Armenia: Poverty Assessment (In Three Volumes) Volume I: A Summary of Findings

Transcription

1 Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Report No: AM Armenia: Poverty Assessment (In Three Volumes) Volume I: A Summary of Findings November, 2003 Human Development Sector Unit Europe and Central Asia Region Document of the World Bank 1

2 FISCAL YEAR January 1 - December 31 CURRENCY EQUIVALENTS Currency Unit = Armenian Dram WEIGHTS AND MEASURES Metric System ABBREVIATIONS AND ACRONYMS CIS Commonwealth of Independent States ECA Europe and Central Asia FPB Family Poverty Benefit GDP Gross Domestic Product ILSC Integrated Living Conditions Survey PRSP Poverty Reduction Strategy Paper WB World Bank FSU Former Soviet Union EPEAS - Education, Poverty, and Economic Activity Survey AHDS Armenia Health and Demographic Survey MOH Ministry of Health IMR Infant Mortality Rate MMR Maternal Mortality Rate Vice President: Country Director: Sector Director: Sector Manager: Team Leader: Shigeo Katsu, ECAVP Donna Dowsett-Coirolo, ECCU3 Annette Dixon, ECSHD Michal Rutkowski, ECSHD Aleksandra Posarac, ECSHD 2

3 TABLE OF CONTENTS TABLE OF CONTENTS 3 ACKNOWLEDGEMENTS 4 EXECUTIVE SUMMARY 5 1. MAIN FEATURES OF ARMENIA S GROWTH PERFORMANCE THE POVERTY PROFILE NON-INCOME DIMENSIONS OF POVERTY RURAL DEVELOPMENT AND POVERTY DEMOGRAPHICS, MIGRATION AND LABOR MARKET PROTECTING THE POOR: SOCIAL PROTECTION PROGRAMS AND THEIR IMPACT ON POVERTY 43 BIBLIOGRAPHY 52 3

4 ACKNOWLEDGEMENTS Armenia Poverty Assessments is a collective effort of a team lead by Aleksandra Posarac (ECSHD) assisted by Edmundo Murrugarra (ECSHD). The multi-sectoral team comprised: Lev Freinkman (macroeconomic developments), Julia Bucknell and Villiam Sutton (rural developments), Hellen Shahriari (subjective aspects of poverty), Dilnara Isamiddinova (social protection) and María Eugenia Bonilla Chacín (education), all from the World Bank, and consultants Gorana Krsitc (poverty profile, labor market), Paulette Castel (social protection, rural developments, private transfers), Stefania Rodica Cnobloch (health), Artsvi Khachatryan (macroeconomic developments) and Hjalte Sederlof. Following a series of team discussions and consultations with the client, this Report was put together by Aleksandra Posarac and Hjalte Sederlof. A number of people contributed by commenting on various drafts and generously providing information: Mark Lundell, Ellen Hamilton, Julian Lampietti, Susanna Hyrapetyan, Peter Nicholas, Mansoora Rashid, Toomas Palu, Toby Linden, Pierella Paci, and Jan Rutkowski. The team owes a special gratitude to the National Statistical Service of Armenia, as well as UNICEF and UNDP Offices in Armenia for cooperation and generous sharing of information. 4

5 EXECUTIVE SUMMARY This report studies poverty in Armenia in 2001 and examines the most recent trends covering the 1998/99 to 2001 period. It looks at the determinants of poverty and analyzes linkages between economic growth, sector policies and poverty. The findings are based on two rounds of the Armenia Integrated Living Conditions Survey (ILCS), one carried out in 1998/99 and the other in The report has contributed to the development of Armenia s national strategy for growth and poverty reduction. Poverty is high, but falling. Poverty is high in Armenia with an estimated 48 percent of the population below the poverty line in Some 20 percent live in extreme poverty, i.e. they are unable to attain the minimum recommended daily energy intake from food consumption. Still, there has been a significant decline in poverty, as poverty incidence has dropped by 12 percent and extreme poverty incidence by 25 percent from their respective levels in 1998/99. The reduction in poverty incidence was accompanied by decreases in the depth and severity of poverty. Income inequality also fell, from 0.64 to 0.54, as measured by the Gini coefficient, but Armenia still remains the country with the highest income inequality among transition (Europe an Central Asia ECA Region) countries. However, consumption inequality was much lower at 0.28 in 2001, which is low compared to countries at similar per capita income levels. Household size and composition are strong indicators of poverty, as are low levels of educational attainment and unemployment. Poverty risk increases with household size, but it also depends on household composition: the presence of children considerably increases the incidence, depth and severity of poverty; and poverty risk is further increased when there is an elderly person present in the household. There also is a strong correlation between poverty and low educational attainment, while the unemployed and non-participants in the labor market face the highest poverty risk and depth and severity of poverty. In rural areas, poverty is positively correlated with the size of landholdings. The drop in poverty reflects a decline in urban poverty since 1998/99. Rural poverty has increased slightly. Yerevan, the capital and the largest urban area in Armenia, registered the most significant reduction in poverty, as poverty incidence there dropped by 23 percent and extreme poverty by over a third, to 45 percent and 20 percent, respectively. While poverty in other urban areas also decreased, their poverty incidence still remains above the national rate, at 52 percent. Poverty incidence is now roughly the same in urban and rural areas, 48.5 percent and 47.9 percent, respectively. The improved situation in urban areas, and in particular Yerevan, reflects successful macroeconomic policies, especially since The economy has averaged a robust 6.3 percent growth per annum over the 1999/2001 period, and reform strategies that have focused on improving the investment climate and broadening the base for economic growth have had favorable outcomes. The employment content of GDP has increased as a result of more rapid growth in labor intensive sectors, especially construction and retail trade, which have created more employment opportunities. 5

