Armenia: Poverty Assessment (In Three Volumes) Volume III: Technical Notes and Statistics

Transcription

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

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

3 Table of Contents Table of Contents 3 ACKNOWLEDGEMENTS 4 Annex I: Poverty profile 6 (i) Household data sources and their comparability 6 (ii) Definition of consumption and poverty lines 8 (iii) Poverty profile descriptive statistics 10 (iv) Poverty regressions 25 Annex II: Non-Income dimensions of poverty 34 A. Education: technical notes and statistics 34 B. Living conditions: statistics 39 Annex III: Demographics, migration and labor market 41 (i) Labor market indicators between LFS and LSLC (ii) Statistical tables 42 Annex IV: Social protection 44 (i) Proxy-means targeting in Armenia 44 (ii) Statistical information 48 3

4 ACKNOWLEDGEMENTS 4

5 Armenia Poverty Assessments is a collective effort of a team lead by Aleksandra Posarac (ECSHD) assisted by Edmundo Murrugarra (ECSHD). The multi-sectoral team comprised: Lev Freinkman (macroeconomic developments), Julia Bucknell and Villiam Sutton (rural developments), Hellen Shahriari (subjective aspects of poverty), Dilnara Isamiddinova, and María Eugenia Bonilla Chacín (education), all from the World Bank, and consultants Gorana Krsitc (poverty profile, labor market), Paulette Castel (rural developments, private transfers), Stefania Rodica Cnobloch (health), Artsvi Khachatryan (macroeconomic developments) and Hjalte Sederlof. Following a series of team discussions and consultations with the client, this Report was put together by Aleksandra Posarac and Hjalte Sederlof. A number of people contributed by commenting on various drafts and generously providing information: Mark Lundell, Ellen Hamilton, Julian Lampietti, Susanna Hyrapetyan, Peter Nicholas, Mansoora Rashid, Toomas Palu, Toby Linden, and Jan Rutkowski. The team owes a special gratitude to the National Statistical Service of Armenia, as well as UNICEF and UNDP Offices in Armenia for cooperation and generous sharing of information. 5

6 Annex I: Poverty profile (i) Household data sources and their comparability The poverty profile is based on the latest available data set for Armenia, Integrated Living Conditions Survey (ILCS) conducted throughout year The sample size included 4,037 households. The questionnaire comprises sections on household composition (including limited information on labor market status), housing conditions, migration and transfers between households, education, health, agriculture, savings and borrowing, and social assistance. It includes also a diary of expenditures and consumption during last 30 days and a section on annual consumption. The survey provides the basis for comparison of poverty measurements with the 1998/99 estimates which were based on ILCS 6

7 carried out in the period between July 1998 and June Previous comparisons of poverty indicators between 1996 and 1998/99 were limited due to differences in the welfare measure available for poverty analysis and time period of the survey. 1 Table I.A.1 provides information on sample size, period of the survey, sampling frame and design, and questionnaire design of the two latest surveys that are used for poverty comparisons between 1998/99 and Given that both surveys were carried over a 12-month period, seasonal fluctuations will not affect comparisons of poverty measurement results. The main drawback of both surveys is the sampling frame, which is based on the Armenian households addresses data base complied in 1996 by the National Statistical Service. Because of intense internal and external migration, the sample is believed to be outdated. 2 The ILSC 2001 sample design was affected by new sampling strategies across urban, rural areas and the capital city of Yerevan. In addition, we found that urban areas were over-represented in the 2001 ILCS data. This conclusion is based on the assumption that the urban/rural structure was not changed significantly in between the two surveys. This is corroborated by the results of the Educational Survey conducted by the UNDP in 2001, which revealed very similar structure of urban and rural areas (Table I.A.2). Therefore, poverty estimates for 2001 used urban/rural structure from 1998/99 ILCS data. Table I.A.1: Armenia: Integrated Living Conditions Survey ILCS 1998/99 ILCS 2001 Sample size 3,600 households 4,037 households Survey period July June 1999 January 2001 December 2001 Sampling frame 1996 data base on households addresses 1996 data base on households addresses Sampling design representative at urban/rural levels representative at Yerevan/other urban/rural levels Questionnaire design - all information for constructing consumption aggregate exist; - a module on employment exists; - a separate module for self-employed. - all information for constructing consumption aggregate exist; - a module on employment does not exist; - a separate module for self-employed does not exist; The 2001 survey contains all information necessary for constructing consumption aggregate comparable with consumption aggregate used in 1998/99 poverty analysis (World Bank, 2002a). While the questionnaire was kept relatively similar in both surveys, some differences between some modules are observed. In contrast to 1998/99 ILCS, the 2001 round has limited information necessary for imputation of the rental value of housing in constructing the consumption aggregate. Also, information on the employment status of individuals is limited in 2001 round (with no module on selfemployment household members), which might affect comparison of poverty by employment categories between the two surveys. Given the fact that the new Labor Force Survey (LFS) in Armenia was conducted in 2001 with the six-month rotation along with the ILCS, further insight into relationship between poverty and labor market can be provided using the latest available LFS data. Table I.A.2: Urban-rural structure by data source ILSC Education Survey 99/ Urban 60.56% 67.90% 58.07% Rural 39.44% 32.10% 41.93% Source: ILCS 1998/99, 2001 and UNDP APEAS In order to achieve comparability in income aggregate between 1998/99 and 2001, we re-calculated household income for 1998/99. In the latest survey, data on labor earnings and on income from selfemployment were only available from the dairy, because the separate modules on employment and self-employment do not exist like in 1998/99. Therefore, both sources of labor income were re- 1 For further details on comparability between 1996 and 1998/99 see Armenia Poverty Update (World Bank, 2002a). 2 However, it is the only available sampling frame. The SDS is currently developing a new data base of households addresses based on the latest 2001 Census. 7

