REGIONAL AND LOCAL DEVELOPMENT STUDIES

DETERMINANT OF URBAN MULTIDIMENSIONAL POVERTY: A

HOUSEHOLD LEVEL ANALYSIS IN THE CASE OF KOLFE
KERANIYO SUB CITY OF ADDIS ABABA CITY

BY

NETSANET AYALEW BELETE

ADDIS ABABA, ETHIOPIA

NOVEMBER, 2021
DETERMINANT OF URBAN MULTIDIMENSIONAL POVERTY: A
HOUSEHOLD LEVEL ANALYSIS IN THE CASE OF KOLFE
KERANIYO SUB CITY OF ADDIS ABABA CITY

By
Netsanet Ayalew Belete

A THESIS SUBMITTED TO COLLEGE OF DEVELOPMENT STUDIES

PRESENTED IN PARTIAL FULFILLMENT OF THE REQUIREMENT FOR
THE DEGREE OF MASTER OF ARTS INREGIONAL AND LOCAL
DEVELOPMENT STUDIES

Advisor
Andualem Goshu, (Ph.D.)

ADDIS ABABA, ETHIOPIA

NOVEMBER, 2021
 Addis Ababa University
School of Graduate Studies

This is to certify that the thesis prepared by Netsanet Ayalew, entitled: Determinant of Urban
Multidimensional Poverty: A House Hold Level Analysis In The Case of Kolfe Keraniyo Sub
City of Addis Ababa and submitted in partial fulfillment of the requirements for the Degree of
Master of Art in Regional And Local Development Studies complies with the regulations of the
university and meets the accepted standards with respect to originality and quality.

Signed by the Examining Committee:

Advisor: _Signature Date

Internal Examiner: Signature Date

External Examiner: Signature Date

I
 DECLARATION

I, the undersigned, declare that the thesis is my original work, has not been presented for degrees
in any other University and all sources of materials used for the thesis have been duly
acknowledged.

Name: Netsanet Ayalew

Signature: _

Date

Advisor: Andualem Goshu, (Ph.D.)

Signature: _

Date

Place of Submission: Collage of Development Studies, Addis Ababa University

Date of Submission: November, 2021

II
 ACKNOWLEDGMENTS

Dr. Andualem Goshu, my adviser, deserves special thanks for his unwavering support and
encouragement during this journey. My sister Selamawit Ayalew, my grandmother Woynitu
Habtewold, and, most importantly, my mother Sr. Mekdes Mekonnen deserve special mention.
Since the commencement of my studies, they have provided me with unwavering moral and
financial support. Furthermore, I would not have been able to accomplish this voyage without the
help and encouragement of my manager, Ato Mikias Kifremariam. My gratitude also goes to the
Kolfe Keranyio Sub City Administrators and residents for their kind cooperation in providing us
with the necessary information.

i
 Table of Content

ACKNOWLEDGMENTS................................................................................................................I

LIST OF TABLE............................................................................................................................V

ACRONYMS...............................................................................................................................VII

ABSTRACT................................................................................................................................VIII

CHAPTER ONE..............................................................................................................................1

INTRODUCTION...........................................................................................................................1

1.1. Backgrounds of the Study.....................................................................................................1

1.2. Statement of the Problem......................................................................................................2

1.3. Research Questions...............................................................................................................5

1.4. Objectives of the Study.........................................................................................................5

1.5. Significance of the Study.......................................................................................................5

1.6. Scope of the Study.................................................................................................................6

1.7. Organization of the Paper......................................................................................................6

CHAPTER TWO.............................................................................................................................7

LITERATURE REVIEW................................................................................................................7

2.1. Theoretical Literature Review...............................................................................................7

2.1.1. Concept and Definitions..................................................................................................7

2.1.2. Measure of Multidimensional poverty............................................................................9

2.1.3. Theories of Poverty.........................................................................................................9

2.1.4. Poverty in Ethiopia........................................................................................................11

2.1.5. Multi-Dimensional Poverty...........................................................................................11

2.1.6. Global Multi-Dimensional Poverty Index 2020............................................................13

2.2. Empirical Literature Review...............................................................................................17

II
 2.3. Conceptual Framework.......................................................................................................19

CHAPTER THREE.......................................................................................................................20

DATA AND METHODOLOGY...................................................................................................20

3.1. Description of the Study Area.............................................................................................20

3.2. Research Design..................................................................................................................21

3.3. Source of Data.....................................................................................................................21

3.4. Sample Size Determination.................................................................................................21

3.5. Method of Data Analysis.....................................................................................................22

3.5.1. Multidimensional Poverty Index (MPI) Estimation Method........................................22

3.5.2. Descriptive Data Analysis.............................................................................................25

3.5.3. Correlation Analysis......................................................................................................25

3.5.4. Econometric Model.......................................................................................................25

3.5.4.1. Logistics Regression............................................................................................25

3.5.4.2. The Model..............................................................................................................25

3.5.4.3. Goodness of Fit of the Model..............................................................................26

3.5.4.4. Model Diagnostics...............................................................................................26

3.5.5. Study Variables.............................................................................................................27

3.5.5.1. Response Variable................................................................................................27

3.5.5.2. Explanatory Variables..........................................................................................28

3.5.6. Instrumental Reliability and Validly.............................................................................29

3.5.6.1. Instrument reliability............................................................................................29

3.5.6.2. Instrument Validity..............................................................................................30

3.5.6.3. Content Validity...................................................................................................30

3.5.6.4. Internal Validity...................................................................................................30

3.5.6.5. External Validity..................................................................................................31

II
CHAPTER FOUR.........................................................................................................................32

RESULTS AND DISCUSSION....................................................................................................32

4.1 Descriptive Analysis.............................................................................................................32

4.2. Result of Multidimensional poverty analysis......................................................................34

4.3. Results of Spearman Correlation Analyses.........................................................................36

4.4. Econometric Analysis..........................................................................................................37

4.4.1. Parameter Estimation....................................................................................................37

4.4.2. Model Checking Techniques.........................................................................................39

4.4.3. Goodness of Fit of the Model.......................................................................................39

4.4.4. Model Diagnostics........................................................................................................40

4.5. Discussions of the Results...................................................................................................41

CHAPTER FIVE...........................................................................................................................44

CONCLUSIONS AND POLICY IMPLICATION.......................................................................44

5.1. Conclusion...........................................................................................................................44

5.2. Policy Implication...............................................................................................................45

REFERENCES..............................................................................................................................47

ANNEX.........................................................................................................................................55

I
 List of Table

Table 2.1: Global Mpi 2020- Dimensions, Indicators, Deprivation Cutoffs And Weights...........15
Table 3.1: Mpi Dimensions, Indicators And Weight.....................................................................24
Table 3.2: Description And Measurement Types Of Explanatory Variables................................28
Table 3.3 Table Reliability Statistics.............................................................................................29
Table 4.1: Descriptive Analysis Of Categorical Variables............................................................32
Table 4.2 Descriptive Statistics Of Continues Variables...............................................................33
Table 4.3 Number Of Household Deprived In Different Indicators..............................................34
Table 4.4: Mpi Estimation.............................................................................................................36
Table 4.5 Relative Contribution Of Dimensions To Mpi..............................................................36
Table 4.6 Spearman Rank Correlation Analysis Of The Explanatory And The Response Variable...........37
Table 4.7 Maximum Likelihood Estimates Of Model Parameters................................................38
Table 4.8 Hosmer- Lemeshow Goodness Of Fit Test...................................................................40

V
 List of Figure
Figure 2.1: Composition of the Global MPI- Dimensions and indicators.....................................14

Figure 2.2: Conceptual Framework...............................................................................................19

Figure 3.1: Map Of Study Area.....................................................................................................20

Figure 3.2: Multiple Deprivations of Indicators of the Households..............................................35

V
ACRONYMS
CSA Central Statistics Authority
EEA Ethiopian Economics Association
ETB Ethiopian Birr
FAO Food and Agriculture Organization
GPRS Ghana Poverty Reduction Strategy
HDI Human Development Index
HH Household
MoFED Ministry of Finance and Economic Development
MPI Multidimensional Poverty Indicators
NGOs Non-Governmental Organizations
SNNPR South Nation, Nationalities and People
SSA Sub Saharan Africa
USD United States Dollar
WB World Bank
WDR World Development Report

V
 ABSTRACT

Poverty is a problem in all countries, but it is particularly acute in Sub-Saharan Africa. Ethiopia
is working to eradicate poverty in all of its forms as part of the nation's 2030 fundamental goals.
As a result, it is vital to measure urban poverty in order to advance the effort. As a result, the
study's goal was to look into the factors that contribute to urban poverty at the household level in
Addis Ababa in the case of Kolfe Keranio sub city. The data for the study were taken from 398
sampled households residing in Kolfe Keranio. Both descriptive and Econometrics analysis are
employed in the data analysis. The determinants of being multidimensionally poor are
investigated using a logistic regression model. According to the descriptive analysis, 65.32
percent of the households in the sample are multidimensionally poor. The intensity of poverty is
89.0% and the adjusted headcount ratio is found to be 64.57%. The living standard component
(34.6%) contributes the most to the overall multidimensional poverty of the sample households,
followed by the education (15.4%) and health dimensions (3.0 percent). Asset indicators (63.8
percent) and cooking fuel (59.0 percent) had the biggest relative contributions to the total
multidimensional poverty index of the study region among the ten multidimensional poverty
index variables. In addition, the results of logistic regression revealed that the being male
household, being married household head, being employed house head and being obtained loan
are statistically significant determinants of households being multidimensionally poor. Policy
implications that prioritize living standards and education, as well as policy implications that
take key elements into account in poverty reduction initiatives, are essential.

Keywords: Multidimensional Poverty, Logistic Regression Model, Urban poverty

V
 CHAPTER ONE

INTRODUCTION

1.1. Backgrounds of the Study

The Sustainable Development Goals (SDGs) now aim for all people to be free of poverty and
hunger by 2030 (Beegle et al., 2016; SDG Center for Africa (SDGC/A), 2019). Despite the fact
that the world's poor people are concentrated in Sub-Saharan Africa and South Asia (UNDP,
2019), Sub-Saharan Africa has the worst situation, with 41 percent of the population living on
less than $1.25 a day and the lowest Human Development Index (HDI) in the world at 0.537.
(UNDP, 2018; IPC-IG, 2019).

By all accepted metrics of poverty, Ethiopia is one of the poorest countries according to various
study findings (UNDP Ethiopia, 2018;UNDP, 2018). Ethiopia's HDI rating is 0.463, which is
lower than the average of Sub-Saharan African nations (0.537) and even lower than the low
human developed countries (0.50), according to the 2019 UNDP human development report.

Poverty is described in different ways. Historically, the concept of poverty mainly evolve around
income and consumption which remains the core of the concept today (Ludi and Bird, 2007). It
has evolved from the 19th century idea about subsistence needs 'what a person needs to survive.
During mid-20th century, conceptualization of poverty is considered lacking basic needs',
extending the subsistence idea by also including basic facilities and services such as healthcare,
sanitation, and education, to the late 20th century understanding of poverty as relative
deprivation ‗, including income and other resources, as well as social conditions: (Ludi and Bird,
2007).

Even though poverty was defined by one-dimensional measures, such as income. As a result of
Conceptual evolution in understanding poverty, it was realized that only one indicator alone can
Capture the multiple aspects that constitute poverty. The multidimensional poverty encompasses
a number of elements that contribute to poor people's deprivation, including bad health, a lack of
education, an inadequate living standard, and a lack of income, disempowerment, terrible
working conditions, and the fear of violence (Arjun.R, 2013).

1
According to OPHI (2020), Around 1.3 billion people live in multidimensional poverty in 107
developing countries. This is equivalent to 22% of the population of the 107 countries. 98.8% of
the 1.3 billion persons who are multidimensionally poor are deficient in at least three indices at
the same time. In Sub-Saharan Africa, 55 percent of the population is poor on several levels. In
comparison to other regions, Saharan African countries have the largest percentages of people
who are multidimensionally impoverished and lack schooling years. (Niger, Burkina Faso, South
Sudan, Chad and Ethiopia) and school attendance (South Sudan, Burkina Faso, Niger, Chad and
Mali) (OPHI, 2020).

In Sub-Saharan Africa, 71.9 percent of rural people (466 million people) are multidimensionally
poor, compared to 25.2 percent (92 million people) in urban regions (OPHI, 2020).Ethiopia is
one of the low-income countries in Sub- Saharan Africa where multidimensional poverty
continues to be a serious problem. According to the head count multidimensional poverty metric,
around 83.5 percent of individuals in the country are in multidimensional poverty. The urban and
rural multidimensional poverty is 16.0 percent and 54.7% respectively. The MPI poverty of
Addis Ababa city administration is 5.9% (OPHI, 2020). As a result, this study attempted to
analyze the determinant aspect of multidimensional poverty in the Kolfe Keranio Sub city of
Addis Ababa. Kolfe Keranio was selected purposely because it has the largest and the diverse
households as compared to other sub cities in Addis Ababa CSA (2007).

1.2. Statement of the Problem

Ethiopia is Africa's second-most populated country, with 115 million inhabitants and one of the
region's fastest-growing economies. According to the World Bank (2020), Ethiopia's economy
grew at a solid, broad-based rate of 9.8% per year from 2008/2009 to 2018/2019, with the
proportion of the population living below the national poverty line falling from 38% to 24%
during that time. The International Monetary Fund (2020) notes that Ethiopia‘s per capita income
has risen by about 200% since 1990 while life expectancy increased by about 10 years in a
decade and infant mortality was reduced by half. Despite this impressive economic growth,
Ethiopia remains one of the poorest countries in the world with a per capita annual income that
the government estimates at $883 (MoFEC, 2019). Findings from the most recent Afrobarometer
survey (2020) show that macroeconomic gains do not tell the whole story. With a high

2
prevalence of lived poverty, many Ethiopians continue to experience frequent shortages of
necessities in their daily lives.

According to the report by OPHI (2020), 83.5% percent of the Ethiopian population was poor as
measured by Head count Multidimensional Poverty Indicators (MPI), which means they were
deprived in at least one-third of the weighted MPI indicators. Hence, as per utilizing the MPI
technique, Ethiopia ranked as one of the poorest countries in the world. Despite this fact,
vulnerability of urban households‘ poverty and related risk factors is not properly addressed.
Esubalew (2006) and Mohammed (2017) are two of the few evaluations of poverty in Ethiopia
that have focused on the determinants of urban poverty and they used income as the sole
indicator of poverty, ignoring other aspects of life such as health, education, and living standards.
However, according to OPHI (2020), poor people's experiences of poverty are defined by factors
other than their income. Lack of education, health, shelter, empowerment, humiliation,
employment, personal security, and other factors are frequently mentioned. No single metric,
such as income or consumption, is capable of capturing all of the factors that lead to poverty.

