Multidimensional poverty in mekong river delta

Key words: Multidimensional poverty measurement, Deprivation, Alkire & Foster Methodology, Mekong River Delta... The study aims at mapping and measuring of multidimensional poverty in ru

UNIVERSITY OF ECONOMICS HOCHIMINH CITY VIETNAM –NETHERLANDS PROGRAMME FOR M.A IN

DEVELOPMENT ECONOMICS

  

MULTIDIMENSIONAL POVERTY IN

MEKONG RIVER DELTA

Ho Chi Minh, April 2014

By

NGUYEN THI LAN ANH

UNIVERSITY OF ECONOMICS INSTITUTE OF SOCIAL STUDIES

VIETNAM - NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

MULTIDIMENSIONAL POVERTY IN

MEKONG RIVER DELTA

A thesis submitted in partial fulfillment of the requirements for the degree of

Master of Arts in Development Economics

Ho Chi Minh City, April 2014

NGUYEN THI LAN ANH

ACKNOWLEDGEMENTS

First of all, I would like to express our deepest gratitude to my supervisor, Dr Nguyen Huu Dung for his invaluable comments, suggestions and engagement through the learning process of the thesis I would also like to thank Dr Pham Khanh Nam for helpful remarks on my TRD Then I am much obliged to Dr.Tran Tien Khai and Dr Truong Dang Thuy for their enthusiasm which has encouraged me a lot to complete my dissertation

I would also like to thank my Research Methodology module lecturer Dr Luca Tasciotti and Assoc Prof Dr Nguyen Trong Hoai who provided me with the background knowledge of research implementation

It is no doubt that I am deeply indebted to my family members for all the kind understanding and spiritual support Finally, I would like to dedicate my concluding words to all class-mate who involved in this study

ABSTRACT

The study provides a critical review of poverty measurement in Mekong River Delta and arguments for utilizing a recent multi-dimensional methodology framework By applying Alkire & Foster’s methodology for Vietnamese Households Living Standards Survey data 2010 (VHLSS 2010), it produces a multidimensional poverty estimation of Mekong River Delta at the aggregate level, provincial level and identify the biggest contributors to multidimensional poverty Twelve indicators corresponding to four dimensions are considered for estimation The study also works out how policy-makers can prioritize the budget spending among provinces and within each province based on decomposable property of adjusted headcount ratio Lastly, the study investigates the effectiveness of using uni-dimension, per capita consumption, in poverty measurement and reaching a conclusion that consumption alone cannot capture deprivation experienced by the poor households

Key words: Multidimensional poverty measurement, Deprivation, Alkire & Foster Methodology, Mekong River Delta

TABLE OF CONTENTS

DECLARATION i

ACKNOWLEDGEMENTS ii

ABSTRACT iii

TABLE OF CONTENTS iv

LIST OF TABLES vii

LIST OF FIGURES vii

ACRONYMS viii

Chapter 1 INTRODUCTION 1

1.1 Problem statement 1

1.2 Research objectives 2

1.3 Research questions 2

1.4 Research hypothesis 2

1.5 Justification of the study 3

1.6 Organization of the research study 3

Chapter 2 LITERATURE REVIEW 4

2.1 Key concepts of multidimensional poverty 4

2.1.1 Poverty 4

2.1.2 Poverty line 5

2.2 Economic theories of poverty 7

2.3 Approaches to poverty measurement 8

2.3.1 Monetary approach 8

2.3.2 Non-monetary approach 8

2.4 Reviews of empirical studies 10

2.5 Chapter summary 14

2.5.1 Empirical literature summary 14

2.5.2 Problems and limitations of previous studies 17

2.5.3 Conceptual framework 18

Chapter 3 METHODOLOGY 19

3.1 An overview of poverty in Vietnam and Mekong River Delta 19

3.2 Data 20

3.3 Methodology 20

3.3.1 Frequently used indicators of poverty measurement 21

3.3.2 Multidimensional poverty index of Alkire and Foster 22

3.3.2.1 Methodology 22

3.3.2.2 Choice of dimensions, indicators and deprivation cut off 23

3.3.2.3 Application of Weights 31

3.4 Treatment of non-applicable population and missing data 33

3.4.1 Treatment of non-applicable population 33

3.4.2 Treatment of missing data 33

3.5 Analytical framework 34

3.6 Chapter summary 34

Chapter 4 RESULTS 36

4.1 Indicator deprivation 36

4.2 “Across dimension” cut-off and MPI estimation 39

4.3 Poverty estimates at provincial level 40

4.4 Which indicator contributes the most to MPI? 44

4.5 Decomposition of adjusted headcount ratio and Policy implications 45

4.6 Comparison between Consumption poverty and Multidimensional poverty 50

4.6.1 Income poverty verses Multi-dimesional poverty 50

4.6.2 Correspondence of consumption poverty and multidimensional poverty 51

4.6.3 Correlation between Consumption and Multidimensional poverty 51

4.7 Chapter summary 51

Chapter 5 CONCLUSION 52

5.1 Conclusion 52

5.2 Limitation and further researches 53

APPENDIX 55

REFERENCES 55

LIST OF TABLES

Table 2.1: GSO – WB poverty lines 6

Table 2.2 MOLISA poverty lines 7

Table 2.3:Summary of empirical studies relating to multidimensional poverty 14

Table 3.1: Households surveyed in each province 20

Table 3.2: Dimensions, Indicators and cutoffs 30

Table 3.3: Weights assigned to dimensions and indicators 32

Table 4.1: Multidimensional poverty estimate on various cut off point 40

Table 4.2: Poverty estimates at provincial level 42

Table 4.3: Spearman correlation between Mo ranks by different k values 46

Table 4.4: Decomposing of adjusted headcount ratio by indicator 49

Table 4.5: Comparing monetary poor and multidimensional poor at aggregate level 51

