MEKELLE UNIVERSITY

COLLEGE OF BUSSINESS AND ECONOMICS

DEPARTMENT OF ECONOMICS

DETERMINANTS OF INVESTMENT IN
MANUFACTURING SECTOR: A MICRO LEVEL
ANALYSIS

(THE CASE OF MEKELLE CITY)

A research proposal submitted to the department of economics in partial

fulfillment of the requirement for the degree of Bachelor of Science in
economics

By: Hannibal Birhane

ID.No: 0239/04

Advisor: Tefera Kebede (MSc.)

May 201
 ACKNOWLEDGMENT
First of all I thank God and his mother Saint Mary. Through their compassion and benevolence I
was able to pass through most of the adversity I encountered.

Next I would like my deepest appreciation to my advisor Mr. tefera kebede (MSc) for his
invaluable comments, suggestions, encouragements and corrections. All this work is a fruit of his
tenacious effort.

Last but not least I would like to express my deepest gratitude to my parents for their moral and
financial support. The man I am today is all their effort and they deserve my credit. I am also
thankful to all my family members. I have no words to express my appreciation but only say
“Thank you”.

ABSTRACT

I|Page
This study examines the factors affecting the possibility of investing in
manufacturing sector in Mekelle city. In addition this paper investigated the major
constraints faced by manufacturers in the city. To accomplish this total of 94
investors who invested in Mekelle spanning from 1996 up to 2005 were asked.
This paper applies heteroskedastic probit model which focuses on the
heteroskedasticity problem that is ignored in most of the probit applications. The
results obtained from the study indicate while education level of the investor,
relative prices and initial capital have a positive significant effect on the decision to
invest in manufacturing average power blackout per weeks have a negative
significant effect. Furthermore many constraints were mentioned by the investors
among them lack of credit, lack of access to land and unreliable infrastructure. The
policy implications to boost investment in manufacturing sector are clear. Make
the provision of infrastructure especially electricity reliable and provide a special
set of incentives for the manufacturing sector in accordance with its peculiarities.

Key words: manufacturing sector, Jorgenson model of optimal capital

accumulation, heteroskedastic probit, probit

II | P a g e
 ACRONYMS

CGE CURRENT GOVERNMENT OF ETHIOPIA

CSA CENTRAL statistics agency

EDR Ethiopian Development Research Institute

EEA/EEPRI Ethiopian Economics Association/Ethiopia

Economic policy Research Institute

FDI Foreign Direct Investment

GDP Gross Domestic Product

GTP Growth and Transformation Plan

LDC Least Developed Countries

MoFED MINISTRY of finance and economic development

NBE National Bank of Ethiopia

UNCTAD United Nations Commission for Trade and

Development

III | P a g e
 TABLE OF CONTENTS
ACKNOWLEDGEMENT I
ABSTRACT II
ACRONYMS III
LIST OF TABLES VI
LIST OF FIGURES VII
CHAPTER ONE 1

1. INTRODUCTION 1
1.1 BACKGROUND OF THE STUDY 1

1.2 STATEMENT OF THE PROBLEM 3

1.3 OBJECTIVES OF THE STUDY 6

1.3.1 GENERAL OBJECTIVES OF THE STUDY 6

1.3.2 SPECIFIC OBJECTIVES OF THE STUDY 6

1.4 JUSTIFICATION OF THE STUDY 6

1.5 SCOPE OF THE STUDY 7

1.6 LIMITATION OF THE STUDY 7

1.7 DATA AND METHOLGY 7

1.7.1 DATA COLLECTION 7

1.7.2 DATA ANALYSIS 8

CHAPTER TWO 10

2. LITERATURE REVIEW 10

2.1 THEORIES OF INVESTMENT 10

2.2 STRUCTURAL CHANGE MODELS 12

2.3 THE ROLE OF MANUFACTURING SECTOR IN ECONOMIC GROWTH 15

2.4 FACTORS THAT AFFECT GROWTH IN MANUFACTURING SECTOR 17

IV | P a g e
 2.5 DEINDUSTRALIZATION AND DUTCH DISEASE 18

2.6 PERFORMANCE OF MANUFACTURING SECTOR IN ETHIOPIA 21

2.7 EMPIRICAL REVIEW 22

CHAPTER THREE 25

3. METHODOLOGY OF THE STUDY 25

3.1 DESCRIPTION OF THE STUDY AREA 25

3.2 THEORETICAL FRAMEWORK 25

3.3 DEFINITION OF VARIABLES 26

3.4 ECONOMETRIC FRAMEWORK 29

CHAPTER FOUR 32

4. RESULTS AND DISCUSSION 32

4.1 PATTERNS OF INVESTMENT IN ETHIOPIA’S MANUFACTURING SECTOR 32

4.2 PATTERNS OF INVESTMENT IN MANUFACTURING SECTOR IN MEKELLE CITY 36

4.3 DESCRIPTIVE ANALYSIS OF THE SURVEY 39

4.4 ECONOMETRIC ANALYSIS 45

4.4.1 THE HETEROSKEDASTIC PROBIT MODEL ESTIMATION RESULT 45

4.4.2 MARGINAL EFFECTS OF INDEPENDENT VARIABLE 48

4.4.3 A MEASURE OF ELASTICITIES OF INDEPENDENT VARIABLE 50

4.4.4 ASSESSING HETEROSKEDASTIC PROBIT MODEL FIT 52

CHAPTER FIVE 55

5. CONCLUSION AND RECOMMENDATION 55

5.1 CONCLUSION 55

5.2 RECOMMENDATION 56

REFERENCES i

APPENDIX I iv

V|Page
APPENDIX II vii

LIST OF TABLES
Table 3.1: Description of variables 28

Table 4.1 Investment projects distribution by sector from 2002/03-2011/12 33

Table 4.2: Investment projects distribution by sectors 1996-2005 in Mekelle city 38

Table 4.3: Personal information of respondents 40

Table 4.4: Educational level of respondents 40

Table 4.5: Initial capital of investors 41

Table 4.6: Perception on the prevailing interest rate 41

Table 4.7: Collateral requirement as percent of principal of last loan 42

Table 4.8: Average power blackout in weeks per hour of sample investors 42

Table 4.9: Discount rate of individual investors in Mekelle 43

Table 4.10: Results of Probit and Heteroscedastic Probit Model 46

Table 4.11: Marginal effects of Heteroshedastic Probit Model 49

Table 4.12: Average marginal effects (ey/ex) of independent variables 50

Table 4.13: Average marginal effects (ey/dx) of independent variables 51

Table 4.14: Average marginal effects (dy/ex) of independent variables 52

Table 4.15: Goodness of fit of the heteroskedastic probit model 53

Table A-1: Results of the heteroskedastc probit model iv

VI | P a g e
Table A-2: Results of the probit model iv

Table A-3: Marginal effects of independent variables v

LIST OF FIGURES
Fig 2.1: The Lewis Model of Modern-Sector Growth in a
Two-Sector Surplus-Labor Economy 13