6 The rural economy so far has not benefited from economic growth. The sector expanded greatly over the 1990s, because it absorbed much of the labor released when state enterprises were restructured or closed. Although some parts of Armenia have good agricultural potential, most of the country s land is basically unsuitable for commercial agriculture. Farms are generally small, and much of agriculture produces only meager returns. Few farmers are able to exploit the potential economies of scale in agriculture. Many households that are not primarily farmers and have and work land plots, could more productively engage in other economic activities. While farming provides the bulk of household income in rural areas, it is unable to allow most rural households to meet their needs unless accompanied by other economic activities, i.e. onand off-farm labor earnings. With little opportunity for off-farm work, many rural households are especially vulnerable. Moreover, much farming in Armenia is dependant on irrigation, irrigation systems are often rundown, and since rainfall in Armenia is subject to high annual variability, agricultural production takes place under significant uncertainty. Particularly volatile weather conditions have reduced yields in 1999 and 2000, and at the same time farmers have faced worsening terms of trade: producer prices for agricultural products have been falling relative to other prices, including those for productivity enhancing farm inputs. Non-income indicators of poverty are deteriorating, including quality of and access to education and health care and adequate housing and public services. The social protection system, while well-targeted and providing an important cushion against poverty, remains modest and resource-constrained. While participation in basic education is high, it is low at other levels, especially among children from poor households. Cut-backs in public spending has led to deteriorating quality of public education and increasing out-of-pocket payments by families. The poor are particularly vulnerable to rising education costs, as they find it difficult to make out-ofpocket payment, and schools in poor areas suffer accordingly in terms of adequate education provision. While health indicators appear favorable, there are significant discrepancies between survey and official estimates, with survey results indicating early childhood mortality rates almost double the official estimates. Rural mortality rates are significantly worse, reflecting reduced access to adequate ante-natal care and supervised delivery in rural areas. Inadequate budget resources increase incidence of formal and informal out-of-pocket payments for health services, especially in rural settings, depressing health care utilization across income groups, and especially among the poor. In spite of modest benefit levels, the social protection system is important in reducing poverty. Pensions and social assistance are relatively well targeted, and simulations indicate that poverty incidence would increase by some 6 percentage points if they were withdrawn. Still budget allocation for social assistance has declined by some 30 percent in real terms since The main program to experience reductions is the family poverty benefit, which currently provides the most effective targeting of the poor. Resources for that program have declined by 35 percent. However, the 2003 draft Poverty Reduction Strategy Paper commits the Government to gradually increase the share of GDP allocated to family poverty benefit. 6

7 Inadequate housing and public services (heating, water, and waste disposal) is widespread and contributes to deteriorating living conditions for the poor. In urban areas, crowding is a problem, especially for the poor. Policy conclusions and recommendations Government policies aimed at macroeconomic stability and diversification of the economy, including elements that increase the employment content of growth, have met with considerable success in reducing poverty in urban areas. These policies, including further improvement of the business climate and facilitating the development of a dynamic private sector, will need to be maintained and expanded to further increase employment and reduce poverty. Generation of offfarm employment will be especially important for tackling rural poverty: off-farm employment will complement rural household incomes, and it should allow a gradual consolidation of farmland as families finding off-farm opportunities rescind marginal plots. Essential complementary policies should include the development of land markets to facilitate the process of land consolidation, and the rehabilitation of irrigation infrastructure. Increasing the level and quality of education and ensuring better health for all are central enabling strategies for poverty reduction. Children from poor families should be offered the same educational opportunities as children from better-off families. Inadequate public expenditures and the need for out-of-pocket payments to maintain adequate education standards place children from poor families at a disadvantage. While public resources will continue to be constrained, efficiencies should be sought in the use of capital and labor in the system, and attention paid to apply resources more equitably with consideration given to specific vulnerabilities of poor children, including measures that will encourage attendance of poor children in post-primary education. Inequities in health care provision arising from underfunding of primary services that especially cater to lower income groups, in particular in rural areas, and from increased out-of-pocket payments, need to be addressed: public resources should increasingly be channeled into cost effective primary care, and private payments, while they need not be discouraged, should be limited to those who can afford them. Social protection plays a crucial role in the poverty reduction strategy, providing permanent assistance to those who cannot directly benefit from the growth process, as well as temporary help for vulnerable households. While benefits are modest and the scope for increasing them limited, there should be a focus on gradually increasing essential benefits pensions and family benefits as well as increasing efficiency in the administration of the family benefit, which is the main social safety net program in Armenia. Consideration should be given to expanding it to cover most of the extremely poor, as well as refining the targeting mechanism. 7