Table 4.6: Spearman coefficient between indicators 50

LIST OF FIGURES Figure 2.1: Dimensions and indicators of the MPI ( by Alkire and Foster, 2010) 10

Figure 2.2: Dimensions and indicators of the MPI (by Cuong Nguyen, 2012) 12

Figure 2.3: Ten dimensions with 16 indicators representative for four livelihood assets (by Khai Tran & Danh Nguyen, 2012) 13

Figure 2.4 Conceptual framework 18

Figure 3.1: Analytical framework 34

Figure 4.1: Comparison percentage of households deprived in each indicator 37

Figure 4.2: Proportion of households deprived in each indicators 38

Figure 4.3: Proportion of household deprived in various numbers of indicators 38

Figure 4.4: Contribution by province of each dimension to MPI 44

Figure 4.5: Contribution to MPI by indicators at aggregate level 45

Figure 4.6: MPI compare to consumption poverty by provincial level 50

ACRONYMS

GSO General Statistics Office

HHP Household Prestige index

HDI Human Development Index

MDG Millennium Development Goal

MPI Multidimensional Poverty Index

VHLSS Vietnam Household Living Standard Survey

MOLISA Ministry of Labor, Invalids and Social Affairs

Chapter 1 INTRODUCTION

1.1 Problem statement

Our life is better but poverty is still knocking at the door, from the world’s wealthiest nations to the poorest According to Human Development report (March 2012), nearly a haft the world, or about three billion, are still living on less than 2USD a day; 1.1 billion people in developing countries with inadequate access to water; 800 million people in the world suffer from hunger and malnutrition; among 2.2 billion children, close to haft of children live in poverty, 15 million children die each year from hunger, 1.4 children die every year from lack of access to safe drinking water and adequate sanitary These unexpected deprivations have severely existed in many countries over centuries Facing with this situation, an issue of how to define the poor has been increasingly concerned by economists during the past decades

Conventionally, poverty is defined and measured in money term but poverty definitely goes far beyond the monetary deprivation During recent decades, poverty has been described as a multidimensional phenomenon due to the fact that people may suffer deprivation in many other aspects of life rather than income For that reason, many alternative approaches to measuring and conceptualizing poverty have been explored to reflect its nature, such as Household prestige score HHP (Rossi, 1974), Human Development Index HDI (UNDP, 1990), the Human Poverty Index (UN, 1997), household asset index (Filmer and Prichett, 1998, Booysen et al 2007)… Such approaches encompassed almost aspects of human life like asset ownership, health, education, security, living standard etc Remarkably, multidimensional poverty index developed by Sabina Alkire and James Foster in 2010 which covers deprivation in various social-economic aspects of human being is widely accepted and internationally used during recent years

Keeping in view international as well as national demand of addressing the poverty through multiple dimensions of deprivation, this paper chooses Mekong River Delta to study poverty in wider and deeper sense through various dimensions The study aims at mapping and measuring of multidimensional poverty in rural areas of Mekong River Delta by estimating incidence and depth

of poverty across provinces through multidimensional poverty index (MPI) developed by Alkire and Foster in 2010 Based on research findings of these, some policy implications will be

discussed and suggested to alleviate efficiently poverty in each province from the multidimensional perspectives

iii To compare between monetary poor and multidimensional poor indexes

iv To propose some necessary/appropriate interventions for alleviating poverty

1.3 Research questions

i How width and depth of multidimensional poverty in Mekong River Delta?

ii Which deprivation contributes the most to multidimensional poverty at aggregated level and provincial level among rural areas of the Mekong River Delta?

iii How difference can be between consumption poor and multidimensional poor indexes?

iv Can consumption alone measure poverty experienced by households?

1.4 Research hypothesis

i Multidimensional poverty is severe in Mekong River Delta

ii Deprivation in per capita consumption, sanitation, cultivated land and housing quality contributes the most to rural multidimensional poverty

iii The poverty rate is much higher in multidimensional poverty than consumption poverty

iv Consumption alone cannot measure poverty experienced by households

1.5 Justification of the study

Given the absence of research on rural poverty in multidimensional sense in Vietnam, this type of research is necessary The study tries to discover the state of multidimensional poverty in rural of Mekong River Delta by using Alkire and Foster methodology and find out which indicator impacts the most to overall MPI Result of study is informative and useful for rural policy maker

in determining scope and aspects which need to be addressed to alleviate poverty and, on that basis, they propose some recommendations on investigation or relief to improve rural living standards in Mekong River Delta

