ARBA MINCH UNIVERSITY

COLLEGE OF BUSINESS AND ECONOMICS

DEPARTMENT OF ECONOMICS

Determinants of Rural Households’ saving: a case of Dale Woreda,

Sidama Zone

A senior essay submitted to the Department of Economics in partial

fulfillment of the requirements for the Degree of Bachelor of ART (BA.) in
Economics.

Prepared BY: MelkamuAbebe

ADVISOR: Mubarek . J (Msc.)

May, 2017
Arbaminch, Ethiopia

Page |0
Declaration
This paper is a presentation of my original research work. Whenever contribution of others are
Involved, every effort is made to indicate this clearly, with due reference to the literature;

Information derived from the published or unpublished work of others has been acknowledging
in the text and a list of reference is given. And I declare that this paper has not been submitting
form for another degree or diploma at any university or other institutions of tertiary education.

Name signature date

1.___________________ __________________ ________________

 ADVISORS’ THESIS SUBMISSION APPROVAL PAGE

COLLEGE OF BUSINESS AND ECONOMICS

ARBA MINCH UNIVERSITY

This is to certify that the thesis entitled determinants of rural household saving: A case of
dale Woreda, submitted in partial fulfillment of the requirements for the degree of Under
Graduation with specialization in Economics, the under graduate program of the
Department of Economics and has been carried out by RBE/436/07, under our supervision.
Therefore, we recommend that the student has fulfilled the requirements and hence hereby
can submit the thesis to the department for defense.

By: -MELKAMU ABEBE

Approval of Board of Examiner

Name signature date

1.___________________ __________________ ________________

(Advisor)

2.___________________ ___________________ ________________

(Examiner)

ii | P a g e
AKNOWLEDGEMENT
Above all, I would like to thank the Omniscient, Omnipotent and almighty God he gave me
courage, strength and impatience to finish the job of this study. Next to God, my special
gratitude goes to my advisor Mubarek Juhar (Msc), for his invaluable support and smooth
communication throughout the preparation of this study.

I am also indebted to my honorable family indeed my father Abebe Hankamo, my mother Aster
Kebede and my lovely brother Adisu and sisters for their moral and financial support. I would
like to thanks my friends Ashagre Alemu, Melkamu Beyene Dawit Datchie, and Markos Mesfun
for their valuable comment and material support to accomplish this study.

Finally, I am very grateful to Dale Woreda microfinance office and households for their
contribution in providing me all the necessary material and relevant information.

ABSTRACT
The study aims at investigating the determinants of rural households’ saving in Dale district
Sidama Zone. Data of 100 respondents are drawn through field survey in 2016/17 by adopting
iii | P a g e
multistage random sampling technique. Questions are asked directly from head of household
about their education level, family size, age, amount of savings per mouth in birr, land size,
income etc. Sample contains information about rural households. The study employed both
descriptive and econometric method of data analysis. It employed censored regression model,
that is, the Tobit model to investigate the main determinants of household saving in Dale
Woreda. Accordingly, it is found that land size and income positively significantly affects
household saving. However, in order to see the impact of years of schooling and family size
negatively significantly affects household saving at five level of significantly. Education has a
negative impact on household savings mainly because of the awareness that occurs with lower
educational levels and at coffee growing time the household head withdraw from education.
Provided awareness how to encourage or facilitate education with productivity in coffee growing
time to goes together in rural area.

Table of Contents
AKNOWLEDGEMENT................................................................................................................iii

ABSTRACT..................................................................................................................................iv

List of Table..................................................................................................................................vii

iv | P a g e
 Acronyms and Abbreviation..........................................................................................................ix

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

Introduction....................................................................................................................................1

1.1 Background of the study....................................................................................................................1

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

1.3. Research questions of the study........................................................................................................4

1.4 Objective of the Study.......................................................................................................................4

1.4.1 General objective............................................................................................................................4

1.5 Significance of the study...................................................................................................................4

1.6 Scope of the study..............................................................................................................................5

1.7 Limitation of the study.......................................................................................................................5

1.8 Organization of the study...................................................................................................................5

CHAPTER TWO............................................................................................................................6

REVIEW OF RELATED LITRATURES......................................................................................6

2.1 Theoretical Literatures.......................................................................................................................6

2.1.1 Definitions of saving..............................................................................................................6

2.1.2 Theoretical evidences about the determinants of household saving patterns..........................7

2.2 Empirical Literatures.........................................................................................................................9

2.3 Research gaps..................................................................................................................................10

CHAPTER THREE......................................................................................................................11

METHODOLOGY ANDDATA SOURCES OF THE STUDY...................................................11

3.1 Description of the study Area..........................................................................................................11

3.2 Methodology...................................................................................................................................11

3.2.1 Model specification.............................................................................................................12

3.2.2 Method of Estimation..........................................................................................................14

3.3 Data type and Data sources of the study..........................................................................................15

v|Page
 3.3.1 Sample size and sampling technique....................................................................................15

3.3.2 Method of Data collection...................................................................................................17

CHAPTER FOUR........................................................................................................................18

DATA ANALYSIS AND DISCUSSIONS OF RESULTS...........................................................18

4.1 Descriptive analysis.........................................................................................................................18

4.1.1 Background of the respondents...........................................................................................19

4.1.2 Socio-economic and demographic characteristics of households.........................................19

4.1.2 Description of the values of the variables of the model.......................................................22

4.1.3 Descriptive analysis of the effects of socio-economic factors on the rural household saving
......................................................................................................................................................23

4.2 Econometric Analysis......................................................................................................................26

4.2.1 Results of the estimated model............................................................................................26

4.2.2 Diagnostic test results of the model.....................................................................................27

4.2.3 Interpretations and discussion of the results of the model....................................................27

CHAPTER FIVE..........................................................................................................................30

CONCLUSIONS AND RECOMMENDATIONS........................................................................30

5.1. Conclusions....................................................................................................................................30

5.2. Recommendations..........................................................................................................................30

REFERENCES.............................................................................................................................32

APPENDEX.................................................................................................................................33

List of Table
Table 3.1 Descriptions and hypothesis of variables......................................................................18
Table 3.2 The distribution of Megera and Wenenata household head in Dale Woreda................20
Table 4.1: Socio-economic and demographic characteristics of households................................23
Table 4.2: Income, Saving and APS of household head by Gender..............................................25

vi | P a g e
Table 4.2: Summary of the values of major variables of the model for sample saving of
households in dale Woreda.....................................................................................................27
Table 4.3: Correlation matrix among variables of the model........................................................29
Table 4.4: Mean saving differences of financial institution customer attracts and marital status
beneficiary in current situation saving, expect full business idea, formal financial institution
and attitude.............................................................................................................................30
Table 4.5: Tobit Estimates of Household saving...........................................................................34

Acronyms and Abbreviation

1. GDP…………………….Gross Domestic Product
2. LDC’s……………………Least Developed Countries
3. MOFED…………………..Ministry of Finance and Economic Development
4. OLS………………………Ordinary Least Square
5. VIF……………………….Variance Inflating Factors
6. WB………………………..World Bank

vii | P a g e
7. CSA……………………….Census Statistical Agents
8. SNNPR…………….……. Southern Nation’s Nationality and Peoples Region
9. SLH……………………..Size of land holdings

viii | P a g e
 CHAPTER ONE

Introduction

1.1 Background of the study

Saving is one among important variables for economic growth of a country. Saving is income
that is not consumed immediately by buying goods and services (Prinsloo, 2000; Manyama,
2007). Saving constitutes the basis for capital formation, investment and economic growth of a
country (Nga, 2007; Nwachukwu and Odigie, 2009).