8 8

9 1. MAIN FEATURES OF ARMENIA S GROWTH PERFORMANCE Following the break-up of the Soviet Union, Armenia lost about 60 percent of its GDP between 1989 and Labor market adjustment was dramatic: unemployment became rampant and wages plummeted; poverty became deep and pervasive. Following macroeconomic stabilization in and sustained Government efforts to advance structural reforms, a steady economic recovery has taken place since The economy has been growing every year since at an average rate of about 6.5 percent, placing Armenia among the best performing economies in transition. Still, real GDP in 2002 amounted to only 70 percent of its pre-transition level. Table 1.1: Armenia: Real GDP growth, (percent change over the previous year) Gross Domestic Product Of which: Industry Agriculture Construction Transport and communication Trade Other services Source: National Statistical Service (NSS). Growth trends The growth performance has been sustained by a number of important factors: international assistance to Armenia has been abundant, accounting for some 7 percent of GDP in annual budget support over the period, mostly through a combination of grants and low-interest credits; comprehensive land reform created a network of family farms that provide a degree of protection against transitional shocks; energy supply has been stabilized and now provides regular around-the-clock access to power, normalizing the environment for business development; and most importantly, Government policy has consistently focused on maintaining a stable macroeconomic environment and pursuing structural reforms to improve the investment climate, broaden the base for economic growth and ensure a wider distribution of the benefits of growth among the population. The early stages of the recovery, over the period from 1994 to 1998, were driven by one-time factors (land reform, energy) and growth in enclave industries (mining, metallurgy) that had a limited impact on the broader economy. At the same time, privatization of state enterprises shed workers that the gradually expanding formal private market was unable to absorb. While some unemployed found jobs in the informal sector, unemployment skyrocketed, reaching 24.4 percent in Lack of job opportunities reduced participation rates and fuelled emigration; it also saw increasing reliance on agriculture in the aftermath of the land reform. While real wages increased overall, reflecting the effects of new skill-intensive businesses, wages in the public sector and agriculture fell in the face of excess labor. 9

10 Figure 1.1: Armenia: Employment trends (in 000) Total employment Private Non-private Agriculture Non-agriculture Since mid-2000, the Government s reform policies have become to make themselves felt, and the country s economic performance has further improved: the average annual rate of economic growth in the three years (2000 to 2002) exceeded 8 percent; and poverty and income inequality declined. Particularly important were improvements in the business environment. These are driving diversification in manufacturing towards higher skill and more labor intensive activity; improved business confidence, including increased inflow of Diaspora and other foreign direct investment; and increased formal sector activity, reflected in increased numbers of registered tax payers. In turn, these developments are contributing to greater tax revenues and facilitating budget management and economic stability. Still, overall net job creation remains slow relative to the economic growth rate. This is consistent with international experience and seems unavoidable for the early stage of economic restructuring. During this stage, private sector expansion occurs in tandem with public sector retrenchment, which dampens the overall positive impact of growth on job creation. In addition, new private sector jobs are mostly those with higher productivity, which also reduces demand for labor in early transition. Policy conclusions and recommendations Armenia s growth performance has been strong over the past eight years, and it has even accelerated in the most recent years, as the structure of economic growth has improved with a gradual diversification of the economy. There has been a gradual diversification of industry as new enterprises have sprung up, and modern services have expanded. These developments indicate that recent government policies have been in the right direction, and the government should continue to follow its current economic strategy over the medium term in order to achieve its poverty reduction goals. It should maintain macroeconomic stability; continue to improve the investment climate; further strengthen its capacity to promote investments and exports; and improve public revenue performance and fiscal management, including raising the efficiency of core services. 10

11 At the same time, it will need to continue to strengthen elements of the strategy which help diversify the economy and increase the employment content of growth. It will in particular need to focus on measures that will further improve the business climate and help create a dynamic private sector. Such measures may include: broadening the dialogue with the private sector to identify major bottlenecks for private sector development, including articulating the Government s role in creating an enabling environment for private enterprise; strengthening cooperation with the Diaspora in order to channel a larger share of charity inflows into productive investment; developing business support services, especially those targeted at small and medium size enterprises; improving efficiency of ongoing micro-credit programs through consolidation and better coordination between donors 11