Fig 2.2: Three Sector Growth Hypothesis 15

Fig 2.3: Share of Agricultural, manufacturing and service GDP 22

Fig 4.1: Percentage share of sectors in total investment

measured by number of projects 34

Fig 4.2: Percentage share of sectors in total investment measured

by investment capital 34

Fig 4.3: Percentage share of manufacturing and non-manufacturing in

total industry investment in terms of number of projects and investment capital 36

Fig 4.4: Percentage share of each sectors’ investment in total investment

registered in Mekelle city measured by number of projects 37

Fig 4.5: Percentage share of each sectors’ investment in total investment

registered in Mekelle city measured by investment capital 37

Fig 4.6: Top business constraints faced by sample investors in Mekelle city 44

Fig 4.7: The relationship between initial capital and investment

in manufacturing sector 48

Fig A-1: Average marginal effects (ey/ex, ey/dx and dy/ex respectively) vi

VII | P a g e
 CHAPTER ONE

INTRODUCTION

1.1 BACKGROUND OF THE STUDY

Investment spending is a central topic in economics for two reasons. First, fluctuations in
investment account for much of the movement of GDP in the business cycle. Second, investment
spending determines the rate at which the economy adds to its stock of physical capital, and thus
helps to determine the economy’s long-run growth and productivity performance. Faster growing
countries generally invest a higher share of their GDP than slower growing countries.
(Dornbusch and Fisher, 1995)

Investment spending plays a key role in long-run growth and also in the short-run business cycle
because it is the most volatile component of GDP. When expenditure in goods and services falls
during a recession, much of the decline is usually due to a drop in investment. While spending on
consumption goods provide utility to households today, spending on investment goods is aimed
at providing a higher standard of living at a later date. Investment is the component of GDP that
links the present and the future. (Mankiw, 2009)

Many policy makers and academicians contend that investment can have important positive
effect on a country’s development effort. In addition to the direct financing capital it supplies,
investment can be a source of valuable technology and know-how while fostering linkages
between firms, which can help jumpstart an economy. Based on these arguments, industrialized
and developing countries have offered incentives to encourage investment in their economies.
(Alfaro, 2003)

As with LDCs, investment is crucial for LDCs in their aspiration for development. LDCs need to
increase investments related to the development of productive capacities (namely productive
resources, entrepreneurial capabilities and production linkages). They also need to significantly
step up investments which are related to productive capacities (especially in infrastructure and
institutions). Only if LDCs overcome these constraints and successfully develop productive
resources, will they benefit from a more favorable process of capital accumulation, technological
progress and structural change. (UNCTAD, 2007)

1|Page
When deciding upon which sectors to invest, the investor has a multiple of choices to invest into.
One of them is manufacturing sector. What is manufacturing? Manufacturing is the production of
goods for use or sale using labor and machines, tools, chemical and biological processing, or
formulation. The term may refer to a range of human activity, from handicraft to high tech, but is
most commonly applied to industrial production, in which raw materials are transformed into
finished goods on a large scale. According to UNIDO it is the physical or chemical
transformation of materials or components into new products, whether the work is performed by
power-driven machines or by hand, whether it is done in a factory or in the worker’s home, and
whether the products are sold at wholesale or retail. The assembly of the component parts of
manufactured products is also considered as manufacturing activities. It includes many fields and
sub fields under it. But according to American bureau of labor statistics it is classified under six
general sectors:

 Clothing and Textiles

 Petroleum, Chemicals and Plastics
 Electronics, Computers and Transport
 Food
 Metals
 Wood, Leather and Paper

Since Ethiopia declared a free market economy policy investment activity started reviving
parallel to the reformed policies of investment (Asmelashe, 2007).Thus, the level of investment
in the country during the last two decades (1975-1994) was fluctuating. In 1994, the share of
private investment total investment was 39.5%. It was 11.7% in 1990 and1991 when dergue was
at the helm of the power. One that should be taken note of is low private investment in
manufacturing sector and the heavy involvement of the public sector in investment. (Workie,
1996)

Tigray regional state is one of the preferred destinations for investment. Though not adequate by
any measure different investors have been flowing into the region. From 2006 to 2009 G.C 2008
projects whose initial capital is 12.8 billion birr and expected to create 92000 jobs gained license
from the respective authority (statistical bulletin, 2008). Mekelle city being the capital city of the
region has been the major recipient of this inflow. Between 1993 and 2002, the city received 65
percent of investment in the Tigray region. One point that should be noted here is the lack of
desire of private investors to invest in the manufacturing sector and the declining trend of
investment in the sector in Mekelle city. Accordingly this paper could be helpful to answer the
above question by asking what determines investment in manufacturing sector at micro level.

2|Page
 1.2 STATEMENT OF THE PROBLEM

The Ethiopian economy is dominated by the agriculture and services sectors—with each
accounting for about 45 percent of gross domestic product (GDP), leaving only about 10 percent
for industry, of which manufacturing accounts for about 6–7 percent. Exports are highly
concentrated, with coffee alone accounting for more than 60 percent of the total. (Investment
climate assessment report, 2004)

Moreover, Ethiopia could hardly be located in the international market for manufacturing
exports, having an industrial export share much less than the already minuscule median for
Africa. The limited change in the structure of the economy, especially with regard to
manufacturing, is partly explained by the low levels of investment flows and the sluggish growth
of the private sector, which was too little to affect its historically low share in labor- intensive
manufactures. Indeed, even after more than a decade of reforms by the current Government of
Ethiopia (GOE) private economic activities in the Ethiopian manufacturing sector remain very
small, even by African standards. (IBID)

The declining trend in the contribution of agriculture to GDP has not been compensated by a
growing share of industry in GDP. Modern manufacturing remained small, stagnating at around
8 per cent of GDP and providing 5 per cent of employment in 1989. Without much notice,
agriculture recently ceased to be the largest sector in the economy for the first time in Ethiopia’s
history. It is forecasted that the services sector—which covers real estate, hotels, transportation,
communication, banking, health and education and retail trade—will make up more than half of
Ethiopia’s GDP in just two years’ time, a development with many implications for Ethiopian
business. (Access capital, 2010)

The current GTP dictates that at the end of the term of the plan the industrial sector should play a
pivotal role in the economic growth of Ethiopia. The government even resorted to using cohesive
power to induce investors in manufacturing sector. Despite this attempt by the government to
increase investment in manufacturing, the investment in manufacturing sector has either
stagnated or decreased.