According to (UN-Habitat, 2008), more than 80% of Addis Ababa residents live in a slum, with
70% of them living in government-owned rental housing (UN-Habitat, 2008). According to the
report, the inner city of Addis Ababa is dominated by congested and decaying dwellings, with an
estimated 70-80 percent of the city's housing stock under "slum conditions. Unemployment
among young people (aged 15 to 29) is a big issue; in Addis Ababa, over 25% of young people
(aged 15 to 29) were unemployed in 2020(Ezana H. 2021).Furthermore, Addis Ababa also sees a
lot of immigration, and most of the newcomers face economic hardship and a poor quality of life
(Ezana H.2021). Rapid urbanization and extraordinary population expansion have put immense
strain on Addis Ababa's ability to offer cheap housing and essential services to its citizens,
particularly low-income households, in recent decades (UN-Habitat, 2017).

Another key indicator of a city's ability to improve its citizens' quality of life is the availability of
urban health services. Although many children in Addis Ababa continue to be malnourished,
there has been steady development. Despite all of the advancements, children from low-income
families are more likely to be stunted (UNCIF 2018). It also demonstrates that children whose
mothers have no formal education have a 31% stunting rate compared to 10% for children whose
mothers have finished secondary or higher school (UNCIF 2018)

3
There is a major water deficit in Addis Ababa. Because of the continual disruption and limited
capacity to replenish public water supplies, residents of informal settlements and low-income
areas are disproportionately affected. (Ezana H.2021) Flooding is becoming more likely as a
result of informal house building, inadequate drainage, incorrect solid waste management, and
the loss of green spaces. Despite the fact that the use of electricity as a source of illumination has
remained largely consistent, power supplies for lighting, cooking, and refrigeration remain
insufficient (UN-Habitat, 2017).

Due to the usage of kerosene and open fire for cooking, the congested settlement pattern, and
poor housing quality in the inner-city slums, many portions of the city are also at risk of fire.
(Ezana H.2021.) In Addis Ababa, power disruptions have become a common occurrence. The
amount of homes reporting power outages for 1-2 days, 3-4 days, and 5-7 days per week (UN-
Habitat, 2017). Despite the city's recent economic boom, the current increased unemployment
rate poses a growing threat to Addis Ababa's urban economy. According to the CSA (2015), the
economically active population includes all people who provide labor to produce products and
services. In terms of gender disparities, males (70.0 percent) had a 15.3 percent greater activity
rate than females (54.9 percent). The city of Addis Ababa's activity rate was reported to be 60.8
percent in 2015, lower than the national estimate of 63.7 percent.

Available literatures also indicated that absence of adequate water and sanitation in slum areas
have resulted in the presence of various diseases (Fry et al, 2002; Water aid, 2008). Despite all of
the facts, the city of Addis Ababa's urban households' exposure to poverty and its risk factors
were not adequately addressed. There is only one research attempt regarding multidimensional
poverty of households at Addis Ababa city level studied by Etaferahu C. and Andualem G.
(2019) using Ethiopia Socioeconomic Survey (2013/2014 – (2015/2016) data. This study has its
own limitation because it didn't reflect the multidimestion poverty status of household at sub city
levels as well as the data used for the study does not shows the current multidimensional poverty
status of the city. Therefore, to fill this gap this study attempted to examine the determinant of
urban multidimensional poverty in Addis Ababa at Kolfe Keraniho sub city by collecting the
primary level data from the households. The study place i.e. Kolfe Keraniho was selected for this
study because it has the diverse and the largest household number as compared to other sub cities
in Addis Ababa city CSA (2007).

4
1.3. Research Questions
The research has the following question
i. What is the level of poverty in Kolfe Keranio Sub-city of Addis Ababa?
ii. What proximate factors are associated with multidimensional poverty in the Kolfe
Keranio sub city?
iii. What are determinates of urban multidimensional poverty in Kolfe Keranio Sub
city?

1.4. Objectives of the Study

General Objective
The general objective of this study is to examine the determinates of multi-dimensional urban
poverty in Kolfe Keranio Sub City of Addis Ababa Administration.

Specific Objectives
 To assess the number of households deprived in different poverty indicators
 To measure the multi-dimensional poverty index of the Kolfe Keranio sub city
 To identify factors correlated with urban multidimensional poverty in the sub city
 To Examine the determinants of multidimensional urban poverty of the sub city

1.5. Significance of the Study

Several studies undertaken in Ethiopia to investigate the elements that contribute to

multidimensional poverty have discovered a number of characteristics that contribute to
multidimensional poverty. But very few of these studies gave attention to multidimensional
poverty at urban level. The goal of this study was to determine the most relevant factors
associated with urban multidimensional poverty and to investigate the impact of these factors on
multidimensional poverty among sampled residents in Addis Ababa's Kolfe Keriniho Sub-city.
This finding will be helpful for better understanding of the determinate factor associated with
urban multidimensional poverty in Addis Ababa as general and in Kolfe Keranio sub city in
particular

5
 i. This study will be helpful for the formulation of policies and strategies towards
alleviating urban multidimensional poverty
ii. Finally, the findings and conclusion for this study can be used as the basis for further
research on the area targeting urban poor peoples

1.6. Scope of the Study

Urban multidimensional poverty was a global issue, particularly in developing countries like
Ethiopia. Even though urban poverty is a concern in many cities in Ethiopia, the scope of the
study is confined to the level of Addis Ababa city in the case of Kolfe Keranio Sub city. The
sample was collected from households in 15 of the sub city's woredas (districts). The study also
looked at the socioeconomic and demographic features of the households and the heads of the
families. The study's core assessments focused on identifying the main causes that led to
multidimensional poverty in urban households

1.7. Organization of the Paper

The rest of the paper is organized as follows. Chapter two contains both theoretical and empirical
literature reviews. Chapter three describes the methodological issues of the study and chapter
four gives the results and discussions. Finally, Summary conclusions and recommendations of
the study are presented in chapter five.

6
 CHAPTER TWO

LITERATURE REVIEW

In this chapter, we present a review of both theoretical and empirical literatures on determinant
of multidimensional poverty. Relevant studies were reviewed giving special focus on findings
and methodological issue in developing countries.

2.1. Theoretical Literature Review

2.1.1. Concept and Definitions

The ‗income paradigm of poverty‘ has been traditionally dominant in the literature. Poverty is
conceptualized following the maximizing utility behavior subject to a budgetary restriction. The
individual choices over a set of goods and services are conceived as an expression of the
revealed preferences of the individuals, rational and autonomous with access to a full set of
information. Under this tradition, poverty is defined as a standard of living below a poverty line
fixed using monetary resources (income, consumption, expenditure).

However, since the late 1970s different authors have identified the limitations not only in the
concept of poverty but also of the indicators used. Townsend (1979) was a pioneer arguing that
poverty was not a lack of income to purchase goods in the market but the lack of resources to
participate in society. He also introduced different indicators, mainly focusing on the access to
goods and services, to account for his concept of poverty.

In 1971, Rawls presented his ‗justice as fairness‘ conceptual framework under the social contract
tradition: ―It conveys the idea that the principles of justice are agreed to in an initial situation
that is fair‖ (Rawls, 1971, p. 11). He describes his approach as based on principles of justice,
using the concept of the original position of equity. For Rawls, the space of equality is the
set of
‗primary goods‘ in which he positions: rights and liberties, powers and opportunities, income
and wealth, and health and vigor.

Also at the end of the 1970s Sen proposes the ‗capability approach‘ as a framework to
understand development and specifies that poverty is a ‗capability deprivation‘. He argues that

7
income is not a satisfactory space to evaluate the well-being of individuals. For him, the

8
monetary tradition is a very limited framework to understand development. Income is just one
particular means among many others but not an end in itself: ―Aristotle had pointed out, at the
very beginning of his Nicomachean Ethics, that income and wealth are only instrumentally1
valued, and we have to go deeper to understand what makes human life rich and human
freedoms effective‖ (Sen, 2006, p. 35).

The proliferation of conceptual frameworks since the late 1970s has shifted the focus of
understanding and measuring poverty from income to other dimensions of development. At the
same time research on its measurement is booming. It is a complex subject to navigate and
although several innovations have been proposed, the ‗income paradigm‘ is still the main
framework used to understand and measure poverty. ―Life is richer and more complex
than buying and selling goods and services, but unfortunately, this creates the need for
alternative measures which are not so readily available and are likely to generate more
controversy and more difficulties in measurement‖ (Fleurbaey, 2015, p. 201). Poverty has
different shapes and magnitudes. Poverty can be defined in terms of income, capabilities, basic
needs, social exclusion, and vulnerability to risk shocks. It can be considered permanent or
transitory. It can be defined in relative or absolute terms. It can be understood in objective or
subjective terms. The discourse in the field of poverty studies has agreed, at least, that poverty is
multidimensional, as noted by Jenkins and Mickelwright (2007, p. 7), ―these developments
reflect the view that poverty is not only about having enough money, and that inequality is not
just about differences in money income‖.

Multidimensional poverty encompasses the various deprivations experienced by poor people in

their daily lives – such as poor health, lack of education, inadequate living standards,
disempowerment, poor quality of work, the threat of violence, and living in areas that are
environmentally hazardous, among others (OPHI 2015).

A multidimensional measure of poverty can incorporate a range of indicators that capture the
complexity of this phenomenon in order to inform policies aimed at reducing poverty and
deprivation in a country. Depending on the context of a country and the purpose of the measure,
different indicators can be chosen to reflect the needs and priorities of a nation, as well as its
constituent regions, districts, provinces, etc (OPHI 2015).

9
A multidimensional poverty measure takes into account both the monetary and non-monetary
aspects of poverty. It also places welfare in the space of freedom and accomplishments that
affect human existence. For instance, individuals must be adequately fed, have access to
education, be in good health, participate in community life, be free, appear in public without
shame, etc. The main dimensions that are often taken into account in multidimensional poverty
analysis can then be summarized as follows: income, education, health, water, sanitation,
nutrition, housing, employment, access to productive assets, access to market, etc. (Asselin,
2009).

2.1.2. Measure of Multidimensional poverty

The measurement of poverty has been under constant scrutiny from academics and policy-
makers. For many years, income (or consumption/expenditures) has been used as a proxy to
understand and measure poverty. Nonetheless, over the last decades, increasingly theoretical and
methodological discussions have shifted the attention to what is now called Multidimensional
Poverty. Starting with the seminal works of Peter Townsend (1979), Amartya Sen (1976) and
John Rawls (1971), social scientists have devised different approaches to understanding poverty
without relying on income. Some of these are the basic needs approach, material deprivation,
subjective well-being, and the capability approach.

Since 2010, the Human Development Report (HDR), the publication containing the Human
Development Index, has included a ranking of more than 100 developing countries, in the form
of an index, referred to as the Multidimensional Poverty (MPI). The publication of the HDR not
only sparked a debate about the interpretation of such rankings and comparisons of poverty
between countries but also about the measurement of poverty in general (See Decanq and Lugo
(2010), Ferreira and Lugo (2013), Ravallion (2011) and Alkire (2011)). The ranking published in
the HDR is constructed on the basis of the most widely used counting index of multidimensional
poverty, the Adjusted Headcount Ratio of the Alkire-Foster (AF) family of indices.

2.1.3. Theories of Poverty

There are several compelling theories of poverty which frequently caught the attention of
researchers when a need arises to anchor the causes of poverty on theories.

1
According to Sen (1987) the value of the living standard lies in the living, and not in the
possessing of commodities. Such an approach to the definition and /or measurement of poverty
suggests a broader set of criteria for assessing poverty than just income and/or consumption. The
measure is said to include publicly provided but non-marketed services; like, sanitation, health
care, education & life expectancy. Sen (1987) also introduced the notion of capabilities in
poverty definition and assessments. Hede fined poverty not only as a matter of low level of well-
being, but also as lack of ability to chase well-being specifically because of lack of economic
means. He favored the capability to function as criteria for assessing standard of living, and by
implication poverty rather than the utility that might be derived from using that capability.

Human capital theories of poverty developed by Becker (1975) and Mincer (1974) explains both
individuals ‗decisions to invest in human capital (education and training) and the pattern of
individuals ‗lifetime earnings, and their different levels of investment in education and training
determine the outcome of an individual either to be poor or non-poor. This theory also explains
why the minorities within the society such as women among others have higher incidence of
poverty due to lower earnings from labor market which in turn caused by low investment in
human capital. But, this theory is too shallow to explain the causes of poverty since earnings are
one of the determinants of poverty (Tasew et al, 2013)

The other dominant theory tried to relate the causes of poverty on the basis of geographical
disparities. This theory explains why poverty is most intense in certain areas and why some
regions lack the economic base to compete. More specifically, remoteness, lack of certain types
of natural resource endowments, political disadvantageousness, and weak integration can all
contribute to the creation of intra-country spatial poverty traps Morrill& Wohlenberg(1971)

The frame of reference for both theories extends across national, cultural, ethnic, racial, and
other boundaries. But both cultural and structural theorists have recognized, explicitly or
implicitly, that there are certain societal characteristics necessary to the development of the
poverty syndrome. Such characteristics are high under- and unemployment for unskilled labor,
low wages, little social organization among the poor, a bilateral kinship system, a value system
stressing the individual accumulation of wealth David Elesh (1970). Since poverty is
multifaceted and multidimensional, it is worthy to construct model of poverty determination

1
taking in to account various causes of poverty mentioned in theories and identifies
the significant causes for a specific area David Elesh (1970).

2.1.4. Poverty in Ethiopia

Ethiopia continues to be one of the poorest countries in the world by different standards and
measures of one-dimensional and multidimensional poverty (Apablaza and Yalonetzky,
2013). Even though there have been improvements in living standards, subjective poverty
measures indicate that poverty remains high in Ethiopia. The United Nations‘ HDI
ranked Ethiopia 174 out of 187 countries where average per capita income was less than
half of the sub-Saharan average (The World Bank, 2014).

Similarly, a young lives multidimensional poverty analysis also indicated that Ethiopia‘s
multidimensional poverty index was very high (Alemayehu et al., 2015). Ethiopia is one of
the poorest countries according to multidimensional poverty measures. Despite some
progress, significant multidimensional poverty reduction has not been observed in Ethiopia.
The OPHI (2013) showed that 87 percent of the population was multidimensional poor in
2011 which made Ethiopia the second poorest country in the world.

