INTRODUCTION
Problem statement
Despite advancements in our quality of life, poverty remains a pressing issue globally, affecting both wealthy and impoverished nations The Human Development Report (March 2012) reveals that nearly half of the world's population, approximately three billion people, live on less than $2 a day In developing countries, 1.1 billion individuals lack adequate access to clean water, while 800 million suffer from hunger and malnutrition Alarmingly, nearly half of the 2.2 billion children worldwide live in poverty, with 15 million children dying each year from hunger and 1.4 million from insufficient access to safe drinking water and sanitation These stark realities have persisted for centuries, prompting economists to increasingly focus on how to define poverty in recent decades.
Poverty is traditionally measured in monetary terms, but it encompasses much more than just financial deprivation In recent decades, it has been recognized as a multidimensional issue, where individuals may experience various forms of deprivation beyond income This has led to the exploration of alternative methods for measuring poverty, such as the Household Prestige Score (HHP), Human Development Index (HDI), Human Poverty Index, and household asset index, which consider factors like asset ownership, health, education, security, and living standards Notably, the Multidimensional Poverty Index, developed by Sabina Alkire and James Foster in 2010, has gained widespread acceptance and is used internationally to assess deprivation across various socio-economic dimensions.
This paper focuses on the Mekong River Delta to explore poverty from a multidimensional perspective, addressing both international and national demands for comprehensive poverty alleviation It aims to map and measure multidimensional poverty in the region's rural areas by assessing the incidence and depth of poverty across provinces using the Multidimensional Poverty Index (MPI) developed by Alkire and Foster in 2010 The findings will inform policy implications aimed at effectively reducing poverty in each province through a multidimensional approach.
Research objectives
This study aims to assess multidimensional poverty across provinces in the rural Mekong River Delta, identifying key deprivations that significantly impact overall and provincial poverty levels Additionally, it seeks to compare the monetary poverty index with the multidimensional poverty index Finally, the research will propose targeted interventions to effectively alleviate poverty in the region.
Research questions
The Mekong River Delta exhibits significant multidimensional poverty, characterized by varying widths and depths across its rural areas At both the aggregated and provincial levels, certain deprivations, such as access to education, healthcare, and living standards, play a crucial role in contributing to this poverty Additionally, there exists a notable disparity between consumption poverty and multidimensional poverty indexes, highlighting that households may experience different poverty levels based on consumption alone Consequently, relying solely on consumption metrics is insufficient to fully capture the complexities of poverty faced by households in the region.
Research hypothesis
Multidimensional poverty is a significant issue in the Mekong River Delta, characterized by severe deprivation in areas such as per capita consumption, sanitation, cultivated land, and housing quality This form of poverty impacts rural communities more profoundly, as the poverty rate associated with multidimensional poverty far exceeds that of consumption poverty Therefore, relying solely on consumption metrics fails to capture the full extent of poverty experienced by households in this region.
Justification of the study
Research on multidimensional poverty in rural Vietnam is essential due to its current absence This study focuses on assessing the state of multidimensional poverty in the Mekong River Delta using the Alkire and Foster methodology, identifying which indicators most significantly affect the overall Multidimensional Poverty Index (MPI) The findings provide valuable insights for rural policymakers, helping them determine critical areas for intervention to alleviate poverty Based on these insights, the study offers recommendations aimed at improving rural living standards in the Mekong River Delta.
Organization of the research study
This study is structured into several chapters: Chapter two explores key concepts and economic theories related to multidimensional poverty, along with a literature review and a conceptual framework that captures various aspects of poverty Chapter three details the data and methodology used for measuring multidimensional poverty, while Chapter four presents the data analysis and results Finally, Chapter five offers a discussion of the research findings, practical recommendations, and suggestions for future research directions.
LITERATURE REVIEW
Key concepts of multidimensional poverty
Poverty is commonly defined by two key organizations: the United Nations and the World Bank The United Nations assesses poverty through multiple factors that impact human well-being, emphasizing a comprehensive understanding of the issue.
Poverty fundamentally denies individuals choices and opportunities, violating their human dignity and preventing meaningful participation in society It manifests as insufficient resources to feed and clothe families, lack of access to education and healthcare, and absence of land for food production or job opportunities This state of deprivation leads to insecurity, powerlessness, and social exclusion, making individuals and communities vulnerable to violence and often forcing them to live in marginal, fragile environments without access to clean water or sanitation.
Similarly, another multi-dimensional poverty definition which is widely used is the one from World Bank:
Poverty is a multifaceted issue characterized by significant deprivation in well-being, including low income and the inability to access essential goods and services for a dignified life It encompasses various dimensions such as inadequate health and education, limited access to clean water and sanitation, insufficient physical security, lack of representation, and a lack of opportunities to improve one’s circumstances.
Besides, in the paper in 2009, Alkire and Foster defined multidimensional poverty as:
A multidimensional poor person is defined as someone who experiences both economic and social deprivation Economic deprivation occurs when an individual's income falls below a specified cutoff, indicating a lack of resources for consumption and wealth Social deprivation is identified when a person's achievements in key areas such as health, education, and social protection do not meet established thresholds This framework emphasizes that each social dimension is vital, and failing to reach the cutoff in any area signifies a violation of fundamental human rights (Alkire and Foster, 2009).
Poverty can be categorized into two main types: absolute poverty and relative poverty Absolute poverty occurs when an individual lacks the essential resources needed to fulfill basic needs such as food, shelter, and clothing In contrast, relative poverty describes the condition of individuals or groups living below the average societal standard, impacting their quality of life compared to others in their community.
The poverty lines, or poverty thresholds, are the cut-off points where individuals below are regarded as the poor There are two main kinds of poverty lines:
Relative poverty line is set by putting one income in relation to the rest of population This establishes poverty lines in terms of a percentage mean or median of income
The absolute poverty line is defined by the minimum amount of money required for survival, with two primary methods for its establishment: the food energy intake method and the cost of basic needs method The food energy intake method, as discussed by Gree and Thorbecke (1986) and Paul (1989), determines the poverty line based on the consumption level needed to meet essential nutrient requirements In contrast, the cost of basic needs method, introduced by Ravallion (1994), calculates the poverty line by assessing the cost of a typical basket of goods consumed by individuals, which includes a specific allowance for non-food items This comprehensive approach makes the cost of basic needs method more widely used than the food energy intake method.
