact o f growth on poverty reduction. A more equitable growth could have a multiplier effect on poverty reduction. To enhance social cohesion, policymakers would need to address these inequities and work towarda unifiedsocial protection system. 6.7 The prospectsfor growth and poverty reduction will crucially depend not just on domestic policies but also on decisions made by the international community. Sound domestic policies are crucial for sustaining growth and attacking poverty, but it i s equally important to acknowledge that, in the short-term, the prospects for growth and poverty reduction in Kosovo will still be heavily influenced by the level o f donor assistance, the clarification by the intemational community o f Kosovo's political status as well as trade and regional stability. An important concern is that delays in the settlement o f Kosovo's status combined with a rapid decline indonor assistance are likely to have severe adverse social and economic repercussions on the population o f Kosovo. B. RAISING HEALTHAND EDUCATIONABILITIES 6.8 Reconciling a prudent macroeconomic andfiscal managementframework with the protection of essential investments in human capital is a challenge that needs to be addressed. The prospects o f an increase in health and education abilities over the next years will dependuponpromoting essential productive investments inhumancapital. The fmancial support o f the intemational community i s essential, as Kosovo remains in great need o f investments inhealth and education as well as inbasic infrastructure services such as water supply, sanitation and waste management which have important implications for social outcomes. Given the level o f uncertainty regarding donor assistance, it is important for 79 national policy makers to undertake the costing o f key pro-poor public policies inhealth and education, establish priorities, and identlfy which interventions will be financed from own resources and can be fiscally sustainable, and which ones would need donor assistance.This costing and prioritization exercise could be developed within the framework o f a Kosovo Development Plan developed by ESPIG. 6.9 Improving the efficiency and equity of public health expenditures. While overall public expenditures onhealth are low inrelation to neighboring countries and domestic needs, fiscal constraints may limit the scope for immediate increase. This makes it even more important for the Government to focus on improving the efficiency o f public expenditure so as to generate savings that can be reinvested to improve the overall effectiveness and equity o f the system. While the forthcoming World Bank Public Expenditure Review will carefully examine issues o f efficiency and fiscal sustainability o f public spending, this report argues that efficiency and equity gains can be realized essentially through a more rational use o f drugs, the reduction o f informal payments, private sector regulation, and a better balance betweenmedical andnonmedicalpersonnel. 6.10 Improving the quality and access by the poor to education. To break down the vicious cycle o f poverty, and avoid reinforcing the disadvantages that poor children already face, policymakers should aim at reducing the disparities between municipalities in the number o f teachers with inappropriate qualification. Additionally, it would be important to expand the coverage o fpre-school education by transferring some teachersto pre-school level and to rationalize the proportion of qualified and less qualified teachers at different levels o f education consistent with desirable and affordable pupilheacher ratios. Finally, while access to primary schools is almost universal, distance to secondary schools emerges as a major barrier to education in rural areas, especially for girls, and it would require particular attention. C. REDUCINGVULNERABILITY 6.11 Reducing environmental risks through continuous efforts on the rehabilitation of basic infrastructure services and the enforcement of environmental safeguards. The focus of policymakers should be on: (i)improving access to safe drinking water, through maintenance, protection and disinfection o f sources o f water systems and wells, and knowledge o f other hygienic and preventive behaviors; (ii) improving sanitation conditions, by improving the handlingo f wastewater and encouraging hygienic behavior; (iii) improving waste management, particularly hazardouswaste; and (iv) strengtheninginstitutional capacity for environmental management, especially to reduce illegal logging and building, and for ensuring stronger compliance with housing security codes that are intended to mitigate the impact o f seismological shocks ina region where the risk o f earthquakesi s rated as relatively high.At the same time, policies and incentives shouldbe developed to ensure that economic growth is environmentally responsible. 6.12 Supporting conflict prevention. An important conclusion o f this report is that inter- ethnic tensions between Albanians and Serbs cannot be explained by socio-economic disparities. Yet, ethnic tensions remain high and need to be addressed by policymakers in a comprehensive manner. While it is crucial to continue with the reconstruction efforts undertaken since the 1999 conflict, and repair the damaged o f the March 2004 riots, it is equally urgent to take measures to prevent further conflicts. Kosovo has already put inplace pluralist institutions that support minority rights and provide the institutional basis for peacefil conflict resolution. Yet, additionalmeasuresmay need to be considered, inparticular efforts to get different ethnic groups - especially among children and youth - to interact with each other, and supporting the role that women often play in many peace keeping and reconstructionefforts. 