The stagnation of manufacturing sector could be attributed to less than optimal investment levels
in manufacturing sector. The reason for this could be credited to Ethiopian business environment
being suitable for investments in the service sector. Both the government and the private sector
invest in the service sector. The reason for low concentration of investment in manufacturing can

3|Page
be attributed to lack of access to finance due to bank policy. In addition investors are happy to
engage in making easy money which contradicts with manufacturing sector which has high risk.
Manufacturing sector is also susceptible to fluctuations in prices and exchange rates which make
it less attractive to prospective investors. (Berihu Assefa, 2010)

The structural transformation we see in Ethiopia is not consistent with the Petty-Clark law which
predicts that the center of gravity in economic activities shifts (through inter-sectorial resource
allocations by the market) from agriculture to the industrial sector, and further, to the service
sector as average per capita income continues to rise (IBID). The question here should not be the
case of Ethiopia following the law or not but what implications could this transformation bring
into the Ethiopian economy.

Though all sectors of the economy are important and indispensable there are some peculiar
characteristics of manufacturing sector which makes it very important for a big and poor country
like ours. These characteristics direct to the idea that manufacturing sector should be given
priority over the other sectors in developing world.

One of the features is that mostly manufacturing is a wealth-producing sector of an economy. In

addition one of the peculiar characteristics of manufacturing sector is its capability to provide
important material support for national infrastructure and for national defense.

An argument in favor of the manufacturing sector is the ‘Development argument’. One can
generally observe that economic development industry goes with: the more developed nations
are better industrialized, and the less developed nations are less industrialized. In addition this
sector has huge potential for employment creation. The industrial sector has more potential to
create job opportunities for the rapidly growing urban populations of developing countries than
any other sector. (Obstfeld, 2008)

A developed manufacturing sector, in general, generates more foreign currency, compared to an

agricultural sector. That is, industrialization helps developing nations to alleviate their balance-
of-payments problems. In addition if industrial development is directed to use local raw
materials, it can create strong linkages among the different sectors of the domestic economy.
This is due to the dynamic nature of the industrial sector which includes positive externalities in
consuming agricultural raw materials.

The more fascinating point with manufacturing sector is its ability to invent and innovate due to
the high level of product differentiation in the sector. Even the industrial revolution that altered
the history of human being for good was as a result the invention of machines that replaced the

4|Page
handicraft production. It is not striking to see countries with high level of R&D are in fact
countries with significant manufacturing base.

The afro mentioned points have been proven right by different researches including the
following one. While LDCs that continue to specialize in non-oil commodities have seen an
increase of extreme poverty in the past decades, the LDCs that specialize in manufacturing have
poverty. Between the late 1980s and the late 1990s, extreme poverty in non-oil commodity
exporting LDCs rose from 67 percent to 69 percent, an increase from the early 1980s, but it fell
in manufacture exporting LDCs from 48 percent to 44 percent. (UNCTAD, 2007)

To have a large manufacturing sector there needs to be huge investment in manufacturing sector
which is obviously lacking in Ethiopia. Mobilization of domestic and foreign investible
resources for industrialization requires a favorable climate which induces both domestic and
foreign investors to commit their capital in industrial development. (Bulti Terfessa, 1990)

In this paper the researcher in line with what have been mentioned above intends to answer the
following questions:

 What is the trend of private investment in manufacturing sector in Ethiopia as a whole and
Mekelle particularly?
 What should local governments and other respective authorities do to create a favorable
climate for manufacturing sector investors?

Much of the prior research on investment has focused on macro level models using aggregate
data. Such models are useful in identifying gross investment relations and have been
instrumental in estimating the effect of changes in national policy variables. However, because
of the level of aggregation of data used, these models tend to abstract from the behavior of
individuals and treat the entire economy as a single entity. Such a procedure treats the differing
behavior of individuals on a net basis, potentially resulting in the appearance of no action when
individual’s actions are offsetting, and fails to identify the behavioral characteristics of those
who are reacting to investment stimuli compared to those who don’t invest or disinvest. In this
paper an attempt will be made to assess micro level factors related with economical and
behavioral characteristics of the investors and their effects in investment decision.

5|Page
 1.3 OBJECTIVES OF THE STUDY
1.3.1 GENERAL OBJECTIVES OF THE STUDY

The general objective of the study is to assess the determinants of investment in manufacturing
sector at micro level in Mekelle city.

1.3.2 SPECIFIC OBJECTIVES OF THE STUDY

The specific objectives of this research are:

 To investigate the status or intensity of investment in manufacturing sector both at national

and city level
 To identify the major determinants affecting investment in manufacturing sector in Mekelle
city
 To examine the challenges that investors in manufacturing sector of Mekelle city encounter

1.4 JUSTIFICATION OF THE STUDY

Much of the research on manufacturing and investment focuses on topics like:

 The nature and sources of the considerable variations that occur cross-nationally in levels of
manufacturing and wider industrial-economic growth;
 Competitiveness; and
 Attractiveness to foreign direct investment

This clearly shows lack of ample researches on the micro level factors affecting investment
decision. Even those available are descriptive by nature. Attempts to quantify the variables have
been limited. Hence this paper comes into picture to fill this gap. Furthermore most of the
researches done in investment decision only consider the external environment facing the
investors. However, in this paper an attempt has been made to study investment behavior.
Accordingly this paper can be considered as an alternative by treating the external environment
which the investor could not control and his/ her preference in which he/she has absolute control
at the same time.

6|Page
In addition this paper can be an important tool for policy makers. In view of the fact that the
problem is wide and needs multi-disciplinary survey this paper also can serve as a basis for
fellow researchers.

1.5 SCOPE OF THE STUDY

As indicated in the objective, the aim of this study was to identify important microeconomic
variables that affect investment in manufacturing sector in the study area. In order to identify the
important variable at micro level, the survey was conducted at Mekelle city by using structured
questionnaire.

This study is has temporal and spatial delimitations. The study is delimited to investors from
1996-2005. Investors who invested prior to that period are excluded. It is also spatially limited
that it only considers investors in Mekelle city. Moreover the study is also only concerned with
microeconomic variables. Macroeconomic variables which have huge impact on the decision of
investors are overlooked.

1.6 LIMITATION OF THE STUDY

Investment decision at micro level is relatively a new topic. The study has been hindered by lack
of ample researches which study the determinants of investment at micro level. The paper was
also limited by the finances of the researcher. Due to this the study was only limited to Mekelle.
This makes it insufficient to conclude to other areas. The results of this study need to be
understood by taking in to account the above limitations.

1.7 DATA AND METHODOLOGY

1.7.1 DATA COLLECTION

The researcher used both primary and secondary data sources. Both sources are crucial for the
research and will be used to analyze their effect on the dependent variable.

Questionnaire was used as the main technique of collecting primary data. The target population
only included investors who participate either in manufacturing or in non-manufacturing but not
in both. If investors who participate in both sectors are included it will raise the question of
statistical dependency between the samples and it could not be estimated using models which
assume statistical independency. This means investors who participate in both sectors are
excluded from the sample.

