6 Inequality and Poverty in the CIS-7 Countries, 1989–2002

Sarosh Sattar, and Clinton Shiells
Published Date:
April 2004
  • ShareShare
Show Summary Details
Jane Falkingham

This chapter examines the evolution and current status of living standards and welfare following a decade of transition in seven of the poorest countries of the former Soviet Union (the CIS-7). At independence, the CIS-7 countries inherited high levels of human capital. Education and health care were universal and provided free at the point of delivery, and there were extensive social services and transfers. High social spending was supported by large budgetary transfers from Moscow. In the late 1980s and early 1990s, such transfers were worth as much as a third of gross domestic product (GDP) in Armenia and more than a fifth in Moldova.1 However, the republics of the Soviet state inherited economic structures that were heavily dependent on Soviet supply and trade networks. The Russian Federation was the main source of inputs and the main market for outputs. Transport and other infrastructure were designed with a view to meeting these trade needs and not necessarily those of the local economy.

The withdrawal of subsidies from Moscow following the breakup of the Soviet Union, combined with the interruption of intercountry trade and the imposition of tight government stabilization policies, resulted in a reduction in output across the region. A series of natural disasters, armed conflicts, and border disputes exacerbated the economic shock accompanying the breakup. The resultant decline in GDP per capita during the early transition years led to large increases in poverty and inequality. Even today, real output remains significantly below pretransition levels in all CIS-7 countries except Uzbekistan.

Trends in Income Inequality and Poverty

In the former Soviet Union, high levels of social expenditure and low wage differentials meant that the distribution of income was significantly more egalitarian than in most market economies.2 Poverty did not officially exist, although there were maloobespechenny (or “underprovisioned”) families—families that had less than a socially acceptable minimum consumption basket.3 A consumption basket reflecting the socially acceptable minimum for a community rather than a subsistence minimum was used to determine who was underprovisioned. On the basis of data from the 1988 Soviet Union Family Budget Survey, with 75 rubles as a proxy for the national poverty threshold, an estimated 11 percent of the population of the Soviet Union was poor.4 The proportion living in poverty varied considerably across the CIS-7, however. Poverty was relatively low (12–14 percent) in Armenia, Georgia, and Moldova, but higher than the Soviet average in Azerbaijan, the Kyrgyz Republic, Tajikistan, and Uzbekistan, where between a third and a half of the population were poor.

Rising Income Inequality

The decline in output since independence has been accompanied by increases in measured inequality. By the end of the 1990s, most of the CIS-7 countries had Gini coefficients well above even the top of the Organization for Economic Cooperation and Development (OECD) range (the United States, at 0.37). The increase has been particularly marked in Armenia, where the income Gini coefficient reached 0.57 in 1998/99. These changes occurred at an unprecedented pace, resulting in an acute sense of relative deprivation for those at the bottom of the distribution. When asked to place themselves on a nine-step ladder, with the first step being the poorest and the ninth step the richest, a majority in Tajikistan placed themselves on the bottom third of the ladder.5

Some researchers have questioned whether comparisons of inequality indices before and after the transition are meaningful.6 Indices may underestimate income disparities in 1989, as the old Family Budget Survey on which they are based was known to exclude people at each end of the distribution.7 However, the survey also failed to capture imputed income from the extensive in-kind benefits provided by state enterprises and local authorities, including subsidized housing, food and utility prices, health, and education. Studies in OECD countries have found that final income is more equally distributed than original or gross income.8 Thus, even if the exact magnitude is hard to measure, there can be no doubt of a significant shift in the distribution of resources within society during the 1990s and a widening of the gap between the top and the bottom of the distribution.

Falling Real Wages and Rising Wage Differentials

Several factors have contributed to the widening of the income distribution during transition, including a shift in the composition of income. In the Soviet Union in the late 1980s, 14 percent of gross income was from private sources, 13 percent from social transfers, and 72 percent from wages, while income from property was nonexistent.9 By 2001, income from wage employment had fallen to 44 percent of monthly per capita income in Azerbaijan10 and 38 percent in the Kyrgyz Republic. Informal sector activity has grown rapidly, much of it small-scale survival activities (private plots, petty street trade, unofficial taxis) undertaken in the absence of formal work opportunities, sufficient wages, or a functioning social security system.11

The growth in income from private sources was accompanied by a rise in earnings inequality (Table 6.1). Between 1989 and 2000, the Gini coefficient for the distribution of monthly earnings increased by more than two-thirds in Armenia and the Kyrgyz Republic and nearly doubled in Azerbaijan. Although nationally representative data are not available for Uzbekistan, a study using data from a survey sponsored by the European Union found that the decile ratio (the ratio of the 90th income percentile to the 10th percentile) had increased from 3.0 in 1986 to 6.5 for all earnings in 1995 and was a staggering 8.3 in the private sector.12

Table 6.1.Gini Coefficient of Earnings in the CIS-7, 1989–2000
Kyrgyz Republic0.2600.3000.4450.4430.3950.4280.4310.4290.4660.470
Sources: 1989, Atkinson and Michlewright (1992); 1991–2000, UNICEF (2002b).Note: The Gini coefficient is based on the distribution of earnings interpolated from group data for monthly earnings, with bonuses, for full-time employees, as reported by employers.

Data for 1993–2000 exclude Transnistria.

Sources: 1989, Atkinson and Michlewright (1992); 1991–2000, UNICEF (2002b).Note: The Gini coefficient is based on the distribution of earnings interpolated from group data for monthly earnings, with bonuses, for full-time employees, as reported by employers.

