Report No. 34324-MK FYR of Macedonia Poverty Assessment for 2002­2003 November 15, 2005 Poverty Reduction and Economic Management Unit Europe and Central Asia Region Document of the World Bank Table of Contents Acknowledgments.............................................................................................................. Key Findings..................................................................................................................... vi Executive Summary......................................................................................................... viii vi1 ..... CHAPTER 1 Introduction..................................................................................................................... OVERVIEW OF LIVING CONDITIONS INFYRMACEDONIA ..... 1 Development Indicators.................................................................................................. 1 2 Social Indicators and Access to Infrastructure................................................................ Conclusions..................................................................................................................... 3 6 ISSUES AND KEY RESULTS .......................................................................................... CHAPTER 2 POVERTY TRENDS INMACEDONIA: METHODOLOGICAL 7 7 Measuring Well-being .................................................................................................... Introduction..................................................................................................................... 7 Who i s Poor and Who I s Not - Choosing a Poverty Threshold ................................... 10 Multi-dimensional Poverty ........................................................................................... Poverty Measures: How Many Poor and How Poor are They?................................... 15 20 Summary ....................................................................................................................... 21 CHAPTER 3 22 Introduction.,................................................................................................................. A PROFILEOF THE POOR IN2003.................................................. 22 HouseholdExpenditures............................................................................................... 22 Household Location, Size, and Demographics............................................................. 23 25 HouseholdHead Characteristics: Gender, Age, andEducation ................................... Housing Conditions and Infrastructure......................................................................... 26 Are There Pockets o f Poverty inMacedonia? .............................................................. Income and Labor Markets ........................................................................................... 27 31 CHAPTER 4 Links BetweenGrowth and Poverty............................................................................. POVERTY, GROWTH, AND INEQUALITY .................................... 32 32 Key economic Trends inMacedonia: 1996-2003 ......................................................... 32 Labor Market Transmission.......................................................................................... Growth-Poverty Elasticity and Inequality .................................................................... 34 Fiscal Redistribution..................................................................................................... 37 40 STATISTICAL ANNEX ................................................................................................ Conclusions ................................................................................................................... 42 52 Poverty lines and Foster, Green and Thorbecke Poverty Indices................................. 52 HouseholdLiving Conditions by Poverty Status.......................................................... 55 Employment Information.............................................................................................. 58 Consumption and Income ............................................................................................. 59 Subjective Poverty ........................................................................................................ 60 List of Tables Table 1.1:Macedonia: Progress on MillenniumDevelopment Goals............................................ 2 11 Table 2.2: Cost-of-basic-needs Poverty Lines for Macedonia, 2002 and 2003 ............................. Table 2.1: Problems with Relative Poverty Lines ......................................................................... 14 Table 2.3: Poverty Headcount Ratio, Poverty Gap Index and Poverty Severity Index (inpercent) ....................................................................................................................................................... 16 111 ... Table 2.4: Poverty Headcount Index for Per Capita Consumption and Equivalence Scales ..... 18 " Table 3.1:Food Consumption Dominates Expenditures of the Poor ............................................ Table 2.5: One and Two Dollar a Day Poverty i s Low inMacedonia........................................... 19 24 26 Table 3.3 :Poverty Ratesby Female and Male-headed Households.............................................. Table 3.2: Housing Conditions of the Poor and Non-poor............................................................ 26 Table 3.4: The Role of InformalEmployment and Social Security for Poorer Households ........-27 Table 3.5: Participation Rates of the Poor are Low and Unemployment i s High.......................... 29 Table 4.1:The Private Sector Still Accounts for Less than Halfof Non-agriculturalEmployment ....................................................................................................................................................... 38 Table 4.2: Workers didnot Move to Higher Productivity Sectorsbetween2002 and 2003 .........39 Table 4.3: The Rural Poor Movedto Better Employment Situations, the UrbanPoor to Worse ..40 Listof Figures Figure ES 1: Income and Social Indicators..................................................................................... i x Figure ES 3: Consumptionand Multi-Dimensional Poverty.......................................................... Figure ES 2: International Comparisons of Poverty inFYRMacedonia, 2003 .............................. x x i Figure ES4: Falling Living StandardsandRisingPoverty........................................................... xii Figure ES 5: Poverty Rates With and Without Social Insurance and Welfare Programs............xiv Figure 1.1: How do health indicators compare with other ECA Countries? ................................... 3 Figure 1.2: International Comparison of Secondary and Tertiary Education.................................. 4 Figure 1.3: International Comparison of Unemployment and Productivity .................................... 5 Figure 1.4: Accessibility and Reliability of Infrastructure Services................................................ 6 Figure 2.1: Problems with Relative Poverty Lines ........................................................................ 11 Figure 2.2: Subjective Poverty Increaseswith Welfare................................................................. 13 19 Figure 2.4: MultipleDimensions of Poverty inFYRMacedonia2003......................................... Figure 2.3: Compared to ECA Countries, FYRMacedonia has Low Levels of Poverty ..............20 Figure 3.1: Larger Households, Especiallywith Many Young Children, are Significantly Poorer than Others..................................................................................................................................... Figure 3.2: PovertyHeadcount Indicesby Age Group, 2003........................................................ 24 Figure 3.3: Poverty Headcount Indicesby EducationLevel of HouseholdHead, 2003 ...............25 Figure 3.5: The Poor have a HighProportion of Occasional Workers .......................................... Figure 3.4: Income Opportunities of the Poor are Hampered by LimitedAccess to Good Jobs..27 28 Figure4.1: Growthhasnot Recuperated,butthe EconomyhasRemainedStable........................ 30 32 Figure4.2: The Economy Shows No Signs ofRestructuring ........................................................ 33 34 Figure 4.4: Poverty, Actual and Simulated, 1996-2003................................................................. Figure4.3: InvestmentandFDILevels are Low by InternationalComparison............................ 35 Figure 4.5: The Poor saw a Reduction inTheir Share of Total Expenditures, .............................. Figure 4.6: The Labor Market Situationhas Worsened Over Time .............................................. 36 37 Figure4.8: Program Coverage and Adequacy ofBenefitsby PopulationQuintile....................... Figure4.7: Poverty RatesWith and Without Social Insurance and Welfare Programs ...............41 42 ListofBoxes Box 2.1: What Do We Meanby Poverty?....................................................................................... 8 Box 3.1: Individual Poverty and Household Characteristics ......................................................... Box 2.3: Equivalence Scales Usedinthe FYRMacedonia Poverty Assessment 2005 .................17 Box 2.2: Monitoring Poverty: The Foster, Greer and Thorbecke Measures of Poverty...............15 23 Box 3.2: Labor Market Indicators inthe FYRMacedonia Poverty Assessment ........................... 30 iv Box4.1:Comparabilityof SurveyData ....,,,,.....,, ..,...,...........,......,.......,,....,,............. , ......,.......,...35 Box 4.2: Why Does GDP GrowthNotTrickleDownto HouseholdExpenditures? ..................... 36 V ACKNOWLEDGMENTS This report is a joint product o f the State Statistical Office of FYR Macedonia and the World Bank. It i s the culmination o f a process which began in 2002 when the two institutions agreed to work collaboratively in order to understand the poverty situation based upon the availability o f better data that would allow drawing robust conclusions on living standards. However, this work represents only the first phase o f cooperation between the World Bank and Government o f FYR Macedonia on this important topic and will be followed by additional in depth work on labor markets inthe coming year and access to social services with the availability o f a richer dataset in2006. The leader o f the team from the State Statistical Office i s Ajrija Causovska (Head), with key members including Snezana Sipovic and Stase Nolev. The work was carried out under the auspices o f the statistical agency's Director, Katerina Daskalova. Not only were the advice and guidance provided by this team on the drafting o f the report critical to its production, but the statistical team has laid the groundwork for improvement in the household data through improvement in the survey questionnaire and broadening the indicators for assessing poverty in the country. The task on the World Bank side was carried out under the leadership o f Sarosh Sattar and Ruslan Yemtsov. The team members were Sara Johansson de Silva, Sasun Tsirunyan, Borko Handziski, Evgenij Najdov, and Janet Owens. Mr. Tsirunyan worked closely with the State Statistical Office on the preparation and statistical analysis o f the survey data and an exposition on methodological issues. Ms.Johansson made major analytical and written contributions to the report. The peer reviewers are Erika Jorgensen and Berk Ozler. Additional comments and guidance were provided by Asad Alam, Sandra Bloemenkamp, and Cheryl Gray. The team would also like to thank Helena Makarenko, kina Partola, Adam Pollack, Jasminka Sopova, and Judy Wiltshire for their administrative assistance to the team. vi KEYFINDINGS PovertySituationin 2003 ... Consumption poverty (measured by cost-of-basic-needs methodology) i s estimated at 21.7 percent. Non-monetary dimensions o f poverty (in particular, poor housing conditions and low education) affect another 30 percent o f the population. Poverty rates in Skopje, secondary urban centers, and rural areas are similar ranging between 20 percent and 22 percent. Determinantsof Poverty . The strongest determinants of poverty are household size and number o f employed . persons in household. Thus, persons with similar characteristics are more likely to be poor ifthe householddependency ratio i s high. L o w educational achievement o f head o f household and living insecondary urban centers are other important determinants o f poverty once we control for other household characteristics. EconomicGrowth andPovertyReductionin2002-2003 ... Poverty and inequality remained stagnant in spite o f a GDP per capita increase o f 3 percent. Poverty rates increased significantly in secondary urban centers and to a lesser extent in Skopje but fell inrural areas. Reasons for the deterioration o f living standards could be the contraction in employment and/or o f low sectoral growth inhighproductivity sectors. Multi-dimensionalAspects of Poverty . Health indicators and school enrollment rates appear to be in line with national income .. per capita. However, compared to the new Central and East European members o f the European Union (EU-8), they are low. The population's access to infrastructure services-water, sewerage, and electricity-is good with the exception o frural access to sewerage services. Reliability and accessibility o f infrastructure services i s below the average rates for EU-8 but are good when compared globally. Effectivenessof the Social ProtectionSystem: WelfareProgramsandPensions .. Pensions are important not only because they mitigate old age poverty, but also support other family members. Though social welfare programs disproportionately cover the poor in comparison to the nonpoor, substantial leakage o f benefits (75 percent) occurs to the nonpoor indicating a waste o f resources. Total current transfers-the majority o f which are social transfers-are an estimated 19 percent o f GDP. This i s highby international transfers and limits resources available for other investments as well as the opportunity for rollingback taxes. vii EXECUTIVE SUMMARY 1. This report is the product of cooperation between the State Statistical Office of Macedonia and the World Bank. The document itself i s the result o f the first stage o f dialogue between the two institutions on how best to understand the poverty situation in the country, develop appropriate indicators for measuring poverty, and to monitor changes in poverty over time. The process o f crafting this report provided a valuable opportunity and vehicle for estimating absolute poverty and more generally assessing living conditions o f the population o f FYR Macedonia. But as importantly, we hope that the process continues through informing the Government as well as the public o f the usefulness o f analyzing individual and household level data inorder to understandthe impact o f economic developments on the daily lives o f citizens. 