SURVEY REPORT URBAN POVERTY AND IN-MIGRATION. Ministry of Labour and Social Welfare UNDP. Population Teaching and Research Center.

Transcription

1 URBAN POVERTY AND IN-MIGRATION SURVEY REPORT Ministry of Labour and Social Welfare UNDP Population Teaching and Research Center Ulaanbaatar, 2004

2 ACKNOWLEDGEMENTS A flow of rural - urban migration, especially migration to Ulaanbaatar, has increased in the last years in Mongolia. On the commission of the Ministry of Labor and Social Welfare with UNFPA financing, the research team of the Population Studies and Research Centre (PTRC) of the School of Economic Studies, National University of Mongolia, has carried out a survey on Urban poverty and in-migration in order to examine the impact of this process on living standards and general situation of Ulaanbaatar population. We are glad to present the survey report based on factual, qualitative and quantitative studies implemented in the course of the survey. Our team would like to express our deep gratitude to the UNFPA, who provided financial support, to international consultants Alicia Menendez and Nils Riemenscheider for their valuable professional advice and recommendations, to the Poverty Research unit consultant Richard Marshall, the senior expert of the National Statistics Office U. Tuul, the staff of the Permanent Representative Office Helene Lilleor, Martina Nikodemova and G. Uyanga, whom we worked with during the survey, to the UNV Abigail Wilmert who done proofreading the final report. Special thanks to Ulaanbaatar city administration and to all individuals in the Ministry of Labor and Social Welfare who contributed to the survey report with their valuable advice and PTRC is available to cooperate work with all of you. Our PSRC staff would also like to thank all individuals, who assisted us in successful implementation of the survey on Urban poverty and migration, Ulaanbaatar citizenrespondents of the study, state and non-government organisations, and students of the School of Economic Studies, NUM, who worked with the survey team as interviewers. Director of PTRC A. Solongo

3 EXECUTIVE SUMMARY Urban poverty and in-migration in Ulaanbaatar is a survey covering 1500 randomly selected households in Ulaanbaatar. We collected data from 4070 people from ger areas and 2777 persons from apartment areas. We gather information on education, employment, income, consumption, social relations, and migration, among other topics. The main objectives of the survey are to determine the living standards of the population in Ulaanbaatar based on their income and consumption expenditures and to study the compexity of poverty and migration. In order to reach to the main objectives the following ussues are explored: - To determine the main sources of household income and consumption expenditures; - To determine the access to and quality of social services; - To establish basic poverty indicators by type of dwelling and by migrants and nonmigrants; - To find out in Ulaanbaatar, the main reasons of migration, and the accessibility of social services for unregistered migrants; - To determine the relation between poverty and migration; - To determine characteristics of poor and identify the people in the greatest need and in order to develop policy recommendation; For each of the topics of the survey, the main questions and findings are reported below. CHARACTERISITCS OF INDIVIDUALS AND HOUSEHOLDS - In total, 59.4 percent of individuals live in ger areas while 40.6 percent live in apartment areas covered by the survey. - The overall level of education of residents of apartment areas is higher than that of residents of ger areas. - Out of people aged years about 14.6 percent migrated during the last four years. The majority of migrants (79.5 percent) live in ger areas. - Majority of all households had 3-5 members (63.4 percent), and in n most households (83.5 percent), there was one nuclear family in the household. - Out of total households, 17.9 percent live in ger dwelling. EMPLOYMENT AND INCOME Key questions - How is the employment participation rate? - How many people are unemployed? How does this look for key-subgroups? - What are main sources of income? Key answers and conclusions: - The employed participation rate is 42.4 percent.

4 - Unemployment rate is 22.9 percent. Unemployment rate is high among the population in ger area (29.5 percent), females (23.2 percent) and migrants (24.6 percent). - There is a clear link between education and employment. - The sources of income are: 67.5 percent of income comes from labour earnings, 18.2 percent from gifts, transfers etc, 8.4 percent from pensions and allowances. ACCESS TO EDUCATION AND HEALTH Key questions - Who are less educated? - Who does not have access to social services? Key answers and conclusions - Education level of men in ger areas are relatively low percent of youth aged have completed secondary education. They face a lack of possibilities to get a professional job. - One third of children from households who have migrated to suburban areas have to travel more than two kilometers to get to their schools. - Due to overcrowded classrooms in the city and lack of schools in the periphery increases the load of the secondary schools. - The rate of medical insurance coverage for informal employees, men and persons in ger areas is unsatisfactory. - Of the migrant households in ger areas, 28 percent are located 3 or more km away from a family hospital. - The focus groups point to a need to improve the quality of the social services. - More investment is needed for the construction of new schools and kindergartens in ger areas. AWARENESS, INFORMATION AND SOCIAL CAPITAL Key questions - How is the information on socio-economic project as well as the programmes advertised in the city? - How is the situation of social capital for various sub-groups Key answers and conclusions - Even though more than half of households were informed of government projects and programmes for the support of livelihood capacity of the population, the beneficiaries of them are estimated to be less than 10%. - The percentage of beneficiary households is higher with green revolution programme and savings and loan projects than other programmes. - There appears to be inadequate information about the availability of programmes for improvement of livelihood of the population in the city. - Low income households have less opportunities to turn trainings into benefits from the training. Even though there is no difference between households in terms of training participation rates, poor households are almost three times less likely than better off households to state that they benefited from the training. - Well to do households participate less in community work

5 POVERTY - The kinship/khuree supports the livelihood of households, but for migrants, people in ger area, poor and very poor households a kinship/khuree is very limited. Key questions - What is the level of poverty in Ulaanbaatar? - Who are the poor? - What are the priorities for intervention? Key answers and conclusions: - 33 percent of the Ulaanbaatar population lives below the poverty line of 25,300 Tug per capita. 10 percent are very poor (i.e. expenditure below 60 percent of the poverty line). - Poverty is higher in Ger areas and among migrants (45 percent and 37 percent). - However, there is less inequality within those living in ger areas. The same can/cannot be said for migrants. - The poor are typically younger, less well educated and more frequently not married. They live overwhelmingly in ger areas, especially Bayanzurkh. Their household size tends to be larger and they are more frequently headed by females. - Target groups for intervention should therefore be households with many members in ger areas, and possibly also households where the head of households has not completed secondary school. - Priorities for action are improving housing and sanitation conditions. Access to health services and education looks comparatively good. The issue there is about improving quality. - Registration is not an issue that is related to poverty. Poor are registered by the same proportion as poor. Moreover, almost 90 percent are registered percent of the Kazakh population (4.3 percent) are poor. This is an issue worth of further investigation. MIGRATION Key questions - Who are the migrants? - What are the main reasons for migration? Have they changed over time? - Does the registration status influence migration flow? - Is there a link between poverty and migration? - What are the priorities for intervention of the government? Key answers and conclusions - Migrants to the city predominantly live in ger area, with a majority of them having less than complete secondary education. - A majority of the migrants are in search of employment, better livelihood, further studies and closer access to markets. Over time, employment, better livelihood and closer access to markets have gained in relevance, wheras educational needs have declined. Hence, migration can be considered to be increasingly need driven. - Out of the migrants, (71 percent) said that there expectations have been met. - Registration status does not help to reduce the migration flow into Ulaanbaatar (half of the migrants are not registered). However, lack of registration is cited by

6 those migrants without health insurance (one third) as one of the main reasons for not having health insurance (a quarter of those without health insurance stated lack of registration as the reason). - Migrants are not poorer because they are migrants, they are poor because they have lower education levels, for example. Migrants seem to face the same opportunities that the non-migrants face, their problem seems to be that they lack some qualifications in a bit greater extent than the non-migrants - In order to reduce the drive for migration it would be best to improve the economic and education situation in the areas of origin.

7 CONTENTS List of tables and figures Executive summary Concepts and definitions iv xi xv CHAPTER 1. Introduction 1 By Solongo Algaa 1.1. Justification Objectives of the Survey Structure of the report 2 CHAPTER 2. Sampling methodology 3 By Solongo Algaa 2.1. Sampling method Quantitative survey Qualitative survey 6 CHAPTER 3. General characteristics of the sample population 7 By Naranchimeg Baatar 3.1. Individual characteristics of the sample population Demographic characteristics of individuals Social and economic characteristics of individuals Household characteristics General characteristics of heads of households Demographic characteristics of households Living conditions of households 15 Conclusion 20 CHAPTER 4. Employment and Income 21 By Narantulga Baatarjav 4.1. Employment Description of terminology 21

8 Employment status of subgroups Formal and informal sector Sources of income 28 Conclusion 29 CHAPTER 5. Access to education and health 31 By Solongo Algaa 5.1. Education Education level of the adults Sex ratio in education School attendance Tuition fees and its types Access to school and kindergarten Health Health status Involvement in health insurance Social welfare and social security service Environment for healthy life 43 Conclusion 43 CHAPTER 6. Awareness, information and social capital 45 By Munkhjargal Battsengel 6.1. Participation of the community Information on projects and programs, and their benefits Participation of citizens in community activities Participation in technical training Social networks 53 Conclusion 58 CHAPTER 7. Poverty 60 By Nils Reimenschneider, Khurelmaa Dashdorj Narantulga Baatarjav 7.1. Standard poverty measures 61

9 7.2. Characteristics of the poor Description of the poor Identifying the target group for interventions Consumption and living standard Determining interventions for the poor Priorities for intervention Alternative ways of looking at poverty Capability poverty Social inclusion poverty How do alternative poverty measures fit with the 90 standard poverty measures? Conclusion 94 CHAPTER 8. Migration 95 By Navch Tumurtolgoi & Khurelmaa Dashdorj 8.1. The characteristics of migrants The characteristics of households and their living conditions Education level of migrants Health status of migrants Reasons for in-migration Registration with resident area Future migration Migration and poverty 111 Conclusion 115 Policy recommendations 116 References 117 Questionnaires 118 List of participants of the survey 132

10 LIST OF TABLES AND FIGURES Tables: Table 2.1. Poverty incidence, by location 3 Table 2.2. Details of the sample 4 Table 3.1. Percentage distribution of respondents, by location according to demographic characteristics and migration status 9 Table 3.2. Percentage distribution of respondents aged 15 and over, by location according to socio-economic characteristics 10 Table 3.3. Percentage distribution of households, by location according to demographic characteristics of head of the household 12 Table 3.4. Percentage distribution of households, by location according to socio-economic characteristics of head of the household 13 Table 3.5. Percentage distribution of households, by location according to demographic characteristics 14 Table 3.6. Percentage distribution of households, by location according to selected characteristics 15 Table 3.7. Percentage distribution of households, by location according to living condotion of household 16 Table 3.8. Percentage distribution of respondents living in GER, according to living condotion 19 Table 4.1. Percentage distribution of respondents, by employment status according to selected characteristics 23 Table 4.2. Percentage distribution of respondents aged 15 and over, by employment status according to demographic characteristics 24 Table 4.3. Percentage distribution of household heads, by employment status of household heads according to household characteristics 26 Table 4.4. Percentage distribution of employed population, by working sector according to selected characteristics 27 Table 4.5. Percentage of distribution of employed people, by health insurance according to working sector 27 Table 4.6. Average wage per month according to working sector 28

11 Table 4.7. Percentage distribution of household incomes, according to selected characteristics 28 Table 5.1. Education level of the adults by age-specific groups 32 Table 5.2. School attendance of respondents aged 7-29, by location and sex according to age group 33 Table 5.3. Percentage distribution of studying population aged 7-29, by location according to living arrangement during study time 33 Table 5.4. Percentage distribution of studying respondents, who pay tuition fees, by location 35 Table 5.5. Health status by living standard 39 Table 5.6. Percent of respondents who not have a health insurance, by location according to selected characteristics 40 Table 5.7. Percentage distribution of respondents, by living standard according to access to health service 42 Table 6.1. Percent of the households, by knowledge of community programs/projects on supporting livelihoods according to selected characteristics 46 Table 6.2. Percent of the households, by getting of benefits from community programs/projets on supporting livelihoods according to selected characteristics 47 Table 6.3. Percentage distribution of households, by participation in community activities according to selected charateristics 50 Table 6.4. Percentage distribution of households, by participation in any technical training according to selected characteristics 51 Table 6.5. Percentage distribution of households, whose members participated in training, by the impact of training on their livelihoods, according to selected characteristics 52 Table 6.6. Percent of households, who supported by kinship in daily life, by occupation of kinship according to selected characteristics 54 Table 6.7. Percentage distribution of households, by supported of khureelel in their daily lives according to selected characteistics 55 Table 6.8. Percent of households supported by khureelel in their lives, by category of khueelel according to selected characteistics 57

12 Table 7.1. Key poverty measures 63 Table 7.2. Inequality measures 65 Table 7.3. Percentage distribution of respondents by consumption expenditure poverty and selected characteristics 67 Table 7.4. Percentage distribution of respondents according to living standard 70 Table 7.5. Percentage distribution of respondents by living standard according to selected characteristics 71 Table 7.6. Percentage distribution of households by living standard according to household characteristics 72 Table 7.7. Percentage distribution of households by living standard according to districts 74 Table 7.8. Percentage distribution of population aged 15 and over, by living standard according to selected characteristics 74 Table 7.9. Percentage distribution of households by living standard according to selected characteristics 76 Table Percentage distribution of respondents by capability poverty index and selected characteristics 86 Table Percentage distribution of respondents by social inclusion and selected characteristics 89 Table Percentage distribution of respondents by greatest need and others and selected characteristics 92 Table 8.1. Percentage distribution of respondents aged 15-64, by location and migration status according to demographic characteristics 97 Table 8.2. Percentage distribution of households, by location and migration status according to demographic characteristics of the household 98 Table 8.3. Percentage distribution of households cwith heads of HH aged 15-64, by migration status according to household dwelling characteristics 99 Table 8.4. Percentage distribution of households living in ger, by migration status according to living conditions 99 Table 8.5. Percentage distribution of households, by migration status according to living conditional selected characteristics 100

13 Table 8.6. Table 8.7. Table 8.8. Table 8.9. Table Table Table Table Table Table Table Table Table Percentage distribution of respondents aged 15-64, by location and migration status according to whether registered in the 101 household as a member Percentage distribution of respondents aged 15 and over, by location and migration status according to socio-economic characteristics 102 Percentage distribution of respondents, by location and migration status according to health indicators 103 Percentage distribution of respondents, by location and migration status according to main reason for not having health insurance 104 Percentage distribution of migrants, by location according to selected characteristics 105 Percentage distribution of migrants, by location according to main expectations of movement 106 Percentage distribution of migrants, by sex according to main expectations of movement 107 Percentage of migrants, by arrival: main expectations of movement 108 Percentage distribution of not registered migrants, by sex according to main reason for not registered 109 Percentage distribution of not registered migrants, by location according to main reason for not registered 110 Percentage distribution of respondents, by living standard and location according to consideration of onward move 110 Logistic regression odds for being poor and standard errors by selected characteristics 112 Logistic regression odds for being in greatest need and standard errors by selected characteristics 114

14 Figures: Figure 3.1. Pyramid of the total population 7 Figure 3.1à. Pyramid of the population living in ger area 7 Figure 3.1á. Pyramid of the population living in apartment area 7 Figure 3.2. Dependency ratio of respondents, by location 8 Figure 3.3. Percentage distribution of respondents aged 15 and over, by location according to education level 10 Figure 3.4. Percentage distribution of households, by location according to way people find out the dwelling 16 Figure 4.1. Illustration of various concepts in the context of employment 22 Figure 4.2. Percentage distribution of economic inactive population, by main reason for not working 23 Figure 4.3. Percentage distribution of respondents, by employment status 25 according to location and migration status Figure 4.4. Percentage distribution by employment status according to education level 25 Figure 5.1. Sex ratio in education level of the respondents aged 15 and over, by location 32 Figure 5.2. Percentage distribution of the households, by distance of two and more km away from the secondary school according to selected characteristics 36 Figure 5.3. Percentage distribution of the school dropped out children aged 7-18, by location and sex 34 Figure 5.4. Percent of people, who have health insurance, by location 38 Figure 5.5. Percentage distribution of the insured people, by location according to type of payment 38 Figure 5.6. Percentage distribution of the households, by dictance of 3 and more km from the family clinic away, according to selected characteristics 41 Figure 5.7. Percent of the people who have access to health service by selected characteristics 41

15 Figure 6.1. Percent of households, who had information about and got benefit from community programs/projects according to household living standards 49 Figure 6.2. Percent of households, whose members attended training and improved livelihood as a result of training according to living standards 53 Figure 7.1. Lorenz curve for Ulaanbaatar 64 Figure 7.2. Distribution of expenditure per capita in ger and apartment areas 65 Figure 7.3. Socio-economic characteristics: Largest differences between poor and non-poor 68 Figure 7.4. Socio-economic characteristics: Smaller differences between poor and non-poor 69 Figure 7.5. Socio-economic characteristics: Smaller differences between poor and non-poor 2 69 Figure 7.6. Percentage distribution of respondents by location according to living standard 72 Figure 7.7. Percentage distribution of respondents by migration status according to living standard 73 Figure 7.8. Percentage distribution of population by living standard according to type of dwelling 75 Figure 7.9. Difference between poor and non-poor: housing 77 Figure Difference between poor and non-poor: education 77 Figure Difference between poor and non-poor: health 78 Figure Difference between poor and non-poor: employment and social inclusion 79 Figure Difference between poor and non-poor: registration and income sources 79 Figure Determining priorities 1 80 Figure Determining priorities 2 81 Figure Determining priorities 3 81 Figure Determining priorities: housing 82 Figure Determining priorities: access to education 82 Figure Determining priorities: access to health 83

16 Figure Determining priorities: housing, education and health 83 Figure Three categories of poverty 90 Figure Three categories of poverty: Overaps 90 Figure Three categories of poverty: including the very poor 91 Figure Comparing the poor and the people in greatest need-1 93 Figure Comparing the poor and the people in greatest need-2 94 Figure 8.1. Population pyramid by migration status 96 Figure 8.2. Percentage distribution of migrants, by living duration in UB 104 Figure 8.3. Percentage distribution of migrants to UB, by aimags 106 Figure 8.4. Percentage distribution of not registered migrants, by districts 109

17 CONCEPTS AND DEFINITIONS Apartment: A dwelling unit with one or more rooms intended for a single household. It is separated from units of other households by a designated wall and has a doorway out with share structures such as stairs, entrance and outside. Building of apartments: A building which is designed and constructed for habitation of two or more households (irrespective of the size, floors and capacity of the building). It has infrastructure provision. Economically active population/labour force: The indicator is defined by total of persons who are employed and unemployed or actively looking for a job. Economically inactive: Population aged 15 and over but not included in labour force, in other words thoes who are not ready to work. These include students, people who are looking after children/sick/older people, too old or retired people, people who are not working as a result of lack of the registration or other related documents, and those who are made redundant and are not looking for new jobs etc., are included in not working people. Employed: Employment is related to population aged 15 and over and people who are employed. Employed persons are those who did some paid work at least one day during the last week preceding the survey. Those who did not work during the last week, but have not cancelled their contracts with the organization and will return to the job after some time. Members of households, who are engaged in HH production or services, for instance herdsmen and agricultural workers. Employment rate: The proportion of the total population of 15 and over years who are employed Informal sector: A sector which is characterized by none permanent, none secure activities which are outside the legal framework including; registry, tax and legal functioning, work conditions, social welfare and protection services. For example own account workers, casual workers and paid household workers are classified as informal employees. Formal sector: A sector comprised of registered public and private owned activities with relatively permanent functioning. It has relatively secure working conditions, social welfare and protection services for the employed. The formal sector is represented by all types of public and private establishments, NGOs and international organizations. Kinship (khuree): People who have close relations with household members such as relatives, friends, acquaintances. Khuree means a dry relation as opposed to a blood relation (e.g. parents, children sisters/brothers etc). A dry relation has equally binding responsibilities towards a person as does a blood relation, if the person is asking for help. The dry relations can be old class mates or parents of an old classmate, friends of the family etc. If a person asks a khuree for help/money/connections, he in principle has to do something. But it should be seen as something a person can approach daily if that person is in difficulties or needs assistance. It is not necessarily something that will just be given i.e. someone food, as soon as someone is

18 hungry, i.e. one has to ask for it. In that sense it can be a social security network one can draw on when times are rough, it can be a means by which one can get a job through connections, it can be the place where one gets credit etc. House: A dwelling which was built and equipped by residents themselves for habitation of one or more households. It has some infrastructure provisions. Housing: A separate structure with one or more rooms designed for a household habitation. Types of housing include houses, apartments, student and public dormitories. Income: Monetary and non-monetary income from different sources to sustain livelihood. Income includes wages, pensions/social allowances, HH business and services income, cash remittances, rentals, interest and others. Labour age: According to the International categories, population aged 15 and over are labour age. Such indicators as employment rate, unemployment rate, labour force participation rate are also related to population aged 15 and over. Labour force participation rate: The proportion of economically active persons in the total population of 15 and over years. Living standard: Indicator which expresses the economic capacity of the population. There are five categories of living standard. Living quarter: Structurally separate and independent places of abode occupied at the time of the survey. It may have a type of non-living quarter. Non-living quarter: A housing unit which was not built for human habitation but occupied by households and individuals as living quarters. It includes housing units in offices, train station and entrances of apartment buildings. Non-migrants: are year old persons who have been residing in Ulaanbaatar for 5 or more years preceding the month in which the survey is conducted. Not working: All population aged 15 and older. In other words, not working is defined by total number of unemployed or people who are looking for a job at the time of the survey and economically inactive population. Number of children of household head: Number of persons who are related to the household head as children irrespective of ages. Number of rooms in housing: The total number of rooms including bedrooms, living rooms, work rooms and other rooms. The number does not include kitchen, corridor, bathroom, or inwall built store. Median age: For example, if the median age of the population covered survey is 24 then 50 percent of them are aged under 24 and the remaining 50 percent are aged above 24. Migrant: year old persons who have migrated from other areas (aimags and overseas) to Ulaanbaatar city in the last four years preceding the month in which the survey is conducted.