Data for 1993–2000 exclude Transnistria.

Wage inequality has been fueled by growth of the importance of private sector and informal earnings and the failure to respect employment contracts, leading to widespread wage arrears. An estimated 30 to 45 percent of workers are in the informal sector, either full-time or as a secondary activity.13

Material Poverty

In addition to exacerbating the disadvantage of the “old poor”—pensioners, families with large numbers of children, and single-parent families—the economic dislocation of transition has given rise to new groups of poor, including the families of workers “on leave without pay,” the long-term unemployed, agricultural workers, young people in search of their first job, and a growing number of refugees.

It is difficult to construct a set of poverty statistics that allow comparisons across time within a country and across countries at any one time, because definitions, survey mediodologies, and coverage all vary. Still, on the basis of two international poverty lines, at the end of a decade of transition, an estimated 31 million people in the CIS-7 countries were living in poverty—almost half of them in extreme poverty (Table 6.2). The problem of low living standards is thus widespread, with the majority of the population living below the US$4.30 (in purchasing power parity terms) a day poverty line. Poverty rates measured according to national poverty lines point to similarly high levels of impoverishment, and the ranking of countries remains more or less the same (Table 6.3).

Table 6.2.Poverty in the CIS-7 Using International Poverty Standards, 1999
CountryExtreme Poverty


Total Poverty


Total Population

Extremely Poor



Poor (millions)
Kyrgyz Republic149.184.12.364.04
All CIS-715.2630.88
Sources: World Bank (2000a); for Uzbekistan, World Bank (2002e).Note: Poverty lines are in purchasing power parity (PPP) terms.

Data are for 1998.

Derived using poverty rates from World Bank (2002e). The actual numbers living below the two poverty lines are likely to be much higher.

Sources: World Bank (2000a); for Uzbekistan, World Bank (2002e).Note: Poverty lines are in purchasing power parity (PPP) terms.

Data are for 1998.

Derived using poverty rates from World Bank (2002e). The actual numbers living below the two poverty lines are likely to be much higher.

Table 6.3.Poverty Estimates in the CIS-7 Countries Based on National Poverty Lines, Various Years

and Year
Poverty Rate

Poverty Gap

Severity of

Poverty (P2)
Kyrgyz Republic
Fall 199651.90.200
Sources: Armenia: World Bank (2002a, 2002f); Azerbaijan: World Bank (1997, 2002b);Georgia: Government of Georgia (2000) and World Bank (2002c); Kyrgyz Republic: World Bank (1995, 2001, 2002d); Moldova: Signoret (2002); Tajikistan: World Bank (2000b) and Falkingham (2000a); Uzbekistan: World Bank (2002e).Note: Because of changes in welfare indicators, poverty lines and survey methodology estimates are not directly comparable over time. Except as noted, the welfare indicator is per capita consumption, and the poverty line is the national subsistence minimum.

Consumption aggregate based on food only.

World Bank estimates.

Government estimates.

Poverty line is taken as the living wage.

Sources: Armenia: World Bank (2002a, 2002f); Azerbaijan: World Bank (1997, 2002b);Georgia: Government of Georgia (2000) and World Bank (2002c); Kyrgyz Republic: World Bank (1995, 2001, 2002d); Moldova: Signoret (2002); Tajikistan: World Bank (2000b) and Falkingham (2000a); Uzbekistan: World Bank (2002e).Note: Because of changes in welfare indicators, poverty lines and survey methodology estimates are not directly comparable over time. Except as noted, the welfare indicator is per capita consumption, and the poverty line is the national subsistence minimum.

Consumption aggregate based on food only.

World Bank estimates.

Government estimates.

Poverty line is taken as the living wage.

For countries with time series data, the proportion of the population living in poverty appears to have peaked in 1999 following the aftershock of the Russian financial crisis, with improvements thereafter, except in Georgia, where poverty has stabilized but has not yet started to recover. The changes in poverty over time appear to be consistent with GDP growth. In particular, Armenia and Azerbaijan both experienced strong economic growth during 2000–02, and both recorded significant reductions in headcount poverty.

Poverty and the 1998 Russian Financial Crisis

The Russian Financial crisis of August 1998 affected economic growth in some of the CIS-7 countries more than others. Azerbaijan appears to have been largely immune to any fallout, with growth rates only slightly lower in 1999 than in 1998 and quickly returning to double digits. Azerbaijan was protected by the large share of oil revenues in GDP. In contrast, Moldova was especially hard hit, with negative growth in both 1998 and 1999, reflecting Russia’s importance as a major market for Moldovan exports.14 As Moldova achieved some macro stability and trading conditions started to improve in 2000, poverty also started to fall. By 2001, headcount poverty rates appear to have returned to their 1998 level, but they remained significantly above their precrisis 1997 level.

