chapter3

This chapter presents a summary poverty profile using the three poverty lines developed in Chapter Two. The group below the NHDR poverty line and above the NHDR intermediate line comprises those families that, technically, receive adequate nourishment. We say "technically" because even though it is feasible for these families to achieve an adequate diet at any point between these two poverty lines, this is subject to choosing the proper dietary components and amounts. As we will see, this is not regularly accomplished, thus some percentage of improperly nourished people should be expected to occur.

This group faces considerable budget constraints in terms of non-food items, but has the potential capacity to switch from the NHDR to the "Recommended Poverty Line" food basket (see Chapter 2), thereby freeing resources for non-food consumption as their budget shrinks. The group below the NHDR intermediate poverty line and above the NHDR extreme line includes families that have serious problems attaining an adequate diet even with knowledge about what to consume and in what amounts. Finally, those falling below the NHDR extreme poverty line are families that do not have an adequate diet and fall way below the necessary amounts to cover minimum non-food needs. This latter group includes those at the margin of society for whom hunger and cold must accompany their daily existence at the time this NHDR is being written and for whom daily survival must be a difficult challenge. We now turn to the poverty headcount for 11 regions of Georgia.

Table 3.1 shows three columns. The first is the headcount, which is the proportion of families that fall below the NHDR poverty line. It shows the percentage of poor in each region and in Georgia as a whole (national average). Next is the poverty gap, which can be understood as a measure of the poverty deficit of the entire population where the poverty deficit captures the resources needed to lift all poor families out of poverty through perfect cash transfers²².

Finally, the severity of poverty, the third column, captures inequality among the poor²³. In general, there is a substantive decrease in the poverty headcount from winter to summer though the change is more pronounced in some regions. While in winter 50% of the country cannot achieve levels of consumption above the NHDR poverty line, this percentage falls to about 40% in summer. Both of these represent very high counts and may help to explain the public mood, which, lately, has not been cheerful.

In terms of poverty and nutrition, the winter percentage of poor in Georgia (50%) is greater than the percentage of people consuming food with a caloric value below 2,200 Kcal/day (41%; see Chapter 2, Table 2.1). However, the average poverty headcount between winter and summer (45%) is close to the average for the 1^st and 2^nd quarters of 2001 (43% of people with less than 2,200 Kcal/day). At the regional level, Tbilisi, Kvemo Kartli, Mtskheta Mtianeti and Shida Kartli present a lower poverty headcount than their respective figures for people with less than 2,220 Kcal. The contrary holds true for Guria, Imereti, Kakheti, Samegrelo and Javakheti. In winter, Adjara shows almost a perfect match between poverty headcount and people with less than 2,200 Kcal/day.

Kakheti and Javakheti show the sharpest decrease in poverty headcount from winter to summer. In Kakheti, both urban and rural families see their consumption increase while in Javakheti urban families benefit more from the arrival of the summer season. Samegrelo is the only exception. Its poverty headcount increases sharply from winter to summer. The enlarged number of poor originates mostly from an unequal pattern of consumption in summer. While spending per capita stays almost unchanged between seasons, poverty increases because those at the bottom presents a lower share of total consumption. The absolute figures for Samegrelo might need to be taken with some caution. Sampling error could account for part of the sharp increase in absolute poverty headcount, but it is unlikely

that it justifies the entire increase. As for Georgia as a whole, the presence of errors in the headcount for Samegrelo would have a negligible impact on the national headcount (less than 1%).

The poverty gap for winter is greater than for summer, although both figures are relatively high. This confirms the results of other poverty profiles in the sense that poverty is not shallow. Poor families are not bunched close to the NHDR poverty line. Tables 3.2 and 3.3, which introduce the poverty headcount for the NHDR Intermediate and NHDR Extreme poverty lines confirms these inequalities among the poor.

The percentage of families unable to reach the consumption level of the NHDR Intermediate Poverty Line accounts for approximately 31% in winter and 25% in summer. These are high poverty headcounts for what is more than just moderate poverty. At the point right on the NHDR intermediate poverty line, an individual conscientious about her diet and knowledgeable about nutrition has already switched completely to the food basket of the "Recommended Poverty line" (see Chapter 2; Table 2.3). From here on, the choices are likely to be in the domain of proper nourishment versus a modest amount of non-food items (see Chapter 2, Table 2.4).

In winter, Kakheti tops the list with 50% of its population below the NHDR Intermediate Poverty Line. The rural population has a much higher poverty rate than the urban population, but since Kakheti is predominantly a rural region the poverty headcount moves towards rural figures. Guria and Imereti follow with 40% of their population in poverty. Samegrelo and Mtskheta-Mtianeti are at the bottom with approximately 18%. In summer, the poverty headcount tends to fall for all regions except Imereti, Samegrelo and Mtskheta-Mtianeti. In summer, Adjara shows the lowest poverty headcount with only 3.67% below the poverty line. In contrast, for the same period, Tbilisi shows almost 18% of its population below the NHDR Intermediate Poverty Line.

