Page 1 1 Report No: 27192-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 Page 2 2 FISCAL YEAR J anuary 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 Page 3 3 T ABLE OF C ONTENTS T ABLE OF C ONTENTS 3 ACKNOWLEDGEMENTS 4 E XECUTIVE S UMMARY 5 1. M AIN F EATURES OF A RMENIA ’ S G ROWTH P ERFORMANCE 1994-2002 9 2. T HE POVERTY PROFILE 12 3. N ON - INCOME DIMENSIONS OF POVERTY 19 4. R URAL D EVELOPMENT AND P OVERTY 31 5. D EMOGRAPHICS , M IGRATION AND L ABOR M ARKET 38 6. P ROTECTING THE P OOR : S OCIAL P ROTECTION P ROGRAMS AND THEIR I MPACT ON P OVERTY 43 B IBLIOGRAPHY 52 Page 4 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. Page 5 5 E XECUTIVE S UMMARY 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 2001. 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 2001. 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 1998. 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. Page 6 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. on- and 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-of- pocket 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 1999. 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. Page 7 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 off- farm 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. Page 8 8 Page 9 9 1 . M AIN F EATURES OF A RMENIA ’ S G ROWTH P ERFORMANCE 1994- 2002 Following the break-up of the Soviet Union, Armenia lost about 60 percent of its GDP between 1989 and 1993. Labor market adjustment was dramatic: unemployment became rampant and wages plummeted; poverty became deep and pervasive. Following macroeconomic stabilization in 1994-95 and sustained Government efforts to advance structural reforms, a steady economic recovery has taken place since 1994. 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, 1994-2002 (percent change over the previous year) 1994 1995 1996 1997 1998 1999 2000 2001 2002 Gross Domestic Product 5.4 6.9 5.9 3.3 7.2 3.3 6.0 9.6 12.9 Of which: Industry 9.6 2.6 1.1 1.3 -2.2 5.2 6.4 3.8 14.2 Agriculture 3.1 4.0 2.0 -4.5 12.9 1.3 -2.3 11.6 4.4 Construction 5.7 -5.8 25.2 3.9 10.6 7.7 28.4 14.5 44.6 Transport and communication -4.1 9.6 17.1 9.2 1.4 0.8 -0.6 16.0 6.0 Trade 60.2 73.0 12.5 5.0 6.7 9.8 8.3 15.5 15.6 Other services -2.2 -2.2 14.6 3.2 2.6 4.5 9.1 5.3 … 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 1995-99 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 1999. 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. Page 10 10 Figure 1.1: Armenia: Employment trends 1990-2001 (in 000) 0 200 400 600 800 1000 1200 1400 1600 1800 1990 1995 1999 2001 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. Page 11 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 Page 12 12 2. T HE 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. 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) Complete Poverty Line (12,276 drams in 98/99) (11,221 drams in 2001) Incidence (P0) Gap (P1) Severity (P2) Incidence (P0) Gap (P1) Severity (P2) Total 98/99 26.8% 6.0% 2.0% 54.8% 16.2% 6.5% (1.94) (0.53) (0.21) (2.26) (0.96) (0.48) 2001 20.0% 4.6% 1.6% 48.3% 13.0% 5.1% (1.53) (0.42) (0.17) (2.42) (0.80) (0.38) Urban 98/99 32.9% 7.8% 2.6% 61.4% 19.3% 8.1% (2.25) (0.66) (0.28) (2.55) (1.13) (0.59) 2001 21.9% 5.0% 1.7% 48.5% 13.7% 5.5% (1.60) (0.36) (0.15) (2.90) (0.82) (0.35) Rural 98/99 18.7% 3.7% 1.2% 46.1% 12.1% 4.5% (2.45) (0.54) (0.21) (3.08) (1.18) (0.54) 2001 17.0% 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, 2001. 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 2002. Annex I.A also describes the major methodological differences between the NSS and that adopted by the World Bank. Page 13 13 Source : ILCS 1998/99 and 2001. 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. Whe n calculated across the poor population (P1/P0) it provides information of the povert y 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 2015. 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 Page 14 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 2005 2015 2005 2015 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 2001. 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, 2000. Page 15 15 Table 2.3: Poverty by gender and age groups in 2001 Extreme poverty incidence 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 0-5 22.2% 52.2% 8.1% 6.8% 7.3% 14.3% 5.9% Aged 6-14 17.3% 45.6% -5.5% 16.7% 15.8% 11.6% 4.3% Aged 15-18 21.1% 49.3% 2.2% 8.3% 8.5% 13.7% 5.5% Aged 19-25 23.7% 51.6% 6.8% 12.1% 12.9% 15.0% 6.2% Aged 26-45 19.3% 47.6% -1.3% 28.6% 28.3% 12.7% 5.0% Aged 46-60 18.9% 44.4% -8.0% 13.3% 12.2% 12.1% 4.8% Aged 61+ 20.4% 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 2001. (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. Table 2.4: Poverty measures by household composition in 2001 Household type Extreme poverty incidence Poverty incidence Relative poverty risk % of population % of the poor 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 2001. 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. Page 16 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 incidence Poverty incidence Relative poverty risk % of Population % 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 2001. (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. Page 17 17 Table 2.6: Armenia: Household income sources in 2001, by regions and consumption (in %) Poorest 2 3 4 5 Average All Households Labor earnings 33.9 27.0 29.3 34.2 39.1 34.0 Self-employment 5.7 8.7 18.8 10.8 16.4 13.4 Farm Income 19.9 25.2 27.1 27.6 21.0 24.0 Remittances 17.3 25.1 13.8 18.6 18.0 18.4 Transfers 22.2 13.7 10.7 8.5 5.1 9.7 Pensions 14.6 9.1 7.6 6.3 4.2 7.0 Other transfers 7.7 4.6 3.1 2.2 0.9 2.8 Assets sold 0.9 0.3 0.4 0.3 0.3 0.4 Total 100.0 100.0 100.0 100.0 100.0 100.0 Yerevan Labor earnings 45.0 47.8 39.0 54.9 60.6 53.9 Self-employment 10.1 17.9 19.4 16.3 14.2 15.3 Farm Income 3.2 3.1 4.4 3.3 2.1 2.8 Remittances 18.1 14.2 24.4 16.8 18.5 18.6 Transfers 22.7 16.6 12.3 8.4 4.1 9.0 Pensions 16.4 12.8 9.3 6.9 3.6 7.0 Other transfers 6.3 3.9 3.0 1.5 0.6 1.9 Assets sold 1.0 0.3 0.5 0.3 0.5 0.5 Total 100.0 100.0 100.0 100.0 100.0 100.0 Other urban Labor earnings 35.5 28.4 37.8 45.1 40.1 38.2 Self-employment 6.1 8.1 20.9 10.5 21.3 14.0 Farm Income 10.6 9.9 14.0 18.3 15.3 14.3 Remittances 20.7 39.2 13.4 15.1 16.7 20.5 Transfers 25.5 14.1 13.6 10.5 6.3 12.6 Pensions 15.7 7.6 9.2 7.3 4.8 8.1 Other transfers 9.8 6.5 4.4 3.2 1.5 4.4 Assets sold 1.5 0.4 0.3 0.5 0.3 0.5 Total 100.0 100.0 100.0 100.0 100.0 100.0 Rural Labor earnings 23.2 14.8 19.0 14.9 14.4 16.2 Self-employment 1.7 4.3 17.3 8.2 16.6 11.7 Farm Income 43.1 49.8 46.9 47.7 45.1 46.6 Remittances 13.2 19.2 8.4 22.1 18.1 16.9 Transfers 18.6 11.8 8.0 7.1 5.5 8.4 Pensions 12.0 8.3 5.7 5.2 4.6 6.1 Other transfers 6.5 3.5 2.4 1.9 0.9 2.3 Assets sold 0.2 0.2 0.4 0.1 0.2 0.2 Total 100.0 100.0 100.0 100.0 100.0 100.0 Source: ILCS 2001. Notes: Quintiles defined as per adult equivalent consumption quintiles. Conclusions Page 18 18 Good economic performance since 1999 resulted in a decreased overall and extreme poverty incidence in Armenia in 2001. 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. Page 19 19 3. N ON - 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) 0.10 0.20 0.13 0.15 0.28 0.16 Basic (grades 1-8) 0.93 0.95 0.93 0.96 0.93 0.94 Upper Secondary 0.52 0.64 0.58 0.67 0.64 0.61 Tertiary 0.22 0.26 0.30 0.43 0.62 0.38 Source : Estimates from EPEAS data 2001. 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 mod ified to account for the fact that the EPEAS was implemented at the second half of the 2000-2001 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). Page 20 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 16-25 years old 0 0.2 0.4 0.6 0.8 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Ye ars of educ ation P r o p . o f s t u d e n t s N on -po or Male N on -po or Fe m ale P oo r Male Poor Fe m a le 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 Page 21 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-of- pocket 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 needs- based 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. Needs- based 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). Page 22 22 ƒ 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; ƒ 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. ƒ 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: ƒ Financing under-funded inputs for basic education that have significant favorable impact on learning, or reduce future costs (such as preventive maintenance); ƒ 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; ƒ 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; ƒ 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. ƒ Revise the tertiary level free place and scholarship system to explicitly favor students from poor families. ƒ 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. Page 23 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 2000. Table 3.2.1: Armenia: Selected health status indicators 1990 1995 1996 1997 1998 1999 2000 Female life expectancy at birth (years) 75.2 75.9 76.2 77.3 78.1 75.5 74.5 Male life expectancy at birth (years) 68.4 68.9 69.3 70.3 70.8 70.7 70.5 Maternal mortality (per 100,000 live births) 40.1 34.7 20.8 38.7 25.4 32.9 52.5 Infant mortality (per 1,000 live births) 18.5 14.2 15.5 15.4 14.7 15.4 15.6 Under-5 mortality (per 1,000 live births) 23.8 19.9 19.5 19.5 18.4 19.2 19.2 Source : UNICEF, Social Monitor 2002, Innocenti Research Center, Florence, Italy, 2002. 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 1991-2000 - 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 26.2 17.9 44.1 4.0 48.0 Urban 23.1 12.8 35.9 1.4 37.3 Rural 29.5 23.3 52.7 6.8 59.2 Source : ADHS 2000, State Department of Statistics et al, 2001. 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 1996-2000) Administrative data ADHS, 2000 IMR 15.0 36.1 U5MR 17.1 39.0 Source : SDS et al (2001), MOH-UNICEF (2002) . Page 24 24 The urban-rural differences may reflect greater difficulty in access to care in rural areas, especially for the poor; and high-risk fertility behavior by rural women: their educational status is often lower, especially in poorer households, they are often quite young or relatively old at child-bearing, they may have inadequate nutrition, and many births are unattended (see Box 1). A recent survey notes that over 70 percent of Armenian women are susceptible to high-risk fertility behavior. Health care utilization in Armenia Utilization of health care has decreased across income groups, and more so in rural areas. Primary care continues to be the most common health care option, but higher income groups, especially in urban areas, are increasingly seeking other care options. These trends may reflect the following factors: the increasing cost of care to individuals; decline in the quality of primary care; and better services in urban hospitals as a result of larger budget allocations to hospitals. Use of health services in Armenia has declined over time. In terms of in-patient admissions and outpatient contacts , the decline is marked (Table 3.2.4): the 2000 levels are only 37 percent of 1990 levels for in-patient admissions and 30 percent for outpatient contacts. Table 3.2.4: Armenia: In and out-patient admissions 1990 1995 1996 1997 1998 1999 2000 Outpatient care: contacts per person per year 6.9 4.2 4.6 3.2 2.4 2.3 2.1 In-patient care: admissions per 100 population 13.9 7.5 7.5 6.7 6.2 5.8 5.1 Source : Health for All. Box 1: Reproductive health and poverty According to the ADHS 2000, a majority (92 percent) of pregnancies that lead to term are subject to antenatal care, but there are significant regional differences: in rural areas, about 11 percent did not receive any antenatal care, compared to 4 percent in urban areas; in some relatively poor areas the share of non-monitored pregnancies may rise to 20 and 30 percent. Geographic differences in antenatal care are confirmed when examining place of delivery and attendance by health