However, some studies have indicated that since 2000, Ethiopia has shown a reduction
(around 33 percent) in the share of its population living in poverty (Apablaza and
Yalonetzky, 2013; Stifel and Woldehanna, 2017; The World Bank, 2014)

In Ethiopia, the proportion of the population living below the poverty line deceased from48
percent in 1990-91 to around 38.7 percent in 2004-05. A notable reduction in the poverty gap
and the depth of poverty was observed in the country in general and in rural Ethiopia in
particular. a study by (Woldehanna and Hagos, 2013). Stifel and Woldehanna (2016) state that
despite a nominal increase in income in Ethiopia over 2000 and 2011, the poorest urban
population experienced no real change in their consumption levels.

2.1.5. Multi-Dimensional Poverty

There has been a shift of focus from one dimension to multiple dimensions of poverty. The
multidimensional nature of poverty has become increasingly important in recent years and
different contributions to this have been (Alkire et al., 2011). In addition to monetary income

1
or consumption expenditure, human lives and well-being are affected by different dimensions
such as health and education.

A one-dimensional measure of poverty using income or consumption expenditure presupposes

that a market exists for all goods and services; however, often markets do not exist for many
goods and services or they function imperfectly (Bourguignon and Chakravarty, 2003;
Thornback, 2008; Tsui, 2002) and therefore, monetary values cannot be assigned to particular
aspects of well-being (Hulme and McKay, 2008; Thorbecke, 2008). Also, having a sufficient
income for purchasing a basic basket of goods does not imply that it is spent on that basket of
goods (Thorbecke,2008). Individual well-being is a multidimensional notion (Stiglitz et al.,
2009), individual scare about many different aspects of their lives, including their material
standard of living, health, and schooling.

As stated by Alkire and Santos (2011) low income, poor health, inadequate education, job
insecurity, disempowerment, and precarious housing are clear manifestations of
multidimensional poverty. The components of poverty change across people, time, and context
but multiple domains are involved. Empirical literature has documented a mismatch between
monetary and non-monetary deprivations (Berenger and Verdire_Chouchane, 2007; Hishe
Gebreslassie, 2013; Tran et al., 2015). This difference is attributed to a possible bias in the single
dimensional measure of poverty. For example, a study in India by Stewait et al., (2007) found
that 53 percent Indian children living in income-poor households were not malnourished and 53
percent of malnourished children were not living in income poor households.

MPI was developed by the Oxford Poverty and Human Development Initiative (OPHI) at the
University of Oxford (Alkire and Santos, 2011; Alkire et al., 2011). It is a comparable
multidimensional measure of acute poverty in over 100 developing countries. MPI acknowledges
that income or consumption is a necessary but not a sufficient measure of gauging social well-
being. In addition to poverty headcount, the depth, persistence, and complexities of poverty must
also be understood. It considers many deprivations faced by severely disadvantaged groups and it
is closely linked to the MDGs‘ or SDGs‘ targets. MPI incorporates alternative indicators; poverty
cut-offs, and weights and is composed of three dimensions made up of ten indicators. Each
indicator is based on international consensus (such as the MDGs) and the minimum level of

1
satisfaction is called a deprivation cut-off. MPI combines the percentage of people who are poor
(headcount ratio) and the average percentage of dimensions in which poor people are deprived.

2.1.6. Global Multi-Dimensional Poverty Index 2020

The 2020 update of the global Multidimensional Poverty Index (MPI) covers 107 countries, 28
low income, 76 middle income and 3 high income and 5.9 billion people in developing regions.
MPI values and data for the MPI‘s component indicators are also disaggregated by age group, for
rural and urban areas and for 1,279subnational regions. Data for 25 countries covering 913
million people have been updated from the 2019 release. The 2020 estimates are based on 47
Demographic and Health Surveys (DHS), 47 Multiple Indicator Cluster Surveys (MICS), 3 Pan
Arab Population and Family Health Surveys and 10 national surveys. All surveys are dated2008–
2019, and data for 83 countries—home to 92 percent of multidimensional poor people—were
collected in 2013/2014 or later.9The global MPI is thus a key resource for recent poverty data
across developing regions OPHI and UNDP (2020)

In the global MPI, people are counted as multidimensional poor if they are deprived in one-third
or more of 10 indicators (see figure 1), where each indicator is equally weighted within its
dimension, so the health and education indicators are weighted 1/6 each and the.

The global MPI begins by establishing a deprivation profile for each person, which shows which
of the 10 indicators they are deprived in. Each person is identified as deprived or non-deprived in
each indicator based on a deprivation cutoff (Table 1). In the case of health and education,
each household member may be identified as deprived or not deprived according to available
information for other household members. For example, if any household member for whom
data exist is malnourished, each person in that household is considered deprived in nutrition.
Taking this approach which was required by the data does not reveal intra household disparities,
but it is intuitive and assumes shared positive (or negative) effects of achieving (or not
achieving) certain outcomes.

1
Figure 2.1: Composition of the Global MPI- Dimensions and indicators

Standard of living indicators are weighted 1/18 each. The intensity of multidimensional poor
people is measured by the average number of weighted deprivations they experience. The
MPI is the product of the incidence of poverty (proportion of poor people) and the intensity
of poverty (average deprivation score5 of poor people) and is therefore sensitive to changes in
both components. The MPI ranges from 0 to 1, and higher values imply higher poverty OPHI
and UNDP (2020).

The global MPI begins by establishing a deprivation profile for each person, which shows which
of the 10 indicators they are deprived in. Each person is identified as deprived or non-deprived in
each indicator on the basis of a deprivation cutoff (Table 1). In the case of health and education,
each household member may be identified as deprived or not deprived according to available
information for other household members. For example, if any household member for whom
data exist is malnourished, each person in that household is considered deprived in nutrition.
Taking this approach, which was required by the data, does not reveal intra household
disparities, but it is intuitive and assumes shared positive (or negative) effects of achieving (or
not achieving) certain outcomes.

Looking across indicators, each person‘s deprivation scores based on a weighted average
of the deprivations they experience. The indicators use a nested weight structure: equal weights
across dimensions and an equal weight for each indicator within a dimension. The global MPI

1
specifies that a person is identified as MPI poor if he or she is deprived in at least one-third of the
weighted indicators. In addition, the measure also identifies those who are close to the one-third
threshold, that is, individuals are vulnerable to multidimensional poverty if they are deprived
in20% to 33.33% of weighted indicators. The measure also specifies a higher poverty cutoff to
identify those in severe poverty, that is, those deprived in 50% or more of the dimensions.
Table 1 : Global MPI 2020- Dimensions, Indicators, deprivation cutoffs and weights

Source Global Multidimensional Poverty Index (2020)

Health
The MPI uses two health indicators .The first indicator looks at nutrition of household members.
Children under 5 years (60 months and younger) are considered undernourished if their z-score
of either height-for-age (stunting) or weight-for-age (underweight) is below minus two standard
deviations from the median of the reference population. Children 5–19 years (61–228 months)
are identified as deprived if their age-specific BMI cutoff is below minus two standard
deviations. Adults older than 19 to 70 years (229–840 months) are considered undernourished if
their Body Mass Index (BMI) is below 18.5 m/kg² (Alkire, Kanagaratnam and Suppa, 2020).

1
The MPI identifies a person as deprived in nutrition if any person under 70 years of age for
whom there is nutritional information is severely undernourished (Alkire, Kanagaratnam and
Suppa, 2020).

The second indicator uses data on child mortality. The second indicator uses data on child
mortality. Most, although not all, child deaths are preventable, being caused by infectious
disease. Child malnutrition also contributes to child death. In the MPI, each household member
is considered to as deprived if a child under 18 has died in the household (Alkire, Kanagaratnam
and Suppa, 2020).

Education
The MPI uses two education indicators that harmonize each other within the education
dimension. The first looks at completed years of schooling of household members, the second at
whether children are attending school. Years of schooling acts as a proxy for the level of
knowledge and understanding of household members. Note that both years of schooling and
school attendance are imperfect proxies. They do not capture the quality of schooling, the level
of knowledge attained or skills. Yet both are robust indicators, are widely available, and provide
the closest feasible approximation to levels of education for household members. (Alkire,
Kanagaratnam and Suppa, 2020).In the MPI, each household member is considered to be
deprived if no eligible household member has completed six years of schooling and/or school-
aged child is not attending school up to the age at which he/she would complete class 8 (Alkire,
Kanagaratnam and Suppa, 2020).

Living standards
The MPI considers six indicators for standards of living. It includes cooking fuel, sanitation,
drinking water, electricity, housing, and assets (Alkire, Kanagaratnam and Suppa, 2020).The
selected deprivation cut-offs for each indicator are discussed below.

Cooking fuel: A household cooks using solid fuel, such as dung, agricultural crop, shrubs,
wood, charcoal, or coal consider to as deprived
Sanitation: A household is considered to have access to improved sanitation if it has some type
of flush toilet or latrine, or ventilated improved pit or composting toilet, provided that they are
not shared.

1
Drinking Water: A household has access to clean drinking water if the water source is any of
the following types: piped water, public tap, borehole or pump, protected well, protected spring,
or rainwater, and it is within a 30-minute walk, round trip.
Electricity: A household considered as deprived if it has no access to electricity
Housing: A household considered as deprived if it has inadequate housing materials in any of
the three components: floor, roof, or walls.

Assets: The household is deprived if; the household does not own more than one of these assets:
radio, TV, telephone, computer, animal cart, bicycle, motorbike, or refrigerator, and does not
own a car or truck. In general for the theoretical foundations of multidimensional poverty; there
are four approaches to multidimensional poverty analyses. First, the theory of the Fuzzy sets/
Fuzzy approach (TFA) is about the need for the characterizations of a whole series of variables/
particular aspect of poverty (Filippone, Cheli, & Agostino, 2001; Betti, Gagliardi, & Salvucci,
2014). Second, the information theory denotes that to identify the household as poor or not
poor, weights are required to be assigned to the indicators (Deutsch & Silber, 2005). Third, the
efficiency analysis approach denotes the need to brought information into composite/aggregation
index (Deutsch & Silber, 2005; Alkire, Foster, Seth, Santos, Roche, & Ballon, 2015). Fourth,
Sen (1981) capability theory and axiomatic derivations approach, aimed at designing poverty
indices (Deutsch & Silber, 2005; Chakravarty, 2006; Alkire & Foster, 2008;2011;).

2.2. Empirical Literature Review

A Multivariate logistic regression study conducted in Nekemte Town, Eastern Wollega Zone has
shown that the sex of household head significantly affect the poverty of the household (Melese,
et al., 2017).The household head being female is positively correlated with the probability of
being poor (Kebede & Sharma, 2014; Teka et al., 2019; Tsehay & Bauer, 2012).Female farm
managers in Ethiopia are23 percent less productive than their male counterparts. They have less
time to spend on formwork and farm less land, more of which is rented (World Bank, 2015).
Female-headed households, especially in rural areas of Ethiopia are likely to have lower
consumption (FDRE, 2018).

1
Age of the household head is measured in years and a study conducted in Ethiopia by Kebede
and Sharma (2014) shows that the age of household head is negatively correlated with the
probability of being poor. Hence, the age of the household head is expected to be positively and
negatively associated with the welfare of the households.

A study done in Southern Nations, Nationalities and Peoples Region of Ethiopia by Mohammed
(2017) shows that urban poverty found to be statistically significant variables with marital status
of the household head. Likewise, a study conducted in Uganda using multinomial logistic
regression analysis reveal that marital status of household‘s head is a significant determinant of
urban poverty Eunice (2011).Education has a clear and positive correlation with consumption, in
both urban and rural areas. Even completing informal education shows significant increases in
consumption, showing that investment in adult education may also pay returns in Ethiopia
(FDRE, 2018).Literacy and schooling are important indicators of the quality of life in their own
right. That is the literacy was found to be negative and significant, which means that literate
household heads are less likely to be poor than are illiterate households. This means that
education might increase earning potential and improve the occupational and geographic
mobility of labour (Kebede &Sharma, 2014; Teka et al., 2019; World Bank Institute, 2005). A
study conducted in Uganda also shows household head education were significantly associated
with the urban poverty Eunice (2011).

The previous study shows that credit is positively associated with the welfare of households
(Teka et al., 2019; Tsehay& Bauer, 2012). Access to and utilization of credit is facilitated to
increase or diversify household incomes and outflow from poverty. Research done by Kassie et
al. (2014) found that possession of assets such as access to credit was found to be positively
related to the well-being of sample households in Malawi. So, access to credit is expected to be
positively associated with the welfare of households.

The total family size of the household affects the welfare of households. Households with
larger family sizes are more likely to be poor (Bersisa & Heshmati, 2016; Kebede & Sharma,
2014).Multi-Dimensional Poverty study finding suggests that standard of living dimension
contributes the most to MPI. MPI has substantial relationships with household head's educational
achievement, access to electricity, and asset stock. Another multi-dimensional poverty study

1
finding suggests that standard of living dimension contributes the most to MPI. MPI has
substantial relationships with household head's educational achievement, access to electricity,
and asset stock.( Ebenezer T.2018).A study done in Ethiopia by Etaferahu C. and Andualem G.
(2019) showed that access to loan and access for own house are negatively significant with
urban multi-dimensional poverty.

2.3. Conceptual Framework

The researcher developed a conceptual framework based on the reviewed literature to analyze
the determinants of urban multidimensional poverty. The three dimensions that is health
dimension, education dimension and living standard dimension are aggregated to the urban
multidimensional poverty index of the household as per Alkere and Forest (2011) procedure. The
urban multidimensional poverty which is the depended variable of the study consists of a dummy
variable categorized as poor and non-poor status of the households. The relationship between the
depended and the independent variables is analyzed by using spearman rank correlation. And to
estimate the effect of the independent and on the dependent variables the logistic regression
model is used.
Figure 1.2: Conceptual Framework

Independent Variables
Health Dimension
Sex of Household Head
Age of Household Head
Marital Status Household Head
Employment Status of Household Head Dependent Variable
Year of Education of Household Head
Household Size
Unemployed Household Members Aged 0-14 Years Old
Household member obtained loan
Household own the house Education Dimension Urban Multidimensional Poverty

Living Standard Dimension

Source: (Reviewed literature, 2021)

2
 CHAPTER THREE

DATA AND METHODOLOGY

3.1. Description of the Study Area

The study was conducted in Kolfe Keraniyo sub-city of Addis Ababa City Administration in
Ethiopia. Kolfe keraniyo sub-city is the largest and the most populous sub-city in Addis Ababa.
According to 2007 national population census conducted by Central Statistical Agency (CSA),
the population of the sub city was 428,895. The sub-city has 15 administrative areas (Woredas‘)
with estimated 104,609 households.