12 2. THE POVERTY PROFILE Good economic performance since 1999 has resulted in improved poverty indicators in 2001 in comparison to 1998/99: overall and extreme poverty incidence, the poverty gap and severity of poverty all declined. Still, poverty in Armenia remains widespread, affecting almost one half of the population. General trends The poverty analysis is based on the comparison of the results of the 1998/99 and 2001 rounds of the Armenia Integrated Living Conditions Surveys (ILCS). 1 This comparison suggests that, while overall poverty has remained high throughout that period, it still has declined by some 12 percent over the period (see Table 2.1). Extreme poverty declined by 25 percent. The reduction in poverty incidence was accompanied by decreases in the depth and severity of poverty and a decline in income inequality. Incidence (P0) Table 2.1: Armenia Poverty indicators in 98/99 and 2001 (standard errors in parenthesis) Extreme (Food) Poverty Line (8,730 drams in 98/99) (7,979 drams in 2001) Gap (P1) Severity (P2) Incidence (P0) Complete Poverty Line (12,276 drams in 98/99) (11,221 drams in 2001) Gap (P1) Severity (P2) Total 98/ % 6.0% 2.0% 54.8% 16.2% 6.5% (1.94) (0.53) (0.21) (2.26) (0.96) (0.48) % 4.6% 1.6% 48.3% 13.0% 5.1% (1.53) (0.42) (0.17) (2.42) (0.80) (0.38) Urban 98/ % 7.8% 2.6% 61.4% 19.3% 8.1% (2.25) (0.66) (0.28) (2.55) (1.13) (0.59) % 5.0% 1.7% 48.5% 13.7% 5.5% (1.60) (0.36) (0.15) (2.90) (0.82) (0.35) Rural 98/ % 3.7% 1.2% 46.1% 12.1% 4.5% (2.45) (0.54) (0.21) (3.08) (1.18) (0.54) % 4.0% 1.5% 47.9% 12.1% 4.6% (3.04) (0.83) (0.34) (4.29) (1.61) (0.77) 1 The ILSC 1998/99 was carried out between July 1, 1998 and June 30, 1999; the ILSC 2001 from January 1, 2001 through December 31, For information on the surveys and their comparability, poverty lines, welfare measure, and poverty indicators and their comparability over time see Annex I.A to the Armenia Poverty Assessment, World Bank 2003, and Armenia Poverty Update, World Bank Annex I.A also describes the major methodological differences between the NSS and that adopted by the World Bank. 12

13 Source: ILCS 1998/99 and Note. Poverty incidence (P0) is the share of the population (households) whose consumption (or income) falls below the poverty line. The depth of poverty (P1) indicates how far households are from the poverty line. It captures the mean aggregate consumption (income) shortfall relative to the poverty line across the whole population. When calculated across the poor population (P1/P0) it provides information of the poverty shortfall or deficit, that is how much, in terms of the percentage of the poverty line, the mean consumption of the poor on average falls short of the poverty line. The shortfall multiplied by the number of the poor and usually expressed as percentage of GDP provides an estimate of what would be the minimum cost of eliminating poverty in the society, assuming perfect targeting. Severity of poverty (or squared poverty gap) captures the inequality among the poor by effectively giving more weight to households that are further away from the poverty line. Although income inequality as measured by the Gini coefficient dropped from 0.64 to 0.54 over the period, Armenia still remains the country with the highest income inequality among ECA countries. Inequality measured by consumption was a much lower 0.28, which compares favorably to countries with similar per capita incomes. Labor earnings represented the largest contributor to income inequality, explaining almost 40 percent of inequality. Government transfers, both pensions and social assistance, considerably reduced income inequality, as was the case in 1998/99. Lower poverty rates reflect a decline in urban poverty, which fell by more than 20 percent. Rural poverty rose marginally, and poverty rates are now similar in urban and rural areas. Yerevan, the capital and the largest urban area in Armenia, registered the most significant reduction in poverty: there poverty incidence dropped by 23 percent and extreme poverty by over a third, to 45 and 20 percent, respectively. Poverty in other urban areas also decreased, but their poverty incidence, at 52 percent, still remains above the national average of 48.3 percent. The shift in urban/rural poverty reflects the effects of more labor intensive growth that favored urban areas. The rural population was faced with increased scarcity of jobs, as bad weather conditions in 1999 and 2000 combined with worsening agricultural terms of trade and declining farm incomes to cause a doubling of rural unemployment. Economic growth and poverty Table 2.2 illustrates the impact of different growth rates and levels of inequality on extreme and overall poverty in 2005 and Two growth scenarios are presented, 3 and 5 percent, with and without an increase in inequality. Three percent growth over the next four years (until 2005 and taking 2001 as a base year) would reduce extreme poverty by 29 percent and overall poverty by 23 percent. The poverty gap would decline even more (32.6 and 28.5 percent respectively). Higher growth rates would have even more favorable effects. The introduction of greater inequality - it is highly probable that economic growth will be accompanied by increasing differentiation in consumption - has a countervailing effect on poverty incidence: at 3 percent growth, a 1.5 percent annual increase in consumption inequality would reduce the growth effect on poverty and extreme poverty by some two percentage points. Over the longer term, differences in poverty incidence would be higher. Cost of poverty reduction To eliminate poverty in Armenia through public transfers (social assistance), would require 53.1 billion drams or 4.5 percent of GDP (in addition to the 1.43 percent of GDP that was already spent on assistance to poor families in 2001), assuming perfect targeting. As perfect targeting is 13