The Kyrgyz Republic also experienced a slowdown in real economic growth in 1998, with growth rates falling from nearly 10 percent in 1997 to just 2 percent in 1998. Since then, however, the country has benefited from a strong economic recovery, with growth exceeding 5 percent a year after 2000. Per capita consumption in rural areas was further aided by egalitarian land reform and liberalization of agricultural markets. As a result, poverty fell 23 percent between 1998 and 2001, though it remains high.15

There is some evidence that the economic shock following the Russian financial crisis affected different groups differently. Between 1997 and 1998, mean consumption fell 33 percent in the Kyrgyz Republic.16 The largest falls were experienced by households at the top of the distribution, with the top quintile suffering a loss of almost two-fifths over the previous year.17 The same picture was observed in Moldova, with better-off households more affected by the economic shock than the less well off.18 Better-off households included those that had benefited disproportionately from the period of growth prior to the crisis—the self-employed and private sector workers. The crisis wiped out their previous welfare gains. In Armenia, the effect of the shock was relatively short-lived and disproportionately affected those living in urban areas.19

A slightly different picture emerges in Georgia. The sharpest increase came in the measure of poverty severity, which captures inequality among the poor—indicating that the poorest are becoming even poorer.20

Who Are the Poor?

Reducing poverty requires knowing which groups are most at risk of being or becoming poor. The growing availability of household survey data makes it possible to establish a detailed socioeconomic profile of the poor in each of the CIS-7 countries. Several common characteristics are associated with an elevated risk of poverty.

Urban and Rural Location

Urban-rural differences in poverty are marked, with the size and direction of the differences being related to the degree of urbanization. In the most urbanized of the CIS-7 countries—Armenia, Azerbaijan, and Georgia—urban households have a higher risk of being poor than rural households. For example, in Armenia in 1998/99, 55 percent of the population in urban areas was poor compared with 41 percent in rural areas (Table 6.4).21 In contrast, in countries where more than half the population lives in rural areas, the risk of poverty is lower in urban than in rural areas. Several factors have been proposed to explain the pattern of poverty in Armenia, and some may be relevant to other CIS-7 countries. First, Armenia experienced early and relatively egalitarian land reform. Collective and state-owned agricultural land was redistributed in 1995, and access to land has provided a self-protection mechanism for rural households, enabling them to grow some food for their own consumption. Urban households did not have this coping mechanism. Second, urban households were more affected by the financial crisis in Russia. Many urban households have relied on migration to Russia as an income diversification strategy. Remittances account for an average of 17 percent of urban household income, compared with 4 percent in rural areas.22 Remittances received a double setback in 1998, as migrant workers in Russia were directly affected by the crisis and their reduced remittances lost value with the devaluation of the Russian ruble. Preliminary results from the 2001 Armenian household survey, however, show a narrowing of the rural-urban gap since 1998/99, as the incidence of poverty rose in rural areas and declined in urban areas (see Table 6.4).

Table 6.4.Rural-Urban Poverty Differentials in the CIS-7, Latest Available Survey Data (In percent)
Country and

Survey Year

Poverty Rate

Share of Population

Living in Rural Areas
Armenia, 1998/9955.040.632.8
Georgia, 199959.341.143.7
Kyrgyz Republic, 200145.462.465.6
Moldova, 199717.2221.158.4
Tajikistan, 199979.499.572.4
Uzbekistan, 200122.530.563.3
Sources: Headcount poverty rates: Government of Georgia (2000) and World Bank (1999, 2000b, 2001, 2002a,e); share of population living in rural areas: UNDP (2002).

Preliminary findings.

Excludes the capital, Ckisinau, where poverty incidence was estimated at 6.7 percent.

Sources: Headcount poverty rates: Government of Georgia (2000) and World Bank (1999, 2000b, 2001, 2002a,e); share of population living in rural areas: UNDP (2002).

Preliminary findings.

Excludes the capital, Ckisinau, where poverty incidence was estimated at 6.7 percent.

In countries where the majority of the population lives in rural areas, reliance on income agriculture offers no guarantee against poverty. In the Kyrgyz Republic, rural poverty was 1.4 times higher in rural than in urban areas in 2001, despite the high growth rate in agriculture value added that accompanied land reform in the late 1990s.

Inhabitants of small towns, without the employment opportunities of larger cities and with reduced possibilities to engage in subsistence agriculture for food consumption, also face an increased risk of poverty. In Moldova in 2001, 80 percent of the population in small towns were living in poverty, compared with 64 percent in rural areas and 44 percent in large cities.23 Residents of one-company towns face particular difficulties. Restructuring or closing these enterprises causes local unemployment rates to skyrocket, and those who are still working are often subject to short hours or administrative leave.

Both urban and rural poverty continue to represent a substantial challenge in all the CIS-7 countries. In urban areas, poverty is often associated with falling employment opportunities and earnings in the formal economy, reflecting the ongoing restructuring of the industrial complex inherited from the Soviet era. So far, outside the capital cities, the emerging private sector in urban areas has been too small to compensate for the collapsed industrial base. Sustainable poverty reduction will depend on enhanced employment creation within a vibrant private sector.

Household Size and Demographic Composition

The risk of poverty increases with household size. Poor households are generally larger than nonpoor households. However, the relationship between poverty and household size is not straightforward. An increase in the number of adults would tend to reduce the risk of household poverty, since the earnings potential of the household should increase with the number of adults. This is found to be the case in the Kyrgyz Republic, where households with three to four adults have lower relative poverty rates than households with only two adults, no matter how many children are present.24 Thus, households with a large number of children—particularly preschool children—are most at risk of poverty.

Several factors explain why poverty increases with the number of children in the household. First, the presence of children reduces the probability of employment for women. Second, large family size is associated with other characteristics, such as low education, that result in low earnings. Third, households with more children simply have more mouths to feed. Thus, as household size increases, total income adjusted for needs declines.25


From studies elsewhere in the world, a strong positive association is expected between a household’s ability to avoid poverty and its asset holdings—including its human capital. Yet many of the early poverty assessments in the CIS-7 countries found only a weak association, if any.26 These results reflected both the magnitude of the economic dislocation associated with transition to a market economy, with all groups being hit regardless of education, and the perverse labor market incentives for human capital acquisition during the Soviet period, when many of the best educated (teachers, professional workers) received lower pay than those with less education (factory workers, miners).