Figure 3.1: Map of Study Area

2
3.2. Research Design
The main purpose of this study was to assess the determinant of multi-dimensional urban
poverty in Addis Ababa the case of Kolfe Keranio sub city. Therefore, to investigate this study
utilized both descriptive and econometric analysis and use multi-dimensional poverty estimation
to calculate multidimensional poverty index.

3.3. Source of Data

The data for this study was obtained from 398 households reside in Kolfe Keranio Sub city of
Addis Ababa City Administration. For this purpose, a survey design was adopted, and a
questionnaire was used to gather information from the target population. The target population of
the study was all households located in Kolfe Keranio Sub city. In the current study, the target
population was big enough to conduct a sampling procedure. Therefore, data were aimed to be
collected from the sample households.

3.4. Sample Size Determination

The sample of this research is calculated by using Taro Yamane (Yamane, 1973) sample size
determination formula with 95% confidence level. Such method is the most appropriate way of
sample size determination for finite population (Yamane, 1973). This is presented as follow:
𝑁
𝑛=
1 + (𝑁 · 𝑒2)
Where,
𝑛is the sample size,
𝑁 is the population size and
𝑒 is the acceptable margin of error.
According to CSA (2007), there are 104,609 households in Kolfe Keraniyo Sub city with a total
population of 428,895, of whom 207,641 (48.4%) are male and the remaining 221,254 (51.6%)
are female. Therefore, the sample size is calculated as

104609
n = 1 + (104609 x (0.05)2) = 398
Therefore, according to the above calculation a total of 398 questionnaires was randomly
distributed to collect relevant data from 398 households in Kolfe Keraniyo sub cities.

2
3.5. Method of Data Analysis

3.5.1. Multidimensional Poverty Index (MPI) Estimation Method

The method used to measure MPI in this paper is the Alkire and Foster‘s (2011)
multidimensional poverty measures, later called the AF methodology. The AF method is
explained as follows:
Let n represent the number of households and m ≥ 2 be the number of dimensions. Each
dimension is represented by wellbeing indicators j where j is between 1 and d. Let Y = |Yij|
denote the n × d matrix of achievements, where the typical entry Yij≥ 0 is the achievement of
household=1, 2,…, n in wellbeing indicator j=1,2, …,d. |Zj| > 0 is the indicators cutoff below
which a person is considered to be deprived in indicator j.

For any given Y, let g = |gij| is a deprivation gap, which denote the 0-1 matrix of deprivations
associated with Y, whose typical element gij is defined by g ij= 1 when Yij<Zj, while gij= 0
otherwise. Clearly, |gij| is an n×d matrix whose ijth entry is 1 when household i is deprived in the
jth indicator, and 0 when a person is not.

After the identification of deprivations, the next step is assigning weights to each dimension. The
AF method implicitly assigned an equal weight to each dimension and similar weights to all
indicators within a dimension. This has been done by assuming that the available chosen
dimensions are relatively equally important (Alkire and Foster, 2011). Similar to the AF method,
this paper used an equal weighting approach to each dimension and similar weights for indicators
within a dimension. Having the weighted deprivation gap (w jgij) for each indicator, finding the
aggregate deprivation score for everyone (Ci) is the next task. Ciis defined as the horizontal sum
of weighted deprivation gaps for everyone, which is written as follows:
𝑑

𝐶i = ∑ Wj𝑔ij.................................... (3.1)
j=1

The last step in the estimation of MPI is identification of those who are poor and not. In a
multidimensional framework, there are three types of identification rules: intersection, union and
intermediate. Under the union approach a person iis said to be multidimensional poor if there is
at least one indicator in which the person is deprived. The intersection approach identifies
person i as being poor only if the person is deprived in all indicators j . AF methodology uses an

2
intermediate cutoff level for Ci that lies somewhere between the two extremes of 1 and j.
Therefore, AF identification includes the union and intersection methods as special cases of
extreme values (Alkire and Santos, 2011). Consider 𝑘as the poverty cutoff and 𝑞as the number
of poor people, then person i is considered poor when the number of indicators in which i is
deprived is at least k. On the other hand, if the aggregate deprivation score falls below the cutoff
k, then person i is non-poor and his/her value will be censored to zero. From eq1, if we censored
all values of Ci to zero which are located below k, we will get a censored aggregate deprivation
score (Ci*). Hence, a person is identified as poor when the aggregate score Ci is above k (or
equivalently Ci*>0). The main challenging task in the intermediate method is the choice of the
appropriate cutoff k among a set of k poverty cutoffs.

The choice of the appropriate k has more of a normative task which is left for the researcher
like the income poverty (Sen, 1979; Alkire et al., 2014). Alkire et al, 2014 suggested two
methods of choosing the appropriate cutoff from a set of alternatives. The first method to select
the appropriate cutoff is to identify the number of poor people based on the available resources.
In this case, the policy maker a priori selects the number of poor segment of the society that
could be accommodated by the available resources. The second method is to use 1/3 to
1/5 of the available indicators. From ―communication‖ point of view, those people who
are deprived of 1/3 to 1/5 of the available indicators are vulnerable of becoming
multidimensionally poor. ―In the MPI, a person is identified as poor if he or she has a
deprivation score higher than or equal to 1/3. In other words, a person‘s deprivation must be no
less than a third of the (weighted) considered indicators to be considered MPI poor‖(Alkire and
Santos, 2011).Following this, the AF family of multidimensional poverty computation has two
main parts
The first one is multidimensional headcount ratio (H) which is the proportion of incidence
(depth) of people who experience multiple deprivations.
𝑞
𝐻=
𝑛
Where 𝑞 is the number of multidimensional poor households and 𝑛 is the total number of
households. The second one is the intensity or width of poverty (A) is the average deprivation
score of those poor segments of the population, written as

2
 ∑𝑛 1
C*
𝐴= i=1 d
𝑞
Where 𝐶* the censored deprivation score of individual i and 𝑞 is the number of people who are
multidimensional poor. Therefore, multidimensional poverty is the product of the above two
terms
𝑀𝑃𝐼 = 𝐻 × 𝐴
This study used three dimensions to measure multidimensional urban poverty: education, health,
the standard of living and each dimension have choice three indicators according to the
availability of data this selection is based on the Global multi-dimensional poverty index 2020.
The following table shows those indicators with the associated cutoffs used to identify deprived
households.

Table 3.1: MPI dimensions, indicators and weight

Dimensions Indicators in each

of MPI dimension Indicators cutoff (Deprived if) Weight

Nutrition Any person under 70 years of age for whom there is nutritional
information is undernourished. 1/6
Health
Child mortality A child under 18 has died in the household in the five-year period
preceding the survey 1/6

Years of schooling No eligible household member has completed six years of

Education schooling. 1/6
Any school-aged child is not attending school up to the age at
School attendance
which he/she would complete class 8. 1/6

Cooking fuel A household cooks using solid fuel, such as dung,

agricultural crop, shrubs, wood, charcoal, or 1/18
coal.
The household has unimproved or no
Sanitation sanitation facility or it is improved but shared with other
households. 1/18
The household‘s source of drinking water is
Living
Drinking water not safe or safe drinking water is a 30-minute or longer walk from
Standard
home, roundtrip. 1/18
Electricity The household has no electricity 1/18
The household has inadequate housing
Housing materials in any of the three components:
floor, roof, or walls 1/18
The household does not own more than one of these assets: radio,
Assets TV, telephone, computer, animal cart, bicycle, motorbike, or
refrigerator, and does not own a car or truck. 1/18
Source (Global MPI 2020)
2
3.5.2. Descriptive Data Analysis
We started with an exploratory data analysis to gain insight into the dataset. Descriptive statistics
were used to observe the measure of MPI at a possible link between explanatory variables and
multidimensional poverty status. The explanatory variables variable were assessed using
frequency tables, mean and standard deviation and the response variable were assessed with the
use of pie chart.

3.5.3. Correlation Analysis

The correlation between variables can be measured with the use of different indices (Hauke J. &
Kossowski T, 2011). The three most popular are Pearson‘s coefficient (r), Spearman‘s rho
coefficient (rs), and Kendall‘s tau coefficient (τ). Kendall‘s tau, introduced by Kendall (1938), is
a correlation coefficient that can be used as an alternative to Spearman‘s rho for data in the form
of ranks. To apply the Pearson correlation analysis the study should meet linearity assumption,
interval or ratio data and finally the assumptions of normality. If these assumptions not fulfilled
it is better to use Spearman rho. In this study, because of the data were ordinal, the researcher
was performed Spearman‘s correlation analysis to identify the association between variables and
to test the research hypothesis (Hogan, and Agnello, 2004).

3.5.4. Econometric Model

3.5.4.1. Logistics Regression

Logistic regression is a popular modeling approach when the dependent variable is dichotomous,
ordinal or multinomial. It allows predicting the log odds of outcomes of a dependent variable
from a set of variables that may be continuous, discrete, categorical, or a mix of any of these
(Agresti, 2002).

3.5.4.2. The Model

For a binary response 𝑌iand a quantitative explanatory variableXji , j = 1,2, … , 𝑚, and Xji , j =
1,2, … , 𝑛 . Let 𝜋i = 𝑃(𝑥i)denote the success probability whenXji take the values 𝑥ji.The problem
with linear model is that the probability mode 𝐸(𝑌) (where is the vectorof parameters to be
estimated) is used to approximate a probability value 𝜋i = 𝑃(𝑥i)within the interval 0 and 1,

2
while is not to be constrained. Therefore, we apply the logit transformation where the
transformed quantity 𝑙𝑛 ( 𝜋i )lies in the interval(−∞, ∞)and it is modeled as
1−𝜋i
𝜋i
𝑙𝑜𝑔i𝑡(𝜋 ) = 𝑙𝑛 ( ) = 𝛽 + 𝛽 X + ⋯ + 𝛽 X … … … … … … … . . (3.2)
i
1− 0 1 1 𝑚 𝑚i

𝜋i
The probability of success can be expressed as
𝑌i
𝜋 =𝑃( ,…, 𝑒𝑥𝑝(𝛽0 + 𝛽1X1 + ⋯ + 𝛽𝑚X𝑚i) ...................
X )= (3.2)
i 𝑚i
1 − 𝑒𝑥𝑝(𝛽0 + 𝛽1X1 + ⋯ + 𝛽𝑚X𝑚i)

X1i
Where the parameter 𝛽jdetermines the rate of increase or decrease of Xjion the log of odds
that𝑌i = 1 , controlling for other variables. Furthermore, exp(𝛽j)is the multiplicativeeffect on the
odds of unit increase in Xji , at fixed level of other Xs (John and Forrest,1984).
With further rearrangement we obtain the odds of success.

Odds(Y1i πi
= 1) = = exp(X'β).................................... (3.3)
1 − πi
The above three equations give suitable representations of log-odds, the success probability, and
odds, respectively. Indeed, these representations facilitate interpretations of parameter estimates.
The parameter refers to the effect of xi on the log odds that Y = 1, controlling the other X's.

3.5.4.3. Goodness of Fit of the Model

The goodness of fit or calibration of a model measures how well the model describes the data.
Assessing goodness of fit involves investigating how close values predicted by the model are to
the observed values. For likelihood-based models, the Hosmer and Lemeshow test is one of the
most commonly recommended tests for overall fit of a binary logistic regression model (Hosmer
& Lemeshow1980).

3.5.4.4. Model Diagnostics

Model building is not the final goal in regression analysis. The model adequacy checking is the
main step of regression analysis after a model fit. It can measure based on diagnosing residuals
and measure of influence. The most used model diagnosis is described here under.

2
Leverage Values (Hat Diagonal) is a measure of how far an observation is from the others in
terms of the levels of the independent variables (not the dependent variable). Observations with
leverage values larger than one are potentially highly influential (Belsley et al., 1980).

2
DFBETAS measure how much an observation has affected the estimate of a regression
coefficient (there is one DFBETA for each regression coefficient, including the intercept). If
DFBETAs is less than unity, this implies no specific impact of an observation on the coefficient
of a particular predictor variable, while DFBETA of a case greater than 1.0, and implies the
observation is an outlier (Cook and Weisberg, 1982).
Cook’s D is a measure of aggregate impact of each observation on the group of regression
coefficients, as well as the group of fitted values. In logistic regression, a case is identified as
influential if its Cook's distance is greater than 1.0 (Cook, R. D., 1998)

3.5.5. Study Variables

3.5.5.1. Response Variable
The dependent variable of this study is the multidimensional poverty status. Following Alkire
and Santos (2011) method of measuring multidimensional poverty, a household‘s deprivation
score (𝑐i)is compared with the multidimensional poverty cut-offs (𝑘). A house is considered poor
if they are deprived in at least one third of the weighted indicators. In other word a household is
identified as poor if it has a deprivation score greater than or equal to one-third (33 percent)
(Alkire and Santos, 2011; OPHI, 2013). Following this we 0.33 cut off point for this study. This
is represented by the binary variable (Yi) that takes the value 1 or 0, as:
1, if multidimensional poverty (𝑐i ≥ 0.33)
𝑌i = { 0 , Not multidimensional poverty

The following Steps were used the study to calculate multidimensional poverty according
to (Alkire and Foster (2007, 2011), those are:-
i. Select dimension:- health, education, living standard dimension
ii. Select indicators for each dimension according to data
iii. Use the first cutoff to determining deprivations (1 = deprived and 0 = non deprived)
iv. Attach weight for three dimensions each has weight .33 and for each dimension
equally distributes the weight among indicator
v. Use second cutoff to determine poor person, since we have ten indicators, a person who
do not have 1/3 of the total 10 indicators considered as poor, So a person who score
below 0.33 considered as non-poor and get value 0 then count number of poor and
calculate headcount index (H) = no of poor/total no of household and the other one

2
 is the intensity of poverty (A) is the average deprivation score of those poor households
*
∑𝑛i=1d1C
i.e 𝐴 =
𝑞

vi. Last calculate Mo = H×A

3.5.5.2. Explanatory Variables

The predictor variables at household and household head level expected to determine urban
multidimensional poverty are listed in the below table.
Table 2.2: Description and Measurement Types of Explanatory Variables

Expected
No Explanatory Variables Measurement Type Sign
1 Sex of Household Head Dummy( 0= Female , 1= Male) +/-
Categorical(0= 15-24 years, 1= 25-34 years, 2 +/-
= 35-44 years,3 = 45-54 years, 4 = 55 - 64
2 Age of Household Head years , and 5 = 65 years and above)
3 Marital status of Household Head Dummy( 0= unmarried , 1= Married) +/-
4 Year of Education of Household Discrete +/-
Employment Status of Household Categorical (0 = Unemployed= Employed
5 Head 2 = Pensioner) -
6 Household Size Discrete +/-
Unemployed Household Members +/-
7 Aged 0-14 Years Old Continuous
8 Household member obtained loan Dummy( 0 = No , 1=Yes) +/-
9 Household own the house Dummy ( 0 =No, 1 = Yes) -
Source (Own Survey, 2021)

Table 3.2 shows that a total of nine independent variable of is proposed based on reviews of
literature described in chapter two of this study. These independent variables are fitted to the
logistic regression model in chapter four to identify the determinant of multidimensional poverty
in kolfe Keranio sub cities of Addis Ababa City Administration. The proposed independent
variables are specifically sex of household head, age of household head marital status of
household head, year of education of household, employment status of household head,
household size, unemployed household members aged 0-14 years old, household member
obtained loan household own the house. And their measurement type and levels are mentioned in
detail in the Table 3.1.