According to Deaton (2005) and Rogg (2006) serious problem confronting poor countries
including Ethiopia is the large saving investment gap. Because of this gap, countries find it
difficult to finance investments needed for growth from domestic saving. It is also common to
see these countries to finance their investment in the short run partly through domestic
government borrowings and/or foreign loan and grants but this would significantly increase the
country’s debt burden and would not be a solution in the long run.

Adam Smith, the father of economics did talk about growth and to him it was governed basically
by capital accumulation and division of labor, and capital accumulation in turn is governed by
the level of domestic saving. Harrod and Domar said that in order to grow economies must save
and invest a certain proportion of their Gross Domestic Product (GDP). Hence, the more they
can save and invest; the faster they can grow (Todaro, 2000). According to Meier (2005) cited in
Todaro (2005), these countries experienced low level of income from which saving is not
expected to finance investment, which in turn leads to low productivity, which is equivalent to
low level of income. Thus, the low level of saving becomes both the cause and consequence for
the poverty they are in.

Private savings can be mobilized through by banking system, development of financial

institutions and the stock exchange. In private savings, large portion of savings is due to large
share of household sector. Though there is an improvement in domestic savings rate in Ethiopia,
it is still at a low level. The prevailing domestic saving rates to be commensurate with the
envisaged huge investment requirements of the country. It is, therefore, imperative that concerted
efforts has to be made to enhance domestic saving (MoFED, 2012).

Household saving is usually the largest component of domestic savings in developing countries,
like Ethiopia (Todaro 2003). This contrasts with the much greater importance of corporate saving
in developed countries. The ability, willingness, and opportunity of households to save over time
can therefore significantly influence the rate and sustainability of capital accumulation and
economic growth in developing countries (Bautista and Lamberte,2000).

1.2 Statement of the Problem

Saving is a very important component which is responsible for combating or meeting any
emergency accrued by the individuals or the households or any corporate agencies. Saving is
more of meant for meeting contingencies but sometimes it also acts as a form of investment. But
sometimes people are not inclined towards saving and the very delicate reason is lack of
awareness. The present study can be a relevant one to know the reason of dis-saving and if
saving occurs then what are the determinants which are responsible for saving. Aggregate saving
in any economy is dependent on a number of variables. For effective economic planning, the
planners should have an idea regarding the volume of saving of different groups of people and
the method by which saving can be improved more over in a better way. To advocate appeals for
saving, there is a need to know about the saving motives of the individuals. An understanding of
the saving preferences also helps in calculating the saving instruments which can efficiently
arouse saving. Dale is a Woreda having a very poor access to the saving need which really has
made a great interest in the minds of looking at the perspective as a whole.

Like other households of rural Ethiopia, most coffee producing rural households in Sidama Zone
spent their income received from the sale of coffee extravagantly (Alemayehu. R 2003).This let
most coffee growing rural households in the zone under the chain of poverty as they lack enough
amount of saving to finance investments needed to improve their productivity in the future. To
formulate an effective policy to break the low saving behavior of these households in the zone
requires the identification of factors that explain the low saving behavior of households in the
zone.

2|Page
Household’s decision to save can be affected by several factors which are internal and external to
the household. Factors which are internal to the households among others include income of the
household, family size, gender and age composition of the household and educational attainment
the household head, size of the land owned by the household, attitude of the household head
toward saving, expected gain from saving and investment by the household head. On the other
hand factors which are external to the household among others include availability of saving
institutions and saving derivatives, the inclusiveness and the level of households’ satisfaction
from the service delivery of the saving institutions and so on.

Previous studies show that households in Sidama zone may have different saving behavior. The saving
level in Ethiopia particularly in rural areas is very low and little is known empirically about its
patterns and determinants. Savings in rural Ethiopia is mainly made out of the income from
agricultural activities. It is also characterized as seasonal and irregular as the cash flow through
sale of agricultural produce and availability of work is seasonal. This reduces their financial
capacity to save or poorly respond to incentives that promote savings in the country (Dejene,
2003; EEA, 2009; EEA, 2009). However, rural households do indeed save in the form of
tangible assets and/or in financial forms which can be potentially utilized by savings institutions
and for investments which is very essential for both households and national wellbeing (Dimova
and Sen, 2010; Karlan and Morduch, 2010).

Previously this saving behavior for the households was dealt at Macro level (see Abu, 2004).
This matter needs to be discussed more at micro level. So that savings in household or micro
level may be analyzed. The micro-evidence on the state of household savings and access to credit
indicates that, particularly in rural Ethiopia, savings in the form of cash is hardly a common
practice (Alemayehu, Abebe and Daniel, 2006). Besides, there isnostudy conducted on
microeconomic level on the determinants of household saving in Dale Woreda, Sidama Zone
Southern Ethiopia and limited studies are found in the country.Therefore, this paper attempted
objectively to identify major micro level determinants of savings at household levelfocusing on
the effects of the socio-economic characteristics of the households on saving behaviors and their
view on income, expenditure and credit institutions.

Considering this problem, it is necessary to study behavior and determinants of household

savings of the Zone and suggest some policies at micro level. As a matter of fact, Sidama Zone

3|Page
one of Southern Nation’s Nationality and Peoples Region (SNNPR) has been given little
attention in the past, because the largest part of this Zone is rural area. Generally, few are
interested to analyze saving behavior in woredas of this area. That is why I have selected Sidama
Zone as my study area. Therefore this study is initiated to address such gap, in the coffee
growing area of the Zone.

1.3. Research questions of the study

The study will address the following research question:

 What is the magnitude of coffee growing households saving in Dale Woreda?
 What are the major household level and ex-household level factors that explain variation
in the amount of saving of coffee growing households in Dale Woreda?
 Do inclusiveness and service delivery of the traditional financial institutions significant
factor in affecting the saving level of coffee growing households in Dale Woreda?

1.4 Objective of the Study

1.4.1 General objective

The general objective of the study is to identify major demographic and socio-economic factors
that affect coffee growing households saving in areas Dale Woreda, Sidama Zone.

Specific objective
 To estimate the average level of coffee growing households saving in Dale Woreda
 To determine the direction and extent of effect of major household level and ex-
household level factors that explain variation in the amount of saving of coffee
growing households in Dale Woreda.
 To determine whether financial inclusiveness and service delivery by formal financial
institutions affect the saving decision of households in Dale Woreda.

1.5 Significance of the study

The findings of the study may initiate other interested bodies to undertake a better and detailed
study in this area. Besides it may used as an input by policy makers to policies and strategies
with which they can improve major household level and ex-household level factors that explain
variation in the amount of saving of coffee growing households in Dale Woreda.

4|Page
1.6 Scope of the study
The study is a cross-sectional analysis to identify the major factors which affect coffee growing
households saving in dale Woreda. It is confined on uncovering how and to what extent
household and ex-household level factors such as land size, financial institution and education
level affect major household level and ex-household level factors that explain variation in the
amount of saving of coffee growing households in Dale Woreda.

1.7 Limitation of the study

The difficulty in obtaining quality data from different sources is one of the limitations of the
study. Because of shortage of time, data was not collected as much as needed. Therefore, the 312
sample size was taken as calculation by sampling techniques but 212 of the questionnaires
excluded or omitted by the case of time limited. Thus only 100 of the questionnaires filled and
returned by the sample households.

In doing this research paper the researcher face the following problem

 The time and finance constraints.