80 extremeAdjusting 6.13 the nascent social assistance system so as to improve the coverage of the poor. Given the cash limit in Kosovo, policymakers may wish to rethink the objective o f the public social assistance program, aiming it at the extremely poor, with the purpose o f raising their living standards at least to the food poverty line, while leaving the situation o f the "non-extreme poor'' to be tackled through family-based safety nets and the overall growth strategy. With gains in targeting efficiency and improved equity, extreme poverty could be eliminated at no additional costs. An extremely poor individualwould need anaverage of 84 eurosper year to reachthe foodpoverty line. Assumingperfect targeting, the cost o f a social assistanceprogram designedto eliminate extreme poverty would be around 25 million euros, or 1.9 percent o f GDP (2003). With a more realistic assumption o f a 25 percent leakage, the cost o f social assistancewould rise to 31million euros, or 2.3 percent o f GDP. In 2003, Kosovo allocated some 32 million euros for the social assistance program, or 2.5 percent o f GDP inthat year. Today, the possibilities o f identlfylng very poor households are much improved since the social assistance benefit was originally introduced. Recent household budget surveys should allow strong poverty correlates to be established, and improvements underway in the administration of the benefit should gradually allow continuous field-level adjustment o f criteria to improve targeting. 6.14 Designing preventive measures to protect individuals with the highest risk of becoming unemployed. The evidence shows that joblessness is disproportionately concentrated among the unskilled and first-job seekers, and thus indicates the importance o f designing preventive measures to reduce the risk of becoming unemployed among these groups. An aggressive program needs to be implemented that will raise the skills and education o f the population and establish mechanisms to facilitate the transition from school to work. 6.15 Considering the introduction of mitigation mechanisms for individuals that are likely to lose their jobs as the result of enterprise restructuring. A realistic social protection strategy for Kosovo will need to acknowledge that current budget resources may be too limited for strengthening substantially the labor market aspects o f the safety net (e. g. unemployment benefits, active labor market programs). Reducing vulnerability to income poverty among the unemployed would need to be addressed primarily through a better- targeted social assistance benefit. Yet, the employment impact o f enterprise restructuringand privatization may soon increase the demand for active labor market measures, increase political pressure, and require a stronger engagement o f policymakers. International experience shows that while several active labor market programs have failed to improve significantly the employment prospects o f the unemployed, some integrated and well targeted programs have been successful. One option that policymakers could consider is to introduce training programs targeted to workers affected by enterprise restructuring. To be cost- effective, these programs would need to be connected with well-identified labor market needs andtargeted to individuals that are neither too youngnortoo old. D. PROMOTING EMPOWERMENT 6.16 There is an urgent need to improve the level of knowledge of youth reproductive health, HW/AIDS, other STIs, and safe sex practices. The Government and many NGOs have made many efforts in trying to increase the knowledge about HIV/AIDS and other sexually transmitted infections among youth. Yet, the evidence brought inthis report suggests that the level o f knowledge o f youth remains very low and that youth still experience difficulties to access quality information about safe sex practices. This should be a major concem for policymakers given that Kosovo faces a high risk for a rapid spread o f HIV/AIDS.Policymakers cando a lotby supporting awareness campaigns onHIVprevention targetedto the youth andledby the youth, andinfluencingpublic debateto reverse the social 81 stigma associated with HIV/AIDS. Other issues to be addressed include: (i) supporting voluntary counseling and testing services; (ii) easing access o f the HIV/STI services among highrisk groups; and (iii) increasing capacities for provision o f the HAART and CD4 count andviral loadmonitoringfor the PLWHA. 6.17 Increasing awareness on environmental issues is equally urgent, especially in environmental hot spots. Despitethe importance of the problem, public opinion polls reveal that people in Kosovo still show a very low concem about environmental depletion and the need to preserve public spaces from environmental degradation. Educational programs and awarenesscampaigns are criticalto provide people ingeneral, and the poor inparticular, with effective means to cope with, or mitigate environmental risks and increase their income, health and personal security. Policymakers should thus consider implementing a program o f environmental awareness all over the territory, with a specific focus on Mitrovica and Prishtinaregions, where major industrialpolluters are concentrated. 