WOLAITA SODO UNIVERSITY

COLLEGE OF BUSINESS AND ECONOMICS

DEPARTMENT OF ECONOMICS

DETERIMINANTS OF URBAN HOUSEHOLD CONSUMPTION BEHAVIOR

(IN CASE OF BICHENA TOWN)

A SENIOR ESSAY SUBMITED TO THE DEPARMENT OF ECONOMICS IN

PARTIAL FULFILLMENT REQUIREMENT FOR THE BA DEGREE IN
ECONOMICS

SET BY: ZEWDU DIRES

ID NO_: CBE/R/018/10
ADVISOR: DERESE D. (MSc)
 DECEMBER 2020
WSU, ETHIOPIA

ACKNOWLEDGEMENT
First and for most I would like to thank the almighty and benevolent GOD that let me
stay in life with a proper health and that gave me limitless aid, costless support, and
strength to pass all the challenges in each phase of this study, Next I would like to
express my heart full thanks to all who supported me in the course of this study
including my advisor Instructor DERESE D. (MSc) for this supervision, valuable
guidance, intellectual encouragement and critical constructive comments.

Last, but not least I am grateful to my family, friends and others who helped me
morally and financially. In addition, I would like to express my profound gratitude
and appreciation to all my friends for their likely discussion in helping me on this
study.
ABSTRACT

This paper aims at investigating the household consumption behaviors in Bichena

town. The data for this study has been collected from 95 respondents of household
through questionnaires by using simple random sampling technique in three kebeles
of the town. The study uses cross sectional approach to investigate the effect of
household disposable income, family size, age, education level of household, saving
status of household, investment of household, expectation of household and other
family dependent on household as source of income by collecting data from the
household. Multiple linear regression models (MLRM) and OLS estimation methods
was determine the consumption behavior of household of Bichena town on cross
sectional data. The econometrics result is showing that, household disposable income,
family size, age, education level of household and other family dependent on
household as source of income, have a positive impact on household consumption
behaviors. However, expectations of household and saving status have a negative
impact on household consumption behavior.
ACRONYMS

CSA= CENTRAL STASTICAL AGENCY

FYGTP=FIVE YEARS GROWTH AND TRANSFOMATION PLAN
USA = UNITED STATE OF AMERICA
OLS= ORDINARY LEAST SQUARE
SDPRP= SUSTENABLE DEVELOPMENT AND POVERITY REDACTION
PROGRAM
 CHAPTER ONE

INTRODUCTION

1.1 Background of study

All around the world, understanding of concept of consumption is very important for
overall economy. Consumption is act of using goods and services to satisfy human
needs and desires. Most business activist in the world was aimed at providing goods
and services for consumption. The value of goods and service families consume
depends almost entirely on its income and wealth. (Hennery J. Aaron. (2001)

Americans spend more than 95% of their after -tax income on consumer goods and
services. About 2/3 of all the goods and services that produced in USA in each year
are used by consumers. The remaining 1/3 was used by the government or invested in
buildings, manufacturing machinery, other forms of capital in America, or other
countries of world. (Hennery J. Aaron. (2001)

According to the idea of Manikew, 2000 Consumption consists of the goods and
household in different product categories buys services. It makes up 2/3 of GDP.
(Manikew, 2000)
In modern term, household consumption is previously defined 70 to 80 percent of
total national expenditure. Even in western capitalist countries, the significance part
of total consumption was determined by expenditure of public authority. (M.C.
Hennery, (2005)
About 85% of total population of Ethiopia lives in rural area and practicing in
agriculture sector with subsistence farming system. Compared to thus production
systems with its population growth rate, it is difficult for each household head to feed
his or her family. In Ethiopia, the household consumption behavior is in the lowest
level. This is because, since most people in Ethiopia produced agricultural products
and these products are both price and income inelastic the return from such types of
production is very low. (Soni, 2007)

According to Sitotaw and Negus, 2005, the household consumption behaviors in the
country even at regional level are different, due to the variation of income and other
variables among nations and people in the world. In Amhara region households faces
problems in all aspects to satisfy their needs. Because individuals cannot give value
for work this leads to he/she have not enough income to maximize his/her living
standard. When individual households` income increase, consumption and saving
were increased and vice versa. (Sitotaw and Negus, 2005)

1.2 Statement of the problem

Theoretically, there are many different factors, which determine the growth of the
overall economy of the country. Among these, the most important determinant is
consumption of household. It has the attention of micro economists who is interesting
in studying the effect of household behavior on the behavior of consumption in
explaining the economy (A. koutsoyinnis, 2006).

According to M. Friedman (2004) in microeconomic term total consumption

expenditure of household is the sum of value of all varieties of goods and service
purchased (i.e. prices times quantity). (M. Friedman, 2004)

According to Harrods and Do mar (1950) as cited by Todaro (2012) saving or

investment are almost proportional and affected by consumption level. That is if
household consumption has such type effect on the economy, the economy has
changed. This leads to conduct a research on household consumption behavior.
Especially, the low-income level in least developed countries like our country
Ethiopia, household is more interested on consumption than investment. Thus, the
consumption levels of household can determine the investment level of country.
(Todaro, 2012)

In addition to this Zelalem (2005) analyzed household consumption behavior in Addis

Ababa by using 871 HHs as a sample size with simple random sampling technique.
His methodology was both descriptive and econometrical. The variables that he used
as explanatory were HH income and family size. Accordingly, both explanatory
variables affect HH consumption positively. Dawit, (2016) also tries to analyze a
study on household consumption behavior in woldya city. He also analyzed his study
by considering household income, family size, age, education and expectations as
main determinants of household consumption behavior.

In general most researchers such as (Zelalem, 2005; Dawit, 2016), conducted with
related to this topic, but they cannot incorporate the variable other family dependent
on household in their studies. There for this study was tire to fill the gap, by adding
this variable. On the other hand, some of the studies have done using descriptive
analysis. However, this study was accomplished by using econometrics model
especially ordinary least square (OLS) estimation methods.

Bichena is one of the town in eastern Gojjam administrative zone, the household
consumption in most part of Keble’s in this town is at low level, due to different
factors like education level, family size, administration problem, socio-economic
problem, income constraint, market problem etc. (BFEDO,2011). However, there is
no research have been done in identifying the major responsible factors that is
responsible for household consumption behavior in study area. Therefore, the study
has been believed to fill the gap by assessing the main factors of household
consumption behavior in the study area.

1.3 Research Question

 What is the consumption pattern of households in Bichena town?

 What are the factors that affect household consumption behavior?

  Which factors are more responsible for determining household consumption
behavior in Bichena town?