 Difficulties of having reliable data.

1.8 Organization of the study

The study will be organized in to five chapters; chapter one deals with introduction parts that
include background of the study, statements of the problem, and objective of the study,
significance, scope, and organization of the study .The second chapter deals with both theoretical
and empirical related review of literature. The third chapter deals with the methodology of study
which includes description of the study area, type of data, method of data collection, sampling
technique, sampling size and method of data analysis. The fourth chapter deals with data analysis
and discussion. Finally chapter five deals with the conclusion of finding from the discussion
section and relevant policy recommendation.

5|Page
CHAPTER TWO

REVIEW OF RELATED LITRATURES

2.1 Theoretical Literatures

2.1.1 Definitions of saving

Saving has been defined differently by various. According to Wikipedia (2011)
saving is income not spent, or deferred consumption. Methods of saving include putting money
aside in, for example, a deposit account, a pension account, an investment fund, or
as cash. Saving also involves reducing expenditures, such as recurring costs. In terms of personal
finance, saving generally specifies low-risk preservation of money, as in a deposit account,
versus investment, wherein risk is a lot higher; in economics more broadly, it refers to any
income not used for immediate consumption.

Saving is closely related to physical investment, in that the former provides a source of funds for
the latter. By not using income to buy consumer goods and services, it is possible for resources
to instead be invested by being used to produce fixed capital, such as factories and machinery.
Saving can therefore be vital to increase the amount of fixed capital available, which contributes
to economic growth.

However, increased saving does not always correspond to increased investment. If savings are
stashed in or under a mattress, or otherwise not deposited into a financial intermediary such as
a bank, there is no chance for those savings to be recycled as investment by business. This means
that saving may increase without increasing investment, possibly causing a short-fall of demand
(a pile-up of inventories, a cut-back of production, employment, and income, and thus
a recession) rather than to economic growth. In the short term, if saving falls below investment,
it can lead to a growth of aggregate demand and an economic boom. In the long term if saving
falls below investment it eventually reduces investment and detracts from future growth. Future
growth is made possible by foregoing present consumption to increase investment. However
savings kept in a mattress amount to an (interest-free) loan to the government or central bank,
who can recycle this loan.

6|Page
In a primitive agricultural economy savings might take the form of holding back the best of the
corn harvest as seed corn for the next planting season. If the whole crop were consumed the
economy would convert to hunting and gathering the next season.

Within personal finance, the act of saving corresponds to nominal preservation of money for
future use. A deposit account paying interest is typically used to hold money for future
needs, i.e. an emergency fund, to make a capital purchase (car, house, vacation, etc.) or to give to
someone else (children, tax bill etc.), Wikipedia (2011).

In economics, savings are defined as income minus consumption. The rate at which people can
be expected to do this is called the marginal propensity to save or average propensity to
save. The rate of savings is directly related to both the interest rate and investment, largely by
way of the capital markets. It is worth noting that some investment is considered savings. If
investment merely replaces depreciated capital stock, rather than increasing the capital stock and
workforce, it is still considered part of savings.

2.1.2 Theoretical evidences about the determinants of household saving patterns

In developing countries, saving is difficult to capture as it can be raised on an informal basis and
as a result, it cannot be completely assessed by the national accounts. On the other hand, in
OECD a country saving is largely made up of property investments, monetary and financial
investments (Schmidt-Hebbelet al., 1996; Bisatet al., 1997; and Sinha, 1999).In developing
countries, households hoard money. This is because these savings are perfectly liquid so they can
be used to face any urgent need or investment opportunity. This becomes all the more important
since households’ confidence in the banking system is low. Moreover, non-financial saving is
important in developing countries. It can take various forms as precious or semi-precious
materials (jewels, carpets, etc...), Robinson (2004).

In Ethiopia, precious or semi-precious materials are accumulated on a regular basis and are
exchanged against liquidities in order to meet lifecycle (education, marriage, immigrations, etc.)
or urgent spending. The non-financial household saving also consists on housing properties and
other forms of ownership (land, livestock, machines, etc.). Robinson (2004) adds building
materials, cereals and harvest to these main forms of savings. More generally, this kind of saving
accounts for a large part of households’ saving. According to Goldstein and Barro (1999)“one of

7|Page
the essential characteristics of non-financial savings is to be able to be easily used in case of
social need or economic opportunity. For cereals stocks or livestock purchases, can add high
motivation of economic profitability”. Therefore, livestock accumulation is a source of profit.
Livestock can be easily sold; some of them produce other consumable and tradable goods (eggs,
milk, wood, etc.) or can be used as agricultural inputs. Nevertheless, this form of savings present
some drawbacks: cattle breading requires resources like water, animal food, pasture, work-time
and can be lost in the case of illness or natural disasters.

Various economic literatures identify a large number of motives for household savings, most of
them derived from two consumption theories: the permanent income hypothesis and the life
cycle hypothesis. Schmidt-Hebbel et al.(1996) discuss the saving determinants ineach specific
theory (which are opposed as far as the sign of some determinants is considered) and how they
are relatedto empirical findings. Among these motives, the most often recurred are the
precautionary behavior, life-cycle considerations, investment opportunities, the preference for
smooth consumption, the need to accumulate resources for large purchases and the bequest
reason.

The permanent income hypothesis predicts that an unanticipated increase in the future income
relative to the current income reduces current savings in contrast to the Keynesianpoint of view.
Most of the empirical studies (Hall, 1978 and Flavin, 1981)Foundthat consumption exhibits
“excess sensitivity” to a change in income.

From the macroeconomic perspective, many empirical studies, both in developed and developing
countries, Investigate the determinants of private saving rates in order to explain the diversity in
saving rates in the world. Losayzaet al.(2000) reviewed drivers of private saving and classified
into demographic and economic variables. The demographic variables include old or young age
and urbanization. Whereas the economic variables include income (temporary/permanent),
uncertainty (political instability), rates of return (interest rate, inflation...), domestic and foreign
borrowing constraints, fiscal policy and pension system. Various model specifications related to
data samples and econometric strategies are also suggested. However, these literatures provide
ambiguous results. Numerous saving determinants are not significant and/or the estimated sign is
not consistent with the theory. A case in point here is the sign of the income level. Moreover, a
large body of empirical macroeconomic work ignores consumer heterogeneity by assuming a

8|Page
representative household agent. These macroeconomic studies cannot deal with “real-world”
features that reflect the diversity of saving behavior. On the other hand, micro-econometric
analysis allows estimating the importance of economic variables and the role of households’
features in the saving behavior. This study tries to keep track with this empirical research field.

2.2 Empirical Literatures

Econometric research on the determinants of household saving based on micro data drawn from
the less developed countries has lagged far behind the pace set in advanced nations. It would
appear that there has been limited hypothesis testing in the LDC's beyond macro formulations of
the consumption function. Furthermore, very little of the development literature attempts to
isolate the impact of structural change on aggregate personal saving, since few studies provide
meaningful disaggregation (Kelley and Williamson, 2009). This state of affairs seems
paradoxical, given the currency of W. A. Lewis's remark that the central problem in development
theory is to explain an increase in domestic saving from 4 or 5 percent of national income to 12
or 15 percent (Lewis, 1954).