The national poverty headcount for winter (NHDR intermediate) closely matches the percentage of people with a diet below 1,800 Kcal/day (28%; see Chapter 2, Table 2.1). This suggests that those falling below the NHDR intermediate poverty line have problems attaining an adequate diet even though it may be feasible technically. This finding indicates that people generally have limitations minimizing food costs while simultaneously meeting safe standards of calorie and micronutrient intake. At the regional level, the percentage of people in Tbilisi on a diet of less than 1,800Kcal/day (40%) is higher than its percentage of population below the NHDR Intermediate Poverty Line (24%), but almost perfectly matches the percentage below the NHDR Poverty Line (38.19%; see Table 3.1). For Adjara, Mtskheta Mtianeti, Shida Kartli, Imereti, Racha and Tbilisi, the percentage of inadequately nourished people (less than 2,200 and 1,800 Kcal/day) in the first quarter of 2001 falls within or close to the interval given by the NHDR and NHDR Intermediate poverty lines. Together, these figures suggest that capacities to achieve a safe diet have a regional component and that poverty lines can be more instructive if seen in conjunction with other indicators, such as adequate nutrition.

Table 3.3 presents the headcount for those falling below the NHDR extreme poverty line. The national headcount is approximately 13% in winter and 8% in summer with similar values of poverty gap. This indicates that the mean distance that separates extreme poor from the poverty line does not vary between seasons. The national average between winter and summer is about 10.5%, a bit below the poverty headcount obtained by means of the Alternative Minimum poverty line, the poverty threshold used in the Poverty Reduction and Economic Growth Program. Yet, an average of 10.5% is not a low level of extreme poverty. These poor are of a different nature than poor below the NHDR and NHDR intermediate. These are people facing serious deprivation in daily life with little capacity to exploit job opportunities that may be available, with little capacity to access health care services, and with seriously inadequate diets. An extreme poverty headcount of 10.5% is high and bad news for Georgian society.

Variation in the number of extremely poor families occurs between regions and seasons. If the three NHDR poverty lines are compared, they provide some insight into the consumption levels in which the poor are grouped. For instance, Racha and Kvemo Svaneti present high poverty headcounts with the NHDR Poverty Line and NHDR Intermediate Poverty Line, but close to zero with the NHDR Extreme Poverty Line. This indicates a high percentage of people in these regions are poor, but they are bunched above the NHDR Extreme Poverty Line. Adjara shows a similar but less marked pattern. We will elaborate more on this subject when exploring inequality in consumption at the end of this chapter.

Guria and Imereti show some of the highest poverty headcounts in all poverty lines explored. Guria, however, starts to show a better headcount than Imereti in summer and this difference grows the lower the poverty line. By the time it is down to the NHDR Extreme Poverty Line, Imereti has almost three times the percentage of extreme poor as Guria and tops all other regions. This happens in the presence of higher consumption per capita in Imereti than in Guria. The distribution of total output in Imereti is more unequal than in Guria and the cost of the food basket slightly higher. Together these account for Imereti’s higher poverty incidence.

Box 3.1: poverty in Imereti

"All regions are in a really dreadful situation, but Imereti is really something special. I have been there this summer and I saw how the whole village was looking forward to Indian corn harvest and hoped that they will have at least something to eat in winter. As for cash, nobody has cash".

(Extension worker; Imereti)

Tbilisi and Adjara also allow for comparisons. They show an almost equal proportion of poor with the NHDR Poverty Line in winter, but Adjara experiences a greater reduction in poverty in summer. In addition, the poor in Adjara are grouped primarily above the NHDR Extreme Poverty Line. Tbilisi, on the other hand, has a percentage of people below the NHDR Intermediate and Extreme poverty lines that is greater than Adjara and this reflects greater levels of inequality. A similar type of exercise can be done for other regions. We now turn attention to the poverty headcount for urban and rural sectors. These are presented in Table 3.4 below.

When the NHDR Poverty Line is applied, urban Georgia has a lower poverty headcount than the rural one. The poverty gap for urban areas is higher at any poverty line indicating that the urban poor fall, on average, further from the poverty line than their rural counterparts. This is also reflected in the comparison between urban-rural as we go down the poverty line. Rural areas have a lower percentage of families whom fall below the NHDR Intermediate and NHDR Extreme poverty lines. Later in this chapter, it will be seen that urban areas have much greater inequality than rural ones and this explains the reverse in rankings. The poverty headcount also decreases from winter to summer as previously observed for Georgian regions. As mentioned in Chapter 1, the gains from winter to summer for urban areas are not equally distributed among cities and towns in Georgia. Approximately 57% of the urban population benefits from the arrival of summer. Tbilisi, Adjara, Kakheti, Samegrelo and Javakheti derive the preponderance of these gains.