19 Migrant household: the head of a household which migrated from other areas (aimags and overseas) to Ulaanbaatar in the last four years preceding the month in which the survey is conducted. Population dependency: Number of children and elders per 100 persons of working age. It is estimated by ratio of the total sum of 0-14 and 65 and above aged persons to the population of years and multiplied by 100. Poverty line/minimum living standard: Criteria or indicator used for dividing the population into poor and non-poor groups and which is used in defining of the living standard of the population. School drop-out: Population of 7-18 years old who were not engaged in schooling during the survey but left schooling without completing the respective level of secondary education. School enrolment rate: The percentage of 7-29 year old persons engaged in studies in the total population by specific ages. Sex ratio: Number of men per 100 women. Social network: Kinship relations which influences the livelihood strategies of households such support from friends, relatives and acquaintances. Student dormitory: A dwelling built for shared accommodation for pupils and students and other a like dwellings (special school, children camp). Unemployment rate: The proportion of persons not working but part of the labour force, i.e. not part of the people who are economically inactive.

20 CHAPTER 1. INTRODUCTION 1.1 Justification In recent years, Mongolia has experienced continued rural to urban migration flows, particular to Ulaanbaatar. According to the 2000 population and housing census (NSO, 2001), one third of the total population of Ulaanbaatar are migrants. The consequences of migration are strongly felt in ger areas of Ulaanbaatar. The rapid growth of population in the peri-urban ger areas places economic and administrative burden on the authorities and deepens the degree of poverty. Most of the inhabitants in ger areas are either unemployed or engaged in under-employed activities in the informal sector. The capacity of the infrastructure to cope with the influx of migrants in arear of intergratiion, support for basic living standards, access to health care and education, presents a serious difficulties that require carefully planned and coordinated responses. According to the micro study of internal migration in Mongolia 2000, the living conditions of in-migrants are lower compared to non-migrants of Ulaanbaatar (MoLSW, PTRC and UNFPA, 2001). Moreover, migration may have an impact on poverty. Poverty increased in actual numbers from thousands in 195 to thousands in 1998, but declined as percentage of the total population from 36.3 percent to 35.6 percent. Half of the total poor population lives in urban areas and one forth of them are concentrated in Ulaanbaatar where the severity of poverty has increased from 5.7 in 1995 to 7.1 in 1998 (Government of Mongolia and UNDP, 2000). A large number of households in areas do not have access to clean water. They live in areas where there is no sewage disposal or regular waste collection. It is not uncommon to have only stoves for heating instead of hot water pipes. A number of the households in ger areas do not have access to electricity either. Policy approches should; focused on protecting the human rights of migrants and thr poor, by providing them social and health services. As a result it is essential to provide baseline information for development of policies and regulations. However, there is lack of information on disparities on levels of income and consumption expenditure inequalities between apartment and ger areas in Ulaanbaatar. 1.2 Objectives of the Survey The main objectives of the survey are to determine the living standards of the population in Ulaanbaatar based on their income and consumption expenditures and to study the compexity of poverty and migration. In order to reach to the main objectives the following ussues are explored: - To determine the main sources of household income and consumption expenditures; - To determine the access to and quality of social services; - To establish basic poverty indicators by type of dwelling and by migrants and nonmigrants;

21 - To find out in Ulaanbaatar, the main reasons of migration, and the accessibility of social services for unregistered migrants; - To determine the relation between poverty and migration; - To determine characteristics of poor and identify the people in the greatest need and in order to develop policy recommendation; 1.3 Structure of the report The report is structured in the following way. It begins with an executive summary that states the key findings of the report. Chapter 1 focuses on justification of the survey and defines objectives of the survey. Chapter 2 discusses a description of the methodolgy used, Chapter 3 presents a description of the characteristics of the sample population. Chapter 4 focuses on employment and income sources of the sample population, Chapter 5 discusses an access to education and health. Chapter 6 describes an access to information and social capital of the population in Ulaanbaatar. Chapter 7 on poverty, characteristics of the poor are established, and the target group for intervention is identified, and as well as priorities for action are outlined. Alternative measures of poverty based on capabilties and social capital are explored in this chapter. Chapter 8 describes the characteristics of migrants and analyses the link between migration and poverty. Finally, we suggest the policy recommendations based on the key findings of the survey.

22 CHAPTER 2. SURVEY METHODOLOGY We used quantitative and qualitative methods for the survey. 2.1 Sampling method This study uses a random sampling survey method for both quantitative and qualitative studies Quantitative survey The sample design used the two-stage simple random sampling. The primary sample was of 30 khoroos, selected from 94 khoroos in Ulaanbaatar with probability proportional to their population. The secondary sample we choose 50 households selected randomly from each of the 30 khoroos. Sample size Sample size was defined according to the following formula: 1 2 e n = 2, where n- Sample size t p( 1 p) e- Sampling error p- Percentage of poor population t- Level of reliability We were accepted the sampling error of 0.04 percentage point and significance level of Table 2.1 Poverty incidence, by location Incidence Location Total population Total poor population thous.per Percent thous.per Percent All urban Ulaanbaatar city Aimag centers All rural National Source: NSO, Living Standard Measurement Survey, Ulaanbaatar. Table-2.1 shows that the poverty level in Ulaanbaatar is 34.1 percent, which is lower than the urban poverty level of 39.4 percent and the poverty level of aimag centers. However, they cannot be disaggregated by housing type. This data was used for designing the sample survey, as it was the most reliable. Using this information and the above formula we determined the representative sample size as 539 households. In order to minimize non-sampling error and to encounter clustering effect we defined an actual sample size as 1500 households. (According to the practical experiences of sample survey the coefficient of typical clustering effect is equal to 2.5.) The coverage

23 Table 2.2 shows the distribution of districts, khoroos, households and the household population selected in the survey. From 1500 households (30 khoroos of 6 districts of Ulaanbaatar) 6847 persons were covered. Table 2.2 Details of the sample Selected District Khoroos Households Persons Percent Bayanzurkh Nalaikh Songinokhairkhan Sukhbaatar Khan-Uul Chingeltei Total The coverage of the sample survey was 100 percent. A total of 4070 persons of selected 816 (54.4 percent) households from ger areas, and 2777 persons of selected 684 (45.6 percent) households from apartment areas were used for the survey. Questionnaire In order to identify demographic, social and economic characteristics of households and household members, the questionnaires were designed consisting of the following modules for the quantitative survey. - Household members and partners - Education - Health - Employment status - Living conditions - Income and consumption - Expenditures - Social services and social networks - Migration The pilot survey A pilot survey was conducted from the 28 th to the 29 th of September 2003 in order to test the design and structure of the questionnaire, as well as possible answers to the questions. Six teams comprising of 14 people conducted the pilot survey. The pilot survey covered ger area khoroos not covered in the main survey that have the highest levels of migration and poverty (18 th khoroo in Songinokhairkhan district), as well as some khoroos in apartment areas (9 th khoroo in Sukhbaatar district). Altogether, 18 households from ger areas and 12 households from apartment areas were covered in the pilot survey. Training of supervisors and interviewers

24 The interviewers were selected from among students of the School of Economics of the Mongolian National University. The selected students were those who fully attended the Social Research Methodology class and past. During October of 2003, a training course together with practice sessions was organized for supervisors and enumerators. A manual was prepared and distributed. Guidelines were developed for conducting individual interviews and focus group discussions with instructions for supervisors, enumerators and note takers. Data collection The survey fieldwork was carried out from November 6 th to the 20 th, 2003, by 7 teams comprising of 42 members. Each team consisted of 5 interviewers and a supervisor. Social workers and administrative officers of khoroos provided assistance to find the selected households. The data collection process by enumerators and supervisors was monitored in order to improve the process, and to identify any mistakes and misunderstandings. Limitations to data collections The biggest difficulties were encountered were: the cold weather; collection of data from unregistered households who live in Gers on the hill sides; and households living in apartment areas in the city center, especially, households in the Nalaikh district. It should be noted that social workers, sanitary inspectors of khoroos and drivers of the survey team greatly helped to overcome these difficulties. Data processing The questionnaires were checked for completeness, internal consistency and correctness of the coding. During the checking and coding of open-ended questions, 47 households (3.1 percent) out of 1500 households were re-interviewed to clarify their answers. Integrated System for Survey Analysis (ISSA) Software Computer Package was used for data entry and Statistical Package for Scientific Survey (SPSS) was used for data processing and statistical analysing of the survey. Descriptive statistics was used for determining distribution frequencies of household and individual characteristics. The main indicators of poverty, namely, the poverty headcount (P0), the poverty severity index (P1), the poverty gap index (P2), the Gini coefficient and Lorenz curve were estimated for Ulaanbaatar, ger and apartment areas and migrants and non-migrants, respectively. Multivariate logistic regression method was used to analyse the variables influencing poor and those in greatest need Qualitative survey The selection of the khoroos for the qualitative survey was based on criteria including khoroos with the highest level of migration and poverty. Based on this criterion, 5 th and 15 th khoroos of Bayanzurkh district and 7 th and 19 th khoroos of Songinokhairkhan district were selected. Data collection The qualitative survey data was collected using individual interviews, focus group discussions and observation methods, from December 18 th to the-23 rd, 2003.

25 Focus Group Discussions A manual for individual interviews and focus group discussions was used for the qualitative survey. Individual interviews were conducted with 8 persons, including governors of ger area khoroos in outlying areas, social workers, family doctors and school teachers. Focus group discussions were conducted among 4 groups differing by their socio-economic characteristics: 1) poor people of ger areas who live semi-urban areas, 2) people who live in apartments in the center of the city, 3) unregistered migrants, 4) registered migrants. In total, 20 people participated in focus group discussions. Data processing We used a simple comparative analysis method for processing the collected data.

26 CHAPTER 3. GENERAL CHARACTERISTICS OF THE SAMPLE POPULATION Chapter 1 focuses on justification of the survey and defines objectives of the survey. This chapter talks about the composition of the population by age-sex structure, marital status, ethnicity, education level, employment status, as well as general characteristics of the studied households. The general characteristics and living conditions of the sample population are characterized by their location namely, by ger and apartment areas. 3.1 Individual characteristics of the sample population Demographic characteristics of individuals The survey shows that 59.4 percent of individuals live in ger areas while 40.6 percent live in apartment areas. Population pyramids showing age and sex distribution are shown by the total population in Ulaanbaatar, as well as separately by ger and apartment areas (see Figure 3.1) Figure 3.1 Population pyramid Total population Male Female Figure 3.1a Population pyramid Ger population Figure 3.1b Population pyramid Apartment population Male Female Male Female The general pattern of age-sex pyramids of the sample population are similar to that of age-sex specific structure of the Mongolian population. The above figures illustrate that for the last decade fertility rate has

27 dramatically decreased in Ulaanbaatar compared to preceding years, but now it tends to be stabilized at that level. No significant differences exist between ger and apartment areas. Of the total sample, 46.6 percent are male, while 53.5 percent are female, resulting in a sex ratio of The median age of individuals was 24 (50 percent of population is under 24 while 50 percent is over 24), whereas the average age was 28. Figure 3.2 Dependency ratio of sample population by Dependency location ratio Ger Apartment Total Dependency ratio Child dependency ratio Elderly dependency ratio The demographic dependency ratio of the sample population is 45.3 percent. The general low level of the dependency ratio is due to the fact that people born during the high-fertility period have now reached the economically active age (Figure 3.2). The overall dependency ratio in the ger areas are 5 percentage points higher than that in the apartment areas. The child dependency ratio of the sample population living in ger areas are 9 percentage points higher than that in apartment areas, whereas the elderly dependency ratio is 4 points lower. The economic dependency ratio (number of economically inactive people per 100 economically active people), calculated on whether individuals did anything to earn income in the week preceding the interviews, was percent. This means that one employed person supports more than two people in addition to himself/herself. Table 3.1 illustrates some demographic characteristics of individuals. About half (50.3 percent) of people aged 15 and over are married and most of the latter (80.7 percent) are officially registered as married. Also, 37.8 percent of them were never married. In terms of household location, the percentage of households officially registered as married was higher in apartment areas compared with ger areas. According to the 2000 Population and Housing Census, for Ulaanbaatar, the percentage of Khalkh, Kazakh, Durvud, Bayad and Buriad ethnicities of Mongolia were comparatively high. Therefore, these ethnicities were classified separately (9 foreigners were included in the category others ) in the questionnaire. The majority of individuals interviewed were Khalkh (86.9 percent). Of the remaining, 3.8 percent were Kazakhs, 2.7 percent were Durvuds, 2.1 percent were Bayads, 1.0 percent were Buriads and 3.5 percent were composed of other ethnical groups. These results are quite similar to that of the census. In terms of location, Khalkhs represented 90 percent of people living in ger areas and 88.2 percent of people living in apartment areas.

28 Questions on migration were asked only from individuals aged Out of these individuals, 14.6 percent migrated during the last four years. The majority of migrants, 79.5 percent, live in ger areas. Table3.1. Percentage distribution of respondents by location according to demographic characteristics and migration status Demographic characteristics Location Apartment Ger Total Marital status* Single Married Living together Separated Divorced Widowed Total Number of people aged 15 and over Ethnic group Khalkh Kazakh Durved Buriad Bayad Others** Total Number of people Migration status*** Migrant Non-migrant Total Number of people aged Note: * Marital status of the sample population aged more than 15 years old. ** Only 9 persons were not Mongolian nationality out of the sample population. Those are included under category Other. *** Migration status of the sample population aged years old Social and economic characteristics of individuals Figure 3.3 shows the percentage distribution of sample population aged 15 and over by location according to education level.

29 Percent Figure 3.3. Percentage distribution of sample population aged 15 and over by location according to education level Ger Apartment Uneducated Primary Noncompleted secondary Completed secondary Technical vocational Special vocational/ Diploma High 10 Thirty two-point-eight percent of the individuals interviewed had secondary education (10 years), 21.5 percent had incomplete secondary 1 education (8 years), 12.3 percent had technical vocational or non-degree tertiary education, 21.3 percent had higher education, and the remaining 12.1 percent had only primary education (3 years) or were uneducated. The level of education of individuals varied substantially depending on household location. The number of individuals with vocational or non-degree tertiary education living in apartment areas is 2 percentage points higher than that of individuals living in ger areas. Furthermore, the rate of individuals with higher education is 27 percentage points higher in apartment areas. At the same time, the percentage of individuals living in apartment areas with education lower than technical vocational is 2 to 15 percentage point lower than that of individuals living in ger areas. The overall level of education of residents of ger areas is lower than that of residents of apartment areas. Table 3.2. Percentage distribution of the respondents aged 15 and over, by location according to social and economic characteristics Social and economic characteristics Location Ger Apartment Total Employment status Employed Not working Total Number of people aged 15 and above Working sector Formal sector Informal sector Abroad Total Working population In the Mongolian education system, the first three years of education are termed primary (age 8-10), completion of the following 5 years (age 11-16) is termed incomplete secondary, after which children can go on to obtain vocation training; completion of the following 2 years of education (ages 17-18) results in completed secondary education.

30 Table 3.2 illustrates the employment status of individuals. Of the sample population aged 15 and above, 42.9 percent answered that they did something to earn money during the week preceding the survey. The number of individuals is 8 percentage points higher in apartment areas compared with ger areas. In terms of employment sector, 67.9 percent of individuals were working in the formal sector, whereas 30.2 percent were working in the informal sector, with the remaining 1.9 percent working abroad. Employment status also varied significantly by location. The percentage of the sample population living in ger areas and working in the informal sector was 12 percentage points higher than that of the population living in apartment areas, whereas the percentage of the sample population living in ger areas and working in the formal sector was 10 percentage points lower than that of the population living in apartment areas.. These results show that ger area residents are more likely to work in the informal sector. 3.2 Household characteristics General characteristics of heads of households Out of all surveyed households, 76.3 percent were male-headed households while 23.7 percent were female-headed households (Table 3.3). The rate of female-headed households in apartment areas were 4 percentage points higher than that in ger areas. Heads of households shown by age were as follows: 22.9 percent were in the age group 20-34, more than half (52.1 percent heads of the household were in the aged group35-54) were people in their prime age, 24.2 percent were elderly (55 and over), and 0.8 percent were under 20. It was observed that it is common among Mongolians to regard a boy of minor age as a household head when the head of household (father) dies. The age composition of heads of households does not vary much by location, except for the fact that the number of households headed by the elderly is 5 percentage points higher in apartment areas than in ger areas. Almost 70 percent of the heads of households were married, while 6.7 percent have never married, 16.8 percent were widowed and the remaining 6.5 percent were divorced. The majority of married heads of households (85.3 percent) were officially registered as married. The number of married heads of households is 6 percentage points higher in ger areas than in apartment areas, whereas the number of heads of households who never married and who were divorced or widowed is 1-2 percentage points higher in apartment areas than in ger areas. The percentage of widowed heads of households was higher in apartment areas than in ger areas, which might be due to the high number of the elderly people, surveyed living in apartment areas. The survey shows that 53.4 percent of heads of households surveyed have 1-2 children, 31.0 percent have 3 and more children, while the remaining 15.6 percent had no children. On average, heads of households had 1.9 children. However, heads of households living in ger areas have 2.3 children on average, which is 0.7 percentage points higher than of those living in apartment areas.

31 The survey illustrates that 13.4 percent of heads of households migrated to Ulaanbaatar within four years preceding the survey, with most of them (76.1 percent) now living in ger areas. Table 3.3. Percentage distribution of households by location according to demographic characteristics of household heads Demographic characteristics of household heads Location Ger Apartment Total Sex Male Female Age Under Marital status Single Married Living together Separated Divorced Widowed Number of children None and over Mean Migration status Migrant Non-migrant Total Number Number of households with heads aged * Table 3.4 shows the social and economic characteristics of heads of households by living area. More than half of heads of households (52.6 percent) had secondary education or lower, whereas 45 percent had education higher than secondary, and 2.4 percent were uneducated. Percentage of heads of households with higher education in apartment areas was over 37 percentage points higher compared with ger areas, whereas percentage of uneducated heads of households in ger areas was 1.5 percentage points higher than in apartment areas. Overall, the survey results shows that the percentage of heads of households with education lower than technical vocational education is much higher in ger areas; correspondingly, the percentage of household heads with higher than technical vocational education was lower in ger areas.

32 Table 3.4. Percentage distribution of households by location according to social and economic characteristics of household heads Social and economic characteristics of Location household heads Ger Apartment Total Education level Uneducated Primary Uncomplete secondary Secondary Technical/vocational Special vocational/ Diploma Higher education Employment status Employed Not working Working sector Formal sector Informal sector Abroad Total Number of household heads Number of household with heads employed More than half, or 59.6 percent of heads of households did something to earn income during the week prior to the survey (this number includes people who did not worked in the preceding week, but had employment contracts). The number of heads of households was 7 percentage points higher in apartment areas compared with ger areas. Six in ten, or 62.5 percent of employed heads of households work in the formal sector, whereas 35.9 percent work in the informal sector and 1.6 percent work abroad. The number of heads of households working in the formal sector was 12 percentage points higher in apartment areas compared with ger areas. But the number of heads of households working in the informal sector was 14 percentage points higher in ger areas compared with apartment areas Demographic characteristics of households Table 3.5 shows demographic characteristics of households by location. The majority of households surveyed (90.4 percent) were registered 2 in their resident sub-district (khoroo) but the number of unregistered households in apartment areas was 1 percentage point higher than in ger areas. Renting of rooms or entire apartments is common in apartment areas, which may be a reason for a somewhat lower rate of registration in apartment areas. 2 Registration is made at several levels: at the municipality level (in which case the individual becomes a resident of Ulaanbaatar), at the district level and at the khoroo level. In order to be registered at khoroo, the individual must be registered at the municipality level and the respective district level. However, those not registered at khoroo levels could be both registered and not registered at the municipality level. The former tend to be those who do not bother or have no time to register at their khoroo, while the latter tend to be migrants.

33 Table 3.5. Percentage distribution of households by location according to demographic characteristics Demographic characteristics Location Ger Apartment Total Registration status Registered Not registered Number of people in the household and over Mean Number of migrants in the household and over Mean Number of nuclear families in the household and over Mean Total Number Six in ten (63.4 percent) of all households had 3-5 members, over 25 percent had 6 and more members, and 11.9 percent were single-headed or two-person households. Percentage of single-headed or two-person households in apartment areas was twice as high as that in ger areas, while percentage of households with 2 to 5 members was 7 percentage points higher in apartment areas compared with ger areas. At the same time, percentage of households with 6 and more members living in ger areas was twice as high as those living in apartment areas. On average, households living in apartment areas had 4.8 members, whereas households living in ger areas had 5.9 members. The majority of households (82.2 percent) interviewed had no migrants, while 10.1 percent had 1 or 2 migrants and the remaining 7.7 percent had 3 or more members who migrated. Less than 10 percent of total households in apartment areas had migrants, whereas in ger areas 12.4 percent of households 1 or 2 migrants, 8.3 percent had 3 or 4 migrants and 3.4 percent had 5 or more migrants. In most households (83.5 percent), there was one nuclear family in the household. Compared with apartment areas, ger areas had more mixed households (4 percentage points higher) Living conditions of households

34 Table 3.6 illustrates types of dwelling and further, types of non-ger dwellings of households by location 3. Table 3.6. Percentage distribution of the households, by location according to selected characteristics Selected characteristics Ger Location Apartment Type of dwelling Ger House* Type of living quarters** House Apartment Public dormitory Non-living quarter*** Type of ownership Owned Not owned Total Number Note: * - including apartments and public dormitories ** - not including ger. *** - places not intended for living The survey shows that most of the households (82.1 percent) live in houses or apartments. In addition, one third of ger area households live in gers. Of those households living in a house, more than half (51.3 percent) live in apartments, 42.3 percent live in stand-alone houses, 3.8 percent live in public dormitories, and the remaining 2.6 percent live in dwellings not intended for living (non-living quarters), such as public lavatories adapted for living, space under stairwells adapted for living, rail carriages, storage areas and office rooms. The majority of households (95.3 percent) in ger areas live in a house whereas, most households (92.2 percent) in apartment areas live in apartment blocks. The percentage of households living in non-living quarters in apartment areas is twice as high as in ger areas. The reason is that almost all apartment blocks have guards/ cleaners living under stairwells in each entrance. Of the households covered by the survey, 87.7 percent live in privately owned dwellings. However, the percentage of households living in dwellings not owned by the household is 1.8 times higher in apartment areas compared with ger areas, due to the tendency in the former to rent out apartments. Figure 3.4 illustrates the ways through which households obtained their dwellings. For instance, residents of ger areas usually (71.1 percent) live in houses that were built by themselves or purchased, whereas residents of apartment areas (52.2 percent) tend to live in dwellings inherited from their parents. The survey showed that apart from renting of apartments, people also often rent fenced houses in ger areas. Total 3 In ger area, households frequently have a ger and a house in a fenced area. The houses in ger areas have neither running water and sewage, nor central heating.