7|Page
The target population was investors in Mekelle who gained license and invested from 1996 up to
2005 E.C. The target population did not include investors who gained authorization to expand an
existing firm, i.e. only new investments were concerned. In line with the above information the
total number of investors from 1996-2005 is 1550. They were divided into two groups, i.e.
investors in manufacturing sector and investors who invested in other sectors. As a result there
are 448 investors in manufacturing and 1102 in non-manufacturing sector.

To figure out my sample size I used the slovin’s formula which is used in determining a sample
size when there is no enough information about the distributional pattern of the population.
Slovin’s formula is expressed as:

N
n= 2
1+ N e

Where: n=number of samples

N=Total population

e=error tolerance∨margin of error

Using 10% level of error or 90% level of confidence 94 samples were taken from the total target
population. Since I was encountered with a heterogeneous population, stratified random
sampling was applied which is a type of probability sampling. I chose stratified random sampling
because it is useful in increasing a sample’s statistical efficiency and in providing adequate data
for analyzing the various subpopulations. The type of stratification thought appropriate for this
research is proportional sampling. After stratification the researcher used simple random
sampling to select individual investors out of total population. So this means there will be 27
respondents for manufacturing and 67 for other non-manufacturing sectors.

Most of the secondary data was collected from Mekelle investment bureau. The data from
mekelle investment bureau portrays the year to year basis investment in mekelle city. In addition
I used NBE annual report, EIA publications and other internet sources.

1.7.2 DATA ANALYSIS

Both descriptive and econometric techniques were employed to study the different variables. By
applying descriptive statistics one can compare and contrast different categories of sample units
with respective to desired characteristics. The descriptive statistics that were included in this

8|Page
study are mean, standard deviation percentage frequency occurrence. Furthermore this research
will make use a deductive approach.

This study deals with categorical variables. There are three most common methods of estimating
categorical variables, viz. linear probability model (LPM), probit (normit), and logit. The LPM
method is plagued by different problems including Unbounded Predicted, Conditional
heteroskedasticity, Non-Normal Errors, and Functional Form. This lives as with probit and logit
as the best techniques of estimating categorical variables. Furthermore the setup of the logit and
probit models is essentially the same. However the shortcomings of both standard probit and
logit model are the priori assumption of homoskeadsticity. But if there is heteroskedasticity in
the variances the parameter coefficients estimated will be biased, inconsistent and inefficient. In
addition the standard errors are wrong. Thus to solve the problem of heteroskedasticity the paper
applied the statistical estimation technique of heteroskedastic probit.

9|Page
 CHAPTER TWO

LITERATURE REVIEW
Investment in manufacturing sector in Ethiopia is not adequate as it should be when a country is
in transitional stage. Especially the private sector investment in manufacturing sector is at low
stage. In addition manufacturing sector remains by far the least attractive sector to indigenous
investors. Though not sufficient some researchers have been done on the treatise of investment in
manufacturing sector. This chapter will deal on some studies related with investment in
manufacturing sector.

2.1 THEORIES OF INVESTMENT

A) New Classical Theory of Capital Accumulation

Jorgenson and others propounded this theory. According to the new classical theory the demand
for factors of production is responsive to changes in relative factor prices to the price of output
(Jorgenson in Tanwar 1978). In other words, the cost of capital induce changes in the investment
behavior by changing the implicit rental price of capital services and further the changes in the
rental price of capital services lead to changes in the desired stock of capital (Tanawar 1978).

The marginal product of capital determines the real rental price of capital. The real interest rate,
the depreciation rate, and the relative price of capital goods determine the cost of capital.
According to the neoclassical model, firms invest if the rental price is greater than the cost of
capital, and they disinvest if the rental price is less than the cost of capital. (IBID)

B) The Profit Theory of Investment

The central theme of the Profit Theory of Investment is, “greater the gross profits, greater will be
the availability of internal funds for investment”. Some empirical studies have found significant
correlation between investment and profit. Eisner found significant correlation between profit
and investment. As we know it’s mainly profit that derives investment in rich capitalist nations.
10 | P a g e
The role of profit as a determinant of investment has been widely accepted and confirmed in
several studies by Roos (1948), Klein (1951) and Grunfeld (1960, as cited in Suresh, 1997) and
Tinbergen (1938). Studies by Hoshae (1991), Band and Meghil (1994) and Tibout (1983, as cited
in Suresh, 1997) conclude profit as a major determinant of investment. They found that there
exists a positive relationship between profitability and investment.

C) Liquidity Theory of Investment

The Liquidity Theory of Investment says, “Desired capital is proportional to liquidity” (Tanawar
1978).

Writing in equation form

¿
k t =αLt

¿
Where k t is the desired capital investment at time‘t’.

 Is the desired rate of capital to the flow of the availability of the internal fund for investment at
time ‘Lt.’

Empirical study by Tanwar (1978) found that, the flow of internal fund 1 is significant in affecting
investment.

D) Acceleration Principle of Investment

J.M. Clark for the first time put forward the Theory of Acceleration. According to the
Acceleration Principle there exist a direct positive relationship between the rate of change in the
flow of output and addition to the stock of capital. Latter on H.B. Chenery modified it. He
considered the level of output instead of the rate of change of output. So the modified
acceleration principle assumes a relation between output and capital stock.

The Principle of Acceleration has been examined in various studies. Studies by Krishnamurthy
and Shastry (1972) and Tanawar (1978) found that acceleration hypothesis has some validity and
is an important determinant of investment.

The accumulation of real physical capital stock has long been regarded as one of the major
factors of economic development (Wai and Wong, 1982). The strong relationship between fixed

1
The availability of internal fund is the net profit.

11 | P a g e
capital formation and GDP growth has been established by many writers including De Long,
Lipsey and Kravis.

Wai and Wong, (1982) in their study examined the modified version of the flexible accelerator
theory of investment. They found that government investment; changes in bank credit to the
private sector are crucial determinants of private investment in developing countries. (As cited in
Santosh Kumar, 2002)

Keynes (1936) in his book” The General Theory of Employment, Interest & Money” has
discussed the role of rate of interest, as a determinant of investment. According to Keynes a rise
in the rate of interest discourages the investment activity.

The majority of the studies shows the validity of acceleration principle and identifies some
important determinant of investment. They are capital, profit, rate of interest, size of the market
etc. (Suresh, 1997).

2.2 STRUCTURAL CHANGE MODELS

Structural-change theory focuses on the mechanism by which underdeveloped economies

transform their domestic economic structures from a heavy emphasis on traditional subsistence
agriculture to a more modern, more urbanized, and more industrially diverse manufacturing and
service economy. It employs the tools of neoclassical price and resource allocation theory and
modern econometrics to describe how this transformation process takes place. This section will
discuss some of the renowned structural change models which can be helpful in understanding
how a country transforms itself from agrarian into industrialized nation.