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 poverty line, calculated by the General Statistic Office (GSO) with support from the World Bank, is based on international standards and was established for two periods: 1993-2008 and 2010 onwards During the first period, the poverty line was determined by the monetary equivalent needed to acquire food for a daily intake of 2,100 calories per person, along with the cost of essential non-food items such as shelter and clothing Individuals whose expenditures fell below this minimum threshold were classified as poor In the second period, the poverty line was adjusted to reflect the cost of basic food for a daily intake of 2,230 calories per person, in addition to necessary non-food expenses Throughout both periods, a single poverty line was applied uniformly to both rural and urban areas, maintaining its value over time in real purchasing power.
Table 2.1: GSO – WB poverty lines
The second approach, developed by the Ministry of Labor, Invalid and Social Affairs (MOLISA), focuses on the economic growth rate and financial resources allocated for poverty reduction Welfare indicators are represented by per capita income, adjusted for regional differences, quantified by the amount of rice allocated per person per day: 15 kg for mountainous countryside and islands, 20 kg for delta countryside and midlands, and 25 kg for urban areas The poverty line was established in the first year of a five-year period, as illustrated in Table 2.2.
(*) There has been no separated poverty line for rural areas and mountain areas since 2006 The poverty line of mountain areas falls under rural areas.
Economic theories of poverty
Economic theories of poverty vary, yet they converge on the idea that poverty represents a state where individuals cannot meet their essential needs for survival These theories can be broadly categorized into two groups: those emphasizing individual behaviors and those highlighting the influence of social structures (Sherraden, 1991, p 35).
Individualistic theories of poverty, as proposed by Schultz (1963) and Becker (1964), suggest that poverty stems from inherent individual characteristics and personal abilities These theories argue that individuals are rational agents who possess the autonomy to act in their own self-interest Consequently, the roots of poverty are found in the choices and decisions individuals make Those who struggle to make effective choices or compete successfully with others are more likely to experience poverty.
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
Poverty is not rooted in the inherent traits of individuals but rather in the structural elements of society (Burton, 1992, p.149) External factors significantly influence individual behavior, leading to disparities in opportunities Those with higher socioeconomic status enjoy greater choices, while those with lower status often lack the means to make beneficial decisions This systemic inequality perpetuates a cycle where the wealthy continue to accumulate wealth, and the poor remain trapped in poverty, regardless of their efforts or values Ultimately, it is the societal structure that fails to provide equitable opportunities, resulting in persistent poverty.
Approaches to poverty measurement
There are two main approaches on the theoretical ground: the monetary approach and the non monetary approach
The monetary approach, also known as the uni-dimensional approach, was developed by Rowntree in the early 20th century In his 1899 survey of poor families in York, England, he established a poverty line based on a minimum weekly income necessary for securing essential needs for a healthy life, which he termed "bare subsistence." This amount encompassed costs for light, fuel, food, clothing, rent, and personal items Rowntree revisited his analysis in 1936, adopting a "relative" concept of the poverty line that focused on the minimum requirements for a healthy life He aimed to assess changes in poverty levels in York over a span of more than 30 years, adjusting for price fluctuations.
This method is popular in economics due to its straightforward numerical calculations and policy applications However, it faces several practical challenges, such as reliance on recall data, difficulties in monetizing products, and seasonal variations in information Most importantly, it fails to adequately address the various dimensions of household living conditions.
Unlike the monetary approach that connects poverty to financial expenditure, the non-monetary approach views poverty through the lens of various deprivations Throughout its extensive history, numerous indicators have been utilized to assess poverty, but this paper focuses on three specific indicators.
The Household Asset Index developed by Filmer and Pritchett (1999, 2001) utilizes asset ownership and housing characteristics to assess household well-being Their research, based on data from India, employed principal component analysis to create an asset index that examines the influence of household wealth on educational enrollment across various states They demonstrated that classifying households using asset indicators closely aligns with expenditure-based classifications while offering enhanced accuracy and validity Unlike traditional money-metric measures, the asset index relies on data gathered through interviews and inspection checklists, leading to superior accuracy and validity compared to income or expenditure-based assessments.
Utilizing assets as indicators for household per capita expenditures is an effective method to maximize information extraction while minimizing measurement errors Economic evidence supports that the asset index serves as a reliable proxy for economic status, particularly in predicting enrollment rates, and is at least as dependable as traditional consumption measurements (Filmer and Prichette, 2001).
In 2008, Booysen et al analyzed poverty trends in seven African countries from 1986 to 1992, utilizing an asset index derived from demographic and health survey (DHS) data Unlike the Filmer and Pritchett study, they employed multiple correspondence analysis (MCA) for categorical variables instead of principal component analysis (PCA) to assign weights to asset indicators For cross-country comparisons, Booysen et al adopted two conventional World Bank poverty lines—the 40th and 60th percentiles—alongside one absolute poverty line, which is based on a weighted sum of assets indicative of subsistence conditions These poverty lines were constructed from aggregate data to ensure comparability, and the authors noted that the asset index is less volatile than income or expenditure, exhibiting slower changes over time.
The Human Poverty Index (HPI), established by the United Nations in 1997, serves as a significant poverty indicator that emphasizes deprivation in three fundamental dimensions: survival, knowledge, and standard of living It is calculated separately for developing countries (HPI-1) and developed countries (HPI-2) to account for varying socio-economic conditions While the HPI addresses a broader range of poverty attributes beyond mere income, it still falls short of encompassing all the diverse aspects of deprivation that individuals may experience.
The Multidimensional Poverty Index (MPI), developed by Alkire and Foster, serves as a global measure of acute poverty, replacing the Human Poverty Index (HPI) in recent years Unlike traditional income poverty assessments, the MPI encompasses a broader range of factors that influence household living conditions, including material standards of living Alkire and Foster (2010) identified two key components of the MPI: the incidence of poverty (H), which represents the percentage of individuals classified as poor, and the intensity of poverty (A), reflecting the average percentage of deprivations experienced by the poor The MPI evaluates poverty across three critical dimensions—health, education, and standard of living—each represented by ten equally weighted indicators.
Figure 2.1: Dimensions and indicators of the multidimensional poverty index
According to Alkire and Foster, there are two primary methods for identifying poverty The "Union" approach classifies individuals as poor if they experience deprivation in at least one dimension of well-being.
“intersection” method which states that someone is poor if he or she deprives in all dimensions.