1.4 Objective of the Study

1.4.1 General Objective

The main objective of the study is identifying the determinants of household

consumption in Bichena town.

1.4.2 Specific Objective

 To assess the consumption pattern of households.

 To identify the factors that affect household's consumption behavior.

 To determine which factors are more responsible for determining household

consumption behavior.

1.5 Significance of the Study

The study has been provide information and frame work to residents of Bichena town
about consumption pattern. The finding of this study could be initiate other interested
bodies to undertake a better and detailed study in this area, and it may also draw
attention of researchers to the importance of understanding and knowing the
household consumption behavior to help in designing policies and strategies for future
development of the country and welfare of the society as a whole.

1.6. Scope of the Study

A household consumption behavior analysis involves a range of activities, which

needs time to cover all details; it requires understanding the resource base of a
particular area in relation to particular group of people. Moreover, the household
consumption behaviors are numerous, and vary from one household to the other.
Likewise, factors that affect household consumption behavior also vary. However, it
is difficult to investigate all consumption behaviors and factors influencing household
consumption behavior the study cover it to the major and common ones, which were
found to be relevant to the topics and research area.

In line with this, for reasons associated with time and resource, the scope of the study
was limited on three selected kebeles in Bichena town such as 01, 02 and 03 kebeles.

1.7 Limitation of the Study

This study has been faced the following limitations with related to time and cost,
another limitation of the study is concerning with data collection and reference
materials. Having thus problems the study has believed to resist these problems and
do the study effectively.

1.8 Organisation of the Study

This study contains five chapters. The first chapter deals with introduction part which
includes background of the study, statement of the problem, research question,
objective of the study, Hypothesis of the study, significance of the study, scope of the
study, limitation of the study and organization of the study. The second chapter deal
about review of literature, the third chapter deals with methodology of the study. The
fourth chapter devote to data discussion and analysis. The fifth chapter presents the
finding conclusion and recommendation.
 CHAPTER TWO

LITERATURE REVIEW

2.1 Review of Theoretical Literature

2.1.1 Definitions of Consumption

Consumption: - is defined as expenditure of household on final goods and services.

Each household consumes food, clothe, home furniture, TV sets, hair cut services, etc.
The consumption of goods and services is to get satisfaction. (C.KBansal, 2002)

Consumption Expenditure: - is defined as expenditure on goods and services to

satisfy needs and wants. It includes expenditure on food products, medicine,
education; culture and social participation such as Idir, miscellaneous which includes
electricity, water, telephone bill (Boys Melvin (2005).

Consumption function: - is the schedule, which describes the amount of

consumption at various levels of income. It is distinguishing from amount of
consumption, which is the amount of at specific level of income. (Dr.H.LAhuja,
2008)

Household consumption equilibrium involves, first, a set of wants it is attending and a

want satisfaction target for each. Second, a collection of activities and activity levels
it is using to meet each want, and a customary balance among the several. Third,
purchases of inputs associated with those activities and third differential employment.
In equilibrium, the mix of activities and their levels, and how the household organizes
its activities, just meets want satisfaction targets, and the purchases of inputs fit within
the household budget and time constraints with little slack. And of particular
importance the household has no tendency to changed circumstance, or changes are
the thing it knows or believes (Nelson et al, 2010).
2.1.2 Theories of Consumer Behaviour

Consumer behavior means how consumers decide on the basket of goods and services
they consume. It is essentially decision making behavior. In most cases, behavior of
consumer measured by its utility.

Utility refers to the numerical score responding the satisfaction that a consumer gets
from a market basket. In other words, utility is device used to simplify the ranking of
market basket. Alternatively, utility of a commodity is the basis of demand for it.

2.1.3 Approaches to the Analysis of Consumer Behaviour

In the study there are four main approaches to the analysis of consumer behavior.

A. Cardinal Utility Approach

It states that utility is measurable in cardinal absolute terms.

Assumptions

 Rationality: - it is assumed that the consumer is rational being in the sense

that he or she satisfies his/her wants in order of their merits.
 Limited money income: - the consumer has a limited money income to spend
on the goods and services he/ she choose to consume.
 Maximization of satisfaction: - every rational consumer intends to maximize
his /her satisfaction from his/her given money income.
 Utility is cardinally measurable: - the cardinalists assume that utility is
cardinally measurable that is it can be measured in absolute terms. For them,
utility of one unit of a commodity equal of the amount paid for it.
 Diminishing marginal utility: - the cardinality assumed that the gained from
successive unit of a commodity consumed decrease as the consumer consumes
a larger quantity of it.
 Constant utility of money: - the marginal utility of money remains constant
whatever the level of consumer income.

1
  Utility is additively separable: - utility derived from various goods and
services consumed by consumer can be added together to obtain the total
utility. It can be expressed through utility function. Suppose that the basket of
goods and services consumed by consumer contains “N” items and their
quantity may be expressed as:-

U=F(X 1, X2, X 3 ……..XN)

K Given the utility function above, the total utility obtained from (n) items maybe
expressed as:-

UN = U1(X1) + U2(X2) + U3 (X3) + ………………………UN (XN)

B. Ordinal (In Difference Curve) Utility Approach

It is only an expression of the consumer preference for one commodity over another
or for one basket of goods over another.

Assumption

 Rationality: - the consumer is a rational being. He or she aims at maximizing

his or her total satisfaction given his or her income and price of goods and
services he or she consumes. They have full knowledge of their own
circumstance and conditions required for rational decision.
 Ordinal utility: - unlike cardinal utility approach, in difference curve analysis
assume that utility only ordinary measurable. That is, the consumers are able
to express only the order of their preference.
 Transitivity and consistency of choice: - consumers’ choice are assumed to
transitive (A.Koutsoyiannis, 2006).

Transitivity of choice means that of, a consumer prefer, commodity X to Y and Y to

Z, then, they must prefer X to Z or they treat X=Y and Y=Z, they treat X=Z
(A.Koutsoyiannis, 2006).

2
Consistency choice means that, if they prefer X to Y in one period, they will not
prefer Y to X in another period or treat them as equal, everything remaining the same.