Tables 3.5 and 3.6 show the average per capita spending and food share partitioned by urban-rural populations in winter and summer.

Spending per capita varies considerably between poverty lines. Table 3.5 and 3.6 highlight the degree of inequality in Georgian society and the fact that this inequality is more preponderant in urban areas. From bottom to top, one can see that extreme poor in urban areas spend half the amount of poor. In turn, non-poor people spend between three and four times more than poor and almost 7 times more than extreme poor. In addition, consumption is much more equally distributed among poor than among non-poor further indicating that the benefits from economic growth have not been reaching all families equally. Rather the lion’s share has been captured by a few²⁴.

Share of food spending in total family consumption provides some surprises. First, the share is low for some groups, particularly urban ones. Second, in winter, as the poverty line is lowered, the share of food spending among urban poor shows little change. This is counterintuitive. One would expect that as people get poorer, the share of spending going to food would increase more than shown in Table 3.6 since food has such a high order of importance. Several things could account for this finding. One, of course, is measurement error. Another is that there may exist non-food expenditures that households find difficult to avoid even at very low levels of total consumption.

Third, and most interestingly, the share of consumption going to food is greater in summer than in winter for all groups except the one above the NHDR poverty line. The increase in spending on food between winter and summer is greatest for the poorest group. What it is interesting is that in this case a greater share of food spending in summer may not indicate greater levels of poverty in summer. The presence of a larger share of food spending takes place together with little change in per capita spending suggesting that households in winter re-direct spending from food to non-food items and that the inverse occurs in summer. This provides additional weight to the idea that families face winter expenditures that are hard to avoid, for example, heating fuel. What is troubling is that the trade-off is taking place at the expense of food for otherwise one would expect food share to decrease in summer. If this line of reasoning holds some truth, it means that on average poor people have less to eat in winter because they have expenses that they cannot avoid. Table 3.7 below shows the share of non-food spending by broad categories.

Table 3.7 shows that among all non-food expenses, communal payments have greater weight the further down the poverty line one goes. For households above the NHDR Poverty Line in winter these payments can take an average of 20% of total non-food spending, for those below the NHDR Extreme Poverty Line the share goes up to 40%. The increased share in communal payments is proportionally heavier as one goes down the poverty line because shares of non-food spending also decrease.

Not surprisingly, Figure 3.1 shows the poorest of the poor spend proportionally more on electricity and kerosene than all other groups. They also spend more on gas than those below the NHDR Poverty Line and more than any other group for the provision of water.

It would not be fair to say that poor people systematically avoid payment of communal services like electricity or gas.

Figure 3.2 shows the bulk of household spending among the poorest of the poor comprises small non-food items (e.g. candles and detergent) and funerals, the latter a difficult expense to avoid in view of ancient customs that characterize Georgian society. Together with expenditures to keep clothes clean and large household materials (e.g. a window) these take almost the entire budget for household expenditures of extremely poor families.

Figure 3.3 shows categories of personal spending for non-poor, poor and extremely poor. Local transportation, which was perhaps a controversial allowance for a non-food poverty basket (see Chapter 2), turns out to amount to 80% of the money spent on medicine and medical services. In fact, poor people spend twice as much on transportation as for medical services and half the amount devoted to medicine. Expenditures on education takes on average a low proportion of the budget. For the poor, it is approximately 5 to 6% of total personal spending and 0.5-0.6% of total spending. This value is so low that it could have passed unnoticed as just statistical error in the measurement of total consumption. For the poor, it highlights how much they depend on the existence of a state-sponsored education system that can provide their children with the necessary tools to escape poverty.

We close this section by looking at sources of household income. Several things catch the eye in Figure 3.4. Most noticeable is the low contribution of take-home money to monthly income.

For the extremely poor, the contribution of pensions to total family income is a major characteristic and shows that delays in their payment can be a devastating blow. Loans and gifts follow in order of importance after take home pay and pensions.

Loans, however, are not true income as they are one-time receipts that must be repaid. In conversations with the public we heard numerous stories of families borrowing 100 GEL, or even less, not to invest in productive activities, but to make up for shortages in monthly consumption, like food. These loans are at interest rates of about 20% monthly, sometimes more, and demonstrate the degree of desperation families can experience. In turn, the category "gifts" represents help from non-household members. When all sources of income are taken into account, the picture is one of high instability. For the poor, income from take home pay accounts for about one third of the household budget while the remainder falls to sources that are either unstable (e.g. pensions), have to be repaid in the future (loans), or are seasonal (food products sold).

Next we present some demographic characteristics of families and their relationship to poverty levels. Demographic characteristics are helpful variables to identify groups at risk. They are also among the most visible traits, which in turn can facilitate accurate targeting. Table 3.8 below shows some selected characteristics for families below the NHDR poverty line. Table 3.8 presents three different measurements of poverty.