3
3.5.6. Instrumental Reliability and Validly

3.5.6.1. Instrument reliability

One of the major requirements of any research process is the consistence of the data and
the results of the research finding (Kothari, 2004). A measuring instrument is reliable if
it provides consistent results (Creswell, 2014). Correspondingly, to Cohen and his colleague
(2003:117), reliability and consistency of the instruments over a group of respondents. By the
same token, Kumar (1999) and Perry (2005) proposed that reliability is the issue of consistency,
stability, predictability and accuracy of the research instrument. Finally, to make sure the
reliability of Cronbach alpha test has been employed. The Cronbach alpha coefficient is the most
common method used for assessing the reliability of a measurement scale (Hayes & Bob, 1998).
The coefficient, which reflects homogeneity among a set of items, varies from 0 to 1.

The literature regarding test and scale construction suggests that an acceptable level of
reliability is a function of the intended use of the test results. Nunnally JC. (1967) suggests
that when a test or scale is used to make decisions about individuals, the reliability coefficients
should be at least 0.90. However, it is impossible to achieve this number, especially assessing
personality and feelings. Others are somewhat less conservative, suggesting that a reliability
coefficient of 0.80 is acceptable for a test or scale that will be used for making decisions about an
individual (Batjelsmit, 1977). Moreover, Saad, et al (1999) argues the following interpretations:
0.90 or higher = excellent, 0.80 to 0.89 = good, 0.70 to 0.79 = adequate and 0.69 and below =
may have limited applicability. This study had a total of 10 variables (one dependent and nine
independent variables) and the overall cronbach alpha value is depicted in Table 3.3.below

Table 3.3 Table Reliability Statistics

Cronbach's Alpha N of Items
.891 10

From Table 3.3 we can see that the overall Cronbachs‘ alpha value of the study was 0.891, which
indicates that there was a good internal consistency in the scale (Saad et al,1999).

3
3.5.6.2. Instrument Validity
Validity is the most critical criterion and indicates the degree to which an instrument measures
what it is supposed to measure. To assure the quality and the acceptability of the research
evaluating validating is the most imperative task for the researcher (Burns, 1999). As Fraenkel
and Wallen (2003) claimed that it is unquestionable to deal with the issues of validity of the
instrument. Because researchers draw a conclusion from the study based on the instrument they
design to collect data from the respondent. In the present study content validity, internal validity
and external validity have been assessed to achieve the overall objectives of the study.

3.5.6.3. Content Validity

Scholars have argued that content validity does not have an exact meaning. Despite the
fact that, most of them outlined content validity as the degree to which the elements of an
evaluation instrument are representative of the dimensions or the variables of the investigation
(Hayness et. al 1995). Along the same line, Polit & Beck (2006) claim that content validity as the
extent to which an evaluation instrument contains an adequate sample of items for the construct
assessed. Similarly, others determine the content validity as the levels of the instrument for a
study has enough samples (Wynd, Schmidt, & Schaefer, 2003). Generally, therefore in this study
content validity understood as the adequate items for each variable and to what extent that items
measured the constructs or the variables (Polit & Beck 2006). To this end, the adequacies of
items for this investigation were checked by the researcher advisor of the study.

3.5.6.4. Internal Validity

As Cook (1976) and Campbell (1979) proposed that, internal validity assessed whether or not an
observation covariation should be considered causal relationships. Internal validity is the extent
that an experimental variable is truly responsible for any variance in the dependent variable
(Kothari, 2004). In line with this other researchers argued that internal validity is the degree to
which a study establishes the cause-and-effect relationship between the treatment and the
observed outcome (Marion & Jolaine, 2001). Internal validity also defined as a logical rather
than statistical issue (Campbell & Stanley, 1963). The logical framework of the research is
provided by the report‘s structure of the study. The method section describes how the study was
designed and what procedures were followed to reduce or eliminate specific threats to internal

3
validity (Campbell & Stanley, 1963). The results section reports the data relevant to establishing
the internal validity and the discussion section provides the investigators‘ assessment of the
influence of bias. In all, in this study, the researcher made internal validity using a logical
process in each part of the research design, results, and discussions.

3.5.6.5. External Validity

External validity is the accuracy with which experimental results can be generalized
beyond the scope of the study (Creswell, 2014). External validity examines whether or
not an observed causal relationship should be generalized to and across different
measures, persons, settings and times (Campbell & Stanley, 1963).

3
 CHAPTER FOUR

RESULTS AND DISCUSSION

This chapter presents, analyzes, and interprets the data received from participants via a
standardized questionnaire. The overarching purpose of the research was to investigate the
factors that influence multidimensional urban poverty in the Kolfe Keranio sub-city. Both
descriptive and inferential statistical data analysis methodologies were used in this investigation

4.1 Descriptive Analysis

In this section, data obtained from 398 households' for the study is summarized
Table 4.1: Descriptive Analysis of Categorical variables

Valid
Variables Category Frequency Percent Percent Cumulative Percent
Sex of Female 87 21.9 21.9 21.9
Household Male 311 78.1 78.1 100.0
Head Total 398 100.0 100.0
Marital Single 35 8.8 8.8 8.8
Status of Married 363 91.2 91.2 100.0
Household
Total 398 100.0 100.0
Head
Employment Unemployed 82 20.6 20.6 20.6
Status of Employed 270 67.8 67.8 88.4
Household Pensioner 46 11.6 11.6 100.0
Head Total 398 100.0 100.0
15-24 2 .5 .5 .5
25-34 66 16.6 16.6 17.1
Age of 35-44 181 45.5 45.5 62.6
Household 45-54 77 19.3 19.3 81.9
Head 55-64 46 11.6 11.6 93.5
>=65 26 6.5 6.5 100.0
Total 398 100.0 100.0
Household No 253 63.6 63.6 63.6
member got Yes 145 36.4 36.4 100.0
loan Total 398 100.0 100.0
Household No 251 63.1 63.1 63.1
own the Yes 147 36.9 36.9 100.0
house Total 398 100.0 100.0
(Source: Own computation, 2021)

3
Table 4.1 shows that from the total households, 87 (21.9%) households are headed by female and
the remaining 331 (78.1%) households are headed by male. This indicates that the proportion of
male household head is larger than of female. From the sample household, it appears that 91.2%
of the household head are married and only 8.8 household head were found unmarried during the
survey time. The employment status of household head counted that, 82 (20.6%) have not
employed, 270(67.0%) employed, and the remaining 46 (37.9%) of the household head are
pensioner. This implies that the large majority of the household head are employed one.

With regards to household head age, the highest proportion of household head was observed
among those whose age group 35-44 years (45.5 percent) followed by age group 45-54 years
(19.3%) as opposed to the smallest percentage (0.5 percent) of household head was observed
among those whose age group of 15-24 years. Of the total household, 253 (63.6%) of them do
not get access to loan services. Only 145(36.4) of the sampled household get loan services. This
infers that majority of the sampled household do not get a chance of loan services.

Finally, the participants of this study were asked about their ownership of the house they live in.
Accordingly, 251(63.1%) households‘ responded the house they live in is not their own house.
Conversely, the remaining 147(36.9%) of the responded they own the house they currently live
in.
Table 4.2 Descriptive Statistics of Continues Variables

N Minimum Maximum Mean Std. Deviation

Year of Schooling 398 4.00 23.00 11.0251 3.17807
Household Size 398 2 9 4.40 1.571
Unemployed Household
Members Aged 0-14 398 0 3 0.79 .517
Years Old
Valid N (listwise) 398
(Source: Own computation, 2021)
It can be observed from Table 4.2 that the average years of schooling of the household head in
the study area is 11.025 years with a minimum and maximum of 4 years and 23 years of
schooling respectively. The result of the descriptive analysis also reveals that the average family
size of the sampled household is 4.4.

3
 Regarding the unemployed household member aged (0 to 14 years), on average 0.79 unemployed
household member aged (0 to 14 years) were found in the household with a maximum of 3
children per household.

4.2. Result of Multidimensional poverty analysis

The results of the table 4.3 show the number of deprived households in health, education and
living standard dimension and their respective indicators. The study used the counting approach
to identify the number of deprived household from non-deprived one that is counting the number
of deprived household for each specific indicator and presented the summarized deprivation
status of the households of the study area as follows
Table 4.3 Number of Household Deprived in Different Indicators
No of deprived
Dimensions Indicators of each Global MPI household and
of MPI dimensions Deprived if (%) Weight

Nutrition Any person under 70 years of age for whom there is

nutritional information is undernourished. 69(17.30)% 1/6
Health
Child mortality A child under 18 has died in the household in
the five-year period preceding the survey 65(16.30%) 1/6

Years of schooling No eligible household member has completed

six years of schooling. 219(55%) 1/6
Education
School attendance Any school-aged child is not attending school up to
the age at which he/she would complete class 8. 233(58.50)5 1/6

Cooking fuel A household cooks using solid fuel, such as dung,

agricultural crop, shrubs, wood, charcoal, or 235(59%) 1/18
coal.
The household has unimproved or no
Sanitation sanitation facility or it is improved but shared with
other households. 221(55.50%) 1/18
The household‘s source of drinking water is
Living Drinking water not safe or safe drinking water is a 30-minute
Standard or longer walk from home, roundtrip. 210(52.80%) 1/18
Electricity The household has no electricity 221(53%) 1/18
The household has inadequate housing
Housing materials in any of the three components:
floor, roof, or walls 231(58%) 1/18
The household does not own more than one of these
Assets assets: radio, TV, telephone, computer, animal cart,
bicycle, motorbike, or refrigerator, and does not own
a car or truck. 254(63.80%) 1/18
(Source: Own computation, 2021)

3
Table 4.3 showed that large number households deprived in assets (63.8%) of the living standard
indicator, followed by cooking fuel (59.0%) of the same living standard indicators and the lower
number of households is depraved in child mortality (16.3%). The table also showed that most of
the households deprived many indicators of the standard of living dimension. This implies that,
there are simultaneous deprivations. The following figure shows households status about
simultaneous deprivations.
.

(Source: Own computation, 2021)

Figure 3.2: Multiple Deprivations of Indicators of the Households

From the total sampled households, 157 households deprived in eight indicators simultaneously.
On the other hand, only 26 individuals faced simultaneous deprivation in 9 indicators.
Furthermore, the above graph also showed that except 46 households all the other 352
households are deprived at least in one indicator. In general, multiplicative deprivation is
unevenly distributed among the households and its inclusion in the estimation would have an
advantage to understand the severity of poverty that arose due to multiple deprivations. This
severity due to multiplicative deprivations is different from the severity of an indicator that
captured by the deprivation gap. As shown in figure 3 above, those 26households who are

3
deprived in 9 indicators simultaneously are suffering more than any other households in the
sample.

Table 4.4: MPI Estimation

Poverty cutoffs H(incidence) A(intensity) MPI

𝑘 = 33% (𝐶i ≥ 0.33) 0.6532 0.890 0.6457
(Source: Own computation, 2021)

As can be seen in the below table 4.4 the poverty headcount, H, is 0.6532 indicating that around
65.8% of the households were deprived in at least three of the indicator dimensions. Once this is
adjusted for the number of deprivations suffered, the MPI is computed as 0.6457. This indicates
that 64.57% of the sampled households are multidimensionally poor.

Table 4.5 Relative contribution of dimensions to MPI

Dimension H(incidence) A(intensity) MPI

Health 0.090 0.330 0.030
Education 0.467 0.330 0.154
Living Standard 0.568 0.610 0.346

The relative contribution of the various dimensions to overall multidimensional poverty is shown
in Table 4.5. The result showed that living standard dimension contribution to the overall MPI
(0.346) is higher as compare to others dimension because it has the highest MPI values. On the
contrary Health dimension has lower contribution to overall MPI (0.03).

4.3. Results of Spearman Correlation Analyses

Based on the result of the spearman correlation analyst in Table 4.6below, urban
multidimensional poverty was found to be significantly associated with sex of household head,
marital status of household head, year of education of household, employment status of
household head, unemployed household members aged 0-14 years old, household member got
loan (loan access)and household own the house at 5% level of significance. Except sex of
household head and marital status of household head, all other significant predictor variables are

3
negatively associated with urban multidimensional poverty whereas in the study,urban
multidimensional poverty does not have significant association with age of household head and
household size. And due to this we exclude the non-significant predictor variables from the
logistic regression model.

Table 4.6 Spearman Rank Correlation Analysis of the explanatory and the response
variable

MPI Sex_HH~d Age_Ca~d Mariti~d Educat~d Employ~d Househ~e

Unempl~f
MPI 1.0000
Sex_HH_Head 0.1571* 1.0000
Age_Catego~d 0.0144 -0.0685 1.0000
Maritial_S~d 0.1860* 0.0719 -0.0251 1.0000
Education_~d -0.2509* -0.0340 -0.0108 -0.0333 1.0000
Employent_~d -0.1516* -0.0055 -0.0310 0.0764 0.0965 1.0000
Household_~e -0.0059 -0.0442 0.4596* -0.0719 -0.0660 -0.0296 1.0000
Unemployed~f -0.1013* -0.0515 -0.3902* -0.0135 0.0241 -0.0157 -0.0859 1.0000
Access_Loan -0.6119* -0.1049* -0.0088 -0.0415 0.1901* 0.0971 -0.0243 0.0662
Home_Owner~p -0.5168* -0.1873* 0.0207 -0.0565 0.1387* 0.0707 0.0116 0.0216

Access~n Home_O~p

Access_Loan 1.0000
Home_Owner~p 0.8702* 1.0000

4.4. Econometric Analysis

Multiple logistic regression analysis was used to identify the effect of each predictor variables on
multidimensional urban poverty status of the households. Since the response variable is binary
meaning zero or one, with one being multidimensional poor and zero otherwise.

4.4.1. Parameter Estimation

The parameter estimation analysis was done by using Stata 14 software package. The resulting
maximum likelihood estimate of model parameters is presented in Table 4.3.