1.6 Organization of the research study

The rest of this study is organized as follows Chapter two introduces some key concepts of multidimensional poverty, some economic theories, a review of literature and relevant research associated with multidimensional poverty In addition to a review of literature, chapter two also presents the conceptual framework and builds on an analytical framework to capture multi aspects

of poverty Chapter three elaborates on the data and methodology for multidimensional poverty measurement Chapter four reflects data analysis and result presentation Chapter five concludes with research’s discussion, recommendation for practice and direction for further researches

Chapter 2 LITERATURE REVIEW

This chapter begins by describing some key concepts of poverty which have been developed over time Economic theories of poverty are then highlighted (section 2) and brief overview of approaches to measure poverty in Mekong River Delta is provided (section 3) Some empirical studies are then presented (section 4) The conceptual framework for multidimensional deprivation

is defined (section 5) and analytical framework to construct multidimensional poverty index is outlined (section 6)

2.1 Key concepts of multidimensional poverty

2.1.1 Poverty

For poverty, two of the most typical definitions accepted world-wide are from United Nations and the World Bank The United Nations measures poverty by various attributes that contribute to human well-being The United Nations states that:

“Fundamentally, poverty is a denial of choices and opportunities, a violation of human dignity It means lack of basic capacity to participate effectively in society It means not having enough to feed and cloth a family, not having a school or clinic to go to; not having the land on which to grow one’s food or a job to earn one’s living, not having access to credit It means insecurity, powerlessness and exclusion of individuals, households and communities It means susceptibility

to violence, and it often implies living on marginal or fragile environments, without access to clean water or sanitation“(United Nations, 1998)

Similarly, another multi-dimensional poverty definition which is widely used is the one from World Bank:

”Poverty is pronounced deprivation in well-being, and comprises many dimensions It includes low incomes and the inability to acquire the basic goods and services necessary for survival with dignity Poverty also encompasses low levels of health and education, poor access to clean water and sanitation, inadequate physical security, lack of voice, and insufficient capacity and opportunity to better one’s life.” (World Bank, 2000)

Besides, in the paper in 2009, Alkire and Foster defined multidimensional poverty as:

“A multidimensional poor person who is if and only if having both 16 economically deprived and socially deprived A person is economically deprived if the person’s income falls below the income cutoff as a shortfall in the space of income (consumption, wealth) A person is socially deprived if any social achievement falls below its respective cutoff It focuses on the social dimensions pertaining to health, education, and social protection, and is based on the assumption that each social dimension is intrinsically important, and that an attainment below the respective cutoff represents a denial of a basic human right” (Alkire and Foster, 2009)

Based on the above concepts, there are two types of poverty Absolute poverty is the state where

an individual found lacking sufficient resources to meet the basic needs like food, shelters, cloths….Relative poverty refers to the situation of population living under an average level of society

In Vietnam, there are two approaches to measurement of poverty line: International poverty line measurement approach and National Poverty lines measurement approach (considered as the official poverty line)

The first approach is based on international standards, in which poverty line was calculated by the General Statistic Office (GSO) with technical assistance from World Bank This poverty line was built for 2 periods 1993-2008 and 2010-later In the first period, poverty line was determined by equivalent money of acquiring food for sufficient nutrient standard of 2.100 calories per person per a day adding the cost of necessary non-food items like shelters, clothes, etc People whose expenditure fall below this minimum level are considered poor In the second period, poverty line was calculated by the cost of basic food for 2.230 calories per person per day plus cost of essential non food items In each period, this approach produces only one poverty line for both rural and urban This poverty line then has been kept constant over time in real purchasing power Table below will present the changes of poverty line over the time:

Table 2.1: GSO – WB poverty lines

The second approach, developed by Ministry of Labor, Invalid and Social Affairs (MOLISA) is based on the growth rate of economy, the financial resources reserved for poverty reduction Then, the welfare indicator is expressed by per capita income adjusted for the regional differences in terms of number of kilogram of rice which is equal to 15kg, 20kg and 25 kg of rice per person per day for mountainous countryside and islands, the delta countryside and midlands and the urban areas respectively The poverty line was built in the first year of five year period, depicted in table 2.2:

Table 2.2 MOLISA poverty lines

2.2 Economic theories of poverty

There are conflict economic theories of Poverty However, they all share the notion that poverty is

an economic status which one cannot satisfy basic needs for subsistence These theories “can be simplistically lumped into two groups: theories that focus on individualistic behaviors and theories that focus on social structures” (Sherraden, 1991 p 35)

The individualistic theories of poverty (Schultz, 1963; Becker, 1964) state that what constitutes poverty are found in attributes inherently existing in individual characteristics as well as in the personal abilities of the poor themselves The explanation is that each individual is a rational being who has totally power to act for maximizing their own interest Hence, determinants of poverty lie

in their behaviors upon making choice/decision Those who fail to make the right choices/decisions and those who cannot compete with others become poor

Structural or situational theories of poverty (Doeringer & Piore, 1971; North, 1990; Sherraden 1991; Burton, 1992; Rank, 1994; etc.), on the other hand, asserted that causes of poverty are found