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
In his 2006 study, Prakongsai utilized data from the Socio-economic Survey (SES) and the Health and Welfare Survey (HWS) to create an asset index for measuring poverty in Thailand during 1998, 2000, and 2002 He analyzed 28 to 30 variables related to household assets and living conditions, such as washing machines, telephones, and sanitation facilities Employing Principal Component Analysis (PCA) as introduced by Filmer and Pritchett in 2001, he derived a household asset index for each household His findings revealed a weak correlation of 0.23 and 0.28 between the asset index and household income/expenditure in 2002, despite 9.8% of households in the first decile being classified as poor under the national poverty line Conversely, the Pearson Correlation indicated a stronger relationship, exceeding 50%, between households classified by the asset index and those categorized by expenditure/income.
Maria Emma Santos and Karma Ura utilized the Alkire and Foster methodology to analyze multidimensional poverty in Bhutan, drawing on data from the 2007 Bhutan Living Standard Survey Their research identified five key dimensions of poverty in both urban and rural settings: income, education, room availability, access to electricity, and access to safe drinking water Additionally, rural poverty was uniquely characterized by two more dimensions: access to roads and land ownership The study employed two weighting structures—one with equal weights and another based on insights from the Gross National Happiness Survey (GNHS) The findings revealed that poverty is primarily a rural issue, with equal weights indicating that deprivation in electricity, room availability, education, and income significantly contributes to overall poverty Conversely, when using GNHS-derived weights, income and education emerged as more influential factors in overall poverty compared to electricity and room availability.
Asselin and Vu (2009) utilized the multidimensional poverty index to analyze dynamic poverty in Vietnam from 1993 to 2002, employing data from the Vietnam Living Standard Survey (VLSS) Their study identified five key dimensions of poverty: education, water/sanitation, health, employment, and housing, using eight specific indicators such as underemployment and chronic sickness A composite indicator of poverty (CIP) was created for each household, based on weights derived from multiple correspondence analysis (MCA) The absolute poverty line was defined as the average of poverty thresholds linked to primary indicators, allowing for the construction of MPI indices By maintaining consistent weights and the CIP absolute poverty line, the study compared multidimensional poverty to consumption poverty and analyzed poverty trends from 1992 to 2002 The results indicated a significant decline in poverty between 1993 and 1998, aligning closely with consumption poverty trends, with the Red River Delta experiencing the most substantial reduction, while the Mekong River Delta showed the least improvement.
In 2010, Cuong Nguyen developed a multidimensional poverty index to assess poverty levels in five major cities of Vietnam: Hanoi, Haiphong, Danang, Hochiminh, and Cantho, utilizing data from the Vietnamese household standard living survey of 2008 The research focused on five key dimensions—education, health, living standards, economic well-being, and employment—measured through fifteen specific indicators Equal weights were assigned both within and across these dimensions The findings revealed a notably high level of multidimensional poverty in all five cities, with Hochiminh city experiencing the most severe deprivation, particularly in areas such as underemployment, housing conditions, working hours, and living space.
Figure 2.2: Dimensions and indicators of the multidimensional poverty index
In 2012, Be Lam developed a multidimensional poverty index to assess poverty levels in the Mekong River Delta, utilizing data from the Vietnamese Household Standard of Living Survey conducted in 2008 The research focused on ten indicators across three dimensions: health, education, and living standards, employing an equal weighting system for the poverty estimation The findings revealed that the Mekong River Delta was the poorest region among eight, including the Red River Delta and other areas, despite ranking third in monetary poverty.
Eastern and Red River Delta His study also concluded that poverty is much severe in multidimensional sense than monetary poverty
Khai Tran and Danh Nguyen (2012) explored the interrelations between monetary poverty and socioeconomic factors in rural Vietnam without constructing a specific poverty index Utilizing the VHLSS 2008 dataset, they identified 23 socioeconomic indicators categorized into human, natural, physical, and financial assets Through principal component analysis and multiple correspondence analysis, they narrowed these down to 16 key indicators for further investigation Their findings highlighted significant distinctions between monetary and multidimensional poverty, revealing that per capita expenditure has the least impact on multidimensional poverty Additionally, they found varying contributions of each dimension and indicator to overall poverty, with physical assets playing a crucial role The study emphasizes the importance of assigning appropriate weights to each dimension, leaving the method of determination an open question.
Figure 2.3: Ten dimensions with 16 indicators representative for four livelihood assets
(by Khai Tran and Danh Nguyen, 2012)
Livelihood asset Dimension Relevant indicators
(1) Human asset Human resource for agriculture
Total number of household labours
Average day of treatment of a household member
Number of household member working non-farm
Number of household member working for others
(2) Natural asset Land resource Total agricultural cultivated area
Livelihood asset Dimension Relevant indicators
(3) Physical asset Housing condition House value; Housing area; Type of house
Type of toilet; Water source; Electric source
Ordinary consumption goods Motorbike; Mobile phone
(4) Financial asset Additional income Remittance received within year
Chapter summary
The table below outlines the commonly utilized dimensions, indicators, indexes, and methodologies for exploring various aspects of poverty Despite differing approaches, all these studies emphasize the essential subsistence needs of human beings.
Table 2.3: Summary of empirical studies relating to multidimensional poverty
Indicators Methodology Data set Results
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
Correlation between the poor classified by asset index and household income/expenditure in
- Pearson Correlation: household classified by asset index is correlated with household classified by expenditure/income at over 50%
Indicators Methodology Data set Results
The study examines five key dimensions in both urban and rural areas: income, education, room availability, access to electricity, and access to safe drinking water Additionally, in rural settings, two more dimensions are considered: access to roads and land ownership.
2007 Bhutan Living Standard Survey data
- Poverty is fundamentally a rural phenomenon
In rural areas, overall deprivation is primarily influenced by electricity access, room availability, education, and income when equal weights are applied However, when weights derived from the GNHS are utilized, income and education contribute more significantly to overall poverty, overshadowing electricity access and room availability due to their higher assigned weights.