3
Table 4.7 Maximum Likelihood Estimates of Model Parameters

Logistic regression Number of obs = 398

LR chi2(8) = 212.48
Prob > chi2 = 0.0000
Log likelihood = -169.0224 Pseudo R2 = 0.3860

MPI Odds Ratio Std. Err. z P>|z| [95% Conf. Interval]

Sex_HH_Head
Male 1.861584 .6453192 1.79 0.073 .9436551 3.672419

Maritial_Status_HH_Head
Married 7.1405 3.69969 3.79 0.000 2.586411 19.71331
Education_HH_Head .889367 .0390928 -2.67 0.008 .8159543 .9693848

Employent_Status_HH_Head
Employed .2417208 .0925516 -3.71 0.000 .1141302 .5119501
Pensioner .4133517 .21576 -1.69 0.091 .1485977 1.149813

Unemployed_HHMem_Aged_lessthan_f .7194878 .1917284 -1.24 0.217 .4267721 1.212972

Access_Loan
Yes .0258045 .0173655 -5.43 0.000 .0069003 .0964992

Home_Ownership
Yes 1.683858 1.111513 0.79 0.430 .4617722 6.140205
_cons 4.475829 3.612345 1.86 0.063 .9202104 21.77007

In Table 4.7, estimated binary logistic regression model fit results sex of household head, marital
status of household head, year of education of household, employment status of household head,
unemployed household members aged 0-14 years old, household member got loan (loan access)
and household own the house. From these predictor variables marital status of household head,
year of education of household head, employment status of household head and household
member obtained loan were found to be statistically significant predictor of multidimensional
poverty status of the households at 5% level of significance and sex of household head was
significant at 10% level of significance. The rest predictor variable was not significantly
predicting the multidimensional poverty status of the households. The State result further reveals
that from the significant predictor variables, sex of household head and marital status of
household head have odd ratios greater than one, which means that these variables are positively
correlated with the probability of being multidimensional poor. On the other hand, predictor
variables such as year of education of household, employment status of household head,

4
household member obtained loan, odd ratios less than one, which means that these variables are
negatively correlated with the probability of being multidimensional poor.

4.4.2. Model Checking Techniques

After a model is fitted the next important step is checking its model adequacy, assuming that we
are primarily satisfied with the final model or model contains variables in their correct functional
form. The objective is to look at how closely model fitted responses approximate observed
responses. It is based on graphical as well as statistical point of view. There are numerous steps
involved in evaluating the appropriateness, adequacy and usefulness of model. First, the overall
usefulness was assessed. Second, the importance of each predictor variable was assessed by
carrying out statistical tests of significance of coefficients then, detecting influential
observations.

4.4.3. Goodness of Fit of the Model

The goodness of fit or calibration of a model measures how well the model describes the data.
Assessing goodness of fit involves investigating how close values predicted by the model are to
the observed values. The diagnostic test results of the Hosmer- Lemeshow test presented in
Table 4.7 below shows a chi-square value of 10.70 with p-value of 0.2192 which is not
significant showing that there is no difference between the observed and the model predicted
values and hence estimates of the model adequately fit the data.

4
Table 4.8 Hosmer- Lemeshow Goodness of Fit Test

. estat gof,group(10)table

Logistic model for MPI_Status, goodness-of-fit test

(Table collapsed on quantiles of estimated probabilities)

Group Prob Obs_1 Exp_1 Obs_0 Exp_0 Total

1 0.0600 1 1.6 39 38.4 40

2 0.0972 6 3.4 36 38.6 42
3 0.1715 3 5.1 35 32.9 38
4 0.4310 12 12.0 29 29.0 41
5 0.6671 23 23.8 16 15.2 39

6 0.7621 32 34.4 15 12.6 47

7 0.8020 34 31.5 6 8.5 40
8 0.8513 25 26.6 7 5.4 32
9 0.9129 41 37.0 1 5.0 42
10 0.9707 33 34.7 4 2.3 37

number of observations = 398

number of groups = 10
Hosmer-Lemeshow chi2(8) = 10.70
Prob > chi2 = 0.2192

4.4.4. Model Diagnostics

The next important step in marginal model building is to perform an analysis of residuals and
diagnostics to study influence of observations. Residuals are used to filter points with outlying
response values. The diagnostic test results for detection of outliers and influential values are
presented in Annex. The DFBETAs for model parameters including the constant term and
Cook‘s influence statistic were both less than unity. DFBETAs less than unity imply no specific
impact of an observation on the coefficient of a particular predictor variable, while Cook‘s
distance less than unity showed that an observation had no overall impact on the estimated vector
of regression coefficients β. A value of the leverage statistic less than one shows that no subject

4
has a substantial large impact on the predicted values of the model. And none of the observation
has standard and deviance residuals larger than 3 in absolute value. The residuals less than 3 in
absolute value show the absence of an outlier observation. Thus, from diagnostic checking we
can say that our model is adequate (See Annex).

4.5. Discussions of the Results

Sex of Household Head: The model's outcome shows that the OR is 1.77. At a 10% level of
significance, this means that a household led by a man has a 1.77 times higher chance of being
multidimensionally poor than a home headed by a woman.. In other word, female-headed
households had a lower probability of being poor as compared to male-headed households. The
result of this study is in line with the study done by Mekonnen B. and Almas H. (2021) that
shows female-headed households had a lower probability of being poor as compared to male-
headed households. The other study by Jayamohan & Amenu (2014) showed that a household
head with more active females has a higher level of living standards and a lower probability of
being poor Jayamohan & Amenu (2014).

This may be associated with women decision making power in the household. a study by
Frederik B. and Sevias G. (2021) showed that assigning greater financial decision-making power
to women also holds the promise of improving household welfare in the longer term via its
impact on resource allocations. The empowerment of women has impacted positively not only on
the wellbeing of women themselves but also on their children‘s education. School enrolment has
been shown to be associated with women‘s decision-making autonomy in Mozambique (Luz and
Agadjanian 2015) and Honduras (Hendrick and Marteleto 2017).However the finding of this
study is inconsistence with previous study conducted in Ethiopia by Anteneh Mulugeta (2020)
and Araya Mebrahtu (2010).

Marital Status of Household Head: The estimated coefficients for marital status of household
head are positive and statistically significant; suggest that at a 1% significance level, household
headed by married household head have a higher probability of falling into multidimensional
urban poverty than the unmarried households head in the study area. The marginal effect
indicates, keeping other factors constant, married household head 7.08 (OR: 7.08) times more

4
likely to be multidimensional poor than their counterpart. The reason might be due to the fact
that if the household head get married household size will increase as new children are born and
expenditures increase which in turn leads to searching for mechanisms of fulfilling additional
needs and necessities for the family.

The finding of this study is in agreement with the findings in other study conducted in Ethiopian
by Melese T. et al. (2017) in of Nekemte town.

Year of Education of Household: With a year increase in education of household head the
likelihood of being multidimensional poor of the household decreased by 11.34% (OR: 0.8866).
This implies education reduces the probability of a household being poor. This is because a
higher level of education provides greater opportunities for a better job and, subsequently, a
higher income. The findings of this study is consistence with a study done in Ethiopia by Kebede
B.(2019) that showed if schooling level of the head of household increases by one grade, the
probability of household falling into urban poverty reduces by 1.863. The finding this study also
in line with the conclusions of other studies, such as Debeli and Endegena (2019) and Bigsten et
al. (2003). Furthermore, a study by World Bank (2002) conformed that an increase in
educational attainment, a household head could secure a job and take opportunities which would
otherwise not be possible and the household less vulnerable to poverty. Education is expected to
lead to increased earning potential and improve occupational and geographical mobility of labor.
Higher levels of educational attainment will provide higher levels of welfare for the household.

Employment Status of Household Head: The employment status of the household head
significantly affects multidimensional poverty status of the household in the area. Specifically, as
shown from the above table, a household headed by employed household head have less
probability of falling into multidimensional poverty status than household headed the
unemployed head. For the given household, keeping the random intercept and other covariates
constant, the likelihood of being multidimensional poor for employed household head was
decreased by0.2375 (OR = 0.2375) times that of unemployed household head in the study area.
This could have been resulted because employment is the source of income for the households.
The same result was observed in the study by Mekonnen (2002) and Debeli and Endegena
(2019). The result of the study further shows that is there is no significance difference between

4
the household headed by unemployed and pensioner household head regarding the likelihood of
being in multidimensional poverty status.

House hold member have accessed to loan: This variable is negative and statistically
significant OR=0.04. That, if other variables are held constant, the higher access to loan services
the lower probability of multidimensionally poor. This implied that a households having access
to loan in any organization, they create a better life by doing their own business and a
household‗s increase living standards, change their life style by getting a better health care or
nutrition, school and standard of living that is the quality of house, access to safe water,
improvement in sanitation, So a household that assess to getting loan have lower probability of
being multi dimensionally poor. As increased access to credit market enhances household
welfare through the provision of investment credit to boost household income (Adugna and
Heidhues, 2000) as well as smooth consumption (Zeller, et al, 1994), which could significantly
influence a household‘s income by helping its members to tap economic opportunities, thereby
assisting them to get out of poverty (Binswinger and Khandker, 1995; Adugna and Heidhues,
2000).

In this study home ownership do not have significant effect on the multidimensional poverty
status. This may be because of two reasons. The first reason is this study did not use
income/consumption as measure of poverty and owning house do not need to increase their
income or decrease their consumptions. The second reason is that this study used MPI to analyze
the poverty of the households and owing the house does not necessarily reflect quality of the
house as per the living standard dimension of multidimensional poverty measurement. The result
of this study also showed that despite the household own the house they live in, most of them
are poor in living standard dimension that is they are in poor in terms of cooking fuel usage,
sanitation usage, access drinking water, access to electricity, adequate housing materials and
ownership of assets.

4
 CHAPTER FIVE

CONCLUSIONS AND POLICY IMPLICATION

In this chapter we discuss the conclusion and policy implication of the study

5.1. Conclusion

This paper attempted to identify and analyze the determinants of the multidimensional poverty in
Kolfe Keranio Sub-City of Addis Ababa city using binary logistic regression model. A primary
level data that were collected from 398 households were used for analysis. The study used
education, health and living standards dimensions to compute multidimensional poverty index of
the household. The main objective was to examine the effects of predictors' variables on the
multidimensional poverty status of the households. Of the total sampled households, 157
households deprived in eight indicators simultaneously. On the other hand, 26 individuals faced
simultaneous deprivation of 9 indicators from the sampled households.

The results of the MPI analysis show that the incidence of poverty status of the sample
respondents is 65.32%, the intensity of poverty is 89.0% and the adjusted headcount ratio or MPI
is 64.57%.Moreover, the large number of households were found to be poor in living standard
dimension as compare to health and education dimension. The result of the Logit model captured
predictor variables that had significant effects on the multidimensional poverty. The model fit
results indicated sex of household head, marital status of household head, year of education of
household head, employment status of household head and household member obtained loan
were found to be statistically significant predictor of urban multidimensional poverty status of
the households at 5% level of significance and rest predictor variable were not significantly
predict the multidimensional poverty status of the households.

This study shows that male and married household heads are more likely to be affected by urban
multidimensional poverty than their female counterparts.. On the other hand, education found to
be an important element in reducing the exposure of urban multidimensional poverty. With an
increase in educational attainment, a household head has a high possibility to secure a job and
take opportunities to increase earning potential through employment and creation job
opportunity. The other important variables found in the study is employment status of household

4
head, a household headed by employed household head has a better chance of escaping from
urban multidimensional poverty. Finally, the crucial predictor variable in this study is access to
loan, As an increased access to loan enhances household welfare through the provision of
investment credit to boost household income, startup business of the household as well as
smooth consumption.

5.2. Policy Implication

Based on the analysis made, results obtained, and conclusions drawn, the following policy
implications are forwarded to the policy makers, concerned government actors and other
stakeholder.
i. To minimize multidimensional poverty, policy implications need be in place that
prioritize living standard components, followed by education and health.
ii. When formulating development strategies, the government, development agencies, and
other interested parties should take into account the major characteristics identified in
each sub city that contribute to multidimensional urban poverty in their country's
development efforts.
iii. The likelihood of the multidimensional urban poverty in the sub city is lower for the
household headed by female household head .Therefore; working more on women
empowerment is crucial steps to alleviate poverty issues at household level. As female is
the heart of the household, supporting and enabling them to generate their own income
has multi effects to improve the lives of the households
iv. Education is a weapon one has to alleviate multidimensional poverty. Hence, expanding
schooling access is a vital for reducing the vulnerability of household multidimensional
urban poverty. Besides, reducing unemployment through job creation, it can provide
technical assistance for self-employed opportunity for the household members.
v. As urban household do not participate in farming activities like rural households, we
highly recommend creating employment opportunity for the household if the policy
makes aims to reduce urban multidimensional poverty
vi. It is also suggested that assisting households in obtaining financial services reduces the
susceptibility of multidimensional urban poverty. Allowing urban residents to receive

4
 credit services will provide financial freedom for household members to launch new
economic operations, reducing the many facets of poverty..
vii. Finally this study has used cross sectional data collected from 398 households dwell in
Kolfe Keranio sub city of Addis Ababa city administration and the outcomes may not be
able to make generalization for other Sub city over a period of time. Therefore, the
researcher recommends conducting further studies to explore other factors affecting the
multidimensional urban poverty that were not covered in this study as well as
investigating the dynamics of urban multidimensional poverty overtime

4
 REFERENCES

Addis Ababa Bureau of Finance and Economic Development (2015) Socio-Economic Profile of
Addis Ababa, for the Years 2000- 2006 E.C. City Government of Addis Ababa,
Agresti, A., (2002).Categorical data analysis. Second edition, John Wiley and Sons, Inc,
publication, New Jersey.ISBN-13: 978-0471360933
ALEMU, B. A., NUPPENAU, E. & BOLLAND, H. (2009). Technical Efficiency Across Agro-
Ecological Zones in Ethiopia: The Impact of Poverty and Asset Endowments.
Agricultural Journal, 4, 202-207.
Alkire, S. Foster, J. Seth, S. Santos, M. Roche, M. and Ballon, P. (2015). Multidimensional
Poverty Measurement and Analysis, University of Oxford/ OPHI, United Kingdom.
Alkire, S., and Foster, J. (2008). Counting and Multidimensional Poverty Measurement, OPHI
Working Paper Series, United Kingdom, P. 33.
Alkire, S., and Foster, J. (2011). Counting and Multidimensional Poverty Measurement. Journal
of Public Economics, OPHI Working Paper Series 95, United Kingdom, (pp. 476-487).
Anteneh Mulugeta Eyasu (2020) Determinants of poverty in rural households:
Evidence from North-Western Ethiopia, Cogent Food & Agriculture, 6:1, 1823652,
DOI:10.1080/23311932.2020.1823652

Araya Mebratu and Zaid Negash(2010) Poverty And Income Inequality In Urban Areas: (Socio-
Economic Analysis Of Households In Wukro Wereda), Master thesis at Mekele
University.