Besides, few studies assess the determinants of saving at the individual level generally due to the
lack of data. Using recent econometric techniques, Carpenter and Jensen (2002) and Kulikov, et
al.(2007) identify how household characteristics affect saving behavior, in Pakistan and Estonia
respectively. Carpenter and Jensen (2002) focus on the role of institutions which collect saving
and stress on the role of formal (banks) and informal institutions (savings committees). They
found that “increased income leads to a greater desire to participate in some form of savings
institutions but as income increases more individuals shift to the formal sector”. They also found
evidence that the urban-rural differences in bank use is negligible which suggests that formal
finance is not primarily restricted to urban households in Pakistan. As opposed to Carpenter and
Jensen (2002) who focus on the savings supply side, where as Kulikovetal.(2007)analyze the
saving determinants on the demand side. Making a distinction between regular and temporary
household income allows the authors to put forward the role of income variability and the
different forms of household assets (financial and non-financial) in a transition economy
(Estonia). Their analysis is based on data from household budget surveys. As in many empirical
studies, they found that the saving rates depend more on the transitory income than regular
income. Among the other variables, the labor market status or the non-financial assets ownership

9|Page
(real estate for instance) and credit access have not significant effect on the household saving
behavior; the durable goods possession (in particular cars) has a negative impact on the saving
rate.

Among the few researches done in developing countries; Klauseet al. (1992)studied households
saving in developing countries and found that income and wealth variables affect saving
strongly. Touhamietal.(2009) also investigate the micro-econometric determinants of households
saving in Morocco. They concluded as income significantly explains the cross-sectional variation
of the saving behavior of households in Morocco. Similarly, Girmaetal.(2013) identified
determinants of rural household savings in East Hararghe Zone, Oromia Regional State Ethiopia
etal.(2013) identified. Nine determinant explanatory variables of rural household savings were
identified which include: household head education level, livestock holdings, access to credit
service, income, investment, training participation, contact with extension, forms of savings and
saving motives.

2.3 Research gaps

The empirical literature review revealed that there are different factors that affect household
savings. Most of these empirical studies focus on aggregate national savings using macro data.
Besides, there is no study conducted on microeconomic level on the determinants of household
saving in Dale Woreda, Sidama Zone Southern Ethiopia and limited studies are found in the
country. Therefore, this paper attempted objectively to identify major micro level determinants
of savings at household level focusing on the effects of the socio-economic characteristics of the
households on saving behaviors and their view on income, expenditure and credit institutions.
The study is also intended to contribute to the existing research gap through a better exploration
of its determinants.

10 | P a g e
 CHAPTER THREE

METHODOLOGY ANDDATA SOURCES OF THE STUDY

3.1 Description of the study Area

Dale Woreda is one of the Woredas in the Sidama zone in the Southern Nations Nationalities and
Peoples' Regional State (SNNPR) of Ethiopia. It located 220 kilometers south of Addis Ababa
the capital of Ethiopia and 45 kilometers south of Hawassa the capital of SNNPR.

The elevation of the woreda varies from about 1200 meters above sea level along the shores of
Lake Abaya to about 3200 meters at its westernmost point. Rivers include the Gidabo. A 2004
survey of the land in Dale shows that 81.9% is arable or cultivable, none used for pasture, 2.7%
forest, and the remaining 15.5% is considered swampy, degraded or otherwise unusable. The
same survey reported important cash crops for Dale include corn, barley, haricot beans, local
varieties of cabbage, and sweet potatoes. Coffee is also an important cash crop in Dale, with
15.38 square kilometers planted with this crop, which produced a total of 9.3 million kilograms
of beans in 2002/03 (5.7 million kilograms in 2003/04). Industry in this woreda includes 57
coffee pulpers.

Based on the 2007 Census conducted by the CSA, this woreda has a total population of 242,658,
of whom 122,918 are men and 119,740 women; 30,348 or 12.51% of its population are urban
dwellers.

3.2 Methodology
The study was used both descriptive and econometric methods of data analyses. Descriptive
methods would be to address the first research question of the study that is estimating the
average saving level of coffee growing households and to summarize the demographic and
socio-economic characteristics of the sample respondents. Econometric method of analysis on
the other hand would be used intensively to address the rest research questions of the study that
is to determine major household level and ex-household level factors which determine the saving
level of coffee growing households in the study area. This is for the fact that descriptive method
of analysis is dubious to make causal-effect type of analysis. To this end a multivariate

11 | P a g e
econometric model relating the level of saving with various characteristics of the household and
other potential external variables which may affect household’s decision to save would be
specified and estimated using appropriate econometric method of estimation.

3.2.1 Model specification

The study was used a multivariate-Tobit model to determine major household level and ex-
household level factors which affect coffee growing households saving in the study are. Because
OLS is both biased and inconsistent when the dependent variable of a model assumes large
number of zeros either for data censoring or truncation problems.(Gujarati,2004). In this study
the dependent variable, household saving, can assume large number of zero values as a result of
households optimal decision not save. Thus, Tobit model is considered as appropriate model than
OLS. The mathematical form of the Tobit model is given by:

Si*= Xiβ+μi, i= 1, 2,....n

Si = Yi* if Si* > 0

Si = 0 if Si* ≤ 0

Si*=Xiβ+ui................................................................................ (3.1)

Where:Si : the observed amount of household savings

Si* : is the latent variable which is not observed

β is Vector of unknown parameters

Xi is vector of independent variable affecting household savings.

Dependent variable is households saving- Si

Independent variable is gender of household head, age of household head, education level of
household head, family size, dependency ratio, land size, , income source of households,
financial institutions, Attitude of the household head about saving, marital status and interest
rate. Thus, the model for the main determinants of household saving can be specified as follows;

12 | P a g e
Si= α + β1Ii + β2AGE +β3GENDER +β4DR+β5FSi+β6LS + β7EDUC +β8AFS
+β9ATTITUDE+ β10IRi+β11MS + β12ExpGfS + β13ExpGfPI
ui…………………….………… (3.2)
Where, α is a constant term and uiis the error term.

3.2.1.1 Descriptions of the variables and expected signs of the parameters of the
model
Si is the household saving takes the values zero for the substantial part of the population and
positive continuous values for the rest of the population. S is a dummy variable which can
assume that saver and non saver (or 1 and 0 respectively).

Ii is personal disposable income of i th of household head. As disposable income of household

increases by 1 birr it will result in an increment in household savings. Income is a continuous
variable and positively affected.

AGE is the age of household heads that are in the middle age (24-64) save more than household
heads that are in the early age and old age. AGE1, 1 if the i th household head age is between 18-
24 years, 0 otherwise. AGE2, 1 if the ith household head age is between 24-65 years, otherwise 0.
AGE3, 1 if the ith household age is >65 years, 0 otherwise. As age of household increases by 1
year it will result in an increment in household savings. Age is a continuous variable which can
assume any age between 24 and >65 years.

FS is the family size increasing by 1 children decreasing the household saving rate in the
significant level. FS is negatively affected of the household saving. FS is a continuous variable.

Gender1 if the ith household head is female 0 otherwise

EDUC1 1, if the household head is illiterate 0, otherwise EDUC2 1, if the household head has
primary education 0, otherwise EDUC3 1, if the household head has secondary education 0,
otherwise EDUC4 1, if the household head has tertiary education 0, otherwise

13 | P a g e
Table 3.1Descriptions and hypothesis of variables

S.No Variable name Description Hypothesized effect

1 Income Continuous +

2 Age Continuous

3 Family Size Continuous -

4 Education Level Continuous -

5 Gender Dummy Indeterminate a priori

6 Forma Financial institution Dummy -

7 Marital Status Dummy Indeterminate a priori

8 Land size Continuous +

9 Dependency Ratio Continuous +

10 Beneficiary current situation saving Dummy -

11 Attitude of household Dummy +

12 Expect full business idea Dummy -

13 Financial institution customer attract Dummy -

Source: Own test (2016/17)

3.2.2 Method of Estimation

The model is estimated using cross-sectional data we may encounter problem of
heteroscedasticity. In order to correct the heteroscedasticity problem we can estimate the robust
standard errors instead of the usual standard errors (Wooldridge, 2002). Thus, the Tobit model
which is used in this study is corrected for heteroscedasticity problem using the robust command
in Stata (robust standard errors are estimated for the Tobit) and from correlation matrix (pair-
14 | P a g e
wise correlation coefficient analysis), it is observed that there is no multicolinearity problem
among independent variables. Therefore, from correlation matrix generated using the survey data
it is shown that there is no series multi colinearity problem in this study.