professionals. While almost all births in urban areas occur in health care settings, 15 percent of births in rural areas occur at home. No significant difference across economic quintiles is found in urban areas. In rural areas, however, more than 21 percent of children are delivered at home in the poorest quintile, compared to less than 9 percent in the richest quintile. Some 30 percen t of home deliveries were unassisted by health professionals. The relatively high share of unassisted home deliveries suggests that while there may be physical access problems regarding health care in rural areas, ability to pay may also be an important f actor that deters pregnant women (and their families) from seeking medical assistance for delivery. Page 25 25 In terms of sick or injured individuals that sought health care, indicators show a similar trend, declining by 15 percent over the period (Table 3.2.5). Utilization varies across socio-economic groups, increasing with income. It is also higher among urban individuals than among rural ones and the gap has widened: urban use was 20 percent higher in 1998/99 and 28 percent higher in 2001. In terms of socio-economic groups, there is less inequality in use in rural populations than in urban ones, and inequality in use has decreased in both cohorts over time: in 1998/99, the highest quintile in urban areas sought health care 1.8 times more than the lowest quintile, while the highest quintile in rural areas did so 2.7 times more; in 2001, the same numbers were 1.6 for urban populations and a dramatically reduced 1.2 for rural populations. Differences in health care utilization are also reflected in the type of health care providers sought (Table 3.2.6). Primary care is the most common health care option in Armenia, accommodating more than half of those seeking care. Most individuals in need of care in the poorest quintile turned to primary care facilities (64 percent), because they are less likely to be charged there. Better-off households tend to seek other (better) facilities. Recourse to private practitioners, while modest, has increased in the higher quintiles and, interestingly, in the poorest quintile. Table 3.2.5: Armenia: Incidence of illness/injury and health care utilization by consumption quintiles 1998/1999 2001 1 2 3 4 5 Total 1 2 3 4 5 Total A. Percent of individuals reporting sickness/injury Rural 19.2 17.8 15.1 14.8 20.0 17.3 11.0 10.3 10.1 11.9 16.2 11.8 Urban 17.7 16.4 15.0 18.2 24.3 18.1 17.1 14.4 12.5 13.0 16.6 14.8 Total 18.1 17.0 15.0 16.4 22.0 17.7 15.0 12.6 11.5 12.5 16.4 13.6 B. Percent of sick and injured seeking health care Rural 16.5 19.6 29.6 32.9 44.3 30.7 22.8 16.5 23.5 37.2 28.2 26.1 Urban 27.8 35.8 34.3 36.2 49.6 36.7 21.9 27.1 37.0 33.6 36.0 30.5 Total 24.5 29.0 32.3 34.7 47.0 34.2 22.1 23.3 31.8 35.0 33.0 29.0 Source : ISLC 1998/99, 2001. Table 3.2.6: Armenia: Choice of providers by consumption quintiles (percent of those seeking care) 1 2 3 4 5 Total 98/99 2001 98/99 2001 98/99 2001 98/99 2001 98/99 2001 98/99 2001 PHC 61.5 60.9 41.9 55.0 56.9 64.3 47.4 43.8 42.7 45.8 47.4 51.5 Hospitals 34.6 34.8 39.5 40.0 25.9 28.6 37.2 41.7 44.8 45.8 38.5 39.5 Rural private/other 3.8 4.3 18.6 5.0 17.2 7.1 15.4 14.6 12.6 8.3 14.1 9.0 PHC 65.1 63.7 63.6 60.5 69.2 66.0 66.7 55.2 56.9 44.4 63.3 56.8 Hospitals 30.3 29.7 18.7 33.3 20.9 22.7 28.1 28.1 36.2 43.7 28.1 32.4 Urban private/other 4.6 6.6 17.8 6.2 9.9 11.3 5.2 16.7 6.9 11.9 8.7 10.8 PHC 64.4 63.0 57.3 59.1 64.4 65.5 58.0 50.5 50.5 44.9 57.3 55.1 Hospitals 31.1 31.0 24.7 35.0 22.8 24.4 32.2 33.6 40.1 44.4 32.0 34.6 Total private/other 4.4 6.0 18.0 5.9 12.8 10.1 9.8 15.8 9.5 10.7 10.7 10.2 Source : ISLC 2001 Note : Figures correspond to percentages of individuals self-reporting illness/injury and seeking healthcare. Page 26 26 Health care financing While budget resources for health remain low, they have been increasing over time. Public policy favors primary care, and targeted free services have been introduced. Still, budgets are regressive, reflecting the emphasis on hospital financing and preference for hospital use by higher income groups; and out-of-pocket payments (OOP) proliferate. Off-budget donor financing is almost at the same level as budget financing, it provides humanitarian aid and finances donor preferred health programs Health care financing in Armenia draws on budget and off-budget financing, the latter essentially external donor support for various health programs, as well as individual OOP. Budget financing is low compared to other countries. Some 1.6 percent of GDP (2001) is publicly budgeted for the health sector, compared with average health care expenditures for the CIS of close to 3 percent of GDP, and 6 to 10 percent of GDP in most developed countries. While budget resources have been increasing in recent years in line with GDP growth, they barely exceed off-budget donor expenditures. Some 64 percent of the budget is allocated to hospitals. Table 3.2.7: Armenia: Out-of-pockets payments for health care by consumption quintiles and location 1998/1999 2001 1 2 3 4 5 1 2 3 4 5 % paying, out of those sick & seeking health care 65.4 69.8 63.8 78.2 77.6 56.5 70.0 82.1 52.1 79.2 Rural average payment for those paying 2324 3147 3007 7259 32393 2346 3039 5130 5580 37047 % paying, out of those sick & seeking health care 33.9 43.0 53.8 72.9 77.0 42.9 48.1 63.9 72.9 73.3 Urban average payment for those paying 2157 2958 2652 5081 25490 1997 3777 5471 7151 19693 % paying, out of those sick & seeking health care 40.0 50.7 57.7 75.3 77.3 46.4 53.7 69.1 64.4 75.3 Total average payment for those paying (in AMD) 2209 3032 2805 6095 28618 2107 3533 5356 6633 25695 Source : ILCS 1998/99 and 2001. Public spending on health is regressive. Due to better access to health care and higher utilization rates, the top quintiles have captured the bulk of public resources for health: almost 32 percent, while only 15 percent went to the bottom quintile. In part, this reflects the greater use of hospital services by upper quintiles and greater use of primary care units by the poorest quintile. With limited budget funding, private expenditures are significant —some 64 percent of ILCS participants in 2001 report making payments for health care, the same as in 1998/99. OOP increase according to income quintile—some 46 percent of persons in the lowest quintile seeking care pay, while 75 percent of such persons in the highest quintile do so. This is to be expected. However, the share of OOP has risen in the three lowest quintiles since 1998/99, mainly in urban areas (Table 3.2.7). Page 27 27 Conclusions and recommendations Health care should play a key role in a poverty reduction strategy. While health indicators have been improving, access to health services appears to be deteriorating, and health outcomes are relatively worse in rural areas and among the poor. This is reflected in worse health status and lower health service utilization rates in rural areas. Outcomes appear to be shaped by the interaction of poverty and a health system that is increasingly dependent on out-of-pocket payments. The limited public budgets appear to be regressive, favoring expensive hospital care, while the majority of the population is still seeking care in underfunded primary facilities. In this context, improving health outcomes represents a serious challenge for Armenia. ƒ Inequalities in health care provision need to be reduced; in particular the primary service network needs to be strengthened in order to deliver adequate basic health care that can have an impact on key health care indicators —maternal and child mortality rates and contain the incidence of infectious diseases, especially tuberculosis and HIV (although HIV and AIDS still appear to be at relatively low levels in Armenia). ƒ Access for the poor to health care, and especially to primary care, needs to be ensured. This may involve further developing the basic benefit package and ensuring that the services are widely available, and at no cost to poor families. ƒ Out-of-pocket payments for health care must be gotten under control. They already are having a tangible effect on the utilization of health services, and for poor households they may constitute the main obstacle to access to health care. If current trends continue, health care in Armenia will increasingly become a privilege for the rich. ƒ Health care resources need to be better managed: 1. While it may well be that current allocations to health care in the budget cannot be increased more rapidly than so far has been the case, their allocation needs to be more carefully examined, so that they better support a primary health care strategy that aims at a healthier population. This includes reviewing hospital services, which absorb the bulk of budget resources and are often, it seems, services that may be better rendered at lower levels of the health care system. Increasingly, attention should be paid to providing appropriate care at the appropriate level of the system; 2. Off-budget health care financing, which currently is provided in an ad hoc fashion, needs to be better coordinated with the Government’s health care strategies, and the health authorities should recognize this as one of their major functions: off-budget financing should seamlessly complement the Government’s strategy; 3. Private payments should not be discouraged, but should be limited to those who can afford them. Hence, (i) strategies should be developed to reduce informal payments, and (ii) public facilities should be sufficiently budgeted to gradually reduce and eliminate the need for patients to provide, in an ad hoc manner, medicines, medical materials, food, bed linen and other services in public hospitals; instead there should be established tariffs and means of subsidizing poor patients. Page 28 28 ƒ To support the kind of strategies mentioned above, the health system requires appropriate tools to measure and monitor indicators over time, including the impact of the system on the poor. Here, the PRSP effort might extend its monitoring role and include health indicators by socioeconomic groups or regions. Indicators can be divided into input, process, output and outcomes. In other government organizations input and output indicators are administratively recorded. In Armenia, the quasi-private nature of health care facilities and their legal status do not allow a precise monitoring of health inputs. Establishing appropriate regulations to monitor health inputs among quasi-private organizations is necessary in order to describe the inputs covering poorer populations vis- à-vis richer ones. ƒ Measurement of outcomes is crucial to assess the impact of any major policy or program in the country. To begin, the measurement of mortality indicators requires actions suggested by MOH-UNICEF (2002): (i) full adoption of WHO definitions and unification of internal procedures related to calculation of IMR; (ii) training and monitoring of health personnel on the use of classification and registration protocols for stillbirths and infant deaths; and, (iii) changes in the regulations that govern the registration of infant deaths so make the infant death registration process easier and convenient. 