“not in the characteristics of the poor themselves, but in the structural elements of the larger society” (Burton, 1992, p.149) The explanation follows that each individual behavior is influenced by external factors, not his/her own characteristics Those with better social economic status/relations have dominant choices while those who have lower social economic/relation usually do not have opportunities to make the right choices Due to this inequality, the rich will become richer and the poor always remains poor regardless of their hard work or their dominant values In other words, it is the failure of structure of society which causes poverty

2.3 Approaches to poverty measurement

There are two main approaches on the theoretical ground: the monetary approach and the non monetary approach

2.3.1 Monetary approach

The monetary approach which is also called uni-dimensional approach has been pioneered by Rowntree (1901, 1941) He carried out a survey of living conditions of poor families in York, England in 1899 and drew a poverty line in terms of a minimum weekly sum of money which enables families to secure the common necessities of a healthy life This minimum sum of money, called a “bare subsistence” covered light and fuel, food, clothing, rent and personal items…He did

it again in 1936 and used a “relative” concept of poverty line “what human needs are for a minimum necessary for a healthy life” Adjust for the movement of price, he wanted to address whether the poverty in York (England) had been decreased or increased over more than 30 years

This approach has been widely used by economics because it is relatively easy to calculate numerically and easily apply policy-based However, it remains many practical problems even though it is well-known; for example, the basis of recall data, difficulties of converting products into money terms and the seasonality of information Above all, it does not explain enough the multiple aspects passing household living conditions

2.3.2 Non-monetary approach

Contrary to the monetary approach which makes a linkage between poverty and expenditure, the non-monetary approach considers poverty from the point of view of multiple deprivations Over its long history, many indicators have been applied to measure poverty This paper only presents three of them

Household asset index of Filmer and Pritchett (1999, 2001) are based on asset ownership and housing characteristics to measure household’s well-being They used data from India and applied principle component analysis for constructing asset index to explore the impacts of household wealth on educational enrollment across states in India Filmer and Prichett illustrated that classification household based on asset indicators are very close to expenditure-base yet more accuracy and validity Different from money-metric measure, asset index are based on data

collected from interview through inspection checklist, on that account, the accuracy and validity of the asset data are better than that of income-based or expenditure-based They also suggested:

“using assets as instruments for household per capita expenditures is most likely the more effective way of extracting the maximum amount of information from the data while reducing the impact of measurement error” and by economic evidences, they convince that “asset index, as a proxy of economic status for use in predicting enrollments, is at least as reliable as conventionally measured consumption” (Filmer and Prichette 2001)

Following Filmer and Prichett study, in the year 2008, Booysen et al assessed trends in poverty in

seven African countries over time (1986 – 1992) and across countries using asset index constructed from nationally data of demographic and health survey (DHS) However, difference

from Filmer and Prichett study, Booysen et al employ multiple correspondence analysis (MCA)

which works better for category variables, rather than principle component analysis (PCA) to

derive the weight of each asset indicator For inter-country comparability, Booysen et al used two

conventional poverty lines suggested by World Bank, the 40th percentile and 60th percentile together with one absolute poverty line The absolute poverty line is derived by using weighted sum of assets which represent an adequate condition for subsistence All the three poverty lines are constructed from aggregate data for the purpose of comparability They also claim that asset index

is not relatively volatile as income or expenditure yet slow-moving due to the slow rate of change

Human poverty index (HPI) developed by the United Nations (UN) in 1997 is another well-known poverty indicator It focuses on deprivation of the most basic dimensions of poverty: Survival, knowledge and standard of living HPI is derived separately for developing countries (HPI-1) and developed countries (HPI-2) to reflect differences in socio-economic features In spite of concentrating on a wider set of attributes than income, HPI have not covered totally the multiple aspects of deprivation which human being may suffer

Multidimensional Poverty Index of Alkire and Foster is an international measure of acute poverty that replaces for HPI during recent years The MPI exhibits a different pattern of poverty than income poverty as it takes into account not only material standards of living but also all attributes which affect household living conditions Alkire and Foster (2010) proposed that MPI combines two aspects of poverty: Incidence (H) - the percentage of people who are poor or the headcount

ratio and intensity of people’s poverty (A) or the average percentage of dimensions in which the poor are deprived Three dimensions are included in the study: health, education, and standard of living These three dimensions are measured by ten indicators Each dimension is equally weighted as well as each indicator in each dimension, briefly represented as below diagram:

Figure 2.1: Dimensions and indicators of the multidimensional poverty index

(By Alkire and Foster, 2010)

Alkire and Foster claim that there are two main approaches to identify the poor “Union” approach defines those who deprive at least one dimension as the poor The other main approach is

“intersection” method which states that someone is poor if he or she deprives in all dimensions

2.4 Reviews of empirical studies

Many studies on multidimensional poverty have been conducted for all countries in the world, especially for developing countries Each of them has different approach