Vu, 2009 five dimensions are presented, including education, water/sanitation , health,
- Poverty has declined significantly during the 1993-1998 period for both multidimensional and consumption poverty concepts
Indicators Methodology Data set Results employment and housing
- Red River Delta obtained most dramatic reduction, the Mekong River Delta shows the lowest improvement
2010 five main dimensions including education, health, living standard, economic well- being, and employment labor
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
2012 ten dimensions representative for four livelihood assets: human asset, physical asset and financial asset multivariate analysis methods as Principle Component Analysis, Multiple Corresponden ce Analysis and Cluster Analysis
- There must be remarkable differences between monetary poverty and multidimensional poverty and expenditure per capital has lowest impact on poverty in multidimensional perspective
Indicators Methodology Data set Results indicator has different contribution level to overall poverty
- physical asset has impressive contribution to multidimensional poverty
- weight assigned to each dimension is very important
Health, Education and Living standard
- 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 the abundance of multidimensional poverty index (MPI) studies in developing countries, empirical evidence for Vietnam remains scarce, with most research focusing on urban areas or the country as a whole, rather than rural regions Notably, existing studies in Vietnam, except for Be Lam, have overlooked the rural context, and even Be Lam's study omitted a crucial aspect - land possession, a vital asset for rural households This research aims to address this knowledge gap by incorporating land possession into the analysis, providing a more comprehensive understanding of multidimensional poverty in rural Vietnam.
Most existing studies on MPI overlook its decomposable property This research aims to enhance the understanding of MPI by leveraging this characteristic to provide insights for resource distribution.
The literature review highlights poverty as a complex, multidimensional issue that impacts various aspects of human life, encompassing both economic and non-economic factors Key indicators identified include education, health, living standards, and economic well-being A household is considered multidimensionally deprived if it experiences deprivation in more than one of these critical dimensions.
METHODOLOGY
An overview of poverty in Vietnam and Mekong River Delta
Vietnam has made significant strides in poverty reduction since the 1990s, with the poverty rate dropping from 58.1% in 1990 to 14.5% in 2008, according to the General Statistical Office The country has been recognized for its effective poverty alleviation policies aligned with World Bank criteria However, despite this progress, poverty continues to be a pressing issue for the Vietnamese Government, as it has widespread and lasting impacts on society.
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.
Data
This research examines the Multidimensional Poverty Index (MPI) by utilizing households as the primary unit of analysis The study relies on secondary data sourced from the Vietnam Household Living Standards Survey (VHLSS) 2010, conducted by the Vietnamese General Statistics Office A total of 1,455 households from the Mekong River Delta were surveyed for this analysis, with detailed data extraction methods outlined in Appendices 1 and 2.
Table 3.1: Households surveyed in each province
Province Number of households Percentage
Methodology
This section outlines the multidimensional poverty index (MPI) developed by Alkire & Foster (AF) as a method for measuring poverty To provide context, a brief overview of commonly used poverty indicators will also be presented for comparison.
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
The headcount ratio is the most widely utilized indicator for measuring poverty, reflecting the incidence of poverty by calculating the proportion of the income-poor population This ratio is determined by dividing the number of individuals classified as poor, defined as those living below the poverty line, by the total sample population The formula for calculating the ratio of people living below the poverty line is n/z = q/y, where 'n' represents the number of poor individuals, 'z' indicates the poverty line, 'q' is the total sample population, and 'y' is the unit of analysis.
H: Headcount ratio N: Total sample population q: Number of poor people
The indicator function I (yi ≤ z) is defined to equal 1 when the income (yi) is less than or equal to the poverty line (z), and 0 otherwise This means that if an individual's income falls below the poverty threshold, the function I will return a value of 1, indicating poverty; conversely, it will return 0 for incomes above this threshold.
The headcount ratio serves as a basic indicator of poverty by counting the number of individuals living in poverty, but it fails to account for the severity of their situation In contrast, the poverty gap index offers a more comprehensive measure by incorporating both the number of impoverished individuals and the depth of their poverty This index is calculated by either dividing the normalized income shortfall of the poor by the total population or by multiplying the headcount ratio by the average income shortfall, providing a clearer picture of poverty's impact.
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
The AF methodology consists of two key steps: first, it identifies the criteria for defining poverty, and second, it aggregates data from all identified poor households to create a comprehensive poverty indicator, as outlined by Sen in 1976.
The AF methodology defines poverty by employing a dual cutoff approach: the 'within' dimension cutoff assesses whether a household is deprived in specific dimensions, while the 'across dimension' cutoff determines if a household experiences multidimensional poverty.
In the context of multidimensional poverty, let f(ci, k) represent the identification function, where ci indicates the total weighted deprivation experienced by household i, and k signifies the cross-dimension cutoff A household is classified as multidimensional poor when f(ci, k) equals 1, which occurs if the deprivation count ci is greater than or equal to k; conversely, if ci is less than k, the household is not considered multidimensional poor.
= 0 if ci < k Therefore, a household is defined as multidimensional poverty if it deprived in at least k weighted dimensions
Given the above identification function, the total number of households which are poor in a multidimensional sense is defined as: q f(c i ,k) (1)
The headcount ratio or percentage of people who are MPI poor:
To assess the extent of deprivation experienced by a household, let d represent the number of weighted dimensions, and c(k) denote the vector of weighted deprivation counts In this context, ci(k) equals ci if ci is greater than or equal to k; otherwise, ci(k) is zero Consequently, the average intensity of Multidimensional Poverty Index (MPI) poverty among the impoverished is computed.
In which ci is the sum of weighted dimensions in which a household i is deprived
MPI can be decomposed into the contribution of sub-groups of population (Alkire & Foster,
In 2011, let A and B represent two data matrices reflecting the achievements of distinct sub-groups, with (A, B) forming the matrix for the overall population's achievements Denote n(A) as the number of households in sub-group 1 and n(B) as the number of households in sub-group 2.
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 Multidimensional Poverty Index (MPI) highlights that deprivation is not one-dimensional; individuals may experience multiple forms of deprivation across various aspects of life, referred to as dimensions Multi-deprivation is assessed through the aggregation of these dimensions, each measured by several indicators It is crucial to clearly identify each dimension and its corresponding indicators, as they significantly influence the MPI This study selects dimensions, indicators, and deprivation cut-offs based on a comprehensive literature review, previous research, public consensus, and available data, which are detailed in the following sections.
Education significantly enhances the quality of human life, influencing both personal and professional development It enables individuals to effectively apply knowledge and theories to accomplish various tasks and objectives Recognized globally as a vital tool for poverty alleviation, education is assessed through two key indicators: years of schooling and school enrollment rates.