2. The report provides a broad assessment o f the status o f living standards in FYR Macedonia during the period 2002 - 2003. This work complements the body o f analysis undertaken by the Government, independent researchers, and donors on economic developments in FYR Macedonia. Instead o f analyzing the successes or failures of policies in engendering growth, the report focuses its attention on examining living conditions as they stand today. Since economic growth i s not an end in itself but rather a means to achieving sustained improvements inlivingconditions ofall citizens, how goodor badoutcomes are ultimatelyreflectthe success or failure o f government policies. Inorder to reach its ultimate objective, the report specifically sets out to do the following: (i)provideabroadperspectiveonhowlivingconditionsinFYRMacedoniacomparesto other countries inthe region; (ii) determine the extent o f poverty among FYR Macedonia's population and outline the profile o fthe poor; and (iii) thelinkagebetweengrowthandchangesinpovertyovertime. analyze 3. Over the last decade, FYR Macedonia has made impressive progress in some key human development areas. Life expectancy increased and secondary school enrollment rates rose sharply. Furthermore, FYR Macedonia has relatively good infrastructure which i s an important national asset-it improves household welfare and facilitates investment. This progress occurred against a backdrop o f economic-and sometimes social-upheaval as the country moved towards a market economy. H o w FYR Macedonia was able to achieve this in a relatively short period o f time remains an unanswered question. It may reflect some combination o f good sectoral policies, targeted public interventions, and a shift inthe population's expectations and preferences on what are acceptable levels o f health care and education. However, greater research into how these successes were achieved remains to be carried out. 4. The analysis o f the available data from multiple sources yields a somewhat unexpected picture o f living conditions in FYRMacedonia. On the one hand, many o f the FYR Macedonia's social indictors are in line with comparable income-level countries-this i s true whether we look at health, school enrollments, or poverty across the Europe and Central Asia (ECA) region. On the other hand, in 2003 consumption poverty was estimated at 22 percent and a broader multi- dimensional definition o f poverty, which includes in addition housing conditions and low educational achievement indicate total poverty at 5 1 percent o f the population. Inequality also i s viii highfor the region, second only to Georgia. Furthermore, at particular riskofpoverty are children and youth who live inhouseholds with few working adults. Figure ES 1: Income and Social Indicators Life Expectancy Income and Secondary School Enrollment .E120 100 80 I 0 20,000 40,000 60,000 80,000 0 20,000 40,000 60,000 GDP per capita in US$FW GCP per capita in US$ (pR3 o EC4 M D + N8-R1wer(EC4) o ECA M D + W-8-Fbwer(ECA) Note: a/ ECA is "Europe and Central Asia region countries", MKD is FYRMacedonia, and EU-8 countries are the new EUmember countries o f Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia, and Slovenia. bi Data for gross secondary enrollment rates. Source: World Bank staff estimates based WDI data, Life expectancy data for 2003 and enrollment rate data for 2001. 5. In dynamic terms, during 2002-2003 poverty and inequality remained stagnant despite moderate GDP growth o f 3 percent. FYR Macedonia's economy suffers from a history o f low economic growth combined with labor market outcomes unfavorable to improving poverty outcomes-that is, job loss, high unemployment, and wage stagnation. Moreover, the Government's attempts at redistributing income to reduce poverty through its social assistance programs are weakened due to significant leakages to the nonpoor and low transfer amounts compared to the needs of the poor. But using the safety net to eliminate poverty would be costly-estimated at an additional expenditure o f 2-3 percent o f GDP annually. 6. The implications o f this complex situation for public policy are such that those involved in deciding or influencing government actions need to take pause-in order to determine the fundamental issue o f whether the current mix o f policies i s sustainable into the medium-term or will they result in an erosion o f progress on many fronts potentially jeopardizing the prize o f joining the European Union in the not-too-distant future. Though this report does not analyze either economic or social policies for efficiency or effectiveness, what it does show i s that continued progress on social indicators and a grappling with the poverty situation are critical for the future o f FYR Macedonia. It also finds that a significant portion o f public expenditures destined to "protect" those in need leak to those who are not among the most in need-leaving even less fiscal space for undertakingnecessary investments that will raise labor productivity and hence real wages inthe future. Social andpoverty indicators are broadly in line with comparator countries... 7. There are differentyardsticks against which to measure living conditions in FYR Macedonia. The internationally recognized set o f targets that are most comprehensive incapturing changes inliving ix conditions are the MillenniumDevelopment Goals (MDGs)since they can be used to compare progress both over time as well as against a comparator group of countries. Another measure o f how well FYR Macedonia is performing is to assess whether social and physical indicators are inline with its level of income as compared to other countries inthe ECAregion, especially the new Central and East European members o fthe EuropeanUnion(EU-8). Figure ES 2: International Comparisons of Poverty inFYRMacedonia, 2003 - Incomeand Poverty Poverty in Selected Countries 0 140 Bb I I I 120 .=0 100 2 80 2C 60 - 40 B 20 d .- I 5,000 10,000 15,000 GDPper capita in 2000 US$ F W o ECA MKD-bwer(ECA) Note: a/ ECA is "Europe and Central Asia region countries" and MKD is FYR Macedonia. bi Poverty i s defined as percent o fpopulationlivingbelow PPP-adjusted $4.30 per day per capita. Source: World Bank staff estimates basedWDI data. Life expectancy data for 2003 and enrollment rate data for 2001. 8. Since the early 1990s, FYR Macedonia has made important progress on several of the MDGs.The share o f population below the minimumlevel o f dietary energy consumption has been reduced, primary completion rates are now near universal, infant and child mortality rates have fallen significantly, as has the incidence o f tuberculosis. Relative to middle income countries and the E C A region as a whole, FYR Macedonia i s also doing relatively well in many aspects o f health and education. The MDGs were designed primarily with low income countries in mind and may not provide a good benchmark for a country at Macedonia's level of development for monitor welfare progress. 9. FYR Macedonia's social indicators are in line with its level of income. Health indicators such as life expectancy and mortality rates are relatively good and are comparable to those o f the EU-8 as seen in Figure 1. School enrollment rates are low compared to the relatively better off EU-8. Though currently broadly in line with its level o f income, analysis o f past enrollment rates show a marked increase over the past decade. Though this i s certainly encouraging, it also points to the fact that the existing human capital stock-especially among the group o f prime age workers-is not high on average which may be one factor in explaining the relatively low level o f worker productivity. 10. The vast majority of the population in FYR Macedonia has adequate access to essential physical infrastructure. This i s true o f access to water, sanitation, and electricity. As expected, a significant share o f the population in rural areas lacks access to sewerage facilities within their dwelling but this is a common phenomenon in developed and developing countries due to the highcost of providing this service ina cost effective manner to a dispersed population. X Analysis o f infrastructure services as reported by businesses gives some insight into quality issues. It appears that there are problems-though relatively modest-with reliability o f services such as water and electricity. Though we cannot extract much information on the quality of the housing stock from the survey, it does appear that a small though significant share lacks some basic amenities such as bathrooms and kitchens-7 percent o f urban households and 24 percent o f rural households lack a kitchen and/or bathroom. 11. Povertyas measuredusingthe internationalpoverty lineof PPP-corrected$4.30 per day is relativelylow comparedto its level of income and to some of the EU-8 countries. In FYR Macedonia, poverty is estimated at 22 percent with this method for 2003. This compares well against some o f the EU-8 countries and certainly against the average o f 55 percent o f the low and middle income countries o f the ECA region as seen in Figure 2. Poverty rates using PPP- corrected $1.08 and $2.15 per capita per day are negligible in FYR Macedonia. Though the poverty gap which measures the depth o f poverty i s not high by E C A standards, the average consumption o f the poor i s about two thirds o f the poverty line indicating that a large-rather than marginal-increase inincomes wouldbenecessary to reducepovertysignificantly. FyRMacetlonia: PovertyaridIriequalii Muni-DiriierisioiialPoverty FYRMaccedom 21 4 SkoPle 199 20 4 Other Uhan 17 5 21 8 Rural 25 3 22 3 hnetluality GmCoeffment 0368 0373 Source: World Bank staffestimatesbasedon HBS 2003. Absolutepoverty afflicts one infive of FYR Macedonia's citizens.. . 12. In order to measure poverty, this report uses an absolute or "cost of basic needs" poverty line. The poverty line estimates the minimum food and non-food basket o f goods and services needed to meet essential needs. It does not guarantee by any means a comfortable existence or perhaps even one that has all the necessities o f life such as sufficient consumption o f utilities or social services. But it does guarantee that enough consumption i s occurring so that families are able to fulfill their food requirements without needing to divert expenditures away to meet what society considers non-food essentials. It should be noted that this poverty line complements the relative and the subjective poverty lines used by the Government in measuring the share o fhouseholds living inpoverty. xi 13. In 2003, an estimated22 percent of the population of FYR Macedonia or about 445 thousandpersonslivedinpoverty.O fthese persons, 113 thousand (6 percent o fthe population) had consumption expenditures below the amount needed to purchase the minimum food basket indicating that they suffered from extreme deprivation and most likely were malnourished. The concentration o f poverty in Skopje, secondary urban centers, and rural areas were remarkably similar ranging between 20 - 22 percent o f their population. Furthermore, poverty was not particularly shallow as the poverty gap in 2003 was 7 percent indicating that the average consumption of the poor would needto rise significantly to reach the poverty line. 14. A multi-dimensional analysis of poverty which includes education and housing poverty in conjunction with consumption poverty results in increasing the poverty rate to 51 percent. Education poverty is defined as that share of the population with less than a secondary degree while housing poverty refers to either crowded living conditions or urban households without sewerage facilities within their dwelling. Education poverty affects 36 percent o f the population indicating that Macedonia's human capital stock i s not large (see Figure 2). On the other hand, housingpoverty i s relatively modest and affects only about 7 percent o f the population. The analysis shows that simultaneous deprivation in three dimensions of poverty i s uncommon indicating that extreme deprivation i s not widespread. FigureES4: FallingLivingStandards andRisingPoverty Consumptiongrowthper quintile, Increase in povertyby region, 2002-2003 2002-2003 -6 .- populationquintile -5 Skopje Otherurban 0Consumptiongrowthperquintile -15 ' I -Average consunption growth, kcedonia Source: World Bank staff computations based on H B S 2003. 15. The key drivers of (consumption) poverty in FYR Macedonia are high dependency ratios-large householdsize combined with low employment among adult members of the household. Persons at the bottom of the distribution are most likely to come from households that have many members-especially children-and adults who are either economically inactive (quite possibly due to withdrawal from the labor market) or unemployed. Other characteristics which increase the probability o f being poor are low educational achievement, and living in a secondary urban center. Though these characteristics are not unexpected, they appear to indicate that poverty i s largely a result o f economic and social policies-inadequate opportunities to find employment (possibly job losses from enterprise shutdowns) and an inadequate emphasis on ensuring completion o f secondary education. 16. The largest demographic group of poor is children, especially very young children. An estimated 30 percent o f the children in the age group o f 0-6 years live in poverty and a slightly less portion o f 6 - 14year olds. This compares to 22 percent for younger adults (aged 15- xii 39 years) and 17 percent for older working age adults (40-64 years). On the positive side, the prevalence o f poverty among the elderly (above 65 years) i s comparatively low. The probability o fpoverty appears to fall steadily over the life cycle. Poverty and inequality remained stagnant in spite of moderate economicgrowth. .. 17. Poverty remained stagnant at around 22 percent during 2002 - 2003 though economic growth during this period was 3 percent. Not only did poverty not fall but average consumption o f the population appears to have fallen as well by about 1.3 percent. Thus, economic growth did not translate into higher consumption for the majority o f the population- whether looking at those at the bottom or top o f the income distribution. 18. The national poverty rate hides large changes at the sector level-rural poverty fell significantly while urban poverty-especially in secondary urban centers-increased substantially. These changes can perhaps be explained by disaggregating economic growth into its sectoral components. Rural poverty's fall i s consistent with the 2 percent growth in the agriculture sector. On the other hand, though services and manufacturing value-added increased (6 and 3 percent respectively), non-manufacturing industrysaw a negative growth rate o f about 1 percent during2002-2003. 19. Employment contracted by 3 percent during 2002-2003 indicating jobless growth which is consistent with an increase in poverty. Between 2002 and 2003, employment fell in spite o f economic growth. Although the pattern i s not very clear, workers do not appear to have moved towards higher productivity-and, hence, high wage-sectors. Sectors with relatively high productivity levels, such as trade, transport and communications, and manufacturing, were shedding labor in this period. Employment contracted by 10 percent in agriculture but it appears that total employment inthe rural areas increased implyinggrowth inoff-farm activities. 20. Inequality also remained stagnant, which also stymied any improvement in the poverty situation between 2002 and 2003. Overall negative consumption growth appears to be one important factor explaining poverty developments, but in addition, the growth pattern did not favor the poor. Between 2002 and 2003, the Gini coefficient remained constant at 0.37, indicating no improvement in the income distribution in FYR Macedonia. As seen in Figure 3, the growth pattern was decidedly unfriendly to the poor. The poorer consumption expenditure quintiles saw the more drastic falls inexpenditures. 21. The Government allocates more than half of its total expenditures to social protection transfers. The Government spends a significant share o f its resources-54 o f total expenditures or 19 percent o f GDP-on current transfers (excluding payment on debt). The majority o f these transfers are for pensions with the remainder for various types o f social welfare transfers. The resources allocated towards social transfers i s a heavy burden on public finances and takes resources away from other important-and higher return-expenditures including investments ininfrastructure. It also limitsthe amount o f additional resources the government can allocate towards redistribution. 