14 unlikely, the real monetary costs would be significantly higher. In market economies, such costs tend to be double those under perfect targeting; in transition countries, the cost of providing the equivalent of 1 US dollar of welfare to the poor ranges from 1.5 US dollars to 8 US dollars (not taking into account administrative costs). 2 As the transfer costs of eliminating poverty would be prohibitive, inclusive economic growth is the only possible way to reduce poverty in a sustainable manner. Table 2.2: Projections of extreme and overall poverty (standard errors in parenthesis) Annual growth Extreme poverty incidence Poverty gap Extreme poverty (A) Inequality remains constant 0% 20.0% 20.0% 4.6% 4.6% (1.5) (1.5) (0.42) (0.42) 3% 14.2% 4.9% 3.1% 0.9% (1.3) (0.6) (0.3) (0.1) 5% 11.1% 1.5% 2.3% 0.3% (1.0) (0.3) (0.2) (0.1) (B) Inequality increases by 1.5 percent annually 0% 22.3% 28.1% 5.7% 9.3% (1.6) (1.8) (0.5) (0.7) 3% 16.2% 12.3% 4.0% 4.1% (1.4) (1.2) (0.4) (0.4) 5% 13.4% 7.1% 3.2% 2.4% (1.2) (0.7) (0.3) (0.3) Overall poverty (A) Inequality remains constant 0% 48.3% 48.3% 13.0% 13.0% (2.4) (2.4) (0.80) (0.80) 3% 37.0% 16.0% 9.3% 3.6% (2.0) (1.4) (0.7) (0.3) 5% 30.3% 6.8% 7.4% 1.3% (1.0) (0.3) (0.6) (0.2) (B) Inequality increases by 1.5 percent annually 0% 49.2% 52.0% 14.4% 18.1% (2.4) (2.5) (0.9) (1.1) 3% 38.7% 24.2% 10.7% 8.0% (2.1) (1.7) (0.7) (0.7) 5% 32.3% 14.1% 8.7% 4.7% (1.9) (1.2) (0.6) (0.5) Source: ICLS 98/99 and Note: Consumption aggregate does not include imputed rental value of dwelling. Standard errors computed with PSU adjustments. Who are the poor? (a) Poverty and extreme poverty is concentrated among the young and the very old. The highest overall poverty incidence, exceeding 52 percent, and the highest poverty risk (8.1 percent over the national average), appeared among children under five years of age. Poverty was deepest and most severe among young individuals between ages 19 and 25. Some 51 percent of the elderly (over 60) were poor with poverty risk above the national average (by 5.6 percent). 2 Braithwaite, J, Grootaert, C., and Milanovic, B., Poverty and Social Assistance in Transition Countries,

15 Extreme poverty incidence Table 2.3: Poverty by gender and age groups in 2001 Poverty incidence Relative poverty risk % of Population % of the poor Poverty gap Severity Gender Male 19.6% 48.4% 0.2% 48.1% 48.2% 12.9% 5.1% Female 20.3% 48.2% -0.2% 51.9% 51.8% 13.2% 5.2% Age groups Children % 52.2% 8.1% 6.8% 7.3% 14.3% 5.9% Aged % 45.6% -5.5% 16.7% 15.8% 11.6% 4.3% Aged % 49.3% 2.2% 8.3% 8.5% 13.7% 5.5% Aged % 51.6% 6.8% 12.1% 12.9% 15.0% 6.2% Aged % 47.6% -1.3% 28.6% 28.3% 12.7% 5.0% Aged % 44.4% -8.0% 13.3% 12.2% 12.1% 4.8% Aged % 51.0% 5.6% 14.2% 15.0% 13.6% 5.3% Total 20.0% 48.3% - 100% 100% 13.0% 5.1% Source: ILCS (b) Household composition and size matter; larger households, households with elderly and/or children and female headed households with children were more likely to be poor. Larger households were more likely to be poor, but this also depended on household composition. Presence of children considerably increased the poverty incidence, gap and severity of poverty; only households with two adults and two children had lower than the average poverty risk. If elderly are included, poverty gap and severity also increased considerably. Female-headed households with children were also more likely to be poor and also displayed deeper and more severe poverty than average. Household type Table 2.4: Poverty measures by household composition in 2001 Extreme Poverty Relative % % of poverty incidence poverty of the poor incidence risk population Poverty gap Severity of poverty single member households 19.9% 37.5% -22.4% 2.5% 1.9% 12.4% 5.5% 2 adults, 2 children 13.9% 40.9% -15.4% 11.3% 9.6% 9.6% 3.3% 2 adults, 2 children, 1 elderly 24.3% 54.4% 12.7% 4.0% 4.5% 15.1% 5.8% 1 adult, with children 23.7% 46.6% -3.4% 2.1% 2.0% 14.3% 6.2% 1 adult, 1 elderly, with children 32.5% 66.2% 37.1% 1.2% 1.6% 19.3% 8.5% 2 elderly, no children 13.8% 46.5% -3.8% 2.6% 2.5% 10.5% 3.6% 2 elderly, 2 children 50.0% 75.0% 55.3% 0.1% 0.1% 22.1% 8.0% Female head, no children 21.3% 46.5% -3.8% 5.9% 5.6% 14.0% 6.2% Female head, with children 27.2% 54.8% 13.5% 15.3% 17.3% 16.9% 7.2% Source: ILCS Note: Children are individuals up to 18 years of age. The elderly are defined as 60 and over. (c) The higher the education level, the lower the overall and extreme poverty risk. 15