8 computed for 1998/99 and were based on the information from the diary. Two other income sources re-estimated for 1998/99 were remittances and government transfers. Although detailed module on transfers between households (internal and external) exists in both surveys, we noticed considerable difference in the reported amount of remittances between the separate module and a diary. Surprisingly, amounts from the diary appeared higher by around a third in 1998/99 and by around a half in 2001 and much closer to the official estimates for remittances. This is the reason why remittances from the 1998/99 survey were re-computed, as they were initially based on the information from the module on transfers between households. Finally, we make government transfers comparable between two surveys using the same reference period for the reported amounts of transfers. 3 (ii) Definition of consumption and poverty lines The consumption aggregate: The consumption aggregate was estimated for the first time in Armenia using the 1998/99 Integrated Living Conditions Survey. This study uses the same definition of welfare measure as adopted for the 1998/99 Armenia Poverty Update (World Bank 2002a). The only difference between consumption aggregate in 1998/99 and in 2001 is dwelling rental value which was not imputed in the 2001 consumption aggregate due to a small number of observations of households who paid rent. 4 Thus, for the purpose of comparisons of poverty measures, consumption aggregate for 1998/99 was re-calculated excluding the imputed rental value of dwellings. The components of the consumption aggregate for 2001 are: (1) value of food and non-food consumption which also includes consumption from home production, aid received from humanitarian organizations and other sources; and (2) rental value of durable goods. The non-food consumption comprises the following categories: clothing and shoes, household goods, transportation, utilities, education, health, and the rental value of durable goods. 5 Given the fact that the 2001 ILCS was carried out throughout the year, the value of consumption from different quarters in 2001 was adjusted for inflation over the observed period. This price adjustment which also takes into account urban/rural price differences was applied for food consumption due to different food price changes between rural and urban areas. Since the National Statistical Service of Armenia does not distinguish between urban and rural food prices, the survey data were used for price adjustment over time and across regions for food items. Food consumption is expressed in the 2001 Autumn-urban price levels. Factors for price adjustments of food consumption are reported in Table I.A.3. 6 The non-food consumption is adjusted for inflation, however, using the official Consumer Price Index for appropriate non-food expenditure sub-groups provided by the National Statistical Service. 7 Total consumption aggregate was expressed in 2001 autumn price levels. Table I.A.3: Factors for price adjustment of food consumption (multiplied by 100) Urban Rural Quarter, 2001 Laspeyres Paasche Fischer Laspeyres Paasche Fischer January-March April-June Although some reported amounts of transfers (pensions and other transfers) refer to different time period (one day, one week, one month or periodically), we find that all reported figures are actually on monthly basis. 4 In the 2001 ILCS, there is no question on rent payment in advance (before the start of the survey month) like in the 1998/99 Survey, which may extend the size of the sub-sample. The sub-sample of households who paid rent amounted to only 3 percent, which was a very small sample size for estimating hedonistic rental equations. 5 For detailed explanation of consumption aggregate and its components see Armenia Poverty Update (World Bank, 2002a). 6 Fisher CPI index was used to make consumption comparable at the Autumn-urban price levels. While Fisher index based on survey data on food shows that annual inflation was 6.4% in urban areas and 13.3% in rural areas, the official overall CPI index for the same period amounted to 97 and CPI for food amounted to For discussion of methodology used see Armenia Poverty Update (World Bank, 2002a). 8