22. In the absence of social insurance and welfare benefits, poverty would increase substantially both at the nationallevel as well as among program beneficiaries. Moreover, 60 percent o f the population benefits from at least one social protection program either directly on indirectly (by residing in the household o f a benefit recipient). Social protection transfers were responsible for reducing the poverty headcount from 35 percent to 22 percent o f the populationin X l l l ... 2003 though this i s largely due to pension income (see in Figure 4). Pensions have the largest impact on poverty: without pensions, the national poverty rate would rise to 32 percent. For all other social welfare programs, the increase in poverty resulting from discontinuing them would have been about 2 percentage points, bringing the national poverty headcount to 24 percent. However, the depth o f poverty would increase significantly since though welfare transfers are not adequate for raising the poor out o f poverty they do significantly supplement the poor's income. FigureES5: PovertyRatesWith andWithout SocialInsuranceand Welfare Programs Social Protection Program Coverage Adequacy of Benefits by Quintile 50 I 3 45 Lc !?a, 40 n .-P 35 .? a, 30 -?a, 0 25 .-3 20 15 *- TJ 0 $ 10 ;C; 5 Pensions Social Child b l t h Other Pensions Social Child Health Other Asst. benefits Insur. Asst. benefits lnsur. N o t e : "Health i n s u r a n c e r e f e r s to "revenues from health insurance"; " s o c i a l asst." i s " s o c i a l assistance"; "other" is "invalid and other subventions". Source: World Bank staff estimates b a s e d on HBS2003. Growth will be key to lower poverty, but a better targeted social protection system is needed to maximize effectiveness.. . 23. In conclusion, FYR Macedonia faces the challenge of raising a large proportion of the population out of destitution-by pursuing an employment generating growth policy. The current economic structure does not favor an equitable distribution o f growth among the population. The lack o f goodjobs available to the poor i s the main explanation for the stagnation in poverty over time. This calls for bold actions to raise output and productivity growth to enhance job creation and sustainable real wage growth over time. For the Macedonian population-and the poor inparticular-in order to access well paidjobs inthe formal sector, the Government of FYR Macedonia will need to identify the key binding constraints to growth- possibly from the menu o f investment climate and reform o f labor and product markets-that inhibit employment generating economic growth. Inparticular, international experience suggests that reliance on massive social transfers and the concomitant high payroll taxes undermine job creation and growth. xiv To better understand the situation of thepoor, more work is needed. .. 24. This report has identified some key remaining knowledge gaps. A better understanding of the following areas will be needed to provide further guidance on identifying policy actions to combat poverty in FYRMacedonia. More details of the structure and functioning of labor markets, especially the informal labor markets, and the role o f labor markets inexplaining the regional convergence inpoverty across rural and urban areas. Inparticular, we know little o f the coping strategies of families insecondaryurban centers-where poverty increased sharply between 2002 and 2003. Determinants of rural income and the role of agriculture and non-farm activities in raising living standards. Rural poverty decreased significantly but whether this is due to increased agriculture income or better off-farm income generating opportunities i s not clear. A greater understanding o fbothincome and non-income dimensions o fpoverty inrural areas will help ininforming any policy dialogue on how to help the ruralpopulation. The impact of health care reforms on living standards, most specifically for the poor. While improvements inthe health care system are critical to ensuring the sustainability o f the system, there i s more work needed to assess the extent these reforms have trickled down to the poor, and what effects there have been from hardening budget constraints on access for certain groups. The role of ethnicity as a determinant of income. As in many other developed and developing countries, different ethnic groups have different average incomes when holding all other observable characteristics constant. Whether this i s due to discrimination or the types o fjobs preferred, requires greater analysis. Vulnerability. Poverty goes hand inhandwith vulnerability and exposure to risks which, for households living close to subsistence minimum, may critically threaten survival. What are the major risks facing households inrural and urban centers, what coping strategies are they presently using, and how could policy help households mitigate the negative impact on shocks both ex ante and expost? The existence of poverty traps. More analysis is needed to understand whether there is a vicious circle o f poverty reproducing over time in FYR Macedonia. There i s reason to believe, for example, that children belonging to poorer households have less access to education and health care services, and as such will be have lower chances o f escaping poverty inthe future. 25. Perhaps more important than filling knowledge gaps, i s the need for the Government to establish what information on households would better inform political debate and shape policies. Household surveys which are done annually provide an opportunity to track changes in welfare for otherwise the Government will only hear the protests or support for reforms from the most vocal groups in society rather than obtaining a broader and more representative view. To confine household surveys to simply the measurement of income or consumption poverty in the country misses an important opportunity for the Government to translate its reform program into specific goals to be achieved, develop indicators for tracking progress, and monitor changes over time. In order to carry this much more useful and pertinent agenda, a more responsive development data monitoring system would need to be established to meet the information needs o f the most current and important political debates inthe country. xv CHAPTER 1 OVERVIEW OFLIVINGCONDITIONSINFYRMACEDONIA INTRODUCTION 1.1. This report provides a broad assessment of the status o f living standards in FYR Macedonia duringthe period 2002 -2003. It complements the body o f analysis undertaken by the Government, independent researchers, and donors on economic developments in FYR Macedonia. Instead o f analyzing the successes or failures o f policies in engendering growth, the report focuses its attention on examining living conditions as they stand today. Since economic growth i s not an end in itself but rather a means to achieving sustained improvements in living conditions o f all citizens, how good or bad outcomes are ultimately reflect the success or failure o f government policies. 1.2. In1999, the WorldBankincooperationwiththe Government ofFYRMacedoniacarried out a similar report which was entitledFormer YugoslavRepublic of Macedonia: Focusing on the Poor. That report also assessed living conditions and provided a snapshot o f poverty in the country during the years 1990 and 1996. However, since then, much has changed. Not only have macroeconomic conditions improved but the country has pulled back from the brink o f a civil conflict and developed a plan for decentralization o f power. Also, FYR Macedonia has embarked on a path o f greater global integrationby submitted in2004 its application for membership to the European Union. Inaddition to social and political developments, significant improvements were made to the poverty monitoring system which allow for a more accurate picture o f existing Conditions. 1.3. In order to fulfill the broad objective o f providing a multi-dimensional assessment of livingconditions, the report aims to carry out the following: 0 provide a broader perspective on how living conditions inFYR Macedonia compares to other countries inthe region; 0 determine the extent o f poverty among FYR Macedonia's population and outline the profile o fthe poor; and analyze the linkage between growth and changes inpoverty over time. 1.4. The outline o f report i s as follows. We begin in this chapter with an international comparison o f key social and infrastructure indicators which provide an overview of how FYR Macedonia is performing relative to its comparator countries in the Europe and Central Asian region. The second chapter discusses the methodological issues pertaining to the estimation o f absolute poverty in FYR Macedonia and provides an estimate o f the proportion o f the population unable to meet a minimumstandard o f living. The third chapter discusses the main features o f the characteristics o f the poor. We close with an analysis o f the relationship between economic growth and poverty with a focus on developments in2002 and 2003. 1 Table 1.1: Macedonia: Progress on MillenniumDevelopment Goals Macedonia ECA I1 MIC 12 Earliestavailable Latestavailable Latestavailable Latest available Goalsrelatedto 1990-1994 2000-2003 2000-2003 2000-2003 Hunger Populationbelow minimum level of dietary energy consumption(%) 15 11 8 10 Schoolingand gender Primarycompletionrate (% ofrelevantage group) 97 99 90 95 Ratioof girls to boys inprimary and secondaryeducation(%) 98.5 98.1 97.3 97.7 Health Infant mortality rate (per 1,000 live births) 32 10 29 30 Under 5 mortality rate (per 1,000) 33 11 36 37 Births attendedby skilled healthstaff ("3oftotal) _ _ 98 -_ 87 Incidenceof tuberculosis (per 100,000 people) 54 31 82 114 Environment CO2 emissions (metric tons per capita) 5.3 5.5 6.7 3.2 Connectedness Fixed line and mobile phone subscribers (per 1,000 people) 148 448 438 403 Internetusers (per 1,000 people) 0.4 24.7 161 115.9 Other GNI per capita, Atlas method(currentUS$) 1,350 1,980 2,580 1,930 Life expectancy at birth, total (years) 71.6 73.6 68.5 69.7 Trade (% of GDP) 62 89 70 63 Source; World DevelopmentIndicatorsdatabase, April 2005. 1. ECA i s Europe and CentralAsia. MIC is the group of middle income countries. DEVELOPMENT INDICATORS 1.5. How does FYR Macedonia fare in terms of non-consumption welfare indicators? Material deprivation i s but one o f the many facets o f poverty. Poor health, inadequate education, and lack o f access to basic amenities and lack o f political and social rights are often strongly correlated with income poverty, but are also indicators o f welfare deprivation in their own right. The Household Budget Survey (HBS) does not constitute a solid basis for an analysis o f non- income dimensions o f poverty in a regional perspective. This chapter draws largely on official World Bank data which i s derived from a number o f sources including specific social and infrastructure surveys. 1.6. FYR Macedonia has made important progress on several of the millennium development goals. The share o f population below a minimum level o f dietary energy consumptionhas been reduced, primary completion rates are now near universal, infant and child mortality rates have fallen significantly, as has the incidence o f tuberculosis (see Table 1.1). The MDGs were designed primarily with low income countries inmindand may not provide a good benchmark for a country at Macedonia's level o f development for monitor welfare progress.' ' The millenniumdevelopment goals (MDGs) include eradicating extreme poverty and hunger, achieving universal primary education, promoting gender equality and empowering women, reducing child mortality 2 Figure 1.1:How do Health Indicators Compare with other ECA Countries? Life Expectancy Infant Mortality Rates 90 2 80 80 z 5 70 em - 75 .$ 60 50 ._ 70 140 30 65 ;2010 60 0 0 0 20,000 40,000 60,000 80,000 0 20,000 40,000 60,000 80,C 0 GDPper capita in US$ (FF) GDPper capita in US$ (FW) 0 ECA MKD + N8-Fuwer(ECA) Note: a/ ECA is "Europe and Central Asia region countries", MKD i s "FYR Macedonia", and EU-8 countries are the new EU member countries o f Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia, and Slovenia. b/ Data for gross secondary enrollment rates. Source: World Bank staff estimates based on WDI data, 2003. 1.7. The 1990switnessedlarge and significant improvements inFYRMacedonia's health and education indicators. The human condition improved across the board in those aspects o f greatest immediate relevance to the population-life expectancy rose and school enrollment rates increased. Improvements began in the early 1990s but continued throughout the 1990s even during difficult economic times. Average life expectancy rose by two years during 1990 - 2003 driven inpart by the large decline in infant and child mortality rates. Secondary enrollment rates increasedrapidly rising from 56 percent to 85 percent. SOCIAL INDICATORS AND ACCESS INFRASTRUCTURE TO 1.8. Health indicators for FYR Macedonia are relatively good in comparison to its level of income. Key indictors o f the population's health status are life expectancy and infant, child, and maternal mortality rates. More disaggregated data on health conditions in middle and upper income countries focus on adult morbidity such as coronary disease incidence. Inorder to provide a broad comparison o f health conditions in ECA, we see that life expectancy rates are above and infant mortality rates below comparator countries. Health conditions are a function o f many factors including prevalence o f potable water, education levels (especially o f women), road and work place safety, and an effective health care system. Macedonia's infrastructure i s well developed enough so that it i s not a source o f spreading infection and disease that lead to prevalent morbidity and mortality, especially among young children and the elderly. and improving maternal health, combating HIVIAIDS, malaria, and other diseases, ensuring environmental sustainability, and developing global partnerships for development. Each goal has a set o f target indicators. For the full set o f indicators, see htttx//www.developmentgoals.ora/. Figure 1.2: International Comparison of Secondary and Tertiary Education Gross Secondary SchoolEnrollment Gross Tertiary Enrollment 90 0 80 0 -5 170 150 0 70 0 - 60 130 50 40 .- 110 90 'E 30 20 70 10 0 20,000 40,000 60,000 0 20,000 40,000 60,000 GDPper capitain US$ (FW) GDPper capita in US$ (ppp) o ECA MKD + N8-Bwer(ECA) Note: ECA i s "Europe and Central Asia region countries", MKD i s "FYR Macedonia", and EU-8 countries are the new EU member countries o f Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia, and Slovenia. Source; World Bank staff estimates based on W D I data, 2003. 1.9. Reformsinthe healthcare system have beena major contributorto decreasingmortality ratesover the past severalyears. As seen fiom Table 1.1, infant andchild mortality rates have fallen significantly by almost two-thirds over the past decade. Progresshas been acheved by the Government through training o f health care professionals in the use o f evidence-basedprotocols and provision o f equipment which was implementedinthe late 1990s.' Reduction inmortality indicators has also led to increasing life expectancy indicators. Further improvement o f the health care system inunderway and the reform agenda is focused on increasing the financial sustainability o f the system over the long-term throughbetter use andallocationo f funds withinthe sector. 1.10. Macedonia's health indicators are comparable to those of the new Central and East European member states of the European Union (EU-8): A quick glance at the Figure 1.1 shows that Macedonia's indicators though usually not as good as most o f the EU-8 countries are inline with its lower GDPper capita-as inmost cases are those ofthe EU-8.The exceptions are male life expectancy and maternal mortality rates to a certain degree. Male life expectancy may be lower in EU-8 countries due to the higher prevalence o f alcoholism and tuberculosis. FYR Macedoniahas lower maternal mortality rates incomparison to Latvia and Estonia. 