Concerning household asset index, in the paper in 2006, Prakongsai used data from two household surveys data, the Socio-economic Survey (SES) and the Health and Welfare survey (HWS), to construct asset index to measure poverty in Thailand over the period 1998, 2000, and 2002 He used a range of 28 to 30 variables (including: Washing machine, Telephone, Video recorder, Refrigerator, Electrical cooking pot, Mobile phone, Bed, Air conditioner, Iron, Water boiler, Sofa

in living room, Electrical iron, Gas cooking stove, Water boiler, Toilet, Bicycle, Light bulb…) which exhibit housing characteristics, ownership of durable and semi-durable assets and, types of household sanitation and water supply Principle Principal Component Analysis (PCA) which was presented by Filmer and Prichett in 2001 was used to derive household asset index for each household By comparing between poor household of the first decile determined by household

asset index and the first decile classified by the national poverty line, he found that the correlation between the poor classified by using the asset index and household income/expenditure in 2002 is only 0.23 and 0.28 respectively This comparison is based on the assumption that household of the first decile determined by household asset index is classified as the poor due to household below the national poverty line of Thailand in 2002 account for 9.8% Meanwhile, Pearson Correlation shows that household classified by asset index is correlated with household classified by expenditure/income at over 50%

Maria Emma Santos and Karma Ura applied Alkire and Foster methodology to estimate multidimensional poverty in Bhutan using 2007 Bhutan Living Standard Survey Poverty are reflected in 5 dimensions in both urban and rural areas: income, education, room availability, access to electricity and access to safe drinking water, and poverty in rural are exclusively reflected in two additional dimensions: access to road and land ownership Two weighting structures are used: one using equal weighting structures and one using weights identified through Gross National Happiness Survey (GNHS) The study found that poverty is fundamentally a rural phenomenon In rural areas, when equal weights are used, deprivation in electricity, room availability, education and income are the ones with the highest share in overall deprivation, followed by access to road, land ownership and water respectively In counterpart, when weights obtained from GNHS, income and education have higher share to overall poverty relative to access to electricity and room availability due to they are assigned higher weights

In the paper of Asselin and Vu (2009, p.124), multidimensional poverty index is applied for analyzing dynamic poverty in Vietnam during the period 1993-2002, using Vietnam living standard survey (VLSS) data In this paper, five dimensions are presented, including education, water/sanitation, health, employment and housing Eight indicators covered are underemployment, chronic sickness; adult illiteracy; under-schooling and; without radio, TV, type of dwelling; and drinking water sanitation A composite indicator of poverty (CIP) on each household was built base on the weight derived for each indicator using multiple correspondence analysis (MCA) technique Then, CIP absolute poverty line defined as “the average of the poverty thresholds specific to the primary indicators” (Asselin 2009, p.142) in view of the union-poverty concept is computed; thus, MPI indices are constructed Keeping the weights and CIP absolute poverty line unchanged over studied period, multidimensional poverty can be compared to consumption

poverty; and also, analysis of poverty trends over the period 1992–2002 was carried out The findings show that poverty has declined significantly during the 1993-1998 period, which match well with consumption poverty Among seven regions, Red River Delta obtained most dramatic reduction; conversely, the Mekong River Delta shows the lowest improvement

Cuong Nguyen (2010) constructed multidimensional poverty index to estimate multidimensional poverty in five central cities in Vietnam: Hanoi, Haiphong, Danang, Hochiminh and Cantho using Vietnamese household standard living survey 2008 In this research, the authors takes into account five main dimensions including education, health, living standard, economic well-being, and employment labor These dimensions are measured by fifteen indicators (figure 2.2) Equal weights are applied within dimension and across dimensions What he obtained was multidimensional poverty is significantly high in five central cites, especially in Hochiminh city and the poor intensively suffer from deprivation of under-employment, types of dwelling, working time and housing space

Figure 2.2: Dimensions and indicators of the multidimensional poverty index

(by Cuong Nguyen, 2012)

Be Lam (2012) constructed multidimensional poverty index to estimate multidimensional poverty

in Mekong River Delta using Vietnamese household standard living survey 2008 In this research, ten indicators pertaining to three dimension health, education and living standards were selected The study also applied equal weighting system for multidimensional poverty estimation The result showed that MRD emerged as the poorest region among 8 regions (Red River Delta, North Eastern, North Western, North Central, South Central, Central Highland, South Eastern and MeKong River Delta) in terms of MPI although it ranked third in monetary poverty, after South

Eastern and Red River Delta His study also concluded that poverty is much severe in multidimensional sense than monetary poverty

Different from the above studies, Khai Tran & Danh Nguyen (2012) did not construct any index for measuring poverty, they instead oriented to investigating interrelations between monetary poverty and other rural households’ socioeconomic in Vietnam through livelihood assets approach and selecting appropriate indicators for measuring multidimensional poverty Data of rural households extracted from VHLSS 2008 data set was employed They began with twenty three socioeconomic indicators which divided into four categories of livelihood assets: human, natural, physical and financial assets The result of Principle component analysis and multiple correspondence analysis methods allow to selecting 16 indicators related to four livelihood assets for further analysis Cluster analysis then will be explored for measuring multidimensional poverty The result exhibits some important findings which give remarkable contribution to multidimensional poverty research in Vietnamese rural households The first finding is there must

be remarkable differences between monetary poverty and multidimensional poverty and expenditure per capital has lowest impact on poverty in multidimensional perspective Each dimension and indicator has different contribution level to overall poverty is the second finding The third finding is physical asset has impressive contribution to multidimensional poverty Finally, difference in variable importance suggests that weight assigned to each dimension is very important, but how is still an open question