The first is the "Poverty Incidence", which is simply the percentage of poor with a given characteristic within the entire population of the country. For instance, poor couples without children represent 4.9% of the entire population of Georgia. The second is "Share in Total Poverty", which shows the percentage in a given category among all poor in Georgia. For example, poor couples without children represent 10.68% of all poor in the country. The last column, "Share in the Selected Group", indicates the percentage of poor people within a given family characteristic. For instance, 40.06% of all couples without children fall below the NHDR poverty line²⁵.

Among all groups considered, pensioners with children and single pensioners are the group most affected by poverty. 75% of pensioners with children and 72.78% of single pensioners fall below the NHDR poverty line. A second group composed of single mothers and single working age adults follows in that order. Extended with and without children and a single adult plus pensioner show almost the same proportions (46%). Finally, couples with and without children appear at the bottom with the lowest percentage of poor within their groups.

Pensioners form a group that consistently shows the highest poverty incidence among all groups at any poverty line. Pensioners and pensioners with children represent two of the most vulnerable and dispossessed groups. These groups include people that cannot work because they are either too young or too old, and whose principal source of income depends on a pension, which as Georgians know, is an extremely unreliable source of income. These results, once again, indicate that failure to pay pensions can have devastating effect on those already enduring great suffering. Budget sequestrations and/or mismanagement in the social funds that directly or indirectly result in pension arrears result in people whose budget can be cut 1/3 or more and result in plain hunger.

Next we consider whether the order of these same groups changes as one goes down the poverty lines. In other words, whether the use of lower poverty lines will help identify those in greatest need. Table 3.9 below shows demographic characteristics for poor people as defined by the NHDR Intermediate and NHDR Extreme poverty lines.

By comparing Tables 3.9 and 3.8 one notices that pensioners with children, single pensioners, single mothers and single working age adults are groups consistently with the highest poverty incidence no matter how far one goes down the poverty line. Lowering the poverty line does not improve accuracy in targeting through demographic characteristics. The rankings of the most vulnerable remain unchanged for all three poverty lines. In terms of absolute numbers, they constitute a sizeable group of about 20 to 30% of all poor depending on the particular poverty line that one prefers to use.

The effect of children on poverty incidence is only clear in the case of pensioners with children. For all other groups the presence or absence of children does not convey a clear message about whether they can serve as visible variables in identifying poor households. If one takes the lowest poverty line, couples and extended families with children have a higher proportion of poor people than their respective categories without. However, these differences are small and could be the result of sampling variation. In the Technical Annex 1, results will be shown that suggest the presence of children as an indicator of higher poverty risk may depend on age. Children aged 0 to 6 are positively correlated to changes in family welfare, while children aged 7 to 17 are negatively correlated.

Table 3.10 explores the incidence of poverty by gender. Only the NHDR and the NHDR Extreme poverty lines are shown. The values of the NHDR intermediate fall within these two and have been omitted.

Table 3.10 shows there is not a disproportional effect on poverty by gender, generally. Only three out of 11 regions, Adjara, Mtskheta-Mtianeti and Kakheti, show disparity in the percentage of males and females affected by poverty. Adjara is an interesting case because as one lowers the poverty line, the percentage of males in poverty increases and subsequently the percentage of females decreases. Mtskheta-Mtianeti and Kakheti show the opposite trend, but in Mtskheta-Mtianeti this is highly marked with females being four times more likely to be in extreme poverty than males. We cannot provide potential explanations here for these disparities although these topics merit further research.

Finally, Table 3.11 shows poverty incidence by marital status. There is no clear variation for single individuals and a slightly lower presence of married couples among extremely poor.

Divorced and widows, however, show a higher presence under the NHDR extreme poverty line. This may reflect increased difficulties in securing adequate income, particularly in the case of females with children.

Poverty incidence in relation to different levels of educational achievement is explored next. Education is an important factor in providing individuals with skills to escape poverty. It allows the acquisition of knowledge, and together with experience, can provide clear signals in the job market about the abilities of the person in question. Table 3.12 divides individuals by categories of education and shows overall poverty incidence, share in total poverty and share of poor in the selected group.

There is a clear decrease in poverty incidence within each group as the level of educational achievement increases. Almost 60% of people with only elementary education fall below the NHDR poverty line while the percentage for those with uncompleted higher education is 44% and goes down to 38% for those with tertiary education and higher. Roughly, those with more education are one third less likely to be in poverty than those with an elementary education. The percentage of poor within the groups of those with secondary and technical educational levels does not differ much. One half of people with secondary education are at risk of being below the poverty line while the percentage for those with technical studies is 47.79%. This is a relatively minor fall in poverty risk and suggests that the gains from pursuing only technical education are limited and that the major prize is in the achievement of a university degree.