9 July-September October December Implicit annual inflation -5.6% -7.3% -6.4% -10.0% -16.4% -13.3% (Survey data) Source: ILSC Note: Factors convert food expenditures into amounts comparable with urban areas during the last Survey quarter (October - December 2001). Food consumption values from different households were multiplied by these factors for the corresponding poverty analysis. Equivalence scales: The consumption aggregate is standardized by the number of adult equivalent members which are based on equivalence scales and size economies. Total consumption was divided by adjusted per-equivalent consumption suggested by Deaton and Zaidi (1999) 8 where typical national household is not affected by changes in α and θ parameters. 9 The estimates of equivalence scales (α=0.68) and scale economies (θ=0.75) estimated and applied in 1998/99 poverty assessment were used, hence avoiding changes in poverty indicators due to changes in those parameters. Typical Armenian household is a five-member household with three adults and two children. Poverty lines: The 2001 poverty lines are based on re-estimated 1998/99 poverty lines. The food poverty line was estimated for the first time using the 1998/99 Survey data and it was based on foodenergy intake method. This method finds the level of equivalent household expenditures that is associated with the household attaining the minimum recommended energy intake from food consumption. The cost of 2,100 calories consumption food basket is estimated at 8,730 drams per adult equivalent per month (291 dram per adult equivalent per day), which was used as 1998/99 extreme poverty line. This line is then adjusted for inflation over the period Spring Autumn 2001 using the official CPI for food provided by Armenia National Statistical Service and expressed in Autumn 2001 values, thus obtaining 2001 extreme (food) poverty line. The real value of food poverty line for 2001 declined relative to 1998/99 because of the negative food inflation between these two periods. It was estimated at 7,979 drams per adult equivalent per month (table below). Table I.A.4: Armenia: Poverty lines in 1998/99 and 2001, in drams Original Re-estimated / /99 Extreme (food) poverty line 8,730 8,730 7,979 Complete poverty line 12,306 12,276 11,221 Source: World Bank 2002a; ILCS 1998/99 and In order to obtain complete poverty line for 2001, which comprises food poverty line and non-food allowance, first we re-estimated 1998/99 complete poverty line, or more precisely, its non-food component. Non-food allowance for 1998/99 complete poverty line was estimated using the Food Expenditure Method (World Bank, 2002a). According to this method, the non-food share is estimated as the non-food share of those households whose food consumption is around the food line. Namely, non-food share in total consumption in 1998/99 was re-estimated due to the fact that original consumption aggregate for 1998/99 was also re-estimated in order to exclude imputed rental value of dwelling. These adjustments were necessary for the purpose of comparisons of poverty measures between 1998/99 and The new share of non-food consumption was estimated at 28.9 percent of the total minimum consumption. Thus, the re-estimated value of complete poverty line for 1998/99 was 12,276 drams per adult equivalent per month. Assuming unchanged structure of poverty line in 1998/99 (non-food share of 28.9 percent), we obtained a complete poverty line for 2001 of 420 drams per adult equivalent per day or 11,221 drams per month. Comparing poverty between 1998/99 and 2001: Poverty indicators for 2001 cannot be directly compared with the previous 1998/99 poverty estimates. The main reason for this limitation lies in the 8 They argued that per adult equivalent consumption measure overestimates total consumption in all household types except in single-adult households. 9 For further description of equivalence scales and size economies used see Armenia Poverty Update (World Bank, 2002a). 9