A) The Lewis Theory of Development

One of the best-known early theoretical models of development that focused on the structural
transformation of a primarily subsistence economy was that formulated by Nobel laureate W.
Arthur Lewis in the mid-1950s and later modified, formalized, and extended by John Fei and
Gustav Ranis. (Todaro, 2012)

In the Lewis model, the underdeveloped economy consists of two sectors: a traditional,
overpopulated rural subsistence sector characterized by zero marginal labor productivity—a
situation that permits Lewis to classify this as surplus labor output—and a high-productivity
modern urban industrial sector into which labor from the subsistence sector is gradually
transferred. The primary focus of the model is on both the process of labor transfer and the

12 | P a g e
growth of output and employment in the modern sector. The speed with which this expansion
occurs is determined by the rate of industrial investment and capital accumulation in the modern
sector.

In conclusion Lewis two-sector model is a theory of development in which surplus labor from
the traditional agricultural sector is transferred to the modern industrial sector, the growth of
which absorbs the surplus labor, promotes industrialization, and stimulates sustained
development.

Fig 2.1: The Lewis Model of Modern-Sector Growth in a Two-Sector Surplus-Labor

Economy

B) Patterns of Development

This model was developed by Hollis B. Chenery. Like the earlier Lewis model, the patterns-of-
development analysis of structural change focuses on the sequential process through which the
economic, industrial, and institutional structure of an underdeveloped economy is transformed
over time to permit new industries to replace traditional agriculture as the engine of economic
growth. (Todaro, 2012)

However, in contrast to the Lewis model and the original stages view of development, increased
savings and investment are perceived by patterns-of-development analysts as necessary but not
sufficient conditions for economic growth. In addition to the accumulation of capital, both
physical and human, a set of interrelated changes in the economic structure of a country are

13 | P a g e
required for the transition from a traditional economic system to a modern one. These structural
changes involve virtually all economic functions, including the transformation of production and
changes in the composition of consumer demand, international trade, and resource use as well as
changes in socioeconomic factors such as urbanization and the growth and distribution of a
country’s population.

Empirical studies, both cross-sectional and time-series of countries at different levels of per
capita income led to the identification of several characteristic features of the development
process. These included the shift from agricultural to industrial production, the steady
accumulation of physical and human capital, the change in consumer demands from emphasis on
food and basic necessities to desires for diverse manufactured goods and services, the growth of
cities and urban industries as people migrate from farms and small towns, and the decline in
family size and overall population growth as children lose their economic value and parents
substitute what is traditionally labeled child quality (education) for quantity with population
growth first increasing and then decreasing in the process of development.

C) Three Sector Growth Hypothesis (Petty-Clark law)

Three-sector Growth hypothesis is an economic theory which divides economies into three
aggregated sectors of activity: agriculture and extraction of raw materials (AGM), and
manufacturing (IND) and services (SRV). In was initially developed by Clark (“Conditions of
Economic Progress", 1940, 1951, 1957). Clark has stressed the dominance of different sectors
(Agriculture + Mineral Industry, Manufacturing and Services) of economy at different stages of
its development and modernization.

Clark introduced the analysis of comparative performance in three main sectors of economy and
initiated the discussion by quoting what he called the Petty's Law; "There is more to be gained by
manufacture then husbandry, and by merchandise than manufacture." He stressed the
significance of the three-sector breakdown and structural change in interpreting economic
growth.

According to Fourastie the main focus of an economy's activity shifts from the primary activity
(Agriculture and mineral industry) through to the secondary activity (Industry) and finally to the
tertiary activity (service). Forastie (1949) saw the process as essentially positive and writes of
the increase in quality of life, social security, blossoming of education and culture, higher level
of qualifications, humanization of work and avoidance of unemployment. To formalize it let

14 | P a g e
 be the average multiplicative rate of change for of these sectors; over time.
The three-sector growth hypothesis, including the Petty' Law means the following ordering:

This process of structural transformation has been followed by current developed countries. But
many developing countries including Ethiopia are following processes that are very distinct from
the above process. The share of services in output is high at relatively low income per capita in
many developing countries in Africa and Latin America. This is not the case for Asian countries
that are mostly following the path of developed countries.

Fig 2.2: Three Sector Growth Hypothesis

2.3 THE ROLE OF MANUFACTURING SECTOR IN

ECONOMIC GROWTH

Economists have for a long time discussed the causes of economic growth and the mechanisms
behind it. In the last two decades, in particular, a revived interest on this topic arose with the
upsurge of ‘new growth’ (or ‘endogenous’ growth) models, after Romer (1986, 1990) and Lucas
(1988).

Nicholas Kaldor was one of the first to consider the role of increasing returns in economic
growth. Contrarily to endogenous growth theory and its focus on supply-side issues, however,

15 | P a g e
Kaldor’s perspective emphasized the importance of the exogenous components of demand in
explaining economic growth in the long run.

Kaldor’s First law (1966), an important stylized fact in the growth trajectory of developed
economies in the postwar period is the relationship between industrial growth and the
performance of the economy as a whole. This observation is the origin of Kaldor’s first law,
which states that there is a close relation between the growth of manufacturing output and the
growth of GDP. Kaldor’s first law can be summed up in the expression “manufacturing is the
engine of growth”, and was first estimated by Kaldor in a cross section of developed countries
over the period 1952-54 to 1963-64. The law can be represented by following regression:

q i ¿ a i+ bi mi (1)

where q and m refers to growth of total output and manufacturing output, respectively.

It is important to note that the correlation between the two variables is not only due to the fact
that industrial output represents a large component of total output. The regression coefficient is
expected to be positive and less than unity, which means that the overall growth rate of the
economy is associated with the excess of the growth rate of manufacturing output over the
growth rate of nonmanufacturing output. This proposition implies that high growth rates are
usually found in cases where the share of manufacturing industry in GDP is increasing, and it
can be tested using the equation:

q i=c i +d i (mi−nm i ) (2)

where nm refers to the growth of non-manufacturing output.

As additional evidence supporting the statement that “manufacturing is the engine of growth”,
Kaldor has argued that the growth is non-manufacturing output also responds positively to the
growth of manufacturing, as described in the following equation:

nm i=ui +v i mi (3)

The explanation for the correlation between the growth of manufacturing output and the overall
performance of the economy is to be found on the impact of the former on the growth of
productivity in the economy. There are two possible reasons for such effect. The first relates to
the fact that the expansion of manufacturing output and employment leads to the transfer of labor
from low productivity sectors (or disguised unemployment) to industrial activities (that present

16 | P a g e
higher productivity levels). The outcome is an increasing overall productivity in the economy
and little or no negative impact on the output of the traditional sectors, given the existence of
surplus labor. According to Kaldor, this process is characteristic of the transition from
“immaturity” to “maturity”, where an “immature” economy is defined as one in which there is a
large amount of labor available in low productivity sectors that can be transferred to industry.