The contribution of groups with uncompleted elementary and secondary education to the total number of poor is not great accounting for 7.39% of the total. A major contribution to the total number of poor comes from those having completed secondary and technical education. Together they account for almost 45% of total people in poverty. This may reflect several things. One is the legacy of the past system in which access to secondary education and technical institutes was widespread. Another is that the skills gained through technical education may be of limited value today. The main sources of employment for these individuals (e.g. state companies, state units for agricultural research and production, etc) are defunct following the collapse of the Georgian economy after 1991. This process also affects those with a tertiary degree (or higher) though their opportunities are more ample.

Next, we examine the poverty incidence of these same groups, when poverty is defined using the NHDR Intermediate and Extreme poverty lines. The objective is to explore whether there are significant variations in the ranking as the poverty threshold is lowered. It is a way of testing whether lowering the poverty lines can help to identify groups in most need when educational categories are used. Table 3.13 shows the results.

In Table 3.13 one can observe that people with elementary and uncompleted secondary education continue to top the list of those with the highest rates of poverty (measured by the percentage of poor within each group). In contrast, Table 3.12 and 3.13 show that people with a university degree are less likely to be in poverty than any other group.

However, the differences among rates of poverty within groups diminish as the poverty line is pushed to lower levels. This may indicate that people in extreme poverty could be in that situation not because of lack of education, but because of other factors at play. Among several examples that come to mind, one thinks of a person who obtained a degree in an agricultural related field and worked during Soviet times in one of the several state farms. The collapse of the Soviet system left this individual jobless with limited prospects for using skills and experience in an agricultural sector that now is hardly more than a subsistence nature.

Another example, a person trained in either physics or mathematics who worked during Soviet times in any one of several high quality research units and now tries to survive on a salary of 40 GEL a month unable to sell his skills on the new job market. These people may have modest additional tools to fight poverty when compared to those with only completed secondary education in a job market that seems to be in need of neither. Note that in these cases education was not of poor quality, outdated, or unable to be of productive use in a market economy. The scores of researchers from the former Soviet Union that are working in western universities testify to the contrary. Rather the highly restricted opportunities in Georgia today appear to erode the relative advantages of higher education for those falling to the bottom of the consumption distribution.

Tables 3.12 and 3.13 appear to indicate that lowering the poverty line may not be the most useful tool to target poor groups when educational categories are used. Elementary and uncompleted secondary education are two groups that show the highest poverty rates using any poverty line. Those with tertiary education always show the lowest.

Public opinion polls show that lack of jobs tops the list of concerns of the general population (see Chapter 1, Figure 1.1). Lack of employment affects poor people more than non-poor, which is a recurrent finding in other poverty profiles carried out in Georgia. Being out of the labor force has a negative effect on short and long-term family well-being. It results not only in less income, but also disconnects people from networks of contacts and increases their social isolation.

Table 3.14 shows the percentage of people that considered themselves to be unemployed and did not work in the week prior to the survey. Their numbers are significant and reflect the lack of success of the Georgian political and economic situation to expand the labor market. The ranks of unemployed in Georgia are not a group characterized by low levels of education.

As shown in Table 3.15 the bulk of unemployed have post-secondary studies either in the form of vocational, incomplete tertiary or tertiary degrees.

Table 3.16 shows the poverty incidence among all poor and the poverty rate for broad categories of employment. Inactive or unemployed unequivocally provide the bulk of the poor. The poverty rate within this group is also the highest with half of them falling below the NHDR poverty line. Part-time workers or temporary employees are the group with the next highest rate of poverty, 47.66%. Employer/entrepreneurs follow with 40.82%. The group with the least percentage of poor is salaried employees with 39%.

One notices that inactive or unemployed and temporary workers are not far away from each other in terms of poverty within the groups. These figures suggest caution is necessary when using national and regional employment statistics as indicators of overall welfare because this group of temporary workers, even if employed for no monetary salary, would show as employed. Their type of employment, however, appears to provide modest help to stay out of poverty. In turn, salaried employees and entrepreneurs share a similar poverty rate within groups. Even though the level of salaries in Georgia is overwhelmingly below the official subsistence minimum, employment opens the opportunity for additional and/or informal income sources. It also keeps alive the network of contacts. The latter two play an essential role in helping a family remain above poverty levels. The group of employers/entrepreneur is a mixed group with a minority that approximates the western image of an entrepreneur in a market economy and scores of people undertaking barely subsistence level activities. For instance, the numerous street vendors of cigarettes and drinks are considered entrepreneurs. Many of them have sold assets to invest in a mobile refrigerator, for example, and pay an informal sum to have it plugged to a lamppost in the street.

Attention is next focused on variation in poverty levels affecting these groups as the poverty line is lowered. Table 3.17 presents the same categories as in Table 3.16 for the NHDR Intermediate and NHDR Extreme poverty lines.

Under the NHDR intermediate poverty line, inactive or unemployed continue to top the list of groups with the highest poverty rates. However, the difference between inactive or unemployed and temporary workers is now insignificant. Salaried employees and entrepreneurs form another distinctive group with little difference in their poverty rates. If one diminishes even further the poverty line, the inactive or unemployed continue to register at the top of the list followed by temporary workers. The risk of salaried employees is now slightly higher than that of entrepreneurs. Lowering the poverty line continues to identify the same groups with the highest poverty rates.