Asmamaw Enquobahrie, 2004, Understanding Poverty: The Ethiopian Context, (A Paper

Assessment, In, the Ethiopian Economy Poverty and Poverty Alleviation, Proceedings of
at the 2nd International EAF Conference, Organized by the Centre for African
Batjelsmit, J. (1977). Reliability and validity of psychometric measures. In R. Andrulis (Ed.),
Adult assessment. New York: Thomas, 29–44.

Belsley et al. (1980). Regression diagnostics: Identifying influential data and sources of
collinearity, John Wiley & Sons, New York, 1980,

4
Bersisa, M., & Heshmati, A. (2016). Poverty and wellbeingin East Africa a multi-faceted
economic approach. Springer International Publishing Switzerland. http://
doi.org/10.1007/978-3-319-30981-1_1

Betti G., Gagliardi F., Salvucci V. (2013), Multidimensional and Fuzzy Measures of Poverty and
Inequality at National and Regional Level in Mozambique, in Betti G., Lemmi A. (eds.),
Poverty and Social Exclusion: New Methods of Analysis. London and New York:
Routledge, pp. 78-105.
Bigsten, A., B. Kebede, A. Shimeles, and M. Taddesse (2003). ‗Growth and Poverty Reduction
in Ethiopia: Evidence from Household Panel Surveys‘. World Development, 31 (1): 87-
96.

Burns, A. (1999). Collaborative action research for English language teachers. Cambridge:
Cambridge University Press.
Campbell, DT. 1979. Quasi-experimental Design: Design and Analysis Issues for Field Settings.
Rand McNally: Skokie, Il.
Campbell. D., & Stanley. J. (1963). Experimental and quasi-experimental designs for research.
Central Statistical Agency) (2015). National Population Abstracts projection for July 2015,
Addis Ababa.
Central Statistical Authority, 2010, Statistical Report on the Urban Employment Unemployment
Central Statistical Authority,( 1994), Statistical Abstract, Addis Ababa

Central Statistics Agency. (2020). Country and regional level consumer price indices (CPI).

Chakravarty, Satya and D'Ambrosio, Conchita, The Measurement of Social Exclusion. Review
of Income and Wealth, Vol. 52, No. 3, pp. 377-398, September 2006, Available at
SSRN: https://ssrn.com/abstract=924476 or http://dx.doi.org/10.1111/j.1475-
4991.2006.00195.x
Cohen, L., Lawrence. M., & Keith. M. (2000). Research Methods in Education.
Cook, R. D. (1998). Detection of influential observations in linear regression. Technimetrics,
19,15-18.
Cook, R. D. and Weisberg, S. (1982). Residuals and influence in regression. New York:
Chapman and Hall.

5
Creswell, W. (2009) ‗Research design: qualitative, quantitative and mixed methods
Approaches‘.Third edition. Los Angeles: Sage
Creswell. J. (2014). Research design: qualitative, quantitative and mixed method approaches (4th
edition ed.).
CSA (Central Statistical Agency) (2013). Population projections for Ethiopia 2007-2037. CSA,
Addis Ababa CSA
CSA (Central Statistical Agency), 2007. Population and housing census 2007. Addis Ababa:
Federal Democratic Republic of Ethiopia.
Decanq, K. and Lugo, M. (2010). Weights in multidimensional indices of well-being: An
overview. Discussion paper, Katholieke Universities Leuven.

Debeli Chala and Endegena Asmamaw (2019). Urban Poverty and its Determinants in
DebreMarkos City: A Household Level Analysis. Economy, 2019, 6(2): 65-75.

Dercon, S. and K. Pramila, 2003. Poverty in rural Ethiopia 1989–95: Evidence from household
panel data in selected villages. In The New Poverty Strategies. London: Palgrave
Macmillan. pp: 179-202
Dercon, S. and M. Taddesse, 1997. A comparison of poverty in rural and urban ethiopia. Mimeo,
Ebenezer T (2018).Poverty Analysis of Urban and Rural House Hold in South Africa, University
of Cape town.

Etaferahu Cheha and Andualem Goshu (2019). Trends And Determinates Of Multi-Dimensional
Urban Poverty In Ethiopia Master thesis conducted in addis Ababa University
Ezana, H. (2021). Addis Ababa City Scoping Study: The University of Manchester: London:

FDRE. (2018).Poverty and economic growth in Ethiopia 1995/96-2015/16. Planning and

Development Commission.

Ferreira, F. H. G. and Lugo, M. A. (2013). ‗Multidimensional Poverty Analysis: Looking for a

Middle Ground‘. World Bank Research Observer, 28(2): 220–235.
Fleurbaey, M. (2015). Beyond income and wealth. Review of Income and Wealth, 61(2):199–
219.
Filippone, Andrea; Cheli, Bruno; D'Agostino, Antonella (2001) : Addressing the interpretation
and the aggregation problems in totally fuzzy and relative poverty measures, ISER

5
 Working Paper Series, No. 2001-22, University of Essex, Institute for Social and
Economic Research (ISER), Colchester
Fraenkel. J., & Wallen. N. (2003). How to design and Evaluate Research in Education (5th
edition ed.).
Frederik, and Sevias Guvuriro (2021). Gender Differences in Intra-Household Financial
Decision-Making: An Application of C Booysen, oarsened Exact Matching. Journal of
Risk and Financial Management 14: 469. https:// doi.org/10.3390/jrfm14100469
Getinet Haile, 2003, The Incidence of Youth Unemployment in Urban Ethiopia. Paper Presented
Goitom Ghirmatsion, 1996, Aspect of Poverty in the City of Addis Ababa: Profile and Policy
Hayes, & B. E. (1998). Measuring Customer Satisfaction: survey design.
Hendrick Emily, and Leticia Marteleto. 2017. Maternal Household Decision-Making Autonomy
and Adolescent Education in Honduras. Population Research and Policy Review 36: 415–
39. [CrossRef]
Hosmer, D. W., and Lemeshow, S. (1980). A goodness-of-fit test for the multiple logistic
regression models. Communications in Statistics, A10, 1043-1069.
Hulme D. and Mckay A. (2005). Identifying and Measuring Chronic Poverty; Beyond Monetary
Measures. CPRC India Working Paper 30
IFAD 2011. Rural Poverty Report 2011. IFAD (International Fund for Agricultural
Development) Rome, Italy

International Monetary Fund. (2020). Ethiopia: Country report No. 20/29.

J. Deutsch, J. Silber(2005). Measuring Multidimensional Poverty: An Empirical Comparison of

Various Approaches . Available at SSRN: https://ssrn.com/abstract=679138
Jenkins, S. and Mickelwright, J. (2007). New directions in the analysis of inequality and poverty.
Working paper 11, ISER.
John , HA andForrest D.N (1984. Linear probability, Logit and Probit models. SAGE
publications, Inc. U.S.A.
Kassie, G. T., Abate, T., Langyintuo, A., &Maleni, D. (2014). Poverty in maize growing rural
communities of southern Africa, development studies research. An Open Access Journal,
1(1), 311–323. http://doi.org/10.1080/21665095.2014.969844

5
Kebede B. (2019).Determinants of Urban Poverty and Coping Strategies Of Household To
Urban Life In Town Of Ethiopia, https://www.researchgate.net/publication/335380804
Kebede, T., & Sharma, M. K. (2014). The determinant of poverty in Ethiopia. Ethiopian Journal
of Economics, XXIII(1), 113–130

Kothari,C.R.(2004),research methodology; methods and techniques,2nd revised edition, New

Age International (P) Ltd., New Delhi

Kumar, R. (1999) Research Methodology: A Step-by-Step Guide for Beginners. Sage

Publications, London, Thousand Oaks, New Delhi.

Luz, Luciana, and Victor Agadjanian. 2015. Women‘s decision-making autonomy and children‘s
schooling in rural Mozambique. Demographic Research 32: 775–96. [CrossRef]
[PubMed]
M.K. Jayamohan & Amenu Temesgen Kitesa (2014) Gender and poverty – an analysis of urban
poverty in Ethiopia, Development Studies Research. An Open Access Journal, 1:1, 233-
243, DOI: 10.1080/21665095.2014.917053

Mattes, R. (2020). Lived poverty on the rise: Decade of living standard gains ends in Africa.
Afrobarometer Policy Paper No. 62.

Mekonnen B. and Almas H. (2021).A Distributional Analysis of Uni-and Multidimensional

Poverty and Inequalities in Ethiopia Social Indicators Research (2021) 155:805–
835.https://doi.org/10.1007/s11205-021-02606-w.
Mekonnen T. (2002). Perception of Welfare and Poverty. Analysis of Qualitative Responses of a
panel of urban Households in Ethiopia. Ethiopian Journal of Economics, Vol.8 No 1,
Ethiopia.

MoFEC (Ministry of Finance and Economic Cooperation). (2019). 2012 FY federal budget
summary volume 1.

MOFED 2002. Ethiopia: Sustainable Development and Poverty Reduction Program. Ministry of
Finance and Economic Development
MOFED, 2005, Ethiopia: Participatory Poverty Assessment 2004-05, Addis Ababa:

5
Nunnally, J. C. (1967). Psychometric Theory. New York: McGraw-Hill.
OPHI (Oxford Poverty and Human Development Initiative) (2015). ―Ethiopia Country
Briefing‖, in Multidimensional Poverty Index Data Bank, OPHI, University of Oxford.
Available at: www.ophi.org.uk/ multidimensional-poverty-index/mpi-country-briefings
Perry. F. (2005). Research in Applied Linguistics.
Polit, D.E. and Beck, C.T. (2006) Essentials of Nursing Research. 6th Edition, Lippincott
Williams & Wilkins, Philadelphia.
Ravallion, M. (2011). 'On Multidimensional Indices of Poverty'. Journal of Economic Inequality,
9(2): 235–248.
Rawls, J. (1971). A Theory of Justice. Original Edition. Harvard paperback. Harvard University
Press.
Saad, S., Carter, G. W., Rothenberg, M., & Israelson, E. (1999). Testing and assessment: an
employer‘s guide to good practices. Washington, DC: U.S. Department or Employment
and Training Administration.
SDG Center for Africa (SDGC/A). (2019). Africa 2030 Sustainable Development Goals Three-
Year Reality Check. Retrieved from www.sdgcafrica.org

Sekaran, U. (2006). Research Methods for Business: A Skill Building Approach. (4ed.).
NewDelhi: Sharda of Sett Press.

Sen, A.K. (1976). Poverty: an ordinal approach to measurement. Econometrica, 44, 219-231.
Sen, A.K. (1981). Poverty and famines: an essay on entitlement and deprivation. Oxford,
Clarendon Press.
Sen, A.K. (1985). Commodities and capabilities. Amsterdam: North-Holland. Sen, A.K. (1992).
Inequality Reexamined, Harvard University Press, Cambridge, Massachusetts
Sen, A. (2006). Conceptualizing and measurign poverty. In Grusky, D. and Kanbur, R., editors,
Poverty and Inequality. Stanford University Press.
Streeten, P. (1998). Beyond the six veils: Conceptualizing and measuring poverty. Journal of
International Affairs, 52(1).
Tassew Woldehanna and Adiam Hagos. 2013. Dynamics of welfare and Poverty in Poor Rural
and Urban Communities of Ethiopia. An International study of childhood poverty.
Working paper 109: pp. 3-9.

5
Teka, A. M., Woldu, G. T., &Fre, Z. (2019). Status and determinants of poverty and income
inequality in pastoral and agro-pastoral communities. Household-based Evidence from
Afar Regional State, Ethiopia, World Development Perspectives,15(2019),
100123.https://doi.org/10.1016/j.wdp.2019.100123

Tesfaye, A. (2006). The Analysis of Urban Poverty in Ethiopia University of Sydney NSW 2006
Australia. Mrheret, A. (ed), Poverty and Poverty Policy in Ethiopia. Proceedings of the
Workshop Organized by Forum for Social Studies. Addis Ababa
Townsend, P. (1979). Poverty in the United Kingdom. Penguin.
Tsehay, A. S., & Bauer, S. (2012). Poverty and vulnerability dynamics: Empirical evidence from
smallholders in the Northern Highlands of Ethiopia.Quarterly Journal of International
Agriculture, 51(4),301–332.

UNDP. (2018). Human Development Indices and Indicators. 2018 Statistical Update. In
United Nations Development Programme (Vol. 27).
UNDP. (2019). Human Development Report 2019 Inequalities in Human Development in
the 21 st Century. Retrieved fromhttp://hdr.undp.org/en/data
UNEASC, 2007, Urban Poverty and the Working Poor, Facing the Challenges of Urbanization
UNECA 2012. Assessing Progress in Africa toward the Millennium Development Goals, United
Nations Economic Commission for Africa

UN-Habitat (2011). The State of Addis Ababa 2011 Report. Nairobi: UN-Habitat.

UNICEF-2018 :Situation Analysis of Children and Women: Addis Ababa

World Bank (2016).Priorities for ending extreme poverty and promoting shared prosperity,
systematic country diagnostic (Report No: 100592-ET). Washington, DC.
World Bank (2020): Ethiopia Macro and Poverty Outlook, October 2020.
World Bank Institute. (2005). Introduction to poverty analysis. Poverty Manual, All, J. H.
Revision of August 8. World Bank.
World Bank. (2015).Ethiopia poverty assessment overview. Poverty Global Practice Africa
Region.

World Bank. (2020). The World Bank in Ethiopia.

5
WORLDBANK 2008. World development report 2008: Agriculture for development, World
Bank UNDP 2012. Africa Human Development Report 2012: Towards a Food Secure
Future. New York, USA: Regional Bureau for Africa (RBA).
Wynd CA, Schmidt B, Schaefer MA. (2003). Two quantitative approaches for estimating content
validity. West J Nurs Res. 2003 Aug;25(5):508-18. doi: 10.1177/0193945903252998.
PMID: 12955968.

Zikmund, W. G. (2000). Business Research Methods (6th edition). USA: Harcourt.