3.3 Living conditions of the poor The Armenian poor face congested housing and inadequate public services (water, waste disposal, heating), which contributes to worsening of their living conditions. Housing: The majority of the urban population resides in apartment buildings (70 percent in Yerevan and 56 percent in other urban areas); in contrast, 83 percent of the rural population lives in houses. In urban areas, crowding is a problem, with 34 percent of households in Yerevan and 24 percent of households in other cities living in crowded circumstances (defined as more than two persons per bedroom). It mainly affects poor households. In rural areas, crowding is less of a problem with 17 percent of households living in crowded circumstances. Crowding often is accompanied by low quality of housing. Self-reporting indicates that over 38 percent of individuals in the poorest quintile live in “bad” housing. Water supply: Urban households receive water from central systems. While deteriorating networks have reduced coverage, it remains high, ranging from 99 percent in Yerevan to 90 percent in other cities. The poor have less operating connections, and they are more dependent on spring water and water vendors. In rural areas, only some 64 percent of households have access to central water systems, and some 50 percent of the poor. In most cases, the operating connection consists of a tap in the yard. The remainder is dependent on spring water and water vendors. While a central system may provide relatively easy access, it does not guarantee continuous supply: on average, a household has access some 7 hours a day, varying from 11 hours in rural areas to 9 hours in Yerevan and 5 hours in other rural areas. Page 29 29 Sewage: Outside Yerevan, where 83 percent of individuals live in dwellings that have indoor plumbing, flushing toilets remain rare, particularly in rural areas: in other cities, more than one third of households have outside toilets and in rural areas 75 percent. Garbage: There is limited garbage collection, garbage chutes in most apartment buildings that have those (40 percent in Yerevan and less than 15 percent in other cities) are clogged, and households dump their garbage outside. In rural areas, garbage is mainly dumped, burned or buried. Heating: In urban areas, the central heating system has collapsed, and in 2001 it was available only to 14 percent of households in Yerevan. Some 76 percent use other sources of heating, and 9 percent had no heating at all. Two thirds of the households nationwide reported using wood for heating; almost all rural household use wood. Electricity is mainly a choice for the rich, although 25 percent in the poorest quintile also use electricity for heating. 6 Wood is the heating source of choice among the poor. The burden of paying for heating is relatively higher for the poor than for the rich. Even though not many of the poor use electricity for heating, electricity bills represent sizeable amounts that cannot be afforded easily and payment is often postponed, leading to the electricity supply cuts (Table 3.3.1). 7 Some 66 percent of the poorest urban and 45 percent of the poorest rural quintile experienced electricity supply cuts. The incidence of service cuts is higher during the winter period, when electricity is most needed and when, obviously, bills are large. Table 3.3.6: Electricity service cutoff (% of individuals) Consumption-ranked quintiles q1 q2 q3 q4 q5 Total Yerevan Yes, for a few days 50.3 38.6 35.9 37.5 25.7 36.8 Yes, for one month 8.7 7.6 5.0 3.9 2.2 5.3 Yes, more than one month 8.3 1.5 2.3 2.7 0.6 3.0 Was not cut 32.8 52.3 56.8 55.9 71.6 55.0 Other urban areas Yes, for a few days 46.2 38.5 41.6 30.2 29.0 37.8 Yes, for one month 9.8 3.0 3.3 4.5 1.2 4.7 Yes, more than one month 12.1 6.3 2.7 0.8 0.2 4.9 Was not cut 31.9 52.2 52.4 64.5 69.5 52.6 Rural areas Yes, for a few days 28.5 25.2 27.2 21.1 14.2 23.2 Yes, for one month 7.2 7.7 2.6 1.9 2.7 4.4 Yes, more than one month 9.8 7.3 6.0 3.5 2.3 5.7 Was not cut 54.6 59.9 64.2 73.5 80.9 66.7 Source : ILSC 2001. 6 Anecdotal evidence indicates that using electricity for heating is the last resort option for poor urban families. Often, they use a very simple type of electric burner and only in one room where the entire family gathers. If the temperature is below freezing, the heating is not sufficient to melt the frost from the windows. 7 The estimates from Table 3.3.1 also indicate how efficient the electricity service is in discontinuing supply in case of non-payment. Given that 25 percent of the top quintile in Yerevan reported cut-offs, it seems that the service is not selective (biased) in discontinuing the service. Page 30 30 Conclusions and recommendations In summary, living conditions among the poor reflect double vulnerability —the effects of deteriorating housing and public services; and the constraints that lack of financial resources places on the ability of poor households to seek more affordable solutions, such as installing in- house water tanks, individual heating systems, etc. Under such circumstances, operational policies should emphasize two elements: 1. Public utilities management . Municipalities, which are responsible for public utilities, should focus on improving the utilities management, so that the systems are repaired, appropriately maintained and fully operational. 2. Targeting support to the poor in the payment of utility bills . The support to the poor in the payment of utility bills should be delivered through the existing cash family poverty benefit. To that end, its coverage of the extremely poor could be extended and the benefit amount increased. Page 31 31 4. R URAL D EVELOPMENT AND P OVERTY Although some parts of Armenia have good agricultural potential, most of the country’s land is basically unsuitable for commercial agriculture. The sector expanded over the 1990s, because it absorbed much of the labor released when state enterprises were restructured or closed. Much of agriculture produces only meager returns; more than two thirds of farmers produce at subsistence levels. Landholdings are small and often divided into several plots with substantial distances between each plot. Since the late 1990s, farm incomes have been squeezed by stagnating agricultural prices and rising farm input prices, and by volatile weather conditions. With little opportunity for off-farm work, many rural households are especially vulnerable. Policies to improve well-being in rural areas should focus first on creating jobs outside the farming sector in rural or in urban areas. Other policies should concentrate on high-potential agricultural areas and should aim to increase agricultural productivity and value-added by improving advice and information to farmers, by improving high potential irrigation systems, by reducing farmers’ vulnerability to weather conditions, and, in the long run, by promoting land consolidation. Although some parts of Armenia have good agricultural potential, notably the Ararat valley, the country is not naturally endowed to be a major agricultural producer, and most of the country’s land is basically unsuitable for commercial farming. Nevertheless, the sector expanded greatly over the 1990s, as large numbers of laid-off workers from declining state enterprises settled in rural areas, partly because they had few other options and partly in response to the Government’s land privatization program that was initiated in 1992. As a result, the share of agriculture grew from 20 percent to 31 percent of GDP, and farming now accounts for 43 percent of total employment. However, this growth did not reflect increased agricultural potential, nor improved exploitation of existing potential, but rather a contraction in other sectors of the economy. The uncertainty of agriculture as a basis for steady economic growth is reflected in the volatility of agricultural production: fluctuations in output averaged close to 11 percent from one year to the next over the 1996-2001 period, resulting in an average annual growth rate of 3.5 percent per year, significantly below the average of 6.4 percent for the economy as a whole (Table 4.1). With almost half the population dependent on the rural economy for its livelihood, productivity and incomes are low and economic risk and vulnerability to poverty is high among rural households. Table 4.1: Armenia: Real GDP changes 1996-2001 (in constant 1996 prices; in billions of drams) 1996 1997 1998 1999 2000 2001 GDP : % change 5.9% 3.3% 7.2% 3.3% 9.6% 10% Agriculture : % change 2.0% -4.5% 12.9% 1.3% -2.3% 11.6% Agriculture: share in total GDP 35% 32% 34% 33% 31% 31% Source : IMF. The unfavorable performance of the rural economy is associated with several factors: small, relatively unproductive landholdings, unfavorable farm input and output price developments, volatile weather conditions, and an absence of alternative employment opportunities outside agriculture. Page 32 32 Land holdings The average landholding is 1.3 hectares, and landholdings are small throughout the country. Wealthier households tend to have larger plots, although the average plot size for households in the top quintile is no more than 1.5 hectares. The small size of the land holdings makes it difficult to exploit potential economies of scale and even to use machinery for some crops; instead, more labor per hectare is employed and higher costs per hectare are incurred. Efficiency is further dissipated, as plots of an individual household are usually not contiguous and farmers must spend large amounts of time simply getting from plot to plot. The small size of landholdings also makes profitable larger-scale specialized production difficult, and therefore production tends to be diversified, averaging three crop varieties and two livestock products. About ten percent of rural households report owning some livestock but not owning or renting any land. These households probably use communal grazing land. As these households tend to be among the poorest and as the livestock is an important part of their consumption basket, grazing land is a crucial safety net for them. Terms of trade Agricultural terms of trade have deteriorated. In recent years, and especially since 1998/99, producer prices for agricultural products have been falling relative to other prices, including those of farm inputs. In 2002, agricultural prices were 9.7 percent lower than in 1997, while consumer prices had increased by 13.1 percent (Table 4.2). Table 4.2: Agricultural and economy-wide price indices 1997 1998 1999 2000 2001 2002 Consumer Price Index 100 108.7 109.4 108.5 111.8 113.1 GDP deflator 100 111.1 110.9 109.2 113.6 116.2 Agricultural Price index 100 104.4 90.4 80.5 88.4 90.3 Source : NSS. Note : Agricultural prices are producers (farm-gate) prices. At the same time, retail prices for farm products have increased faster than producer prices for some key crops (Figure 4.1), which indicates widespread problems with getting products to market. The country never had an extensive network of processing and storage facilities in rural areas, and many of those that did exist have now broken down. In addition, the security situation blocks much agricultural production from its closest markets. Producers and middlemen must transport their crops often over long distances to market (World Bank 1999a; Chluba and Schmidt-Kallert 2001). Much anecdotal evidence highlights problems transporting crops: few vehicles are available, gasoline is expensive and bribes frequently necessary. These problems appear to be particularly severe in areas other than the Ararat valley. Page 33 33 Figure 4.1: Difference between producer price and retail price for key crops, 2002, expressed as a percentage of the producer price Source : NSS. Retail price information was not available for wheat and barley. The difference in the margins for the different crops may relate to their relative perishability. Grapes, when ripe, must go to market quickly and this may force farmers to accept lower prices than they might for less perishable crops. Vulnerability to weather Figure 4.2 About 45 percent of Armenia’s 420,000 hectares of cultivable land is irrigated, with the rest depending on rainfall. In many instances, the irrigation system has deteriorated, making water often unavailable in sufficient quantities at the times the farmers need it. Rainfall, therefore, is extremely important to Armenian farmers, and its high variability, as shown in Figures 4.2 and 4.3, makes many villages vulnerable to floods and droughts. Farmers in mountainous areas are particularly vulnerable. During the most recent drought, in 1999-2000, some 15 percent of farm families lost virtually all their crops, and many more suffered significant losses. 0% 50% 100% 150% 200% 250% 300% 1998 1999 2000 2001 2002 potatoes grape milk Dilijan (Tavoush) Monthly Rainfall -90 -70 -50 -30 -10 10 30 50 70 90 110 1 9 9 8 . 