While employment is one of the top preoccupations of the general public, so are low incomes. Caution should be exerted in equating access to a job with access to sufficient income to meet the family needs. Approximately 70% of extremely poor people earn less than 60 GEL/month while 35% earn less than 10 to 30 GEL/month, which is a sum unable to cover even the cost of basic food.

Disparity in salary exists among poor and non-poor and between males and females. While Table 3.10 (Gender and Poverty Incidence) did not convey the existence of a gender dimension in poverty, Figure 3.6 does indicate a gender dimension in access to better paid jobs.

Figure 3.6 appears to be reflecting an unequal access to better paid jobs. Males have a greater access to well paid positions than females. For both males and females, low salaries from the primary source of employment oblige them to look for a secondary income source, which as Figure 3.7 shows, provide low earnings. A significant percentage of people that take a second job do so for no salary at all.

Of all Georgians engaged in secondary income activity, 35.7% receive no salary. This closely matches the percentage of those working in agriculture. This type of employment often amounts to tending a home garden for purposes of producing food for both home consumption and sale of what is leftover. Another 4% take secondary employment for a salary in the range of 1-10 GEL, while an additional 26% work for a salary in the range of 10-60 GEL. Two thirds of people with secondary employment work for a salary below the NHDR Intermediate Poverty Line and half the minimum official subsistence income.

Health care is a crucial factor affecting the Life Expectancy Index, one of the three components of the Human Development Index. It is also a crucial variable because the health status of the individual greatly affects her/his ability to obtain income while an illness can send a family down a spiral of debt and poverty. Those with greater income access health care services more often, as would be expected and as it is shown in Figure 3.8.

Access to medical assistance not only shows inequality, but also seasonality. Figure 3.8 shows the percentage of non-poor, poor and extremely poor that sought medical assistance during the winter and summer in 2001. Note that in Figure 3.8, poor people sought less medical assistance in summer than in winter. However, one should resist the temptation of concluding that the summer, with its more benign weather, improves the general health of the population. The percentage of people above the poverty line whom accessed medical assistance is almost the same between seasons. Since there is no reason to believe that in summer poor people enjoy better health conditions than rich people, the differences in trends in Figure 3.8 should have another explanation. A potential one could be that poor people tend to postpone a visit to the doctor until it is really needed, often with dire effects on the health of the individual and the budget of the family in question. Summer, with its more benign weather, may result in a greater percentage of poor postponing treatment. We have no hard evidence for this save that of consultations with the general public in which this behavior was explained.

Figure 3.8 also reflects seasonal variation in spending on medical services and medicine by different groups. On average, a bit less than 10% of people spend money on health care. Poor people make less use of health services and spend less money on medicine in summer than in winter.

Those at the bottom of the social scale reported poorer health than rich households. This patterns holds for both urban and rural areas and it shows a seasonal component as observed in Figure 3.8. The number of poor reporting bad health is greater in summer than in winter. Again, we do not have a solid explanation for this trend, which is to some extent counterintuitive.

Table 3.18 shows the percentage of poor and non-poor people that needed to visit a doctor, but did not. As the poverty line goes down, the percentage of those that chooses not to access medical services increases. This pattern is unchanged between winter and summer. Note that both poor and non-poor people have problems obtaining proper health care.

Both in winter and summer, the percentage of extremely poor that felt they needed to visit a doctor but chose not to is approximately 1.6 times greater than among the non-poor. For those below the NHDR Poverty Line and NHDR Intermediate Poverty Lines the increase is approximately 1.3 times. Overwhelmingly, the reason to skip a visit to the hospital or a doctor is lack of money, as Table 3.19 below shows.

Table 3.19 shows that the reason to avoid treatment is simply lack of money and that access to health care is beyond the reach of a substantive share of Georgians. Take the group below the NHDR Poverty Line for example. On average, 10% of them requested some kind of medical assistance while 90% did not. Out of this 90%, 27.5% did not even though they believed medical assistance was necessary.

In turn, 90% of them could not visit the doctor simply because they could not pay. Putting these percentages together, one can see that while 10% of poor requested medical assistance, more than double that figure did not because they lacked money. This means that skipping a visit to the doctor because of lacks of funds is the rule rather than the exception.

The problems, however, do not end here. A percentage of those whom located the resources necessary to pay for medical services encounter another barrier to full treatment. Many lacked the money needed to pay for medicine. Table 3.20 presents the share of non-poor and poor whom after visiting the doctor could not afford the medicines prescribed.