Years of schooling is a key indicator of educational attainment, reflecting an individual's highest level of completed education and serving as a proxy for literacy and numeracy skills These skills are essential not only for urban populations but also for rural communities, where basic literacy and numeracy are critical for agricultural productivity Enhanced literacy and numeric skills enable farmers to better understand information, engage with government support programs, calculate appropriate input rates for cultivation, and adopt new technologies and seeds to boost productivity Statistical evidence shows that educated farmers are more efficient than their uneducated counterparts Consequently, achieving universal primary education is a priority for both the poorest and wealthiest nations, aligning with the United Nations Millennium Development Goals This study identifies primary education completion as a benchmark for assessing educational deprivation, following the criteria proposed by Alkire and Foster A household is deemed deprived in education if no member has completed five years of schooling, while it is considered non-deprived if at least one member has reached this level This approach aligns with Basu and Foster's concept of "effective literacy," highlighting the benefits that a literate household member brings to all members, regardless of their individual educational levels.
School enrolment serves as a crucial indicator for assessing educational deprivation alongside years of schooling, highlighting its significance in helping children navigate various challenges This issue affects both urban and rural areas, making it essential to address children's school enrolment A household is deemed deprived if at least one child aged 6 to 15 is not attending school, which can negatively impact the household's current and future educational levels Consequently, this leads to a decline in knowledge and capabilities within the family Notably, school enrolment is more responsive to policy changes compared to years of schooling, underscoring the need for targeted interventions.
Health and health care deprivation
Health is a vital asset that significantly influences our lives, enabling individuals to engage in active services, achieve professional success, and enjoy happiness Those in good health can maximize their performance, enhance their work capacity, and improve employment opportunities, leading to increased earnings and overall life satisfaction Consequently, a long and healthy life is essential for human development and serves as a key dimension in assessing multidimensional poverty To evaluate health deprivation, three critical indicators are considered: affordability of medical treatment, food security, and health insurance.
Health and poverty are closely interconnected, as poverty can lead to detrimental health outcomes through various channels such as malnutrition, lack of awareness, and limited access to quality healthcare A significant factor affecting physical health is the lack of affordable health services; when individuals cannot afford primary medical treatment, they often face more severe illnesses that require costly care Consequently, the affordability of medical treatment is a critical aspect of health, and households with at least one member unable to pay for medical expenses are considered deprived of necessary medical care.
Treatment of non-applicable population and missing data
3.4.1 Treatment of non-applicable population
Some indicators in the Multidimensional Poverty Index (MPI) are not relevant for all households, such as child enrollment, which does not apply to households without school-age children, and affordable medical treatment, which is irrelevant for households without patients Despite these limitations, we include these indicators in the MPI to enhance measurement accuracy To address this, a household is considered non-deprived in any indicator that does not pertain to its applicable population.
Missing values are a common issue in survey data When all members of a household have missing information for a specific indicator, that household is excluded from the analysis However, if only some members of a household have missing data, we utilize the available information from the other members for reference.
In assessing educational deprivation within households, we classify a household as non-deprived if at least one member has completed five years of schooling, regardless of the educational status of other members Conversely, if more than two-thirds of members have not completed five years of schooling and data for the remaining members is missing, the household is deemed deprived in education If fewer than two-thirds have not completed five years and over one-third of members have missing schooling data, the household is assigned a missing value for this educational indicator.
In assessing child enrolment indicators, if enrolment information is absent for all school-age children in a household, it is classified as missing data Conversely, if only some school-age children lack enrolment information, the household is deemed non-deprived provided that all observed children are enrolled in school; otherwise, the household is classified as deprived.
Analytical framework
This article explores four key dimensions of well-being: wealth, education, health, and living standards, which are assessed through 12 specific indicators These indicators include per capita consumption, cultivated land ownership, years of schooling, school attendance, food security, health insurance, medical affordability, type of dwelling, access to electricity, access to clean drinking water, sanitation, and asset ownership.
Chapter summary
The Mekong River Delta (MRD) is identified as the most impoverished region in Vietnam, highlighting the need for in-depth research This chapter outlines the data and methodology employed, utilizing a representative sample of 1,455 households from the MRD.
The study on education deprivation utilized data from the VHLSS 2010 and employed the Multidimensional Poverty Index (MPI) developed by Alkire and Foster Based on a comprehensive literature review, prior research, public consensus, and available data, four dimensions with twelve indicators were selected An equal weighting system was implemented for the estimation process.
RESULTS
Indicator deprivation
Figure 4.1 illustrates the comparison of deprivation indicators between rural households in the Mekong River Delta (MRD) and those in rural Vietnam overall The data reveals that rural households in the MRD experience equal or greater deprivation across most indicators compared to the national average, with the exception of three indicators related to health and consumption.
Appendix 3 shows that the biggest difference is house quality In rural of MRD, percentage of households live in temporarary or deficient houses is 45 percent, 26 percent point greater than in rural of Vietnam The second biggest difference is sanitation Percentage of households do not have standard toilet or have no toilet is 66 percent, 21 percent point higher than in rural of Vietnam This is followed by years of schooling and asset ownership which level of deprivivation in MRD is around 4 percent higher than in rural of Vietnam
Figure 4.1: Comparison percentage of households deprived in each indicator
(Mekong River Delta vs Vietnam)
Source: Calculated from sub-sample set VHLSS 2010 (n= 1455)
Over 60 percent of households in the Mekong River Delta are deprived of adequate sanitation, highlighting a significant neglect of this issue in rural areas Many households lack standard toilets, often using facilities that discharge directly into water sources, thereby polluting underground water and posing health risks to nearby residents Additionally, 40 percent of households do not have access to clean drinking water, relying instead on unprotected dug wells, rivers, and rainwater Asset ownership and housing quality are also concerning, with 45 percent of households deprived of adequate housing and 40 percent lacking sufficient assets Despite a reliance on agriculture, nearly 30 percent of the population is deprived of cultivated land Using the World Bank's poverty line of 653,000 VND per capita per day, 22 percent of households are considered deprived, compared to 15.9 percent using the national poverty line of 400,000 VND On a positive note, access to electricity is widespread, with only 1.3 percent of households lacking this essential service, and only a few households are deprived of children's school enrollment.
Figure 4.2: Proportion of households deprived in each indicators
Source: Calculated from sub-sample set VHLSS 2010 (n= 1455)
Figure 4.3 shows the proportion of households suffer from deprivation on a number of indicators
A small number of households show no signs of deprivation across any indicators, while the majority experience deprivation in one to nine areas Notably, over half of the households are affected by at least three indicators of deprivation.