An indifference curve is the locus of points, each representing a different combination

of goods, which yield the same utility or level of satisfaction to the consumer. So that
they are indifferent between any combinations of goods when incomes to making a
choice between them. It also known as ISO utility or equal utility curve (Hiwot, 2007)

C. The Permanent Income Hypothesis

This hypothesis is primarily developed by Milton Friedman (2004) winner of prize in
university of Chicago, Friedman’s point of view is the rejection of the usual concept
of current income and its replacement with permanent income. A household
permanent income in any one year is in no sense indicated by its current income to be
received over a long period of time, stretching out over a number of future years.
Friedman puts this permanent income is to be interpreted as the mean income regard
as permanent by the consumer unit in question which in turn depends on its horizon
(Dorn bush, 2002)

Households measured or observed income in any particular year may be larger or

smaller than its permanent income. Friedman divides the households measured yearly
income into permanent and transitory income, so that is measured income is large or
smaller than its permanent income depending on the sum of positive and negative
transitory income components. In the same way, Friedman divides measured
consumption in to permanent and transitory components. A good purchased because
of an attractive sale price or a normal purchase deferred due to unavailability of the
good are examples of positive and negative transitory consumption (Mankiew,2000).

D. the Life Cycle Hypothesis

This hypothesis developed by Franco Modigliani (2003) a novel prize winner in
economic science. It is like Friedman’s permanent income hypothesis in that the
individual consumption in any given time period does not depend on to a significant

3
degree on his income during that period but depends on value of expected income
wealth.

The life cycle hypothesis is based on the argument that the rate of consumption in any
given period is a facet of plan, which extends, over his life cycle, although his income
displays a quit different pattern over the same year.

According to him, age structure of the population is an important determinate of

consumption pattern of different households in the economy. Consumption over some
one’s life time cannot exceeds his life time income unless that person is born wealth
then according to Franco consumption spending is financed by life time income and
wealth.

One important reason that income varies over a person life is retirement. Host people
plan to stop working at about age 65 and expect their income to fall when they retire.
Yet they do not want large drop in their standard of living, as measured by their
consumption. To maintain consumption after retirement, people must save during
their working years (Manikew, 2000).

Based on the life cycle hypothesis, when a household enjoys an increasing or

decreasing in income, there will be little effect on consumption.

2.2 Review of Empirical Literatures

Zelalem (2005) analyzed household consumption behavior in Addis Ababa by using

871 HHS as a sample size with simple random sampling technique. His methodology
was both descriptive and econometrical. The variables that he used as explanatory
were HH income and family size. Accordingly, both explanatory variables affect HH
consumption positively.

Additionally, Tadess kuma, (2010) analyzed changes in consumption patterns in

urban Ethiopia, Addis Ababa by using working lesser demand function as a dependent
variable, per capital income and other demand factors like dependence ratio, age and

4
family as independent variables. His methodology was both descriptive and
econometrical. According to his finding, age affects food demand negatively.

The demand for high value food items (example: meat, milk vegetables and fruits)
increase with higher income. They are also expensive source of energy. This implies
that poor households are unlike to access them. This is largely because poor
households prioritize to fulfill their basic energy requirement to avoid hunger. This is
mainly because high value food to be expensive source of energy for them.
Households with income near to subsistence level, consume large quantities of grains
and starchy staples and few fruits, vegetable, meat, milk, and milk products.
Consumer preferences on the other hand, shape the decision of consumer what to do
consume or not. Poor households, until they meet physiological need to satisfy
hunger, they have little choice but to focus on cheap sources of energy as grains and
starchy staples. Once they satisfied their basic energy needs, households start to
diversify their diets by including animals’ food sources, dairy products and fruits and
vegetable (Ruel et al, (2005) cited in Tadess kuma, (2010).

Dawit, (2016) used descriptive statement to identify the determinants of household

consumption behavior in woldia city. He used both primary and secondary data, the
result in household income, family size, age, education and expectations of household
as main determinants of household consumption behavior that determine a household
do not consume appropriately.

Generally, even though many researchers try to investigate effectively on household

consumption behavior in Ethiopia, they were not considered other family dependent
on household as a determinant of household consumption behavior. However, the
study has been tried to investigate by adding this variable.

5
 CHAPTER THREE

METHODOLOGY OF STUDY

3.1 Description of Study Area

Bichena is a town in west central Ethiopia, located in the east Gojjam zone of Amhara
region. It is located at 169km to north of the city of Addis Ababa and 155km to south
east of Bahir Dar city. Astronomically, it is located 10.24⁰N-10.34⁰N latitude and
38.9⁰E-38.4⁰E longitude. Attitudinally, it is more of woina Dega being located
between 1480m-2400m above sea level. The topography of the town is gentile slope.
Currently Bichena town has 8 kebeles. It is the administrative canter of Enemay
Woreda. It has a total land mass of 6945 hectares. The town has been experiencing
rapid urbanization and population increase. For instance, the 1984 and 1994
population censes put the population size of the town at 12570 and 16340,
respectively-an increase of 30% of the population in just 10 years. According to the
population and housing census of 2007, the population of the town reached about
27966, the population in 2007 increased by71.2% from its level in 1990s showing
also the very rapid rate of urbanization taking place in the town (BFEDO, 2011).

The climatic condition of the town is such that the mean annual maximum and
minimum temperature is 20.7°c and 7.5°c, respectively. The mean annual maximum
and the mean minimum rainfall of the town are recorded to be about 1200 mm and
2200 mm in the order of mention. The highest rainfall concentration occurs from the
middle of May to the end of September showing that the duration is short but the
intensity is relatively high in the town. (BFEDO, 2011)

Despite all the efforts made by the regional government in the provision of
infrastructure, still the town lacks the basic services and facilities such as water
supply, electric power, standard roads, sewerage and drainage facilities. (BFEDO,
2011)

6
3.2 Research Design

Under this section the study used different information gathering technique such as
observations, Questioners and interview, the subject of the study, tools of the data
gathering methods and procedures of data analyses and interpretation would be
discussed.

3.3 Types of Data

Cross sectional data would be used. Cross sectional data or cross section of a study
population in statistics and econometrics is a type of data collected by observing
many subjects (such as individuals) at the same point of time or without regard to
difference in time. Cross sectional data can be used in cross sectional regression,
which is regression analysis of cross sectional data (Brady et al, 2008). since it is
important for the values one or more than one variable are collected for several
sample unites at the same points in time, the researcher collect the data from the
respondents directly in a particular time.

3.4 Source and Method of Data Collection

3.4.1 Data Source

In this study both primary and secondary sources of data has been used. The source of
data’s included in central statics agency of Bichena town of finance and economic
development office, written document, books, journal report etc.

3.4.2 Method of Data Collection

As both primary and secondary data was used, the appropriate method of data
collection for primary data is questionnaires for sample households, interview with
town administrative and other response bodies. Likewise the secondary data was
collected from books, journal reports and other related documents.