0 1 1 9 9 8 . 0 4 1 9 9 8 . 0 7 1 9 9 8 . 1 0 1 9 9 9 . 0 1 1 9 9 9 . 0 4 1 9 9 9 . 0 7 1 9 9 9 . 1 0 2 0 0 0 . 0 1 2 0 0 0 . 0 4 2 0 0 0 . 0 7 2 0 0 0 . 1 0 2 0 0 1 . 0 1 2 0 0 1 . 0 4 2 0 0 1 . 0 7 2 0 0 1 . 1 0 2 0 0 2 . 0 1 2 0 0 2 . 0 4 2 0 0 2 . 0 7 2 0 0 2 . 1 0 m m ( d e v i a t i o n s f r o m n o r m ) Page 34 34 Figure 4.3 Figures 4.2 and 4.3 compare rainfall in a mountainous area (Tavoush) and a flatland region (Ararat) in 2000. Ararat has less precipitation on average (about 20 mm/day) than Tavoush (55 mm/day), but both regions have similar variability over the year. Ararat, however, has higher coverage of irrigation, while crops in Tavoush are mainly rain-fed. In both regions the average monthly rainfall in 2000 was significantly below the historical norm. Irrigation thus becomes an important variable in determining overall agricultural development, and it is positively correlated with household consumption. Off-farm employment While farm income provides the bulk of household income in rural areas, farming is unable to meet the needs of most rural households unless accompanied by other economic activities, i.e. on- and off-farm labor earnings. With rural job opportunities scarce, most rural households end up close to subsistence levels, and less than 30 percent of households that use land and produce net agricultural output receive some farm income. Less than 10 percent of households that use land hire farm labor. Under these circumstances, self-employment has become an important route out of rural poverty, but it appears to be an outlet mainly for better-off households. Poorer rural households are more dependent on labor earnings, and transfers and remittances. Transfers and remittances account for some 31 percent of household income of poor households in rural areas. Simulations using Integrated Living Conditions Survey data emphasize the importance of off- farm activities to household well-being. A high concentration of off-farm activities increases labor market participation with consequent positive effects on household consumption. If a household member is involved in off-farm activities, household consumption per adult equivalent increases by 13 percent. Living in a dynamic area has an even greater impact. If the percentage of households in the community involved in off-farm activities increases by one percent, consumption per adult equivalent in that community increases by 42 percent (for regression description and results, see Armenia Poverty Assessment Volume 2, Annex IV). Rural strategies (i) Land consolidation. Encouraging development of a dynamic land market to increase the average size of landholdings is an important long-term goal, but appears unlikely to succeed in Ararat Monthly Rainfall - 35 -25 - 15 -5 5 15 2 5 35 1 9 9 8 . 0 1 1 9 9 8 . 0 4 1 9 9 8 . 0 7 1 9 9 8 . 1 0 1 9 9 9 . 0 1 1 9 9 9 . 0 4 1 9 9 9 . 0 7 1 9 9 9 . 1 0 2 0 0 0 . 0 1 2 0 0 0 . 0 4 2 0 0 0 . 0 7 2 0 0 0 . 1 0 2 0 0 1 . 0 1 2 0 0 1 . 0 4 2 0 0 1 . 0 7 2 0 0 1 . 1 0 m m ( d e v i a t i o n s f r o m n o r m ) Page 35 35 the short run unless a more dynamic off-farm economy develops that can absorb some of the labor released from farming. Land consolidation to increase the overall size land holdings of rural households that wish to remain in agriculture is an important long-term goal to lift rural households out of poverty. A starting point for a land consolidation policy could be to see how much land households in rural Armenia would need to meet their basic needs. However, poor households vary significantly in their land needs to meet their basic needs, and land consolidation strategies need to differentiate between the relatively modern and productive producers and the more traditional and poorer farmers. Land consolidation would not necessarily be pro-poor. If a land market were stimulated to promote land consolidation, the most likely sellers are non-poor rural households that produce below subsistence levels. These are households that need their land the least. However, if they did sell their land and ceased agricultural production, 83 percent of them would fall into poverty. This would increase the poverty headcount from 48 percent at present to 60 percent and would cause the poverty gap of the rural population to jump from 13 to 73 percent. Before they would choose to sell their land, such households would need to have alternative source of income. Under-employment plus a flat off-farm economy in most rural areas of Armenia make that a long-term proposition. An aggressive policy of land consolidation therefore seems unlikely to succeed in the current economic climate. If it were compulsory, households would almost certainly be pushed into poverty. If, as is more likely, it were voluntary, households would be unlikely to sell until other options were available to them. The large poverty impact of losing farm income shows both how agriculture plays a key role in shielding rural households from poverty, even those that do not appear to derive a large proportion of their livelihoods from agriculture. This suggests that policies to promote land consolidation should be carefully sequenced and ideally come after other measures to stimulate the rural economy, particularly the off-farm sector. The government’s policy of leasing or selling arable and some pasture land that is currently owned by national or local governments is appropriate. Communal grazing land, however, should probably remain in public ownership, as it appears to provide an important safety net for some of the poorest households. The government may wish to consider interventions that would increase the quality and productivity of the pasture. (ii) Irrigation. Irrigation is crucial to the country’s agriculture, since irrigated farming accounts for some 80 percent of agricultural GDP (World Bank 2001b). It is particularly important in the fertile Ararat valley, since farms with irrigation generate about 60 percent more sales per hectare than farms without irrigation (World Bank 1999a). Ninety-eight percent of farms in the ILCS report having irrigation infrastructure serving around one third of their land (though the 1998 Farm Survey indicates the figure to be closer to 60 percent of farmers having irrigation on an average of half of their land). The 1998 survey of family farms reveals that farmers with irrigation generated net returns of about US$370 compared to US$40 for those without irrigation. The 2001 ILCS indicates lower returns, but it also shows that above-subsistence farmers are able to realize net returns over 50 Page 36 36 percent higher than above subsistence farmers without irrigation. The difference for below- subsistence farmers is 8 percent. This difference relates to productivity of land, skills of farmers, quality of irrigation services and other factors. Irrigation infrastructure needs considerable investment. The government strategy, supported by the Bank, is to invest in rehabilitation of irrigation systems in areas of high agricultural potential and encourage farmers to take over responsibility for operating and maintaining the systems. Both the government and the Bank strategies involve increasing the fees that water users must pay. While the investments will entail increasing the costs to the farmer of irrigation, ILCS survey data indicate that farmers are willing to pay more for more reliable service. Other sources indicate that the situation is more complex. A social assessment for a World Bank-financed investment in irrigation has contradictory findings. Qualitative methods indicate that increasing water fees to the amount necessary to cover operation and maintenance costs would lead to charges of around 10 percent of gross income from farming (an average of two thirds of farmers’ net income) and would not be affordable (8 percent of gross farm income is the benchmark for what is reasonable to pay for irrigation water). The Poverty and Social Impact Analysis for Water Sector Reform in Armenia indicates, under the most likely circumstances, doubling the irrigation water price would create critical problems in Tavush, Vayots Dzor and Lori marzes, that fruit cultivation would become unprofitable in five of the country’s ten marzes, and that forage crops would decline with resulting impacts on livestock. They estimate that a sharp increase in the price of irrigation water would increase rural to urban migration, particularly in outlying border areas, and would increase food insecurity in the country. The report recommends that an increase in irrigation fees be coupled with efforts to promote the off- farm economy in rural areas. Detailed calculations for one marz, however, come to quite a different conclusion. There, irrigation water would amount to less than 2 percent of gross revenues and would be affordable. The social assessment indicates that there is a great difference between increasing fees in rehabilitated and non-rehabilitated systems since system losses and thus water used per farmer are far lower in former. Clearly, water fee increases should come only with increase in service and rehabilitation of infrastructure. Policy considerations Reducing rural poverty will require increasing rural households’ access to off-farm employment and improving the productivity of agriculture. Public policies should seek to reduce the dependence on agriculture, particularly for those households that have neither skills nor interest in farming. Thus promoting off-farm or even urban employment will be key. At present, however, off-farm employment opportunities are scarce in rural areas and rural households are typically not able to access those urban jobs that are emerging. Rural well-being is also reduced by the absence of adequate public infrastructure and services. Within agriculture, farm productivity is low and excessively subject to the vagaries of weather; farms are often too small to allow even efficient farmers to meet their needs through agriculture and agricultural markets appear inefficient. Specific measures to reduce rural poverty include: ƒ Increasing opportunities to generate off-farm income would provide options for households that do not wish to remain in agriculture to leave the sector, and allow the Page 37 37 average land holdings for those who wish to remain to increase. Some parts of the country have potential for developing rural tourism industries and increasing opportunities for trade. In other areas, off-farm activities will grow through goods and services associated with agriculture, such as marketing of inputs, advisory services, intermediation with markets, establishing storage and processing facilities. The government’s role is principally in activities such as training, education, promotion and measures to help potential entrepreneurs establish businesses. Achieving such goals is not easy. Simple models that successfully achieve such objectives have proved elusive. However, carefully tailored interventions can make a difference, and investigating appropriate measures for rural Armenia is therefore a high priority. ƒ Policies that facilitate storage, processing and transport in rural areas would relieve important bottlenecks to generating more income from agricultural production and could have major short-term benefits in terms of poverty reduction and stimulating the rural economy. ƒ Policies to increase access to agricultural information and technologies are important to increase agricultural productivity. Current policies to improve extension services appear to be a good start in encouraging use of inputs and equipment, and increasing their coverage out of marz capitals is likely to have a positive impact on agricultural production and on rural poverty. ƒ Land consolidation to increase the overall size land holdings of rural households is an important way to lift agricultural households out of poverty in the long term. But it needs to be pursued with caution, and it needs to go hand-in-hand