1.11. Macedonia's secondary school and tertiary education gross enrollment rates are on the low side thoughstill broadly inline with its incomelevel. Secondary andtertiaryenrollmentrates are 84 and 27 percent respectively as see inFigure 1.2. Though these enrollment rates are somewhat below average, they are not particularly poor for the country's level of development. However, given Macedonia's desire to become a member o fthe EU, enrollment rates are well below the secondary and tertiary averages o f the EU-8 (98 and 54 percent respectively). Furthermore, low secondary school enrollment rates are particularly worrisome because for a modem economy, especially one with the 'Macedonia: Health Sector Management Project, Project Appraisal Document. Report 27760-MK. EU-8 countries are Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia, and Slovenia. 4 intention o f integrating into Europe, the key to prosperity will lie inthe productivity o f the labor force which is inpartdeterminedby adequatetrainingandeducationlevels. Figure1.3: InternationalComparisonof UnemploymentandProductivity UnemploymentRates 35 140 30 - Worker Productivity 28 120 25 0 100 .d 2 5 20 .-cb 80 .- 15 0 60 I O 5k 40 c 5 B 20 0 0 20,000 40,000 60,000 0 20,000 40,000 60,000 80,000 GDPper capita in US$ (W) GDPper capita in US$ (FW) 0 ECA MKD + W8-Rwer(ECA) 0 ECA MKD + EM-mwer(ECA) Note: ECA i s "Europe and Central Asia region countries", MKD is "FYR Macedonia", and EU-8 countries are the new EU member countries o f Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia, and Slovenia. Source: World Bank staff estimates based on WDI data, 2001. 1.12. In comparison to the EU-8, FYR Macedonia has exceptionally high unemployment indicatingan economic system which under-utilizeshuman capital. Unemployment rates in 2003 are an estimated 31 percent in FYR Macedonia. Even when household survey data are used, unemployment rates continue to be high and well above the average for ECA or the EU-8. FYR Macedonia has historically had highunemployment rates indicating that structural rigidities continue to remain despite significant reforms. Moreover, a better measure to use may be employment rates simply because long-term employment may have resulted to significant numbers o f people withdrawing from the labor force altogether. The current set o f policies does not appear to be leading to an increaseinthe level o finvestment that would allow for greaterutilization o funemployedlabor. 1.13. Access of essentialinfrastructureservices appearsadequate.Very limitedinformation i s available on the population's access to infrastructure services such as electricity, sewerage, and water-much o f that information comes from the analysis o f the household budget survey. This information indicates that in general the vast majority o f the population has access to these three services. However, there i s one group which has low access to sewerage disposal within their dwellings and that i s the rural population (48 percent). However, this i s a common phenomenon in both developed and developing countries due to the prohibitively high cost of providing network sewerage services to a dispersed population. 1.14. Reliabilityof services appearsto berelativelygoodbutproblemsstill remain.Inaddition to availability o f services a related issue i s their reliability and here the only information available i s from the Investment Climate Survey. As can be seen in Figure 1.3, businesses inFYRMacedonia do report power, water, and telephone outages. Even though the reported averages for FYR Macedonia 5 are significantly higher compared to the EU-8, the number o f days i s well below the averages seen in the ECAregon-which p e r f o mrelatively well compared to the rest o fthe world. Figure 1.4: Accessibility and Reliability of Infrastructure Services 12 b wkIn210 8 Lc u 6 0 n 4 $ E .-2S 2 0 Delay in obtaining Electrical outages Water supply Delay in obtaining Telephone an electrical failures a telephone outages connection connection a Macedonia, FYR (2002) EU-8 Europe& CentralAsia Romania(2002) EJ Source:World Bank's Investment Climate Survey. CONCLUSIONS 1.15. FYR Macedonia has made major improvements across the board inMDGs as well as in more advanced social indicators in the 1990s. This progress was made during a period that saw economic turbulence as the country embarked on the process o f market reforms and a contraction in the Government's role in the economy. How FYR Macedonia was able to achieve this in a relatively short period o f time remains an unanswered question. It may reflect some combination of good sectoral policies, targeted public interventions, and a shift inthe population's preferences towards more education and better health care (resulting from a shift inpeople's preferences with the opening up o f the country to the world). Greater research into how these successes were achieved could provide valuable information to the Government for designing effective policies for the future. 1.16. Though progress was indeed significant and development indicators are broadly in line with its level o f income, FYR Macedonia still lags behind many European countries. Consequently, significant improvement still i s needed for it to catch up to the EU and even the EU-8. Of the three areas reviewed in this section-health, education, and infrastructure-the areas inwhich FYRMacedonia appears to need further concerted effort in order to bringit inline with other countries in the region are education-especially tertiary and secondary. However, whether education i s the binding constraint to economic growth or to welfare improvement i s not clear given the highlevels o f unemployment and would need further analysis. 6 CHAPTER 2 POVERTYTRENDS INMACEDONIA: METHODOLOGICALISSUESAND KEYRESULTS INTRODUCTION 2.1. The purpose of the Poverty Assessment is to provide a greater understanding of the poverty situation in FYR Macedonia and to identify measures to strengthen the poverty monitoring system. Widespread poverty is one of the key development challenges facing the country. Macedonia's main medium term objectives include establishing stable higher growth rates, sharing the benefits o f growth more equitably between groups and regions, and moving towards greater integration with Europe. Increasing the impact o f economic growth on living standards o f the population, especially those at the lower end o f the distribution, requires a clearer understanding o f the poverty situation. 2.2. Livingin poverty can be defined as livingin a state of "unacceptabledeprivationin well-being". Though few would perhaps disagree with this description from the World Development Report 2000, in practice, making poverty comparisons i s a more complex issue. What i s deemed an acceptable standard o f living will often differ depending on the socio- economic context. But there also exist some notions o f basic poverty that prevail in most societies: a person who i s undernourished, who has nowhere to live and who cannot get basic medical help if needed, would most likely be considered poor anywhere. There i s wide spread agreement that such basic levels o fpoverty still exist inFYRMacedonia today. 2.3. This chapter will summarize methodological issues in monitoring poverty in FYR Macedonia and summarize the key poverty trends and patterns in 2002 and 2003. The chapter addresses the following questions: H o w can we best define and measure "well-being"? At what threshold does "deprivation" become "unacceptable" and how do we ensure proper poverty comparisons between different people and over time?4 And using best practice methods, how has poverty developed across regions and over time inMacedonia? MEASURING WELL-BEING Conceptualapproaches to measuringwell-being 2.4. Poverty is a multidimensional concept and closely linked to the socio-economic environment of the individual. The main conceptual approaches to measuring welfare can be separated into welfarist and non-welfarist approaches, which differ quite substantially both in theory and practice (see Box 2.1). In practice, welfarist poverty measures usually focus on This chapter draws largely on the methodological overview in Ravallion (1992) and on the framework presented inthe World Development Report 200012001. 7 consumption expenditures or income levels. An advantage o f these indicators i s that when appropriately estimated, they can be used to compare poverty across time and space-which i s a key purpose o f poverty monitoring. On the other hand, they risk missing out many important dimensions o f poverty beyond consumption and income. Non-welfarist poverty measures include a variety o f non-income indicators, including education levels and health status. In spite o f the differences in approach, consumptiodincome and non-income poverty indicators tend to reflect similar levels o f attainment-people at the subsistence minimum generally have low levels o f education, higher mortality rates, and have little power over their own lives. I Box 2.1: What Do W e Mean by Poverty? The main conceptual approaches to measuring poverty can be separated into welfarist and non-welfarist arguments (Ravallion, 1992). In the welfarist approach, individuals are considered rational and make choices that maximize their utility and sense o f wellbeing or "welfare", given the constraints they face, for example limited income. From a pure welfarist perspective, a person who is happier i s considered better off irrespective o f whether he i s adequately nourished, has shelter, or adequate education. However, since measuring (and comparing) people's own sense o f wellbeing is difficult, a standardized indicator o f welfare i s used. Consequently, in practice, welfarist poverty comparisons are usually based on money-metric indicators o f income or consumption as proxies for utility. Poverty then becomes equivalent to "lack o f command over commodities". The non-welfarist paradigms focus o n "basic needs" and "capabilities" needed to reach a certain level o f fimctioning. The first approach-basic needs-focuses on means which are linked to reaching a particular quality o f life: minimumquantities o f basic amenities such as water, food, and shelter that are essential to avoid states o f illhealth and starvation. The second approach addresses the capabilities a person has to lead the life he or she chooses. Sufficient income, for example, means that a person has the capability to consume at least adequate levels o f food and other essential goods and services; or adequate levels of education to findproductive employment. Both of these approaches have some drawbacks. A criticism against the welfarist approach i s that in its extreme form it i s entirely subjective. The multidimensional nature o f basic needs and capabilities approaches to poverty also pose some problems interms o f poverty comparisons. One such issue i s how to aggregate, weight, and compare qualitative indicators o f poverty-is a person lacking access to sanitation poorer than an illiterate person, for example? Source: World Bank (2000),Ravallion (1992), Duclos, 2002. 2.5. This report primarily uses data on household consumption expenditures to estimate poverty levels in FYR Macedonia. Along the lines o f the World Development Report 2000/2001, it also recognizes that the concept o f poverty extends in many other dimensions o f well-being besides the amount o f commodities consumed at a point in time. The poor are often more exposed to economic downturns, natural disasters, and epidemics, and lack means o f protecting their families from such shocks. Generally, the poor also have fewer human capital assets: they have lower levels o f education attainment and are more prone illness. Public services, be it for infrastructure, education or health, are less accessible for low income households. Beingpoor i s associated with social stigma and lack o f voice and decision-making in society. And these poverty characteristics often combine to deprive the poor o f opportunities to improve their current and future living standards, or that o f their children. The difficulty in 8 capturing non-consumption aspects o f poverty lies in the unavailability o f data, especially for a nationally representative sample. 2.6. Consumption expenditures are a better indicator ofwelfare than income.There is an important choice to be made between using income or consumption as the welfare indicator. In principle, income could be a good indicator o f a person's capability to reach a certain level o f welfare. However, its major drawback i s that survey respondents tend to under-report their income for reasons pertaining to confidentiality. Furthermore, income may not be an appropriate measure for agricultural households, especially since consumption o f own-produced food would not be captured. Moreover, income often fluctuates from year to year, resulting in greater volatility o f income-based poverty rates. People's consumption o f food and non-food goods and services i s an alternative measurement o f welfare. It i s based upon household expenditures though i s not identical to it. Consumption is considered a better measure o f both current and long-term living standards than income. Consumption can better reflect an individual's ability to safeguard against temporary income losses. For example, members o f a household with savings can sustain consumption levels for some time even ifincome temporarily drops. 2.7. Consumption poverty measures are complemented by non-income poverty indicators. Consumption patterns can provide a comprehensive though not complete understanding o f the level o f deprivation a household i s faced with. Lack o f financial resources i s a key explanation for why some people cannot access reasonable health care or adequate shelter. Butmore money cannot necessarily buy better public services or betterjob opportunities-it may not even buy higher social status if it results only from public social welfare transfers. In the Poverty Assessment report, the analysis o f consumption data i s therefore complemented by an overview o f other non-monetary welfare indicators which trace social development, such as education, health, and labor market indicators. Yet there are many aspects o f poverty which will not be covered inthis report, including notions o f vulnerability to shocks, self-empowerment, and social capital. Constructingthe consumptionaggregate 2.8. What expenditures represent true consumption?Once consumption is chosen to be the most appropriate measurement of welfare, household expenditures must be translated into a consumption aggregate that i s a meaningful representation o f current welfare levels based upon available data. In order to be as robust a measurement o f present welfare as possible, it should reflect consumption in the period under consideration rather than future consumption- consequently, total expenditures should not be used as a welfare aggregate. The consumption aggregate should ideally include the following categories: food expenditures, non-food non- durable commodity expenditures, service expenditures, and use value o f durable goods. Food expenditures include both actual expenditures as well as the valuation o f self-produced food that i s consumed. Non-food consumption includes clothing, utilities, services, personal care and hygiene items, communication and transportation, and other non-food expenditures. Durables are items such as cars, washing machines, and housing and usually represent large expenditures. Though the purchase cost o f these items would not be included in the consumption aggregate since these items provide welfare well into the future, their annual "use value" should be included. 