7
3.5 Sample and Sampling Technique

There are two main sampling techniques in modern research study. These are
probability sampling and non- probability sampling. The study used simple random
sampling techniques with the purpose of homogenous population in various aspects,
like, culture, living standard etc. it is the simplest and easiest method of probability
sampling technique. It`s sampling procedure in which each element of the household
has an equal time chance of being selected in to sample. From the entire 8 kebeles of
Bichena town the study would select three kebeles such as, (01 kebeles, 02 kebeles
and 03 kebeles) in order to allocate the sample size via simple random sampling in
each element of household has equal chance being selected in to the sample.

Based on the data obtained from (BFEDO, 2011) the total numbers of household in
Bichena town are 4960, from thus this study takes 1900 households randomly for
three selected kebeles as a target population.

To determine the sample size the study might consider factors like cost, time etc. so
total size of 95 respondents are selected randomly by using Yamane sample size
formula of 1963.

n=N/1+N (e2) Where: n= sample size

N=total number of households

e= expected error that the study committed when gathering data.

The number of household is equal to 1900 (610, from 01 kebeles, 623, from 02
kebeles, and 667, from 03 Keble) and take 10% error term.

n =1900/1+1900 (0.12)

8
 =1900/1+1900 (0.01)

=1900/1+19

=1900/20

=95

Among this, from 01 Keble (610*95/1900) =31, from 02 Keble (623 *95/1900) =31,
and 03 Keble (667*95/1900) =33 respondents was selected randomly. The technique
would apply as gives equal chance to all household, that selected as a sample and they
represents the entry of household in all Keble’s due to homogeneity of Keble’s.

Table 3.1 determining of sample size

No. of kebeles No. of house hold Sample size

01 610 31

02 623 31

03 667 33

Total 1900 95

Source; own survey, 2019

3.6 Model Sspecification

As mention earlier the study has been applied econometrics tools in order to analysis
in the determinants of household consumption behavior, ordinary least square (OLS)
method was employed with dependent variable. OLS is particularly powerful as it
relatively easy to also check the model assumption such as linearity, constant variance
and the effects of outliers using simple graphical methods (Hutcheson and Sofroniou,
1999). In this study the dependent and independent variables have linear relationship.
Therefor OLS method is appropriate to show the linear relationship between the two
variables.

9
 CT =β0+ β1yd +β2 Ni+β3Educi+β4Agei+ β5Si

+β6Expc+β7Invst+β8Odf+ Ui

CT = total household consumption per month

yd. =household disposable income per month

Ni =household family size

Educ =education level of household head in years

Age = age of household head in year

Si = household saving status

Expec = expectation of household head

Invest= investment status of household

Odf= other family that dependent on the household

Ui =error term

Bi=coefficient of explanatory variables

B0 =autonomous consumption or constant

3.7 Expected Sign (Hypothesis) and Description of Variables

3.7.1 Expected Sign (Hypothesis)

The variables, which are incorporated in the model with expected signs, are given by the

following tables

No Variable name Symbol Description Expected sign

1 Income yd. Income of household +

2 Family size Ni Total number of household +

10
 member

3 Age Age Age of household head +

4 Education level Educ Educational level +

household head

5 Saving status Si Household saving _

6 Expectation Expec Expectations of the _

household head

7 Investment Invest Investment level of +

households

8 Other dependent family Odf Other dependent family on

household head +

Source; self-developed (2019)

3.7.2 Description of Variable

3.7.1.1 Dependent Variable

Total household consumption (CT):- It is the sum of all expenditure of household to

satisfy their wants. It is a continuous variable with the following components:
expenditure on food and drink (cereal, vegetable, fruit, milk, and milk product,
alcohol, chat and cigarette), expenditure on clothing including foot ware (shoe),
expenditure on housing (building materials, housing maintain ace and rent),
expenditure on culture and social (expenditures on different ceremonies, contribution
on Idir, donation to religious institution), expenditure on miscellaneous (water bills,
electricity bills, telephone bill including mobile card payment), expenditure on school
fee (school fees and other education expenditure), expenditure on medical service

11
(health treatment, durable goods fuel, utilities, personal care, transportation and non-
durable goods).

3.7.1.2 Iindependent variable

Household disposable income (yd.):- It is the income after tax (net income of
household from different source of income).This variable is continuous and the
expected sign is positive.

Family size (Ni):- It is the total numbers of household members. This variable is
continuous and the expected sign is positive.

Age of household head (Age):- It`s one of the determinant of household

consumption. It measure the age of the household head in year. This variable is
continuous and the expected sign is positive.

Education level of household head (Educ):- It is the number of years of schooling

attained by the simple household head up time of the survey. This variable is
continuous and the expected sign is positive.

Saving status of household (Si):- It is a deposited income after consumption. It is a

continuous variable and the expected sign is negative.

Expectation of household (Expec):- is one of the determinants of household

consumption, it is a belief about what might happen in the future. This variable is
dummy (1= certain and 0= uncertain) and the expected sign is negative.

Investment of household (Invest):- it is the purchase of goods, which not consumed

today but used in the future to create wealth for households. This variable is dummy
(1 =have a plan to invest and 0 = have not a plan to invest) and its expected sign is
positive.

Other dependent family on household (Odf):- It represents the other household

depend on the household and expected sign of this variable is positive.

12
3.8 Method of Data Analysis

After the collection of the necessary data, data analysis was carried out by using both
descriptive and econometric analysis. In the descriptive analysis different table, pie
chart and percentage were used to describe the information clearly. For econometrical
analysis, the study would concentrate on the variables. such as, household disposable
income (yd.), household family size (Ni), age of household head (Age), household
head education level (Educ), total household saving status (Si), expectation of
household (Expec), investment of household (Invest) and other dependent family on
household (Odf) as explanatory variable and total household consumption (CT) as
dependent variable and used ordinary least square (OLS) estimation method.
Additionally, t-test (variable significant), prob>F, Multicollinearity and
Heteroscedasticity problems can be tested. This estimation technique is working out
with the help of STATA software program method. The study uses ordinary least
square estimation method (OLS) the reason is that OLS is the most frequently used
estimator due to the fact that, it has some optional properties; best, linear and unbiased
estimation(BLUE) than any other estimators(Gujarat, 2004).

13
 CHAPTER FOUR
DATA PRESENTATION, ANALYSIS AND INTERPRETATION

4.1DESCRPTIVE ANALYSIS

4.1.1 DESCRPTIVE ANALYSIS OF PRIMARY DATA

Most of the time, market price of a commodity and income (budget) of a household
considered as a major determinants of household consumption. However, broadly
thinking, there are many other important determinants of consumption that we have
seen in day today activity of purchasing and consuming different consumption and
none consumption commodities. These are government tax, family size, income of
household, weather condition, education level, age of household head, saving of
household, expectation of household, investment of household and other different
demographic factors, among these the following were considered as major
consumption determinants in this study. In these study 95 households has selected as
a sample to response the questionnaire.

