All tables and figures with source cited as "UNDP" indicate that data originate in UNDP-supported aurveys.See Technical Annex6 for a description of sampling procedure

See for example, "Georgia: Poverty Update 2001". Report No. 22350-GE. World Bank, Washington DC

See Georgian Economic Trends; Second and Third Quarters of 2001.

ILO "loose" methodology counts discouraged workers as unemployed while the "strict" methodology does not

As reported in several UNDP-sponsored round tables with the public

See Technical Annex 5 for a description of the computation of the HDI as well as the treatment of income in the computation of the GDP index

See Human Development Report, 2001. "Making New Technologies Work for Human Development". United Nations Development Programme. New York. Oxford University Press. Available at www.undp.org

In view of lack of data of GDP for Georgian regions, some adaptations to the original method for computing the HDI have been made. These changes are described in Technical Annex 5

In order to see these changes, however, it is necessary to use spending per capita adjusted for PPP as the basis for the GDP index instead of the method described in Technical Annex 5 (regional shares in total consumption). This is because we lack reliable figures of national GDP for winter and summer quarters from which shares could be calculated

Poverty lines for other groups (e.g. working age female; children, etc) are derived from the poverty line of a working age male

We have chosen this poverty food basket as a starting point since it contains the products most widely used by poor Georgian families. We have increased some of its contents and added others until a chosen level of calorie and nutrient intake was attained. For a description of the formulation of the food basket in the Recommended Poverty Line see "Georgia: Poverty and Income Distribution". Report No 19348-GE (volume II technical papers). World Bank, Washington D.C. 1999

See "Georgia: Poverty and Income Distribution". Report No 19348-GE (volume II technical papers). World Bank, Washington D.C. 1999

Therefore the non-food component of the poverty line can vary between winter and summer because of different heating costs in different seasons

One notices that poverty lines may also vary between summer and winter because of expenses on education

See World Bank, Uzbekistan: Adjusting Social Protection. Report No 13023-UZ. Cited in Lanjow, Jean Olson. "Demystifying Poverty Lines". Paper available at the UNDP web page http://www.undp.org/poverty/publications/pov_red/Demystifying_Poverty_Lines.pdf

The use of fixed shares can also lead to seriously wrong conclusions depending on the composition of the family and the distribution of expenses among family members. See for example Deaton A. and Muellbauer "On Measuring Child Costs: With Applications to Poor Countries". Journal of Political Economy Vol. 94 (4) 1986

Other methods also exist. For instance, one can determine a non-food poverty line by observing the value of non-food expenditures of those households whose total consumption equals the cost of the food basket. This amount is considered to represent the essential non-food consumption because the household is trading off consumption of an already minimal food basket. The application of this method is also heavily dependent on the definition of the food basket and the resulting non-food allowance may or may not be compatible with local perceptions of minimum non-food needs. See Ravallion, M. Bounds for a Poverty Line. Mimeo. Policy Research Department, World Bank, Washington DC. 1995

Insulation, of course, comes at a cost. However, household items such old newspapers and pieces of cloth can provide acceptable insulation to windows and door cracks, which constitute the most important sources of heating loss

Economies of scale are usually accounted for by the parameter Theta. This takes the form of nq where n is the number of family members and q represents the parameter of economies of scale. For the NHDR poverty line, the implicit economies of scale in non-food items are approximately equal to a theta of 0.50. If food and non-food components are taken together, the implicit theta of the NHDR poverty line would be approximately 0.65. Regression analysis that controls for family size and composition is used to estimate the value of theta. See Lanjow, Peter and Martin Ravallion "Poverty and Household Size". The Economic Journal 105. Nov. 1995; see also Lanjow, Peter, Branko Milanovich, and Stefano Paternostro. "Poverty and Economic Transition: How do Changes in Economies of Scale Affect Poverty Rates of Different Households?" Working paper. Development Economic Research Group. The World Bank. Washington DC

See Table A4.4 in Technical Annex 4 for a description of poverty thresholds for families of 1-4 members under different poverty lines

Even though the method of fixed shares and the direct definition of the non-food poverty basket would have yielded similar NHDR poverty lines (approximately 6 GEL difference between them), we have chosen to use the direct approach, even at the risk of exposing ourselves to criticisms of paternalism and arbitrariness. We believe that for the purpose of this study the direct definition of non-food components provides a clearer understanding of the kind of life being lived by those whom fall below this or that poverty line. If this objective is accomplished, then the risk is worth taking

The poverty gap, which is also called the depth of poverty, is the mean distance separating the population from the poverty line. If the poverty gap for Georgia is 0.18 it would mean that the cash transfers to lift each family out of poverty would represent 18% of the poverty line. If the mean income of Georgia had been the same as the poverty line, then the cash transfer would amount to 18% of the country’s mean income.

While the poverty gap takes into account the distance separating the poor from the poverty line, the yseverity of poverty takes the square of that distance into account. It is important to apply the poverty depth and the severity of poverty as complementary indicators to poverty headcounts. Note that the underlying evaluation of the poverty headcount would consider as more effective policies those that lift families closer to the poverty line out of poverty. This effect would arise since the transfers to those close to the poverty line are lower than for those far away from it. On the basis of the poverty gap and severity of poverty one can put emphasis on helping those at the bottom of the poor group.

Note that the mean per capita spending for urban and rural non-poor is much greater than the median indicating that the distribution of spending has a longer tail to the right. Or in other words, that the figure for average spending is being pushed upwards by the contribution of a relatively small number of people with high levels of spending. In contrast, mean and medians are of similar values for all groups falling below any poverty line for winter or summer in urban or rural areas.