16 Households whose heads had primary or incomplete secondary education did not only have poverty rates well above the average, their poverty was also deeper and more severe. In contrast, households where the head had higher education had the lowest poverty incidence, almost half that of those with only primary education. This is to be expected in a transition environment, where returns to (higher) education tend to be high. Table 2.5: Poverty and education of the household head in 2001 Extreme Poverty Relative % of poverty incidence poverty risk Population incidence % of the poor Poverty gap Severity Primary 29.3% 60.4% 25.1% 9.7% 12.1% 17.8% 7.8% Incomplete Secondary 25.3% 59.6% 23.4% 13.0% 16.1% 16.8% 6.8% Complete Secondary 22.8% 53.6% 11.1% 31.7% 35.2% 14.7% 5.7% Technical 17.1% 43.0% -10.9% 26.6% 23.7% 11.4% 4.4% Higher Education 10.9% 32.8% -32.1% 19.0% 12.9% 7.7% 2.8% Total 20.0% 48.3% - 100% 100% 13.0% 5.1% Source: ILCS (d) The unemployed and labor market non-participants face a high probability of being poor. While the situation somewhat improved in urban areas, it significantly worsened in rural. The unemployed and non-participants face the highest poverty risk and the highest levels of depth and severity of poverty, as was the case in 1998/99. The unemployed, and especially those living outside Yerevan, are most at risk. Non-participants also experienced positive relative poverty risk, but their risk was significantly lower than that of the unemployed. Pensioners, the largest group among non-participants, faced the highest poverty risk among non-participants. Sources of household income Labor market earnings were the major source of income of the poor and non-poor alike (Table 2.6). The second largest source of income of the poorest households was government transfers (pensions, social assistance and other transfers). About 22 percent of household income in the poorest quintile was derived from government transfers. The third largest source of income of the poorest households was income from farm activities. Around 20 percent of the income of the poorest quintile was derived from farm incomes. For the poorest in rural areas subsistence agriculture provided a safety net in coping against extreme poverty. Remittances made up about 17 percent of the poorest households income. Income from self-employment was not an important source of income for the poor: it represented only about 6 percent in the poorest quintile. 16

17 Table 2.6: Armenia: Household income sources in 2001, by regions and consumption (in %) Poorest Average All Households Labor earnings Self-employment Farm Income Remittances Transfers Pensions Other transfers Assets sold Total Yerevan Labor earnings Self-employment Farm Income Remittances Transfers Pensions Other transfers Assets sold Total Other urban Labor earnings Self-employment Farm Income Remittances Transfers Pensions Other transfers Assets sold Total Rural Labor earnings Self-employment Farm Income Remittances Transfers Pensions Other transfers Assets sold Total Source: ILCS Notes: Quintiles defined as per adult equivalent consumption quintiles. Conclusions 17

18 Good economic performance since 1999 resulted in a decreased overall and extreme poverty incidence in Armenia in Although the reduction in poverty incidence was significant in 2001 relative to 1998/99, poverty was still widespread among Armenian population in 2001, as almost one half of the population lived in poverty and one fifth lived in extreme poverty. Only urban households benefited from the economic growth. Rural households experienced a reduction in consumption, and an increase in poverty incidence. As a result, there was no clear distinction between the poverty incidence in urban and rural areas. However, urban poverty was deeper and more severe than rural, particularly in urban regions outside Yerevan. The Armenian poverty profile remained almost unchanged in 2001 relative to 1998/99. There were no gender differences in poverty risk. Poverty affected more the young and the elderly. Larger households, households with elderly and/or children and female headed households with children were more likely to be poor. Less educated were more vulnerable to poverty. The unemployed faced the highest poverty risk. Non-participants in the labor market were also more likely to be poor relative to the national average. Key determinants of household consumption were household size and composition, education of the household head, participation in the labor market of the household members, unemployment of the household head and household members, access to land, possession of livestock and household location. Poverty is tightly linked to labor market developments: while the situation somewhat improved in urban areas, it significantly worsened in rural. Yet, the labor market in Armenia remains depressed. Given persistence and pervasiveness of poverty in Armenia, the key elements of the strategy for its reduction are: (i) sustaining economic growth and stability; (ii) fostering private sector development, which would result in more jobs and provide opportunities for those Armenians who are willing and capable to participate in the labor market through gainful employment; (iii) protecting and developing human capital across socioeconomic groups, through public interventions in education and health; (iv) assisting the poor and vulnerable in managing social risks better and coping with them when they materialize, though social protection policies. 18