6.18 Promoting women's influence in the economic, social and political life is essential to enhancing the status of women in Kosovo and reducing gender inequity. This report has shown that not only women in Kosovo are disadvantaged in terms o f lower education achievements, high maternal mortality, worse labor market outcomes, and a greater exposure to domestic violence and human trafficking, but also that despite increased representation in formal state institutions, their effective involvement in decision making remains at stake. A legal framework already exists at both elected and non-elected government that promote gender equity, but other direct public measures may need to be considered to raise women's empowerment, including: (i) using public resources to subsidize girl's education and/or tighteninglinks between social assistanceandschool attendance; (ii) supporting microfinance schemes for women; and (iii)working toward a better representation o f women in government leadership positions. 6.19 Supporting youth voices. Policymakers would need to: (i)encourage the development o f youth policies in collaboration with national and local youth organizations; (ii) consultativegroupswithyouthorganizationsonissuesthataffectyouth;and(iii) establish support youth servicing NGOs, especially inthe area o f political right, health counseling, and business skills development. E. SUPPORTINGAN EVIDENCE-BASED DIALOGUEONPOVERTY 6.20 What is needed is to ensure that the poverty dialogue becomes more constructive and contributes to the mainstreaming of poverty issues into government policy? The PISG andUNMIKestablished inMay2003 the Economic Strategy andProject Identification Group (ESPIG) with the objective o f developing the PISG's capacity to formulate a comprehensive economic strategy and plan future investments. Despite its inter-ministerial setting, ESPIG does not have, at the moment, a representation o f the social sector ministries and does not have a specific focus on poverty reduction. As Kosovo gradually embarks on the process o f building a Kosovo Development Plan, ways o f strengthening the poverty focus within such a plan will need to be explored. Policymakers could support: (i) inclusion of social the ministries in ESPIG; and (ii) set-up o f consultative groups to give civil society groups a the great voice inthe development o fthe NDP. 6.21 Improving the statistical infrastructure isfurther essential in establishinga poverty dialogue based on robust evidence.I n thefirst place, this requires conducting a population census. To maintain and improve upon the quality o f the sampling frame methodology utilized inthe household survey program, it i s crucial to conduct a population census. This would allow: (i) obtaining a current population estimate o f individuals and households; and (ii) thesamplingframeofbothprimarysamplingunits(PSUs)andhouseholdunitsin refining 82 any new surveys. Inaddition, it is recommended that when publishingits poverty estimates, SOK starts reporting standarderrors that account for the stratified samplingdesign effect (the most recent version o f SPSS incorporates new features that facilitate the computation o f corrected standard errors) 6.22 Important knowledge gaps on key socio-economic outcomes and their distribution across different population groups need to be fdled-in within SOK's current survey program. To improve the capacity for poverty monitoring, SOK may consider: (i) improving inthe LFS the coverage of the informal sector and adjust timing to reflect agricultural employment activities; (ii) monitoring morepreciselyinthe H B S keynon-income dimensions of poverty; and (iii) disaggregating inpublished statistics all relevant indicators by gender, ethnicity, age categories, rurayurban location, and income groups. 6.23 A decision needs to be taken as to the overall approachfor monitoringpoverty with the household statistics system. There are different approaches that can be taken to improve the capacity of poverty monitoring with SOK household statistics system: (i) an LSMS-type survey, which would offer in-depthcoverage of important socioeconomic indicators, but may not be sustainable in the long run given the financial and human resources necessary to operate such a survey; and (ii) H B S format that incorporates rotating modules to explore an critical indicators in-depth while maintaining core questions that are repeated ineach survey round. 6.24 Introducing a core HBS survey with rotating modules may offer a holistic approach to monitoring multi-dimensional aspects of poverty that is inherently more sustainable. An improved HBS format as discussed in Box 2 with rotating modules may be preferable to the LSMS-type option for sustainability reasons. 6.25 The monitoring of critical socio-economic outcomes requires unfettered access to household survey data among government ministries and thepolicy community. Affording access to non-identifiable household-level data among government ministries and other policy makers is essential to promote transparency in poverty analysis, ease data analysis and improve the production o f statistical data. To ensure access, SOK may needto establish a data program. 