The advantage conferred by primary and secondary education remains low in the CIS-7 compared with OECD and Central and Eastern European countries.27 However, there are signs of change. In Armenia, education of the household head is found to be associated with increased consumption, with greater gains for those with technical secondary and higher education.28 In Uzbekistan, households in which the head has a higher education were half as likely to be poor as those in which the head had nine or fewer years of education.29 In Azerbaijan in 2001, living in a household headed by a person with higher education reduced the risk of being poor from 50 percent to 42 percent.30

In Georgia and the Kyrgyz Republic, education also reduced the negative impact of the economic shock associated with the Russian financial crisis, with households whose head had a vocational and technical education suffering less than others.31 It is likely that as the labor market becomes more efficient, the link between education and earnings will strengthen. That makes it vitally important that children from poor households be guaranteed the opportunity to benefit from education—an opportunity that appears to be under threat (see below).


Early poverty assessments found that employment was not a guarantee against poverty, although having a job was better than being unemployed or not in the labor force at all.32 Low differentials in poverty rates between workers and the unemployed reflected substantial wage arrears and large numbers of workers on unpaid leave or shortened hours. In all CIS-7 countries, a substantial group of new poor emerged: the working poor.

Many enterprises have adjusted wages and hours rather than employment. Thus, although official unemployment rates remain low, there is substantial concealed unemployment. Wage arrears reached epidemic proportions in many of the CIS-7 by the end of the 1990s. Arrears in the public sector alone were at 1 percent of GDP in Georgia, 1.6 percent in Moldova, and 2.7 percent in Armenia at the end of 1999.33 In Tajikistan in 1999, it is estimated that more than 35 percent of the workforce suffered from some wage arrears.34 The earnings of the poor are more subject to delays than those of other groups. An analysis of the Survey of Georgian Households found that 22 percent of poor wage earners were owed back pay by their employers, compared with 8 percent of the nonpoor.35

Despite the late payments, many workers have been slow to leave their jobs. In part this is because there are few opportunities elsewhere. Employees also stay to retain access to other job-related benefits—particularly housing—-and to preserve routine and maintain self-respect. And there is always the hope that the enterprise will revive and become viable again.

Refugees and Internally Displaced Persons

Refugees and internally displaced persons are particularly vulnerable to poverty, often experiencing loss of property and harsh living conditions in collective centers. They may also experience discrimination in the job market. Children are particularly hard hit by armed conflict. The consequences for them include disruption of schooling, exposure to war trauma, loss of family, and poverty.36

Armed conflict has affected all the CIS-7 countries at some stage since 1989 (Box 6.1). Up to a million people have had to flee their homes, and though many of them have returned, hundreds of thousands of people are still registered as internally displaced within their countries or living as refugees in foreign countries (Table 6.5). In 2001, there were more than a quarter million refugees registered in Armenia and more than half a million internally displaced persons in Azerbaijan and a million in Georgia.37 The region also has been affected by the recent conflict in Afghanistan, with refugees seeking safety in Uzbekistan, Tajikistan, and the Kyrgyz Republic. Georgia has a growing number of refugees from neighboring Chechnya. An estimated 8,000 Chechen refugees—the vast majority of them women, children, or elderly—are living in the Pankisi valley.

Table 6.5.Refugees and Internally Displaced Persons in the CIS-7, 2001
CountryRefugeesInternally Displaced Persons
Kyrgyz Republic9,700
Source; UNHCR (2002).

Includes 6,500 asylum seekers.

Some 2,300 of these are returned refugees.

Source; UNHCR (2002).

Includes 6,500 asylum seekers.

Some 2,300 of these are returned refugees.

Despite the great number of internally displaced persons within the region, relatively little is known about their poverty status. In May 2000, a special survey of internally displaced persons was conducted in Georgia under the auspices of the International Federation of Red Cross and Red Crescent Societies. Not surprisingly, it found that internally displaced persons experienced a substantial disadvantage with respect to ownership of productive assets (especially land) and employment opportunities.

Trends in Capability Poverty

It is sometimes argued that estimates of output in transition economies are unreliable and that declining income and rising poverty are exaggerated.38 It is useful, then, to look at evidence of other dimensions of poverty to get a fuller picture of the changes in welfare during transition. Capability poverty looks at an individual’s capacity to live a healthy life, free of avoidable disease, with adequate nourishment, being informed and knowledgeable, capable of reproduction, enjoying personal security, and able to participate freely in society.39 Material resources at some level are generally necessary for some of these activities, but they are not sufficient.

Box 6.1.Armed Conflicts in the CIS-7

  • Armenia and Azerbaijan, 1988–94. As a result of the war over the territory of Nagorno-Karabakh, about one million people have been uprooted from their homes.

  • Central Asia, the Ferghana Valley, 1989–91. There has been ongoing tension in the Ferghana Valley, which straddles the borders of the Kyrgyz Republic, Tajikistan, and Uzbekistan. Events escalated in 1989, when violent outbreaks resulted in hundreds of deaths and extensive property damage.