35 Figure 3.4. Percentage distribution of households by location according to the way people find out the dwelling Inherited Own built/purchased Rented Privatized Other Total Apartment Ger Percent The survey shows, that (6.1 percent) of the households look after other peoples apartments or houses, as a result they live free of charge. These type of households are twice as high in apartment areas compared with ger areas, which may be because some households, would be looking after other peoples apartments or living together with others (relatives etc.) in their apartments. Table 3.7 illustrates housing conditions of households by living area. All households living in apartment areas obtain electricity from the centralized power source, whereas 5.1 percent of the households in ger areas are not supplied with electricity, using candles instead. Table 3.7. Percentage distribution of households, by location according to living conditions Living conditions Ger Location Apartment Electricity Central power supply Candle Main sources of drinking water Central: hot and cold water cold water only Well : protected unprotected Other Heating system Central Non central Other Total Type of fuel* Electricity Wood

36 Charcoal Dung Gas Other Toilet location Indoor flush toilet Public indoor flush toilet Pit-latrine None Water sewage system Central Non central Hole Open space Waste disposal Dedicated waste disposal sites Open spaces, holes, canals, etc Number of rooms** Presence of storage/garage Yes No Telephone Have Don t have Total Number Note: *- calculated based on multiple response- questions. **- not including gers In terms of drinking water supply, households (94.1 percent) living in apartment areas have centralized supply of both hot and cold water, whereas the majority of ger area residents (82.0 percent) obtain drinking water from protected wells. Out of all households surveyed, 44.3 percent get water from centralized water supply (hot and cold water), 45.7 percent get water from protected wells, 3.4 percent get water from unprotected wells, and 6.6 percent get water from other sources. The number of households who use water from insecure sources is 30 times higher in ger areas than in apartment areas. Moreover, households living in ger areas are located meters away from their source of drinking water, on average. The survey reinforces the urgency of the issue of poor water supply in ger areas.

37 Focus group discussions showed that people have to line up for 1-2 hours to get their drinking water. Sometimes, even after queuing, they can not get water because of the shortage of well water. It is necessary to have bath-houses, many people desire them. The well water is really difficult. Because there are not enough water supplies at wells, there is always a long queue, and sometimes you end up not getting any water. (B. female, age 45, teacher, Bayanzurkh district) Another serious problem is that some residents in apartment areas do not always have access to drinking water from the centralized water system. These are people who live in former public dormitories of workers, usually located in apartment areas. Every district had such households living in public dormitories that do not meet minimum requirements for living conditions (e.g. no centralized water system, toilet or waste disposal site). Because there is water in those public dormitories, people always have to go knock on doors to ask for water. But because water is metered not everyone gives them water. (D. female, age 53, khoroo governor, Songino-Khairkhan district) Another big difference between living conditions in ger areas and apartment areas is the heating system. The majority of households (97.7 percent) living in apartment areas have a centralized heating system, while most households (97.4 percent) living in ger areas have an ordinary heating system (coal-fired stoves). The most commonly used sources of energy for heating are: first, electricity used by 60.7 percent of households, second, wood percent, and third, coal percent. Also, a small percentage of households use dung or gas for fuel. In addition, there are several households that use other kinds of materials such as waste paper or litter for their energy. All inhabitants in ger areas burn wood and coal, whereas residents in apartment areas mostly use electricity. One of the key indicators of living conditions is the type of toilet/restroom. The survey shows that 42.5 percent of households surveyed have inside toilets, while 2.4 percent have outside toilets (in other words public toilet). Also, 53.7 percent have latrine pits, while the remaining 1.4 percent have no toilets at all. When we look at the distribution by living areas, the majority of residents in apartment areas (93.1 percent) have inside toilets, whereas most of the residents in ger areas (97.3 percent) have pit latrines. Of all households living in apartment areas, 5.1 percent have public toilets, a figure 5 percentage points higher compared with ger areas. But 0.1 percent of all households living in apartment areas have no toilet at all. These households live in public dormitories, and the issues of toilets, solid waste and waste water disposal present the utmost urgent challenges to them. This situation not only leads to contamination of air and soil in the surrounding areas, but also affects safe and secure living conditions of citizens. Only the residents of the three discard buildings throw their rubbish everywhere. Because they do not have toilets they relieve themselves and throw their sewage outside, which causes a lot of air pollution and soil degradation. The neighbors complain that they have to breathe their defecation. (D. female, age 53, khoroo governor, Songino-Khairkhan district)

38 Almost half (48.7 percent) of households surveyed answered that they dispose their sewage into holes. Forty-four point nine percent use a centralized sewage system, 0.5 percent use individual/local sewage systems, and 5.9 percent dispose sewage in open spaces. Also as the survey shows, most ger area residents (99.1 percent) dispose their sewage into holes, whereas 10.0 percent disposes sewage in open spaces, and less than 1 percent uses individually built sewage systems. In case of apartment areas, almost all residents (98.5 percent) dispose sewage through a centralized system. It is worth noting that, regardless of living area, the majority of households surveyed (94.7 percent) dispose their solid waste into dedicated disposal sites. However, 9.4 percent of households in ger areas and 0.4 percent of households living in apartment areas replied that they throw rubbish into the open or into a ditch or canal. The majority of households in apartment areas (81.2 percent) live in 1-2 room apartments (excluding kitchen). The percentage of households with a one-room house in ger areas is 3 times higher than in apartment areas, whereas the percentage of households with 3 and more room apartments is 4 times higher when compared with ger areas. Of all households, 32.8 percent own barns, garages and extra gers besides their main living place. The number of these households is twice as high in ger areas compared with apartment areas. The survey also shows that ger area residents tend to have an extra ger and a shed, whereas residents in apartment areas tend to have garages. The survey illustrates that telecommunication has been rapidly developing in Mongolia, with the use of telecommunication services extending rapidly. To some extent, the use of telephones indicates living conditions and living standards of households. Six in ten, or 61.4 percent of all households have a telephone at home, including those who have landlines, cellular phones, or both. During the survey, 84.5 percent of households in apartment areas answered that they have telephones, which is twice as high when compared with ger areas. Table 3.8. Percentage distribution of households living in gers, according to dwelling conditions Selected indicators Percent Number of walls in a ger* 4 and less and over 12.7 Tuurga** Single 40.5 Double 59.5 Burees *** Single 39.8 Double 60.2 Floor**** Yes 77.7 No 22.3 Total Note: * Walls are the panels of the wooden skeleton of a ger. The number of walls conditions the size of the ger. The smallest ger has 3 walls, and then it s really tiny. ** Felt cover of the ger for insulation, put around the wooden structure of the ger. Usually, single tuurga is used in the summer and double is used in the winter. *** Several layers above the tuurga for protection against precipitations. **** Usually a folding wooden floor. In the summer many households do without.

39 Of 1500 households surveyed, 269 households live in gers. Of those, more than half (64.3 percent) live in gers with 5 walls, 23.0 percent live in gesr with 4 or less walls, and 12.7 percent live in gers with 6 and more walls (see Table 3.8). The type of tuurga, burees and the use of floors in cold seasons is an essential indicator of living standards. Of households living in gers, 59.5 percent use double burees, 60.2 use double tuurga, and 77.7 percent put in a floor during cold seasons. The numbers show that there are still many households living without these basic needs in the cold winters. Conclusion As seen by the age structure of the population of the capital city there are two working persons (aged 15-64) per non-working person (child and elder). However, in reality, there are two dependents per employed person which means he or she has to feed 2 more persons in addition to him/herself. Particularly, economic dependency is higher in ger areas than in apartment areas. The unemployment rate in Ulaanbaatar city is high and especially in ger areas it is the double the rate than the apartment areas. Considerable sex disparity in labour force participation rate is caused by a higher proportion of women in economically inactive population. At the same time, compared to men, women in the city are more engaged in unpaid household duties such as household daily chores and looking after children and elders. The education level of persons in ger areas is lower that the ones in apartment areas. This is also the case with the proportion of persons engaged in work in which ger areas are far lower than apartment areas. However, ger areas have a higher proportion of informal employees. Or to be particular, there are more people employed in the informal sector which does not require high education levels and proffesional skills. Living conditions, particularly the clean water provision is poor in ger areas. Households which consume unprotected (untreated) water is 30 times more in ger areas than in apartment areas. There are cases where people can t get water even though they stand up in long queues for it. The housing condition of half the percentage of households living in gers fails to meet even the minimum requirement. Such households live in gers without floors, double outer and inner cover during the winter time. One in every ten households dumps their garbage into the open, thus contributing to deteriorated living environment and conditions in ger areas. There are quite a few households in the apartment areas with the same or even worse living conditions as in ger areas. Such households are the ones living in residential houses, the condition of which has severely deteriorated to the level, that it can t meet the basic housing requirement. People in suburban areas have to put up with noise from power stations and smell and smoke generated by leather factories. A majority of the households living in apartment areas have an access to telephone whereas this is true for less than half the percentage of people in ger areas. The poor access of communications in ger settlements, results in the poor capability to use the communication facilities in case of emergency or other urgent matters.

40 CHAPTER 4. EMPLOYMENT AND INCOME Chapter 4 focuses on employment and income sources of the sample population. This chapter starts with determining the employment rate and the reasons for not working. It then looks at differences between the formal and the informal sector and finally determines sources of income. Key questions - How is the employment participation rate? - How many people are unemployed? How does this look for key-subgroups? - What are main sources of income? Key answers and conclusions: - The employed participation rate is 42.4 percent. - Unemployment rate is 22.9 percent. Unemployment rate is high among the population in ger area (29.5 percent), females (23.2 percent) and migrants (24.6 percent). - There is a clear link between education and employment. - The sources of income are: 67.5 percent of income comes from labour earnings, 18.2 percent from gifts, transfers etc, 8.4 percent from pensions and allowances. 4.1 Employment Description of terminology In order to determine the employment rate, it is useful to first describe the concepts used in this section. First, a distinction between economically active and inactive people can be made. The economically inactive are those who: are not ready to work, including students, people who are looking after children/sick/older people, elderly or retired people; are those not working as a result of lack of the registration or other related documents; and those who are made redundant and are not looking for new jobs etc.. The economically active population are those who are either employed or who are reported to be looking for a job when asked why they do not earn income. The combination of the two is considered to be the labour force. The unemployment rate only refers to those people who are part of the labour force (i.e. not economically inactive) but still not working. This leads to the following categories: People who are in the labour force. The labor force participation rate is calculated by taking the economically active and dividing by the total population of 15 years and over. people who are part of the labour force and employed. The employment rate is calculated by taken the number of people employed and dividing by the total population 15 years and older.

41 People who are part of the labour force and not employed. The unemployment rate is the share of unemployed relative to the labour force. People who are not part of the labour force, i.e. who are not economically active. It should be noted that employment rate plus unemployment rate do not make up 100 percent of the total. This is because they,, refer to two different bases: The employment rate refers to the total sample, and the unemployment rate only to the labour force. The following chart illustrates these concepts and how they are connected. Figure 4.1 Illustration of various concepts in the context of employment Labour force participation rate = (Economically active) / (People of working People of working age (15 and above years old) Economically active (=Labour force) Economically inactive Employed Unemployed (Reasons for not working: Looking for a job ) Reasons for not working: -Students -Looking after children -Too old/retired etc. Employment rate = (Employed) / (People of working Unemployment rate = (Unmeployed) / (Economically active) In the following sections, we only look at people who are 15 years and older. These are people who could be employed. Although it can be argued that this still involves a higher number of people in education, and that, therefore, the minimum age line should be higher, this age line follows the standard of other studies in Mongolia and makes the numbers more easily comparable to them. Description of Results Of the respondents aged 15 years and above, 42.4 percent are engaged in work which generates incomes for them (Table 4.1). Of the employed people 7.0 percent are represented by administrators and managers, 11.9 percent by professionals such as doctors and teachers, 7.4 percent by engineers and technicians, 23.5 percent by service workers, 15.2 percent by sales and trades workers, and 22.5 percent by simple workers like porters, waiters etc. Almost 23 percent of the economically active population or the labour force in Ulaanbaatar are unemployed. The unemployment rate in ger areas is almost 16 points higher than in apartment areas. This means that population available for employment or is looking for a job is higher in ger areas than in apartment areas.

42 55.6 percent of the population aged 15 years and above are economically active population or labour force. The labour force participation rate by sex shows that men exceed women. This is because the women are opting for their pension earlier or are more engaged in studies. In addition, the number of women engaged in housework is relatively higher. Table 4.1 Percentage distribution of the respondents, by employment status according to selected Characteristics Respondent s characteristics Labour force indicators Labour force participation rate Employment rate Unemployment rate Sex Male Female Location Ger area Apartment area Migration status Migrant Non migrant Total Reasons for being economically inactive

43 Figure 4.2 Percentage distribution of economic inactive population, by main reason for not working Migrated/Lack of document 3,8% Too old/retired 26,9% Student 44,5% Sick 10,8% Looking after child/older/sickman 9,6% Other 4,4% Figure 4.2 illustrates the structure of the economic inactive population. In total, 2245 people or 44.4 percent of the total population are economically inactive and almost half of them are students. Being economically inactive because of engagement in studies is related to age the structure of the population. The proportion of people who are sick and not working is higher by 38.6 percent in ger areas than apartment areas. This comparison may be related to the health status of people. Also the proportion of the population aged 65 and above in apartment areas is higher than in ger areas. Soto is the proportion of retired/too old people higher in apartment areas. One reason for migration used to be a search for higher education. Accordingly, the percentage of migrants who are not working because of engagement in studies are 4.5 points higher than that of non-migrants. In the chapter on migration it will be seen, that half of the migrants are not registered. Accordingly, the percentage of migrants who are not working because of the lack of documents and registration is 5.7 times higher than that of non-migrants Employment status of subgroups Table 4.2 illustrates the employment status of the respondents by demographic characteristics. As is generally expected, the highest rate of employment is in the age group of In this age group seven out of every 10 persons are employed. Table 4.2 Percentage distribution of respondents aged 15 and over, by employment status, according to demographic characteristics Demographic characteristics Employment status Number of cases: Total population aged 15 Employed Not working and over Age Less than

44 Table 4.2 Percentage distribution of respondents aged 15 and over, by employment status, according to demographic characteristics and over Sex Male Female Marital status Single Married Living together Separated Divorced Widowed Total Of the population who never married, about 60 percent are under 20 years of age and are living with their parents. Accordingly, the employment rate was found to be the lowest for this age group. The proportion of employed in ger areas is 8 points lower than that amongst the population in apartment areas (Figure 4.3). In ger areas every two out of five persons are employed, while every three out of five persons are employed in apartment areas. This has to do with lower education levels in ger areas (see Figure 4.4 below on the link between education and employment).

45 Figure 4.3 Percentage distribution of the respondents by employment status, according to location and migration status Percent ,5 52,5 59,4 53, ,5 47,5 40,6 46,7 0 Ger area Apartment area Migrant Non migrant Employed Not working The reasons for unemployment among migrants vary. The main reasons are that they cannot work in the traditional way, if they came to Ulaanbaatar due a loss in livestock as a consequence of the winter storms (dzud). Accordingly, lack of working experience in jobs in an urban setting is one of the major reasons for not working. In addition, their low level of the education and no ID/passport or registration in UB also leads to unemployment. (See also the chapter on migration). As the education level rises the employment rate also tends to rise (Figure 4.4). In terms of employment rate, higher educated people are more frequently employed than those without education and those with vocational education. Thus, the survey concludes that better education level serves as one of the requirements for getting employed. Figure 4.4 Percentage distribution of the respondents by employment status, according to education level Employed Not working Higher 68,8 31,2 Special technical/incomplete high 50,7 49,3 Vocational 54,6 45,4 Complete secondary 39,8 60,2 Incomplete secondary 33,8 66,2 Primary 13,2 86,8 Not educated 10,3 89, Percent

46 When the employment rate of the household heads was studied by household characteristics (Table 4.3) half of the heads of households with 1-2 members (46.9 percent) are employed whereas this holds true for almost two third of heads of households with 3-5 members. Households with more members are more vulnerable to poverty for their employment rate is low while the dependency ratio is high in comparison with households with fewer members. Table 4.3 Percentage distribution of the household heads, by employment status of HH heads, according to HH characteristics Household charateristics Employment status of the HH heads Total Number of HHs Employed Not working Number of children None and over Number of people in the HH and over Number of family in the HH and over Registration status in the Khoroo Registered Unregistered Total The employment rate of household heads registered with respective administrative units remains lower compared with the unregistered ones. This could be seen as a sign that people have less chance to find a job through the mediation of respective khoroo, which have an image of not cooperating well enough with employment offices.

47 4.2 Formal and Informal Sector This section looks at the differences between the formal and the informal sector. Table 4.4 shows the percentage distribution of employment population by employment sectors. Of them, 68 percent are engaged in the formal sector and 30 percent are in the informal sector. The remaining two percentages are working abroad. The scope of informal sector is believed to be expanding, fuelled by open market, high rate of urbanization, and market needs. As shown in the survey the livelihood sources of one third of the population, in the capital city, rely on the informal sector. Higher educated non-migrants in the city are formally employed. Low educated migrants work in the informal sector to make their income. Working in the informal sector does not require specific qualification and professions and cannot provide secure social protection and insurance services. (See Table 4.4) Table 4.4 Percentage distribution of employed population, according to the working sector, by selected characteristics Selected characteristics Working sector Total Number of cases: Formal Informal Abraod employed population Sex Male Female Education level* Less than Incomplete secondary Higher than complete secondary Location Ger area Apartment area Migration status** Migrant

48 Non migrant Total Note: * the group of people who are not educated is too small (n=14) to provide relaibale results **-migrants aged As seen in Table 4.5 the medical insurance coverage of the population varies by employment sectors. While out of the employed in the public and private sectors, 8-9 persons in every 10 are covered by medical insurance, this is true, for 5-6 persons in every 10 working in the informal sector such as those working for families on contract, trading in open markets and streets. Table 4.5 Percentage distribution of employed people, according to health incurance by working sectors Working sector Health insurance Total Number of employed Involved Not involved population Formal Informal Total A small number of people (N=41) is working abroad. The number is too small to provide reliable information, and is, therefore, excluded. Table 4.6 illustrates that there is a disparity in wages among the employees in the informal sector. For example, wages of informal laborers who make up 2/3 of the total employees in the informal sector (vegetable growing, petty trades in open markets, scavenging, porters etc.) are lower by 79 percent in comparison with the wages of people engaged in private business. Those people who had lost their livestock have to work for others here in this market to make money. While more businesslike people found a job with the help of their relatives and friends, some of them are living on the money that they made from sales of their livestock back in their hometown. (Focus group discussion with registered migrants) Moreover, it is interesting to note that people working for themselves, i.e. those who run their own business in the informal sector claim to earn more money than people working in the formal sector. Table 4.6 Average wage per month according to the working sector Working sector* Average wage/income (tugrugs) Number of employed population Formal sector

49 ...work for one employer Informal sector work for himself/herself do odds/casual work *Three categories are not shown, as the numbers are too small to provide reliable results. N=27 people work for more then one employer (counted towards the formal sector). N=7 people work for another household member (counted towards the informal sector), and N=41 work abroad (not counted as either the formal or informal sector). According to another study, the scope of the informal sector is expanding because it is relatively easy to get employed in this sector and there are more opportunities to generate income (ILO, The informal sector in Mongolia). Moreover, the working conditions in the informal sector are regarded to be frequently poor. Therefore, the government needs to take actions towards the improvement of work conditions and social protection and insurance services for informal sector employees. 4.3 Sources of Income Of the total households in Ulaanbaatar city, 7.7 percent have one income source and the remaining 92.3 percent have more than one income source. Of the total households almost 82 percent earn wages and salaries while only 10 percent make earnings from household production and services. The percentage of households whose income is comprised of cash remittances, gifts, rentals and interest earnings is almost three times as much of that of the households which generate income from agriculture production. The composition of average monthly monetary income of the survey households is defined as follows (Table 4.7). Table 4.7 Percentage distribution of household incomes according to selected characteristics Selected characteristics Wage Pension/Social allowances Income structure Cash remittance/gifts/ rentals/interest revenue Business income Income of agricultural production Total Average income per month (tugrugs) Location Ger area Apartment area Migration status Migrants Non migrant Total The gap in the average income per month of the households as per the location was relatively higher than as per the migration status. The average income per month of the households in apartment areas, and non-migrant households is higher than that of average income of households in Ulaanbaatar. For instance, the average income per month of the households in ger areas was 1.5 times lower than that of the households in apartment areas and 1.3 times lower than the average of Ulaanbaatar households. Although, the average income per month of the migrant households is 14.3 percent highter than that of non-migrant households.

50 Average monthly income of households is composed mainly from wages and salaries, cash remittance, gifts, rentals, interest revenue, and pensions and social allowances. Wages of more than one third of the people in ger areas are composed mainly of earnings from employment in the informal sector such as petty trade in open markets and streets, chopping wood and providing some small services for others. But wages of more than three fourths of the people in apartment areas are comprised of earnings from employment in the formal sector. Cash remittance/rental/interest earnings makes up a higher share of income sources in apartment areas than in ger areas as it has been quite common to rent flats or rooms in this block. More than one fifth of in-migrants in the city rely on assistance and support from relatives. This percentage is higher by 43.1 percent compared to non-migrant households. Since the employment rate of migrants is lower than that of non-migrants, the share of wages from income sources for the former was 6.4 points lower compared with the latter. And the share of income from livestock herding of total incomes was different by almost 2.0 times. This shows that in-migrants from rural areas live on income from sales of their livestock to sustain their livelihood for some time. Migrants are more likely to fall victim to poverty since income such as remittance from others and sales from livestock do not provide stable and reliable income sources. Conclusion Almost 2 in every 5 people of working age in Ulaanbaatar are employed. Employment rate is high among males, higher educated population, population in apartment areas and non migrants. More than one fifth of the population aged 15 and above are included in labour force and looking for a job or unemployed. Unemployment rate in ger areas is higher than in apartmant areas. This is related to the lower education level of population in ger areas. Considerable sex disparity in labour force participation rate is caused by a higher proportion of women who are classified as economically inactive population. In the meantime, compared to men, women in the city are more engaged in unpaid household duties such as household daily chores and looking after children and elders. Higher educated non-migrants are employed in the formal sector while lower educated migrants tend to work in the informal sector. Of those, working in the informal sector, that have less requirement for education and skills, poor social welfare and protection services, one in two employees are covered by medical insurance. Nine in every ten households have more than one income source. Major income sources of households come from salaries, gifts and transfers etc, and pensions and allowances. The remittance/rental/interest earnings of people in apartment areas are mainly comprised of earnings from rental of flats or rooms in the block. But for migrants, income is comprised of earnings from sales of their livestock and support and assistance of relatives and friends play an important role in getting settled by rural migrants in the city.