3.3 Data type and Data sources of the study

To address the research questions raised by the study a cross-sectional primary data set is used.
Secondary data about the total number of coffee producing households in Dale Woreda is also
used to determine target population and the size of sample for survey. On the other hand sample
survey would be conducted to obtain appropriate information about the dependent and
independent variables of the model.

3.3.1 Sample size and sampling technique

The target population is the total number of coffee producing household’s in Dale Woreda. In
Dale Woreda there are 32 kebeles. From out of 23150 households of coffee producing 1655
households of coffee producers are selected, Megara and Wenenata kebeles in 2016/17 Census
conducted by the CSA. So that, the study would be conducted only two kebeles. The other
kebeles are excluded from the survey because of this two kebeles are more coffee producing .
Multi-stage sampling method would be used to select the sample. Within Sidama zone
purposively Dale Woreda select and then Megara and Wenenata kebeles. Since these two kebeles
has approximately equal size of households which is about 830 and 825 respectively. A random
sample of 156 householders finally select from each of the two kebeles by proportionally. Since
the sampling procedure was fairly random, the samples adequately represent the targeted
populations in the area of study.

Table 3.2 The distribution of Megera and Wenenata household head in Dale Woreda

15 | P a g e
 Woreda Kebele No. households Sample proportion Sample size

Dale Megera 830 50% 156

Wenenata 825 50% 156

Total=1655 Total=100% Total =312

Source: Dale Woreda the CSA office (2016/17)

To determine the sample size of the survey the following tow formula as suggested by Cochronis
(1963) Taro Yammen (1967: 886) are would be used jointly. The first formula is used to
determine the adequate size of the sample for a given amount of population proportion and
sampling error (level of significance). After the adequate sample size for a given values of
population proportion (P) and sampling error (e) has been determine the second formula as
suggested by Yammen(1963) is used to determine the exact sample size of the survey
considering the total number of the target population.

First Cochronis (1963) is applied to determine the adequate size of the sample for a given value
of P and e as follows:

no = Z2P(1-P)/e2

Where: Z is Z– value at the chosen Level of confidence (sampling error e)

no -adequate sample size at the chosen Level of confidence and population proportion

e –margin of error (sampling error)

P -population proportion

To arrive at adequate sample size 5% level of significance (margin of error) and 0.5 values for
the Population proportion are will be use. That is 5% risk of committing sampling error and
symmetrically distribute population are assume in the sampling procedure. The Z value at the
chosen level of significance (5%) is 1.96. Then plugging these values into the above formula the
adequate size of the survey is determined as follows:

no = (1.96)2 0.5 (1-0.5) /(0.05)2

16 | P a g e
 = (3.8416) (0.25) /0.0025

= 384

Then, using Yammen formula and 1655 size of the population the exact sample size of the
survey is determined follows:

ns=no/1+no /N

Where: ns is the sample size

Nis the total population of the target group

no is the error term

ns=no/1 +no /N N=1655 e=5%

ns=384/1 + 384/1655

=312

3.3.2 Method of Data collection

To collect the necessary raw data for the study a self-administered questionnaire would be
distributed to households included in the survey. Also the primary data would be collected via
questionnaires in March 2017. The questionnaire comprise, among others, household
characteristics, monthly and/or annual income, wealth in its various forms, location (area of
residence) of the questionnaires interest rate, absence or presence of financial
institutions/intermediaries, financial management habit and knowledge of respondents, which are
considered to be important variables that affect household saving behavior on a priori theoretical
ground.

17 | P a g e
 CHAPTER FOUR

DATA ANALYSIS AND DISCUSSIONS OF RESULTS

In this chapter major results about the determinants of rural household saving in Dale Woreda
obtained from both descriptive and econometric methods of data analyses would be presented.
Besides, some discussions on major findings of the study would also be made. This chapter
proceeds as follows. In the first section of this chapter summaries of results obtained from
descriptive method of data analysis would be presented. In this regard, in the first section
summaries of the response rate of the questionnaire distributed, the backgrounds of the
respondents and the descriptions of the independent variables of the model would be presented.
In addition to these the relationship of the independent variables with the dependent variable
would be described using descriptive method of analysis. The second section would present
results obtained from econometric method of data analysis. In that part the results of the
estimated model, diagnostic test results of the model and discussions on major findings of the
study would be presented.

4.1 Descriptive analysis

To get appropriate data required addressing the research questions of the study a survey of
saving and some selected factors of saving of rural households has been conducted taking 312
randomly and proportionately selected rural households from two kebeles (Megera and
Wenenata) in Dale Woreda. A self administered structured questionnaire with a total of 16
questions in it has been distributed to the 100 sample rural households included in the survey. Of
the 100 questionnaires distributed 100 of the questionnaires filled and returned by the sample
households. This implies more than 100% response rate has been achieved. Thus the analysis is
based on the responses of the remaining 100 of the sample households whose responses have
been found consistent when they cross checked against the control questions. This is to minimize
the adverse effect of wrong information on the analysis. The results from the responses of the
100 sample rural households included in the survey are summarized and presented below.

18 | P a g e
4.1.1 Background of the respondents
For the above case, a summary of age, gender, years of schooling, marital status of households,
family size, land size, income of households, dependency ratio, the beneficiary in the current
situation of saving, expect gain from full business idea, financial institutions attract of customers,
formal financial institution in area and attitude of about saving other socio-economic,
demographic and family backgrounds of the respondents that you think as necessary would be
summarized and presented in a single table. Then short and précis elaboration of the summaries
would also be given in the following table.

19 | P a g e
4.1.2 Socio-economic and demographic characteristics of households
Table 4.1: Socio-economic and demographic characteristics of households

Variables Obs Average Min Maxi

saving(birr)

Gender M 83 173.63 0 1400

F 17 231.17 0 1000

Age (24-45) 62 185.046 0 4000

(46-55) 25 209 0 1400

(56-64) 8 144.325 0 350

>64 5 37.4 0 350

Years of Illiterate 14 136.07 0 1400

schooling
Pre.edu(1-8) 43 204.81 0 1000
(Education
level) Sec.edu(9-10) 17 176.47 0 900

Ter.edu(11-12) 17 166.4 0 600

Higher.edu(>12 9 200 0 800

)

Income (0-500) 3 36.6 - 475

(500-1000) 23 55.21 - 1000

(100 0-2000) 59 161.72 - 2000

(2000-3000) 13 484.61 - 3000

>3000 2 1200 - 4000

Source: Own Survey, 2016/17

20 | P a g e
In this study, it is found that, household heads that are in the middle age (24-64) save more than
household heads that are in the early age and old age. The mean saving of middle age, early and
old age household heads is about Birr 185.046, 206.206 and 144.345 per month respectively.