10 fact that the original consumption aggregate for 1998/99 reported in Armenia Poverty Update (World Bank, 2002a) included imputed rental value of dwelling in contrast to the consumption aggregate for The imputation of rental value of housing in the 2001 consumption aggregate was not possible due to a small number of households that paid rent. Therefore, consumption aggregate and complete poverty line for 1998/99 were re-estimated to exclude imputed rental value of dwelling and new poverty estimates for 1998/99 were obtained. There are no significant differences between original and re-estimated 1998/99 poverty indicators. 10 (iii) Poverty profile descriptive statistics Table I.B.1: Armenia Poverty Indicators 1998/99 (standard errors in parenthesis) Extreme (Food) Poverty Line (8,730 drams) Complete Poverty Line (12,306 drams) Incidence (P0) Gap (P1) Severity (P2) Incidence (P0) Gap (P1) Severity (P2) Total 25.4% 5.5% 1.8% 53.7% 15.5% 6.1% (0.81) (0.23) (0.10) (0.92) (0.36) (0.19) Urban 31.2% 7.0% 2.3% 60.4% 18.4% 7.6% (1.11) (0.33) (0.14) (1.14) (0.49) (0.27) Rural 17.7% 3.4% 1.1% 44.8% 11.6% 4.2% (1.14) (0.30) (0.13) (1.46) (0.51) (0.26) 10 Table A2.2 in Annex II reports poverty measurements in 2001 using the structure of urban and rural areas from the 2001 ILCS. 10

11 Source: Armenia Poverty Update (World Bank, 2002a). Note: Consumption aggregate includes imputed rental value of dwelling. Incidence (P0) Table I.B.2: Armenia Poverty Indicators in 98/99 and 2001 (standard errors in parenthesis) Extreme (Food) Poverty Line (8,730 drams in 98/99) (7,979 drams in 2001) Gap (P1) Severity (P2) Incidence (P0) Complete Poverty Line (12,276 drams in 98/99) (11,221 drams in 2001) Gap (P1) Severity (P2) Total 98/ % 6.0% 2.0% 54.8% 16.2% 6.5% (1.94) (0.53) (0.21) (2.26) (0.96) (0.48) % 4.7% 1.6% 48.3% 13.2% 5.2% (1.42) (0.38) (0.16) (2.38) (0.74) (0.35) Urban 98/ % 7.8% 2.6% 61.4% 19.3% 8.1% (2.25) (0.66) (0.28) (2.55) (1.13) (0.59) % 5.0% 1.7% 48.5% 13.7% 5.5% (1.60) (0.36) (0.15) (2.90) (0.82) (0.35) Rural 98/ % 3.7% 1.2% 46.1% 12.1% 4.5% (2.45) (0.54) (0.21) (3.08) (1.18) (0.54) % 4.0% 1.5% 47.9% 12.1% 4.6% (3.04) (0.83) (0.34) (4.29) (1.61) (0.77) Note: Consumption aggregate does not include imputed rental value of dwelling. Poverty indicators for 2001 are based on 2001 urban/rural structure. Standard errors computed with PSU adjustments. Source: ILCS 1998/99 and

12 Figure I.B.1. Rural and Urban Cumulative Consumption Distribution in Armenia, Log per adult equivalent consumption Note: The higher line refers to rural and lower to urban cumulative consumption distribution. Source: ILCS

13 Table I.B.3: Poverty and extreme poverty incidence by regions in 98/99 and 2001 (standard errors in parenthesis) Extreme poverty incidence, % Poverty incidence, % 98/ / Aragatzotn (3.50) (2.32) (3.72) (2.06) Ararat (2.31) (2.00) (3.05) (4.18) Armavir (1.89) (4.92) (2.76) (5.21) Gegharkunik (2.41) (9.79) (3.31) (14.5) Lori (2.95) (3.41) (2.93) (3.17) Kotayk (2.94) (5.22) (3.06) (10.50) Shirak (2.91) (5.40) (2.33) (12.31) Syunik (3.78) (1.19) (4.09) (4.66) Vayots Dzor (4.78) (10.4) (5.90) (9.55) Tavush (3.79) (1.30) (4.58) (1.27) Yerevan (1.60) (0.00) (1.66) (0.00) Other urban (1.58) (2.58) (1.55) (4.07) Rural 18.7% % 47.9 (1.16) (3.04) (1.46) (4.29) Total (0.82) (1.42) (0.92) (2.38) Source: ILSC 1998/99 and Note: Consumption aggregate does not include imputed rental value of dwelling. Poverty indicators for 2001 are based on 2001 urban/rural structure. Standard errors computed with PSU adjustments. Table I.B.4: Age Groups by Quintiles in 2001, in % Poorest Total Composition of quintiles by age groups Children Aged Aged Aged Aged Aged Aged Total Composition of age groups by quintiles Children Aged Aged Aged Aged Aged Aged Total Source: ILCS