with the development of a more dynamic off-farm economy. If households do not have opportunities outside agriculture, they will either not sell their land—thwarting the objective of allowing efficient farmers to build up larger land holdings—or risk being pushed into poverty. Policies to promote land consolidation therefore may work best if carefully sequenced with other income-generating opportunities. Continuing to lease or sell state-owned arable or pasture land does appear to be an appropriate policy. The sale of grazing land, however, should be approached with caution. Communal grazing land provides an important safety-net for the poor and should remain in public hands, ideally taking some measures to make the grazing land more productive. ƒ Where land has high agricultural potential, making irrigation services more reliable will be vital for reducing rural poverty. The Government’s policy to rehabilitate irrigation schemes accompanied by an increase in the price users pay for irrigation water appears appropriate, as long as fee increases are sequenced together with improvements in service quality. Farmers appear to be willing and able to pay for more reliable irrigation water. In order for the farmer to generate sufficient revenue from farming, the investments and associated fee increases should take place in areas of highest crop potential. Farmers in areas less suitable to irrigated crop production should be encouraged to specialize in livestock or other products. Page 38 38 5. D EMOGRAPHICS , M IGRATION AND L ABOR M ARKET While economic growth has been vigorous since the economic turnaround in 1994, the employment content of growth has been modest. This is reflected in low labor force participation and increasing unemployment, especially in rural areas, where structural deficiencies have made households particularly vulnerable to external shocks. Urban areas have fared somewhat better, particularly Yerevan. One outcome of a demand deficient labor market has been migration: there are significant labor movements both inside the country and abroad. 5.1 Demographics and migration trends The Armenian population has declined as people have sought work abroad, essentially in Russia, and as the fertility rate has dropped to a low 1.1 child per woman. Internal migration is also significant, as people seek better job opportunities. Internal migration is mainly a risk management strategy of poor households, while migration abroad is more frequently used by members of more affluent households. The population of Armenia has declined throughout the 1990s as a result of migration abroad and falling fertility rates. The latest population census, conducted in October 2001, found that the population was 3.2 million. This differs from official population projections based on the 1989 census, which forecast a population of more than 3.8 million people in 2001. The difference reflects high levels of migration abroad, especially in the early 1990s. In addition, Armenia experienced a decline in fertility to 1.1 children per women (UNICEF, 2002), reflecting worsening economic conditions, as well as out-migration of child-bearing age adults. Migration —within Armenia as well as abroad—has been a key coping strategy of households when faced with poverty and in the absence of labor market opportunities. Evidence from ILCS 2001 shows how prevalent migration has become for Armenians: 22 percent of the population lives in households with permanently absent members. About 40 percent of migrants in 2001 moved within the country, a significant proportion, 34 percent, from the poorest households. Migrants were typically young individuals between 18 and 20 years of age. Most of the migration seems to be into- and out of neighboring (non- Yerevan) urban areas. There is less rural-urban migration, and such migration is usually to close-by cities other than Yerevan. Over half of migrants from Yerevan go to other urban areas. 5.2 Labor market developments Despite significant economic growth since 1994, employment has been slow to respond to growth and even shows a downward trend during the 1994-2001 period (Figure 5.2.1). With a low employment content of GDP, unemployment has remained high. Page 39 39 Figure 5.2.1: Armenia: Real GDP and employment 1990-2001 (1990=100) 0 20 40 60 80 100 120 1990 1995 1999 2001 GDP Total employment Employment in industry & constr. Source : Statistical Yearbooks of Armenia. Major structural changes in registered employment occurred in the mid 1990s, with a sharp rise in employment in agriculture. Labor shedding from the state sector was underway, and rural land privatization allowed the rural sector to absorb laid-off workers, albeit inadequately: in most cases, plots given to workers were insufficient to maintain families even at subsistence, and often the workers had little farming experience. Altogether, some 50 percent of laid-off workers were absorbed into agriculture. As labor shedding continued through the latter part of the 1990s, an already saturated agricultural sector could no longer absorb the surplus labor that was being released, and the new private sector was not able to expand fast enough to compensate for the job losses in the state sector. The transition resulted in significant un- and underemployment. In the period af ter 1999, positive changes have begun to appear, as the Government’s efforts to expand the investment climate and broaden the base for economic growth have begun to bite. The employment content of GDP has increased as a result of more rapid growth in labor intensive sectors, especially construction and trade, which have created more employment opportunities. The improved situation has not, however, extended to the rural population, and rural unemployment has almost doubled since 1999. Official unemployment, based on the number of unemployed registered by the National Employment Service, is relatively low. However, Labor Force Survey (LFS) data for 2001 indicate that registered unemployed accounted only for 18 percent of the total number of unemployed, reflecting the absence of incentives for registration. Hence, the unemployment rate using the LFS data for 2001, according to the ILO definition, is almost three times higher than the official unemployment rate. Page 40 40 Data on registered employment also differs depending on the data source. NSS data set the total number of registered employed persons at 1.27 million in 2001 (Table 5.2.1). The State Social Insurance Fund (SSIF) reported that the total number of registered employed in early 2003, excluding farmers, was 0.47 million. When those employed in agriculture according to official data are included, the SSIF number would rise to one million. Table 5.2.1 Armenia: Labor resources, official data (in thousand people or percentages) 1995 1996 1997 1998 1999 2000 2001 Labor resources, total 2,133 2,154 2,173 2,251 2,288 2,357 2,357 Economically active population 1,582 1,584 1,538 1,476 1,462 1,447 1,427 % of labor resources 74.2% 73.5% 70.8% 65.6% 63.9% 61.4% 60.5% Growth rate 0.1% -2.9% -4.0% -0.9% -1.0% -1.4% Employed 1,476 1,436 1,373 1,335 1,298 1,278 1,265 - industry 303 255 229 209 195 183 170 - agriculture and forestry 552 586 567 568 562 556 570 - transport and communication 53 48 49 51 48 48 44 - construction 76 68 60 57 54 53 41 - trade and public catering 100 110 116 113 109 108 111 - other services 392 368 351 340 330 336 329 Private sector 721 860 846 916 931 931 942 Non-private sector 755 576 527 421 367 347 323 Employment rate (% of labor resources) 69.2% 66.7% 63.2% 59.3% 56.7% 54.2% 53.7% Registered unemployed 106 148 166 139 164 170 147 Unemployment rate (% of active population) 6.7% 9.3% 10.8% 9.4% 11.2% 11.7% 10.3% Source : Statistical Yearbook of Armenia 2001; Social and Economic State of RA for January-December of 2001. Survey based information on labor market in Armenia. While overall participation rates appear stable, regional shifts occurred between 1999 and 2001: participation rates increased in urban areas, particularly in Yerevan, and decreased in rural areas. The increase in urban areas is primarily the result of increased participation of the young (16 to 25) and those over 50 years of age. The drop in participation rates in rural areas by 14 percent is mainly explained by a decrease in the participation of people 26 to 50. They may have found work in agriculture unattractive or faced difficulties in finding employment in either farm or non-farm sectors due to limited job creation in these sectors. There are significant gender differences in participation rates, with that of men exceeded that of women by over a third in 1999 and by over half in 2001. The unemployment rate increased considerably between 1999 and 2001. With an unchanged participation rate, this reflects a reduction in employment. In 2001, about 3 out of 10 active individuals were unemployed. Young people, mostly new labor market entrants, experienced the highest unemployment rates. Reflecting trends in participation rates, the unemployment rate dropped in urban areas, particularly in Yerevan, and increased in rural areas. As the unemployed face the highest risk of poverty, a large increase in the number of unemployed can be expected to have a strong impact on rural poverty. Still, about 76 percent of the unemployed live in urban areas, 42 percent in Yerevan. There do not appear to be significant gender differences, although more males appear to be unemployed in rural areas. A relatively large share of highly educated persons is unemployed, particularly in Yerevan. There was also relatively high proportion of young among the unemployed, again in rural areas. Almost half of the unemployed are new labor market entrants, indicating that new entrants exceed the absorptive capacity of the Armenian labor market. Page 41 41 Unemployment in Armenia is predominantly long-term in nature, and over half of the unemployed have been looking for a job for over a year. Long-term unemployment was highest in urban areas outside Yerevan, where almost two thirds of the unemployed have been seeking a job for over a year. The structure of the long-term unemployed by gender and education is very similar to the structure of all unemployed. The differences only appear by age groups with less youth and more middle age individuals being long-term unemployed compared to the overall number of unemployed. Structural aspects of unemployment may be indicated by the relatively high number of individuals with higher education that are unemployed, for instance they constitute 31 percent of all the unemployed in Yerevan. This could reflect that there are quite a number of people whose academic qualifications make them unsuitable for the needs of the current economy. Poor households in urban areas appear to have benefited more than richer households from the positive changes in the labor market that resulted in a considerable decline in urban poverty over the 1999-2001 period, as the increase in urban participation rates was more pronounced for the poor than for the rich. At the same time, the reduction in participation rates in rural areas was more pronounced for poor than for the rich households. The Armenian labor market is characterized by a great deal of informal activity, driven by the relative absence of remunerative formal sector jobs. According to the ILCS 2001 estimates, informal employment made up 40 percent of overall employment in Armenia, with the self- employed accounting for over 60 percent of informal employment. Subsistence agriculture accounts for around one fourth of total informal employment. This result confirms recent empirical work (World Bank, 2003), which estimated that the Armenian informal sector accounts for some 45 percent of all jobs in the country, one of the highest informal employment rates in the CEE and the CIS. Engagement in informal activity appears equal at the bottom and at the top end of consumption distribution; almost half of the informal workers are highly educated. A similar