51 CHAPTER 5. ACCESS TO EDUCATION AND HEALTH Chapter 5 discusses an access to education and health. This chapter introduces education level of the urban population, accessibility and quality of health services, social welfare and social security services. Thus, the chapter consists of two parts; 1) the education sector services by education level of the adults, school attendance, school drop-out and its causes, and location of secondary school; and 2) the health sector services by health status of population, health insurance, availability of health workers, and location of family clinics. Key questions - Who are less educated? - Who does not have access to social services? Key answers and conclusions - Education level of men in ger areas are relatively low percent of youth aged have completed secondary education. They face a lack of possibilities to get a professional job. - One third of children from households who have migrated to suburban areas have to travel more than two kilometers to get to their schools. - Due to overcrowded classrooms in the city and lack of schools in the periphery increases the load of the secondary schools. - The rate of medical insurance coverage for informal employees, men and persons in ger areas is unsatisfactory. - Of the migrant households in ger areas, 28 percent are located 3 or more km away from a family hospital. - The focus groups point to a need to improve the quality of the social services. - More investment is needed for the construction of new schools and kindergartens in ger areas. 5.1 Education Education level of the adults As mentioned earlier in chapter 3.1.2, the education level of adults (aged 15 and over), covered by the survey, in urban ger areas, are different from that of apartment areas. The difference seems to be immensely big among people with high education. Table 5.1 presents the education level by age group. Youth from ages are still studing in educational institutions; therefore we did not give emphasis on this cohort in the report. Out of people aged years about 38.4 percent had completed secondary education which is the highest share compared to other age groups. They can be considered to be a group at risk. Either they continue with higher education or they try to enter the labour market. Given that they have no professional training it can be expected that they either become unemployed or don t get a professional job. However, the population aged who have special vocational education is relatively low compared to other cohorts. Due to poor economic development the opportunity of getting a job is relatively low for the people of labour-force age with no professional education. Due to economic system crisis, a number of large-scale industries including construction and agriculture have a lesser demand for labour, therefore, there is restricted demand for specialized personnel. Moreover, the prospects for vocational training

52 also declined. Since the beginning of 1990s a number of technical vocational schools were closed. The possibility to train the graduates from 8 th to 10 th grades at vocational schools of all levels in the former socialist countries got lost. However, plenty of public and private universities or institutions were established offering a wide-range topics which served as a catalyst for the secondary school graduates to enter into one of them to get a profession. This offered new opportunities for higher education. Without it, the number of people going into higher education would probably have dropped. Now, they are staying on the same level. Table 5.1 Education level of the adults by age-specific groups Education level By age group (percentage) Total Non-educated Primary Incomplete secondary Complete secondary Technical Vocational Special vocational/ Incomplete secondary High Total Number Sex ratio in education Figure 5.1 shows sex ratio in the education of the adults by level of education and location. Figure 5.1 Sex ratio in education level of the sample population aged 15 and over by location Sex ratio Non educated Ger Primary Uncomplete secondary Apartment Secondary Technical vocational Special vocational/ Diploma Education level High The proportion of males is higher in incomplete high schools and technical vocational schools. In addition, the sex ratio of people with technical vocational or high education, who are living in ger and apartment areas, varies. For instance, there are 46 non-educated males per 100 non-educated females living in apartment areas, whereas there are 63 non-educated males per 100 non-educated females living in ger areas. Also, there are 81 males per 100 females with higher education in apartment areas, whereas it is 58 males in ger areas. It shows that the education level of men living in ger areas is low compared to that of women living in ger areas as well as men and women living in apartment areas.

53 5.1.3 School attendance Table 5.2 indicates school attendance rate of people aged 7-29, which shows that 58.5 percent of males and 61.6 percent of females are attending school at any level. School attendance at any level of education is points higher in apartment areas than in ger areas. According to the Education law of Mongolia, the age of school enrollment is eight; however of the children enrolled in schools aged 6-7, the proportion of girls is higher than that of boys (48.3 percent of girls and 35.3 percent of boys). Of these the rate of enrollment of children living in apartment areas (60.3 percent in apartment areas, 32 percent in ger areas) remains high. In other words, girls compared to boys and children living in apartment areas compared to children living in ger areas tend to be enrolled in school comparatively early. School attendance rate of children aged eight is 95 percent in apartment areas while it is 93 percent in ger areas. Table 5.2 School attendance rate of the sample population aged 7-29 by location and sex according to age groups Location Total Age groups Ger area Apartment area Male Female Male Female Male Female Total School attendance rate of people aged over 18 remains to be high in apartment areas too. For example, school attendance rate of people of years is twice higher in apartment areas than in ger areas. Table 5.3 Percentage distribution of the studying population aged 7-29, by location according to living arrangement during the period of study Living arragement Location Total Ger area Apartment area At home Dormitory* Relatives Rented an apartment Total Number of school attanding population aged Note: * One case living in Care Center is included. The location of students was identified in order to specify the influence of relatives on the budget spending of households regardless of household members. As Table 5.3 shows 89.5

54 percent of students are staying at home, but for the people living in ger areas (91 percent) it is slightly higher compared to the apartment areas (87.6 percent). The survey shows that 9 percent of students are likely to rent apartments or live together with relatives in their houses. In total, 4.8 percent out of all children or 88 children aged 7-18 years, covered by the survey, dropped out of school, of whom majority are males living in ger areas (see Figure 5.2). Figure 5.2 Percentage distribution of thc school dropped out chidren aged 7-18 by location and sex (N=88) Ger are 86% Apartment area 14% Location Male 67% Female 33% Sex percent Children mainly dropped out of schools when they were in primary (51 children or 60 percent from 1-4 class) and secondary (30 children or 34 percent from 5-8 class) classes. Major causes of school drop-out were extreme poverty of families (26 percent), health status (25 percent), and a lack of money (13 percent). In total, 8 children living in ger areas dropped out of school due to no registration or ID cards while no similar case has been reported in apartment areas. As migrants do not register in their newly settled residential areas and as parents do not get children s certificates from the previous schools, children drop out of schools. 21 children (24 percent) who dropped out of school were involved in non-formal training. The rate of involvement in non-formal training in ger areas (23.7 percent) is similar to that of apartment areas (25 percent) Tuition fees and its types 42.3 percent of students attending schools at all-levels pay tuition fees but it varies by location (Table 5.4). Fifty-seven percent of people living in apartments pay tuition fees, whereas only 30.7 percent of the ger area inhabitants pay that. Depending on training quality, class capacity and learning environment, parents are more likely to send the children to private schools for general education or to schools for advanced studies (state-owned but with tuition fees). However, children of the households with sufficient income are likely to go to private schools. Although the Constitution of Mongolia declares to give complete secondary education free of charge for all, it can not be always realized in life. Due to a number of reasons such as uniforms of pupils, necessity of various textbooks and notebooks, celebration of learning,

55 class fund and class restoration, a big sum of money is collected by schools which make a burden to the poor households. The celebration of learning reveals the income difference of children s family since children from wealthy family are likely to celebrate it spending much money which makes children from poor family excluded. Therefore, poverty strongly influences on the psychology of children reducing their interests in learning. (T. female, age 50, teacher, Songino-Khairkhan district) When I wanted to transfer my grandchild to the second year of urban school it was required tugrugs. As authorities of the school explained, it was required for extra practice since rural children are usually poor in study. Thus, I gave that money as they were scaring me not to promote my grandchild to the second year. But still my grandchild has not been given any extra practical lesson. (B. female, age 65, pensioner, Songino-Khairkhan district) Table 5.4 Percentage distribution of the studying population,who pay tuition fees, by location Required tuition Location Total Ger area Apartment area Not paid tuition fees Paid tuition fees Total Number of the total studying population percent of students, who pay tuition fees, are sponsored by their families while 11 percent of students pay their tuition fees with the support of the state fund. There is no big difference in the source of finance for people living in ger areas and in apartment areas Access to school and kindergarten Figure 5.3 shows distance between households and secondary schools by district, location and migration status. The average distance of the households to the nearest secondary school for the respondents covered by the survey is 1.4 km. The longest average distances have been observed in Songinokhairkhan (1.7 km) and Nalaikh (1.7 km) districts whereas the shortest average distances have been found in Sukhbaatar district (0.9 km). The survey showed that 35 percent of households in Nalaikh district, 26 percent of households in Bayanzurkh and Songino-Khairkhan district, and 22 percent of households in Chingeltei district are a distance of two or more km away from the secondary school. The above mentioned districts covering many ger khoroos are located in the outskirts of city. As ruralurban migrants are moving to the outskirts of city year after year, in-school children of some households of ger areas have to walk more than two km of way.

56 Figure Percentage distribution of the households by distance of two and more 5.3 km away from the secondary school according to selected characteristics Khan-Uul (N=100) 9 Sukhbaatar (N=350) 13 Chingeltei (N=250) 22 Sonigokhairkhan (N=500) Bayanzurkh (N=200) Nalaikh (N=100) 35 Non-migtant (N=1139) 19 Migrant (N=176) 34 Apartment (N=684) 8 Ger (N=816) It can be also seen that 34 percent of migrated households living in ger areas reside two or more km away from the nearby school percent children of the households involved in the survey go to school or kindergarten located in their sub-district. However, the rest of 26.2 percent or 439 children go school or kindergarten located in another district or in sub-district due to poor training quality and resource basis (72.1 percent), overloaded capacity of classroom (13.1 percent), unwillingness of transferring to another school (12.1 percent) and no registration or ID cards (9.3 percent). The major challenges posed to the children of ger areas, who go to school or kindergarten located in another district or in sub-district, are no registration or ID cards (12.7 percent) and far remote distances between home and school or kindergarten (7.4 percent). It is assumed that these chalanges of ger area inhabitants are greater compared with those of residents in apartment areas (4.3 percent and 1.8 percent respectively). During focus group discussions the participants mentioned that rural-urban migrants with no registration or IDs make donations, namely money to school in order to enroll their children. Other children of households living in ger areas stay with their relatives or in other familiar families because of the long distance from the school. The above mentioned fact shows that there is a difficulty in access to school or kindergarten for migrants and inhabitants of ger areas. Due to overcrowded classrooms in the city and lack of schools in the periphery, increases the load of the secondary schools. Although an estimated number of students in one classroom are from 25-30, in actuality there are over students studying in one classroom; namely, three students per desk. This is due to increase in migration. Overcrowding in classrooms has negatively influences not only on quality of study but also on the possibility of spreading contagious diseases.

57 As our school has 3 rounds of lessons a day it is impossible for students to spend leisure time properly. The excess of loading very negatively influences on quality of study. For instance, when a lesson of the first year pupils is over teachers are supposed to give extra time for reading until their parents come. But in fact as soon as the first round lessons finish the second one starts in 5 minutes which gives no possibility of doing reading practice (Ts. female, age 35, teacher, Bayanzurkh district) Participants in individual discussions emphasized that experienced teachers of secondary schools are likely to work for private schools that offer comparatively higher salary, whereas recently graduated, inexperienced and poor educated teachers tend to work for schools on the periphery. Moreover, teachers and citizens of districts stated that the schools are poorly constructed; there is a lack of training facilities, dirty conditions of toilets and an inadequate and unsupportive learning environment. 5.2 Health Health status When we asked people whether they have any chronic conditions that was detected by a doctor or medical specialist, 15.3 percent (16.3 percent of females while 14.2 percent of males) out of respondents answered yes. The rate of chronic illnesses remains high among residents of apartment areas (16.3 percent), non-migrants (17.1 percent), and informal sector workers (15.9 percent). Furthermore, the older the person, there is a higher incidence of chronic illnesses. As living standards decreases the rate of people, who were diagnosed of such chronic illnesses by doctors or health workers, tends to be increasing (Table 5.5) Involvement in health insurance Health insurance covers the following services; examination, treatment, but excluding the medicine expenditure. Health insurance is provided by the family doctors as well as public hospitals for the citizens registered at districts. Moreover, health insured people may get cheaper services at the government clinics. A total, of 82 percent of people interviewed, had health insurance. According to the survey 78 percent out of people living in ger areas have health insurance which is 8 points lower than those in apartment areas. In other words, for people staying in ger areas, the proportion of people having health insurance is low (Figure 5.4). The low percentage of people having health insurance in ger areas is connected to some extent with the prevalence of informal sector labour. However as family doctors say, it reduces their possibility of access to health service. The poor who are mainly concerned about their daily foodstuff or the rich who approach the hospitals that have high quality service do not pay health insurance fee. Probably for the rich there is no need for insurance. In case of illness they simply go to private clinic with good service paying for service, whereas the poor tend to spend their little money for food solely.

58 (O. female, age 41, family doctor, Bayanzurkh district) As people are not much concerned about health insurance the family clinics are required to involve the residents of respective districts in health insurance on a whole. And depending on its outcome the family doctors are duly paid, for instance, the payment is deducted in case of unsatisfactory results. Nevertheless, there is no allocated financial resources for advocating the benefits of health insurance among family clinics. Figure 5.4 Percentage of people, who have health insurance, by location Ge r 78 Apartmen 86 Total Health insurance fees are paid for 74 percent of the total population by the government or respective organizations. Of this percentage, for residents (76 percent) of ger areas, the proportion is higher than that (72 percent) of apartment areas (Figure 5.5). This might be related with the fact that health insurance fees of children and of the elderly are paid by the government since there is a comparatively high percentage of this group of people especially, children in ger areas (Chapter 3, Population pyramid). Figure 5.5 Percentage distribution of the insured people by location according to type of the payment Ger (N=2397) 76 State/organization Individual 24 Apartment (N=3179) Total (N=5576) Although it is said that non-fee treatment is available if you have health insurance, the poor who earn less income, do not have access to these clinics and to quality health service due to standardized rates at the central and district hospitals, viz. consultation with tertiary specialists or check up through ultra-sound scanner and X-ray.

59 The possibility of getting health insurance directly depends on the standard of living of households (Table 5.5). For instance, 31.4 percent out of extremely poor households do not have health insurance which is twice greater compared with non-poor low level households. Table 5.5 Health status by living standard Selected characteristics Living standard Very poor Poor Non poor (low) Non poor (medium) Non poor (high) Total Chronic illness Have Don t have Number Total Health insurance Have Don t have Number Total Table 5.6 shows the percentage of people who do not have health insurance by location.. Number of males (20.2 percent) who do not have health insurance is greater than those of females (15.1 percent). It might be explained by the following reasons that men pay less attention to their health or they rarely go to hospital. Based on the level of education, people with technical vocational education are unlikely to have health insurance. This is because of the change in the economic system, where numerous graduates of technical vocational institutes were sacked resulting in them getting work in the informal sector. Unfortunately, the proportion (43.6 percent) of informal sector workers, who do not have insurance is 31.6 points higher than that of formal sector workers. The major reasons why people do not having health insurance are as follows: the lack of money (32.7 percent), no job (24.0 percent), no registration (18.7 percent), no need to have insurance (13.4 percent), and no interest (8.5 percent). These causes appear to be similar for both men and women. However, as Kazakh ethnics have lack of money and low income they can not have health insurance. The proportion of people of this ethnic group is the highest compared to other ethnic groups. Access to health services was evaluated based on the distance between family clinics and the households. Figure 5.6 shows that a number of households located three or more km away from family clinics is very high in Bayanzurkh district (19 percent). For non-migrant households, the distance from family clinics was less than that of migrated households. For instance, 28 percent out of migrated households are residing more than three km away from family clinics, whereas for non-migrated this percentage was only 10 percent.

60 Table.5.6 Percentage of respondents who not have a health insurance, by location according to selected characteristics Location Total Selected characteristics Ger Apartment Percent Number Percent Number Percent Number Sex Male Female Age groups Less than and over Religion Khalkh Kazakh Durvud Buriad Bayad Other Education (N=1162) Non educated Primary Uncomplete secondary Secondary Technical/vocational Special vocational/ Diploma High Employment status (N=1130) Employed Not working Working sector (N=466) Formal Informal Migration status (N=1126) Migrant Non-migrant Total

61 Figure 5.6 Percentage distribution of the households by dictance of 3 and more km from the family clinic away, according to selected characteristics Nalakh (N=100) Sukhbaatar (N=350) Chingeltei (N=250) Songinokhairkhan (N=500) Khan-Uul (N=100) Bayanzurkh (N=200) Non-migrant (N=1139) Migrant (N=76) Apartment (N=684) Ger (N=816) Residents of apartment areas (96 percent) and non-migrants (95 percent) have more possibility of getting access to family clinics, health services and professional counseling services (Figure 5.7). However, only 78 percent of migrants answered positively about access to these kind of health services. Table 5.7 Percentage of people who have access to health service by selected characteristics Non-migrant (N=4024) Migrant (N=688) Apartment (N=2777) Ger (N=4070) Due to heavy workload, family doctors (especially of ger areas) do not often visit their respective patients. In addition it is difficult for the elderly and the poor to go to the family clinics as they need to go by bus and they do not have money to pay for the buss fare.. As our family clinic is located far away I need to catch a bus and pay for the bus fare. But for the elderly like me it is also impossible to walk to clinic. It would be better to have a family clinic located in a short distance. (B, female, aged 65, pensioner, Songino-Kharikhan district)

62 In total, 83 percent out of households of ger areas often consult with family doctors while it is 56 percent in apartment areas. There is still a lack of accessibility and availability of health services as a result of the rapid growth of population of the city. The district health alliance was established in 1983 with over 300 beds. Due to a shortfall of beds, patients who are in need of inpatient care have to wait at least 7 days. (I. female, aged 50, family doctor, Bayanzurkh district) Table 5.7 shows that 16.2 percent of the extreme poor people and 9 percent of poor people lack access to health service and counseling, whereas for non-poor households it is 4-6 percent. Prior to counseling, one is required to meet a minimum standard of hygiene; namely, to have a bath and wear clean clothing. But as a family doctor says, for the extreme poor people, it is the main reason for which they often do not go to hospital. Table 5.7 Percentage distribution of the population, by living standard according to access to health service Access to health service Living standard Very poor Poor Non poor (low) Non poor (medium) Non poor (high) Access to health service Yes No Dictance of HH from the family clinic away Less than 3 km and more km Total Number Total Social welfare and social security service As chapter four emphasized, for vulnerable groups of people, e pension funds and allowances provided from the social welfare fund are considered to be the main income source of their livelihood. Some elder people complained that although there are a number of families that are financed by pension or allowance of one or two family members there is still a big gap in the pension system of the elderly. The pension gap should be eradicated. Although I have worked for the state-owned sector for 37 years I have a pension of tugrugs. The amount of pension has increased since If I had my pension fixed after 1995 I would have a pension of tugrigs. (Focus group discussion among the migrated group of people with registration) Some vulnerable groups of people do not have access to social security service. According to a provision of the law on Social welfare, vulnerable groups of ger areas have to be provided with fuel, namely logs and coal, whereas inhabitants of apartment areas are granted a discount

63 on monthly fees. Even though citizens claim that the inhabitants living in a house with the same facilities like gers are not granted any discount in fuel. We have not been provided with fuel such as logs or coal that is provided in accordance with the social welfare law as they cover our sub-district in the apartment area. But in fact, the residents living in houses with poor conditions have no heating system which makes them live in circumstances like in ger area. (D, female, aged 53, district governor, Songino-Khairkhan district) Environment for healthy life In recent years due to overwhelming settlements of population that has exceeded the carrying capacity in Ulaanbaatar, there have been numerous negative consequences in the areas of ecology and housing facility. Majority of people involved in the focus group discussion were concerned about negative environment consequences, such as air pollution, rubbish, mess of buildings, fence and dam among others. All agreed that air pollution is the most urgent issue for inhabitants of both ger and apartment areas on the periphery. Air pollution has badly influences on the health of people; furthermore is has even become one of challenging issues of human rights. Air pollution is much worse in ger areas. When we go shopping in the evening the hair absorbs the smell of pollution. Early morning there is smog and the visibility is very poor. It is just too difficult to breathe. (Focus group discussion among the migrated group of people with no registration) Rubbish of factories and industries and their impact on activities exert considerable negative influences on relaxed life and health of the people. As the industries of Ulaanbaatar are usually located in residential areas, it has adverse effects on the health and comfort level of residents. For instance, noise of power plants and the smell of hide processing factory etc. When it is windy from the west it stinks as there is a water treatment plant located in the western part of Orbit. The noise of nearby power plant also bothers. Sometimes we have to scream in order to speak even when we are staying in one room. It would be much better if the noise of power plant is ceased during night if not during the day time. It is a clear evidence of violation of human rights since it does not allow us even to sleep. (E. female, aged 29, social worker, Songino-Khairkhan district) Inhabitants of ger areas are very concerned about their comparatively poor water supply. They have to line up for 1-2 hours to get water. Sometimes due to a shortage of water they can not even get water after queuing. Open-air rubbish in the streets deteriorates the surrounding environment. In particular, for ger areas on the periphery rubbish remains an urgent issue. In view of it, local administrative officials and some other citizens blame the public conscience. Conclusion The fact that the education level of population in ger areas is relatively low, particularly among men, draws the attention. School enrolment is far higher in apartment areas than in

64 ger areas and school entering age is lower for girls compared with boys. The school enrolment of young persons (23-29) is two times lower in ger areas. There is little opportunity for people in ger settlements to send their children to quality and private schools. The cost of secondary schooling heavily impacts on the children of poor households and serves as a reason for some of them to drop out of school. Investment in the social sector, which should follow the intensive growth of the urban population is extremely inadequate in ger settlements. For example, one third of children from households migrated to suburban areas have to travel more than two kilometers to get to their school. This long distance poses difficult circumstance for children especially those of primary grades. The population in ger areas have poor access and availability of schools and kindergartens. Because of the poor quality of teaching and material base coupled with overloaded classes households can t enroll their children in schools in their respective districts and khoroos. Stemmed from migration, the overcrowding of a class is excessive with more than twice than the accepted norm. Boys in ger areas are more likely to drop out of school. This is not only due to economic difficulties but also due to parents ignorance of getting the school transfer certificate from originating areas. Although there are a number of projects and programmes to address the issue of school drop out, only 1/5 of the school drop outs were covered by informal training. The rate of medical insurance coverage for the population is unsatisfactory (particularly for informal employees, men, and persons in ger areas). It wouldn t be exaggerating to say that placement of charges on specialized medical services, not only private hospitals but also by state hospitals deprived the population, who barely meet their basis needs, particularly the poor ones, of the access to medical services. Apart from poor access to medical services, one third of the migrants are not medically insured. More than half of the percentage of those who are insured pay the fees on their own. Yet, the health of the migrants is slightly better than that of non-migrants. This might be explained by two angles. One is the higher likelihood of engagement in migration by relatively healthy people. Another is migrants might have been unable to have their health examined due to poor access of specialized medical check-up in their original place. Of the migrant households in ger areas, 28 percent are located 3 or more km away from family hospitals. Thus, it is impossible for them to benefit from medical services especially family hospital services in the real sense. Furthermore, the high work load of family doctors especially in ger areas adversely affects the access to medical services.