19 3. NON-INCOME DIMENSIONS OF POVERTY 3.1 Human capital and the poor: education The poverty profile emphasizes the welfare gains from education: non-poor households have higher levels of educational attainment than poor ones, and welfare gains increase with higher educational attainment. The future welfare of today s children depends crucially on equitable access to good education. Equity of the education system - access While public resources earmarked for education have been scarce, Armenia has been able to maintain high levels of participation in basic education during the transition. But at pre-school and post-basic levels, participation is low, especially among children from poor households. Coverage. Participation in basic education in Armenia remains high at a net enrollment rate of 94 percent in 2001, with all consumption quintiles exceeding 90 percent. 3 However, rates are considerably lower at other education levels, with the poorest quintile having noticeably lower enrollment rates than others (Table 3.1.1). Table 3.1.1: Armenia: Net enrollment rates across school levels and consumption quintiles Consumption quintiles Poorest II III IV Richest Total Pre-school (3-6 years) Basic (grades 1-8) Upper Secondary Tertiary Source: Estimates from EPEAS data Note: Net enrolment rate is defined as the ratio between the number of children in an age group currently enrolled divided by the total number of children in that age group. The age brackets were modified to account for the fact that the EPEAS was implemented at the second half of the academic year. For instance, the age bracket for basic education used was all children age 8-14 plus half of the children age 7 and half of the children age 15. The low enrollment rates in pre-school education are particularly disturbing. They may reflect both supply and demand side constraints: faced with scarce budget resources, the Government has shifted responsibility for pre-school education to local governments with often negligible fiscal capacity, leading to closure of many pre-schools; while the introduction of fees for public pre-school participation has reduced demand (and access), especially among poor families. This must be set against the importance of pre-school education for educational attainment later on; children from poorer families seem to gain especially from attending pre-school. Reasons for non-enrolment: There do not appear to be large supply side barriers to access, except for quality of instruction: low quality of instruction may discourage parents from educating their children beyond the compulsory level. On the demand side, while enrollment is high in basic education, it declines at higher levels, and urban-rural and income-related 3 Estimates based on a UNDP 2001 Education, Poverty, and Economic Activity Survey (EPEAS). 19

20 differences begin to appear. The drop-off in rural enrollment rates may reflect lower returns to post-primary education in rural areas, as curricula inadequately correspond to the demands of rural labor markets, and limited employment opportunities outside farming discourage further study. The drop-off in attendance by children from poor families (Figure 3.1.1) may reflect higher opportunity costs of post basic-education for children from poor households, especially boys, as poor families take their children out of school to earn necessary income for the household. It is noteworthy that only 76 percent of boys from poor households continue beyond basic education, while 90 percent of poor girls do so. As poor children reach tertiary education, there are further dramatic drops, with less than 20 percent of children from the poorest quintile entering higher education. The latter may reflect systemic deficiencies, where children from poorer families are more likely to end up in academically less demanding secondary vocational schools, from which access to higher education is limited. Figure 3.1.1: School attainment across consumption and gender, cohort years old Prop. of students Years of education Non-poor Male Non-poor Fem ale Poor Male Poor Fem ale Source: Estimates from EPEAS 2001 Note: Poor in this figure refers to the people living in households in the poorest 40 percent of the consumption distribution. Education financing Public spending on education is inadequate, and it is reflected in cut-backs in essential services, deteriorating educational quality and increasing education costs for families. The poor are particularly vulnerable to rising education costs, and this contributes to low participation rates by children from poor households beyond compulsory basic education. While public spending on education has made impressive gains over the past five years and now accounts for 2.9 percent of GDP (2001), it still is insufficient in relation to education needs. This is reflected in severe cutbacks in public services: pre-school education, out-of-school institutions, and student stipends; and in negligible resources for new construction, maintenance of existing facilities, or purchase of new equipment and teaching materials. It can also be seen in 20

21 international comparisons, where Armenia compares unfavorably to other CIS countries. 4 To this should be added that actual expenditures consistently fall short of budgeted amounts. In 2000, the executed education budget was only 89 percent of the adjusted budget that year. Public spending on education is increasingly being complemented by private payments, and it is estimated that the average Armenian household with school-age children spends close to 10 percent of total monthly household expenditures on education (ISLC 2001). Households in the poorest consumption quintile spend 7.5 percent, almost all of it on basic education. Households provide for transport, textbooks and materials, in many instances for maintenance and repair of educational facilities 5, and increasingly for private tutoring. An increasingly important expenditure item is private tutoring, as insufficient public expenditures have adversely affected the quality of the instruction. Estimates using the ISLC 2001 show that more than one in every ten students in primary education receives private tutoring. Fees for private tutoring exceed the average total household expenditure per student in basic education, and the often prohibitive cost of tutoring puts it out of reach of many poor families, placing their children at a disadvantage. Increased private spending on education services may have significant effects on school participation. Multivariate regression analysis confirms that school enrollment in Armenia is highly correlated with income; and among children age 14 to 17, affordability was indicated as the main reason for not going to school. As all levels of education increasingly require out-ofpocket payments in order to access education services, these are likely to constitute significant barriers to access, especially for children from poor households, and at higher levels of the education system. At tertiary levels, state order places providing publicly financed free education and needsbased scholarships mitigate the affordability barrier. Free state order places are in principle untargeted and based on entry exam scores. A remarkable number of poor students are awarded free places close to 80 percent of tertiary level students from the lowest quintile receive free places, compared with 30 percent of students from households in the richest quintile. Needsbased state scholarships also play an important role in supporting attendance by the lowest quintiles. They appear well-targeted with some 50 percent of tertiary level students from households in the two poorest consumption quintiles receiving scholarships, while less than 15 percent of tertiary level students from households in the highest quintile receive scholarships. Conclusions and recommendations Increasing the level and quality of education is a central strategy for poverty reduction. Improving the educational opportunities of children from poor families should aim to improve their school attainment and increase their opportunities to break out of poverty. 4 Georgia, Kyrgyz Republic, Azerbaijan, Tajikistan, Uzbekistan, Armenia and Moldova. 5 Data from the EPEAS shows that close to 50 percent of all grade 1-10 students report making payments for maintenance and equipment of their school. Similarly, 12 percent of students in colleges and 4.5 percent of those in universities also report having this type of expenditure. This is not surprising as analyses of government budget shows deferred maintenance and under-funding of utilities at all levels of education (Berryman et al., 2002). 21