21% of people below the NHDR poverty line who visited a doctor found themselves unable to afford the prescribed medicine. This percentage climbed to 27.9% in the case of the extreme poor. One had already estimated that while 10% of poor people requested medical services, double that figure could not because they did not have the funds. Now, out of the 10% that managed to visit the doctor, one fifth could not buy the medicine prescribed. Putting these numbers together, the percentage of people that had a complete course of treatment (visit to the doctor and purchase of required medicine) was only 24% of all poor in need of medical assistance. Of all extremely poor that would have needed medical services, only 15.6% received a complete course of treatment. In spite of several years of reform, health care services are still inaccessible to the great majority of those that are in need of them.

In the final section of this chapter, attention is turned to the measurement of inequality in three components of the Human Development Index. Inequality in per capita consumption is explored by region and by urban-rural areas and we also present measurements of inequality in the consumption of health care and education services. Table 3.21 below presents Gini and Theil coefficients for winter and summer in 11 regions of Georgia.

Box 3.2: Measuring Inequality

There are several ways of measuring inequality and we briefly explain three of them that are applied in this report: the Gini Index, the Theil Index, and the Lorenz Curve.

The Gini Index is a widely used measure of inequality and varies between 0 and 1 with 0 being total equality and 1 total inequality. The Theil Index starts with 0 but can go beyond 1. Its numerical interpretation is not straightforward. However, it has the advantage that it can be decomposed and therefore it can highlight the contribution of different factors to total inequality. Finally, the Lorenz Curve shows inequality in the form of a graph with cumulative distributions (e.g. income or consumption). For instance, by means of a Lorenz Curve one can see that a tiny 2% of Georgians account for 50% of all private expenditures in education.

The national average Gini indexes for Georgia, 0.39 in winter and 0.41 in summer, are high and indicate a worrying degree of inequality in the distribution of consumption. For purposes of comparison, the Gini indexes for Latin American countries, which are characterized by high levels of inequality, start around 0.4. Javakheti in winter and Tbilisi in summer have the highest Gini indexes. Two regions show great change between seasons. One is Javakheti whose Gini index is cut by more than half from winter to summer. Another is Samegrelo in which the Gini index increases approximately 1.5 times in the same period. In the case of Samegrelo, the increase in inequality from winter to summer is solely responsible for the sharp increase in its poverty headcount in the same period as presented in Table 3.1. For Javakheti, the opposite holds true though the arrival of summer also brings a slight increase in per capita consumption.

Adjara shows little difference in its Gini indexes between winter and summer and this contributes to its decrease in poverty headcount in the same period. A similar thing happens in the case of Kakheti though its Gini coefficients are much higher than Adjara’s. The higher level of inequality and lower per capita consumption results in six times greater percentage of extremely poor in Kakheti than in Adjara in summer. Tbilisi follows a similar pattern. The capital of Georgia shows an increase in average consumption from winter to summer of around 30%, which is a significant increase and above that of Adjara. Yet, Tbilisi fails to distribute the summer windfall equally and has a percentage of extremely poor people that is six times that of Adjara. Racha and Kvemo Svaneti is an example of a poor region with a Gini coefficient that is low in relation to the national average. This relatively better distribution of total output results in a significant percentage of poor under the NHDR and NHDR Intermediate poverty lines, but almost none under the NHDR Extreme.

The high Gini indexes for both winter and summer reflect an important obstacle to the Poverty Reduction and Economic Growth Program. The obstacle lies in that with this degree of inequality, economic growth will need to be quite high and sustained in order to reach the poorest strata of the population. The Gini indexes are in line with the perception of the general population that the benefits from economic reforms have only reached a few.

Table 3.22 decomposes the Theil indexes of Table 3.21 between and within regions’ inequality. The totals at the bottom of Table 3.22 show the contribution to total inequality that originates from within and between regions. It can be seen that the bulk of inequality comes from within (not between) regions and that Tbilisi is a major contributor to both between and within inequality. The decomposition of the Theil Index says that Georgia is a country in which the capital, Tbilisi, captures the bulk of economic gains (thus its high contribution to "between region" inequality) and distributes these gains in an unequal manner. Inequality worsens in summer and this is reflected in the increased contributions of Tbilisi to between and within region inequality. Imereti, Kakheti and Adjara follow though far from the values of the capital.

In summer, inequality in rural areas is much lower than in urban ones. The effect of the cropping season plays a major role.

Recalling for a moment the description of the HDI indexes for urban and rural sectors for winter and summer (Chapter 1; Figure 1.9), we saw that the HDI index of urban areas grew significantly between seasons (the index for rural areas stayed almost unchanged). This seasonal windfall is not equally shared among urban families and is reflected in an increased Gini index for urban areas in summer. In turn, this increased inequality is reflected in the increased urban contribution to total between and within sector inequality from summer to winter as shown in Table 3.24.

Inequality in consumption of educational services as measured by the distribution of expenditures on education in the total population is explored next. Figure 3.11 shows a Lorenz curve for education expenditures in winter 2001. It shows that education is the most unequal type of expenditure. A tiny minority of 2% of Georgians account for about half the total private expenditures on education. Figure 3.11 below shows only the curve for winter. The one for summer is even more unequal. The reason for this increased inequality is that educational expenses in summer are mostly tutors for general support and preparation for entrance exams. Only the better off invest in these services.