  • Georgia, 1990–94. Fighting in South Ossetia in late 1990 resulted in the displacement of 50,000 to 90,000 people. This was followed by conflict in Abkhazia, when hundreds of thousands of people were uprooted. The situation began to stabilize in 1994, although permanent political settlement remains elusive.

  • Moldova, 1992. Following a short and intense conflict, the country split, with the area of Transnistria, which contains much of the industrial base of the country, declaring independence.

  • Tajikistan, 1992–97. The civil war in 1992–93 was followed by continuous unrest until the signing of the peace accord in 1997. During this period, an estimated 50,000 people lost their lives and up to 700,000 people were displaced. Intermittent unrest continues, and refugees continue to return From other countries of the CIS.

Capability poverty can be measured directly in terms of capabilities themselves (percentage of children who are underweight, percentage of adults who are illiterate) or indirectly in terms of access to opportunities (access to trained health personnel at birth, access to education and other public services). This multidimensional nature of poverty is explicitly recognized by the internationally agreed Millennium Development Goals, which address aspects of poverty that are important in their own right and that interact and reinforce other aspects of poverty.

Government spending as a share of GDP has declined in the CIS-7 over the past decade. Since one of the aims of transition is to reduce the once all-encompassing role of the state, a decrease in government spending is to be expected. However, in several countries, the share of government spending is now so low that the functioning of vital state services may be impaired. Public expenditures on health (and education) are now at critically low levels in Armenia, Georgia, and Tajikistan, having fallen below 2 percent of GDP (Table 6.6)—most OECD governments spend 5 to 7 percent of GDP on health care. With the drop in government expenditure, private spending has necessarily increased. However, the introduction of charges for health care, textbooks, and school lunches and the increasing cost of public transport all mean that access to basic social services has been eroded severely. This situation is likely to be reflected in a deterioration in the indicators of capability poverty over time.

Table 6.6.Public Expenditure on Health and Education as a Share of GDP in the CIS-7, 1990–2000 (In percent)
Health care
Kyrgyz Republic3.
Kyrgyz Republic6.
Source: UNICEF (2002b).Note: Public expenditures represent current and capital expenditures on health and education by local, regional, and national governments. Household contributions are normally excluded.
Source: UNICEF (2002b).Note: Public expenditures represent current and capital expenditures on health and education by local, regional, and national governments. Household contributions are normally excluded.


Health is an important dimension of well-being. Several factors can affect health status, from access to health services and safe drinking water to adequate nutrition. This fact is reflected in the three Millennium Development Goals that address health: reducing infant and child mortality, improving maternal health, and combating HIV/AIDS, malaria, and other diseases. By all these measures, health status has deteriorated in the CIS-7.

Access to Health Care

An important determinant of good health is access to good-quality health care. During the Soviet era, health care provision was extensive and free at the point of delivery. Since independence, however, health services have deteriorated rapidly in the face of severe financial constraints, exacerbated in some areas by extensive damage to infrastructure during the armed conflicts that have wracked the region.

Trends in the use of health care services over time are difficult to interpret, being affected by both demand and supply factors. It is clear, however, that the widening gap between health care budgets and the actual costs of care has resulted in diminished quality of services and an increased burden on households, including under-the-counter payments outside official payment channels.40 Such informal user charges for consultations are frequently imposed to help subsidize salaries, despite the fact that such consultations officially remain free of charge. And although medical supplies and drugs required as part of inpatient treatment remain free in principle, patients increasingly have to purchase them because they are unavailable in medical facilities.

There is a small but growing body of evidence that this increase in out-of-pocket payments is affecting access to health care. In a 2001 household survey designed to collect information on the use of health care services in the Kyrgyz Republic, 7 percent of men and 12 percent of women reported that they had sought medical assistance in the past 30 days.41 Another 10 percent of men and 16 percent of women reported that they had needed medical assistance but had not sought treatment. The main reason given for not seeking health care was self-medication using pharmaceuticals (57–58 percent) or herbs (13–14 percent). However, 14 percent of men and 15 percent of women reported that they did not seek medical assistance because it was “too expensive.” Among those who consulted medical providers, total payments for the consultation constituted an average of 10 percent of household monthly expenditures for the poorest households and 5 percent for the richest.42

Of all households reporting that they needed health care in the past year in the Kyrgyz Republic, 18 percent said it had been very difficult to pay for such care, and a further 42 percent reported that it had been difficult. Over half the households reported reducing current consumption to meet health care costs, a third received help from relatives, and 27 percent borrowed money. The picture elsewhere in the CIS-7 is similar.

The current mixture of unregulated prescription charges and payments for consultations that has emerged across the CIS-7 is both inefficient and inequitable. Costs to users vary widely and arbitrarily. Ability to pay for health care is now a major problem for the poor, and there is growing evidence that despite informal systems of targeting, access to health care is being affected. The challenge facing policymakers is to ensure equity in access to health care. The long-term development prospects of a country rest on its human, intellectual, and social capital.

Access to Safe Drinking Water and Other Services

Good health also is affected by environmental factors, particularly access to safe drinking water. Again, households that suffer material poverty are likely to be disadvantaged with respect to access to utilities. The recent Household Budget Survey in Uzbekistan found that poorer households live in homes with less access to running water and public sewerage. Similar patterns are found elsewhere in the region.

There are also significant differences in environmental and health conditions between urban and rural areas. In the Kyrgyz Republic, the average access of the richest quintile in rural areas to each of four health and sanitary indicators was worse than that of the poorest quintile in urban areas. Moreover, the poorest two quintiles in rural areas live in houses without a flush toilet or shower.