Figure 2.3: Ten dimensions with 16 indicators representative for four livelihood assets

(by Khai Tran and Danh Nguyen, 2012)

(1) Human asset Human resource for

Job diversification ability

Number of household member working non-farm

Livelihood asset Dimension Relevant indicators

(3) Physical asset Housing condition House value; Housing area; Type of house

(4) Financial asset Additional income Remittance received within year

2.5 Chapter summary

2.5.1 Empirical literature summary

The following table summarizes the common used dimensions and indicators as well as indexes and methodologies applied for investigating multi aspects of poverty Although these studies are different by approaches, they all focus on the basic subsistence needs of human being

Table 2.3: Summary of empirical studies relating to multidimensional poverty

No Author Dimensions &

1 Prakongsai,

2006

30 variables which exhibit housing characteristics, ownership of durable and semi-durable assets and, types of household sanitation and water supply

Household asset index with Principle components analysis

Thailand, 1998,2000,

2002

- Decile compare:

Correlation between the poor classified by asset index and household income/expenditure in

2002 is only 0.23 and 0.28 respectively

- Pearson Correlation: household classified by asset index is correlated with household classified

by expenditure/income at over 50%

No Author Dimensions &

income, education, room availability, access to electricity and access to safe drinking water,

in rural the study considered two additional dimensions:

access to road and land ownership

MPI

2007 Bhutan Living Standard Survey data

- Poverty is fundamentally a rural phenomenon

- In rural areas, when equal weights are used, deprivation in electricity, room availability,

education and income are the ones with the highest share in overall

deprivation In counterpart, when weights obtained from GNHS, income and education have higher share to overall poverty relative to access to electricity and room availability due to they are assigned higher weights

3 Asselin and

Vu, 2009

five dimensions are presented, including education, water/sanitation, health,

MPI, multiple correspondenc

e analysis

Vietnam, 1993-2002

- Poverty has declined significantly during the 1993-1998 period for both multidimensional and consumption poverty concepts

No Author Dimensions &

employment and housing

- Red River Delta obtained most dramatic reduction, the Mekong River Delta shows the lowest improvement

MPI

Hanoi, Haiphong, Danang, Hochiminh and Cantho, VHLSS2008

- Multidimensional poverty is significantly high in five central cites, especially in Ho Chi Minh city

- The poor intensively suffer from deprivation of under-employment, types

of dwelling, working time and housing space

multivariate analysis methods as Principle Component Analysis, Multiple Corresponden

ce Analysis and Cluster Analysis

Vietnam rural household, VHLSS2008

- There must be remarkable differences between monetary poverty and multidimensional poverty and expenditure per capital has lowest impact

on poverty in multidimensional perspective

- Each dimension and

No Author Dimensions &

indicator has different contribution level to overall poverty

- physical asset has impressive contribution

to multidimensional poverty

- weight assigned to each dimension is very

MPI

Mekong River Delta, VHLSS2008

- MPI is much higher than monetary poverty

- Although monetary poverty in MRD is 14.6 percent, ranked as the third poorest region in Vietnam but in MPI, MRD emerges as the poorest region

2.5.2 Problems and limitations of previous studies

Despite many studies constructing multidimensional poverty indexes for developing countries, empirical evidence for Vietnam is very limited Most of the studies focus on MPI in largest cities

or Vietnam as a whole rather than in rural area, except for Be Lam However, the study of Be Lam skipped over land possession, an important asset of rural households To overcome such drawback, this study takes into consideration land possession

Besides, most of the studies do not utilize the decomposable property of MPI This study will enrich the current studies on MPI by employing this property in giving implication for distributing resources

2.5.3 Conceptual framework

Figure 2.4 Conceptual frameworks

The literature review attempted that poverty is conceptually a multidimensional issue, distinguishing the various aspects of human life affected by poverty, including both economic and non-economic factors Many dimensions/indicators are identified in the literature review Among

of them, education, health, living standards and economic well-being are the most focus A household is deprived in multi-dimensional sense if it deprived in more than one dimension

Education Deprivati

on

Living standard Deprivation

Health Deprivatio

n

Wealth Deprivatio

n

Multidimensional

poverty

Chapter 3 RESEARCH METHODOLOGY

The previous chapter discusses on theoretical basis of poverty measurement This chapter presents

an overview of poverty in Vietnam and Mekong River Delta, the data used and methodology applied for poverty measurement The first sub-section of methodology is often used poverty indicators The second one is multidimensional poverty index of Alkire & Foster as well as the argument for indicators selection and weighting application

3.1 An overview of poverty in Vietnam and Mekong River Delta

There is a growing concern relating to poverty dilemma in Vietnam In the years of 1990s and 2000s, Vietnam seems to have great progress Specifically, Vietnam is recorded as one of the country which has good policy for poverty reduction in accordance with method and criteria given

by World Bank Poverty rate in of Vietnam reduces remarkably from 58.1% in year 1990 to 37.4%

in year 1998 and 14.5% in the year 2008 (GSO), using the poverty line of Vietnamese General Statistical Office Despite the fast poverty reduction, poverty still remains one of the most serious problems facing Vietnamese Government due to its massive and pervasive impacts