Another important determinant of household saving is educational level of household heads. This
is because of the fact that, as the level of education increase the awareness of households
concerning saving also increase. Table 4.1. Showed that mean saving of households with higher
educational level on average save more than households with no or lower educational level. The
mean saving of illiterate household heads is Birr 136.07 where as household heads with primary
education, secondary education, tertiary education and higher education level (university) on
average saves Birr 204.81, Birr 176.47, Birr 166.4 and 200 per month respectively. Hence, as the
educational level increases, the average household saving also increase and this finding is
consistence with the empirical results of other researchers.

Gender is identified as an important variable in the household saving behavior. This by gender
analysis is relevant because women are usually expected to save more part of their disposable
income than men do. However, the finding showed that women save more than men. The
average saving of women is Birr 231.17 per month but, the mean saving of men is Birr 173.63
per month.

Even if the average saving of men is less than the average saving of women, it is not logical to
conclude as generally men save less than women by simply observing at the amount of average
saving. This is because, we have to consider other saving measurement mechanisms like average
propensity to save (APS) and marginal propensity to save (MPS). Average propensity to save is
expressed as the ratio of total saving to total personal disposable income. Here, to measure the
average propensity to save of men and women household heads, it is must to measure the
average income of both household heads. Therefore, the mean income of men and women
households is Birr 1461.26 and 1429.70 respectively. Hence, table 4.2 showed that average
propensity to save (APS= saving / personal disposable income) of women is more than men.
Average propensity to save of women and men is 0.16 and 0.11 respectively, calculated as
231.17/1420.7 and 173.63/1461.26. Therefore, on average, women save a major part of their
disposable income than men, keeping other factors constant.

21 | P a g e
Table 4.2: Income, Saving and APS of household head by Gender

Variable Obs Mean APS

Income Saving

Male 83 1461.26 173.63 0.1188

Female 17 1429.70 231.17 0.1616

Total 100 2890.96 404.8 0.2804

Source: Own survey 2016/17

Another most important determinant of household saving is disposable income of the household.
Disposable income is expressed as personal income minus personal income tax and household
tend to spend a certain part of their disposable income on consumption and save the rest. Thus,
disposable income is the summation of household's consumption and their saving. Both
theoretical and empirical literatures on saving have consistently outlined that income is one of
the major determinants of household saving. The relationship between savings and income has
been a major subject of discussion in the growth literature. Subsistence consumption theories
suggest that countries with higher income levels tend to have a higher saving rate and the
empirical evidence strongly supports this conclusion.

In order to clearly show the impact of household disposable income on household saving,
income is categorized in to five groups [0 – 500], (500 – 1000], (1000 – 2000], (2000 – 5000],
>5000 Birr). Based on this classification of income, the mean saving of household heads with an
income level of, (([0 – 500], (500 – 1000], (1000 – 2000], (2000 – 3000], >3000 is Birr 36.6,
55.21, 161.72, 484.61 and 1200 respectively.

4.1.2 Description of the values of the variables of the model

In this sub section summary of the values of major variables of the model obtained from the
survey would be presented and discussed in a table. In this regard used descriptive summary
statistics such as mean, variance (or standard deviation), minimum and maximum values to
describe the values of major variables of the model obtained from the survey. Then the
summaries would also be given under the table.

22 | P a g e
Table 4.2: Summary of the values of major variables of the model for sample saving
of households in dale Woreda

Variable Obs Mean Std. Dev. Min Max

saving 100 183.42 239.4686 0 1400

age 100 44.66 10.22496 24 75
gender 100 .83 .3775252 0 1
ys 100 7.19 4.629058 0 17
ms 100 .11 .314466 0 1

fs 100 5.79 1.887198 2 12

ls 100 1.0745 .594837 .25 3
income 100 1455.9 700.7688 100 4000
dr 100 .3818 .1683812 0 .75
bcss 100 .74 .440844 0 1

exfbi 100 .51 .5024184 0 1

fica 100 2.08 1.186456 1 4
ffi 100 .71 .456048 0 1
attitud 100 .68 .4688262 0 1

Source: own survey, 2016/17

4.1.3 Descriptive analysis of the effects of socio-economic factors on the rural household saving
In this part descriptive methods of date analysis are employed to have a quick look at the effects
of effects of socio-economic factors on the rural household saving in Dale Woreda. To show the
effect of the continuous independent variables we computed the person’s correlation coefficient
between saving and each continuous independent variables of the model based the data obtained
from the sample survey.

23 | P a g e
4.1.3.1 The effects of age, income, family size, education level, land size and dependency ratio
on rural household saving pattern in dale Woreda.
Since all age, income, family size, education level, land size and dependency ratio are continuous
variables, in this part the researchers used persons correlation coefficients to show the effects of
these independent variables on saving. The result is summarized below in the table.

Table 4.3: Correlation matrix among variables of the model

Saving Age Income Family Education Land Dependency

size level size ratio

Saving 1

Age 0.0352 1

Income 0.0781 -0.5135 1

Family size 0.0624 0.0083 0.2825 1

Education level 0.4935 0.0144 0.2065 0.4396 1

Land size 0.5992 -0.1831 0.3964 0.3973 0.5052 1

Dependency ratio 0.1996 0.0404 -0.1846 -0.4737 -0.2252 -0.0498 1

Source: Own computation based on own survey, 2017

4.1.3.2 The effect of gender, marital status, beneficiary in current situation saving, expect full
business idea, financial institution customer attract, formal financial institution and attitude
The following table summarizes the mean saving of rural households included in the survey
based on their gender, marital status, beneficiary current situation saving, expect full business
idea, financial institution customer attract, formal financial institution and attitude of about
saving.

It is evident that from table 4.4 below most of the family’s exercises saving with commercial
bank, 42 percent of the sampled households was found to have saving account with the
commercial or other banks. Off course families saves their surplus money after consumption in

24 | P a g e
different form and place 24 percent of the respondents save their money at any other which they
consider it highly liquid and easily accessible without any additional cost at any time. The left
over households 2% saved their money within Bunna International Bank. The sum of households
who saved their after consumption money with Rural microfinance and Credit and saving
association are constitute 32 % of the sampled households. About 97 percent of the respondents
were married and the remaining percent were single.

Table 4.4: Mean saving differences of financial institution customer attracts and
marital status beneficiary in current situation saving, expect full business idea,
formal financial institution and attitude .

Variables Freq. Percent Com.

Financial Commercial bank 42 42.00 42.00

institution
Omo microfinance 32 32.00 74.00

Bunna I. Bank 2 2.00 76.00

Other 24 24.00 100.00

Total 100 100

Marital status Married 89 89.00 89.00

Single 11 11.00 100.00

Total 100 100

Expect full No 49 49.00 46.00

business idea
Yes 51 51.00 100.00

Total 100 100.00

Formal financial No 29 29.00 29.00

institution
Yes 71 71.00 71.00

25 | P a g e
 Total 100 100.00

Beneficiary in No 26 26.00 26.00

current situation
Yes 74 74.00 100.00
saving

Total 100 100.00

Attitude No 32 32.00 32.00

Yes 68 68.00 68.00

Total 100 100.00

Source: Own computation based on own survey, 2017

4.2 Econometric Analysis

4.2.1 Results of the estimated model

The data analysis of multiple tobit regression result shows that among the thirteen (13)
explanatory variable that entered in the model, the expected sign or the coefficient for land size,
family size, income, years of schooling (education level) and expect of full business idea are
statistically significant at 5% and age, gender, marital status, dependency ratio, beneficiary in
current situation saving, financial institution customer attract, formal financial institution and
attitude are statistically insignificant at 5%. From the regression result, the coefficients of
determinants which is the measurements of goodness the model, is 45%. This shows that 45 of
the total variation of output explained by the change of age, education level, family size, gender,
land size, income, expect of full business idea marital status, dependency ratio, beneficiary in
current situation saving, financial institution customer attract, formal financial institution and
attitude . This mains independent variable has expected 45% of the variation in the dependent
variable. The above regression model also show the number of oxen, land size, improved seed
family size and access of credit were a positively related and the other is negatively related with
the wheat production.