pattern is also observed for informal workers in Georgia (Bernabe, 2002), while in most transitional countries in Eastern Europe the majority of informal workers are poorly educated. The large share of highly educated informal workers reflects the overall high educational level of the Armenian population, but it is also driven by scarcity of jobs for highly educated individuals in formal sector. Conclusions and recommendations Armenia has made significant progress in implementing structural reforms and needs to continue policies that foster inclusive economic growth. This includes completing the privatization process and enterprise restructuring, and injecting more flexibility into the labor market to allow it to better adjust to changing economic conditions and external shocks. The regulatory framework needs simplification to ensure that the benefits to enterprises of formalization outweigh the costs of working within the rules of the formal economy. Related bureaucratic regulations that impede small-scale enterprise development should be minimized. Policies that create a macroeconomic environment conducive to the development of the new formal private Page 42 42 sector will also influence integration of the informal into the formal economy. The emphasis should be on small-scale business outside agriculture. A large informal sector in Armenia has several important policy implications. First, it causes loss of public revenues. Second, it leads to distortions in the main economic indicators such as GDP, employment, unemployment, etc., with implications for policy making. Third, as informal employment has become an important coping strategy for poor households, income targeting for the purposes of social assistance becomes more difficult. Fourth, large numbers of educated workers become engaged in low-skill informal activities that may erode skills and undermine the capacity of such workers to adjust to the demands of a new market-based economy. This may in turn lead to marginalization and exclusion of certain groups of workers. Given the scale of informal sector activity in Armenia, it is important to emphasize the importance of reliable data sets that can facilitate a rigorous analysis of informal labor market activity. A major limitation of the analysis of informal activity in Armenia is the reliance on data sets which were not initially designed for this purpose. Therefore, surveys customized to elicit more and better information on both formal and informal economic activity would be beneficial. Page 43 43 6. P ROTECTING THE P OOR : S OCIAL P ROTECTION P ROGRAMS AND THEIR I MPACT ON P OVERTY With high poverty rates, a well-targeted and effective public social protection system becomes central to family welfare among the poor. The current social protection system in Armenia has three main components: ƒ social insurance providing old- age, disability and survivor’s pensions, as well as sickness, maternity, and unemployment benefits (comprising an unemployment compensation benefit and various training programs). The benefits are funded through mandatory social insurance (payroll) contributions and administered by State Social Insurance Fund; ƒ various non-contributory, state budget financed social assistance benefits that include the family poverty benefit; a one-time cash payment to poor families; a social pension for individuals who do not qualify for a labor pension; a one-time cash payment for all newborns; a child care allowance for employed mothers until their child reaches age two; and public works schemes (including “food for work” type of programs); 8 and ƒ social care services to certain categories of the population, including medico-social rehabilitation for veterans and disabled persons, services at home for frail elderly, and institutional care for vulnerable individuals, such as children deprived of birth parental care, children from destitute and/or dysfunctional families, and children and adults with disabilities. The labor pension and the family poverty benefit are the main cash benefit programs in terms of coverage, resources and poverty impact. Table 6.1 provides summary administrative data on all the social protection cash benefit programs in Armenia. Social protection cash benefit programs in Armenia have broad coverage: in 2002, more than 1.2 million Armenians, or 40 percent of the population, received at least one benefit. Coverage has declined over time: from 53 percent of the population in 1999 to 40 percent in 2002, reflecting improvements in living standards and better targeting of social assistance, but also a decrease in resources allocated to and eligibility for the family poverty benefit. Social protection programs are important as a source of family income, particularly for low income families. According to the 2001 ILCS, they made up 22 percent and 14 percent of the income of the two bottom quintiles, respectively. 8 Social assistance programs also include price-discounts for some groups of the population. Previously common and widely spread, they were substantially reduced in 1999 and currently consist of a very few privileges mostly for veterans of WW II. Page 44 44 T able 6.1: Armenia: cash social protection programs 1999-2002 (administrative statistics) 1999 2000 2001 2002 Recipients (in 000) 1630.9 1384.7 1306.6 1213.7 Pensions 1 568.2 560.1 551.0 543.6 Other social insurance benefits 2 89.7 76.8 68.0 74.1 Family poverty benefit 3 657.1 667.9 598.6 532.0 Other social assistance benefits 4 315.9 79.9 89.0 64.0 Resources (in bill of constant drams, 1999=100) 55.28 53.87 56.22 58.13 Pensions 32.17 34.38 38.27 41.84 Labor 26.47 26.45 29.13 31.89 Social 1.93 1.92 1.97 1.91 Military 5 3.77 6.00 7.17 8.03 Other social insurance benefits 1.87 1.20 1.05 1.58 Family poverty benefit 6 21.10 17.86 16.48 14.36 Other social assistance benefits 7 0.14 0.44 0.43 0.35 Resources by programs in % (Total=100) 100.0 100.0 100.0 100.0 Pensions 58.2 63.8 68.1 72.0 Labor 47.9 49.1 51.8 54.9 Social 3.5 3.6 3.5 3.3 Military 5 6.8 11.1 12.8 13.8 Other social insurance benefits 3.4 2.2 1.9 2.7 Family poverty benefit 6 38.1 33.2 29.3 24.7 Other social assistance benefits 7 0.2 0.8 0.8 0.6 Share in GDP (in %) 5.60 5.18 4.89 4.43 Pensions 3.26 3.31 3.33 3.19 Labor 2.68 2.54 2.53 2.43 Social 0.2 0.18 0.17 0.15 Military 5 0.38 0.58 0.62 0.61 Other social insurance benefits 0.19 0.12 0.09 0.12 Family poverty benefit 6 2.14 1.72 1.43 1.09 Other social assistance benefits 7 0.01 0.04 0.03 0.03 Source : Administrative data from the State Social Insurance Fund (SSIF) and Ministry of Social Security (MSS). Notes : 1/ Includes labor, social and military pensions. Retired Army officers not included (data not available). 2/ Includes unemployment benefit, maternity leave compensation (under assumption that it lasted on average 140 days) and sick-leave compensation (under assumption that average sick-leave was 10 days). 3/ Includes only regular monthly family poverty benefit. 4/ Includes one-time social poverty assistance payment, child care allowance and newborn allowance. 5/ Includes pensions of soldiers and officers. 6/ Includes resources for the one time poverty assistance. 7/ Includes child care allowance and newborn allowance. Resources allocated to social protection have not kept pace with GDP growth. As a result, their share of GDP has declined by some 20 percent over the 1999-2002 period, falling from 5.6 percent in 1999 to 4.4 percent in 2002. The family poverty benefit has experienced the sharpest drop, falling 32 percent in real terms; its share of GDP dropped from 2.1 percent in 1999 to 1.1 percent in 2002. There has been a corresponding drop of some 20 percent in the number of beneficiaries. The budget for 2003 indicates a further reduction in the share of the family poverty benefit in GDP —to 0.9 percent. This policy is driven by fiscal constraints and assumes efficiency gains from improved targeting and better administration, as well as reduced poverty incidence from economic growth. Still, in light of persistently high levels of poverty and the importance of the benefit for the poor (see below), this trend is worrisome. While spending over 2 percent of GDP on a single family poverty benefit may be unreasonable in a low-income country with limited budget resources, spending too little may have long term negative consequences for legitimately poor families who are denied access to the benefit, such as chronic poverty, social exclusion, anti-social behavior, etc. As far as other transfers are concerned, Page 45 45 spending on military pensions has increased substantially —2.1 times in real terms—while the number of beneficiaries has remained almost unchanged, reflecting the priority given to this particular group. Spending on labor pensions has increased in real terms as well, but at a much lower rate—10 percent between 1999 and 2002. Poverty impact of social programs The poverty impact analysis focuses on the two major benefit groups—pensions and social assistance, with social assistance comprising all non-pension social transfers. The impact is analyzed comparing with and without transfer situations. It includes a sensitivity analysis for household ability to compensate for loss of social transfers at the 0, 25, 50 and 75 percent level. The analysis indicates that if social transfers were eliminated, incidence of poverty and extreme poverty would increase by 6.0 and 7.4 percentage points, respectively. Table 6.2 presents estimates of what would happen if households were able to compensate 0, 25, 50, and 75 percent of eliminated social transfers. Table 6.2: Armenia: Poverty reduction impact of social transfers Mean Standard Error 95 percent Confidence Interval Change in poverty incidence (percentage points) Poverty incidence Post-transfers rate 48.3 2.4 43.4 53.2 0.0 Pre-transfers rates depending on the percent of transfers substituted by other households resources 75 percent 51.2 2.5 46.2 56.2 +2.9 50 percent 52.6 2.5 47.6 57.6 +4.3 25 percent 53.4 2.4 48.5 58.3 +5.1 0 percent 54.3 2.4 49.5 59.1 +6.0 Extreme Poverty Incidence Post-transfers rate 19.9 1.5 16.9 23.0 0.0 Pre-transfers rates depending on the percent of transfers substituted by other households resources 75 percent 22.6 1.6 19.4 25.8 +2.7 50 percent 24.3 1.7 20.9 27.8 +4.4 25 percent 26.1 1.7 22.7 29.5 +6.2 0 percent* 27.4 1.8 23.8 31.0 +7.4 Source : ILCS 2001 Note : * pre-transfer poverty rate significantly higher than post-transfer poverty rate. The elimination of social transfers would not only push non-poor into poverty, but it would also increase depth and severity of poverty. Table 6.3 shows how depth and severity of poverty would change if social transfers were eliminated and households were able to compensate the loss by 0, 25, 50 and 75 percent using other resources. In all the cases, except when households are assumed to be capable of compensating 75 percent of lost social transfers, poverty shortfall and severity of poverty are significantly higher as a result of the elimination of social transfers. In other words, social transfers have a significant poverty alleviating effect on households, even if they are not sufficient to lift them out of poverty. Page 46 46 Table 6.3: Armenia: Poverty alleviation impact of social transfers Mean Standard Error 95 percent Confidence Interval Change Poverty shortfall (P1/P0) Post-transfers shortfall 27.0 0.9 25.3 28.7 0.0 Pre-transfers poverty shortfall (P1/P0) depending on the percent of the transfers substituted by other households resources 75 percent 30.0 0.9 28.3 31.8 +3.0 50 percent* 31.8 0.9 29.9 33.6 +4.7 25 percent* 33.5 1.0 31.5 35.5 +6.4 0 percent * 35.2 1.1 33.1 37.3 +8.2 Severity Post-transfers severity 10.7 0.6 9.5 11.8 0.0 Pre-transfers severity of poverty depending on the percent of the transfers substituted by other households resources 75 percent 13.0 0.6 11.7 14.3 +2.4 50 percent * 14.3 0.7 12.9 15.7 +3.6 25 percent * 15.7 0.8 14.2 17.3 +5.1 0 percent * 17.4 0.9 15.6 19.2 +6.7 Source : ILCS 2001 Note : * pre-transfer poverty