22 3 Inadequate public expenditure on education and rising private costs are particularly damaging for children of poor households. While public resources are limited, it is imperative that current budgets, which are among the lowest in any of the countries of the former Soviet Union, are increased if the generational transmission of poverty is to be broken; 3 While the Government needs to spend more on education, it also needs to seek to generate resources by way of improved efficiencies in the use of capital and labor in the system. The recent Armenia Public Expenditure Review (World Bank, 2003) draws attention to maintenance strategies and economies of scale in the use of plant and efficiency gains in the use of labor inputs that could generate education budget savings to be applied to pro-poor education initiatives. 3 Particular attention needs to be paid to the relevance of the education system for the labor market in light of changes in skills demands as a result of new structures of economic production, where basic work competencies may no longer be sufficient. Any additional resources should focus on: 3 Financing under-funded inputs for basic education that have significant favorable impact on learning, or reduce future costs (such as preventive maintenance); 3 Possible poverty-targeted stipends to increase enrollments of children from poor families into secondary education; as well as considering other measures that might reduce the opportunity cost of post basic education to poor families; 3 Improving the quality of education. Many students receive private tutoring to complement their school instruction. This handicaps those who cannot afford it, i.e. children from poor families. In addition, the low quality of education, especially at the upper secondary level, and its lack of relevance for the labor market, may be a reason for low enrolments. It may be necessary then, in order to retain children in school past the elementary level, to improve instruction. This includes strengthening teacher training programs, and improving teacher morale by providing the necessary tools (inputs) and by ensuring adequate and regularly paid salaries; it may also include a review of current curricula, in particular with regard to those students who may not be continuing into tertiary education but instead venture into the labor market; 3 Examine more closely the current structure of secondary education to facilitate the entry of poor children into higher education. International observation has concluded that differentiated programs or school types lead to clustering of students with particular socio-economic characteristics in separate facilities. Such also appears to be the case in Armenia where children from poor households mostly end up in secondary vocational education facilities, while non-poor children tend to end up in academic secondary schools. As a result, since the cohort of poor children entering post-basic education is small, the secondary vocational schools tend to have small enrollments and high unit costs, in addition to offering few opportunities for entering the university. 3 Revise the tertiary level free place and scholarship system to explicitly favor students from poor families. 3 Research why a third of the upper secondary school age children that are not currently enrolled in school, are not interested in going to school. 22

23 3.2 Human capital and the poor: health While there has been a slow improvement in most health status indicators over the past decade, increasing health care costs to individuals and declining quality of health care may over time jeopardize this positive trend. Already, use of health services by sick or injured individuals has decreased, dramatically in rural areas. High and growing maternal mortality also points to inadequate pre-natal care and a decreased rate of births attended by health professionals, particularly in rural areas. Health status In Armenia, indicators for life expectancy and mortality, with the exception of maternal mortality, have remained relatively stable throughout the 1990s. Only maternal mortality has experienced a significant increase in the past few years, rising from 20.8 to over 52 per 100,000 live births over the relatively short period between 1996 and Table 3.2.1: Armenia: Selected health status indicators Female life expectancy at birth (years) Male life expectancy at birth (years) Maternal mortality (per 100,000 live births) Infant mortality (per 1,000 live births) Under-5 mortality (per 1,000 live births) Source: UNICEF, Social Monitor 2002, Innocenti Research Center, Florence, Italy, Figures are based on official Armenian sources. Mortality indicators hide significant differences between urban and rural areas, with rural infant mortality almost 50 percent higher, and rural early childhood mortality over 20 percent higher (Table 3.2.2). Mortality indicators need to be treated with some caution, however, as results from survey data (Armenia Demographic and Health Survey 2000) differ from administrative sources (Table 3.2.3). Table 3.2.2: Armenia: Early childhood mortality rates in urban and rural areas in ten-year averages based on ADHS - Neonatal mortality Post-neonatal mortality Infant mortality (0-1 yrs) Child mortality (1-5 yrs) Under five mortality (0-5 yrs) Armenia/Total Urban Rural Source: ADHS 2000, State Department of Statistics et al, Note: Post neonatal mortality is computed as the difference between the infant and the neonatal mortality. Table 3.2.3: Armenia: Official and survey estimates of child mortality (five-year averages for ) Administrative data ADHS, 2000 IMR U5MR Source: SDS et al (2001), MOH-UNICEF (2002). 23