65 CHAPTER 6. AWARENESS, INFORMATION AND SOCIAL CAPITAL Chapter 6 describes an access to information and social capital of the population in Ulaanbaatar. This chapter is about community participation of urban residents, the relationship among relatives and friends, khureelel or social networks, awareness of projects and programs aimed at increasing incomes and improving livelihoods of people, as well as benefits from these projects and programs. In short, the chapter looks at the role of the social capital in improving livelihoods of people. Key questions - How is the information on socio-economic project as well as the programmes advertised in the city? - How is the situation of social capital for various sub-groups Key answers and conclusions - Even though more than half of households were informed of government projects and programmes for the support of livelihood capacity of the population, the beneficiaries of them are estimated to be less than 10%. - The percentage of beneficiary households is higher with green revolution programme and savings and loan projects than other programmes. - There appears to be inadequate information about the availability of programmes for improvement of livelihood of the population in the city. - Low income households have less opportunities to turn trainings into benefits from the training. Even though there is no difference between households in terms of training participation rates, poor households are almost three times less likely than better off households to state that they benefited from the training. - Well to do households participate less in community work - The kinship/khuree supports the livelihood of households, but for migrants, people in ger area, poor and very poor households a kinship/khuree is very limited Participation in the community Information on projects and programs and their benefits In recent years, numerous projects and programs have been implemented by the government and international organizations for the improvement of livelihoods of the people. Questions to identify urban households knowledge and awareness of these projects and programs were included in the questionnaire, and results are summarized in Table 6.1. The survey shows that two-third of households was aware, in one way or another, of projects and programs implemented to improve people s livelihoods. One household knows of two programs and projects on average; and about one-third of total households knew of large-scale projects and programs by multiple answers. Among the projects and programs the Green Revolution program stands out as the most widely known (60.2 percent), almost twice as many compared with other projects and programs. The awareness of the Household livelihood capacity support program is the poorest (27.4 percent). When we look at the indicator by district, households of Khan-Uul and Sukhbaatar districts had more knowledge of projects and programs, whereas households of Bayanzurkh and

66 Nalaikh districts had the least knowledge. For instance, one household knew of 1.8 projects and programs on average in Bayanzurkh district, which is 1.3 times lower than the standard of the capital. However, the awareness among households in Khan-Uul district (on average 3.1) was 1.3 times much higher compared to the average level in Ulaanbaatar. The share of households who did not know of any projects or programs was the highest in Bayanzurkh district (40.5 percent) and the lowest in Khan-Uul district (20 percent). Table 6.1. Percent of households by knowledge of community programs/projects on supporting livelihoods, according to selected characteristics Selected characteristics Programs/projects Household Livelihood Capacity Support Program Green revolution White revolution Restocking livestock Saving/Credit Employment promotion Others Don t know any programs /projects Average number of programs /projects known Number of all households District Bayanzurkh Nalaikh Sukhbaatar Songinokhairkhan Khan-Uul Chingeltei Location Ger area Apartment area Living standard Very poor Poor Non-poor (low) Non-poor (medium) Non-poor (high) Migration status Migrant * Non-migrant * Total Note: * - Migration status had been defined for population years old When we look at numbers by location, households of ger areas ( percent) have not been much informed about the projects and programs compared to those in apartment areas ( percent). The higher the living standard of a household, the more likely was the household to know of projects and programs. For example, half of the very poor households were aware of implementation of projects and programs while for the same figure for the wealthy households (non-poor, high income) were 80 percent. Moreover, the average number of programs and projects known by very poor households (1.3) was twice as little compared to that of wealthy households.

67 The awareness by migrant households was less than that of non-migrant households. Nonetheless, their knowledge of the livestock-restocking project was higher than that of nonmigrant households, due to the focus of the project solely on rural areas. One interesting result of the survey was that the knowledge of small-scale projects was high among households of ger areas (3.2 percent) and remote Nalaikh district (7.0 percent), as well as among very poor (4.7 percent) and poor (3.9 percent) households. It can be inferred from here that small-scale projects and programs have more benefits for the disadvantaged households compared with large-scale projects. Projects and programs ranked as follows by popularity (by percentage of households who knew about the project) first, Green Revolution program (60.2 percent), second, savings/credit project (38.9 percent), and third, the livestock restocking project (37.5 percent). The knowledge of the Household livelihood capacity support program and of the Employment promotion program was comparatively low among all groups of households, probably due to the focus of the advertisement campaign for these projects solely in selected districts and subdistricts of the capital. The majority of households (91 percent) said that they gained no benefit from any project or program at all (Table 6.2). Table 6.2. Percent of households by getting benefits from community programs/projects on supporting livelihoods according to selected characteristics Benefited from programs/projects Selected characteristics Household Livelihood Capacity Support Program Green revolution Saving/ Credit Employment promotion Others Didn t benefit Number of total households District Bayanzurkh Nalaikh Sukhbaatar Songinokhairkhan Khan-Uul Chingeltei Location Ger area Apartment area Living standard Very poor Poor Non-poor (low) Non-poor (medium) Non-poor (high) Migration status Migrant * Non-migrant * Total Note: There are no any households who got benefit from White revolution and Restocking Livestock projects among the households covered in the survey * Migration status had been defined for population years old The share of such households was higher in central districts of the capital, such as Chingeltei (93.6 percent) and Sukhbaatar (93.4 percent), compared to other districts (80-92 percent).

68 Furthermore, this percent was higher in apartment areas than in ger areas. Participants in focus group discussions pointed out that households living in sub-urban apartment areas are likely to be left out of projects or programs. The poverty rate is comparatively high in our area. Although it is lower than in ger areas, it is high compared with that in apartment areas. However, all projects and programs in support of livelihoods of the people are mostly focused on people in ger areas. We have many poor households in our khoroo, but it is common for us to be left out. (D, female, 53 years old, khoroo governor, Songino-Khairkhan district) Of migrant households, 93.8 percent answered to have gained no benefit from any projects or programs, which is 3.5 percentage points higher compared with non-migrant households. Since projects and programs reach out to people through khoroos, unregistered or new households are less likely to benefit from them compared with non migrant-households. There are relatively many (4.7 percent) households in Ulaanbaatar that have benefited from the Green Revolution program. They are followed by households (2.9 percent) that benefited from the savings and credit project. At the same time, the percentage of households who benefited from the Household livelihood capacity support program and the Employment promotion program was less than one percent. The qualitative survey showed that there are many people who would like to be involved in these kinds of projects and programs. It is very difficult to find a job. They ask for good education and qualifications and there is impossible. If I take whatever job I can find, the payment is so low that I will spend all my money just to go by bus from here to the city centre. Instead, I would cook meals for my children at home. Jobs are really needed. (J, female, 43 years old, unemployed, Songino-Khairkhan district) Households living in apartment areas and non-migrant households get more benefits from large-scale projects or programs with the exception of the Green Revolution program. But households mostly living in the peripheral districts with high poverty rate, such as Nalaikh (8.0 percent), Bayanzurkh (5.0 percent), Songino-Khairkhan (4.4 percent) and ger areas (5.8 percent), as well as non-migrant households (4.0 percent) are more likely to benefit from the Green Revolution program. Figure 6.1 shows households who know of projects and programs and those who benefited from them, by living standards. The figure illustrates that more than a half of households in all groups are aware of projects and programs; whereas the number of beneficiaries is much less (under 10 percent). Although the extent of benefits of projects and programs is generally equal, the very poor group (9.4 percent) and the non-poor, high-income group (10.4 percent) have benefited more, whereas those in the middle (non-poor, medium income percent) have benefited less.

69 Figure 6.1. Percent of households, who had information about and got benefit fro community programs/projects according to household living standards non-poor (high) non-poor (medium) non-poor (low) poor very poor had information about programs/projects got benefit from programs/projects per cent Participation of citizens in community activities 4 The quality of people s life is dependent not only on incomes of households and individuals, but also, to some extent, on their participation in social life. Therefore, the data was collected through the questionnaire to enable to assess how poor are households in terms of their participation in social life. Out of all households, 55.7 percent said that their participation in community activities announced and undertaken at district/khoroo levels (Table 6.3). This shows that almost half of households in Ulaanbaatar are inactive in terms of their participation in community activities. When we take participation in community activities by household characteristics, there are no big differences in location, since over half of households living in both ger areas and apartment areas (55.5 percent and 55.8 percent respectively) participate in community work. However, for migrant households (32.4 percent), the figure is 25-percentage points lower than for non-migrant households. Participation of households registered in their khoroos (58.8 percent) was almost twice as high compared to that of unregistered households. Large households participate more in community work. As seen in Chapter 4, large households are more likely to be poor. Moreover, poor people tend to have more children, so the number of people living in a household is greater. Thus, the poorer the household, the more likely it is to participate in community activities. Furthermore, the younger the age of head of household, the lower the participation of households in community activities. For instance, for household with the head of household aged under 35, the above indicator is 36.3 percent, whereas for household with the household head aged 65 and over, it is around 70 percent. However, there is no great difference in gender-specific pattern of household head. We can conclude that smaller and younger households are less likely to participate in community activities. 4 Meetings of residents of khoroo, cleaning company and advocacy activities etc.

70 Table 6.3. Percentage distribution of households by participation in community activities, according to selected characteristics Selected characteristics Whether participated in any community activities organised by khoroo/district Total Number of total households Yes No Location Ger area Apartment area Migration status Migrant * Non-migrant * Whether registered in khoroo Registered Unregistered Number of people in the household and more Living standards Very poor Poor Non-poor (low) Non-poor (medium) Non-poor (high) Sex of the household head Male Female Age of the household head Under and over Total Note: * Migration status had been defined for population years old As the living standard of a household improves, the participation in community activities decreases. For example, 60.4 percent of the very poor households and 57.7 percent out of poor households said that they participate in community activities, while the same figure it is less than 50 percent (47.8 percent) among wealthy (non-poor, high income) households. It might be that households with less income are more likely to participate in community activities undertaken at the district or khoroo level to get assistance or support by introducing themselves.

71 6.1.3 Participation in technical training 5 Based on the assumption that involvement of a member from a household in technical training will raise the standard of living standard of the household, data was collected through the questionnaire. Table 6.4 shows the results of the survey by characteristics of households. Table 6.4. Percentage distribution of households, by participation in any technical training, according to selected characteristics. Selected characteristics Participation in any technical training by any member of household Total Number of total Yes No households Location Ger area Apartment area Migration status Migrant * Non-migrant * Registration status at khoroo Registered Unregistered Number of people in the household and more Living standard Very poor Poor Non-poor (low) Non-poor (medium) Non-poor (high) Sex of the household head Male Female Total Note: * Migration status had been defined for population years old One-fifth of households answered that at least one household member attended technical training. According to characteristics of households, members of households living in apartment areas (22.4 percent) and members of non-migrant households (21.9 percent) are more likely to attend such technical training. At the same time, for households with registration in their khoroos (21.0 percent) attendance rate was higher than that of unregistered households (17.4 percent). 5 Short-term apprentice or professional trainings. For instance, hairdresser, cook, dressmaker etc.

72 As the number of people living in a household increases, the participation in various training programmes increases too. For example, over 10 percent of households with 1-2 members have at least one person trained, whereas for households with 3-5 members this figure is 20.0 percent and for households with 6 and more members, it is 27.8 percent. It is probably due to targeting of short-term technical training programmes aimed at poverty alleviation and reduction of unemployment to vulnerable households. By sex, members of female-headed households (19.4 percent) are less likely to participate in training compared to male-headed households (21.1 percent). Table 6.5 shows how participation in short-term technical training programs impact households livelihoods. Table 6.5. Percentage distribution of households, whose members participated in training, by the impact of training on their livelihoods, according to selected characteristics Livelihoods improved after attending Selected characteristics technical training Total Yes No Location Ger area Apartment area Number of households whose members participated in training Migration status Migrant * Non-migrant * Number of people in the household and more Living standards Very poor Poor Non-poor (low) Non-poor (medium) Non-poor (high) Sex of the household head Male Female Total Note: * Migration status had been defined for population years old. Less than half of households (42.9 percent) believe that their livelihood improved to some extent because of the training. The percentage of households that were able to improve their livelihoods is relatively high among; households living in apartment areas (49.7 percent), households with few members (52.9 percent), and among male-headed households (44.8 percent). However, there is not much difference between migrant and non-migrant households (43.8 and 43.4 percent respectively).

73 Figure 6.2 shows the percentage of those households whose members attended various training programmes; and the corresponding percentage of households, who believe that their livelihoods improved as a result of trainings. The numbers show that the rate of households enrollment in training programmes is generally alike (17-24 per cent) across various levels of living standards. However, the way households were able to use the training to improve their livelihoods varies substantially. Figure 6.2. Percents of households, whose members attended training an improved livelihood as a result of training according to living standard very poor 60 non-poor(high) poor households, whose members attended training households who improved livelihood thanks to training non-poor(medium) non-poor (low) For instance, 18 percent out of the very poor households had someone of its members participated in training, while only 21 percent of them were able to improve their livelihoods using the knowledge and skills they acquired. Meanwhile, 24 percent out of wealthy households had at least one of their members participate in this kind of training, with 57 percent of them saying that they improved their livelihoods as a result. Thus, the extent to which training programmes affects livelihoods of households in different living standard groups show that poor households, in general, are less able to utilize their knowledge and skills, even after they had been trained. One of the reasons could be the lack of financial resources, because even after undergoing vocational training they still lack money to run businesses and to improve their lives. Another reason could be, although it does not show up from the survey, that wealthier people have access to higher quality and a wider choice of training possibilities, whereas the poor people s choices are constrained by only those programmes that donors or the government provide. 6.2 Social networks For countries with populations as small as Mongolia s, it is common that social networks are powerful and pervasive, be it networks of kinship or friendship. In order to investigate whether such kinds of networks really exist, we asked households whether they have any

74 acquaintances who may help them in their daily life. Those households who said they do were asked who those people are and whether they help their daily life. These questions are useful to study how acquaintances or khureelel helps households. Literally, khureelel means a social circle. It includes both blood relations (e.g. your parents, children, sisters/brothers etc) and dry relations. A dry relation has equally binding responsibilities towards you as a blood relation, if you ask for help. The dry relations can be old classmates or your parents old classmates, friends of the family etc. If you ask a khureelel for help/ money/ connections they in principle have to do something. However, it should be seen as something you can approach if you re in difficulty or need assistance, not something that will just be there giving you food as soon as you re hungry i.e. you basically have to ask for it. In that sense it can be a social security network you can draw on when times are rough, it can be what gives you a job through connections, and it can be the place where you get credit. Table 6.6. Percent of households, who supported by kinship in daily life, by occupation of kinship according to selected characteistics Occupation of kinship Selected characteristics Doctor Nurse Secondary school teacher Administra tive officer Director/manager of private company Policeman Inspector of market Number of household Location Ger Apartment Living standards Very poor Poor Non poor (low) Non poor (medium) Non poor (rich) Migration status of head of the HH Migrant * Non migrant * Sex of heads of the HH Male Female Education of heads of the HH Non educated Primary Uncomplete secondary Secondary echnical/vocational Special vocational/ Diploma Employment status of heads of the HH Employed Not working Total * Migration status of heads of the HH aged years old. So the total case of them is lower than 1500 Table 6.6 shows detailed information about households, who have acquaintances that hold jobs/ positions or possess qualifications, which may be, useful to the household s daily life in

75 question. The types of jobs/ qualifications include doctors, nurses, secondary school teachers, and government administrative officers, directors/managers of private companies, policemen and market inspectors. The table shows that percent of all households said that they have acquaintances who may help in big ways to the household. The percentage of households of such acquaintances who work as doctors (46.1 percent), nurses (35 percent) and secondary school teachers (31.5 percent) is the highest, whereas the percentage of households who know policemen (20.5 percent) and market inspectors (8.8 percent) is the lowest. The percentage of households with such acquaintances is higher among households living in apartment areas ( percent) and among non-migrant households ( percent)., as well as among those households where the head of the household is employed ( percent). In addition, as the education level of the head of household increases, the likelihood of knowing useful people increases also. For instance, among those households where the household head is uneducated, only percent have such acquaintances, whereas for households whose heads have higher education the same figure is twice as high. By living standards, wealthy households are more likely to have useful acquaintances ( percent), which is three times greater than that of the poor households and those with medium living standard. In view of above, for the poor and very poor households, migrants and ger area households, the opportunities to obtain assistance from acquaintances are more limited. Table 6.7 gives information on whether the khureelel (it may not include only the acquaintances mentioned earlier) supports the households in their daily life. Table 6.7. Percentage distribution of households, by supported of "khureelel" in their daily lives, according to selected characteristics Whether supported by khureelel in Selected characteristics daily life Total Number of total households Yes No Location Ger area Apartment area Migration status Migrant * Non-migrant * Number of people in the household and more Living standards Very poor Poor Non-poor (low) Non-poor (medium) Non-poor (high) Sex of the household head

76 Male Female Employment status of household head Employed Not working Total Note: * - Migration status had been defined for population years old. As the table shows, almost two-third of all households get support from their Khureelel in their daily lives. This percentage is 3.2 percentage points higher for households in apartment areas (63.6 percent) compared with those in ger areas, it is also 6.6 percentage points higher in female-headed households (66.9 percent) compared to male-headed households. But the percentages for migrant and non-migrant households are same (60.8 percent). The survey finds that the fewer the number of household members, the greater is the likelihood of getting support. By living standards, the percentage (45.3 percent) of very poor households that get support for daily life is 3.7 percentage-points lower than that of non-poor (but with low income) households. At the same time, it is 25.7 percentage-point lower than the percentage of non-poor but with average income households, and 21.3 percentage-point lower compared to that of wealthy households, respectively. To sum up, the lower the livelihood of household the smaller the likelihood of getting helps. From the above, it can be concluded that households better off in various respects, such as households living in apartment areas, households with few members or households with higher income, where heads are employed, tend to get more support in their daily lives. However, it is necessary to study further whether these households are living in fact much better with support of khureelel. Although the sample survey did not attempt to study the means of providing support to households, the qualitative research, and particularly, the discussion among migrants provided some information on this. The source of incomes of migrants is very similar to that of ger area people. However, migrants are likely to receive support from relatives or from the countryside. Furthermore, it s easier for those migrants who have friends or relatives in the city, but for those with few relatives or friends, it is very difficult. (Report of focus group discussion, page 17) The qualitative research shows that although the scope of support of khureelel is very broad, they tend to provide more support for migrants in finding a new job or in helping them settle down and start a business in their new place. I had borrowed 1 million tugrugs with an interest rate of 8-9 from relatives to run a small private business. Almost one year has passed since then, and by now I have almost paid off my debts. (D, female, aged 29, running private business, unregistered migrant, Songino-Khairkhan district) It is very difficult to find a job unless you have acquaintances. My husband for example still can t find a job (S, female, aged 26, hairdresser, Songino-Khairkhan district)

77 We moved here from Bulgan aimag two years ago. I got a job as soon as I came here. It was easy for me because my acquaintance helped me. (Ts, female, aged 39, Songino-Khairkhan district) Also, khureelel helps households to send their children to school or get some health services. After we moved it was very difficult to enroll our children at school. Schools would refuse because they have too many children. Finally, I had to ask my acquaintance and enrolled in a school located far away from our home (Ts, female, aged 39, Songino-Khairkhan district) I go to my acquaintance who is a doctor when I need. I have never been in our family clinic. (Ch, female, aged 33, unemployed, unregistered migrant, Songino-Khairkhan district Table 6.8 shows those who form the khureelel. Relatives are the ones most likely to help (89.4 percent), close friends (29.1 percent) and colleagues (28.8 percent) come the next on the list. The pattern tends to be similar among all households and, on average, there are two supportive khureelels per household. Table 6.8. Percent of households supported by "khureelel" in their lives, by category of khureelel according to selected characteristics Category of khureelel Selected characteristics Close/distant relatives Old class mates at school Old class mates at university Close friends Military company/soldier Home town/village Old/current colleagues Others* Average number of khureelel Number of households supported by khureelel in their lives Location Ger area Apartment area Migration status Migrant ** Non-migrant ** Number of people in the household and more Living standards Very poor Poor Non-poor (low) Non-poor (medium) Non-poor (high) Total

78 Note: * neighbors, other acquaintances ** Migration status had been defined for population years old. Compared to households living in apartment areas, households in ger areas are more likely to get support from close friends, from military service, and from people of their home town. But households living in apartment areas appear to get relatively more support from relatives, classmates or from colleagues. Smaller households (1-2 members) get more support from relatives compared to larger households. However, larger households tend to receive more support from secondary school classmates, close friends, colleagues or from people from their hometown. As the number of household members increases, the average number of khureelels increases too. The average number of supportive khureelels is lower for migrant households (1.6) than for non-migrant households (1.8), even though they get more support from people from their home town or from close friends compared to non-migrant households. When we look at living standards, the khureelels of very poor or poor households are smaller. Non-poor (low income) households, i.e., those at high risk of becoming poor, get more support from their khureelel compared with other groups. As the living standard improves, the average number of khureelels that provide support increases too. For instance, for a very poor household this number is 1.2, whereas for wealthy households it is 2.0. Put differently, rich households are likely to have more friends and relatives. Thus, the survey shows that households with few members, households living in apartment areas, and nonmigrant households or households with higher living standards are likely to get more support in their daily lives from khureelels that includes their relatives, close friends or colleagues. Conclusion Even though more than half of households was informed of government projects and programmes for the support of livelihood capacity of the population, the beneficiaries from them are estimated to be less than 1/10 percentage. The survey respondents are aware of two programmes and projects on average. The percentage of beneficiary households is higher with the green revolution programme and the savings and loan projects than other programmes. There is almost no household who benefited from programme for livelihood capacity support and employment promotion programmes. Thus, there is inadequate information about the availability of programmes for improvement of livelihood of the population in the city. Almost half the households in the capital city do not take up all opportunities for social participation. Especially, the involvement of high income households is the lowest in community activities hence showing that they are not interested in such activities. On the contrary, poor households more actively participate in community activities organized by district and khoroo administration. In other words, it might be quite possible that poor residents more actively participate in meetings and commuity activities organized by district and khoroo administration in hope of getting involved in projects and programmes targetted at vulnerable groups of the population. However, the percentage of households, any member who was involved in vocational skills and work training is similar in all groups of income, but the percentage of poor households benefiting from the trainings remains almost three times

79 less than that of better off households. This shows that low income poor households are fragile not only in economic resources and opportunities but also in capacity. This, itself, leads to conclusion that people s current involvement and activity are not satisfactory enough to improve their livelihood. Hence, it can be said that low income households have less opportunities to turn trainings into benefits from the training. Even though there is no difference between households in terms of training participation rates, poor households are almost three times less likely than better off households to state that they benefited from the training. A certain percentage of total households (less than half) have an acquaintance with people who may have important influence on their lives such as doctors, teachers, administrators, private company directors and inspectors of market places. But as with migrants, people in ger areas, poor and very poor households such a kinship/khuree is very limited. Almost 2/3 of the total households receive assistance and support from their close friends and acquaintances. But the most substantial support in their livelihood is given by close relatives, friends and colleagues. In general, advantaged households such as those with fewer members, higher income, employment and non-migrants in apartment areas receive relatively more support and assistance from others to support their day to day livelihood. The average number of acquaintances who give support to households is on the rise as the income group gets higher. To conclude, in reality, the kinship and acquaintance exists to support the livelihood of households.