2.9. In this report, the Consumption aggregate includes food and non-food non-durable goods and services but excludes the use value of all durables. Construction o f a consumption aggregate i s limited by the availability o f data collected. Detailed information i s collected on expenditures of food and non-food goods and services including on assets such as cars and other 9 household durable goods. However, in order to compute the (annual) use value o f durable household goods, information i s needed on the age o f the durable goods-which i s not collected in FYR Macedonia. Another important item not included is the imputed value of housing. The reason for this is that most families in FYR Macedonia own their housing and consequently too few observations are available on rent to estimate the use value o f housing. 2.10. The consumption aggregate also takes into account spatial price differences. Another important, though often overlooked, issue i s that prices may differ among geographical areas o f the country, perhaps significantly inpoorer and larger countries where markets are less integrated across regions. In this report, different food price indices are therefore used for urban and rural regions. Price data collected from the household survey indicates that food prices are marginally higher in urban areas than in rural areas. Non-food prices are not corrected for spatial price differences due to the lack o f adequate comparable price information on goods and services. WHO I S POOR AND WHO I S NOT CHOOSING A POVERTY THRESHOLD - 2.11. The second important issue in analyzing poverty patterns is where to establish the poverty line, i.e., the level o f consumption below which a person i s considered poor. As noted, the notion o f where deprivation becomes unacceptable will tend to differ across societies, although some fundamental concepts o f poverty-inadequate nourishment and shelter-are common to most. It should be noted that a minimally acceptable level o f consumption does not mean the ability to achieve a comfortable standard o f living or to have adequate resources to have a buffer against exogenous shocks (such as job loss or major illnesses). However, it does mean that the person i s living in conditions where he i s consuming sufficient food (as determined by caloric value) and consumes some bundle o fnon-food goods and services. 2.12. Reliable and consistent poverty data are essential to monitor poverty over time and understand who is poor and who is not. The choice o f method for calculatingpoverty lines will ultimately depend on the characteristics o f the country as well as how this information will be used. From the perspective o f poverty monitoring, a poverty line should be consistent-two persons at the same standard o f living should be considered poor or not poor, irrespective o f the time or place being considered. Against this background, three alternative approaches to constructing poverty lines are discussed below: relative poverty lines and subjective poverty lines, both o f which are currently used by the Government o f Macedonia, and cost-of-bnsic-needs poverty lines, which are used inthe remainder o f this report. Relative poverty lines 2.13. Relative poverty lines imply that poverty is a question of inequality only. Under this approach, whether a person i s judged poor depends not on whether he has a certain level of consumption or income but entirely on how little he has compared to others. Relative poverty lines are frequently applied in high income countries; they are standard in measuring social welfare in the European Union (EU). A common approach i s to use some proportion o f the median o f the distribution o f income as the poverty line. In EU countries the population with income levels below 60 percent o f the median income i s used as a poverty benchmark (Eurostat, 2004).5 The EUuses income and not expenditures as the basis for calculating the relative poverty lines. 10 2.14. The official poverty line in FYR Macedonia is a relative poverty line set at 70 percent of median household expenditures. Inline with Macedonia's desire to align closer with EU practices and recommendations, beginning in 1996, FYR Macedonia,adopted a relative poverty line. It was established at 70 percent o f median household expenditures (rather than 60 percent o f income-as inthe EU) ofthe current year per equivalent adult, usingthe OECD scale to adjust for household size and composition (discussed in more detailed below).6 In both nominal and real terms, the denar value o f relative poverty lines has increased over time as seen inFigure 2.1. However, it should be noted that the dramatic increase between 2001 and 2002 is most likely due to changes in the survey instrument and sampling methods. Ggure 2.1: Problemswith Relative Poverty Table 2.1:Problems with Relative Poverty Lines Lines I FYR Macedonia: RelativePoverty Lines, 1997.2003 FI'R RIacedoi&i: CDP and Headcornit In(liccs Clmnges 70,000 (1998 - 2003) 60,000 FI. ni . GDP growth Change inheadcount index/d 50,000 (m percent) (m percent) 40,000 1998 3.4 8 9 -.- L 1999 4.3 1.4 30,000 2000 4.6 6.2 20,000 2001 -4.5 1.8 10,000 2002 0.8 33.0 2003 3.1 0.0 I 1997 1998 1999 2000 2001 2002 2003 61 Constant 1997denars Mninaldenars a/ Based upon official relative poverty lines Note: Poverty lines are for household and not Source: World Bank staff calculations based o n individual level. SSO and WDI data. Source: Macedonia SSO, 2004. 2.15. Relative poverty lines are a good indicator o f how highdisparities are at a particular point intime by the standards o f any particular country. But relative poverty lines also have several disadvantages, especially inthe context of a developing country such as FYRMacedonia. 2.16. Relative poverty lines disregard unfulfilled basic needs. First, the choice o f relative poverty lines i s necessarily arbitrary-what determines the choice o f 70 percent o f the median, rather than 60 or 50 percent? Second, in countries where absolute deprivation exist (which includes middle income countries such as Macedonia), relative poverty lines do not allow for identifying a level o f consumption which would be considered unacceptable at any time, and irrespective o f the income di~tribution.~ As will be discussed below, the new estimates produced by the State Statistical Office together with the World Bank suggest that a considerable share o f 6 The relative poverty line using OECD scale and this report's welfare aggregate i s MKD 58,655 which i s comparable to the SSO's poverty line o f MKD 63,197 for 2003. At the extreme, in a country where virtually nobody can satisfy their basic needs, relative poverty-lines may indicate little or no poverty if income i s evenly distributed, but people are, by any standards, poor. 11 the Macedonian population is in fact living below such a subsistence minimumlevel. As noted, the EU uses relative poverty lines for analyzing poverty in its member countries. Generally, however, the level o f absolute deprivation within high income EU counties i s very low, which makes comparisons o f unfulfilled basic needs less useful for policy purposes. And even within the EUit i s recognized that relative poverty lines do not imply that people below them are in fact poor, but only that they are at risk ofpoverty.' 2.17. They also result in inconsistent poverty comparisons over time. A related problem i s that relative poverty lines cannot be used for monitoring changes in poverty as they do not allow for consistent poverty comparisons over time. Furthermore it i s not clear whether economic growth and relative poverty rates would move in the same or opposite direction systematically since recessions could potentially be accompanied by a compression o f the income distribution. In2002-2003 for Macedonia, the median consumption per capita fell by 2.1 percent whereas the share of the population living inrelative poverty declined by over 5 percent. 2.18. The role of relative poverty rates in informing policy is unclear. Anti-poverty interventions tend to be targeted to households living in deprivation and relative poverty lines do not-and are not supposed-to inform policy makers whether more or less people are inneed o f assistance to fulfill fundamental needs. As noted earlier, since economic recession could lead to a reduction in relative poverty rates due to income compression, relative poverty could decline suggesting that safety nets are less needed in bad times than in good, although poorer people's consumption levels have fallen. A look at the official Macedonian poverty lines relative to economic cycles i s revealing as seen in Table 2.1. If anything, there i s a positive correlation between GDP growth and the annual change in the official poverty headcount index (at least in 2000,200 1and 2002) suggesting that higher growth rates coincide with higher poverty. 2.19. Despite the drawbacks of relative poverty lines, consistency with EUnorms remains a valid reason to continue to use them in conjunction with other measurements of poverty. Relative poverty lines are an important welfare indicator for the EU. Given Macedonia's strong interest in joining the EU, computation and official publication o f relative poverty should continue. However, for the reasons just discussed, it i s imperative to track absolute poverty in order to be able to develop policies for reducing poverty as well as to determine whether economic growth i s pro-poor. Subjective poverty lines 2.20. Subjective poverty lines are based on people's notions of what is a reasonable income level to escape poverty in their particular context. They are estimated entirely on the basis o f people's subjective perceptions o f what i s a minimum income to reach an acceptable level o f well-being intheir particular context. Subjective poverty lines tend to be based on survey responses to questions of the type: "what income level do you consider to be absolutely minimal to make ends meet?" They tend to be more popular indeveloped countries, rather than developing countries. Indeveloping countries, the concept o f income may be much vaguer, especially where subsistence agricultural production and informal employment could make total income more difficult to gauge. "Having an income below this threshold [60 percent o f the median income] i s neither a necessary nor a sufficient condition for being in a state o f poverty: indicators are consequently referred to as measures o f poverty risk." (Eurostat, 2004.) 12 2.21. The subjective poverty line for FYR Macedonia is about 70 percent of GDP per capita. The State Statistical Office o f FYR Macedonia collects information on subjective perceptions o f minimumincome levels which for 2003 are presented in Figure 2.2. When asked what the monthly minimumincome required for a household to make ends meet is, the average was MKD 7,557 per capita. This translates into a subjective poverty line o f MKD 85,579 (US$ 1,575) per person per year compared to a GDP per capita o f US$ 2,277 and private consumption per capita o f $1,688. Even when we look only at the bottom half o f households, the monthly minimumincome requiredis stated as MKD5,843 per capita. 2.22. Subjective poverty lines can be an interesting complement to more objective poverty lines, but cannot be used for poverty monitoring or for informing policy choice. While telling an interesting story in terms o f each environment's notion o f poverty, and how that changes, subjective poverty lines are not suited for targeting purposes, not least because the perceptions o f what i s a minimum income needed tend to increase with a household's income level. Alternative methods have been proposed for developing countries focusing on perceived consumption adequacy as a means o f estimating the subjective poverty line. In Nepal and Jamaica, this method rendered subjective poverty lines which were closer to other poverty lines based on more objective deprivationcriteria (Pradhan and Ravallion, 1998). Figure 2.2: Subjective Poverty Increases with Welfare Subjective Poverty Lines by Decile ~ .-C dn 8,000 I' 8 B 6,000 C .Q c 4,000 Q 5E 2,000 s 1 2 3 4 5 6 7 8 9 Population deciles --eMinirmrnmnthlyincormneeded &Real mnthly consumption per capita Source: World Bank staff computationsbasedon HBS2003. Cost-of-basic-needs 2.23. Most of this report focuses on a narrow definition o f poverty: a household lacking the ability to consume a bundle of food and nonfood goods and services that are deemed sufficient to satisfy the basic needs o f its members. The methodology used for estimating poverty in the FYR Macedonia PovertyAssessment 2005 is referred to as the cost-of-basic-needs approach. This poverty line is based on the minimum expenditures needed for basic food and non-food needs, on the assumption that individuals try to maximize their welfare given current prices and their budget constraints. Unlike relative or subjective poverty lines, it i s fixed in terms o f living 13 standards, meaning that two persons at the same standard o f living in different regions or in different years (adjusting for price differences) will always be considered poor or non-poor and treated the same. Once the consumption aggregate has been defined, the steps in calculating the minimumcost of basic needs-the poverty line-are as follows: 0 Based on the United States Department o f Amiculture (USDA) caloric values, the food poverty line i s calculated, ie., the yearly expenditure level required, given prices and consumption patterns, to be able to buy a minimumlevel o f calories (2100 calories per day per person) needed for an adult person to function adequately. Inorder to represent relevant basic needs consumption, the cost o f the food component o f the poverty line reflects food habits consistent with the consumption patterns in the low-income population. The basket i s adjusted for food prices by differentiating across urban and rural regions. a The share o f non-food items i s estimated following Engel's law. Basic non-food items are estimated based on the share o f expenditures on non-food items made by households whose food expenditures are equal to the food poverty line, i.e., those who are spending just enough on food to meet the caloric requirement per capita. Inthe case o fMacedonia, the non-food share i s around 44 percent for these households. a The cost-of-basic-needs poverty line i s the sum o f the food poverty line and the cost o f basic non-food items. This poverty line i s estimated on a per capita basis. The complete poverty line i s MKD 35,580 per year per capita (in2002 nominal denars). a In order to ensure intertemporal consistency, the poverty line for 2003 i s calculated by pricing the 2002 food basket with 2003 prices by applying the consumer price index.g As seen, the cost-of-basic-needs method results in poverty lines that appears significantly lower than the relative poverty lines calculated by the statistical office. However, this can be misleading as the cost o f basic needs poverty line i s based on an expenditure bundle which includes durables. Table 2.2: Cost-of-basic-needs Poverty Lines for Macedonia, 2002 and 2003 (inMYD) 2002 2003 Foodpoverty line " 20,009 20,249 Complete poverty line" 35,580 36,029 Memo: Relative poverty lines '* 64,946 63,197 Note: 'Annual per capita expenditures in denars. Based on official poverty lines from Statistical Office. These adult equivalent poverty lines are based on unweighted data, all consumption expenditures and do not include different price indices for rural and urban areas. Ifthe consumptionaggregate i s calculated as in this report and relative poverty lines at 70 percent o f the median, they amount to MKD42,997. Source: FYR Macedonia State Statistical Office 2004 and World Bank staff calculations. Note that if instead a new poverty line was calculated for 2003 using a new basket based o n 2003 values, the poverty comparison between 2002 and 2003 would not be valid. Since the consumptionbaskets are not identical we would-perhaps literally-be comparing apples in one basket with pears in another. There would be then no saying whether expenditures have increased or not. The same argument holds for comparisons across locations: although food and non-food consumption patterns may differ for example between rural and urban areas or between regions, the baskets must be the same for poverty comparisons betweenthese areas to be valid. 