The second reason for the relation between manufacturing growth and productivity relates to the
existence of static and dynamic increasing returns in the industrial sector. Static returns relate
mainly to economies of scale internal to the firm, whereas dynamic returns refer to increasing
productivity derived from learning by doing, ‘induced’ technological change, and external
economies in production.

Kaldor’s Second Law (Verdoorn’s Law) Law refers to the statistical relation between the growth
of manufacturing output and the growth of labor productivity in manufacturing, where causality
runs primarily from the former to the latter. The basic argument is that an initial growth in output
induces productivity gains that allow for reduction of unit labor costs and, given a mark-up
pricing rule, for fall in prices, increasing the competitiveness of a country or region. These gains,
in turn, allow for further output expansion through increasing exports, which reinitiate the cycle.
In conclusion, once a country or region acquires a growth advantage, it will tend to keep it
through the process of increasing returns and consequent competitive gains that growth itself
induces. (Libanio and Moro, 2009)

Different empirical studies has confirmed kaldor’s first and second law (Libanio and Maro for
Latin American countries) there by confirming statement that manufacturing sector is the engine
of growth. So for developing countries to have robust economic growth, a manufacturing sector
growth is necessary condition.

2.4 FACTORS THAT AFFECT GROWTH IN

MANUFACTURING SECTOR

The central core of economic growth of any nation is influenced by the growth rate of the
industrial sector specially manufacturing industries. The major components of this activity
involve labor productivity, capital formation, improvements in technology and the market
environment (UNIDO, 1996). (As cited by Urgaia Rissaa)

Labor productivity: Output per worker. Labor productivity in economic growth refers to how
much output per hour, week and so on, results from labor-input. Labor productivity depends
upon a number of factors including the population size (quantity) and degree of education and

17 | P a g e
skills (quality) of labor supply, the capital stock other resources; each laborer has to work with
and the technology available for production.

Capital formation: The promise to invest capital on some activity. It is the initial stage of
accumulation of money either through savings domestically or borrowing from financial
institution. Labor productivity is enhanced when capital stock available to labor growth is faster
than the labor supply. This capital is said to be deepening capital.

Improvements in technology: Technological growth and invention, improvements in methods by

which goods and services are produced and sold is another key to economic growth. The
contemporary high living standard of the developed world is unthinkable without the invention
of items such as steam engine, assembly – line production and etc. Technology and invention
have been responsible for nothing less than the modern world.

The market environment: Most economists believe that the market environment surrounding the
growth in natural resources, labor productivity, capital formation and so on, are a large
contributing factors to economic growth. A competitive market system may encourage invention
and rapid innovation. All private and public restraints on competition, including excessive or
unnecessary regulation or tariffs or quotas on goods exchanged in an international and trade, will
trend to reduce the rate of growth in real per capital GNP.

Industrial Policy: Industrial policy concerns with the effective and coherent implementation of
all policies, which enhance the structural adjustment of industry to promote competitiveness.
Then after, the infant industries (like Ethiopian manufacturing case) can develop and prosper by
remedying the structural deficiencies and addressing areas where the market mechanism alone
fails to correct. Though industrialization in Ethiopia was initiated three -quarters of a century
ago, manufacturing sector is still in an incipient stage characterized by unbalanced and distorted
internal and external structural linkage, backward technology and lack of technological
capability (Kibre and Worku, 2003/04).

2.5 DEINDUSTRALIZATION AND DUTCH DISEASE

The Petty-Clark law states at some time the tertiary sector will replace the secondary sector as
the most important sector in the economy in terms of GDP or employment. The point where this
substitution takes place is called deindustrialization. This phenomenon has been evident in
developed countries like USA, Europe and Japan. It has also been observed more recently in
developing countries like the Four Tiger economies of East Asia (Hong Kong, China, Korea,
Singapore, and Taiwan).

18 | P a g e
Though this phenomenon can be an opportunity for developing countries which have low labor
costs (although Ethiopia didn’t use this opportunity) it is of huge concern in advanced countries.
Many regard deindustrialization with alarm and suspect it has contributed to widening income
inequality in the United States and high unemployment in Europe.

During deindustrialization, the declining share of employment in manufacturing appears to

mirror a decline in the share of manufacturing value added in GDP. At first glance, this decline
would suggest that domestic expenditure on manufactures has decreased while expenditure on
services has increased.

Closer analysis, however, reveals that this conclusion is misleading. Expenditure on services in
current price terms has indeed grown in the advanced economies. But this growth can be
accounted for by the fact that labor productivity (output per worker) has grown more slowly in
services than in manufacturing, pushing up the relative price of services and making
manufactures relatively cheaper. When output in the manufacturing and service sectors is
measured at constant rather than at current prices, however, the shift in expenditure away from
manufacturing to services is nothing like the scale of the shift away from employment in
manufacturing to services. Indeed, at constant prices (in contrast to its steeply falling current-
price share), the share in GDP of value added by manufacturing in the advanced economies was
roughly unchanged between 1970 and 1994. (Rowthorn and Ramaswamy, 1997)

Continued deindustrialization has important implications for long term growth prospects in the
advanced economies. Most obviously, as mentioned in the introduction, if more of the workforce
moves into the service sector, productivity growth within services will probably determine the
outlook for living standards overall.

Certain industries are more amenable to technological progress (that is, have high productivity
growth rates, usually because of their potential for standardization), as opposed to those that are
less amenable to such progress. Manufacturing would appear to be, by its nature, technologically
progressive—with a systematic tendency to find ways to produce more goods with fewer
workers.

Technological developments will likely make it feasible for some services to grow faster than
others, and the service sector will thus undergo significant internal structural changes. Product
innovation in manufacturing will continue to be important, inasmuch as it provides spillover
effects to productivity growth in services.

19 | P a g e
Deindustrialization is also likely to have important implications for industrial relations in the
developed world, and particularly for the role played by trade unions. Trade unions have
traditionally derived their strength from industry, where the modes of production and the
standardized nature of the work have made it easier to organize workers. In services, workers are
typically more difficult to organize (with the possible exception of public services) and
unionization has thus been less prevalent, owing not least to wide differences in the types of
work available.

Another issue linked with deindustrialization is Dutch disease. While deindustrialization has
been hampering the growth of manufacturing sector in the more advanced countries Dutch
disease has been plaguing the manufacturing sector in the resource endowed developing
countries.