80 CHAPTER 7: POVERTY Chapter 7 on poverty, characteristics of the poor are established, and the target group for intervention is identified, and as well as priorities for action are outlined. This chapter addresses the question of how large and widespread poverty is, and whether it affects particular groups. If so, these would be the groups that should be targeted for intervention. Next we explored, which kind of interventions would be most beneficial to them. In addition, this chapter explores alternative ways of determining poverty by using concepts of capabilities and social deprivation. In this chapter we report and describe standard poverty measures. We also compute two measures of inequality, the decile dispersion ratio and the Gini coefficient. This is followed by a description of the socio-demographic characteristics of the poor. We focus on providing identifiers of the poor, i.e. characteristics that are attributed mainly to poor people and describe the target group for interventions. The next section offers a detailed examination of living standards of poor and non-poor, and a summary of the key differences between them in terms of housing, education, health, etc. This provides the basis for determining priorities for interventions. Which issues demand immediate improvement and are the most urgent ones. Which interventions are needed that will target the poor in particular, rather than the total population. Finally, alternative poverty measures based on concepts of capabilities and social deprivation are presented. We analyze how they link with more traditional poverty measures based on expenditure. Key questions - What is the level of poverty in Ulaanbaatar? - Who are the poor? - What are the priorities for intervention? Key answers and conclusions - 33 percent of the UB population lives below the poverty line of 25,300 Tug per capita. 10 percent are very poor (i.e. expenditure below 60 percent of the poverty line). - Poverty is higher in Ger areas and among migrants (45 percent and 37 percent). - However, there is less inequality within those living in ger areas. The same can/cannot be said for migrants. - The poor are typically younger, less well educated and more frequently not married. They live overwhelmingly in ger areas, especially Bayanzurkh. Their household size tends to be larger and they are more frequently headed by females. - Target groups for intervention should therefore be households with many household members in ger areas, and possibly also households where the head of households has not completed secondary school. - Priorities for action are improving the housing and sanitation conditions. Access to health services and education looks comparatively good. The issue there is about improving quality. - Registration is not an issue that is related to poverty. Poor are registered by the same proportion as poor. Moreover, almost 90 percent are registered.

81 - 60 percent of the Kazakh population (4.3 percent) are poor. This is an issue worth of further investigation. 7.1 Standard Poverty Measures In order to get a sense of the width and depth of poverty it is useful to turn to standard poverty indicators. This section starts with a description of these indicators. Readers familiar with the indicators are advised to turn directly to the tables with the results. Description of Poverty Measures We start with the poverty rate (P0) which shows the percentage of poor people. In addition we use measures of the poverty gap (P1), also referred to as the depth of poverty, and the measure of severity of poverty (P2). In this section these measures are briefly explained and results are provided. Poverty rate (P0) measures the number of people in poverty. Poverty is defined as people living below the poverty line. The poverty line in turn is usually constructed by determining the costs of a consumption basket that covers basic dietary needs (2100 calories per adult per day) plus non-food expenditures. In the case of Mongolia the National Statistics Office (NSO) has determined a poverty line of 25,300 tug per capita. In our survey, data on expenditure on food items and non-food-items has been collected at the household level. Based on these figures, it is then possible to determine how many people have an expenditure per capita that is below the poverty line. This number is the P0 value 6. While P0 is the most intuitive poverty figure, it does not say anything about the depth of poverty. Are the poor people far away from the minimum they need for a living or are they just barely below the poverty line? Moreover, if a large number of people are slightly above the poverty line and another segment is just below the poverty line, a small shift in the poverty line will substantially change the P0-value. These issues are addressed by the depth of poverty measure P1, usually referred to as the poverty gap. The depth of poverty measure P1 takes into account how far people are below the poverty line. The measure effectively gives more weight to those with lower income levels 7. The 6 Poverty rate (P0) is the proportion of people under the poverty line. PUPL P 0 = 100 P total pupl Number of people with expenditure less than the poverty line Ptotal Total population 7 The measure of the poverty depth (P1) indicates the average distance of the poor below the poverty line. PL Poverty line npl 1 P = * n i= 1 ( PL Ci) 1 PL 100

82 resulting number is less intuitive than the P0 headcount but it can be seen as a per capita measure of the gap in total individual welfare with respect to the poverty line. A higher P1 implies a larger gap. However, even the poverty gap measure P1 does not satisfy all considerations. There may be a situation where the poorest person looses income to the least poorer person. In this scenario, the depth of poverty measure (P1) would not change (it would only increase if the transfer went to a person above the poverty line). For that reason an additional measure is used, the severity of poverty measure (P2) 8. This measure incorporates inequality among the poor. Again the lower the number the better. Adult equivalence The poverty line provided by the NSO is based on expenditure. For that reason, this analysis also estimates poverty based on expenditures. For computing the poverty figures the concept of adult equivalencies has not been used. This is in line with Mark D. Brenner detailed argument in A Stragtegy for Poverty Reduction in Mongolia 9. The consequence for interpreting the results is mainly that one should bear the following in mind: Families in larger households, i.e. with more family members tend to be poorer. They are in fact poorer, but giving that economies of scale or the presence of children are not taken into account, there may be a slight overestimate of the poverty for large families. Key Poverty Measures: the Results The table below reports the main poverty figures for key subgroups. The overall fraction of people below the poverty line in the sampled region (Ulaanbaatar) is 33 percent. Table 7.1 Key Poverty measures Selected characteristics Ci Poverty Meassures based on expenditure per capita Fraction Very poor Depth of Severity of below (<60% of Poverty Poverty Poverty poverty (P1) (P2) Line (P0) line) Number of individuals Total 33% 10% Location Consumption of the poor person i n number of people in the sample npl number of people below the poverty line 8 The severity of poverty (P2) indicates the inequality of consumption distribution of the poor (notation as in previous footnote). P npl 1 = * n i= 1 ( PL Ci) 2 2 PL 100

83 Ger khoroolol 45% 14% Appartment 16% 5% khoroolol Migration status* Migrant 37% 12% Non-migrant 31% 9% Note: * - Migration status had been defined for population years old Poverty seems to be more concentrated and much more severe in the ger areas. The groups of migrants also shows a higher proportion of poor than their counterparts. Measuring inequality Another aspect worth looking at is the degree of inequality in a society. We will use two measures, decile dispersion ratio and Gini coefficient. Again, readers familiar with these inequality measures are advised to turn directly to the table with the results. Description of Inequality Measures The decile dispersion ratio is a simple and intuitive measure. It provides the ratio of the mean expenditure of the top 20 percent to those of the poorest 20 percent (mean expenditure of top 20 percent divided by mean expenditure of poorest 20 percent). The higher the measure the higher the gap between the rich and the poor. For example a number of 4 says that the rich are spending 4 times as much as the poor. The decile dispersion ratio has the drawback of focusing only on the poorest and the richest. What about the people in the middle? Are the rich rather a separate class far above the rest or are differences between rich, people in the middle people and the poor rather fluent? Gini coefficient and Lorenz curves measure how far a society is away from total equality. In a world of total equality each group of 10 percent of people would spend 10 percent of the expenditure. In reality the poorest 10 percent have less than 10 percent share of the total expenditure. The share of expenditure can be computed for the poorest 10 percent, for the second poorest 10 percent and so on. This is then plotted on a graph. Figure 7.1 Lorenz curve for Ulaanbaatar

84 Lorenz Curve Total 100% share of consumption expenditure (cumulative by individual) 90% 80% 70% 60% 50% 40% 30% 20% 10% 45 degree line = total equality in terms of income distribution Distance to equality, meassured by Gini coefficient (the smaller, the better) 0% 0% 20% 40% 60% 80% 100% population share (cumulative by individual) The 45 line shows the world of total equality. Each 10 percent of the population earns 10 percent of total income. The curve is always below the 45 line. It shows that, e.g. the poorest 10 percent earn less than 10 percent of the total income, and that even the poorest 20 percent (taken together) earn less than 10 percent of total income. The closer the curve is to the line, the lower inequality. The average distance between the curve and the 45 line is the Ginicoefficient. It lies between 0 and 1. The smaller it is, the lower the inequality. Inequality Measures: the Results Turning towards the results of the inequality measures there is an interesting finding concerning ger and apartment areas. As we have seen people in ger areas are poorer and there is also less inequality among them. Table 7.2 Inequality measures Selected characteristics decile dispersion ratio Inequality based on expenditure per capita Expenditure of top 20% Expenditure of bottom 20% Ginicoefficient Number of individuals Total tug tug Location Ger khoroolol tug tug Appartment khoroolol tug tug Note: * - Migration status had been defined for population years old This can already seen by the comparably low expenditure of the top 20 percent. It is also reflected in the lower Gini-coefficient. A comparison of the expenditure distribution shows

85 that most ger residents are close to the poverty line. Even the non-poor are not far above the poverty line. In apartment areas residents are richer and more people are further away from the poverty line. It should be noted, though, that inequality in Ulaanbaatar is low by international standards. Figure 7.2 Distribution of expenditure per capita in ger and apartment areas Distribution of expenditure per capita: less inequality in Ger Khoroos 800 People below poverty line 600 Poverty line Poverty line Number of responents People above the poverty line Std. Dev = Mean = N = Std. Dev = Mean = N = GER KHOROOS: consumption expenditure monthly per capita APT KHOROOS: consumption expenditure monthly per capita Comparison of poverty figures with other data Ulaanbaatar is still a place of substantially less inequality and severity of poverty than many other parts of the world. Another question is whether poverty and inequality have increased over time. This is the first time this particular survey was carried out, so a comparison would have to be made with other poverty studies. According to the Living Standard Measurement Study of 1998 the poverty rate in Ulaanbaatar was 34 percent, whereas this survey reports a poverty rate of 33 percent. This suggests that poverty had basically stayed the same or, at best, has slightly improved. However, differences in data and methodology do not make the figures strictly comparable Characteristics of the Poor Who are the poor? What are their characteristics? An overview is given in this section. 10 For example, differences in methodology result from a number of sources: different questions, different composition of consumption basket, differences in weighing children and adults, different assumptions about depreciation rates of durable goods, etc. Hence, it cannot be said whether the decrease in the poverty figures is due to a real decline in poverty or just the result of using two different methodologies.

86 Structure of this section The section begins with a table showing what proportion of various sub-segments are poor. It gives a first indication of which characteristics are linked to poverty. The next section determines how closely certain characteristics are linked to poverty. It does so by means of logistic regression. Finally the question is addressed, which segments should be target groups for intervention. This does not only take into account, which characteristics are particularly prone to be linked with poverty but also the size of the segment. When interventions are designed, the target group should be sufficiently large (or, in the case of e.g. the disabled, be in particular need for assistance) Description of the Poor The following table shows what proportion of various sub-segments are poor. It answers questions such as: how many females are poor?, how many people in Bayanzurkh district are poor? and so on. It does an individual basis, i.e. it looks at poor people, rather than poor households. It gives a first indication of which characteristics are linked to poverty. Table 7.3 Percentage distribution of respondents by the consumption expenditure poverty and selected characteristics Consumption expenditure poverty Characteristics Poor Non poor Total Numbers Migration status** Migrant Non-migrant Location Ger area Apartment area Districts Bayanzurkh Nalaikh Sukhbaatar Songinokhairkhan Khan-Uul Chingeltei Age Below Sex Male Female

87 Ethnicity Khalkh Kazakh Durved Buriad Bayad Other Marital status* Single Married Living together Separated Divorced Widowed Number of HH members Education* Not educated/ primary Incomplete secondary Complete secondary Vocational/Special technical/high Registration status Registered Not registered Working sector*** Formal Non-formal Abroad Type of living quarter Ger Conventional dwelling Number Total * calculated for the population aged 15 and above (n=5053) ** calculated for the population aged (n=4712) *** calculated for the working population aged 7 and above (n=2166) The table 7.3 shows that particular subgroups have an above average proportion of poor people, such as people in ger areas, especially Bayanzurkh and Nalaikh, people living in gers

88 (not all dwellings in ger areas are necessarily gers and gers can also be located in apartment areas) and people with little education. Moreover, poor people tend to have more children which in turn means that more children are poor as well as households with more family members. Moreover, some results are interesting in the sense, that a characteristic is not linked to poverty, although it frequently is in other places, notably being female and working in the informal sector. The first finding may be seen as a confirmation that in Mongolia in general the gender relations are relatively good. Concerning the informal sector, usually people working in a non-formal sector are considered as more vulnerable. However, it has also been noted in the chapter about employment that income in the non-formal sector is high among those who run their own business, thereby reducing the proportion of poor in the non-formal sector. Finally the high proportion of poor among the Kazak population should be noted. While Mongolia is a relatively homogeneous country with small variations in the ethnic composition as well as cultural variations between ethnicity, it is still the case that around 60 percent of the population of kazak nationality are poor. While the number of kazaks in the sample is rather small (n= 173) the representativeness is still satisfactory when compared to the share of kazaks in the national population of 4.3 percent (The 2000 Population and Housing Census, NSO, 2001). This suggests, that ethnicity is an issue worthy of further analysis Identifying the Target Group for Interventions In order to identify the characteristics of the target group for intervention the following aspects should be considered: 1. The difference between poor and non-poor with respect to the characteristic should be large. 2. The characteristic should apply to a sizable number of households. Some particularly vulnerable groups, such as disabled or orphans, who (fortunately) do not constitute large sub-groups, are not considered in this analysis, even though there is no question that they should benefit from special programs. The charts below present the results in such a way that these aspects are satisfied and that conclusions can easily be drawn. The first chart shows for which characteristic poor and nonpoor household differ most. It also shows the size of the sub-segments within the groups of poor and non-poor. Figure 7.3 Socio-Economic Characteristics: Largest differences between poor and non-poor

89 Socio-Economic characteristics: poor and non-poor differ most with respect to HH size and location Reading example: of all poor households (HHs), 47% have 6 or more people. Of the non-poor HHs, only 17% have 6+ members poor 40% 47% 79% 6 or more people per HH Ger areas Ger khorool and 6+ people per HH non-poor 10% 17% 46% 0% 20% 40% 60% 80% 100% Base: poor: n=337 households non-poor: n=1163 households Above it has been seen that 45% of people in ger areas are poor (37% of HHs). If the focus is moved from looking at the ger area to looking at all the poor (both in ger and apt. areas) it is found that almost all poor HHs (79%, represent 81% of the people) are concentrated in ger areas The chart already provides a number of insights Almost all poor households (79 percent, representing 81 percent of the poor people) are located in ger areas The households with 6+ members and located in ger areas account for 40 percent of the poor households (representing 52 percent of the poor people). The probability to find a poor person in a ger area household with 6+ people is 6 out of 10, which is very high given that only one third of people are poor.

90 Figure 7.4 Socio-Economic Characteristics: Smaller differences between poor and non-poor Socio-Economic characteristics: part 2 (where differences are smaller) Reading example: 47% of poor households are have an education level of uncompleted Secondary or less. However, 22% of non-poor households are also less well educated. poor 31% 47% education level: uncompleted Secondary at most (head of household) 59% Ger living quarter* 22% non-poor 13% 50% age up to 25 (individuals) 0% 20% 40% 60% 80% 100% *Ger living quarter: this refers to people who actually live in a ger. Ger khoroolol by contrast just states that the respondent lives in an area (khoroolol) that is considered to be a ger area. He may live in an apartment, but still be located in a ger khoroolol. Base: households, unless stated otherwise Figure 7.5 Socio-Economic Characteristics: Smaller differences between poor and non-poor 2 Socio-Economic characteristics: part 3 (where differences are still smaller) Reading example: 30% of poor households are female headed. However, 22% of non-poor households are also female headed. poor 5% 17% 30% 45% female headed households not married (head of household) non-poor 12% 22% 39% migrants (individuals aged 15-64) 2% Kazakh 0% 20% 40% 60% 80% 100% Base: households, unless stated otherwise Figure 7.2 and 7.3 show that in addition to focusing on large households in ger areas, a second focus should be on education level. Those with a lower education level can also be seen as a target group for intervention.

91 With respect to the other characteristics that are linked with poverty, it is debatable whether they should be part of a description of a target group for intervention. The number of female headed households is not that large within the group of poor. The same applies to migrants. Concerning marriage the difference between poor and non-poor households is not large. An intervention focusing on not-married people would also benefit a large number of poor. Concerning Kazaks the number is so small that an intervention just targeted at Kazaks would reach too few people. Nevertheless, it should be noted that Kazaks appear to be much poorer. Hence, the target groups for intervention should be - households with many members in ger areas - and possibly also: households where the head of the household does not have finished secondary school. 7.3 Consumption and living standard The following sections describe consumption levels and living standards of the poor versus the non-poor. It has proved useful to distinguish both groups in subgroups in order to provide a more nuanced understanding of both the poor and the non-poor. A broader look at the sample is taken. For the analysis the following groups are considered: 1. Very poor: per capita household expenditure is less than 60 percent of the official poverty line; this is less than 15,180 tugrugs. 2. Poor: per capita household expenditure is between 15,181 and the official poverty line of 25,300 tugrugs. 3. Non-poor (low): per cap household expenditure ranges between the poverty line and up to 60 percent higher; this is 25, tugrugs. 4. Non-poor (mid): per capita household expenditures are between the highest level of non-poor low expenditures and up to 60 percent higher: this is between 40, tugrugs. 5. Non-poor (high): per capital household expenditures are above tugrugs. Table 7.4 presents the living standard of the city population by consumption expenditure. Of the total population, 33 percent have the consumption estimated lower than the official poverty line. Following our classification, 10 percent of the total population are very poor with per capita expenditure estimated at less than tugrugs. Table 7.4 Percentage distribution of respondents according to living standard Living standars Percentage distribution Number of population Very poor Poor Non poor (low) Non poor (medium) Non poor (high) Total

92 Table 7.5 shows several sociodemographic caracteristics of the groups. Looking at various age groups we find that poverty is lowest (26.4 percent) among the population of 65 and over years old. A possible explanation is that older individuals receive pensions, which does not apply to the younger age groups. Table 7.5 Percentage distribution of respondents by living standard and according to selected characteristics Living standards Selected characteristics Very poor Poor Non poor (low) Non poor (medium) Non poor (high) Total Number of cases Age Less than and over Sex Male Female Marital status* Single Married Living together Separated Divorced Widowed Total The proportion of men and women in poor and non-poor groups is the same. Divorced and separated individuals have the lowest living standard while married people seem to be better off. We can see in Table 7.6 that characteristics such as number of children, household size, and number of families in the household, etc. have a significant impact on living standards. There is a reverse relation between the number of children and household size, and living standard. When the number of children and household size are big, the living standard of households

93 tends to drop. For instance, only 9.3 percent of households without a child live poor while 44.3 percent of households with more than three children live in poverty. Table 7.6 Percentage distribution of the households by living standard, according to household characteristics Living standards HH characteristics Very poor Poor Non poor (low) Non poor (medium ) Non poor (high) Tota l Number of children None and over Number of people in the HH and over Number of family in the HH and over Registration status in the Khoroo Registered Unregistered Number of HHs Total Compared to nuclear or single household, mixed or extended households have relatively lower living standard. This is associated to the fact that 2/3 of mixed households have families of six and more members. There are more unregistered households in the very poor category than registered ones. However we also see that the proportion of unregistered households is larger at the top of the distribution.