14 2.24. Cost-of-basic-needs poverty lines, like others, are based on the judgments o f those who construct them, for example, picking the reference group for the consumption pattern i s often a subjective choice. But their interpretation i s more transparent than that o f relative poverty lines. Absolute deprivation appears to exist in Macedonia, and these poverty lines reflect our best assessment of what that level is. A person whose food consumption does not reach the food poverty line i s not fulfilling his basic food needs, and most would agree that this i s indeed a key aspect o f poverty. Moreover, this method ensures consistency inpoverty comparisons over time. POVERTY MEASURES:MANY HOW POOR AND HOW POORARE THEY? 4.1 Poverty patterns and trends in2002 and 2003 2.25. This section looks at the situation o f poverty in FYR Macedonia in the years 2002 and 2003, Using the cost-of-basic-needs poverty line, three indicators are presented: the headcount ratio, the poverty gap index, and the poverty severity index (see Box 2.2). For comparative purposes, some of these indicators are also presented for relative poverty lines calculated for this report (i.e,, based on the same consumption aggregate as the cost-of-basic-needs poverty lines). 2.26. More than one in five persons lived in poverty in FYR Macedonia in 2003.'' An estimated 21.7 percent o f the population livedbelow the cost-of-basic-needs poverty line in2003. This means that about 445 thousand persons inFYRMacedonia were unable to meet their basic food and non-food needs. Of these persons, 113 thousand had consumption expenditures below the amount needed to purchase the minimum food basket indicating that they suffered fi-om extreme deprivation and most likely were malnourished. 2.27. Poverty remained stagnant between 2002 and 2003 in FYR Macedonia. The share o f the population living inpoverty increased slightly from 21.4 percent to 21.7 percent. Though the increase i s not statistically significant, it i s consistent with the decrease in average living standards seen inFYRMacedonia during that period. Box 2.2: Monitoring Poverty: The Foster, Greer and Thorbecke Measures of Poverty The headcount ratio (PO) measures the prevalence of poverty. It gives the percentage share o f poor people inthe population, i.e., the proportion o f the population whose consumption falls below the poverty line. This most basic o f poverty measures cannot tell us how poor the poor are. If between two years the poor become less poor, but still have consumption expenditures below the poverty line, the headcount ratio will not change. The poverty gap index (Pl) measures the depth of poverty. It measures the distance o f the actual expenditures o f the poor to the poverty line. Unlike the headcount ratio, the poverty gap index can tell us whether the poor are on average very poor, Le., very far from reaching the consumption expenditures at the poverty line, or whether there are a lot o f people clustered around the poverty line. If poverty is very shallow, a small increase in consumption would lift a large portion o f the population out o f poverty (and thus lower the poverty headcount index significantly). The poverty severity index (P2) measures the level of inequality among the group of poor by giving more weight to the depth o f deprivation o f the poorest people. Unlike the headcount ration and the poverty gap index, the poverty severity index will pick up changes in the distribution o f income within the g r o w o f Door. loInwhat follows, we refer to the complete poverty line only. 15 2.28. Between 2002 and 2003 there was a convergence in poverty headcount indices across Macedonia's urban and rural areas.'' The small changes in poverty headcount ratio at a national level mask important regional trends as seen in Table 2.3. In Skopje, the share o f its population living in poverty increased marginally, from 19.9 to 20.4 percent. Poverty rates increased quite significantly in other urban areas, from 17.5 percent to 21.8 percent, while falling inrural areas, from 25.3 to 22.3 percent. As aresult, povertyrates were, by2003, almost uniform across the country.'* 2.29. Poverty also became deeper in other urban areas, and shallower in rural areas, but the relative situation of the very poorest households worsened inboth Skopje and secondary urban centers. Other poverty measures confirm the worsening poverty situation inurban centers excluding Skopje, and the improvements seen in rural areas. In Skopje and rural areas, poverty became shallower, meaning that poor people saw their consumption expenditures rise closer to the poverty line. This led to a reduction in the poverty gap at a national level. In urban areas excluding Skopje, the poverty gap index increased by more than one fourth. The poverty severity index, measuring the level o f inequality among the poor, also improved in rural areas, and worsened in other urban areas. In Skopje, the poverty severity index actually worsened. This suggests that although many people saw their consumption expenditures increase to approach or cross the poverty line, those at the very bottom o f the income distribution were actually worse off. Table 2.3: Poverty Headcount Ratio, Poverty Gap Index and Poverty Severity Index (in percent) Poverty Headcount Poverty Severity Ratio Poverty Gap Index Index 2002 2003 2002 2003 2002 2003 Costof basic needs 'I: FYR Macedonia 21.4 21.7 6.3 6.7 2.7 2.8 Skopje 19.9 20.4 6.3 6.7 2.7 3.0 Other urban 17.5 21.8 5.0 6.9 2.1 2.9 Rural 25.3 22.3 7.4 6.5 3.2 2.6 Memo: Relative poverty lines" FYRMacedonia 'Per 31.5 29.7 10.4 9.8 4.7 4.4 capita. Per capita. These relative poverty lines are based o n the same consumption aggregate as the cost-of-basic-needs poverty lines. (see footnote 5). Source; World Bank staff calculations. Accountingfor householdsize and demographics 2.30. Households of different size and demographic composition may reach different welfare levels with the same level of expenditures. In the poverty measures above, total household consumption expenditures have been transformed into per capita consumption (by simply dividing by total number of household members). We assume, in other words, that a In 2002 21.7 percent of the population lived in Skopje, 35.2 percent in other urban areas, and 43.1 percent inrural areas. l2It i s worth noting that although poverty as calculated by the complete poverty line is lowest in Skopje in all o f Macedonia, poverty according to the food poverty line is actually higher inSkopje than inother urban areas. 16 household with eight people that i s consuming the same as a household with two people must be worse off. Though using per capita i s more transparent and requires fewer subjective decisions, the approach ignores several important factors. One issue i s the demographic composition o f the household: for example, children may need less food than a worlung adult. Economies o f scale pose a similar problem. Some goods can be shared across members-for example, housing or consumer durables, up to a certain point-which will lower the expenditure per person compared to a smaller household; yet, the level o f utility derived will be the same. Hence, with the per capita approach, households with many members and/or many children could come across as poorer than they actually are. 2.31. The issue can be addressed by adjusting the consumption aggregate by adult equivalent scales which rescale household consumption into the equivalence of adults, depending on size and composition. There i s potentially an indefinite number o f scales to use for such purposes. One clear disadvantage o f equivalence scales i s that the choice o f allocation rule necessarily will be arbitrary and there i s no agreement on which one to apply. The preferred practice i s therefore to use the per capita measures as a base-case and check the robustness o f poverty measures and poverty profiles vis-his the application o f equivalence scales. Two options are evaluated here: the OECD scale and a version used regularly inWorld Bank poverty assessments (see Box 2.3). Box 2.3: EquivalenceScales Used in the FYRMacedonia Poverty Assessment 2005 The scales referred to inthis report can be expressed ina general form as: AE = (1 S(A- 1) + +..cy where AE is the number o f adult equivalents inthe household, A i s the actual number o f adults, C i s the number o f children, CY measures the difference inconsumption between adults and children, and 6 or 0 adjusts for the level o f economies o f scale. These parameters are set between 0 and 1 (and either 6 or B will be 1). The weights attached to economies o f scale and demographics can thus be varied. Two versions are evaluated here: The OECD scale sets 6 to 0.7, CY to 0.5, and B to 1, resultingin: AE =1 0.7(A-1) 0.5C + + The general form o f the World Bank adjusted scale sets 6 to 1, and CY and B can be adjusted according to whatever i s considered reasonable for the particular context: AE = (A +..cy The closer to 0 the parameters, the stronger are the respective economies o f scale and demographics effects. Children are arguably less expensive in developing countries, even in relative terms, suggesting that CY should be small, reflecting such items as lower school fees and no entertainment costs. But with a high food share, children's consumption should not come with a large discount, as the caloric requirements o f all but very small children are almost as high as those o f adults. Moreover, when there are few shared goods inthe consumption aggregate (perhaps because food i s an important share or fewer durables), there i s less justification for malung large adjustments for economies o f scale. Source: Adapted from Deaton and Zaidi, 2002. 2.32. The OECD scale currently used by the Macedonian State Statistical Office to adjust the welfare aggregate does not provide a good representation of the impact of size and 17 demographics on household consumption. This scale is significantly different from the per capita approach as it implies substantial economies o f scale and much smaller consumption needs o f children. But it i s not clear that this scale adapts well to the Macedonian case. First, following the cost-of-basic-needs poverty lines, food accounts for the majority o f consumption expenditures (54 percent). Food cannot be shared, so economies o f scale do not apply, and children moreover need almost the same amount o f calories as adults. Further, in the consumption aggregate used here, durables-which typically are shared goods-have been excluded. In view o f these shortcomings, an alternative scale, drawing on Deaton and Zaidi, 2002, i s also applied where parameters have been chosen to reflect smaller adjustments than inthe case o f OECD. 2.33. Overall, the poverty profile is robust to different equivalence scales. Table 2.4 compares the poverty headcount indices for 2003 across per capita scales, the OECD scale, and the adjusted World Bank scale. As expected, the OECD scale increases poverty rates for people living in small households relative to larger ones, and for older people relative to younger ones. But the ranking remains robust to the scale. Large households are ingeneral poorer than smaller households, and it i s still the case that a higher share o f children than elderly live inpoverty. Table 2.4: Poverty HeadcountIndexfor Per CapitaConsumptionand Equivalence Scales I' Per capita OECD adjusted" Adjusted generalformula '' Food poverty 5.5 6.3 5.1 Complete poverty 21.7 24.8 21.7 By Region Skopje 20.4 23.0 19.9 Other urban 21.8 25.3 22.6 Rural 22.3 25.3 22.0 By Household size 1 5.1 13.5 8.9 2 5.7 11.9 8.3 3 10.5 15.3 12.3 4 15.3 19.0 15.8 5 22.3 24.4 22.3 6+ 34.6 36.2 32.9 BY Age Less than five years 29.5 29.6 28.9 6-14 28.2 29.4 26.1 15-24 24.1 28.1 24.6 25-39 23.1 25.3 22.6 40-64 18.1 21.7 18.7 65 and over 13.9 19.8 15.4 Note: 1. Adjusting for household size and composition increases the overall poverty, but not the ranking within the poverty profile. 2. Normalized equivalence scales. Source; World Bank staff computations based on H B S 2003. H o w does FYR Macedonia compare with other countries? 2.34. Poverty indices based on national poverty lines are not internationallycomparable, due to differences insurvey designs, inconsumptionpreferences, and the assumptions underlying the estimates of welfare indicators and poverty lines, as the notions o f what i s minimum subsistence and basic needs increases as a country becomes richer (Ravallion et al., 1991). To facilitate internationalcomparisons, the World Bank calculates poverty lines ininternational (PPP adjusted) dollars fixed at about 1, 2 and 4 U S dollars per day which should buy the same standard 18 o f living ineach ~ o u n t r y .These are presented for FYRMacedoniainTable 2.5, using2000 PPP ' ~ conversion factors. As can be seen, a negligible proportion o f the population i s livingon less than one or two dollars a day, and the poverty gap and poverty severity indices are also very low using these international poverty lines. Instead, Macedonia's actual poverty lines appear to be fairly close to PPP-adjusted 4.30 U S dollars per day, yielding very similar poverty numbers. Table 2.5: One and Two Dollar a Day Poverty is Low in Macedonia Poverty headcount, poverty gap, and poverty severity indices for FYR Macedonia using international poverty lines, 2003 '' Headcount Poverty Ratio Poverty Gap Severity International poverty lines $1.08 per day 0.13 0.04 0.03 $2.15per day 3.8 0.8 0.3 $4.30 per day 21.7 6.7 2.8 Cost-of-basic-needs 20.5 6.1 2.5 Using2000 PPP numbers. Source; World Bank staff calculations Figure 2.3: Compared to ECA Countries, FYRMacedonia has Low Levels of Poverty Income and Poverty Povertyin Selected Countries p 140 (u 120 d 80 .',._ 100 C ! C E 6o 40 5 20 aec ._ o 5,000 10,000 15,000 GDPper capita in 2000 US$ FW 1o ECA M(D-Fwer(EC4) Note: I aiECA is "Europe and Central Asia region countries" andMKDis FYRMacedonia. b' Poverty i s defined as percent o f population living below PPP-adjusted $4.30 per day per capita. Source: World Bank staff estimates basedWDI data. Life expectancy data for 2003 and enrollment rate data for 2001. l3 The $1.O per day poverty line-which actually represents purchasing power parity adjusted $1.08 dollar-was chosen because at the initiation o f the work o n international poverty comparisons, it was representative o f poverty lines found inthe poorest countries. For middle income countries, it is clearly too low to be meaningful for policy analysis. Instead, two alternative poverty lines, twice and four times as high (2.15 and 4.30 international dollars a day) are typically used. PPP conversions are not straight- forward, however, and caution is still warranted for international poverty comparisons. 19 2.35. FYR Macedonia has fairly low levels of poverty compared to other low and middle income countries in Eastern Europe and Central Asia. Usingthese internationally comparable numbers, Macedonia's poverty numbers are modest relative to its income at least within ECA. Two important comparators for FYR Macedonia are Romania and Bulgaria. Romania has higher GDP per capita (PPP) than FYRMacedonia but muchhigher povertyrates as well as can be seen inFigure 2.3. On the other hand, Bulgaria has a lower income per capita but also higher poverty than Macedonia, at almost the same income levels, have substantially higher proportions o f the population living on less than $4.3 per day. - MULTI-DIMENSIONAL POVERTY 2.36. Deprivation is not limited to consumption and can includeimportant aspects such as limited education and poor housing. Though consumption i s indeed the most immediate and possibly devastating loss households suffer when income i s too low, deprivation along other aspects can have equally negative affects for peoples' welfare. This section will provide some flavor o f multi-dimensional aspects poverty and how different aspects affect or do not affect the same group o f persons. W e examine two types o f poverty, housing and education, because o f their importance for people's ability to adequately meet their immediate and future needs. The results are presented inFigure 2.4. 