Statistics on access alone can be misleading, overstating the proportion of the population enjoying safe water and utilities. A survey in Azerbaijan found that although 95 percent of the households in Baku reported being connected to the water system, water was available only 22 days of each month, on average, and for only four hours a day. Moreover, 87 percent of households believed that the piped water they received was unsafe. The poor in Baku were even less likely to have water every day and throughout the day and were more likely to report illness attributable to water.43 Similarly, while most households in Armenia have piped water, service is unreliable. About 12 percent of households report being without water for three months or more.44 Significant investments in infrastructure, particularly in rural areas, will be needed if the Millennium Development Goal target of reducing by half the proportion of the population without access to an improved water source is to be achieved by 2015.

Mortality and Life Expectancy

The most fundamental measure of the well-being of a population is how long its members can expect to live, on average. By this measure, it is clear that the health of the population deteriorated in most of the CIS-7 during 1990–1995, with declines in life expectancy for women in four of the seven countries and declines for men in five countries (Table 6.7). Mortality rates began to improve in the mid-1990s.

Table 6.7.Life Expectancy at Birth, CIS-7, 1989–2000(In years)
Kyrgyz Republic72.470.472.464.361.464.9
Source: UNICEF (2002b).



Source: UNICEF (2002b).



Life expectancy rates are heavily influenced by trends in infant and child mortality. However, there is sometimes considerable discrepancy between mortality rates drawn from vital statistics registries and those from surveys that ask women about their reproductive health.45 A comparison of infant and child mortality rates from these two sources confirms that rates calculated from survey data are consistently higher than those from official data. In Azerbaijan and Tajikistan they are more than three times as high, reaching levels close to those in India or parts of Africa (Table 6.8)46

Table 6.8.Infant and Child Mortality Rates in Five of the CIS-7 Countries, Latest Available Survey Data(Deaths per thousand live births)
Country and YearSurveyReference YearInfant Mortality Rate (per 1,000 live births)Under-Five Mortality Rate (per 1,000 live births)
2000Demographic and Health Survey19935355
2000Multiple Indicator Cluster Survey199676106
Kyrgyz Republic
1997Demographic and Health Survey19907179
2000Multiple Indicator Cluster Survey199389126
1999Tajikistan Living Standards Survey199882
1996Demographic and Health Survey19934959
2000Multiple Indicator Cluster Survey19995269
Sources: Vital statistics registration data from UNICEF (2002b), except for Tajiktstan 1997, which is from the National Human Development Report.Note: For Demographic and Health Survey, reference years an for midpoint of five-year reference category.Several factors have been put forward to explain the discrepancies, including differences in the definition of live birth and the decline in the proportion of births registered because of increased fees—around 3 percent of average monthly wages in Azerbaijan and the Kyrgyz Republic, 10 percent in Georgia, and a staggering 53 percent in Tajikistan.
Sources: Vital statistics registration data from UNICEF (2002b), except for Tajiktstan 1997, which is from the National Human Development Report.Note: For Demographic and Health Survey, reference years an for midpoint of five-year reference category.Several factors have been put forward to explain the discrepancies, including differences in the definition of live birth and the decline in the proportion of births registered because of increased fees—around 3 percent of average monthly wages in Azerbaijan and the Kyrgyz Republic, 10 percent in Georgia, and a staggering 53 percent in Tajikistan.

Several factors have been put forth to explain the discrepancies, including differences in the definition of live birth47 and the decline in the proportion of births registered because of increased fees—around 3 percent of average monthly wages in Azerbaijan and the Kyrgyz Republic, 10 percent in Georgia, and a staggering 53 percent in Tajikistan.48

Data on adult mortality are more reliable. The mortality crisis among Russian men has attracted considerable attention, with much being written about its causes and the relationship to poverty, unemployment, depression, and alcohol.49 The fact that male mortality also increased elsewhere in the former Soviet Union has received less comment.

Adult male mortality rose significantly in the mid-1990s in the Kyrgyz Republic and Moldova and remained above the pretransition level at the end of the decade (Figure 6.1). The impact of the civil war in Tajikistan is also clearly visible in the data. Indicators of women’s health are, on average, significantly worse in the CIS-7 than in developed countries. Maternal mortality rates are particularly high in Tajikistan, at 74.0 per 100,000 live births (1993), Georgia, 56.9 (2000), and Armenia, 52.5 (2000). By comparison, the average is under 10 in most European Union (EU) countries. Before independence, rates in the Kyrgyz Republic were similar to those in neighboring republics, at 55.6 per 100,000 in 1991, but since then, rates have improved to 33.6 in 1998, almost certainly because of a new reproductive health program. Maternal mortality also has fallen significantly in Uzbekistan (from 30.1 in 1992 to 14.7 in 1999), again largely the result of increased contraceptive use and better reproductive health services.

Figure 6.1.Male Mortality (ages 40–59), CIS-7, 1989–2000

Nutritional Status

There is concern that the level and depth of income poverty faced by households within the region may be affecting children’s nutritional status, with subsequent long-term developmental consequences.