26 | P a g e
4.2.2 Diagnostic test results of the model
When we used cross-sectional data we may encounter problem of heteroscedasticity (Greene,
2008). In order to correct the heteroscedasticity problem we can estimate the robust standard
errors instead of the usual standard errors (Wooldridge, 2002). Thus, the Tobit model which is
used in this study is corrected for heteroscedasticity problem using the robust command in Stata
(robust standard errors are estimated for the Tobit) and from correlation matrix (pair-wise
correlation coefficient analysis), it is observed that there is no multicollinearity problem among
independent variables. According to (Gujarati, 2004) rule of thumb, multicollinearity is a serious
problem, when a pair wise correlation coefficient between two independent variables is greater
than or equal to 0.8. Therefore, from correlation matrix generated using the survey data it is
shown that there is no series multicollinearity problem in this study.

4.2.2.1. Test for muliticollinearity

An indication for a liner relationship between independent variable is called multicollinearity
(Gujarati 2004). To test the existence or not-existence of multi-collinear problem, variance
inflation factor (VIF) techniques should be employed. The variance inflation factor is a measures
of the reciprocal of the complement of the inter correlation among the predictors. VIF=1/ (1-r 2)
where r2 is the multiple correlations between the predictor variables and other predictors.

A decision rule for multicollinearity test for the model stated is a variable whose VIF value are
greater than 10 indicate the existence of multicollinerity. As it is observed from table 4.7 there is
no VIF value greater than 10. The researcher concluded that there are no problems of
Multicollinerity between the explanatory variables.

4.2.3 Interpretations and discussion of the results of the model

As reported in table 4.5, the log transformed disposable income positively and significantly
affects household saving. That is, as the disposable income of the household head increases,
household saving changes in the same direction. This result is consistent with the hypothesis and
is also in agreement with the other empirical studies by (Kibet et al., 2009; Horioka and Wan,
2007; Abdelkhalek et al., 2009; Loayza et al., 2000)

Education affects the household saving behavior negatively and statistically at the level of 5%
significance. Educational level of the household head is also another important determinant of
27 | P a g e
household saving. From the Tobit estimates of table 4.5, education coefficient is negatively and
significantly determined household saving. In order to see the impact of education on saving it is
categorized in to four (illiterate, primary school, secondary school and university level).
Therefore, being illiterate (EDUL1) and household saving are negatively and significantly
related. But, as the level of education decreases household saving increases. Thus, EDUL3
(secondary school) and EDUL4 (tertiary school) negatively and significantly determine
household saving. Education has a negative impact on household savings mainly because of the
awareness that occurs with lower educational levels and at coffee growth time the household
head withdraw from education. The household should not be spent their time on unnecessary
activate in rural area.

The other determinant of household saving is family size. As it is clearly shown in the Tobit
estimates of table 4.5, family size negatively and significantly affects household saving. A higher
dependency ratio implies a greater burden of consumption expenditure, that is, expenditure on
food, education, cloth and so on and hence, the more the allocation of household budget towards
consumption expenditure leads to lower saving. On the other hand a reduction in the number of
children relative to the working age population alleviated household budget constraints, thereby
boosting savings rates.

Finally, Land size of households positively and significantly affects households saving. Indirect
relationship is found between size of land holdings (SLH) and household savings. The results
suggest that households having more land holdings can save more than the households have less
land holdings. Households with one more hectare of land can reduce savings by 2760.04 birr on
the average. Study area is basically the agriculturist area. Coffee, Enset, Avocado and Maize etc
are main crops of this area. People having more land holdings can spent very large amount of
money throughout the year. They have less capacity to save in response of more land holdings.
This result is different with, Khalek et al. (2009).

Table 4.5: Tobit Estimates of Household saving

28 | P a g e
. tobit saving age gender ys ms fs ls income dr bcss exfbi fica ffi attitud, ll

Tobit regression Number of obs = 100

LR chi2(13) = 100.69
Prob > chi2 = 0.0000
Log likelihood = -532.34114 Pseudo R2 = 0.0864

saving Coef. Std. Err. t P>|t| [95% Conf. Interval]

age 2.54532 2.108557 1.21 0.231 -1.645666 6.736305

gender -52.57427 64.9536 -0.81 0.420 -181.6766 76.52805
ys -9.967498 4.89996 -2.03 0.045 -19.7067 -.2282968
ms -17.74631 76.73325 -0.23 0.818 -170.2619 134.7693
fs -31.46035 12.40214 -2.54 0.013 -56.11096 -6.809748
ls 115.0038 34.51895 3.33 0.001 46.39369 183.614
income .260981 .033134 7.88 0.000 .1951235 .3268385
dr 43.16616 115.1105 0.37 0.709 -185.6284 271.9607
bcss 71.97949 59.31938 1.21 0.228 -45.92421 189.8832
exfbi -88.53566 54.24601 -1.63 0.106 -196.3555 19.28416
fica -29.20671 18.29484 -1.60 0.114 -65.56968 7.156267
ffi -23.44382 49.03758 -0.48 0.634 -120.9113 74.02368
attitud 49.79779 49.57908 1.00 0.318 -48.74601 148.3416
_cons -144.1116 147.4439 -0.98 0.331 -437.1724 148.9491

/sigma 158.0729 12.83748 132.5571 183.5888

Obs. summary: 20 left-censored observations at saving<=0

80 uncensored observations
0 right-censored observations

29 | P a g e
 CHAPTER FIVE

CONCLUSIONS AND RECOMMENDATIONS

5.1. Conclusions
The study analyzes determinants of household savings based on data collected from Dale district
(Woreda) through multistage random sampling technique in 2016/2017. Generally, based on the
findings of the study, it is possible to conclude the effect of the explanatory variables on
household saving as;

 Higher income households save more than those of lower income. Income was found to
determine saving positively and significantly.
 Educational level of households negatively and significantly affects rural households
saving in Dale Woreda. As level of education decrease households would be aware of the
merit of saving as a result, household saving increase.

The study reveal that illiterate households negatively and significantly determine saving.
Household heads with secondary and tertiary school negatively and significantly affects saving.

 The family size also determines saving negatively and significantly. That is, households
with large family size save less than households with small family size.
 Finally, Land size of households positively and significantly affects households saving.
The results suggest that households having more land holdings can save more than the
households have less land holdings.

5.2. Recommendations
It is obvious that the level of saving in Ethiopia is very low; even it is less than from saving of
low income sub-Saharan Africa countries. It is not secret that saving contributes a lot for
economic growth and development.

Thus, what measures should be taken to improve saving in the country in general and in Dale
Woreda in particular?

Based on the study findings the researchers recommend the following policy recommendations.

30 | P a g e
  Since income is the major determinant of saving then, due attention should be given to
increase income of households. Income could be increased by implementing policies that
increases the employment opportunities and reduce underemployment and disguised
unemployment.
 Women are expected to save more than men. But, most of the time men dominate women
and control income and assets. Thus, in order to improve saving the government should
enhance women empowerment and pave the way for women to increase their income.
Moreover, there should be equality especially on the control of income and other assets
among men and women.
 Financial sector development in Ethiopia in general and in Dale Woreda in particular is at
its infant stage.