4.1.1.1 Gender of respondents

Table 4.1 gender of respondents
Gender Number of respondents
FER %
Male 58 61
Female 37 38.9
Source: own computation survey 2019
Household grouped under both male and female. According to the above table
58(61%) of respondents are male and 37(38.9%) of respondents are female.

4.1.1.2 Age of household (age)

It is one part of determinants of household consumption behavior in the world. Then,
age of population can be classified into three major groups, that is, children, adult and
old age. Even if, the study considers age of household, these age groups have their

14
own contribution on household consumption level.
Table 4.2 Composition of Household Head’s Age
Number of respondents Total respondents
Male Female
Age of HH
FER % FER % FER %
< 18 0 0 0 0 0 0
19 – 64 55 57.9 36 37.9 91 95.8
>65 3 3.15 1 1.05 4 4.2

Source: own computation survey 2019

As a result, most household grouped under both adult and old age groups. According
to the above table 91(95.8%) of respondents are adult age group from this 55(57.9%)
of respondents are male and 36(37.9%) of respondents are female and 4(4.2%) are old
age group from this 3(3.15%) of respondents are male and 1(1.05%) of respondents
are female. This indicated that adults are more engage in production system with
higher consumption. Therefore, these age groups have either negative or positive
effect on household consumption level.

4.1.1.3 Education Level of Household Head (Educ)

Education level determines household consumption level. Because when each
household start education, they need different things or materials for achieving these
objectives. On the other hand, when they already educated, they know different items
or commodities that are available for consumption. This means that education level of
a household has its own effect on household consumption behavior (analysis).

Table 4.3 Educational Levels of households

Number of respondents Total
Educational level of HH Male Female
respondents

15
 FER % FER % FER %
Illiterate 1 1.05 0 0 1 1.05
Primary school 19 20 15 15.78 34 35.78
literate
Secondary and 14 14.7 9 9.47 23 24.2
preparatory
College and 24 25.3 13 13.68 37 38.94
above
Source: own computation survey 2019
From the above table, most households under this study area educated that is, at least
they can read and write. From respondents 34(35.78%) are educated primary
education (1-8 grade) from this 19(20%) of respondents are male and 15(15.78%) of
respondents are female, 23(24.2%) are educated secondary and preparatory education
(9-12) from this 14(14.7%) of respondents are male and 9(9.47%) of respondents are
female, 37(38.94%) are educated college and above from this 24(25.3%) of
respondents are male and 13(13.68%) of respondents are female and 1(1.05%) are
illiterate that male household. Most household heads are grouped under higher
education level. As a result, I can conclude that household education (weather they are
already educated or learning now) has great on their consumption.

4.1.1.4 Family Size (Ni)

It is the total numbers of household members and the major determinants of
consumption next to disposable income. This is because; it is mostly related with
human needs and wants. Household consumption in relation with needs can be
differing according to age and sex.

Table 4.4 family size of household

Number of respondents Total
Family size of HH Male Female
respondents
FER % FER % FER %
1-3 17 17.89 10 10.52 27 28.42
4-6 38 40 23 24.21 61 64.21

16
 7-9 3 3.15 4 4.2 7 7.36
>9 0 0 0 0 0 0

Source: own computation survey 2019

According to the above table, 27(28.42%) of family sizes found (1-3) range from this
17(17.89%) of respondents are male and 10(10.52%) of respondents are female,
61(64.21%) are family size found (4-6) range from this 38(40%) of respondents are
male and 23(24.21%) of respondents are female and 7(7.36%) are family size found
(7-9) range from this 3(3.15%) of respondents are male and 4(4.2%) of respondents
are female. Most household heads family size found (4-6) range. As a result, I can
conclude this family size consume more than the other family size range. This means
that family size of a household has its own effect on household consumption behavior
(analysis).

4.1.1.5 Household Disposal Income (Yd.)

It is the income after tax net income of household from different source of income. It
is the most determinants of consumption compared to others. Most of household’s
income especially those households, live in low-income level of developing countries
are shared consumption purpose. Any change in disposable income may have effects
on total expenditure. The direction of the effect of this change depends on the income
elasticity of the given commodity that is the consumption pattern of welfare of the
household.
Household income is one of the main determinants of household consumption level.
According to Mankiw, income and consumption have positive relation, means when
income increase at the same time consumption also increases.
Table 4.5 household disposable income

Number of respondents
Male Female Total respondents
Income of household`s
FER % FER % FER %

< 1000 11 11.57 7 7.36 18 18.95

1000 - 2000 17 17.89 10 10.52 27 28.42
2001 – 3000 5 5.26 6 6.31 11 11.57

17
 3001 – 4000 3 3.15 6 6.32 9 9.47

>4000 20 21.05 10 10.52 30 31.5

Source: own computation survey 2019

As shown on the above table from respondents 18(18.95%) are earn below 1000 birr
monthly from this 11(11.57%)of respondents are male and 7(7.36%) of respondents
are female, 27(28.42%) respondents income level is found in from 1000-2000 birr per
month from this 17(17.89%) of respondents are male and 10(10.52%) of respondents
are female, 11(11.57%)of respondents get 2001-3000 birr income per month from this
5(5.26%) of respondents are male and 6(6.31%) of respondents are female, 9(9.47%)
of respondents earn income 3000-4000 birr monthly from this 3(3.15%) of
respondents are male and 6(6.32%) of respondents are female and the remaining
30(31.5%) of respondents earn above 4000 birr per month from this 20(21.05%) of
respondents are male and 10(10.52%) of respondents are female. Most household
grouped under from 3000-4000 birr income level. As a result, I can conclude that
household disposable income has great effect on their consumption.

4.1.1.7 Household Saving Status (Si)

Saving is deposited income after consumption. According to Harrods- Do mar, it is
considered as proportional to investment. It is more sensitive for interest rate (its
saving rate is depends on interest rate of the financial institutions).

Table 4.7 household saving status

Number of respondents Total respondents
Do you save? Male Female
FER % FER % FER %

YES 58 61.05 37 38.9 95 100

5
NO 0 0 0 0 0 0

Source: own computation survey 2019

18
According to the data, we have collected for this study all households (95), has their
own saving account in different financial institutions, which is found in bichena town
from this 58 (61.05%) of respondents are male and 37(38.95%) of respondents are
female. From this, I can conclude that saving by itself has its own negative
contribution for household consumption.