The Dutch disease is the apparent relationship between the increase in exploitation of natural
resources and a decline in the manufacturing sector. The mechanism is that an increase in
revenues from natural resources (or inflows of foreign aid) will make a given nation's currency
stronger compared to that of other nations (manifest in an exchange rate), resulting in the nation's
other exports becoming more expensive for other countries to buy, making the manufacturing
sector less competitive. While it most often refers to natural resource discovery, it can also refer
to "any development that results in a large inflow of foreign currency, including a sharp surge in
natural resource prices, foreign assistance, and foreign direct investment". (Wikipedia)

The classic economic model describing Dutch Disease was developed by the economists W. Max
Corden and J. Peter Neary in 1982. In the model, there is the non-traded good sector (this
includes services) and two traded good sectors: the booming sector, and the lagging sector, also
called the non-booming tradable sector. The booming sector is usually the extraction of oil or
natural gas, but can also be the mining of gold, copper, diamonds or bauxite, or the production of
crops, such as coffee or cocoa. The lagging sector generally refers to manufacturing.

A resource boom will affect this economy in two ways. In the "resource movement effect", the
resource boom will increase the demand for labor, which will cause production to shift toward
the booming sector, away from the lagging sector. This shift in labor from the lagging sector to
the booming sector is called direct-deindustrialization. However, this effect can be negligible,
since the hydrocarbon and mineral sectors generally employ few people. The "spending effect"
occurs as a result of the extra revenue brought in by the resource boom. It increases the demand
for labor in the non-tradable, shifting labor away from the lagging sector. This shift from the
lagging sector to the non-tradable sector is called indirect-deindustrialization. As a result of the
increased demand for non-traded goods, the price of these goods will increase. However, prices

20 | P a g e
in the traded good sector are set internationally, so they cannot change. This is an increase of the
real exchange rate.

There are two basic ways to reduce the threat of Dutch disease: by slowing the appreciation of
the real exchange rate and by boosting the competitiveness of the manufacturing sector. One
approach is to sterilize the boom revenues, that is, not to bring all the revenues into the country
all at once, and to save some of the revenues abroad in special funds and bring them in slowly. In
developing countries, this can be politically difficult as there is often pressure to spend the boom
revenues immediately to alleviate poverty, but this ignores broader macroeconomic implications.

2.6 PERFORMANCE OF MANUFACTURING SECTOR IN

ETHIOPIA

Ethiopia has a long tradition in the development of handcrafts and cottage manufacturing
activities such as weaving, blacksmithing, pottery, and woodwork. But the introduction of
modern industries began at the end of the 19th century (History of Ethiopian economy textbook).
Particularly, the following two major early 20th century events contributed to the introduction of
modern manufacturing industries in Ethiopia:

 The emergence of a strong central government, which resulted in political stability and
 The construction of the Ethio-Djibouti railway.

According to the International Standards for Industrial Classification (ISIC), the Ethiopian
industrial sector is composed of mining and quarrying, manufacturing, electricity, water supply,
and construction. During the years 1991/92 - 1997/98, large- and medium-scale manufacturing
contributed 38.4% of the gross value of industrial production, while small-scale manufacturing
and handcrafts contributed around 18.7%. (IBID)

The performance of the manufacturing sector can be measured, among other things, by:

 Gross value of output (GVO), which refers to the total output produced during a given
period of time;
 Value added at factor cost (VAFC), which is the difference between the gross value of
output and the value of intermediate inputs, such as the cost of raw materials;
 Value added at current market price (VACMP), which is the sum of value added at factor
cost and indirect taxes, regardless of any subsidies
 GDP contribution: the share of manufacturing sector in the total output of a country

21 | P a g e
But this paper mainly focuses on the GDP contribution of manufacturing sector to measure the
sector’s performance.

Fig 2.3: Share of Agricultural, manufacturing and service GDP

The above figure demonstrates how stagnant the manufacturing sector (in terms contribution to
GDP) has been for the last three decades. It has never been able to pass its benchmark of 8
percent. But as different theories mentioned above predict, the share of manufacturing sector
should have been rising. On the contrary the economy is rapidly moving from agrarian into
service based economy against the development experience of the now developed countries. The
decline in the share of agricultural GDP was compensated more by the increase in the
contribution of the service sector than that of manufacturing. Furthermore this trend does not
seem to change in the near future because there is no much effort by the public or the private
sector to revive the manufacturing sector.

2.7 EMPIRICAL REVIEW

Many researches have been published related with investment in manufacturing sector, albeit not
in micro level. This section will provide with some of the related literature.

Alan K.Severn (1972) in his study of “Investment and financial behavior of American investors
in manufacturing sector” based on the model of the theory of investment and financial behavior

22 | P a g e
of the firm found that changes in the domestic economic circumstances of individual firms
appear to have little direct impact on the balance of payments of the United States, since
domestic liquidity offsets the substitution between net outflow and domestic investment plus
dividends.

K.A.Bohr (1953) in his paper “Investment criteria for manufacturing sector in underdeveloped
countries” discussed about the selection of manufacturing sector that can be adapted to the
conditions that exist in the underdeveloped world. He concluded that any proper evaluation of
the suitability of an industry must weigh advantages against disadvantages in the context of the
particular situation. The situation can be for example the relative abundance of capital or labor.

Gilberto Libanio and Sueli Moro in their paper entitled “Manufacturing industry and economic
growth in Latin America: A kaldorian approach” analyzed the relation between manufacturing
output growth and economic performance from a Kaldorian perspective by estimating Kaldor’s
first and second growth laws for a sample of eleven Latin American economies during the period
1980-2006. Their results confirm the “manufacturing is the engine of growth” hypothesis, and
suggest the existence of significant increasing returns in the manufacturing sector in the largest
Latin American economies.

Janvier D. Nkurunziza (2005) analyzed issues relating to credit in African manufacturing, not
directly tested for the effect of credit on firm growth. For Nkurunziza, the use of bank credit
could affect firm growth in two opposite ways. The effect might be positive if credit allowed a
firm to address its liquidity constraint and increase profitability. However, if macroeconomic
shocks such as increases in interest rates made firm debts unsustainable as experienced in Kenya
in the 1990s, indebted firms might be shrunk or even collapsed. The researcher used
microeconomic data on the Kenyan manufacturing sector; the study found that conditional on
survival, the firms that used credit grew faster than those did not.

A study by the center for the Study of African Economies (1997) shows that there is very low
level of investment in Africa’s manufacturing sector. Moreover a positive effect from profits
onto investment is identified in a flexible accelerator specification of the investment function
controlling for firm fixed effects.

Urgaia Rissa (2007) in his thesis growth of industrial manufacturing in Ethiopia indicated the
long-term growth rate of investment in manufacturing sector is positively related to the weight
placed on growth of the sector. While that of labor engaged in production, is the short-run effect.
Hence the manufacturing sector in Ethiopia is characterized as labor intensive. In the meantime,
he showed the sector growth is negatively influenced by total factors of production that represent

23 | P a g e
the obsolete uses technological level in manufacturing activities accounts for the sector’s
stagnant growth.