94 Figure 7.6 Percentage distribution of respondents according to living standatds and by location Percent ,3 31,7 28,7 31, ,6 Ger Apartment 20 16, ,7 4,6 10,8 6,8 0 Very poor Poor Non poor low Non poor mid Non poor high As seen in Figure 7.6 there is drastic disparity in living standard between ger and apartment areas. The proportion of poor households in ger areas is around 45 percent, while in apartment areas we find just over 15 percent. Figure 7.7 Percentage distribution of respondents according to living standards and bymigration status Percent Migrant Non migrant , , ,9 21,7 18, ,4 8,6 10,3 5 0 Very poor Poor Non poor low Non poor mid Non poor high The difference between migrants and non-migrants is less marked, but it still is there. (Figure 7.7). Particularly, 37.3 percent of migrant population and 30.6 percent of the non-migrant population live in poverty or absolute poverty. This finding has been supported by focus group discussions. About percent of the poor in our khoroo are made up from in-migrants from rural areas. They lost their livestock to zud and accordingly moved to the city to set their ger in open field without permission. They tend to seek the assistance form the government. (A. 40 year old women, khoroo governor, Songinokhairkhan district)

95 Of the migrants to the city 72.7 percent (Table 7.7) moved into Bayanzurkh, Songinokhairkhan and Chingeltei districts which are loated in the outskirts of the capital city. The availability of lands or possibility of extending the territories made these districts as the main zone for receipt of migrants. Accordingly, these districts are having high poverty level. The districts Chingeltei, Bayanzurkh and Songinokhairkhan are home to 74.6 percent of the total ger areas in the capital city. As for Nalaikh district, the in-migrants account for the lowest percentage or 2.8 percent but with highest poverty level and lowest living standard. The districts are under-developed compared with central districts of the city and about 40 percent of the total households in Nalaikh district are poor. We will explore later in this report the individual characterstics of migrants and why they tend to be poorer. Living standards of households in apartment areas are relatively high, and accordingly, districts with large apartment areas show lower levels of poverty. For example, 18.9 percent of the total households in Sukhbaatar district are poor. This is 18.1 points lower than the figures for Nalaikh district. The percentage of non-poor households is higher than the city average of the city. In contrast, Nalaikh district shows the lowest percentage of non-poor households. Table 7.7 Percentage distribution of the households, by living standard, according to distircts Living standards Tota Very Poor Non poor Non poor l poor (low) (medium) Districts Non poor (high) Number of total HHs Bayanzurkh Nalaikh Sukhbaatar Songinokhairkhan Khan-Uul Chingeltei Total Table 7.8 shows education level and employment status of the respondents. As expected, higher levels of education are associated with lower poverty incidence.

96 Table 7.8 Percentage distribution of the population aged 15 and over by living standard, according to selected characteristics Selected characteristics Very poor Poor Living standards Non Non poor poor (low) (mediu m) Non poor (high) Tota l Number of the popluatio n aged 15 and over Education level Not educated Primary (grade 4th class) Uncompleted secondary (grade 8) Completed secondary (grade 10) Vocational Special vocational/uncompleted 100. high Higher/post graduated Employment status Employed Not working Working sector* Formal Informal Total Note: *Employed population only. A small number of people (n=41) is working abroad. The number is too small to provide reliable information, and is, therefore, excluded. Among those that work 5.8 percent is very poor the percentage goes up to 11.3 in the case of those who do not work (unemployed and non participants). Approximately, eight in every ten persons who are employed belong to non-poor category whereas this is six in every ten

97 persons who do not work. Wages and equivalent income sources make up an important source of their livelihood. From the focus group discussion one can see that working or getting a job is one step which has been made to escape poverty. I think that we can live and work well if we try and step on the right path. First it was very difficult for my husband and me when we moved into the city to improve our life. Sometimes we felt very discouraged. We sewed to make clothes and sell them. Now we bring the clothes from China for sale. To start our business we got a loan, now we have paid out loan. We live much better than we did back in rural areas. (D. female, 29 year old, private tradeswoman, Songinokhairkhan district) Limited employment opportunities for poor people contribute to an increase in poverty. There is an age limit, plus an assessment by appearance. So when applying for a job I put on my best clothes. But they don t employ me even in postion of a cleaner. These days money has become a pivot in relations with each other. I got to understand that no matter how much I try I will never be employed. (N. female, 49 year old, unemployed, Bayanzurkh district) Of the informal employees 27.4 percent are poor whereas this indicator is 23.2 percent for the formal employees. This situation exposes the disparity in wages and salaries in sectors (see Table 7.8). In general, as the living standard of the population falls, the housing conditions tend to deteriorate. Moreover, people living in poor housing conditons are likely to be more vulnerable to sickness. And such situation leads to poor life for these people. The poverty is the lowest in apartment areas where housing conditions tend to be better. (Figure 7.8). Moreover, the focus groups showed that most of the poor in apartment areas are those who live in suburban areas of the city, rather than close to the centre. 100 Percent Figure 7.8 Percentage distribution of the population by living standards, according to type of dwelling Poor Non poor ,8 61,2 60,4 45,9 28, ,9 39,6 54,1 71,1 0 13,2 Apartment/ N=2578/ Public dormitory /N=211/ House /N=2598/ Ger /N=1356/ Non living quarter /N=104/

98 Table 7.9 Percentage distribution of the HHs by living standard, according to selected Characteristics Living standards Selected characteristics Very poor Poor Non poor (low) Non poor (medium) Non poor (high) Tota l Numbe r of HHs Ownership of dwelling* Owned Not owned Do you have another ger/house/cellar/garage * Have Do not have Do you have TV and refrigerator Have both Have one of the two Do not have Total 0 Note: * Dwellings not intended for living are excluded Interestingly, ownership of dwelling does not vary much by living standard (Table 7.9). The same goes for ownership of garages, cellars and barns in addition to housing. But it is worth to mention that these items vary by building materials, costs, design, designation and size depending on living standards of households. It is little surprising that ownership of durable consumption goods, such as TV and a refrigerator increases with higher living standards. From Table 7.9 one can see the existence of these items becomes an indicator showing the living standard of households. Of the very poor households, 29.3 percent have none of these items. Among the households which own both of these items the very poor households make up only 3.6 percent. Better off households own all necessary items starting from refrigerators ending with transport means (vehicles), while the poor people live with a shortage of basic household items. Whereas almost one in every four household has a vehicle, ¾ belong to 2 better off groups of the non-poor.

99 7.4 Determining Interventions for the Poor In the previous sections the living conditions of poor people relative to non-poor people have been described in detail. This section serves the purpose of ensuring that the key messages are not lost in the detail. It provides a graphical overview of key aspects of housing, education, health, etc. The charts show to what extent the poor are deprived of certain aspects of living conditions, and whether this deprivation is an aspect of poverty or whether it is an issue also for the non-poor. Those issues that affect both groups in almost equal measure can be considered as laudable interventions in their own right. However, for the purpose of designing interventions that target the poor, the focus should be on those deprivations that affect the poor to a much larger extent then the non-poor. The following charts show, which issues affect both groups, and which issues only affect the non-poor. Among the poor, 85 percent lack access to sanitation (measured by lack of access to flush toilet). The chart also shows that only half of the non-poor have flushing toilets. The difference between the two groups is rather large, and sanitation can be considered a problem that affects the poor more severely. A different example is access to clean water. This can be considered to be not an issue that affects the poor in decisively greater numbers. While 17 percent of the poor do not have access to clean water, not even to a protected well, 8 percent of the non-poor also lack access to clean water. Hence it could be argued that an improvement of water sources would be a laudable intervention on its own merits. As an intervention that is targeted at the problems of the poor, it is likely to be of less consequence than other measures. Figure 7.9 Difference between poor and non-poor: housing Housing: differences between poor and non-poor are mainly to do with the toilet and the dwelling poor non-poor no sanitation (i.e. no indoor flush toilet, base n=1500 households) 51% 85% poor dwelling (ger or living quarter) 14% 36% poor water sources (not even a protected well) poor disposal unit (no centralized dust-hole) 17% 8% 10% 3% Reading example: 17% of the poor have access to poor water sources only. Of the non-poor these are still 8%. poor electricity 7% 1% 0% 20% 40% 60% 80% 100% Base: poor: n=337 households non-poor: n=1163 households Figure 7.10 Difference between poor and non-poor: education

100 Education: differences between poor and non-poor: poor non-poor drop out of school (base n=287, aged 6-18, not attending school) 17% 39% Highest education level: uncompleted Secondary (base n=5053, age 15+) 26% 50% distance to school > 2km 18% 32% not attending school (of base n=1352, aged 8-16) 2% 11% 0% 20% 40% 60% 80% 100% It is hardly surprising that the poor are less well educated and are less frequently attending school In this way, we can also look at other aspects, such as education. While access to school is good, the concern should be more with the level of education completed. Compared to other developing countries, Mongolia as an enviably high school attendance rate. In other countries it would already be a success if all children attended and finished Primary School. Does that imply that not much further attention needs to be given to education? Hardly. These charts don t say anything about crowded classrooms or about quality of education. Figure 7.11 Difference between poor and non-poor: health Health: differences between poor and non-poor poor non-poor no health insurance 16% 24% no access to health professional 5% 11% distance to hospital >3km 11% 15% chronic illnesses (diagnosed by a doctor) 14% 16% 0% 20% 40% 60% 80% 100% Health services are comparatively well supplied. The largest difference is in terms of health insurance Base: poor: n=2258 individuals non-poor: n=4589 individuals As with education, the access to health services looks promising. Even among the poor, 75 percent have health insurance. Again, this is a value that can be considered to be quite an achievement. By the same token, there is still one quarter of the population (not necessarily all

101 poor) uninsured, a state of affairs that demands correction. However, given the relatively small difference between poor and non-poor improving health insurance coverage is an urgent cause in itself, but not one that would focus on poor. It should also be noted that this graph refers to access to services. Quality of services is another question, not covered in the quantitative survey. Figure 7.12 Difference between poor and non-poor: employment and social inclusion Employment and social inclusion: differences between poor and non-poor poor non-poor not working (base n=5053, age 15+) 53% 67% employment in informal sector (n=2166, working population) 34% 29% no khuree support in daily live 37% 52% not participated in community work no access to information about projects 40% 43% 39% 27% 0% 20% 40% 60% 80% 100% The poor do (or can do) less to earn income. They have less khuree support in their daily lives Base: poor: n=2258 individuals non-poor: n=4589 individuals Concerning employment there is a marked difference between poor and non-poor. This is to be expected, as employment is the best way out of poverty. On that basis, it can be argued that more needs to be done to increase employment opportunities. What the poor do not tend to

102 have to a comparable degree is the support by means of khuree, which in turn may very well be a contributing reason for their poverty. Figure 7.13 Difference between poor and non-poor: registration and income sources Registration and income sources: differences between poor and non-poor poor (n=2258) non-poor (n=4589) not registered in khoroo 10% 9% not registered full residence 8% 7% HH has no other house/ger/cellar/garage/ambaar 66% 64% HH not owns living quarter (base individuals) 11% 12% 0% 20% 40% 60% 80% 100% There are hardly any differences in this respect Base: total of individuals It is surprising that there are hardly any differences between poor and non-poor. One would have expected that the non-poor have more frequently another house/ger/cellar and that they more frequently own their living quarter. But the poor do have that in equal measure. One would also have expected that lack of registration is far more frequent among the poor, leading to a greater difficulty in getting access to social services, as the focus groups have shown. The finding above suggests that registration is not an issue that is related to poverty. Non-poor are not registered by the same proportion as poor. Moreover, almost 90 percent are registered. However, in the context of migration it will be seen that registration does make access to social services more difficult, and that it is important to make registration easier (see Section 8.6). 7.5 Priorities for Intervention How do these findings translate into priorities for intervention that are targeted at the problems of the poor? Each of the dimensions discussed (housing, education, health, employment, registration) offers itself to urgent and extremely necessary action. However, not everything can be done at the same time, nor should it be tried. The challenge is therefore to determine among all these areas of action the ones that have the highest priority. How can this be done? Two dimensions are important in determining priorities for interventions that are focused on the poor: - Degree of deprivation: Which are the services/improvements that are lacking most among the poor? - Gap between poor and non-poor: Which are the issues that target especially the poor? Where is the gap between the poor and the non-poor especially large? Figure 7.14 Determining priorities 1

103 Housing: an overview of how much is already there for the poor and how big the gap towards the nonpoor is gap between poor and non-poor Reading example: 64% of the poor have a good dwelling. Yet there is still a large gap compared to the non-poor (gap=22%) 0% 33% 67% 100% % of poor people (that have access to the shown indicators) sanitation (variable used: indoor flush toilet) good dwelling (house, apt., dorm.) good water sources (incl. a protected well) centralized dusthole Figure 7.15 Determining priorities 2 On average the poor have access to 66% of some 10 key indicators (in housing, health, education). The gap to the non-poor is 15% on average. gap between poor and non-poor mean mean 0% 33% 67% 100% % of poor people (that have the shown indicators) saniation (proxy: indoor flush toilet) good dwelling (house, apt., dorm.) good water sources (incl. a protected well) centralized dusthole By looking at both dimensions simultaneously it is possible to determine priorities.

104 Figure 7.16 Determining priorities 3 Now there are four areas ( quadrants ), each suggesting a different course of action. In the top left quadrants are those items, where the poor have little access (and where the gap to the non-poor is large). These can be reagarded as areas for improvement difference poor versus non-poor improve mean mean 0% 33% 67% 100% % of poor people (that have access to the shown indicators) saniation (indoor flush toilet) good dwelling (house, apt., dorm.) good water sources (incl. a protected well) centralized dusthole Figure 7.17 Determining priorities: housing

105 Items in the right quadrants provide already relatively good access for the poor. For those in the bottom right quadrant the gap to the non-poor is not even large. The items in the right quadrants should at least be kept in the relatively good state. difference poor versus non-poor improve mean mean lower priority (comparatively good already) 0% 33% 67% 100% % of poor people (that have the shown indicators) keep saniation (variable used: indoor flush toilet) good dwelling (house, apt., dorm.) good water sources (incl. a protected well) centralized dusthole Figure 7.18 Determining priorities: access to education Education: An area for improvement could be the length of education, wheras attention is already on a rhigh level difference poor versus non-poor improve mean mean keep lower priority (comparatively good already) 0% 50% 100% % of poor people (who have access to the shown indicators) Highest education level: completed Secondary and more distance to school < 2km attending school (of base n=1352, aged 8-16)

106 Figure 7.19 Determining priorities: access to health Health indicators are already on a high level (all of them refer to access, not to quality, though) difference poor versus non-poor improve mean mean keep 0% 50% 100% % of poor people (that have access to the shown indicators) health insurance access to health professional distance to hospital <3km Figure 7.20 Determining priorities: housing, education and health gap between poor and non-poor Bringing it all together: the focus of intervention should be on improving housing conditions and possibly also a little on the education side improve mean mean keep 0% 33% 67% 100% % of poor people (that have access to the shown indicators) Highest education level: completed Secondary and more distance to school < 2km saniation (variable used: indoor flush toilet) good dwelling (house, apt., dorm.) good water sources (incl. a protected well) health insurance distance to hospital <3km Conclusion of the section Hence, priorities for action are improving housing and sanitation conditions. Access to health services and education looks comparatively good, although we cannot assess quantitatively whether there are differences in the quality of these services. The qualitative study has shown, however, that there are substantial quality issues (as shown in chapter 5 on access to social services).

107 7.6 Alternative ways of looking at poverty One of the main difficulties in studying the poverty is the definition of poverty. The common ways of defining and measuring the poverty are methods related to the income and expenditure (see above). However, the human well-being cannot be measured by the income only. The Nobel Prize economist Amartya Sen has noted that it is important to measure the poverty not only in terms of income but also in terms of capability including accessibility and availability of a person to get educational and health services etc Capability poverty Education and health indicators as well as housing are competent indicators of the person's well-being. These indicators will determine the poverty in terms of capability. The index to measure the capability poverty has been calculated based on three groups of indicators: housing conditions; access to the education and health services. The following 9 variables were included in that indicator - type of dwelling: house, apartment, dormitory (not ger and not dwellings not intended for living) - electricity - drinking water from protected well or centralized - centralized dust hole as garbage disposal unit - indoor flush toilet - health insurance - distance to school: less than 2km - distance to health professional: less than 3km - access to health professional If at least two thirds of these aspects apply to the household, then the household members are considered non-poor. The construction of this indicator is necessarily arbitrary. It is acknowledged that there are weighting issues. For example, education only enters once, whereas housing enters with 5 variables. The guiding principle has been the concept of accessibility. Do people have access to education? This depends entirely on the closeness to the school. Access neither has to do with rate of completion nor with quality of education (the classroom still may be overcrowded). It could be argued that access to education should really be about access to good education. But that would take the concept too far. Where does good education begin? At a classroom size of 25? Or is the qualification of the teacher decisive? Given that this leads to a wide range of additional arbitrary choices, access was understood to simply mean whether a person could get a minimum level of education, health service etc. if he wanted to. A look at the variables that make up the indicator show that many of them are already provided on a relatively high level. This in turn supports the following statement: Mongolia has relatively good development of education and health systems. It has impressive indicators of number of schools and hospitals; number of persons per doctor and per teacher; school enrolment; literacy etc., (Government of Mongolia, UNDP, 2003). The conditions in which people live are good indication of the well-being and poverty. Table 7.10 shows the capability poverty index according to demographic and socio-economic characteristics.

108 Table 7.10 Percentage distribution of respondents by capability poverty index and selected characteristics Capability Characteristics Poor Non poor Total Numbers Migration status** Migrant Non-migrant Location Ger area Apartment area Districts Bayanzurkh Nalaikh Sukhbaatar Songinokhairkhan Khan-Uul Chingeltei Age Below Sex Male Female Ethnicity Khalkh Kazakh Durved Buriad Bayad Other Number of HH members Education* Not educated/ primary Incomplete secondary Complete secondary Vocational/Special technical/high

109 technical/high Registration status Registered Not registered Working sector*** Formal Non-formal Abroad Employment status* Employed Not working Type of living quarter Ger Conventional dwelling Number Total * calculated for the population aged 15 and above (n=5053) ** calculated for the population aged (n=4712) *** calculated for the working population aged 7 and above (n=2166) The capability poverty is higher for the migrants. More than 50 percent (54.8 percent) of migrants are poor in terms of capability that is twice higher than share of poor non-migrants. It can be concluded that social services are not distributed equally to the migrants and nonmigrants. With respect to districts, the capability poverty is the highest in Bayanzurkh (37.3 percent) and Nalaikh (31.7 percent) districts. It is observed that these districts are poor as in terms of per capita household expenditures and in terms of capability as well. The lowest share of poor is in Khan-Uul (around 14 percent) district. It is possible that percent of migrants in this district is lower. There are no much variations between age and sex groups regarding the capability poverty. Notable differences can be observed in capability poverty of ethnic groups. For instance, 45.7 percent of kazak are poor while less than 10 percent of buriad are in the same group. Furthermore, 31 percent of durvuds are poor in terms of capability. Although capability poverty might depend on cultural and traditional particularities of the certain ethnicity, it shows that there is a need for deeper analysis of ethnic variations and differentials in all the poverty aspects. Survey results indicate that education is an important factor for the poverty and well-being. The highest percentage of poor in terms of capability is among the respondents with complete secondary education (34.7 percent). Out of those who have the education level higher than completed secondary education 14.8 percent are poor.

110 Half of the not registered at the khoroo respondents are live in poor conditions. Previous studies state that registration is important to access the health and education services. It appears that it is true for the sample of the current survey as well. Regarding the type of living quarter we may see large gaps in the percentage of poor. Although ger is convenient dwelling for the nomadic Mongols, in the urban environment it cannot provide sufficient housing condition Social inclusion poverty The accessibility to information, friends as well as social support, and networking are determinants of the well-being. The social inclusion indicator a summary index that has been constructed based on three variables. These variables are: whether the household gets assistance from their khuree 11 in their everyday life; participation in the community work; and lastly, knowledge about the existing projects and programs that are implementing to improve the living standards of the population 12. The existence of khuree support indicates networking of the household with their friends, relatives and others, which is very important in the context of Mongolia and probably everywhere, particularly in the developing countries. Besides getting assistance from others, participation in community work by individuals and households can be regarded as a means of social participation or social inclusion. For that reason it is included in the indicator. The access to information about projects and programs on improvement of the living standards is also considered as indicator of social inclusion. The indicator of poverty in terms of lack of social inclusion is therefore constructed using khuree, community involvement and access to information. If a person has khuree is non-poor (reflecting the importance of khuree). If khuree is missing, the person is considered to be poor, unless both other factors are given (I.e. community involvement and access to information). 11 As explained in chapter 6, khuree means a dry relation as opposed to a blood relation (e.g. parents, children sisters/brothers etc, which, however, is also summarized in khuree). The use of it is that a dry relation has equally binding responsibilities towards a person as a blood relation, if the person is asked for help. The dry relations can be old class mates or parents of an old classmate, friends of the family etc. If a person asks a khuree for help/money/connections he in principle has to do something. But it should be seen as something a person can approach if that person is in difficulties or need assistance, not necessarily something that will just be there giving someone food as soon as someone is hungry, i.e. one has to ask for it. In that sense it can be a social security network one can draw on when times are rough, it can be what gives one a job through connections, it can be the place where one gets credit etc. 12 These programs are for example the following: Household Livelihood Capacity Support Program, Green revolution, Saving/Credit, Employment promotion (see also chapter 6, e.g. table 6.2 for further information)

111 The construction of this indicator is arbitrary. Especially, it can be argued that the three aspects should not be bundled together in the first place. Community involvement, for example, decreases with higher income. So, are well-off people more likely to be socially excluded? Only if they miss out on khuree, which in turn reflects on the significance of khuree in Mongolian society. The index is shown according to the demographic and socio-economic variables in the Table Overall, 24.3 percent of the respondents are poor in terms of social inclusion. More than 30 percent of migrants and 23.6 percent of non-migrants are also in a poor category of people. It is well known the process of adaptation; creation of own khuree of friends and relatives takes time. Therefore, migrants are more likely to be poor in terms of social inclusion compared to non-migrants. On the other hand, the study on internal migration states that networking is a very important factor for the migration (MOLSW, PTRC & UNFPA, 2001). This might explain relatively small gap between migrants and non-migrants. Unlike income poverty, there are no much differences between those living in ger and apartment areas with regard to the social inclusion poverty (26.1 percent vs percent). In other words, although residents of ger areas are poor in terms of expenditure, they are not as poor in terms of access to the information, khuree or networking as well as participation in a community work. Table 7.11 Percentage distribution of respondents by the social inclusion and selected characteristics Social inclusion Characteristics Poor Non poor Total Numbers Migration status** Migrant Non-migrant Location Ger area Apartment area Districts Bayanzurkh Nalaikh Sukhbaatar Songinokhairkhan Khan-Uul Chingeltei Age Below

112 Sex Male Female Ethnicity Khalkh Kazakh Durved Buriad Bayad Other Number of HH members Education* Not educated/ primary Incomplete secondary Complete secondary Vocational/Special technical/high Registration status Registered Not registered Working sector*** Formal Non-formal Abroad Employment status* Employed Not working Type of living quarter Ger Conventional dwelling Number Total * calculated for the population aged 15 and above (n=5053) ** calculated for the population aged (n=4712) *** calculated for the working population aged 7 and above (n=2166) Regarding the districts, except Khan-Uul district other districts have percentage distribution of the poor and non- poor enough close to the general pattern. In the Khan-Uul district the majority or 89.4 percent of the residents are non-poor.