2.37. Education poverty affects 36 percent of the populationindicating that Macedonia's human capital stock is not large. The definition used for education poverty i s persons above the age o f 19years with less than secondary school. This category also includes persons who may not have completed primary school or have no education at all, though these are uncommon phenomena in FYR Macedonia. The relatively low educational attainment o f a significant share of the adult population i s a result o f the low secondary school enrollment rates that continued until the early 1990s. For example, in 1990, gross secondary school enrollment rates were 56 percent and increased to 85 percent in2000. 2.38. Housing poverty affects 7 percent of the population. The definition used for housing poverty i s crowded conditions (less than 6 square meters per person) or no sewerage disposal in dwelling (inurban areas). The share o f the population experiencing this type o f poverty i s limited and inpart this may be a reflection o f the limited indicators used for assessing housing conditions or the alternative that the country has indeedput adequate resources inthis area. Figure 2.4: Multiple Dimensionsof Poverty inFYRMacedonia 2003 Education Consumption Housing Non Poor=49% I Source: World Bank staff computations basedon HBS 2003. 20 2.39. Simultaneous deprivation in three dimensions of poverty is uncommon indicating that extreme poverty is not widespread. As can be seen from Figure 2.4, the subset of persons who are consumption poor as well as housing and education poor i s 1percent. The largest overlap between different dimensions o f poverty lies between consumption and education poverty. An estimated 9 percent o f the population suffers from both forms o f poverty. The second largest overlap between the consumption and housingpoor i s at 3 percent o f the total population. 2.40. According to a broader definition of poverty, the share of the population living in poverty is 51 percent. This multi-dimensional poverty is driven primarily by education and consumption poverty. The greatest concerns that arise when viewing poverty through this lens i s that poverty becomes a larger problem than first envisioned. In addition, education poverty i s probably the most difficult to address because o f the limited opportunities as well as demand for education among adults. As we will see in Chapter 3, poverty rates are highest among persons living inhouseholds headed by someone with less than secondary education. SUMMARY 2.41. This chapter has presented methodological issues and key results on poverty measurements inFYRMacedonia. Its main conclusions are: 0 The methodology currently used by the Government of FYR Macedonia does not permit consistent and accurate poverty monitoring.First, the consumption aggregate must be adjusted to better measure household welfare. Second, cost of basic needs poverty lines are better suited than relative poverty lines to a country where absolute deprivation exists. The main differences between the methodology used for the FYR Macedonia PA 2005 and that o f the Government are outlined inTable 2.6. 0 M o r e than one in five persons cannot satisfy their basic food and non-food consumption needsin FYR Macedonia.Although rural poverty remains higher than in urban areas, there was an important regional convergence between 2002 and 2003. Poverty headcount indices and the poverty gap fell inrural areas compared to Skopje and other urban areas. 0 Poverty numbers tend to increase with household size and younger demographic profile,butpoverty trends andpatterns remainrobustto equivalence scales. Table 2.6: Government ofFYR Macedoniavs. FYRMacedoniaPoverty Assessment 2005: butdoes not include imputedvalue ofrent 21 CHAPTER3 A PROFILE OFTHE POORIN2003 INTRODUCTION 3.1. This chapter presents the key characteristics of the poor inFYRMacedonia in 2003. The household survey data collected by the State Statistical Office o f FYR Macedonia i s used to identify some important correlates o f poverty. We focus only on those characteristics for which data are available such as on age, gender, education, and employment. However, many important aspects o f poverty--especially those pertaining to health conditions-are not discussed due to the lack o f information. 3.2. The daily lives of the poor differ significantly from those of the non-poor. Consumption patterns differ and household composition are not the same. The poor live in worse conditions and less access to basic infrastructure services. They have less labor market access, and income opportunities o f the mainbreadwinner, given by age, gender, and education levels are a key predictor of poverty. 3.3. Poverty is a household characteristic, and individuals are considered poor because he belongs to a poor household. Measuring individual poverty dimensions using household poverty status and characteristics has some consequences for the poverty profile, which should be kept in mind when interpreting the results (see Box 3.1). Moreover, a profile based solely on simple correlates o f poverty can be misleading. An analysis o f the conditional correlates o f per capita consumption-the relationship between characteristics o f households and consumption, controlling for other variables-is provided inAnnex 2. HOUSEHOLD EXPENDITURES 3.4. Food consumption absorbs a very large share of poor households' expenditures leaving few resources for other expenses. Expenditure patterns differ between poor and non- poor, and, since poverty incidences differ, also between different regions as seen in Table 3.1. Food expenditures make up three quarters o f poor households' total expenditures, compared to just over half for the national average household. Together with alcohol-which accounts for a larger share o f poor people's expenditures than o f those o f non-poor-food expenditures amount to 80 percent o f all expenditures. These high food shares leave little room for other essential expenses. 3.5. Expenditures on other essential non-food items are very low compared to national averages. In particular, clothing, housing, and utilities account for a smaller share (and much lower expenditures in absolute terms) among the poor compared to more affluent households. Means o f transportation and communication are also underrepresented in the poor's expenditure patterns relative to others. Not only does the lower amount o f resources available for non-food consumption items indicate less money can be spent on improving living conditions, but also the difficulty o f poor households accessing markets and jobs for earning a living (e.g., transportation). 22 Box 3.1: IndividualPoverty and Household Characteristics Inthis report, income/consumption poverty is a household characteristic, and individuals are poor because they belong to a poor household. However, applying a household measure to each individual has some consequences for the poverty profile. First, studies have shown intra-household distribution to be far from equal and so some members may therefore be poorer than others within any single household. Infact, overall poverty rates may be substantially higher if intra-household inequality in distribution i s taken into account (Haddad and Kanbur, 1990). Second, given the construction o f household surveys, poverty profile analysis tends to be based on the characteristics o f the head o f the poor household rather than the whole population o f the household. Yet, the household head may not be the person who de facto determines the income opportunities o f the household. For example, an older and inactive person can be considered the household head inthe survey on account o f seniority, but i s perhaps not the mainbreadwinner, and his education and labor market characteristics have little impact on consumption levels. In households with multiple income earners, it i s also difficult to designate one person in the household to establish the profile o f the poor. Related to this issue i s the problem o f household size and demographics discussed inthe chapter on methodology. The per capita approach (dividing total household consumption by number o f members) can overestimate poverty in larger families, who could benefit from economies o f scale, relative to smaller ones. It could also overestimate poverty rates among children relative to older age groups. These caveats shouldbe kept inmindwhen interpreting the results. HOUSEHOLD LOCATION, ANDDEMOGRAPHICS SIZE, 3.6. Poverty i s higher in rural areas than in urban areas, but poverty is increasing in urban areas outside Skopje. Unfortunately, the HouseholdBudget Survey (HBS) does not allow for detailed geographic analysis. As shown in Chapter 2, poverty i s highest in rural areas, and lowest inthe capital city. However, the differences are not large and poverty levels appear to have converged between 2002 and 2003. As a result, the concentration o f poverty in other urban areas i s very similar to that in rural areas at 22 percent. Moreover, regressions on per capita consumption suggest that it i s not locationper se, but other typical characteristics o f poverty (e.g., lower education or size of household) that keep poverty levels higher in rural areas. When these characteristics are accounted for, living in other urban areas has in fact a stronger negative relationship with consumptionthan living inrural areas. 3.7. Poverty rates are generally higher among households that are large and have many dependent children. Households consisting of one or two people have considerably lower poverty rates than larger households. Beyond four person families (the "modal" household in Macedonia), poverty rates increase dramatically; indeed, one third (34 percent) o f all people living in households with six or more people are poor. Having a large number o f young children-who are too young to contribute to household income-has an even stronger impact on poverty. Almost half (47 percent) o f the population living in households with three or more children under age 14 are poor, compared to 13 percent for those without young children. There i s a very high penalty, in poverty terms, from having more than three children (as seen in Figure 3.1). 23 Table3.1: FoodConsumptionDominatesExpendituresof the Poor Household expenditures by main spending category, 2003 Poorest 20 National Rural Other percent" Average Skopje Urban Food & Beverages 73.7 53.7 57.0 49.2 53.2 Alcohol 6.4 5.3 5.6 5.3 5.2 Clothing 1.9 6.0 6.1 5.7 6.0 Housing & Utilities'* 3.1 10.2 7.9 12.8 10.8 HouseholdFurnishings 4.6 4.7 5.0 5.0 4.3 Health 3.6 4.1 4.0 4.0 4.2 Transport/communication 2.4 9.1 8.5 10.9 8.5 Other" 5.4 9.5 8.6 9.8 10.3 Source: World Bank staff estimates based on HBS 2003. Note: Firstexpenditure quintile. Excludes imputedand actual rent. Includes recreation, ' education, restaurants and hotels, and miscellaneous expenditures. Figure 3.1: Larger Households,Especiallywith ManyYoung Children,are Significantly Poorer than Others Poverty headcount indices by household size (left) and number of children aged less than 14(right) -s-2i50 _. I 45 40 35 30 8 25 U g 20 .c P$ l5 10 3+ 8 5 i 0 T 1 2 3 4 5 6 t None 1 2 No. household members No.children age 14 or less Source: World Bank staff estimates based on HBS L 3.8. It follows that relativelymore children experience poverty firsthand than older groups. This is especially the case for very young children, of whom 30 percent live inpoverty. This compares to 22-23 percent for younger adults (aged 15-39 years) and 17 percent for older working age adults (40-64 years). On the positive side, the prevalence o f poverty among the elderly, above 65 years), i s comparatively low. The probability o f poverty appears to be falling monotonically over the life cycle (as seen inFigure 3.2). 3.9. As discussed inChapter 2 and Box 3.1, the per capita approach does not take into account the age and gender composition or size o f households which can affect total household consumption levels. Adjusting for economies o f scale and different consumption needs o f children tends to compress the difference between poverty rates along the age and size scale, but poverty rates still fall with age and increase with household size. The ranking and conclusions above do not substantially change. 24 Figure 3.2: PovertyHeadcount Indices by Age Group, 2003 - 35 - I-I 3 0 - .-8 2 5 - I. s 2 0 - L 8 1 5 - c 0, 1 h 10- 5 - n O t T 15- 24 25 - 39 65 + Under 5 40 - 64 Age groups Source: World Bankstaffestimatesbasedon HBS2003. HOUSING CONDITIONSAND INFRASTRUCTURE 3.10. Poor and non-poor families own their houses but housing conditions differ. Home ownership i s high in Macedonia, around 90 percent, for the poor as well as the non-poor. This i s consistent with patterns observed in other transition countries. Indeed, housing often provides the only means o f asset accumulation available to the poor. Unsurprisingly,housing conditions differ between the poor and the non-poor. The poor live in more crowded dwellings (14 m2vs. 21 m2 per capita) and also have smaller holdings o f cultivated land on average (12 m2vs. 5 m2per capita). Though a majority o f the poor do have access to water, sewerage and electricity, they lag behind the non-poor, and there are relatively more poor households that do not even have a kitchen or a bathroom (Table 3.2). Few households in FYR Macedonia have access to central heating, and 78 percent o f the non-poor and 89 percent o f the poor heat their dwellings with stoves, using solid fiels. At the time o f the household survey, less than halfof the non-poor, and less than one quarter of the poor households, believed that they had enough resources to heat their houses adequately. 3.1 1. The poor consume lower-and probably inadequate-levels of infrastructure services. As seen inTable 3.1, the poor spend 3 percent o f their total consumption expenditures on utilities compared to the national average o f 10 percent. This large difference may be a reflection o f the poor's inability to pay for utilities, the poor's lack o f access to these services, or that the poor consume these services but do not pay for them. Inorder to determine what the real reason is, a targeted housing survey would need to be administered. However, irrespective o f why utilityexpenditures are low, what is clear is that the poverty line useddoes not assuremore than a minimumaccess to utilities-and consequently, the conditions o f the people who are just at the poverty line may not actually meet socially acceptable-let alone comfortable-living standards. 25 Table 3.2: Housing Conditions of the Poor and Non-poor Proportionof non-poor persons living in householdswith no Proportionofpoorpersonsliving access in householdswith no access (in percent) Water 5 11 Seweragedisposal 22 36 Electricity 0.5 1 Central Heating 88 96 Kitchen 5 12 Bathroom I 15 Source: World Bank staffestimatesbased on HBS 2003. HOUSEHOLD CHARACTERISTICS: GENDER, HEAD AGE,AND EDUCATION 3.12. There are no major gender differences in poverty levels, and poverty also decreases with age of household head. In 2003, 21 percent o f the population living in male-headed households was poor, compared to 17 percent for female headed households (Table 3.3). This should not be taken to imply that poverty i s lower among the female pop~lation.'