Hence, due attention should be given by government and private institutions in the expansion of
financial institutions in order to mobilize saving. Besides, not only the limited outreach of banks
and microfinance institutions but also their focus is on providing loan than mobilizing saving.
Therefore, there should be “Saving lead credit” strategy than “Credit lead saving”.

 A lot should be done in Dale especially in the family planning. This is because the family
size in this Woreda is very high. This is one reason for the existence of low household
saving within the Woreda. Therefore, appropriate measures should also be taken for the
implementation of family planning.
 Education is reducing aggregate household saving in rural. Education has a negative
impact on household savings mainly because of the awareness that occurs with lower
educational levels and at coffee growth time the household head withdraw from
education. The household should not be spent their time on unnecessary activate in rural
area. Though government has already provided free elementary and high school
education in public schools, their outreach is limited. Also the government should have
provided awareness how to encourage or facilitate education with productivity of coffee
growth time to goes together. However, efforts should be made by government, private
institutions and the public to increase the number of schools.

31 | P a g e
Nowadays, our government increases tertiary education by increasing the number of Universities
at an alarming rate. Efforts to lessen drop-outs and advocate for more students to continue and
finish high school should also be maintained.

32 | P a g e
33 | P a g e
REFERENCES

Abu, G. (2004). Determinants of Domestic Saving in Ethiopia. Paper for the Second
International Conference on the Ethiopian Economy, EEA.

Alemayehu, G., Abebe, S. and Daniel, Z. (2006). Finance and Poverty in Ethiopia: A Household

Level Analysis. UNU-WIDER no. 2006/51.

Bautista, Romeo M. and Lamberte, Mario B. (1990). Comparative Saving Behavior of Rural
and Urban Households. Journal of Philippines Development, Volume XVII No. 2
Brata, A. G. (1999). Household Saving Behavior: The case of rural industry in Bantul. Analysis

CSIS,28(1),75-86.
Burney NA, Khan AH (1992). Socioeconomic Characteristics and Household Savings: An

Analysis of the Households’ saving behavior in Pakistan.

Gedela, Sureya P. (2012). Determinants of Saving Behavior in Rural and Tribal Households

(An Empirical analysis of Visakhapatnam District). International Journal of

Research in Social Sciences. 2(3):108-128

Gujarati, D. N. (1995). Basic Econometrics. New York: McGraw Hill, Inc. 3rd Edition.

khalek, T. A., Arestoff, F., El Mekkaoui N. de Freitas, and Mage, S. (2009). A Micro

econometric Analysis of Households Saving Determinants in Morocco.

Université Paris Dauphine, France.

MoFED (2012). Macroeconomic Developments in Ethiopia. Annual Report 2010/11

Rehman H, Faridi Z, Bashir F (2010). Households Saving Behavior in Pakistan: A Case of

Multan District. Pakistan Journal of Social Sciences (PJSS). 30(1): 17-29.

Bersales, Lisa Grace, S., and Mapa, Dennis S. (2006). Patterns and Determinants of Household

Saving In the Philippines. Retrieved June 14, 2011, from pdf.usaid.gov/pdf

docs/PNADH621.pdf

Todaro, M.P. (2000). Economic Development. 7th ed. New York.

34 | P a g e
 APPENDEX
APPENDEX - 1

ARBA MINCH UNIVERSITY

COLLEGE OF BUSINESS AND ECONOMICS
DEPARTMENT OF ECONOMICS
QUESTIONNAIRES
This questionnaire is designed by third year economics students in Arba Minch University,
College of Business and Economics in order to fulfill the study in DaleWoreda. The aim of the
study is to determinants of rural coffee producing households saving in Dale Woreda.As it is for
pure academic purpose telling the truth should not affect you in one or another way. Yours
genuine and honest response is highly regarded for the paper success.
Instruction:-
 You are not supposed to write your name
 Your response should be confidential
 Please tick ( √ ) in the box according to your choice explain or circulate letter of your choice
Thanks for your advancement

1. Age of household head………………………years

2. Gender: Male Female

3. Years of schooling……………………….years.

4. Marital status?

A. Married B. Single C. Divorced D. Widowed

5. Family Size………………………………..

Male…………………….. female………………………….

6. Total land size you owned by hectares…………..

7. Total income that the family received from different resource per-month in birr……………

35 | P a g e
8. Based on question number 5. Are there any dependent on your families from the following age
category?

A. 0-14 B. 15-64 C. Above 65

9. Do you agree that of the current situation saving is beneficiary?

Yes No

10. Have you expect gain from any full business ideas?

Yes No

11. Which financial institutions have more customers attract?

A. Commercial Bank B. Omo Microfinance C. Buna Bank D. Others

12. Based on Question number 7

12.1 Total expenditure on food per-month in birr……………

12.2 Total expenditure on cloth per-month in birr…………...

12.3 total expenditure on education per-month in birr………..

12.4 Total expenditure on housing per-month in birr………….

13. How long it takes you before you reach the nearest financial institution…………….minute?

14. Based on question number 12, If you have choose above, how many birr is saved in different
institution in birr………………………

15. Do you think the formal financial institution around your area?

A. Agree B. Disagree

16. What is the attitude of about saving?

A. Good B. Bad

Thank you

36 | P a g e
APPENDEX – 2

REGRESSION

Table 4.6 Robust Regression

Robust regression Number of obs = 100

F( 13, 86) = 14.11
Prob > F = 0.0000

saving Coef. Std. Err. t P>|t| [95% Conf. Interval]

age 3.015967 1.522197 1.98 0.051 -.0100596 6.041994

gender -67.23008 48.41821 -1.39 0.169 -163.4823 29.02214
ys -1.093934 3.50244 -0.31 0.756 -8.056554 5.868687
ms -63.86752 56.92603 -1.12 0.265 -177.0327 49.29769
fs -32.2194 8.739552 -3.69 0.000 -49.59305 -14.84575
ls 141.2988 24.84878 5.69 0.000 91.90107 190.6965
income .1503656 .0237155 6.34 0.000 .1032207 .1975106
dr 12.34829 84.51295 0.15 0.884 -155.6579 180.3545
bcss 2.058083 41.58256 0.05 0.961 -80.60531 84.72148
exfbi -39.94596 39.93409 -1.00 0.320 -119.3323 39.44039
fica -25.28683 12.93307 -1.96 0.054 -50.99693 .4232699
ffi -56.56287 34.25847 -1.65 0.102 -124.6664 11.54071
attitud 32.53046 35.24062 0.92 0.359 -37.52558 102.5865
_cons 6.886869 106.0738 0.06 0.948 -203.9809 217.7547

Table 4.7 The result of VIF

Variable VIF 1/VIF

exfbi 2.44 0.409443

expgs 2.26 0.442572
gender 2.26 0.443303
ms 2.21 0.452400
ys 1.79 0.560069
ffi 1.74 0.575525
age 1.72 0.580009
income 1.71 0.586217
fs 1.68 0.594120
ls 1.48 0.676570
fica 1.39 0.719219
dr 1.24 0.803932

Mean VIF 1.83

37 | P a g e
 Kernel density estimate

.003
.002
Density
.001
0

0 500 1000 1500

Saving
kernel = epanechnikov, bandwidth = 56.4403

38 | P a g e