113 Variations between age groups and sexes are close to the general pattern. Only for those older than 60 years the share of non-poor exceeds 80 percent. Older people might have broader khuree, networking as well as higher social participation. Marked variations in the social inclusion poverty have been observed between ethnic groups. Around 24 percent of the respondents of major ethnic group- khalkh are poor in terms of social inclusion whereas for the kazakhs this percent is 41. Some researchers propose the hypothesis of language barriers to explain the differences. However, the ethnic differences in poverty aspects should be studied further. In general, Mongolia has not been experiencing ethnic or nationality conflicts or struggles historically. Nevertheless, ethnic relationships should be carefully studied in the context of the social and economic changes of the last 20 years period. Out of not registered in a khoroo respondents 36.8 percent are poor while the share of the poor registered respondents is 23.1 percent. People who are better linked may have less difficulty in getting the necessary steps done for registration How do alternative poverty measures fit with the standard poverty measure? How do the two alternative poverty categories overlap with poverty based on expenditure? Are poor people in terms of capabilities for example very different from poor people in terms of social inclusion? Or are always broadly the same people poor? This question can be best illustrated by the following picture: Figure 7.21 Three categories of poverty

114 3 ways to look at poverty poor people in terms of expenditure (33%) poor people in terms of capabilities, mainly referring to housing, but health and education also considered (24%) poor people in terms of lack of social inclusion (24%) To what extent do the circles overlap? To what extent are people poor in more than one way? Those people, who are poor in more than one way, can be regarded as the people in greatest need Figure 7.22 Three categories of poverty: Overlaps 3 ways to look at poverty poor people in terms of expenditure (33%) 44% not poor in these categories based on individuals 15% 8% 8% 5% 5% 11% 3% poor people in terms of lack of social inclusion (24%) poor people in terms of capabilities, mainly referring to housing, but health and education also considered (24%) Around 21 percent of our sample (in the shaded areas) is poor in more than one way. This group can be regarded as people in greatest need. Given that there is a group of people in greatest need, then, surely, the very poor in terms of expenditure belong to that group? Who would be in greater need than the very poor (in terms of expenditure)? The following graphic shows, that out of the 10 percent very poor, 60 percent already belong to the group of people in greatest need. The remaining 40 percent,

115 (who are not poor in the other 2 categories) are included into that group of people in greatest need on the strength of the argument, that they are, indeed, in particular great need of assistance. Figure 7.23 Three categories of poverty: including the very poor Where is the group of the very poor located (<60% of the poverty line)? Very poor people in terms of expenditure (10%) 11% 2% 4% 2% 2% 3% poor people in terms of capabilities (24%) poor people in terms of lack of social inclusion (24%) One should also note that some of the people in greatest need are, by the very intention of this perspective above the poverty line (3 percent of total). Demographic and socio-economic characteristics of those in greatest need are presented in the Table Table 7.12 Percentage distribution of respondents by greatest need and others and selected characteristics Selected characteristics Others People in greatest need Total Numbers Migration status** Migrant Non-migrant Location Ger area Apartment area Districts Bayanzurkh Nalaikh Sukhbaatar Songinokhairkhan Khan-Uul Chingeltei Age Below

116 Sex Male Female Ethnicity Khalkh Kazakh Durved Buriad Bayad Other Number of HH members Education* Not educated/ primary Incomplete secondary Complete secondary Vocational/Special technical/high Registration status Registered Not registered Working sector*** Formal Non-formal Abroad Employment status* Employed Not working Type of living quarter Ger Conventional dwelling Number Total * calculated for the population aged 15 and above (n=5053) ** calculated for the population aged (n=4712) *** calculated for the working population aged 7 and above (n=2166)

117 The survey findings reveal that 39.1 percent of the migrants are in greatest need. This is 17 percentage points higher non-migrants. In the previous sections we have been mentioning that although share of poor varies between migrants and non-migrants, the gap is relatively small except only for the differences in capability poverty. It appears that the overall poverty in three ways affect migrants more than non-migrants. Substantial variation can be seen between residents of ger and apartment areas. More than one third or 37 percent of respondents living in ger areas are in greatest need whereas the percentage of those in the greatest need for apartment area is 7.3 percent. From the previous sections we have seen that Nalaikh is one of the poorest districts of Ulaanbaatar. The highest share of people in greatest need (38.2 percent) has been identified in Nalaikh followed by Bayanzurkh where 33.2 percent of the respondents are in greatest need. The survey also shows that a larger proportion of young people tend to be poor. Out of the respondents aged below 25, around 27 percent are in greatest need whereas only 10 percent of those aged 60 and above are in the group of people in greatest need. The percentage distribution of people in greatest need by ethnicity reveals that out of 173 kazaks covered by the survey more than half (56.6 percent) are in vulnerable position. We have noted before that although the number of cases is relatively small kazaks have quite good representativeness in the survey. For other minority ethnic groups such as bayad and those categorized as others, the share of people in greatest need is also high comprising percent. Figure 7.24 Comparing the poor and the people in greatest need 1 Comparing the poor in terms of expenditure and those greatest in need. To what extent are they different? poor in terms of expenditure (n=2258) in greatest need (n=1709) 6 or more people per HH 3 or more children (base HH) 62% 62% 54% 49% Ger khoroolol 81% 88% Ger living quarter 33% 45% age up to 25 59% 59% 0% 20% 40% 60% 80% 100% Base: total of individuals The charts show that the difference between the poor in terms of expenditure and the greatest in need is small. The only meaningful difference can be seen with respect to ger as a living

118 quarter. But that is expected, since living in a ger contributed to being considered greatest in need. Figure 7.24 Comparing the poor and the people in greatest need- 2 Comparing the poor in terms of expenditure and those greatest in need. To what extent are they different? poor in terms of expenditure (n=2258) in greatest need (n=1709) age up to 25 59% 59% female headed HHs (base HHs) less than 6 months in UB (base migrants, n=688) migrants (aged 15-64) 29% 23% 23% 23% 17% 23% Kazakh 5% 6% 0% 20% 40% 60% 80% 100% Base: total of individuals Conclusion Standard poverty measures neglect considerations of capabilities and social inclusion. For that reason, two alternative indicators of poverty have been constructed. It has been checked to what extent these indicators overlap. I.e. do people who are considered poor by one indicator also considered to be poor by another indicator. It has been seen that of the people who are poor in at least of one of the indicators (56 percent), more than one third are poor in at least two dimensions (21 percent of the total). The latter were considered as being those in greatest need (combined with those who are very poor in terms of expenditure). Additional analysis has shown that the characteristics of the people in greatest need do not differ widely from those of the poor in terms of expenditure. Hence, the target group for intervention is largely the same, independent of which approach is used. Hence, even if the standard expenditure based poverty definition does not take into account capabilities and

119 social inclusion, it is in the context of Ulaanbaatar a sufficiently good approximation and leads to similar results, as if they were taken into account.

120 CHAPTER 8. MIGRATION Chapter 8 describes the characteristics of migrants and analyses the link between migration and poverty. This chapter presents in detail the discrepancies in livelihood and living conditions of the population who migrated and didn t migrate, their access to social services and problems encountered to migrated population. By conducting this survey the households which in migrated to Ulaanbaatar city in the last 4 years have been studied. In particular, households the members of which aged and household head that participated in migration were comparatively studied. Key questions - Who are the migrants? - What are the main reasons for migration? Have they changed over time? - Does the registration status influence migration flow? - Is there a link between poverty and migration? - What are the priorities for intervention of the government? Key answers and conclusions - Migrants to the city predominantly live in ger area, with a majority of them having less than complete secondary education. - A majority of the migrants are in search of employment, better livelihood, further studies and closer access to markets. Over time, employment, better livelihood and closer access to markets have gained in relevance, wheras educational needs have declined. Hence, migration can be considered to be increasingly need driven. - Out of the migrants, (71 percent) said that there expectations have been met. - Registration status does not help to reduce the migration flow into Ulaanbaatar (half of the migrants are not registered). However, lack of registration is cited by those migrants without health insurance (one third) as one of the main reasons for not having health insurance (a quarter of those without health insurance stated lack of registration as the reason). - Migrants are not poorer because they are migrants, they are poor because they have lower education levels, for example. Migrants seem to face the same opportunities that the non-migrants face, their problem seems to be that they lack some qualifications in a bit greater extent than the non-migrants - In order to reduce the drive for migration it would be best to improve the economic and education situation in the areas of origin. 8.1 The characteristics of migrants Of the total population covered by the survey 14.6 percent comprised of persons who migrated to the city in the four years preceding the survey (13.4 percent of the households). Among the total households 11.7 percent are households which moved in and 17.8 percent are households with members who migrated in the city. Out of the migrants, 79.5 percent live in ger areas and 20.5 percent in apartment areas. Figure 8.1 shows the age-sex structure of the population who migrated in (hereinafter referred as migrants) and didn t migrate (hereinafter referred as non-migrants).

121 Male Figure 8.1 Pyramid Migrant and non-migrant Non migrant male - Non migrant female - Migrant male - Migrant female Table 8.1 illustrates the characteristics of migrants and non-migrants by location. The proportion of women is higher than men among migrants and non-migrants. More than half of migrants (58.7 percent) are represented by young people between the ages of years. This indicator is 10 points higher compared to non-migrants, but in terms of location it shows no significant discrepancy. However, when it comes to those living in the apartment areas the proportion of young people aged is lower by 14.6 points than their non-migrant counterparts. They are also lower by 6.2 points and 13.5 points respectively than their non-migrant peers in the ger areas and peer migrants. What this shows is that the percentage of people of older ages is higher in the apartment areas where as in the ger areas a high percentage consists of young persons. In terms of marital status, single migrants present 45.9 percent in the ger areas and 49.6 percent in the apartment areas. Whereas in the apartment areas the percentage of non-migrant singles (37.1 percent) exceeds the migrant singles by 12.5 points. This supports the earlier fact that more elder aged people, who are non-migrant and who have been in the city relatively longer are found in the apartment areas. As seen in the table the proportion of widowed people is higher in the apartment areas. The ethnicity of migrants is presented as follows: 81 percent are Khalkh, 4.2 percent are Bayad, 2.8 percent are Durvud, 2.3 percent are Buryad, and 1.7 percent are Kazakh. Table 8.1 Percentage distribution of the respondents aged 15-64, by location and migration status according to demographic characteristics

122 Demographic Ger Apartment Total characteristics Migrant Nonmigrant Migrant Nonmigrant Migrant Nonmigrant Sex Male Female Sex ratio Age Marital status Single Married Living together Separated Divorced Widowed Ethnic group Khalkh Kazakh Durved Buriad Bayad Others* Total Number Note: * Only 9 persons were not Mongolian nationality out of the sample population. Those are included under category Other. 8.2 The characteristics of households and their living conditions The characteristics of households covered by the survey are shown by location and migration in Table 8.2. Most households whether migrants or non-migrants have three to four family members. In other words the number of members in a household for migrants and non-migrant is on average 4.3 percent. The proportion of migrants living in households with 5 or more members is higher than that of non-migrants live in households of the same size. It is

123 interesting to note that the number of families in one household varies a lot by location. According to the survey there are more cases in the apartment areas whereby more than one family, who are non-migrants, share the same house. The explanation may go with two facts. Firstly, newly wed young couples tend to live with their parents because of their preference in living in apartment areas. Secondly, households in apartment areas are likely to have tenants in one of the rooms in the apartment. Also, the number of children in the household is two times higher in ger areas than in apartment areas. Table 8.2 Percentage distributions of the respondents, by location and migration status, according to demographic characteristics of the household Ger Apartment Total Demographic characteristics Migrant Nonmigranmigranmigrant Migrant Non- Migrant Non- Number of people in the HH and over Mean Family size/mean/ Number of migrants in the HH and over Mean Number of families in the HH and more Number of child /Mean/ Total Number of head of HH aged Of migrant households 48.3 percent live in a house (including a house in the ger areas) and 51.7 percent live in a ger (Table 8.3). The proportion of non- migrant households in a ger (87.4 percent is almost 4 times less than that of migrant households (12.6 percent). As witnessed by the survey a significant percentage of migrants from rural areas live in a nonliving quarter. While 36.5 percent of migrant and 50.2 percent of non-migrant households live in a separate house or flat, 11.8 percent of migrant households live in a non living quarter (excludes ger). This percentage is only 2.2 for non-migrant households.

124 Table 8.3 Percentage distribution of the households with heads of the HH aged 15-64, by migration status according to household dwelling characteristics Migratin status Dwelling characteristics Migrant Non-migrant Total Type of dwelling House* Ger Total Number Type of living quarters** House Apartment Public dormitory Non-living quarter Total Number Type of ownership*** Owned Not owned Total Number Note: *- including apartment and public dormitory. **- not including ger

125 ***-not including non-living quarter and other Table 8.4 Percentage distribution of the households living in ger, by migration status according to living conditions Living conditions Migration status Migrant Non-migrant Total Number of ger walls Less than and over Ceiling panel Single Double Wall panel Single Double Floor With floor Without floor Total Number of the households living in gers with head of the household aged The migr ant and non-migrant households have very similar living conditions. For example, more than half of the total households (65 percent) live in a ger with five walls, and there is no significant difference in this term by migration status. In winter time out of all migrant households, 47.3 percent live in a ger with a single felt cover, 46.2 percent in a ger with a single inner cover and 29.7 percent live in a ger without floor. In general, the number of migrant households living in a ger with single felt cover, single inner cover and no floor is found to be not so many. Table 8.5 Percentage distribution of households by migration status according to living conditional selected characteristics

126 Living conditional selected Migration status characteristics Migrant Non-migrant Total Electricity Central Candle Main sources of drinking water Central: hot and cold water cold water only Well Greek stream Heating system Central Non central Other Garbage disposal unit Centralized dust-centre Open places Type of fuel* Electricity Wood Charcoal Other Toilet location Central: indoor flush toilet outdoor flush toilet Pit-latrine None Telephone* Have Don t have Have cellar/garage Have Have not Total Number of the households living in gers with head of the household aged

127 Note: *- calculating from multiple responses. Although, being settled in Ulaanbaatar where the infrastructure is relatively well developed, 8.5 percent of the migrant households and 2.0 percent of the non-migrant households have no access to the centralized electricity network (Table 8.5). The reason is that a large number of migrant households which have settled in so-called open areas are found in ger areas. In addition, the proportion of migrant households is twice as much as the non-migrant households which lack access to centralized hot and cold water systems. Less than half of migrant households and 64.5 percent of non-migrant households are linked to the telephone network, one of the basic necessities in the present time of high technology development. The survey asked questions from individuals on whether they are registered as a household member, the result of which is illustrated in Table 8.6. Table 8.6 Percentage distribution of the population aged 15-64, by location and migration status according to whether registered in the household as a member

128 Whether registered in the Ger Apartment household as a member Migrant Non-migrant Migrant Non-migrant Registered Not registered Total Number The proportion of persons registered as a household member is not that diverse by location. But there is quite a difference between migrant and non-migrant households. In particular, in the ger areas the proportion of persons not registered as a member of migrant households (19.4 percent) are four times higher than those in non-migrant households. Whereas in the apartment areas, this indicator is three times higher (19.9 percent) in migrant households as opposed to non-migrant households. 8.3 Education level of migrants The education level of the population covered by the survey is presented by migration status and location in Table 8.7. The education level of non-migrant persons is generally high. While people with no education and lower than non-completed secondary education level comprise of 38.7 percent among migrants, this percentage is 30.6 percent or 8.1 points lower among the non-migrants. As per high education level the proportion is 11.0 points higher among non-migrants than that of migrants. Explaining this difference is to do with one of the major reasons for migration which is for further schooling or needs of attainment of education. It is worth to draw attention to the high percentage of people without education among the migrants from rural areas. Furthermore, 32.6 percent of migrants in the apartment areas are with high education whereas this percentage is only 6.8 percent for migrants in the ger areas. As the education level of migrants rises, the proportion of migrants in the apartment areas tends to go higher. Hence, it can be concluded that highly educated people often live in the apartment areas. Table 8.7 Percentage distribution of the respondents aged 15 and over, by location and migration status according to socio-economic characteristics

129 Socio-economic Ger Apartment Total characteristics Migrant Nonmigranmigranmigrant Migrant Non- Migrant Non- Education level Non educated Primary Uncomplete secondary Secondary Technical vocational Special vocational/ Diploma High Employment status Employed Not working Total Number School drop out According to the qualitative survey the opportunity to schooling is much lower for children in migrant households compared to those in non-migrant households. Schooling for children is the biggest problem after migration. Schools refuse to recruit children on pretense of oversized classes. So we have to send our children to faraway schools through the connection. (Ts. Female, 39 years old, non-registered migrant, Songino-Khairkhan district) Although the previous surveys concluded that one of the major reasons for school drop out is migration, they omitted to present it in terms of difference by location. To fill in this gap the current survey has studied the school drop out by location. There is a relatively high rate of school drop out among children in the ger areas, but it differs less by migration status. Compared to apartment areas, the number of children who dropped out of school is 8 times higher in ger areas. And more than half the percentages of school drop out children are from non-migrant households in the ger areas. Boys constitute more than half of school drop outs among the migrant households in the ger areas. Of the children in migrant households in the ger areas, 33.3 percent dropped out of school because of no registration. School drop out caused by this reason is not observed in the apartment areas. It is important to note that half of the children in migrant households in the ger areas drop out of school when they are in grades 5-8. Almost all school drop out children in migrant households are not covered by informal training (with only 4 children attended). 8.4 Health status of migrants

130 The survey has revealed the poor access and opportunity to medical services for migrants. The migrants associate this fact with their poor awareness and knowledge of how to access medical services and also, with the unsatisfactory medical services in the ger areas and in suburban areas of the city. Migrants who were not registered with respective administrative units do not have access to basic medical examination and diagnosis. They are covered by medical services only when they get seriously sick. Our family hospital is very far. To be honest, doctors are very arrogant and angry. May be they get irritated by examining many patients all day long. They can t afford to examine all their basic patients not to mention us, migrants. (Ch. Female, 61 years old, pensioner, non-registered migrant, Songino-Khairkhan district) In addition, doctors and staff of the hospitals complained about the poor conditions and facilities in their work. Working condition is very appalling. In less than two years since the move into this building the roof has been damaged and floods often. There is no water system in injection room and no separate room to keep outer garments in. Work condition in the family hospital will be improved when water supply system is made better and a link to central heating system is made. (E. female, 50 years old, family doctor, Songino-Khairkhan district) Table 8.8 Percentage distribution of the respondents, by location and migration status according to health indicators Ger Apartment Total Health indicators Migrant Nonmigrant Migrant Nonmigrant Migrant Nonmigrant Chronic illnesses Healthy Nonhealthy Disability Have got Haven t got Whether have health insurance Have Don t have Don t know Total Number Of the migrants, 13.9 percent reported to have poor health and suffer from chronic diseases diagnosed by the professionals. This answer was shared by 17.2 percent of non-migrants. Such answer by migrants and non-migrants differed by 7.7 points in the apartment areas. Thus, the migrants have slightly better health status than the non-migrants.

131 One third of the migrants are not covered by medical insurance and such situation is found more prevalent among the migrants in the ger areas (35.6 percent) than in the apartment areas (24.1 percent). Table 8.9 Percentage distribution of the respondents, by location and migration status according to main reason for not having health insurance Ger Apartment Total Main reason for not having health insurence Migrant Nonmigrant Migrant Nonmigrant Migrant Nonmigrant Not working Lack of money Not registreted No need Don t know Don t how to have it Uninitiated Working in the private sector Total Number When the individuals were further asked, why they were not covered by medical insurance, the major responses for migrants were as follows: 38.0 percent of respondents said it was a lack of money; 27.1 percent said it was due to, no registration; and 13.1 percent said it was due to unemployment. For non-migrants the major reasons were: the lack of money (33.7 percent), unemployment (28.4 percent) and no need to look for a job (14.7 percent). No medical coverage due to unemployment was the highest percentage among reasons stated by the non-migrants in the ger areas (Table 8.9). Hence, it can be seen that registration prevents migrants from having access to social services, especially health insurance. 8.5 Reasons for in-migration Of the migrants covered by the survey 72.1 percent moved in Ulaanbaatar city 1-4 years ago (Figure 8.2). Figure 8.2 Percentage distribution of migrants, by living duration in UB Less than 6 months 20 6 months to one year years 3-4 years Percent

132 Table 8.10 Percentage distribution of migrants, by location according to selected charateristics Selected characteristics Ger Location Apartment Total Region which living before in migrating Khangai Western Central Eastern Abroud Location which living before in migrating Aimag center Village Soum center Remote rural Abroad Main motives to move Economic Non economic Mixed Total Number When we looked by regions, 30.7 percent of the migrants came from the Central region, 27.3 percent from the Western region, 26.9 percent from the Khangai region and 12.8 percent from the Eastern region (Table 8.10). Of the migrants; 30.7 percent are from the Central region, 27.3 percent of migrants are from the Western region, 26.9 percent are from Khangai region and 12.8 percent are from the East (Table 8.10). According to the percentage of migrants by aimags, the hishest percent of migrants are from Tuv (12.8 percent) and Uvs (9.2 percent) aimags (Figure 8.3). By location, about half of the migrants originated from soum centers. The fact that almost all persons who came from overseas live in the apartment areas draws the attention. It can be explained as follows. On the one hand, international migration is quite supportive to the livelihood of households and on the other hand, people in apartment areas, with more opportunities are involved in international migration. We expected that economic factors would be the main driving reason for migration. However, this was true for only 37.8 percent of migrants while 43 percent associated their movement to the city as non-economic. Yet, the percentage of those who moved to the city driven by economic reasons was found higher among the in-migrants in the ger areas. Those people who lost their livestock and suffered from deprivation come here to make some money by working for others. And some businesses like people move to the city to have a closer access to cultural and educational services. (D. female, 53 years old, khoroo governor, Songino-Khairkhan district)

133 Table 8.11 Percentage distribution of migrants by location according to main expectations of movement