5
Annex Descriptive Statistics

N Minimum Maximum
DFBETA for constant 398 -.19535 .21311
DFBETA for Sex_HH_Head 398 -.08321 .06894
DFBETA for Maritial_Status_HH_Head 398 -.17581 .12076
DFBETA for Education_HH_Head 398 -.00708 .01680
DFBETA for Employent_Status_HH_Head 398 -.04260 .05477
DFBETA for Unemployed_HHMem_Aged (0-
398 -.06501 .06442
14)
DFBETA for Access_Loan 398 -.23421 .43090
DFBETA for Home_Ownership 398 -.44781 .25109
Valid N (listwise) 398

5
5
5
6
Dear Respondent

This questionnaire was designed to collect information from households reside in Kolfe Sub City
of Addis Ababa City Administration and aimed to analyse"Determinant of Poverty in Addis
Ababa: A Household Level Analysis: In the Case of Kolfe KeraniyoSub City" as a research
subject for the partial fulfilment of the requirements of Master of Art in Regional and Local
Development studies). Your response would have been used only for academic purpose and
kept confidential.
Thank you in advance for your co-operation.
Netsanet Ayalew
MA in Regional and Local Development Studies Students
Email:[email protected]
Tell_0911268770
Addis Ababa University
Addis Ababa, Ethiopia

General Directions:
i. You are kindly requested to give genuine responses.
ii. You don ‗t need to write your identification
iii. Please put a tick (√) in the appropriate box.
iv. Put the numbers you agree with to those questions which are not multiple choices.

Part-One : Question Related to Household Head

A. Personal Characteristics of the Household Head
1. Sex of Household
Head Female Male
2. Age of Household Head
3. Marital Status household
head SingleMarried

6
 4. Year of Education
B. Employment /Occupation
5. Employment Status of Household Head
Employed Unemployed Pensioner

6. If your response is employed to Question 5 what is your main occupation?

Self-employed Government employee Private Employee NGO
employee Daily Labourer

7. If your response is Self-employed to Question 6, which type of self-employed you

are engaged in?
Petty-trade/Gulit Trade Metal /Wood Work
Hotel and restaurant Investor Consultant
Contractor Other Specify

Part -Two: Question Related to Household

C. Household Characteristics
8. Household Size
9. Number of household members aged between 0-14 years
10. Number of household Members aged between 15-64 years
11. Number of household Members aged 65 years and above_
12. Number of unemployed household member aged 14 and above_

D. Income of Household
13. What is your household‘s monthly income (in Birr)?
14. How much is your household‘s monthly expenditure (in Birr)?
15. How much is your household‘s monthly saving (in Birr)?
16. Household member has obtained loan or credit?
Yes No
17. If "yes" to Q. 15 inquire, why you need loan or credit.
Start-up business Housing rent Food Medical Purpose

6
 Education fee Ceremony Other
18. Household own the
house Yes No
19. How many rooms are there in your house?
Part -Three: Question Related to Multidimensional Poverty Indicators
E. Health Dimension ( Child mortality and Nutrition)

20. Any person under 70 years of age for whom there is nutritional information
is undernourished?
Yes No
21. A child under 18 has died in the household in the five-year period preceding the
survey?
Yes No
F. Education Dimension (year of schooling and school attendance)
22. No eligible household member has completed six years of schooling?

Yes No
22. Any school-aged child is not attending school up to the age at which he/she would
complete class 8?
Yes No
G. Living Standard
23. The household cooks with dung, wood, charcoal or
coal? Yes No
24. The household‘s sanitation facility is not improved (according to SDG guidelines
i.e if it has no some type of flush toilet or latrine, or ventilated improved pit or
composting toilet) or it is improved but shared with other households
Yes No

6
25. The household does not have access to improved drinking water (according to SDG
guidelines i.e. if it isn't piped water, public tap, borehole or pump, protected well,
protected spring or rainwater) or safe drinking water but at a 30-minute walk from
home, round trip
Yes No
26. The household has no electricity
Yes No
27. At least one of the three housing materials for roof, walls and floor are inadequate:
i.e. if floor is made of mud/clay/earth, sand or dung; or if dwelling has no roof or walls
or if either the roof or walls are constructed using natural materials such as cane,
palm/trunks, sod/mud, dirt, grass/reeds, thatch, bamboo, sticks or rudimentary
materials such as carton, plastic/ polythene sheeting, bamboo with mud/stone with
mud, loosely packed stones, uncovered adobe, raw/reused wood, plywood, cardboard,
un burnt brick or canvas/tent

Yes No
28. The household does not own more than one of these assets: radio, TV, telephone,
computer, animal cart, bicycle, motorbike or refrigerator, and does not own a car or
truck.
Yes No

6
ውድ መሌስ ሰጪ

ይህ መጠየቅ የተዘጋጀው በሪጅናሌ ኤንድ ልካሌ ድቨልፕመንት ስተዱስ የማስትሬት ዱግሪከፉሌ

ማሟያመመሪቂያ ጽሑፍ ሆኖ በ“የአዱስ አበባ ከተማ የድህንት መኪያዎች ርዕስ በመኖሪያ ቤት ዯረጃ
የተዯረገ በአዱስ አበባ ከተማ አስተዲዯር የኮሌፌ ቀራኒዮ ክፍሇ ከተማ ሊይ ያ ተ ኮረ ጥናት” በኮሌፌ
ክፍሇ ከተማ ውስጥ ከሚገኙ ነዋሪዎች መረጃን መሰብሰብ ነው፡፡ ምሊሾችዎ ሇትምህርት አሊማ ብቻ
የሚውለ እና ሚስጥራዊነታቸው የተጠቀበ ነው፡፡

ሇትብብራችሁ በቅድሚያ እናመሰግናሇን፡፡

ነፃነት አያላው

በሪጅናሌ ኤንድ ልካሌ ድቨልፕመንት ስተዱስ ተማሪ

Email:[email protected]

ስሌክ_0911268770
አዱስ አበባ፣ ዩኒቨርስቲ
አዱስ አበባ፣ ኢትዮጵያ
ጠቅሊሊ መመሪያዎች
i. እውነተኛ ምሊሾችዎን እንዱመሌሱሌን በትህትና እንጠይቃን
ii. ማ ንነት ዎ ን እንዱፅፈ አሌፇሌግም
iii. በተገቢው የመሌስ ሳጥን ውስጥ (√) ያ ስቀ ም ጡ

iv. ምንም የምርጫ ጥያቄዎች ሊሌሆኑት የሚስማሙበትን መጠን በቁጥረ መሌስ ይስጡ ክፍሌ
አንድ፡- ከቤተሰብ ሃሊፉ ጋር ተያያዥ የሆኑ ጥያቄዎች
ሀ. የቤተሰብ አስተዲዲሪ ግሊዊ ባህሪያት
1. የቤተሰብ አስተዲዲሪ ፆታ
ሴት ወንድ
2. የቤተሰብ አስተዲዲሪ እድሜ
3. የቤተሰብ አስተዲዲሪ የጋብቻ ሁኔታ
ያሊገባ ያገባ

6
4. የትምህርት ዯረጃ
ሇ. ስራ ቅጥር/ሙያ
5. የቤተሰብ አስተዲዲሪ የስራ ቅጥር ሁኔታ
ተቀጣሪ ስራ አጥ ጡረተኛ

6. ሇጥያቄ ቁጥር 5 መሌስዎ ተቀጣሪ ከሆነ ዋናው ሙ ያ ምንድን ነው?

በግሌ የሚሰራ የመንግስት ሰራተኛ የግሌ ድርጅት ሰራተኛ
መንግስታዊ ያ ሌሆነ ድርጅት ሰራተኛ የቀን ሰራተኛ

7. ሇጥያቄ ቁጥር 6 መሌስዎ በግሌ የሚሰራ ከሆነ የትኛው አይነት በግሌ የሚሰራ ስራ ሊይ
የተሰማሩ ናቸው ?
የጉሉት ንግድ ንግድ እንጨት/ብረት ስራ
ሆቴሌ እና ሬስቶራንት ባሇሃብት አማካሪ
ተቋራጭ ላሊ ካሇ ይግሇጹ
ክፍሌ ሁት፡ - ከቤተሰብ ጋር ተያያዥ ጥያቄዎች ሐ.
የቤተሰብ ባህሪያ ት
8. የቤተሰብ ብዛት
9. ከ 0-14 አመት እድሜ የቤተሰብ አባሌት ብዛት
10.ከ 15-64 አመት እድሜ የቤተሰብ አባሌት ብዛት
11.65 አመት እድሜ እና ከዚያ በሊይ የቤተሰብ አባሊት ብዛት
12.ከ 14 አመት እና ከዚያ በሊይ ስራ አጥ የቤተሰብ አባሊት ብዛት
መ. የቤተሰብ ገቢ
13.ወርሃዊ የቤተሰብ በወር ገቢ ምን ያ ህሌ ነው (በብር)?
14.ቤተሰቦዎ በወር ምን ያህሌ ወጪ ያወጣሌ (በብር)? _
15.ቤተሰበዎ ወርሃዊ የቁጠባ መጠን ምን ያ ህሌ ነው (በብር)?
16. ከቤተሰብ አባሌ ብድር የወሰዯ ሰው አሇ?
አሇ የሇም
17. ሇጥያቄ ቁጥር 16 መሌስዎ አዎ ከሆነ ብድሩን ሇምን አሊማ ፇሇጉት ንግድ
ስራ ሇመጀመር ሇቤት ግንባታ ሇኪራይ ሇምግብ
ሇህክምና

6
 ሇትምህርት ወጪ ሇክብረበዓሌ ዝግጅት ላሊ
18.የቤተሰብዎ ቤት የግሌ ነው
አዎ አይዯሇም
19. በቤት ውስጥ ምን ያህሌ ክፍልች አለ? _
ክፍሌ ሶስት፡- ርጀ ብዙ የድህንት ጠቋሚዎች ጋር ተያያዥ ጥያቄዎች ሠ.
ከጤና አኳያ (የህፃናት ሞት እና የአመጋገብ ሁኔታ)
20.ማ ንኛ ው ም ከ 70 አመት በታች እድሜ የሆነው ግሇሰብ ሆኖ የአመጋገብ ሁኔታ መረጃ
አስፇሊጊውን መጠን በታች የተጓዯሇው አሇ?
አሇ የሇም
21.ከዚህ ጥናት በፉት ባለት 5 አመታት ውስጥ በቤተሰብ ውስጥ ከ 18 አመት በታች እና
በሞት የተየ ህጻን ነበር?
አሇ የሇም
ረ. ከትምህርት አኳያ (የትምህርት አመታት እና ትህምርት ስሇመከታተሌ)

22.እድሜው ሇትምህርት የዯረሰ የቤትብ አባሌ ሆኖ 6 አመት የትምህርት ጊዜ ያ ሊ ጠናቀ ቀ አ ?

አሇ የሇም
23. ማ ንኛ ው ም እድሜው ሇትምህርት የዯረሰ ህጻን ሆኖ ነገር ግን 8 ክፍሌ ማ ጠናቀ ቀ እያሇበት
ትምህርት ቤት ያ ሌገባ?
አሇ የሇም
ሸ. የኑሮ ዯረጃ
24. ቤተሰቡ ምግብ የሚያበስሇው ኩበት፣ እንጨት፣ ከሰሌ በመጠቀም ነው? አዎ
አይዯሇም
27. የቤተሰቡ የጽዲት መገሌገያ ያ ሌተ ሻ ሻ ሇ (በንጽህና መመሪያዎች መሰረት ማ ሇት ም የውሃ
ማ ው ረጃ ያ ሇው መጸዲጃ ቤት ወይም ባሇ ጉድጓድ መጸዲጃ ቤት ወይም አየር ማ ስወጫ
ማስገቢያ ክፍተት ወይም በመቅበሪያ ጉድጓድ የላሇው አይነት መጸዲጃ ወይም የተሻሻሇ ሆኖ
ከላልች ቤተሰቦች ጋር በጋር የሚጠቀሙበት)
አዎ አይዯሇም

6
28.ቤተሰቡ ንጽህና የተጠበቀ የመጠጥ ውሃ አያገኝም (በኤስጂጂ መመሪያዎች መሰረት ማ ሇት ም
የቧንቧ ውሃ ያ ሌሆነ የመንግስት የቧንቧ መስመር የጉድጓድ ውሃ ወይም በፓምፕ ውሃ
መሳቢያ የሚቀርቡ ውሃ፣ ንጽህናው የተጠቀበ የውሃ ጉድጓድ፣ ንጽህናው የተጠበቀ የምንጭ
ወይም ዝናብ ውሃ ያ ሌሆነ) ወይም ዯህንነቱ የተጠበቀ የመጠጥ ውሃ ሆኖ ነገር ግን ከቤት
ሇዯርሶ መሌስ 30 ዯቂቃ በእግር የሚያስኬድ እርቀት ሊይ የሚገኝ
ነው አይዯሇም
29.ቤተሰቡ የኤላክትሪክ አቅርቦት አያገኝም
አዎ ያገኛሌ
27. ቢያንስ ሇጣሪያ ፣ ግድግዲ እና ወሌ መስሪያ ከተጠቀሙት ሶስት የቤት መስሪያ ቁሳቁስ በቂ
አይዯለም፡፡ ማሇትም ወሇለ ከጭቃ/ሸከሊ አፇር/ አሸዋ ወይም ከእበት የተሰራ ወይም
የመኖሪያ ቤቱ ወይም ጣሪያ ግድግዲ ከእነዚህ አንዲቸው የተፇጥሯዊ የሆኑ እንዯ አገዲ፣ ሸንበቆ፣
ያ ሌተ ጠረበ እንጨት ፣ የጭ ቃ ም ርጊ ት ፣ ሳር፣ ጨርቅ ፣ ቀ ርክሃ፣ ጭራሮ ወይም ተራ እንዯ
ሇካርቶን፣ ፕሊስቲክ፣ ፖሉቲን ፕሊስቲክ ጭቃ እና እንጨት የመሳሰለት ቁሳቁሶችን
በመጠቀም የተገነባ የወዲዯቀ ያ ሌተ ሸ ፇነ ክፍተት ያ ሇው በድጋሚ እንጨት ፣ ጣውሊን በመጠቀም
ወይም በእሳት በዯረቀ ጡብ ወይም ተራ እንዯ ድንኳን ባሇ ቁሳቁስ የተገነባ ነው፡፡

አዎ ኤዯሇም
29.ቤተሰቡ ከእነዚህ ሃብቶች አንደ ወይም ብዙዎቹ የለትም ሬዱዮ፣ ቴላቪዥን፣ ስሌክ፣
ኮፒዩተር፣ የእንሰሳት ጋሪ፣ ባይስክሌ፣ ሞተር ሳይክሌ፣ ወይም ፍሪክ እና ተሸከርካሪ ወይም
ከባድ መኪና የሇውም
አዎ አው