14 Table I.B.5: Household size by quintiles in 2001, in % Poorest Total Composition of quintiles by household size Number of household members or more Total Composition of household size by quintiles Number of household members or more Total Source: ILCS Table I.B.6: Poverty by education of the household head in 98/99 and 2001 Extreme poverty incidence, % Poverty incidence, % 98/ / Primary Incomplete Secondary Complete Secondary Technical Higher Education Total Source: ICLS 98/99 and Table I.B.7: Education of the household head by quintiles in 2001, in % Poorest Total Composition of quintiles by education of the household head Primary Incomplete Secondary Complete Secondary Technical Higher Education Total Composition of education of the household head by quintiles Primary Incomplete Secondary Complete Secondary Technical Higher Education Total Source: ILCS

15 Table I.B.8: Poverty by labor force participation of the household head; 1998/99 and 2001 Extreme poverty incidence, % Poverty incidence, % 98/ / Non-participants Seasonally unemployed Unemployed Salaried worker Self-employed Other employment Total Source: ICLS 98/99 and Table I.B.9: Composition of Quintiles by Labor Force Participation in Armenia, in % (population over 16) Poorest Total All Participants Unemployed Salaried worker Self-employed Other employment Non-participants Pensioners Students Other non-part Total Yerevan Participants Unemployed Salaried worker Self-employed Other employment Non-participants Pensioners Students Other non-part Total Other urban Participants Unemployed Salaried worker Self-employed Other employment Non-participants Pensioners Students Other non-part Total Rural Participants Unemployed Salaried worker Self-employed Other employment Non-participants Pensioners Students Other non-part Total Notes: Quintiles defined as per adult equivalent consumption quintiles. Source: ILCS

16 Table I.B.10: Labor Force Participation in Armenia by Quintiles, in % (population over 16) Poorest Total All Participants Unemployed Salaried worker Self-employed Other employment Non-participants Pensioners Students Other non-part Total Yerevan Participants Unemployed Salaried worker Self-employed Other employment Non-participants Pensioners Students Other non-part Total Other urban Participants Unemployed Salaried worker Self-employed Other employment Non-participants Pensioners Students Other non-part Total Rural Participants Unemployed Salaried worker Self-employed Other employment Non-participants Pensioners Students Other non-part Total Notes: Quintiles defined as per adult equivalent consumption quintiles. Source: ILCS

17 Table I.B.11: Consumption Components in Armenia by Quintiles in 98/99 (drams per month, Spring 1999 prices) Poorest Total All Households Food 4,428 6,408 8,112 10,169 14,318 8,726 Clothing and shoes , Household goods , Transportation , Utilities , Health , Education , Rental value of durables 1,086 1,395 1,571 1,778 2,147 1,600 Average 6,282 9,112 11,677 15,114 27,058 13,928 Yerevan Households Food 4,135 5,878 7,339 9,141 12,661 7,855 Clothing and shoes , Household goods , Transportation , Utilities , Health , Education , Rental value of durables 1,135 1,513 1,788 1,945 2,422 1,763 Average 6,300 9,108 11,683 14,904 27,439 14,069 Other Urban Households Food 4,364 6,227 7,959 9,714 12,792 7,542 Clothing and shoes , Household goods , Transportation , Utilities , Health , Education ,096 2, Rental value of durables 1,144 1,490 1,656 1,837 2,246 1,589 Average 6,223 9,104 11,704 15,073 27,150 12,260 Rural Households Food 4,906 7,018 8,796 10,951 16,071 10,309 Clothing and shoes , Household goods Transportation Utilities Health , Education , Rental value of durables 922 1,209 1,345 1,660 1,921 1,486 Average 6,364 9,120 11,650 15,245 26,764 15,140 Notes: Consumption measured per adult equivalent. Quintiles defined as per adult equivalent consumption quintiles. Source: ILCS 1998/99. 17