Figure 4.3: Proportion of household deprived in various numbers of indicators
Source: Calculated from sub-sample set VHLSS 2010 (n= 1455)
4.2 “Across dimension” cut-off and MPI estimation
The selection of the "cut-off" value, or the number of indicators indicating a household's multidimensional poverty, is a crucial step in constructing the Multidimensional Poverty Index (MPI) There is no consensus on the optimal cut-off value, as it largely depends on the researchers' or policymakers' subjective perspectives, the social context, and the study's objectives A low cut-off value may result in a high number of households being classified as multidimensionally poor, potentially leading to confusion for policymakers Conversely, a high cut-off value could identify fewer households as multidimensionally poor, which may provide misleading insights for decision-making.
Table 4.1: Multidimensional poverty estimate on various cut off point
Adjusted Head Count Ratio (Mo=H*A)
Source: Calculated from sub-sample set VHLSS 2010 (n= 1455)
According to the findings in Table 4.1, a poverty cut-off of k = 0.3 is established, indicating that households are considered multidimensionally poor if they are deprived in at least 30 percent of the weighted sum of indicators This threshold identifies 30 percent of households as multidimensional poor, with these households experiencing an average deprivation of 45 percent across the indicators The potential deprivation of a poor household, relative to all possible deprivations, accounts for 13 percent The data reveals a negative correlation between cut-off points and both the headcount ratio and adjusted headcount ratio; as the cut-off point decreases, both ratios increase, and vice versa Conversely, average poverty exhibits a positive correlation with higher cut-off points, indicating that lower cut-off points lead to a reduction in the average deprivation experienced by poor households.
Poverty estimates at provincial level
Table 4.2 illustrates the estimation of provincial poverty based on a 30 percent deprivation threshold across various indicators, ranked by headcount ratio from highest to lowest poverty levels Soc Trang has the highest poverty incidence, with 43 percent of households classified as multidimensionally poor, followed by Tra Vinh at 39 percent Ca Mau and Kien Giang each report a 34 percent incidence, while Dong Thap and Hau Giang maintain a significant poverty level with a headcount ratio of 33 percent.
Long An province has the lowest poverty rate in the region, with only 16% of households classified as multidimensional poor, likely due to its proximity to Ho Chi Minh City Vinh Long and Tien Giang follow closely, with 21% and 22% of households, respectively, facing significant deprivation Ben Tre ranks fourth with 25% of households experiencing poverty, while Bac Lieu also shows low poverty levels The trend indicates that these provinces, being near Ho Chi Minh City, benefit from better access to education, healthcare services, and employment opportunities, contributing to their lower poverty rates.
The rankings for the adjusted headcount ratio, which reflects the depth of poverty, closely mirror those of the headcount ratio Soc Trang leads with an adjusted headcount ratio of 21%, followed by Tra Vinh at 18% Hau Giang ranks third in poverty depth, while Ca Mau and An Giang occupy the fourth and fifth positions, respectively Conversely, Long An has the lowest depth of poverty, succeeded by Vinh Long, Tien Giang, Bac Lieu, and Ben Tre.
Soc Trang has the highest level of deprivation among poor households, with each multidimensional poor household experiencing, on average, 50 percent deprivation based on the weighted sum of indicators In comparison, Hau Giang and An Giang have a lower incidence of poverty at 33% but still record the second highest level of deprivation, with multidimensional poor households facing an average deprivation of 47 percent Tra Vinh and Ca Mau follow closely, with average deprivations of 46 percent and 45 percent, respectively The lowest levels of deprivation are found in Bac Lieu, followed by Can Tho and Long An.
Table 4.2: Poverty estimates at provincial level
Source: Calculated from sub-sample set VHLSS 2010 (n= 1455)
The maps below will represent the incident, depth and breadth of poverty in each province
Map 2: ADJUSTED HEADCOUNT RATIO (Mo)
Figure 4.4 illustrates the contribution of each dimension to the Multidimensional Poverty Index (MPI) across various provinces The contribution of each dimension is calculated by summing the contributions of all indicators within that dimension Among the four dimensions, "wealth" deprivation consistently accounts for the largest share in most provinces Notably, Long An, identified as the least poor province, exhibits the highest level of wealth deprivation Additionally, deprivation in living standards ranks as the second most significant contributor in most provinces.
Figure 4.4: Contribution by province of each dimension to MPI
Source: Calculated from sub-sample set VHLSS 2010 (n= 1455)
Which indicator contributes the most to MPI?
Figure 4.5 illustrates the contributions of various indicators to overall deprivation, highlighting that cultivated land, per capita expenditure (PCE), years of schooling, and sanitation are the top four factors Notably, land possession stands out as the largest contributor, accounting for 17 percent of the Multidimensional Poverty Index (MPI) This significant contribution is likely due to the unequal distribution of cultivated land at both national and provincial levels These findings align with expert predictions discussed during the seminar "Using Land Resources in Vietnam with Urban and Rural Settlements," organized by the Union of Vietnam Science and Technology and the Institute of Settlements Research.
The contribution of four key indicators to multidimensional poverty in Vietnam is calculated by taking the product of the number of people who are both multidimensionally poor and deprived in each indicator, multiplied by the indicator's weight, and then dividing by the product of the total number of indicators and the overall development of institutions and infrastructure Vietnam's limited cultivated land contributes to a high incidence of multidimensional poverty, particularly in households reliant on agriculture The second largest contributor to the Multidimensional Poverty Index (MPI) is per capita expenditure, reflecting the country's overall low consumption levels Years of schooling rank as the third biggest contributor, although rising child school enrollment indicates improvements in educational access Sanitation emerges as the fourth largest contributor, while other indicators such as insurance, housing quality, and asset ownership also significantly impact deprivation among multidimensionally poor households.
Figure 4.5: Contribution to MPI by indicators at aggregate level
Source: Calculated from sub-sample set VHLSS 2010 (n= 1455)
Decomposition of adjusted headcount ratio and Policy implications
From a policy standpoint, the decomposable nature of Mo makes it a valuable tool for resource allocation across provinces and various indicators Provinces with higher Mo scores should receive larger budgets, but it's essential to recognize that Mo rankings depend on selected dimensions, indicators, weights, and cutoff values Therefore, effective resource spending must be based on reliable Mo rankings To ensure robustness, we employ dominance analysis of Mo rankings, considering all potential values of k A high correlation coefficient indicates strong consistency in Mo rankings, confirming their reliability across different k values.