The percentage of poor in a selected group can also be understood as the risk of poverty in that group. For instance, there is a 40.06% chance that a (randomly selected) couple without children will be poor.

26

Easterly, William. "The Effect of International Monetary Fund and World Bank programs on Poverty". Mimeo. World Bank. Washington DC. 2000; pp2

27

See for example, UNICEF, The State of the World’s Children, 1989. Oxford University Press. 1989

28

The transfer of resources from one part of the world to another was massive. Between 1982 and 1985, Latin America paid out net annual transfers of an average 4.2% of GDP, almost double what vanquished Germany paid to the victorious Allies after World War I. See Sachs, Jeffrey and Felipe Larrain. Macroeconomics in the Global Economy. Prentice Hall, Englewood Cliffs, New Jersey USA. 1993.

29

ESAFs allowed countries to receive assistance over a period of up to three years with reimbursement stretched out over a period of 10 years at an interest rate of 0.5%.

See Technical Annex 3 for a summary of strategy, policy instruments and tools of the Poverty Reduction and Economic Growth Programme for Georgia (version October 2001).

There is an interesting debate in the literature about the effectiveness of structural adjustment programs. Those interested in further reading can see (i) Schadler, Susan at al. "IMF conditionality: Experience under Stand-By and Extended Arrangements". IMF Occasional Paper No 128. International Monetary Fund. Washington DC. 1995; (ii) Conway, Patrick 1994. "IMF Lending Programs: participation and impact". Journal of Development Economics 45(2). 1994; (iii) Mireaux, Louis et al. "Evaluating of IMF lending to low-income countries". Journal of Development Economics 61. 2000; (iv) Easterly, William "The Effect of International Monetary Fund and World Bank Programs on Poverty". World Bank mimeo. Washington DC. 2000; (v) Garuda, Gopal. "The Distributional Effects of IMF Programs: a cross country analysis". World Development 28(6). 2000; (vi) Killick, T. et al "What can we know about the effect of IMF programmes?" World Economy 16. 1993; (vii) Pastor, M. "The Effects of IMF Programs in the third world: debate and evidence from Latin America". World Development 15. 1987.

Examples are Brazil and Pakistan. See for instance (i) World Bank. Brazil: Public Spending on Social Programs: issues and options. Vol. I and II. Washington DC. 1988; (ii) Malik, Sohail "Poverty in Pakistan: 1984-85 to 1987-88" in Michael Lipton and Jacques van der Gaag, eds. Including the Poor, Washington DC. World Bank.

See Dreze, Jean and Amartya Sen. Hunger and Public Action. Oxford, Clarendon Press. 1990

However, see Dollar, David and Art Kraay. Growth is Good for the Poor. Development Research Group. World Bank. Washington DC. 2000.

See also Danziger Sheldon and Gottschalk P. "Do Rising Tides Lift all Boats? The impact of secular and cyclical changes on Poverty". American Economic Review 76(2). 1986.

See Lyn Squire. "Fighting Poverty". American Economic Review, 83(2). 1993

Anand, Sudhir and Ravallion M. "Human Development in Poor Countries: on the role of Private Incomes and Public Services". Journal of Economic Perspectives 7(1). 1993.

Leakages in the Georgian system of education can be lower than in Africa. See, for example, Ritva Reinikka and Jakov Svensson 2001. "Explaining Leakages of Public Funds". Working Paper 2709. World Bank. Washington DC.

Includes only doctors, nurses and midwives

The Gender Development Index is a composite index that comprises values for (i) the ratio of female non-agricultural wage to male non-agricultural wage, (ii) percentage share (male and female) of the economically active population, (iii) Male and Female Population, (iv) life expectancy at birth (male and female separately), (v) adult literacy rate (male and female separately), and (vi) combined gross enrolment ratio (male and female separately).

The Gender Empowerment Index is a composite index that comprises values for (i) Percentage share of Parliamentary representation, (ii) Percentage share of administrative and managerial positions, (iii) Percentage share of professional and technical positions. See Human Development Report 2001 for an explanation of how to calculate these indexes (available at www.undp.org).

The Human Poverty Index II is a composite of (i) probability at birth of not surviving the age of 60, (ii) adults lacking functional literacy skills, (iii) population below income poverty line (50% of median disposable household income), (iv) long-term unemployment rate. See Human Development Report 2001 for an explanation of how to calculate the HPI-II (available at www.undp.org).

Some restrictions in the dataset were introduced. Households with total spending less than 80% of reported income were discarded. Households with spending more than 5.24 times their reported income and non-poor were also taken out of the analysis. This results in 15% loss of observations.

All results to be presented meet the condition of having residuals with a normal distribution. Variables presented pass the test of collinearity.

In brief, when using forward induction, a variable is first selected at random, its fit is observed, and then others are added until the best fit is obtained. In a sense, this is an attempt at understanding variations in family welfare without a pre-conceived "model" at hand. Whatever helps best to explain the dependent variable remains in the final model and whatever does not is eliminated. The best combination of variables remains as the final "model".

Of course, we remember that for this particular analysis, welfare is a function of consumption and the poverty line, and this says nothing about how happy those families with a large share of children aged 7-17 can be even if having a lower disposable income.

For further reading about institutions and economic development see North, Douglass, 1990. Institutions, Institutional Change and Economic Performance. Cambridge. Cambridge University Press.

A difference exists between the national average computed by the NHDR and the one presented at the beginning of this section. These are due to different values of Georgian GDP and different estimations of the Georgian population. The main purpose of the regional HDI is therefore to compare human development across regions.

For those interested in obtaining the complete description of the sampling procedure and full sampling plan, please contact the UNDP office in Georgia at fo.geo@undp.org.ge att: Ms. Nato Alhazishvili, Programme Analyst.