18 Table I.B.12: Consumption Components in Armenia by Quintiles in 2001 (drams per month, Spring 1999 prices) Poorest Total All Households Food 4,064 5,575 6,551 8,021 11,531 7,231 Clothing and shoes , Household goods , Transportation , Utilities , Health , Education , Rental value of durables 1,192 2,159 2,755 3,246 3,596 2,618 Average 6,162 9,246 11,543 14,525 25,042 13,506 Yerevan Households Food 3,458 4,633 5,667 6,909 10,877 6,785 Clothing and shoes , Household goods , Transportation , Utilities ,036 1,896 1,078 Health , Education ,305 2,727 1,268 Rental value of durables 1,372 2,479 3,006 3,216 3,589 2,809 Average 6,040 9,202 11,533 14,547 26,750 14,993 Other Urban Households Food 4,164 5,584 6,563 7,947 10,984 6,852 Clothing and shoes , Household goods , Transportation Utilities , Health , Education , Rental value of durables 1,304 2,277 2,862 3,443 3,921 2,702 Average 6,271 9,274 11,535 14,445 23,144 12,422 Rural Households Food 4,428 6,119 7,058 8,701 12,571 7,860 Clothing and shoes , Household goods , Transportation Utilities , Health , Education , Rental value of durables 931 1,880 2,520 3,095 3,384 2,414 Average 6,137 9,250 11,557 14,580 24,581 13,379 Notes: Consumption measured per adult equivalent. Consumption values expressed in Fall 2001 prices were divided by CPI between Spring 1999 and Fall 2001 (99.2) in order to obtain consumption in Spring 1999 prices. Quintiles defined as per adult equivalent consumption quintiles. Source: ILCS

19 Table I.B.13: Composition of Quintiles by Consumption Components in Armenia in 2001 (drams per month, Spring 1999 prices) Poorest Total All Households Food Clothing and shoes Household goods Transportation Utilities Health Education Rental value of durables Average Yerevan Households Food Clothing and shoes Household goods Transportation Utilities Health Education Rental value of durables Average Other Urban Households Food Clothing and shoes Household goods Transportation Utilities Health Education Rental value of durables Average Rural Households Food Clothing and shoes Household goods Transportation Utilities Health Education Rental value of durables Average Notes: Consumption measured per adult equivalent. Quintiles defined as per adult equivalent consumption quintiles. Source: ILCS

20 Table I.B.14: Consumption Components in Armenia by Quintiles in 2001 (drams per month, Spring 1999 prices) Poorest Total All Households Food Clothing and shoes Household goods Transportation Utilities Health Education Rental value of durables Average Yerevan Households Food Clothing and shoes Household goods Transportation Utilities Health Education Rental value of durables Average Other Urban Households Food Clothing and shoes Household goods Transportation Utilities Health Education Rental value of durables Average Rural Households Food Clothing and shoes Household goods Transportation Utilities Health Education Rental value of durables Average Notes: Consumption measured per adult equivalent. Quintiles defined as per adult equivalent consumption quintiles. Source: ILCS

21 Table I.B.15: Average Household Income Sources in Armenia by Quintiles in 98/99 (drams per month, Spring 1999 prices) Sources Total All Households Labor earnings 5,695 10,751 12,079 17,688 20,558 13,408 Self-employment 1,201 2,546 9,213 5,491 17,500 7,271 Farm Income 3,868 7,483 10,936 13,548 20,011 11,235 Remittances 3,081 4,393 6,771 6,228 12,596 6,657 Transfers 3,698 4,068 3,082 2,981 2,504 3,262 Pensions 2,312 2,954 2,335 2,205 1,725 2,305 Other transfers 1,386 1, Assets sold Total 18,441 30,682 42,523 47,454 74,051 42,866 Yerevan Labor earnings 11,007 18,263 22,581 32,732 38,745 24,583 Self-employment 1,898 6,063 7,511 4,802 33,335 11,316 Farm Income Remittances 3,933 6,573 9,706 12,176 18,709 10,270 Transfers 4,305 3,669 3,674 4,098 2,944 3,713 Pensions 3,186 2,743 2,962 3,171 2,094 2,805 Other transfers 1, Assets sold , Total 22,027 35,272 44,217 59,917 96,225 51,940 Other Urban Households Labor earnings 4,844 11,773 13,211 28,649 21,698 14,327 Self-employment 920 1,458 15,648 3,540 14,470 6,433 Farm Income 567 1,055 2,610 4,281 4,266 2,222 Remittances 3,429 5,917 8,579 9,455 22,407 8,581 Transfers 3,838 5,353 3,379 2,859 2,731 3,748 Pensions 2,178 3,550 2,170 1,835 1,325 2,295 Other transfers 1,660 1,803 1,208 1,024 1,406 1,453 Assets sold Total 15,105 29,452 44,132 49,132 66,445 36,870 Rural Households Labor earnings 538 3,492 3,597 3,909 7,902 4,340 Self-employment ,010 6,912 8,177 4,913 Farm Income 14,306 19,348 25,555 25,335 39,763 26,478 Remittances 1,407 1,185 3,144 1,384 4,367 2,442 Transfers 2,695 3,251 2,408 2,468 2,113 2,541 Pensions 1,455 2,598 2,025 1,906 1,647 1,939 Other transfers 1, Assets sold Total 19,805 27,913 39,953 40,073 62,412 40,821 Notes: Quintiles defined as per adult equivalent consumption quintiles. Source: ILCS