Table 4.3 illustrates the Spearman correlation coefficient between the rankings of Mo provinces with varying k values, revealing that the coefficient exceeds 0.8 in most instances, with the exception of a maximum value of 0.7 This indicates that the rankings of Mo provinces remain consistent across different k values for measuring the Multidimensional Poverty Index (MPI), except in the case of the extreme value of 0.7 Consequently, it can be concluded that the estimates for Mo presented in Table 4.2 are robust and can effectively guide budget allocation decisions.
Table 4.3: Spearman correlation between Mo ranks by different k values k = 0.1 k = 0.2 k = 0.3 k = 0.4 k = 0.5 k = 0.6 k = 0.2 0.9725 k = 0.3 0.9396 0.9396 k = 0.4 0.9011 0.9286 0.9451 k = 0.5 0.8242 0.7967 0.8516 0.8407 k = 0.6 0.8791 0.8516 0.8956 0.9176 0.9286 k = 0.7 0.7912 0.7253 0.7692 0.6923 0.7143 0.8022
All correlations are significant at 1 percent
According to the adjusted headcount ratio (Mo) by province, Soc Trang has the highest level of deprivation, making it a top priority for poverty alleviation programs, followed closely by Tra Vinh province, with Hau Giang ranking third.
Decomposition by indicators enables policymakers to identify the percentage contribution of each indicator to overall deprivation in provinces Decision-making will prioritize indicators with higher contributions to overall deprivation, guiding the allocation of resources effectively.
The analysis indicates that the lack of cultivated land resources and low per capita expenditure (PCE) significantly contribute to overall poverty, accounting for approximately 30% of the Multidimensional Poverty Index (MPI) in many provinces Given the strong positive correlation between these two factors, it is essential to address them simultaneously through a comprehensive policy approach This involves minimizing land recall and possession issues, restricting the conversion of agricultural land to non-agricultural uses, and implementing training programs to help farmers adopt new technologies that enhance both the quantity and quality of agricultural products Additionally, collaboration between field and provincial management is crucial for effective policy implementation Research on suitable crops for different land types should be integrated with outreach efforts to ensure that farmers understand government planning, fostering trust and compliance with policies that can boost productivity and income Furthermore, promoting trade to increase export volumes is vital for enhancing farmers' earnings.
In Soc Trang, years of schooling and sanitation are critical factors influencing resource allocation, with education being the third largest contributor to overall development A significant portion of the population, particularly ethnic minorities, faces educational deprivation; nearly 40% of households are Khmer, with 38% lacking sufficient schooling To address this, local authorities must enhance educational opportunities for ethnic minorities and promote awareness of education's role in improving living standards Additionally, investments in building schools near minority communities and attracting qualified teachers through incentives are essential Sanitation is another pressing issue, as over half of households either lack toilets or use inadequate facilities To improve sanitation, authorities should educate the public on environmental protection and hygiene, promote the use of proper sanitation facilities, and provide subsidies or funding for upgrades.
In Tra Vinh province, insurance is a significant contributor to poverty, ranking alongside land and PCE, with sanitation following closely Notably, 43% of households lack any form of insurance, which is double the average rate in the Mekong River Delta This situation contributes to 13% of the overall poverty rate To enhance health insurance coverage, local government must actively promote health insurance policies and legislation within communities, improve the quality of healthcare services, and foster trust in voluntary health insurance participation Additionally, it is crucial to monitor compliance among formal employers regarding compulsory health insurance Similar to Soc Trang, the substantial impact of sanitation on poverty necessitates increased budget allocation for sanitation initiatives in Tra Vinh.
In Hau Giang, the lack of education and inadequate sanitation are significant factors contributing to poverty, with educational deprivation accounting for 10.6% of overall poverty levels Notably, 17% of households lack a member who has completed five years of schooling, slightly above the Mekong River Delta average of 14% To tackle this issue, local authorities must prioritize education funding, focusing on increasing primary school enrollment and student retention rates Additionally, similar to the challenges faced in Soc Trang and Tra Vinh, the high sanitation-related poverty necessitates substantial budget allocation for sanitation improvements To effectively address these concerns, local authorities should enhance public awareness of environmental protection and consider providing direct cash grants or favorable consumer loans to support households.
The provinces of Long An, Vinh Long, and Tien Giang share common factors contributing to overall deprivation, with land ownership and per capita expenditure (PCE) being the primary issues To address these challenges, these provinces must focus on land reform and expanding agricultural economic opportunities In Long An, the significant deprivation in years of schooling and insurance highlights the need for substantial investment in education and health insurance Vinh Long's high levels of food insecurity, along with inadequate water and sanitation, necessitate an increased budget allocation for these essential services Similarly, Tien Giang's poverty can be alleviated by directing more funds towards education and health insurance.