In the CIS-7 countries, the percentage of children classified as stunted (low height for age) is significantly greater than the World 1 Health Organization (WHO) standard of 2.3 percent of children below two standard deviations of the reference population (Table 6.9). The rate of stunting in Tajikistan is particularly worrying, despite the improvement in 2002. Contrary to expectations that access to food would be higher in rural areas and malnutrition lower, the reverse is true, with stunting and wasting (low weight for height) worse in rural areas. More than a quarter of children under 5 in rural Kyrgyz Republic and nearly a third of children in rural Uzbekistan exhibit signs of chronic undernutrition. Around a fifth of rural Azeri children have poor nutritional health.

Table 6.9.Share of Children Under Five Severely or Moderately Undernourished1(In percent)
Country and YearSurveyUnderweight

(low weight for age)

(low height for age)

(low weight for height)
Armenia 2000Demographic and Health Survey2.613.02.0
Azerbaijan 2000Multiple Indicator Cluster Survey16.819.67.9
Georgia 1999Multiple Indicator Cluster Survey3.111.72.3
Republic2 1997Demographic and Health Survey11.024.83.4
Tajikistan3 2002National Nutritional Survey30.94.9
Uzbekistan2 1996Demographic and Health Survey31.311.6
Source; UNICEF (2002b).

Latest available survey data.

Rates are for children 0 to 35 months.

Rates are for children 6 to 59 months.

Note: Moldova not included
Source; UNICEF (2002b).

Latest available survey data.

Rates are for children 0 to 35 months.

Rates are for children 6 to 59 months.

Note: Moldova not included

Countries with high levels of material poverty, such as the Kyrgyz Republic and Tajikistan (see Tables 6.2 and 6.3), also have more malnourished children (see Table 6.9). Children from the poorest 20 percent of households were twice as likely to experience chronic malnutrition as children from the richest 20 percent.50 In addition to being positively correlated with other indicators of well-being at a national level, child nutritional status is strongly associated with indicators of socioeconomic status at the household level. Children whose mothers have a postsecondary school education are less likely to suffer from malnutrition than children whose mothers have only a secondary education or less. There is also a clear gradient in nutritional status by household economic status as measured by ownership of selected assets, characteristics of the dwelling, and access to amenities.

A key Millennium Development Goal is to reduce child malnutrition by half by 2015. It is clear that doing so will require improving both material living standards and other dimensions of poverty.


Education is another important dimension of well-being. Two of the Millennium Development Goals concern education: achieving universal primary education by 2015 and eliminating gender disparities in education at all levels. The CIS-7 countries began the transition with an enviable record on education, with near-universal literacy. School attendance was compulsory from ages 7 to 15, and there was an extensive kindergarten system as well as technical and vocational schools for postcompulsory education. However, there have been serious reversals over the past decade in several countries, and it is unlikely that the high literacy rates of the past will be sustained.

Economic decline may affect education outcomes in three ways.51 First, decreased access (and increased costs) may reduce enrollment. Parents who are unable to afford textbooks, uniforms, or even shoes may simply withdraw their children from school altogether. Second, even if enrolled, children may not actually attend school regularly, whether for the reasons given above or because they are needed as family labor. Finally, children may be enrolled and may be attending school, but may not be benefiting from the education. Teachers may be absent on a second job, there may be no textbooks, it may be too cold to concentrate, or the child may be anemic or malnourished and too lethargic to learn. There is very little evidence concerning learning outcomes, so this section focuses on trends in kindergarten, primary, and secondary school enrollment and attendance and on emerging gender disparities.

School Enrollment

A worrying trend is the decline in the proportion of children ages 3 to 6 enrolled in pre-primary school education (Table 6.10). Prior to independence, attendance at kindergarten was widespread. Since independence, rates have fallen dramatically. This drop is in part due to the closure of employer-provided kindergartens. However, enrollments have fallen by more than the drop in capacity, suggesting a fall in demand for kindergarten places as well as their supply.52 A survey in Tajikistan found that mothers increasingly prefer to take care of their children at home. Reasons given included lack of personnel at child-care facilities, concerns over food safety, and rising costs.53 Such trends are of concern given the role that kindergartens can play in raising household welfare, by freeing parents to participate in paid employment as well as by offering preschool education and nutritional and health interventions.

Table 6.10.Changes in School Enrollment Rates, CIS-7, 1989–2000(Percentage of relevant age group)
Pre-primary Enrollment RateBasic Education
Country19892000Change 1989–200019892000Change 1989–2000
Azerbaijan21.615.8-2787.989.6+ 2
Kyrgyz Republic31.38.7-7292.295.9+ 4
Uzbekistan36.818.2-5192.097.0+ 5
Source: UNICEF (2002b).


General Secondary (15–18 years old)Higher Education (19–24 years old)
Country19892000Change 1989–200019892000Change 1989–2000
Kyrgyz Republic36.723.5-3613.234,6+ 162
Moldova27.422.7+ 116.221.1+ 30
Source: UNICEF (2002b).


Source: UNICEF (2002b).


Primary education remains compulsory. But while enrollment rates in basic education generally have remained high, they have fallen in Armenia and Georgia. Data from the UNICEF TransMONEE (2002b) project indicate that only four of five children ages 7 to 15 are enrolled in school in Armenia. Furthermore, enrollment rates tell only part of the story, as there is also a growing problem of declining school attendance.

Postcompulsory education enrollments have dropped dramatically. The proportion of 15- to 18-year-olds attending general secondary schools has fallen more than 40 percent in Tajikistan and a third in Azerbaijan, the kyrgyz Republic, and Georgia (see Table 6.10). In contrast, there has been significant growth in higher education in four of the countries, particularly in the Kyrgyz Republic, Moldova, and Azerbaijan. Virtually all the growth in higher education has come in the private sector with the proliferation of new private colleges, especially business studies, economics, and law.