Among 8 regions in Vietnam, Mekong Delta River is one of the poorest one even though being known as a biggest “rice bowl” of Vietnam According to Vietnamese household living standard survey 2010 conducted by General Statistic, poverty rate in Mekong River Delta reduce significantly from 15.3% in year 2004 to 8.9% in year 2010 However, this ratio is still rather high, higher 6.5% compare with Red river Delta and 2.2% compare with the Southeast; Per capita income of Mekong River Delta is 1,247.2, significantly lower than the overall average rate of the country which is equal 1,387.2; For education and health, MRD emerged as the poorest region, 34.4% population of Mekong River Delta have not finished primary school; 47.9% of its population has suffered medical treatment This puts pressure on local authority in alleviating poverty Although poverty exists in both rural and urban areas, the characteristics of those who live in these two places and the severe level of poverty are distinctly different among places People who live in remote rural areas have less power to influence on decision-making for service delivery like school, electricity distribution system, clean water system, sanitation… Hence, they are lack of opportunities for studying, access to electricity, clean water, sanitation…which in turn

result in deprivation of education, electricity, clean water, sanitation… Therefore, the study chooses rural of MRD for targeting the poor

3.2 Data

The research explores MPI measure with household as a unit of analysis to address the research questions The main data is the secondary data extracted from the Vietnam Household Living Standards Survey (VHLSS) 2010 of Vietnamese General Statistics Office The study used data with sample of 1455 households surveyed in the Mekong River Delta Details of data extraction from VHLSS 2010 are presented in the appendix 1 and 2

Table 3.1: Households surveyed in each province

Province Number of households Percentage

3.3.1 Frequently used indicators of poverty measurement

In the old days, several poverty indicators have been developed, most remarkably head count ratio and poverty gap index

Headcount ratio is by far the most used in terms of poverty measurement This indicator measures the incidence of poverty and is expressed by the proportion of the income poor population computed by dividing the number of poor by the total sample population, where individual is the unit of analysis The poor people are defined as the people under the poverty line The formula for calculating ratio of people living below the poverty line is as follows:

n

q z y I n

1

1

Where:

H: Headcount ratio N: Total sample population q: Number of poor people

I (yi ≤ z): indicator function which takes the value 1 if the expression (yi≤z) is true

or the value 0 otherwise So, if income (yi) is less than poverty line (z), I equals to 1; otherwise, I euquals to 0

Headcount ratio is the simplest poverty indicator but it only takes the number of the poor and ignores the intensity of poverty Poverty gap index, a more sophisticated poverty indicator, combines both number of the poor and the depth of poverty It is constructed by dividing normalized income shortfall of the poor by total population or multiplying headcount ratio by average income shortfall:

y z n

PGI

1

.

I

1

1

Where:

PGI: Poverty gap index

I: average income shortfall

n: Total sample population

z: Poverty line

yi: Income of the poor i

3.3.2 Multidimensional poverty index of Alkire and Foster

3.3.2.1 Methodology

AF methodology involves two steps: Identification – ‘defining the criteria for being considered as

the poor’ and aggregation – ‘Obtaining an overall indicator of poverty by bringing data of all poor

household together (Sen, 1976)

(1) Multidimensional poverty identification

To define the poor, AF methodology uses a dual cut off, ‘within’ dimension cut off to identify whether a household is deprived in each dimension and ‘across dimension’ cut off to find out whether a household is suffered multidimensional poverty

Denote f(ci, k) as identification where ci represents weighted deprivation counts or sum of weighted deprivation household i suffered, k refers to ‘cross dimension’ cut off A household is identified as multidimensional poor or f(ci, k) = 1 if ci ≥ k and not multidimensional poor or f(ci, k)

= 0 if ci < k Therefore, a household is defined as multidimensional poverty if it deprived in at least k weighted dimensions

(2) Multidimensional poverty aggregation

Given the above identification function, the total number of households which are poor in a multidimensional sense is defined as:

The headcount ratio or percentage of people who are MPI poor:

),(1

1

k c f n n

q H

n

i i







To reflex how breadth of deprivation a household suffered, denote d as the number of weighted dimensions considered and c(k) as the vector of weighted deprivation counts such that ci(k) = ci if

ci ≥ k ; otherwise, ci(k) = 0 Then, average intensity of MPI poverty across the poor is calculated:

( )/( )

1

qd k c A

n

i i





In which ci is the sum of weighted dimensions in which a household i is deprived

Multidimensional poverty index:

MPI can be decomposed into the contribution of sub-groups of population (Alkire & Foster, 2011) Let A and B be two data matrices of each sub-groups achievement respectively; (A, B) be the matrix of total population achievement Denote n(A) as number of households in sub-group 1 and n(B) as the number of households in sub-group 2, we have:

),(),(

)(),(),(

)(),,

B A n

B n z A M B A n

A n z B A

In which z = [zj]: row vector of cut-offs where zi is the threshold below which a household is considered multidimensional poor