4.1.1.8 Investment of household (Invest)

Investment is to allocate money (or sometimes another resource such as time) in the
expectation of some benefit for example, investment in durable goods.
Table 4.8 response about investment
Do you have a plan to Number of respondents Total respondents
Male Female
engage in investment
FER % FER % FER %
activity in near future?

YES 25 26.31 18 18.9 43 45.26

NO 33 34.7 19 20 51 52.7

Source: own computation survey 2019

As shown on the above table, 4.8, 51(52.7%) of the respondents respond that, they
have no a plan to engage in investment activity in near future from this 33(34.7%) of
respondents are male and 19(20%) of respondents are female and the remaining
43(45.26%) of them answered that they have a plan to engage in investment in near
future from this 25(26.31%) of respondents are male and 18(18.9%) of respondents
are female, so from the responses it can be implied that, most of respondents doesn’t
have a plan to engage in investment, because of high level of consumption (low level
of saving)

4.1.1.9 Other families dependent on household (Odf)

Table 4.9 Response about Dependent Families
Are there any households
Number of respondents Total respondents
Male Female
dependent on you as a
FER % FER % FER %
source of income?

19
 YES 31 32.63 15 15.7 46 48.42
NO 27 28.42 22 23.5 49 51.57

Source: own computation survey 2019

The above table (4.9) shows that, 49(51.57%) of sample respondents replied that there
is no any families dependent on them as source of income. From this, 27(28.42%) of
respondents are male, 22(23.5%) of respondents are female, and the remaining
46(48.42%) of respondents answered that, there is any families dependent on them as
source of income. From this, 31(32.63%) of respondents are male and 15(15.7%) of
respondents are female. this, implies that most of the households aren`t depend on
their own source of income.

4.1.1.11 Expectation of household (Expec)

Table 4.10 response of household expectation
Number of respondents Total respondents
What do you expect about Male Female
FER % FER % FER %
your future income?

Certain 35 35.84 16 16.8 51 53.68

Uncertain 23 24.21 21 22.1 44 46.31
Source: own computation survey 2019
The above table (4.11) indicates that, 51(53.68%) of sample respondents respond that,
their expectation about the future income have certain. From this 35(35.84%) of the
respondents are male and 16(16.8%) of the respondents are female and the remaining
of 44(46.31%) of sample respondents respond that their expectation about the future
income have uncertain expectation from this 23(24.21%) of the respondents are male
and 21(22.1%) of the respondents are female. This, implies that most of the household
certain expectation about future income.

4.1.1.6 Household`s consumption expenditures

In this section of the study, more attention is given to show the share of household’s
consumption expenditure on different commodities or items from the expenditure.

20
Figure 4.1 expenditure of household

8.24%
5.39%
5.44%
5.70% 45.80%
6.32%

11.10%
12.00%
total food consumption expendituer on housing
expendituer on cloth expendituer on social , cultural life
miscellaneous expenditure expenditure for school
expenditure on medical services other expenditure

Source: own computation survey

The result of this study shows that the consumption patterns of each household in the
town distributed on different commodity items in the following table.
Table 4.11 share of consumption expenditures
No Expenditure items Percentage share
1 Expenditure on food and drink 45.8%
2 Expenditure on housing 12%
3 Expenditure on cloth 11.1%
4 Expenditure on social, cultural life 6.32%
5 Miscellaneous expenditure 5.71%
6 Expenditure on school 5.44%
7 Expenditure on medical service 5.39%
8 Other expenditure 8.24%
Source: own computation survey 2019
According to the above, Expenditure on food and drink is one the most basic needs
for everyone to survive in life on the earth. It constitutes around 45.8% of the total
consumption expenditure of the household. It does not mean that food is the one only
consumption commodity rather it is compared to with others. Mostly it is known that
the share of consumption in poor countries are higher than that of rich countries , but
this gap declines (consumption share of poor countries decline) as its national income
increases.

21
Expenditure on house includes house building materials, maintenance and rent
expenditure, which constitutes 12% of in the total household expenditure. This shows
that most households (residents) of bichena town spend high rent for housing. This is
due to the imbalance between the income they received and the cost of materials for
building, because of existence of price shocks (inflation) and lack accessibility of
land.

Expenditure on cloth including footwear is unavoidable expenditure item in nature.

Because it is one of the basic needs. As shown in the above table, it constitutes around
11.1% of the total household expenditure.

Expenditure on Cultural and Social Life this expenditure item (expenditure on

different ceremonies, donation religious institution, contribution to idir etc.)
comprises 6.32% in the total household expenditure. Culture and social heritages are
one of major difficult for economic development in least developing countries.
Because this lead the household to spent more of their income for such types of
activities.

Miscellaneous Expenditure this category of expenditure item in the household

expenditure includes water bill, electricity bill, and telephone bill, and mobile card,
which constitutes 5.71% of the total household expenditure. It is the fourth major
consumption expenditure item from others.

Expenditure on School Fee means (other educational expenditure example, fees for
maintenance and for school fee). It accounts 5.44% of the total household
expenditure. Now a day’s people know the importance of education for sustainable
economic development of the country and improve social economic welfare of the
society as whole. They give more attention for children education in private school in
addition to government schooling system. This leads the household to spent more
income on school fee.

22
Expenditure on Medical Services is includes both modern and traditional medical
system, which accounts 5.39% of the total household consumption expenditure. This
result shows that either the might be in the household live in good health condition or
the household might use cheap medical treatment system. Expenditure on Other
Goods and Services this categories of expenditure item includes kitchen equipment,
durable goods like (TV and radio), fuel, utility personal care etc that constitutes
8.24% of the total household consumption expenditure item next to food and housing.
Durable goods are luxury, expensive and server the household for long period.
4.2 Econometric Analysis
4.2.1 OLS model result and interpretation
As stated in the methodology of this study the model that used to analyze the
household consumption expenditure can be written (describe) as follows.