22 Table I.B.16: Average Household Income Sources in Armenia by Quintiles in 2001, (drams per month, Spring 1999 prices) Sources Poorest Total All Households Labor earnings 7,081 8,772 10,851 15,138 28,249 14,282 Self-employment 1,195 2,815 6,977 4,808 11,870 5,646 Farm Income 4,149 8,166 10,058 12,243 15,139 10,090 Remittances 3,614 8,152 5,115 8,235 13,023 7,723 Transfers 4,644 4,436 3,953 3,779 3,669 4,082 Pensions 3,047 2,940 2,803 2,790 3,038 2,922 Other transfers 1,597 1,496 1, ,159 Assets sold Total 20,873 32,429 37,099 44,321 72,191 41,981 Yerevan Households Labor earnings 9,852 14,100 14,515 24,879 49,340 25,415 Self-employment 2,209 5,273 7,218 7,401 11,538 7,210 Farm Income ,636 1,479 1,702 1,328 Remittances 3,967 4,195 9,105 7,631 15,091 8,752 Transfers 4,968 4,904 4,569 3,815 3,372 4,225 Pensions 3,591 3,766 3,453 3,113 2,919 3,319 Other transfers 1,376 1,138 1, Assets sold Total 21,905 29,496 37,242 45,352 81,415 47,154 Other Urban Households Labor earnings 7,123 10,010 11,756 18,419 22,122 13,566 Self-employment 1,233 2,850 6,487 4,278 11,711 4,982 Farm Income 2,134 3,509 4,367 7,488 8,418 5,065 Remittances 4,156 13,832 4,175 6,186 9,208 7,266 Transfers 5,118 4,969 4,234 4,290 3,475 4,459 Pensions 3,146 2,688 2,867 2,987 2,655 2,887 Other transfers 1,972 2,281 1,367 1, ,572 Assets sold Total 20,074 35,299 31,106 40,844 55,102 35,516 Rural Households Labor earnings 4,859 4,731 7,961 6,948 10,716 7,087 Self-employment 356 1,359 7,236 3,807 12,318 5,101 Farm Income 9,020 15,920 19,647 22,244 33,473 20,368 Remittances 2,754 6,127 3,537 10,294 13,472 7,380 Transfers 3,881 3,759 3,362 3,328 4,105 3,670 Pensions 2,513 2,652 2,369 2,444 3,419 2,674 Other transfers 1,368 1, Assets sold Total 20,918 31,948 41,900 46,669 74,239 43,698 Notes: Quintiles defined as per adult equivalent consumption quintiles. Income values expressed in Fall 2001 prices were divided by CPI between Spring 1999 and Fall 2001 (99.2) in order to obtain income in Spring 1999 prices. Source: ILCS

23 Table I.B.17: Household Income Sources in Armenia by Quintiles in 2001, in % Poorest Total All Households Labor earnings Self-employment Farm Income Remittances Transfers Pensions Other transfers Assets sold Average Yerevan Households Labor earnings Self-employment Farm Income Remittances Transfers Pensions Other transfers Assets sold Average Other Urban Households Labor earnings Self-employment Farm Income Remittances Transfers Pensions Other transfers Assets sold Average Rural Households Labor earnings Self-employment Farm Income Remittances Transfers Pensions Other transfers Assets sold Average Notes: Quintiles defined as per adult equivalent consumption quintiles. Source: ILCS