The skewed distribution in consumption of educational services is reflected in an extremely high Gini index for education. As Table 3.25 shows, the index is close to 1 and signals almost total inequality. Almost the same value for the index can be obtained when urban and rural sectors are considered. As previously seen, the contribution to total inequality comes primarily from within regions with the bulk of it originating in Tbilisi.

Next a Lorenz curve for health care expenditures for winter 2001 is presented. The curve reflects the appalling disparities in access to health care by poor and rich Georgians. A high level of inequality characterizes the Lorenz curve for health expenditures. Figure 3.12 shows that a tiny minority of 5% of Georgians account for 50% of all private expenditures in health care.

The Lorenz curve is skewed to the right because of inequality in access to health care services and because not all people need to spend money on health care in the same month. One could argue that the Lorenz Curve is being pushed to the right because of an artificial contribution to total inequality arising from people that do not spend money on health, not because they cannot but because they do not need to. Unfortunately, this does not seem the case. The limited capacity of poor people to spend on medical assistance and medicines appears to be the main contributor to inequality in access to health care services. The computation of the Gini index for only the subset of individuals who have spent at least 3 GEL on health care gives a value of approximately 0.76, not far from the 0.77 shown in Table 3.26.

As seen in the case of inequality in overall consumption and inequality in expenses on education, the major contribution comes from inequality within regions with Tbilisi, once again, contributing the bulk of it. Imereti, Kakheti, Samegrelo and Adjara follow Tbilisi though their values are far from that of the capital. It indicates that a minority in Tbilisi accounts for a disproportionate share of spending on health care.

The current pattern of distribution in consumption, education and health care presents a picture of a highly unequal country. Georgia is rapidly approaching levels of inequality comparable to those observed in Latin America, a region characterized by a highly biased distribution of total output. As it has been observed in that region, a highly unequal distribution of income erodes the legitimacy of much needed reforms, as only a few benefit while many are ignored. A fairer distribution of economic gains is an overdue item in the plan of economic reforms. There is little meaning to the concept of economic growth in the absence of distribution.

Poverty is a serious issue in Georgia and poverty headcounts are high in all regions of Georgia. Approximately half the population cannot achieve a level of consumption above the NHDR Poverty Line suffering serious constraints for the provision of food and non-food items. More than 25% of the population of Georgia is eating less than the minimum requirements while more than 10% of our fellow citizens are just hungry and in a desperate

situation. Poverty in Georgia shows a marked regional and seasonal component. Poverty varies significantly among regions and between seasons. The poor fare badly in winter as they are oblige to decrease expenditures on food in order to make room for expenditures in commodities like heating.

Sources of income for the general population are highly unstable. Formal salaries account for between one third and one half of total family income. Pensions, gifts and debts are important components of monthly income. More than half the family income depends on sources that are highly unreliable (pensions and gifts), and/or seasonal (selling food products) and/or that have to be repaid in the future (loans).

Pensioners are a highly vulnerable group. There is no need to lower poverty lines to realize this. Pensioners show the highest poverty risk at any poverty line. Even at the risk of being repetitive, it should be re-emphasized that failure to pay pensions can constitute devastating blows to a group with hardly additional means of subsistence.

At the national level, males and females have similar poverty risk though in some regions disparities are noticeable. In Adjara, males are more likely to be poor than females while in Kakheti and Mtskheta-Mtianeti the reverse holds true. Mtskheta-Mtianeti is a case in question with females being four times more likely to be poor than males. There are also gender disparities in salaries probably due to barriers in the access to well paid jobs by females.

Those with tertiary education show the lowest risk of poverty while those with elementary and uncompleted secondary education show the highest. However, high education is not a safe passport to higher standards of living. The group of the unemployed is also characterized by high levels of education with 2/3 of them having post-secondary studies. For a further discussion about the contribution of high education to family well-being see Technical Annexes 1 and 2.

Even for those that are employed, subsistence can mean overcoming great difficulties. Salaries from the primary job are regularly below the subsistence income while 40% of people with a secondary job receive in-kind payments and no monetary remuneration at all.

Except for a tiny proportion of the population, access to health care services is a difficult thing. Poor and non-poor alike face severe restrictions to access medical assistance. Lack of money stands as the overwhelming reason keeping people out of private doctors and public hospitals. For every person that manages to visit a doctor, there are two that could not simply because of lack of means. For the extremely poor, the situation is desperate. 3 out of 4 people in need of medical assistance cannot receive a complete treatment (visit a doctor and buy the required medicines).

Distribution of consumption in Georgia is highly biased. Education is the most unequal item with a Gini Index close to total inequality. The level of inequality in Georgia should be a matter of concern as it erodes the legitimacy of economic reforms.