~ And indeed, controlling for other household characteristics, women-headed households appear to have slightly lower per capita consumption than men, leaving the conclusion uncertain. People living in households with older household heads also tend to be less poor. Poverty rates are 25 percent for the population living in households where the head i s less than 40 years old, compared to 20 and 18 percent ifthe headi s 40-64 or older than 65, Table 3.3: Poverty Rates by Female and Male-headed Households Share of poorhouseholds Povertyheadcountindex Share of all households (inpercent) (inpercent) (inpercent) Male headed 88 90 22 Femaleheaded 12 10 18 Source: World Bank staff estimatesbased on HBS 2003 3.13. There is a strong relationship between education and poverty. As i s generally the case, the education level o f the household head i s negatively correlated with poverty incidence. As seen in Figure 3.3, there i s a very high premium to the head having at least secondary education relative to lower levels o f education. Living ina household where the headhas primary education i s also considerably better than if he/she has no education whatsoever, but the most significant difference appears to kick in at higher levels o f education-areas where FYR Macedonia has a deficit given its past under-investment in education. Multivariate regressions corroborate this result. l4Gender-poverty analysis based on gender of household head is very incomplete, as it is impossible to separate out individual consumption levels of men and women (and boys and girls) inhouseholds headed by either men or women. Indeed, only 12 percent of all households are headed by females. Yet, women, andperhapsespecially girls tend to be disadvantagedinintra-householddistributiono f goods or investment in human capital. (see e.g., Quisumbing and Maluccio, 1999 and Case and Deaton, 2003.) Non-income aspects ofpoverty therefore tend to be a better way o funderstandingthe links betweengender andpoverty. 26 Figure3.3: PovertyHeadcountIndicesby EducationLevel of Household Head,2003 I , 1 Age groups Source: World Bank staff estimates based on HBS 2003 INCOMEAND LABOR MARKETS 3.14. Poor households depend more on informal employment and public transfers than wealthier households. Sources o f income change with household average wealth. Thus, poorer households have relatively less income from non-agricultural employment or self-employment, and relatively more from social transfers. Dependence on income from informal employment i s also higher than for richer households. Importantly, the share o f public transfers i s higher for poorer groups, with a big difference between the poorest 20 percent and other households, while pension income account for more or less the same share along the expenditure continuum. In absolute terms, the poorest households receive about twice as many denars worth o f public transfers as the average Macedonian household, including mostly various social assistance schemes, but also child benefits, health insurance and support for the disabled, scholarships, and alimentary support and subsidies. The poor are less likely to benefit from the public pensions system than the non-poor, however. In denars, pensions income for the 20 percent poorest households amounts to only three-quarters o f that of the average Macedonian households. Table 3.4: The Roleof InformalEmployment and Social Security for Poorer Households Share o f HouseholdIncome by Source By expenditure quintile Source o f Income Poorest 2 3 4 Richest Agricultural Income I' 8 9 9 8 7 Nonagricultural Income I' 52 58 58 59 61 Formal Income 41 50 50 54 5 1 Informal 11 9 7 5 4 Public social security 28 24 23 22 22 Pensions 22 22 21 20 21 Public Transfers 6 3 2 2 1 Private Transfers 9 7 8 9 7 Other '` 3 2 3 2 3 From employment or self-employment. 2Scholarships, patents, insurance, and income from property. Source: World Bank staff estimates based o n HBS 2003. 27 3.15. Low-income families are burdened by many non-working members. As seen in Figure 3.4, poor households have fewer members working on a regular job, even though they have much larger families. On average, poor households have more than twice as many dependent, non-working members as non-poor households and fewer working members to support them. In fact, poor households have, on average, only 0.6 members with a regular job, compared to 0.9 for non-poor families. Importantly, in both poor and non-poor families, there is on average one person who derives income in theform of unemployment bene$& pensions and social security income, or rental income (summed up as personal income in the figure below), i.e., more than the number o f people with a regular job.'' Figure3.4: Income Opportunities ofthe Poor are Hamperedby LimitedAccess to GoodJobs Household Members by Working Status r I I Non-poor Poor 0 Regularjob 0 Inagriculture 0 Self-enployed 0 Personalincome 0 Dependent. Source: World Bank staff estimates based on HBS 2003. 3.16. The poor have limited access to jobs.16Clearly, an important determinant of a household's poverty status is the amount o f income members is able to secure in order to support the entire household. This is in turn a function of many different factors: the number o f children and elderly to support, the share o f household memberso f working age who are actually active in the labor market, how manyo fthese are able to secure ajob, andhow highthe wage rate i s (see Box 3.2). InMacedonia, as discussed earlier, the poor have somewhat higher dependency rates than the none-poor, but the difference is not remarkable. Employment options appear to play a more important role than demographics inthis sense. Employment rates-the share o fworking age populationthat is employed- are substantially lower for poor than for non-poor people, whether in rural or urban areas as seen in l5Given that most people own their own dwelling, the rental market i s likely to be relatively thin. Hence, rental income is likely to be a relatively small and unimportant source o f income. l6Analyzing the labor market with household budget survey instead o f labor force/employment surveys has some implications interms o f interpretation and comparisons. Unlike H B S data, Labor Force Survey data i s collected specifically for the purpose o f measuring labor market activity and contains much more detailed information on wages, hours worked, and other important aspects. This also means that labor market indicators cannot be derived in the same way from the two sources and so will also differ. H B S data, however, permits analysis along poverty dimensions, Although labor force data cannot be used for measuring poverty directly, it does contain information on what has been shown to be important poverty correlates inMacedonia, such as ethnicity and education levels. , 28 Table 3.6. This is a consequenceboth o f the fact that much fewer poor are active inthe market and that when they do look for a job, they are much less successful in securing one. Whde roughly half o f the non-poor working agepopulationis active, labor force participationrates for the poor are only 27 percent in Skopje and 32 and 35 percent in rural areas and other urban areas respectively. Conversely, unemploymentrates are muchhigherfor the poor. 3.17. Unemployment rates among the poor in Skopje are very high at 43 percent even when compared to the national average of unemploymentamong the poor (28 percent). In Skopje, only about three intwenty poor o f worhng age are actually employed, compared to eight out o f twenty non-poor. Why unemployment rates are so much higher in Skopje requires further investigation but could be related to such factors such as higher relative wages or better access to social assistance programs. 3.18. Though the vast majorityof the nonpoor classifiedas "unemployed" by the survey receive benefits, most of the poor unemployedeither receive no benefits or are engaged in occasional labor. The definition o f unemployment used here includes those with access to official unemployment benefits, and those who have seasonal or occasional work only. It excludes those classified as unemployed without benefits who instead are classified as inactive. Alternative definitions o f unemployment only shift people between inactive and unemployed, however, and it remains the case that the poor are seriously disadvantaged in the labor market." Figure 3.5 shows unemployment rates if unemployed without benefits are included among the unemployed (and thus inthe active population). As seen, the poor have a disproportionate amount o f people with access to occasional jobs only. The differences are most pronounced in Skopje, although the non-poor there have more people that are unemployedwithout benefits than the non- poor. Table3.5: ParticipationRates of the Poor are Low andUnemploymenti s High Dependency Employment, Labor Force Participation, and Unemployment Rates for Households inRural and UrbanAreas by Poor andNon-Poor Status (inpercent) Labor Force DependencyRates total 15-64yr (as % of Participation(as % of Unemployment(as % (1 and <=2 19.1 7.6 2.1 0.8 N o . o f Ratio >2 and <=4 31.9 17.5 4.6 1.8 employed) Ratio >4 25.4 33.4 10.5 4.4 N o employed 22.1 27.7 9.7 4.5 Access to land N o 63.9 24.5 7.8 3.3 Yes 36.1 _. 16.8 4.8 1.9 *Group of occupation ofpersons -income recipients: pensioners, war and civil disable persons, social benefits recipients, scholarship recipients, property income recipients, other income recipients **Group of dependentpersons: housewifes, children below 7,children inprimary and secondary school, students, Elderly people, Other dependents Source: HouseholdBudget Survey, 2003 54 HOUSEHOLD CONDITIONSBYPOVERTY STATUS LIVING Table $5. Householdcharacteristics Macedonia Povertystatus Nonpoor Poor Total number of householdmembers 4.0 3.8 5.2 Number of memberswith regularjob 0.8 0.8 0.6 Number of agriculturists 0.2 0.2 0.2 Number of self-employed 0.03 0.03 0.04 Number of members with personalincome 0.9 0.8 1.o Number of sustentativemembers 2.1 1.8 3.3 Source: HouseholdBudgetSurvey, 2003 Table S6. Type of flat (Percentof population) Macedonia Povertystatus Nonpoor Poor Room 0.8 0.7 1 Apartment 0.3 0.3 0.5 One-roomdwelling 5.2 4.7 7 Two-room dwelling 26.3 25.4 29.8 Three-roomdwelling 29.9 29.4 31.8 Four-roomdwelling andmore 37.3 39.4 29.7 Not living area 0.1 0.1 0.1 Total 100 100 100 Source: HouseholdBudget Survey, 2003 Table S7. Ownershipof flat (Percentof population) Macedonia Povertystatus Nonpoor Poor Owner 91.4 91.9 89.5 Leaseholder ofthe state dwelling 0.7 0.5 1.1 Leaseholder (of the whole dwelling) inprivate property 0.4 0.3 0.9 Leaseholder (of the partof dwelling) in privateproperty 0.1 0.1 0.1 Relationship(living at relative'shouse) 7.2 7 8.1 Other 0.2 0.2 0.3 Total 100 100 100 Source: HouseholdBudgetSurvey, 2003 55 Table S8. Dwelling area Macedonia Povertystatus Non poor Poor Average m2 per household. 79 80 73 Dwelling areaper capita (m2) 19.7 21.3 14.0 Source: HouseholdBudget Survey, 2003 Table S9.1113. Installations indwelling(Percent of population) Macedonia Povertystatus Nonpoor Poor Water supply 93.7 94.9 89.4 Sewage disposal 74.8 77.8 64.0 Electricity 99.4 99.5 98.9 Centralheating 10.1 11.9 3.6 Telephone line 90 93.3 78.1 Source: HouseholdBudgetSurvey, 2003 Table S10. Amenitiesindwelling (Percent of population) Macedonia Povertystatus NonDoor Poor Kitchen 93.3 94.9 87.7 Bathroom 91.2 92.9 85 Terrace 71.3 76.6 52.2 Garage 26.7 30.1 14.5 Other 53.9 56.8 43.3 Source: HouseholdBudgetSurvey, 2003 Table S11. Way of heating(Percent of population) Macedonia Povertystatus NonDoor Poor Centralheating 4.9 5.9 1.1 with private installation- on electricity 1.3 1.4 0.9 with private installation on solid fuels - 1.6 1.8 0.7 with private installation- on liquid fuels 0.8 1 0.1 with private installation- combined 0.3 0.4 0.1 Heatingby stove on electricity - 10.3 11.2 7.2 Heatingby stove - on solid fuels 80.3 77.8 89.4 Heatingby stove - on liquid fuels 0.5 0.5 0.4 Total 100 100 100 Source: Household BudgetSurvey, 2003 56 Table S 12. Householdsupplies with durablegoods (Percentof households) .* MaCeclOnla * . Povertv status Non poor Poor Solid fuel cooker 76.2 74.3 85.4 Electric cooker 93.4 94.2 89.4 Gas and combined cooker 3.1 3.4 1.7 Solid fuel stove 15.9 15.8 16.3 Electric stoves (all kinds) 34.2 36.5 23.1 Oil stove 1.4 1.6 0.6 Gas stove 0.8 0.8 0.9 Boiler 88.6 90 81.2 Refrigerator 96.4 97.1 93.3 Freezer 82 84.7 68.6 Washing machine 80.2 83.4 64.3 Iron 93.7 94.4 90.5 Dish-washer 4.5 5.1 1.9 Sewing machine 33.1 35.9 19.7 Vacuum cleaner 82.9 85 72.8 Air conditioner 5.8 6.7 1.2 Radio 35.3 36.2 30.6 Record player 10.2 10.5 8.7 Cassette recorder 49.6 50.2 46.3 Hi-fisound equipment 23.2 25.1 14.1 Black and white television set 4.8 4.4 6.9 Color television set 94.8 95.4 92.2 Video recorder 42.6 45.2 29.6 Video camera 3.1 3.6 1.1 Camera 31 33.8 17.1 Personal computer 10.6 11.8 4.9 Telephone 88 90.6 75.3 Mobile phone 32.9 34.7 23.9 Bicycle 34.6 36.4 25.4 Motorcycle 5 5.2 3.8 Motorcar 45.8 49.7 26.7 Auto trailer 1.3 1.5 0.3 Boat 0.5 0.6 0.1 Motorboat 0.4 0.5 0.1 Guitar 2 2.3 0.6 Piano 0.4 0.5 0 Source: HouseholdBudget Survey, 2003 57 EMPLOYMENT INFORMATION Table S 13. HouseholdEmploymentInformation. 0 22.1 20.4 28.3 1 41.4 40.1 46.1 2 28.1 30.7 18.9 3 6.3 6.4 5.7 4 1.8 2.0 0.8 5 0.3 0.3 0.2 Total 100.0 100.0 100.0 Average numberof employed members in household 1.25 1.31 1.05 Source: HouseholdBudgetSurvey, 2003 Table S14. Distributionof PopulationAged f 5-64 by Labor Market Status. Skopje Other Urban Rural Non-poor Poor Non-poor Poor Non-poor Poor TOTAL MACEDONIA 246260 59349 424952 116894 430964 117648 EMPLOYED 99751 9060 174256 29276 177830 29807 Employed 95280 8508 142033 22286 60565 12058 Agriculturist 352 97 3773 1492 37562 6809 Self-employed 2477 174 3716 2596 2828 806 Employedand work on own land 1642 115 19594 2546 35793 3548 Self-employedandwork on own land 353 1537 Unpaidfamily worker on own land 166 4187 356 39545 6586 UNEMPLOYED 17040 6809 35306 11812 31232 7611 Unemployedwith official unemploymentbenefit 6888 1907 20733 3657 10445 1636 Seasonal, not officially worker or occasionalworker 10152 4902 14573 8155 20787 5975 INACTIVE 129469 43480 215390 75806 221899 80229 Unemployedwithout benefit 38540 10112 65568 23335 38747 14466 Pensioner 26922 4223 33888 6446 14104 3926 Socialand other benefitsearner 3472 4360 9516 9733 14400 7603 Supportedperson(disable, housewife, kids) 60446 24785 101017 35829 142759 52313 Agricultural pensioners 40 62 Elderlypeoplewith own landand benefitfrom land Pensionerworking on own land 89 4864 9053 503 Social and other benefitsearner and work on own land 537 463 2796 1356 ~~ ~~ Source: HouseholdBudget Survey, 2003 58 Y SUBJECTIVE POVERTY Table S17. The satisfactionlevel bv the monthlv income of the household(HHlevel) Question: Having on mindall your monthlyincomks, does the household make'ends mee;? Macedonia Poverty status Nonpoor Poor Completely satisfied . . 3.1 3.7 0.2 More or less satisfied 28.9 32.2 12.3 More or less not satisfied 43.1 43.7 40.5 Not satisfied at all 24.9 20.4 47.0 Total 100 100 100 Source: Household Budget Survey, 2003 Table S18. The minimummonthlyincome of the household neededto make ends meet (HH level) Macedonia Poverty status Nonpoor Poor The minimumhouseholdmonthly income level needed to make ends meet (MKD) 25024 24985 25215 The minimummonthly income level per capita needed to make ends meet (MKD) 7047 7403 5284 Source: Household Budget Survey, 2003 Table S19. The rate of the current financial situationof the household(HHlevel) Question: Compared to one year ago, how would you rate your current financial situation? Macedonia Poverty status ' Nonpoor Poor Much better 0.2 0.2 0.0 Somewhatbetter 3.5 3.8 1.9 About the same 46.3 49.3 31.4 Somewhatworse 29 28.9 29.2 Muchworse 21.1 17.8 37.5 Total 100 100 100 Source: Household Budget Survey, 2003 Table S20. Can the household satisfy any of following needs (Percent of Households) Macedonia Poverty status Nonpoor Poor Keeping home adequately warm 41.4 45.2 23.0 Oneweek holiday 19.7 22.7 4.6 Replacing furniture 4.4 5.2 0.5 Eating meat or fish every second day 35.2 39.1 16.1 Having friendsifamily for drink or meal at least once amonth 30.8 34.2 13.8 Source: Household Budget Survey, 2003 Table S21. Duringthe past3 months has the householdbeenable to (Percent of Households) Macedonia Poverty status Non poor Poor Pay rent for accommodation 14.5 16.1 6.9 Pay utility bills (elec., water, tel., central heating) 71.6 76.5 47.5 Repay credit or loans 14 15.8 4.7 Source: Household Budget Survey, 2003 60