CT =β0+ β1ydi +β2 Ni+β3Educi+β4Agei+ β5Si +β6Expeci+β7Invst+β8Odf+ U

Table4.12 OLS Regression result

OLS regression
Number of observation=95
R2= 0.9441
Explanatory variable Coefficient Standard error t-value P > /t/
Age 3.367767 2.045581 1.65 0.103
Education level 41.88108 16.80465 2.49 0.015*
Family size 1.848763 20.69021 0.09 0.929
Disposable income 0.4889863 0.0524707 9.32 0.000**

HH
Saving status -0.2957617 0.0544141 -5.44 0.000**
Investment of HH 156.4234 101.9756 1.53 0.129
Expectation of HH -6.523321 21.65024 -0.30 0.764
Other family 153.791 70.82933 2.17 0.033*

23
dependent
Constant -65.69391 143.6872 -0.46 0.649

Note that: - * 5% level of confidence

** 1% level of confidence
The format of total household consumption function were specified as-

CT= -65.69391+0.4889863Ydi+1.848763Ni+41.88108Educi+3.367767Age

(-0.46) (9.32) (0.09) (2.49) (1.65)

--0.2957617Si -6.523321Expec+156.4234Invst+153.791Odf+Ui

(-5.44) (-0.30) (1.53) (2.17)

Note: - the numbers that attached below the coefficient are the t-value.
As shown in table 4.12 above, most of the estimated coefficients have their own
expected result (signs). The estimated coefficients were interpreted as follows;
Household disposable income:-The coefficient of household disposal income is
statistically significant at 1 % level of confidence, including that, keeping other things
constant if disposable income of the household in bichena town increased by 1 birr,
then the mean consumption expenditure of the household increased by 0.4889863
than the previously monthly consumption expenditure level.
Family size:-The coefficient of family size to household consumption is positive that
is 1.848763. However, it is statistically insignificant at 5% level confidence.
Educational level household:-Education level of each household head is statistically
significant at 5% level of confidence. And has a positive relation with household
consumption. Generally, keeping other variables constant if the education levels of
household were increased by 1 year, consumption expenditure of households were
increased by 41.88108.
Age of household:-Age of household head is statistically insignificant at 5% level of
confidence with the coefficient of 3.367767.
Saving status of household:-Saving status of household is statistically significant at
1% level of confidence with a coefficient of -0.2957617. It has also a negative relation

24
with consumption.
Expectations of households:-The relation between expectations of household-to-
household consumption is positive with the coefficient of -6.523321. In addition, it is
statistically insignificant at 5% as well as 1% level of confidence. Investment of
household:-Investment of household is statistically insignificant at 5% level of
confidence and positive relation with household consumption at the coefficient of
156.4234, as the regression result shown, but logically, the relation between
investment and consumption is negative.
Other family dependent on household:-Other family dependent on household as
source of income is statistically significant at 5% level of confidence and it has a
positive relation with consumptions of household. This implies that, if the other
family dependent on household increased by one individual person, the mean
consumption expenditure of household goes up by 153.791.
All the regression results have sign that according to my prior expectation except
saving status and expectations of household, others have positive contribution for
household consumption expenditure.
The constant term (B0) indicates that, autonomous consumption of household means
even they have not their own income per month, the household continue their
consumption. These types of consumption come from different sources like loan from
different financial sector, which accounts -65.69391 of the total consumption
expenditure.
4.2.2 Test
4.2.2.1 Overall level of significant test
The overall significance of the variable is tested by Prob> F. If Prob> F less than 0.05
(5%), then the explanatory variables included in the model are jointly statistically
significant and the explained variable. As a result show Prob> F= 0.0000 then, it is
statistically significant.

25
The goodness of fit of the model is measured by coefficient of determination,
which shows the percentage or power of explained variable to express by the

explanatory variables. This is expressed by R–Square = 0.9441, which states that

94.41% of the variable of the explained (dependant) variable is due to the
variation of explanatory variables included in the model and remaining variation
(5.59%) is explained by variables which are not included in our econometrical
model.

4.2.2.2 Diagnostic test

In this section of test, inter relationship between the explanatory variables. In
addition, natures of the variance of error term were tested.

4.2.2.2.1Multicollinearity
This method shows the interrelation ship between the explanatory variables.
The presence of Multicollinearity affects the OLS estimator and make inefficient and
inconvenient. Therefore, it must be tested using variance inflation factor (VIF)
method. It shows how the variance of estimator influenced by the existence of
Multicollinearity. By using rule of thumb:
 When VIF > 10 or tolerance is close to zero, then, the explanatory variables
are correlated, accept alternative, has Multicollinearity problem.
 When VIF <10 or tolerance is greater than zero, then accept the null
hypothesis, no Multicollinearity. Based on this decision rule, the result shows
that there is no Multicollinearity problem (APPENDIX 2).

4.2.2.2.2Heteroscedacity

Heteroscedasticity means there is not constant variance, when you`re given the
first test, the Breusch-pagan one, first check the null hypothesis, constant
variance. This is saying that if the null hypothesis rejected then we have
heteroscedasticity, if the p- value 0.05 is smaller, then the null hypothesis
rejected and there is significant evidence and there is heteroscedasticity, so in
your example below as the p-value is less than 0.05 you have
heteroscedasticity. This test concerned with variation of variance.

26
Decision rule: if Prob> chi2 is less than 5%, accept H1 (reject alternative hypothesis)
and accept the null hypothesis Ho, that is 0.000 is less than 0.05. Therefore, there is
variation of variance in the model; the model has heteroscedasticity problem.to
overcome the problem use the method of heteroscedasticity –Robust standard errors
(APPENDIX 3).

CHAPTER FIVE
Conclusion and Recommendation

5.1 Conclusion
The study was attempted that to show household consumption behavior in bichena
town by describing households, consumption expenditure item and by showing the
effect of different determinants of consumption like household disposable income, age
of household head, education level of household, family size of the household, saving
status of household, expectation of household, investment of household and existence

27
of other family dependent on household as source of income. In this study 95
household are taken as a sample at random. To do this OLS estimation was employed.
The descriptive part of the study result shows that households spent their income on
eight major commodity items. From the total budget of household, most part of the
budget is used for consumption purpose. From this, the study can conclude that the
saving habit of each household in the town is low even if they participate in saving
activity.
Because of this low saving habit most of town’s residents have not a plan to engage in
investment activities. This study can conclude from the regression result, that the
households spend their income for unnecessary and for extravagant activities.
Adjusting and giving emphasis for such type of condition helps to achieve the goal of
government that is put under the statement of the problem like SDPRP and FYGTP.
That is, it is easy for the concerned body (government or public) to give solution for
such types of goals (achieving SDPRP, FYGTP and improve human living standard).
As indicated in discussion part, households spent their income for extravagant
activities such as nuptial, brith day ceremonise, tezkar, festive meal and other social
and cultural life.
The OLS result shows that household consumption positively affected by disposable
income, family size, age of household head, education level of household, family size
of the household, investment of household and existence of other family dependent on
household as source of income. The study concludes the overall significant of the
model (R2) it is seems good fit. But it indicates that the study left other explanatory
variables which are essential to determine household consumption behavior.
5.2 Recommendation
After taking above result in to consideration, the study recommended household give
more attention for consumption than saving, it needs a vast investigation on
household consumption behavior to enhance (improve) saving and design proper
policy for integrated economic development. During the consumption expenditure
(budget expenditure) households are better to use their own notebook (record habit of

28
expenditure) because it is good for plan saving and expenditure. The government
should have promoted saving habit of household by expanding different financial
institution, Awareness should have given to the household to save part of their income
and reducing the extravagant and cultural consumption habit to enhance the
sustainable development of country, food agency should set out various proclamations
in relation with household consumption behavior or pattern, Increasing investment of
household should be important for household consumption, should have created
awareness to the household to invest in investment goods like durable goods, The
household should have to allocate money (or sometimes another resource such as
time) in the expectation of some benefit for example, investment in durable goods.
Generally, Study in this area is a process, not the end. Then it gave the way to other
interested bodies to do the study on this area by including other explanatory variables
which is very essential for consumption determinant, not included in this study.