shortfall and severity of poverty significantly higher than post-transfer shortfall and severity of poverty. Table 6.4 suggests that social transfers are effective in reaching the pre-transfer poor and extremely poor population, i.e. the coverage of the poor and extremely poor is relatively high, while coverage of the non-poor is relatively low. This appears to be the case even for pensions, which are not poverty targeted. Table 6.4: Armenia: Effectiveness of social transfers —coverage of pre-transfer poor and non-poor population Pensions Social assistance Mean 95 percent Confidence Interval Mean 95 percent Confidence Interval Percent of “pre-transfer” population covered by pensions or social assistance (standard errors in parentheses) Poor* 56.9 (2.8) 51.2 62.6 19.7 (2.4) 14.8 24.5 Extremely Poor* 64.3 (3.3) 57.6 70.9 28.3 (3.0) 22.2 34.4 Non-poor** 38.6 (1.8) 35.2 42.0 6.4 (1.1) 4.1 8.7 Source : ILCS 2001 Note : *coverage of the poor and extremely poor is significantly higher than the coverage of the non-poor. ** coverage of the non-poor by social assistance (family poverty benefit) is significantly lower than in the case of pensions. Social transfers also contribute to to the reduction of inequality in the distribution of consumption by about 9 percent. The pre-social transfers Gini coefficient for consumption distribution is reduced by 6 percent when pensions are added to consumption, and by a further 3 percent when social assistance is added (Table 6.5). Table 6.5: Armenia: “Pre-social transfers” and “post-social transfers” Gini coefficients for consumption Post-social transfers (with social assistance) 0.278 Pre-social assistance (post-pensions) 0.289 Pre-social transfers 0.305 Source : ILCS 2001 Page 47 47 Pensions Pensions are an important source of income for the elderly and their families. About one third of Armenians live in extended families with one or more pension-age elderly; another 11 percent live in households composed only of pension-age elderly. In many of these families, pension income is important: it is estimated that some 6 to 12 percent of such families would be poor and an additional 5 to 14 percent extremely poor, if pensions were not available. Pension payment in a timely manner reduces poverty and extreme poverty by an estimated 1.7 and 1.4 percent, respectively. Still, pensions are low and insufficient to keep most elderly households out of poverty. While at 3.3 percent, their share in GDP is comparable to neighbouring countries. In 2001, the average pension represented 65 percent of the extreme poverty line, 46 percent of the poverty line and 22 percent of the average wage respectively. This reflects a declining trend, as the corresponding percentages in 1999 were 51, 36 and 24, respectively. A 20 percent increase in pensions (0.6 percent of GDP) would reduce the poverty incidence by 3.4 percentage points. In addition, it appears that a significant number of elderly above the statutory retirement age are not receiving pensions. Only 85 percent of them report receiving a pension, which is low for a country where every elderly person is entitled to either a labor pension or a social pension. In part, this reflects the effect of the working elderly, but even when adjusting for those individuals, coverage rises only to 90 percent. Still, given that a social pension is available to anyone age 65 and over; coverage of less than 100 percent is surprising. It may reflect that some elderly may not be aware that they are entitled to a social pension, or do not know how, or bother, to apply for it. Social assistance Among the various programs that comprise social assistance, the overriding one in terms of resources and coverage is the family poverty benefit (FPB) —a regular, monthly, proxy means- tested cash benefit introduced in 1999. The FPB is administered by a nationwide network of territorial centres for social services. Over the years, procedures for benefit award, as well as monitoring and supervision have been improved. The Table 6.6 summarizes administrative data on the FPB. Since the inception of the family poverty benefit, the Government has pursued a policy of decreasing resources allocated to the program, as budget resources have become constrained. Correspondingly, the number of beneficiaries has been reduced, as has the average benefit. The decisions about which beneficiaries to exclude appear to have been driven by the need to adjust the number of beneficiaries to the decreased funding. For instance, although unemployed are among those facing the highest poverty risk, their eligibility weight has been reduced by a quarter: the explanation given by the officials was that the unemployed are capable of working and should look for a job more actively instead of relying on the FPB. Similarly, the increase in pensions led to the exclusion of thousands of pensioners who, by getting additional 700-1,000 drams per month in pensions, lost 3,000 drams per month in family poverty benefit. The average Page 48 48 benefit has been reduced: as of January 1, 2002, only children are paid a supplement to the family base benefit (1,500 drams per child, while the family gets 4,000 drams). Previously, each family member was given a supplement. The poverty and welfare implications of these measures, and consequently their longer term costs to the individuals and families, as well as to the economy as a whole, have not been examined. It is likely, however, that these measures may hide significant social costs. Table 6.6: Armenia: Family poverty benefit 1999-2003 1999 2000 2001 2002 2003 Beneficiaries Regular monthly benefit (as of January 1) Number of families 211555 199456 174800 149603 141218 Number of individuals 657071 667897 598616 532014 505560 One-time benefit Number of families 66980 11797 15917 10140 14889 Number of individuals 289711 44935 54139 30544 39456 Average benefit Regular monthly benefit: drams per household 7193 7196 7712 6554 7099 In % of the average wage 38.8 34.3 32.2 26.0 25.6 Regular monthly benefit: drams per family member 2313 2148 2255 1841 1983 In % of the poverty line 18.8 N/a 20.0 n/a n/a In % of the extreme poverty line 26.5 N/a 28.3 n/a n/a In real terms (constant 1999 prices) 2313 2166 2207 1781 1862 One-time benefit (drams per household) 3500 3500 3500 4000 4000 Resources Total (nominal in bill drams) 21.08 17.72 16.85 14.85 13.23 In real terms (1999=100) 21.08 17.86 16.48 14.36 12.42 In % of GDP 2.14 1.72 1.43 1.09 0.89 Source : Ministry of Social Security of the Republic of Armenia and Armenia State Department of Statistics. Notes : For 2003, beneficiaries as of January 1. Average benefit and resources are planned values. Total resources spent in 2002 comprise 12.252 bill drams in payments for 2002 and 2.598 bill drams for payment of two months of the arrears accumulated in 2000. While the social assistance system is well targeted, it only covers a limited share of the poor, and in many instances it is badly calibrated. There is a strong pro-poor focus with 45 percent of recipients come from the bottom “pre-social assistance” consumption quintile, receiving 51 percent of the social assistance budget (Table 6.7). Likewise, the poorest population receives the highest level of benefit. Where leakage to the non-poor does occur, analysis of the survey data indicates that it is most significant in instances where a household corresponds to the profile of households with the highest pre-social assistance poverty and coverage rates. In such instances, screening may be particularly difficult. However, while leakage is small, only some 20 percent of “pre-social assistance” poor and 28.3 percent of the extremely poor receive benefits. According to the poverty estimates presented in this report, there were in Armenia in 2001 about 700,000 “pre-social assistance” extremely poor and about 800,000 “pre-social assistance” poor people. At best, the social assistance programs would then cover about 650,000-700,000 people. Inefficiencies in calibration imply that families receive more than they need to be lifted out of poverty. Thus, almost 20 percent of recipients of the FPB are not only lifted out of poverty, but actually end up in the four highest consumption deciles. The reduction in the average family poverty benefit that took effect at the beginning of 2002 may have to some extent reduced this effect. Page 49 49 Table 6.7: Armenia: Distribution of social assistance recipients and funds across the “pre-social assistance” consumption quintiles Quintiles Bottom 2 3 4 Top Recipients (%) 45.0 23.7 15.2 9.0 7.1 Resources (%) 51.0 21.3 13.2 8.2 6.3 Source : ILSC 2001. How much is needed to eliminate poverty in Armenia? A perfectly targeted social assistance system would require about 63. 6 billion drams, or 5.3 percent of the GDP to eliminate “pre- social assistance” poverty in Armenia (18.4 billion drams, or 1.5 percent of GDP would be required to eliminate “pre-social assistance” extreme poverty). The total budget for the family poverty budget in 2001 was 16.9 billion drams. With perfect targeting and assistance given only to the extremely poor, this should have been almost sufficient to lift everyone out of the extreme poverty. However, perfect targeting exists only as a theoretical possibility. In practice, even extremely well designed and implemented systems experience significant leakage. Assuming leakage of 40 percent, 9 the amount needed to lift everyone out of extreme would be 26 billion drams (about 2 percent of GDP). Determinants of social assistance: Econometric analysis shows factors most closely related to the incidence of social assistance (Table VI.B.6 in Annex VI.B of the Main Report). Household composition plays an important role in explaining social assistance. Demographic categories most affected by poverty are also categories most likely to receive social assistance. The children and elderly appear to be more likely to receive social assistance relative to other age categories. The larger the share of children of all age groups in the household, the higher the probability that the household receives social assistance relative to the reference category (share of adults between 45 and 60), keeping the household size constant. The share of elderly has also significant positive effect on receiving social assistance. Female-headed households, which are among those most affected by poverty, are more likely to receive social assistance than male- headed households, being similar in other characteristics. The labor status of the household members is closely associated with the incidence of social assistance, as unemployed individuals are most affected by poverty--the larger the share of the unemployed members in the household, the higher the likelihood of receiving the social assistance. Consumption per equivalent adult has significant negative effect on social assistance incidence. Rise in the consumption per adult equivalent by 1 percentage point decreases probability of receiving social assistance by 3 percentage points. Households residing in temporary lodgings are more likely to receive social assistance relative to those living in houses or apartments. In contrast, car ownership reduces probability of incidence of social assistance. Likelihood of social assistance also declines if the share of land holding owned by the household shows significant size. Finally, household location is significant in explaining incidence of social assistance. Conclusions and recommendations Social protection plays a crucial role in a poverty reduction strategy, providing permanent assistance to those who cannot benefit from other poverty reduction strategies, as well as temporary help for poor households. 