3.3.2.2 Choice of dimensions, indicators and deprivation cut off

The main idea of multidimensional poverty index (MPI) is that deprivation is multidimensional In other words, each person may be deprived in one or more than one aspect of their life, called dimension Multi-deprivation is the aggregation of these dimensions in which each dimension deprivation will be measured by many indicators Therefore, each dimension and its indicators, with their own impact on MPI, should be clearly identified In this study, dimensions, indicators and deprivation cut off have been selected on the basis of literature review, previous studies, public consensus and data available The selected dimensions, indicators and their deprivation cut-offs are illustrated below

Education Deprivation Dimension

It is undoubted that education impacts overall quality of human life, both personally and professionally Specifically, education links to the capability of applying knowledge, information and theories to achieve a wide range of tasks and goals For that reasons, education, internationally admitted as one the most powerful instruments for poverty reduction, is selected To measure education deprivation, the study uses two indicators: years of schooling and school enrolment

Years of schoolings

Years of schoolings is defined as the highest level of education which an individual has completed It has been the most used indicator of education measurement and can be taken as a proxy for literacy skill, numeric skill and acknowledgment of information These skills are not only important for urban citizens but also for rural citizens In rural areas, basic literacy and numeracy is particularly necessary in agricultural production Improving literacy and numeric skills may help the farmers to understand information, to catch up with policies/Government’s supporting programs which benefit them, to calculate appropriate input rate for agricultural cultivation and to apply new technologies, new seeds, etc for improving productivity Many statistical studies also proved that educated farmers are more efficiency than non-educated farmers For reducing poverty and growing purpose, achieving Universal primary education is the first priority in the poorest countries as well as in the wealthiest It is the second goal in the United Nations Millennium Development Goal Hence, in targeting the poor, this study uses primary education completion as the cut of point The cut-off is also suggested by Alkire and Foster in

recent studies Following their suggestion, a household will be considered deprived in education if

no household member has completed five years of school In contrast, a household will be deprived in education if at least one household member has completed the fifth level The choice

non-also coincides with the concept of “effective literacy” of Basu, K and Foster, J (1998) that there

is intra-household externality arising from literate member in which every household member will benefits from a literate member regardless of the other’s actual level of education

School enrolment

Beside years of schooling, school enrolment of children is used as complementary indicator for assessing education-deprived Children education plays an important role in enabling them to meet the variety of challenges and obstacles of the world around them It is relative to urban, as it is in

rural Therefore, school enrolment of children should be addressed A household is considered

deprived if at least one school-aged child (6-15 years old) is not attending school This is relative

to external effects When a child is not attending school, the household education level in current and future will be reduced This brings about a decrease in their knowledge and their ability In comparison with the years of schooling, school enrolment is much more sensitive to the changes

in policy

Health and health care deprivation

Health is an asset of human and occupies most part of our life Without health, people cannot join any active service; cannot be a successful professional, nor can live a happy life In reverse, people with good health are able to perform to the best of their ability in any activities, increase the capacity to work and employment opportunities which in turn increase earnings and above all, people with good health can live his life happily Therefore, having a long and healthy life is a prerequisite for human development As a result, health is one of important dimension in measuring multidimensional poverty For measuring health deprivation, the study has included

three following indicators: medical treatment affordability, food security and health insurance

Medical treatment affordability deprivation

It is well-known in the literature review that health and poverty are interconnected Poverty could lead to worse health through various paths including malnutrition, lack of awareness, less chance

to have good health care ….Above all, a lack of affordable health service impacts seriously on physical health Once a person cannot afford for primary health treatment, he/she usually suffers more serious diseases and expensive care Hence, medical treatment affordability is considered as

one dimension of health A household with at least one member cannot afford medical payment is

supposed to be deprived in medical treatment

Food security deprivation

Food is one of the most essential elements for life of human being It has a significant impact on one’s health In the absence of level of health information, it can be seen as a proxy for this

dimension A household is supposed to be deprived in food security if it does not have access to

enough food

Health insurance deprivation

Health is the most significant concern in our peace time However, health risks are unpredictable and incalculable Hence, medical treatment expenses do not only appear to be a burden for patients but also for their family According to GSO’s statistics, there is about 150 millions participations come to the hospital for medical examinations every year in which more than 7 millions participations must hospitalized and 6 millions are out-patient Among them, many patients cannot afford medical treatment and barely empty because of hospital fees Many even refuse to be hospitalized or to take surgeons because medical expenses are too high in compared with their income Facing this problem, health insurance, a form of risk-sharing, is increasingly recognized

as an indispensable cure to protect us against financial consequences of diseases Without insurance, people may suffer a severe financial burden due to unpredicted risks of health Moreover, being uninsured also links to health problems Uninsured people, as explained above, are limited access to health care and used to delay care for primary health problem which may leads to more serious diseases and expensive care Thus, health insurance is a necessary It is proved by the fact that percentage of insured people has been increasing day after day, from 1% in

1993 to 5.5% in 1997 and 64% in 2010 and currently on the way of reaching to the target of health insurance coverage for the whole population With demographic characteristics of older, poorer and less formally educated population, health insurance is particularly important for rural

population A household is considered health insurance deprived if no household member has

insurance coverage of any type

Living standards Deprivation

Standard of living is one of the fundamental determinants of individual level well-being Obviously, improving standard of living constitutes an indispensable goal of program/policy of