School Attendance

Tight education budgets and capitation funding formulas provide an incentive to overstate enrollments in order to maintain teaching posts and schools. Thus, enrollment rates may overestimate the number of children attending school. Furthermore, attendance appears to be well below enrollment rates, making data on official enrollments a poor indicator of the number of children actually receiving educational services. The highest rate of school attendance (for the school year) reported by respondents to Uzbekistan’s 2000/01 Household Budget Survey was 82 percent, and average attendance rates for children ages 7 to 15 (for whom education is theoretically compulsory) were 79 percent for boys and 77 percent for girls (Figure 6.2). This rate is significantly lower than the 97 percent reported for Uzbekistan in official data (see Table 6.10) and indicates a serious erosion of the country’s human capital.

Figure 6.2.School Attendance by Age and Gender, Uzbekistan, 2000 and 2001

Source: World Bank (2002e).

Both demand- and supply-side factors can influence attendance. For example, in Georgia, more than half of children who did not attend school for at least one day during the survey period cited the absence of a teacher, lack of heating, or excessive distance as the primary reason.54 Supply-side factors such as lack of textbooks and other materials were also cited in the poverty assessments for most countries.

However, there is evidence that absenteeism is more prevalent among the poor and that economic factors are an important determinant of school attendance. Traditionally, children from rural areas have been absent at the beginning and end of the school year to help with the harvest and planting. But there is evidence that withdrawing children from school to supplement household income is becoming more widespread. In Armenia, only 52 percent of poor rural households reported that their children attended school very regularly, compared with 68 percent of nonpoor households.

With rising costs, it is difficult to resist the conclusion of a recent United Nations Development Programme (UNDP) report that the education systems in the CIS-7 are beginning to reflect the increasing socioeconomic stratification in these societies. In Azerbaijan, half the university students come from the two richest quintiles.55 Qualitative social assessments find that students and their families increasingly are expected to pay bribes to ensure admission to elite schools, to receive good grades, and to secure entry into a university, effectively excluding students from poor families.56 In Georgia, the richest 20 percent of households spent an average of 22 times as much on educating children as the poorest 20 percent.57 Access to quality education is substantially confined to those who can afford private fees and private tuition.

Gender and Education

Gender equality in education was a key achievement of the Soviet era. Gender ratios of gross enrollment at different levels of education show that in primary school, the ratio of girls to boys is almost exactly one (Figure 6.3). In most countries, the ratio increases in secondary school, indicating that more girls than boys are enrolled. Beyond compulsory education, the ratio increases even more in most countries. The main exceptions are Armenia and Tajikistan, where the ratio falls. In Armenia, survey data indicate that boys are more likely than girls to drop out of postcompulsory education, suggesting caution in interpreting the ratio. In Tajikistan, however, if the current trend of falling enrollments and increasing gender differentials remains unchecked, there is a risk of losing one of the main social achievements of the Soviet period. A new countrywide Cash Compensation Program of targeted social assistance to poor families to stimulate school attendance was introduced at the beginning of 2002.

Figure 6.3.Gender Composition of Enrollment

Source: Adapted from Paci (2002).

Policy Responses to Poverty

Reducing poverty in the CIS-7 and achieving the Millennium Development Goals will require efforts to promote pro-poor growth and to protect and build human capabilities.

Promoting Pro-Poor Economic Growth and Reducing Income Inequalities

Achieving pro-poor growth must remain at the heart of efforts to alleviate poverty. However, it will be important simultaneously to address the economic, institutional, and political causes of poverty. Economic growth alone is not sufficient.

Two factors behind the growth of poverty over the past decade have been falling real wages and rising income inequality. Not only is inequality undesirable in its own right, but there is mounting evidence that it is bad for growth,58 undermines confidence in the government, fosters social exclusion, and impairs the functioning of democracy.59

Where disparities in incomes are large, it will be difficult to increase the incomes of poor families substantially in the short to medium term without some reduction in income differentials. But rather than reducing inequality, government policies and actions often have had the reverse effect. A key route to reducing income inequality is the establishment of transparent and efficient product markets. But progress has been hampered by the slow pace of restructuring (or closing) many of the large nonviable state enterprises and by the capture of these and other state institutions by vested interests. No country in the CIS-7 has achieved a score of more than 2.0 on the European Bank for Reconstruction and Development’s index of enterprise reform (the index ranges from 1, low, to 4, high).

To promote pro-poor growth, the CIS-7 countries need to foster a business environment that is conducive to private sector development by improving governance and reducing the “bribe tax,” streamlining administration and reducing the “time tax,” strengthening credit markets, simplifying taxation and reducing incentives to engage in tax avoidance and evasion, and strengthening the legal framework.

Protecting and Building Human Capabilities

In addition to working to promote economic growth and reduce inequality, public action needs to protect and build capabilities so that the poor can take advantage of new income-generating opportunities. That means reversing the trend of rising differentials in health and education between the poor and the better-off.

There is an urgent need to invest in schools and primary health care facilities. But unless governance issues are also tackled, higher social spending may do little to enhance the human capital of the poor. Many countries need to strengthen public expenditure management systems to ensure that public funds are used for their intended purposes and to reduce the rent-seeking behavior of health and education professionals. It is essential to break the vicious cycle of bribes and corruption within the social sectors.



    Other Resources Citing This Publication