9 According to the Integrated Living Conditions Survey, extremely poor receive 48.7 percent of the social assistance. Page 50 50 Improving the social protection system should focus on the following: ƒ Social assistance, i.e. the family poverty benefit program that represents most of the social assistance in Armenia, is a very important poverty reduction and poverty alleviation policy instrument and as such should be maintained, strengthened and further developed. ƒ The program should be expanded to cover most of the extremely poor. In order to achieve this: 1. the targeting formula should be reviewed annually based on the findings of the ILCS. The Ministry of Social Security has a Policy Department with staff capable of carrying out the analysis with appropriate analytical tools; 2. the budget should be increased and, taking into account fiscal constraints, maintained at about 1.5 percent of GDP, which is the approximate amount currently needed for elimination of extreme poverty (assuming perfect targeting); 3. the screening mechanism should be further developed with benefit administrators’ visits to families becoming an obligatory element of the eligibility assessment process. In addition, territorial centres for social services should be obliged to randomly check 5 percent of the family poverty benefit recipients, as part of the auditing process. The findings should be reported to the respective local social protection councils and the Ministry of Social Security; 4. in screening applicants, particular attention should be paid to the households with illiterate household head, rural landless households and rural families whose only source of income was farming —they were found to be facing very high poverty risk, but their coverage by the family poverty benefit was low; 5. the role of the social protection councils should be further strengthened, by authorizing them to officially audit the decisions of the territorial centres for social services, including random field checks of the recipient households. ƒ While pensions are not a poverty reduction instrument, they still play an important role in a poverty-focused safety net in Armenia. In that respect: 1. Care should be exercised that pensions consistently are paid on time in view of the significant positive poverty impact of pensions; 2. The average labour pension should be increased gradually over the medium term to about 25-30 percent of the average wage, which would imply that the share of labor pensions in GDP would increase from 2.5 percent in 2001 to 3 to 3.8 percent over the medium term. The average social pension should be kept at about 60-70 percent of the average labor pension. Pensions at this level would begin to approach the level in more advanced transition countries and provide a reasonable pension for the elderly population; 3. The administration of social pensions may need to be reviewed to facilitate the application and award process to the extent some of the current non-applicants are deterred because of lack of information and/or administrative procedure. Page 51 51 ƒ Regarding other benefits, those related to employment —the unemployment benefit, employment promotion programs, public works—deserve attention. 1. The current unemployment insurance scheme is modest, and with widespread unemployment it is likely to remain that way for some time to come—creating sustainable funding arrangements is best accomplished when an economy is experiencing prosperity, and this is presently not the case in Armenia. Unemployment assistance provides a safety net of last resort for those unemployed who have reached the end of eligibility for unemployment insurance. The benefit is small, and the Government should consider if it should be retained or rolled into the family poverty benefit, while retaining the unemployed as eligible for the family poverty benefit. 2. In their current form, employment services seem to offer little opportunity for placement—there appears to be little monitoring of national (or local) demand and supply for labor, and there is great uncertainty as to the job placement outcomes of training programs that the employment services provide. Nonetheless, employment services can play an important intermediary role, assuming that (i) there are jobs to be had; (ii) the services have adequate information about the labor market, drawn from labor force surveys, employers and administrative data; and (iii) and they continuously monitor labor market developments and adjust programs accordingly. Obviously, this may require resources beyond those currently allocated to the employment services, and a difficult trade-off needs to be faced: should moneys continue to be spent on the present employment services, or should they be closed and the moneys spent on them allocated to some better use; or should they gradually be built up into legitimate European-style services, perhaps initially through a selected set of model offices. While public works have been, and continue to be, an important tool in the Government’s employment policy, one needs to be cautious about the cost effectiveness of public works programs, and one needs to be clear about the objective of such programs: are they meant to transfer income to poor households, build infrastructure, or do both. It may well be that from the point of view of poor families facing income and consumption constraints, public works schemes represent a diversion of resources away from their direct welfare needs. Page 52 52 B IBLIOGRAPHY American Bar Association (2002) CEDAW Assessment Tool Report. Armenia . July 2002. Washington D.C.: American Bar Association-Central European and Eurasian Law Initiative. Avanesyan, V. and Freinkman, L. (2003). Costing-out the Big Bang: Impact of External Shocks on the Armenian Economy at the Outset of Transition. Paper prepared for the Armenian International Policy Research Group conference “Armenia: Recent Economic Trends and Growth Prospects”, January 25, 2003, Washington, D.C. Braithwaite, Grootaert, J. C., and B. Milanovic (2000), Poverty and Social Assistance in Transition Countries. St. Martin’s Press, New York. Bernabè, S. (2002,b), A Profile of Informal Employment: The Case of Georgia, Working Paper # 2002/9, Employment Sector: International Labour Office, Geneva. Chamberlain, G., (1994). Quantile Regression, Consoring and the Structure of Wages in C. Sims and J.J. Laffont eds., Proceedings of the Sixth World Congress of the Econometrics Society , Barcelona, Spain, New York: Cambridge University Press. Chluba, K. and Schmidt-Kallert, E. (2001). Strategy for Land Consolidation and Improved Land Management in Armenia: Pre-Feasibility Study. Food and Agriculture Organization of the United Nations, Sub-regional office for Central and Eastern Europe, Budapest. December 2001 Cnobloch, R., E. Murrugarra and P. Paci (2003) Gender dimensions on health and poverty: a case study from Armenia . Europe and Central Asia Region. PREM-HD. The World Bank. Processed. Cnobloch, R. and E. Murrugarra (2003) Health Status and Health Care Dimensions of Poverty in Armenia . Europe and Central Asia Region. Human Development Department. Processed. The World Bank. Cord, L., Lopez, R., Huppi, M. and Melo, O. (2003). Growth and Rural Poverty in the CIS7. Case studies of Georgia, the Kyrgyz Republic and Moldova. World Bank March 2003. Datt, G. and M. Ravallion (1992), Growth and redistribution components of change in poverty measures, Journal of Development Economics, Vol. 38, pp. 275-295. Deaton, A. and S. Zaidi (1999). Guidelines for Constructing Consumption Aggregates for Welfare Analysis . Research Program in Development Studies Working Paper No. 192. Princeton University. October. De Neubourg, C., C. Namazie and Z. Loza (2002), Part B – Employment and Informal Activities in Four Rural Regions in Tajikistan 2002 – Social Protection Sector Review, The World Bank, Washington DC. Page 53 53 DFID, (2002). Poverty and Social Impact Analysis for Water Sector Reform in Armenia. A pilot study. Department for International Development (DFID UK) conducted by Economic Research Institute of the Ministry of Economy and Finance of the Republic of Armenia. June 2002. Dow, W., P. Gertler, R. Schoeni, J. Strauss, and D. Thomas (1997) “Health care prices, health and labor outcomes: Experimental evidence.” Labor and Population Program Working Paper Series 97-01, RAND, Santa Monica, Calif. Economic Trends (2001) Armenia. Quarterly Issue. (several issues). EBRD (2000) Transition Report. European Bank for Reconstruction and Development. EBRD (2002) Transition Report. European Bank for Reconstruction and Development Elbers, C, J.O. Lanjouw, P. Lanjouw and P.G. Leite (2002) Poverty and Inequality in Brazil: New Estimates from Combined PPV-PNAD Data . Washington D.C.: The World Bank. Processed. FAO Aquastat FAO Stat Filmer, D. and L. Pritchett (2001) “Estimating Wealth Effects without Expenditure Data- or Tears: An Application to Educational Enrollments in States of India,” in Demography , Vol. 38(1), pp. 115-132. February. Hentschel, J. and P. Lanjouw (1996) Constructing an Indicator of Consumption for the Analysis of Poverty. Principles ad Illustrations with Reference to Ecuador . LSMS Working Paper No. 124. World Bank. Washington, D.C. IMF (2002), Country Report No. 02/225, October. Koenker, R. and Bassett, G. (1978). Regression Quantiles, Econometrica 46(1): 33-50. Lacko, M. (2000), Hidden Economy – An Unknown Quantity? Economics of Transition , Vol.8, pp.117-149. Lamb, R. (2003). Inverse Productivity: land quality, labor markets and measurement error. Journal of Development Economics 71 (2003) 71-95. Lerman. Z., Lundell, M., Mirzakhanian, A., Asatrian, P. and Kakosian, A. (1999). Armenia’s Private Agriculture: 1998 Survey of Family Farms. September 1999. ECSSD Working Paper No 17. The World Bank. MOH-UNICEF (2002) Infant Mortality in Armenia: A Review of Procedures for Registration, Classification, and Related Medical Services . Ministry of Health and UNICEF/Armenia. December 2002. Page 54 54 ORC-Macro (2001) Armenia. Demographic and Health Survey 2000 . National Statistical Service (Yerevan, Armenia), Ministry of Health (Armenia), and ORC Macro (USA). Paci, P. (2002), Gender in Transition, Eastern Europe and Central Asia Region: The World Bank, Washington DC. Reilly, B and G. Krsti, (2003), Employees and second -job holding in the Federal Republic of Yugoslavia: an empirical analysis, The Economics of Transition , Vol.11, No. 1, pp. 93-122. Schneider, F. (2002), The Size and Development of the Shadow Economies of 22 Transition and 21 OECD Countries, IZA Discussion Paper No.514, Bonn. World Bank (1999) Improving Social Assistance in Armenia, Report No. 19385-AM. June. World Bank (1999a) Armenia’s Private Agriculture. 1998 Survey of Family Farms . ECSSD Environmentally and Socially Sustainable Development Working Paper No. 17. A study by Z. Lerman, M. Lundell, A. Mirzakhanian, P. Asatrian, and A. Kakosian. World Bank (2000) Making Transition work for Everyone. Poverty and Inequality in Central Asia . The World Bank. Washington, D.C. World Bank (2001) Armenia: Growth Challenges and Government Policies , Main Report, Report No. 22854-AM. World Bank (2001a) Kyrgyz Republic: Poverty in the 1990s in the Kyrgyz Republic , Report No. 21721-KG. June. World Bank (2001b). World Bank Irrigation Rehabilitation Project Implementation Completion Report. Aide Memoire from ICR mission, July 2001. World Bank (2002) Armenia Poverty Update . Report No. 24339-AM. Human Development Department. Europe and Central Asia Region. Washington DC. World Bank (2002a) Armenia: Towards Integrated Water Resources Management. February 2002. ECSSD Working Paper No 35. The World Bank World Bank (2002b) Armenia Trade Diagnostic Study . ECSPE. June 2002. World Bank (2002c) Coping with the Cold. Heating Strategies for ECA´s Urban Poor . Europe and Central Asia Region. Environmental and Social Sustainable Development Department. Document prepared by J. Lampietti and A. Meyer. World Bank (2003), A Study of informal labor market activity in the CIS-7, Conference Volume (Forthcoming), The World Bank, Washington DC. Yemtsov, R. (2001), Labour Markets, Unemployment and Growth, Mimeo , World Bank: Washington. wb56473 C:\Documents and Settings\wb56473\My Documents\ARMENIA\QAG\AMPA_Volume1_Summary of Findings.doc Page 55 55 02/25/2006 5:31:00 PM