24

25 (iv) Poverty regressions Table I.C.1: Differential effects of determinants of ln(consumption) (estimated coefficients and standard errors for quantile regressions) 10% 25% 50% 75% 90% Fraction age (0.116) (0.119) (0.116) (0.089) (0.144) Fraction age (0.095) ** (0.098) ** (0.092) ** (0.070) (0.105) Fraction age (0.099) (0.101) (0.098) (0.077) (0.121) Fraction age (0.093) (0.095) (0.092) (0.070) (0.108) Fraction age (0.065) (0.066) (0.065) (0.051) (0.080) Fraction age (0.057) ** (0.061) * (0.059) * (0.045) ** (0.069) ** ln(household size) (0.037) ** (0.038) ** (0.036) ** (0.028) ** (0.043) ** Age of head (0.001) ** (0.001) * (0.001) (0.001) (0.002) Female head (0.026) ** (0.027) ** (0.027) * (0.021) * (0.033) Incomplete Secondary (0.045) (0.046) (0.045) (0.036) (0.058) Complete Secondary (0.040) ** (0.043) * (0.042) (0.033) (0.053) Technical (0.043) ** (0.045) ** (0.043) ** (0.034) * (0.055) * Higher Education (0.044) ** (0.047) ** (0.045) ** (0.035) ** (0.056) ** Non participant (0.031) ** (0.031) ** (0.030) * (0.022) (0.034) Unemployed (0.041) ** (0.042) * (0.041) * (0.032) (0.051) Selfemployed (0.041) ** (0.043) ** (0.043) (0.033) (0.055) Other employment (0.111) (0.152) (0.145) (0.119) * (0.182) % Unemployed (0.037) * (0.037) ** (0.037) ** (0.029) ** (0.046) ** % Self-employed (0.041) ** (0.043) ** (0.043) (0.034) * (0.057) * % Other employment (0.085) * (0.171) (0.177) (0.141) (0.228) Total land (0.009) (0.010) (0.009) (0.006) ** (0.008) ** % owned (0.036) (0.033) (0.032) (0.024) (0.038) % irrigated (0.041) (0.038) (0.036) (0.027) (0.043) Received credit? (0.039) (0.040) (0.040) (0.032) (0.052) ** Has livestock? (0.031) ** (0.029) ** (0.029) ** (0.022) ** (0.034) ** Aragatzotn (0.048) (0.050) ** (0.051) ** (0.041) ** (0.064) ** Ararat (0.041) ** (0.041) (0.042) (0.033) ** (0.053) ** Armavir (0.040) (0.039) * (0.038) ** (0.029) ** (0.046) ** 25

26 Differential effects of determinants of ln(consumption) (estimated coefficients and standard errors for quantile regressions) 10% 25% 50% 75% 90% Gegharkunik (0.052) * (0.049) ** (0.048) ** (0.037) ** (0.059) ** Lori (0.036) (0.035) ** (0.035) ** (0.028) ** (0.044) ** Kotayk (0.046) (0.045) (0.044) (0.034) * (0.056) ** Shirak (0.034) (0.037) ** (0.038) ** (0.031) ** (0.052) ** Syunik (0.048) * (0.050) ** (0.052) ** (0.044) ** (0.073) ** Vaiots Dzor (0.063) (0.068) * (0.065) ** (0.049) ** (0.072) ** Tavush (0.051) * (0.051) ** (0.050) ** (0.039) ** (0.059) ** Constant (0.102) ** (0.106) ** (0.105) ** (0.082) ** (0.126) ** 26

27 I.C.2: Determinants of Poverty Headcount Probit estimates Number of obs = 4037 LR chi2(46) = Prob > chi2 = Log likelihood = Pseudo R2 = poor df/dx Std. Err. z P> z x-bar [ 95% C.I. ] sage sage sage sage sage sage lhhsz hdage hdfemale* hdedseci* hdedsecc* hdedtehn* hdedhigh* hdlfp0* hdlfp1* hdlfp3* hdlfp4* slfp slfp slfp lndtot slndown slndirr agcred* aglvstk* marz1* marz2* marz3* marz4* marz5* marz6* marz7* marz8* marz9* marz10* month1* month2* month3* month4* month5* month6* month7* month8* month9* month11* month12* obs. P pred. P (at x-bar) (*) df/dx is for discrete change of dummy variable from 0 to 1 z and P> z are the test of the underlying coefficient being 0 27

28 I.C.3: Determinants of Poverty Gap Tobit estimates Number of obs = 4037 LR chi2(46) = Prob > chi2 = Log likelihood = Pseudo R2 = pgap Coef. Std. Err. t P> t [95% Conf. Interval] sage sage sage sage sage sage lhhsz hdage hdfemale hdedseci hdedsecc hdedtehn hdedhigh hdlfp hdlfp hdlfp hdlfp slfp slfp slfp lndtot slndown slndirr agcred aglvstk marz marz marz marz marz marz marz marz marz marz month month month month month month month month month month month _cons _se (Ancillary parameter)