REFERENCE

 A.koutsoyiannis (2006). Modern Micro Economics University of

ottawaoontario.

 Bemnet and Woliams, (2008). Analysis of change in food consumption

patterns in urban Ethiopia. Ethiopian development research institute AA,
Ethiopia.

 Central statistical agency (2007). Census report of Ethiopia, AA, Ethiopia.

 Dawit H. (2016). Determinants of household consumption behavior,

Mekelle.

 Dourn bush, (2002).Macroeconomics, consumption and saving Chicago

University, USA.

 Guajarati D. (2004). Basic econometrics, United States of America, Mc

Graw-Hill press.

29
 Henery J. Aaron. (2001) world book, vol.4USA, Chicago

 M. Friedman (2004), a theory of the consumption function (Princeton, N.J.:

Princeton university press.

 M. Gregory Mankiew (2000), Macroeconomics, Harvard University press.

 Modigliani F. (2003) the life cycle hypothesis of saving, the demand for
wealth and the supply of capital, social research, Issue 33, pp.106-217.

 Nelson R. and David, (2010), an evolutionary theory of HH consumption

behavior, Colombia University New York, USA and university of
Manchester, UK

 Robert Mc Henery, (2005), economic theory world book, vol.17.

 Sitotaw and Nigus, (2006), introduction to economics. AA, Kuraz

international publisher.

 Soni, R.N (2007). Leading issues in agricultural economics theoretical and

applied New Delhi.

 Tadess k, (2010). Analysis of change in food consumption patternes in

urban Ethiopia. Ethiopia development research institute AA, Ethiopia.

 Zelalem T, (2005), analysis of household consumption behavior. Ethiopia,

AA university.

30
 Appendix
Questionnaire
Wolaita Sodo University
College of business and economics
Department of economics
Research questionnaire
This questionnaire is prepared by students at Wolaita Sodo University in department
of Economics. The purpose of this questionnaire is to study the expenditure behavior
(pattern) of households and to know the major determinants of consumption under a
title “Urban household consumption behavior analysis” in Amhara region Bichena
town. The questionnaire is going to use for academic and economic research purpose
only. Write and choice
Instruction:
 Choose the best for you from the alternative and put ‘X’ mark in the box of your

choice.
 Explain open ended question shortly and briefly.
 No need of writing your name.
Thank you in advance

Part one: demographic information’s

1. Age___________________________
2. Sex A male B female
3. Education level (specify the grade in number)

31
A) Illiterate B) Primary school (1-8) C) secondary and preparatory (9-12)
D) College and above
E) Specify your education level in number_______________
4. Total number of family size ______male ________female______
5. Are there any households dependent on you as a source of income?
A Yes B No

6. If you say yes in question 5, how many other family size dependents on the
household? Male________ female_____________
Part two: information on occupation and expenditure of household
7. How much the total net income that the household that received from different
source per a month in birr? _________
8. Based on question no 7 answer the following questions.
8.1. Total expenditure on food and drink per month (including cereal, vegetables,
Milk, onion, oil etc.) in birr _______________________
8.2. Total expenditure on cloth including footwear per month in
birr_________________________
8.3 Total expenditure on housing (building, rent etc.) in birr____________________
8.4Total expenditure on social and culture life (different ceremonies, such as, Idir etc.)
in birr________________________________________________
8.5 Total expenditure on miscellaneous (such as water bill, electricity bill, Telephone
bill including mobile card) per month in birr________________________
8.6Total expenditure on school fee (fee and expenditure for building or repairmen of
old school) in birr____________________________________
8.7 Total expenditure on medical services (treatment) per month in birr
___________________
8.8 Total expenditure on others (including utility, kitchen equipment, transportation,
etc.)In birr_______________________________
9. Do you save? A Yes B No
10. If you say yes in question 10 above, how much you save per month (in birr)

32
___________________________________?
11. Do you have a plan to engage in investment activity in near future?
A Yes B No
12. What do you expect about your future income?

A. certain B. uncertain Thank you!

33
APPENDEX 1

. Regress CT Age Educ Ni Yd Si Invest Expec Odf

Source | SS df MS Number of obs = 95

-------------+---------------------------------- F (8, 86) = 181.39
Model | 66037321 8 8254665.12 Prob > F = 0.0000
Residual | 3913657.99 86 45507.6511 R-squared = 0.9441
-------------+---------------------------------- Adj R-squared = 0.9388
Total | 69950978.9 94 744159.351 Root MSE = 213.33

------------------------------------------------------------------------------
CT | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Age | 3.367767 2.045581 1.65 0.103 -.6987133 7.434247
Educ | 41.88108 16.80465 2.49 0.015 8.474543 75.28761
Ni | 1.848763 20.69021 0.09 0.929 -39.28202 42.97955
Yd | .4889863 .0524707 9.32 0.000 .384678 .5932946
Si | -.2957617 .0544141 -5.44 0.000 -.4039333 -.1875901
Invest | 156.4234 101.9756 1.53 0.129 -46.29743 359.1442
Expec | -6.523321 21.65024 -0.30 0.764 -49.56258 36.51594
Odf | 153.791 70.82933 2.17 0.033 12.98691 294.595
_cons | -65.69391 143.6872 -0.46 0.649 -351.3347 219.9469
------------------------------------------------------------------------------

APPENDEX 2
. estat vif

34
 Variable | VIF 1/VIF
-------------+----------------------
Yd | 18.78 0.053243
Educ | 12.52 0.079874
Si | 6.43 0.155457
Invest | 5.38 0.185927
Ni | 2.81 0.355974
Odf | 2.62 0.382321
Expec | 1.36 0.736444
Age | 1.06 0.939187
-------------+----------------------
Mean VIF | 6.37

APPENDEX 3

. estat hettest

Breusch-Pagan / Cook-Weisberg test for heteroscedasticity

Ho: Constant variance

35
Variables: fitted values of CT

chi2 (1) = 43.43

Prob > chi2 = 0.0000

36