Table 4.4: Decomposing of adjusted headcount ratio by indicator
Education Health Living standards Wealth
Electric- ity Water Sanita- tion Asset Land PCE
Soc Trang 0.024 0.011 0.014 0.014 0.013 0.017 0.005 0.009 0.021 0.014 0.036 0.033 0.212 Breakdown 11.2% 5.1% 6.7% 6.7% 6.3% 8.0% 2.4% 4.4% 9.8% 6.7% 16.8% 15.7% 100% Tra Vinh 0.015 0.009 0.012 0.013 0.023 0.017 0.001 0.004 0.020 0.012 0.021 0.030 0.177 Breakdown 8.5% 5.2% 6.5% 7.4% 13.0% 9.3% 0.8% 2.4% 11.2% 6.8% 11.8% 17.0% 100% Hau Giang 0.017 0.01 0.0065 0.0065 0.0083 0.0129 0.0007 0.0126 0.0165 0.0113 0.0306 0.0278 0.157 Breakdown 10.6% 4.4% 4.1% 4.1% 5.3% 8.2% 0.4% 8.0% 10.5% 7.2% 19.4% 17.7% 100%
An Giang 0.025 0.01 0.0025 0.0019 0.0094 0.0128 0.0031 0.0098 0.0097 0.0114 0.0322 0.0246 0.152 Breakdown 16.2% 6.2% 1.7% 1.2% 6.2% 8.4% 2.1% 6.5% 6.4% 7.5% 21.2% 16.2% 100% Kien Giang 0.013 0.01 0.0078 0.0043 0.0092 0.0140 0.0032 0.0124 0.0167 0.0134 0.0224 0.0267 0.151 Breakdown 8.5% 5.6% 5.2% 2.8% 6.1% 9.2% 2.1% 8.2% 11.0% 8.8% 14.8% 17.6% 100% Dong Thap 0.018 0.01 0.0102 0.0048 0.0084 0.0121 0.0000 0.0121 0.0147 0.0091 0.0290 0.0190 0.145 Breakdown 12.5% 5.0% 7.1% 3.3% 5.8% 8.4% 0.0% 8.4% 10.1% 6.3% 20.0% 13.1% 100% Can Tho 0.018 0.00 0.0052 0.0035 0.0121 0.0094 0.0010 0.0094 0.0125 0.0094 0.0339 0.0130 0.130 Breakdown 14.0% 2.0% 4.0% 2.7% 9.3% 7.2% 0.8% 7.2% 9.6% 7.2% 26.0% 10.0% 100% Ben Tre 0.012 0.00 0.0042 0.0048 0.0120 0.0082 0.0012 0.0118 0.0100 0.0091 0.0199 0.0145 0.110 Breakdown 10.7% 2.5% 3.8% 4.4% 10.9% 7.4% 1.1% 10.7% 9.1% 8.2% 18.1% 13.2% 100% Bac Lieu 0.012 0.01 0.0099 0.0010 0.0059 0.0113 0.0012 0.0000 0.0119 0.0095 0.0119 0.0193 0.100 Breakdown 11.9% 6.0% 9.9% 1.0% 5.9% 11.3% 1.2% 0.0% 11.9% 9.5% 11.9% 19.4% 100% Tien Giang 0.014 0.01 0.0045 0.0040 0.0090 0.0065 0.0000 0.0068 0.0082 0.0071 0.0170 0.0162 0.098 Breakdown 13.9% 5.2% 4.6% 4.0% 9.2% 6.7% 0.0% 6.9% 8.4% 7.3% 17.3% 16.5% 100% Vinh Long 0.009 0.00 0.0095 0.0022 0.0087 0.0054 0.0004 0.0094 0.0093 0.0070 0.0143 0.0154 0.093 Breakdown 9.5% 2.4% 10.2% 2.4% 9.4% 5.8% 0.5% 10.2% 10.1% 7.6% 15.4% 16.6% 100% Long An 0.011 0.00 0.0013 0.0006 0.0071 0.0023 0.0004 0.0023 0.0070 0.0050 0.0145 0.0097 0.066 Breakdown 16.2% 7.4% 2.0% 1.0% 10.8% 3.5% 0.6% 3.5% 10.6% 7.7% 22.1% 14.7% 100%
Source: Calculated from sub-sample set VHLSS 2010 (n= 1455)
Comparison between Consumption poverty and Multidimensional poverty
4.6.1 Income poverty verses Multi-dimesional poverty
Figure 4.6 illustrates the divergence between consumption poverty and multidimensional poverty, with the bar line representing the multidimensional poverty headcount and the zigzag line indicating consumption poverty headcount Notably, consumption poverty is generally lower than multidimensional poverty across most provinces, highlighting significant disparities This difference can be attributed to multidimensional poverty's comprehensive approach, which assesses deprivation in health, education, living standards, and other critical aspects of life Additionally, variations in households' capacity to translate income into tangible outcomes in education, health, or living standards may also play a role Furthermore, the reliance on recall data for measuring consumption can lead to an underestimation of consumption poverty due to potential data gaps or forgotten expenditures Lastly, inadequate healthcare services, educational facilities, and poor living conditions significantly contribute to the widening gap between multidimensional and consumption poverty in the MRD.
Figure 4.6: MPI compare to consumption poverty by provincial level
Education Health Living standards Wealth PCE
Source: Calculated from sub-sample set VHLSS 2010 (n= 1455)
4.6.2 Correspondence of consumption poverty and multidimensional poverty
Table 4.5 illustrates a significant disparity between multidimensional poverty and monetary poverty, with 30 percent of households classified as multidimensionally poor compared to only 22 percent below the monetary poverty line This difference arises because monetary poverty primarily focuses on consumption deprivation, while multidimensional poverty encompasses a broader spectrum of deprivations The data also indicates that 83 percent of households maintain consistent poverty status across both measures However, the table highlights that multidimensional poverty measurement (MPI) is more comprehensive; only 4.9 percent of households identified as monetary poor are classified as non-multidimensionally poor, whereas 12.2 percent of multidimensionally poor households are deemed non-monetarily poor, revealing a discrepancy 2.5 times greater.
Table 4.5: Comparing monetary poor and multidimensional poor at aggregate level
4.6.3 Correlation between Consumption and Multidimensional poverty
The Spearman correlation analysis reveals a correlation coefficient of 0.57, indicating a statistically significant relationship at the 1 percent level However, this moderate correlation suggests that relying solely on consumption as an indicator of poverty may not be sufficient, emphasizing the need for a multi-indicator approach to accurately measure poverty.
Estimating poverty levels based on consumption is commonly justified by its strong correlation with other indicators of deprivation This method implies that measuring poverty through per capita consumption inherently accounts for other forms of deprivation However, this approach may not be suitable in the context of MRD As indicated in Table 4.6, the Spearman coefficients reveal that, aside from water deprivation, the correlations between consumption deprivation and other indicators are statistically significant at the 1 percent level, yet they remain relatively low The highest correlations are with food security, asset ownership, housing quality, and sanitation, with coefficients of 0.29, 0.24, 0.24, and 0.23, respectively These findings underscore the necessity for a multidimensional approach to poverty measurement, as consumption alone fails to capture the diverse deprivations faced by the poor.
Table 4.6: Spearman coefficient between indicators
Medical affordability Insurance Housing quality Electricity Water Sanitation
* Correlation is significant at 10 percent
** Correlation is significant at 5 percent
** * Correlation is significant at 1 percent
Chapter summary
This chapter highlights that rural households in the MRD are generally more deprived compared to their counterparts across Vietnam, with Soc Trang identified as the poorest province and Long An as the least impoverished Key findings reveal that land possession and per capita consumption are the primary contributors to the Multidimensional Poverty Index (MPI) at an aggregated level A provincial analysis, based on the adjusted headcount ratio, provides essential insights for policy implications Furthermore, the comparison between monetary poverty and MPI indicates that consumption poverty is typically lower than multidimensional poverty in most provinces While there is a notable similarity in categorizing households as poor or non-poor based on both measures, the correlation suggests that a comprehensive study of multidimensional poverty is essential, as consumption alone fails to capture the various deprivations faced by the poor.