6.26 Creating a data user group can also stimulate theproduction of relevant and timely socio-economic data. A data user group may be created, in the spirit o f the Data Users Groups implemented by DBD in Bosnia Herzegovina, which could include line ministries, policy makers, the international donor community, and academia. Such a group would create real demand for statistical data and eventually financial and technical support for SOK in developing and analyzing the data. Without communication and attachment to data users, SOK could risk producing irrelevant or inadequate data about its population. Data users can articulate needs and provide continual feedback about the type, amount, and quality o f the prevailing data. The involvement o f the international donor community in data user groups may also invite the possibilityo f financial support for the development o f key modules within arotatingmodule survey. However, the singular interest o f a particular donor or policy maker should not completely determine the type o f data collected and analyzed within a particular module. Rather, it should serve as aninput into the decision-making process. 83 SELECTEDREFERENCES Backgroundpapers Owens, Janet (2004a): "The Household Survey PrograminKosovo Improving the Systemfor Poverty Monitoring". Owens, Janet (2004b): "Gender, Ethnicityand PovertyinKOSOVO" PRISMresearch (2004): "KOSOVO Qualitative PovertyAssessment" Sederlof, Hjalte (2004): "KOSOVO:The Public Safety Net. An Analysis of its efficiency and effectiveness". Tamborrini, Tiziana (2004): Poverty and EnvironmentinKOSOVO". " Tamburlini, Giorgio (2004) HealthandPovertyinKosovo". " Tsirunyan,Sasun(2004): "Poverty andInequalityinKosovoyy. Verme, Paolo (2004): "Multidimensional poverty and social cohesioninKosovo". Verme, Paolo (2004): "Poverty and the life cycle inKOSOVO". OfficialWorldBankdocuments World Bank (2000): World Development Report 2000/2001: Attacking Poverty. Washington DC World Bank(2001): Kosovo Poverty Assessment; ReportNo 23390-KOS, WashingtonDC. World Bank (2002a): Kosovo, FR Yugoslavia: Medium-TermPublic Expenditure Priorities; ReportNo 24880-KOS. World Bank (2002b): Kosovo, Federal Republic of Yugoslavia: Transitional Support Strategy. ReportN o 24275-KOS. WashingtonDC. World Bank(2003b): TrackingNon-Income Dimensions of Poverty in ECA, WashingtonDC. World Bank (2003d): Kosovo Labor Market Study: Policy Challenges of Formal and Informal Employment. ReportNo 25990-KOS. Washington DC. World Bank (2004a): Kosovo Economic Memorandum. Report No 28023-KOS. Washington DC. World Bank (2004b): Kosovo, Federal Republic of Yugoslavia: Transitional Support Strategy. ReportNo 28006-KOS. WashingtonDC. 84 World Bank (2004~):YoungPeople inSouthEasternEurope: FromRiskto Empowerment. World Bank (2004d): Human Trafickingpom, through and to South East Europe: A Stock- Taking and Scoping Exercise. Social Development Unit, Europe and Central Asia Region, Washington, DC. 85 . 1 e .. 0 . e m e e e e e e e 00 00 e .I I I 4 FI C c , .iE .- . .- . .-.-. e . . . e e e e e . e . . . . . e e . a a I 0 * . 0 . * . . . . e . . . Table 39: Non-age Specific Quantitative MultidimensionalPovertyIndicators 92 Table 39: Continued 2000 2001 I 2002 2003 Security Environmentalvulnerability K o fHHwith secured (owned) tenure 1.6a K o f HH with more than 3 persons per room (crowding .ndicator) % o f HHwith major damaged houses K o fHHwith access to improvedsanitation: iv. Connection to the sewage system or availability o f a septic can v. Available in-house flush toilet *49 a vi. Protected outside pit latrine K o f HHwith access to improvedsource o fwater: 1.4 li.I54.1 iv. Available regular connection to the Central Pipeline Pipe Water ii.10.9 iv. 3 v. Public standpipe vi. Protected well I -1 3/0 o f HHwith access to solid waste collection % o fHHwithgarbage dumpedhuriedbumed 62.8a % o f HH who depend directly on natural resources for their Livelihood I I % o f HHusing solid fuels (wood, coal) 194.9 Proportiono fLand area covered by forests I National protectedarea (percent o f total land area) 4.27 e Number o f endangered species Uncontrolledlogging GDP per unit o f energy use (US$/Kgoe) C02 t per capita Level o flead and other h e a w metals (inair. soil and foodstuff II ~, Ambient concentration o f pollutants in urban areas Level o f other toxic gasses inthe air Rate o fillegal building growing Income insecurity I I % o f HHwith only one employed working age member % o f extreme poor HHnot receivingpublic social assistance % o f extreme poor HHnot receivingprivatetransfers I 82.2 Isolation % o f female headedhouseholds I 4.7O I I 93 Table39: Continued 2003; UNECE, 2002; World Conservation Union (IUCN Red List of Threatened Plant, UNECE e 2002 Estimates from EIA conductedin 2002 (EAR, 2003); Human Development Indicators Survey b (HDIS); ' Labor Force Survey (LFS), Demographicand Health Survey (DHS);kNIPHK, Ministry of ' Health (MoH); Statistical Office of Kosovo (SOK); Kosovo Early Waming Report (KEWR); UNMIKpolicereports; cases reportedto Centrefor Protection of Women and Children, 2000-2003 94 Education Pre-schoolenrollment rates (3-6) Primary enrolmentrate (6- 91.7' 9O.Saa95.4d 14) Rate of children 73e completing compulsory education Ratio of girls to boys in 0.92' 0.92' primary General secondary 59.5' 75.2 enrolments Ratio of girls to boys in 0.72' 0.79' secondary Girls reachinggrade 9 58.1' Highereducation 17.3' 19.Sd enrolments Ratio of girls to boys in 0.79 0.82 95 Riskof exploitation or worst forms of labor Incidenceof child labor 4.5c (age 10-14) % of wage employedwith no 30.2b Source:`HBS; bLFS;`LSMS. Table 42: Poverty IndicatorsWithin the Life Cycle: Empowerment Source:aPSI/KApSurvey, as reportedinthe UNICEF "Youth inKosovo" report; b KEarly~Warning Reports. ~ ~ ~ ~ 96