Investment climate assessment report in Ethiopia surveyed by World Bank (2004) reveals three
pivotal features of the Ethiopian manufacturing sector. First, labor costs in Ethiopia are very low
compared with those of potential competitors in and outside Africa (for example, labor costs in
Ethiopia are almost one-third of those in China). Second, despite this huge cost advantage,
exporting Ethiopian firms account for less than 8 percent of the population of firms which is very
small by any standard. Third, the Ethiopian manufacturing is also dominated by small private
firms, which suggests limited firm growth. The manufacturing firms also identified access to
land, access to reliable infrastructure service, banks’ cumbersome credit allocation and the
favorable treatment given to state-owned enterprises (SOEs) as a major business impediment.

24 | P a g e
 CHAPTER THREE
METHODOLOGY OF THE STUDY

3.1 DESCRIPTION OF THE STUDY AREA

Mekelle is one of the largest cities in Ethiopia and the capital of the Regional State of Tigray. It
has a population of 215,546 by 2007 (CSA census, 2007). The city lies in the Ethiopian
highlands 783 km north of Addis Ababa. The city has a temperate climate and low malaria
prevalence due to its elevation (over 2,200 meters).

It is important to note that Mekelle is the primary economic hub in the Tigray region. Within a
100 km radius of the city, there are rich and fertile farmlands to the south, significant mineral
deposits to the east and west, and over one hundred rock-hewn churches throughout the region
that serve as important tourist destinations. Between 1993 and 2002, the city received 65 percent
of investment in the Tigray region.

Most of the investors in manufacturing sector are found on the outskirts of the city. They are
usually located on investment zones arranged by the city administration. Semein sub city has the
highest industrial concentration among the sub cities.

3.2 THEORETICAL FRAMEWORK

One approach to identification of the basic forces influencing investment is to start with the firm
level neoclassical model of optimal capital accumulation (Jorgenson) where net worth (N) of the
firm is given by:

N= ∫ e−rt ¿
0

Where: p= price of production

Q=quantity of output produced

w= price of variable inputs

l=quantity of inputs used

25 | P a g e
 q= price of capital

i=investment ∈durable goods

r =discount rate∨time value of money

Since investment is based on expected future income streams which are not known with
certainty, expected income is probabilistic in nature. To reflect the fact that operators may value
non uniform probabilistic income streams differently, the model must be placed in a utility
framework. Thus, the utility of expected income becomes a basic factor which may influence
investment.

From this model, it is clear that investment is a function of the prices of output, inputs and
capital, the production function which establishes the level of output as a function of the amount
of the amount of the amount of inputs and capital used and the time value of money or discount
rate. This model is theory of optimal investment perse.

This model is very helpful in explaining the working of the economic environment especially the
investment environment. But this doesn’t mean we can totally rely on this model to explain the
entire real world phenomenon. Strict adherence to the classical analysis leaves very important
variables which appear to affect the dependent variable significantly.

In addition this paper is not concerned with the decision to invest or not or how much to invest. It
is more related with which sector to invest. This means this paper will not use the model as it is.
There will be alterations to the optimal investment theory of Jorgensen in this paper. Some
variables will be left out and other variables will be added. In the following section the variables
which are tested against the dependent one will be explained rigorously.

3.3 DEFINITION OF VARIABLES

As explained above though there are many investment theories which can explain how
investment occurs, the Jorgensen model is better at explaining it at micro level. This research’s
objective is to investigate factors that led investors to invest in manufacturing sector.

The dependent variable is invested in manufacturing sector. This variable has a value of 1 if
some random investor invested in manufacturing sector and 0 if he/she invested in other sectors
(i.e. in service or agriculture). The study focuses on investors who invested in one of the sectors
but not both.

26 | P a g e
The model includes six economic and one behavioral (psychological) explanatory variable. One
of the independent variable is education level of the investor. This variable is measured by the
year of education the investor had gained. It is expected to have positive impact on investment in
manufacturing sector. This is because manufacturing sector needs more technological expertise
than other sectors. In addition since manufacturing sector is thought to be riskier business an
educated investor would be better off by making the right decision to minimize risk.

It is assumed manufacturing sector needs more finance than other sectors. This also leads to the
conclusion that if investors are provided with money from different financial institution the
probability that they invest in manufacturing sector increases. In this research access to finance is
measured by a proxy variable in the name of loan. Loan measures the frequency of times loan is
given by bank to an investor in the last 10 years. It is thought to have positive correlation with
investment in manufacturing sector. This is because if investors are provided with better finance
they would be in the sector that have higher profit margin which is manufacturing sector.

Another variable that is assumed to affect investment in manufacturing sector is access to

reliable electricity. An infrastructure the status of electricity has a huge influence on investment
decision. Though much of the firms’ have access to electricity its reliability differs from firm to
firm. A disruption in the transmission of power could have a huge impact on the manufacturing
sector. As a result to measure the reliability of electricity the variable blackout has been used.
Blackout quantifies average power blackout in weeks per hour for a random firm. It is expected
to have a negative coefficient. The explanation for this could be with high power blackout more
and more investors will abstain from investing in manufacturing sector.

One of the constraints facing investors in Ethiopia is access to land. Especially for the
manufacturing it is one of the main factors hampering the growth of the sector. In Mekelle too
there is problem of land administration, ownership and provision. The variable area is used as a
proxy to access to land. It measures the total area of the firm’s production plant or store in meter
squares. The variable area is expected to have positive sign because with more provision of area
the more will be the probability of investment in manufacturing sector.

Initial capital is another variable which will be verified whether it has an effect on decision of
investment in different sectors. It is also the variance function. A variance function specifies the
variables on which the variance is assumed to depend. Initial capital was chosen as a variance
function because the recurrent occurrence of outliers. This means an there is an observation
which is too large or too low in relation to the observation on the sample.

27 | P a g e
Another variable included in this model is relative price of domestic goods over similar imported
goods. In market economy price conveys cost of production and the reward for producers. Many
capitalists base their investment decision on price. It is expected to have positive sign. Because if
there is high reward for internally produced goods investors will be more likely to produce that
good than import it from other country.

The behavioral model included in the model is the discount rate of the investor. Discount rate
measures the rate at which we are willing to trade future benefits to present benefits. The need
for this variable arises because the return for manufacturing sector is most of the time realized at
latter stages. This variable can decide us whether a random investor chooses immediate profits or
higher profits at latter stages. This variable is estimated using the choice-based method. Choice-
based methods generally present participants with a series of binary comparisons and use these to
infer an indifference point, which is then converted into a discount rate. This variable is expected
to have negative coefficient. The reason for this can be investors with high discount rate (present
oriented) are expected to neglect the manufacturing sector which takes longer span for the
returns to be realized.

Table 3.1: Description of variables

Variable Short name Expected sign Description

Invested in Invmf Dependent 1= if in manufacturing
manufacturing variable
0= if in other sectors
Education level of Edu + -
the investor
Loan given by bank Loan + -
in the last 10 years
Average power Blackout - -
blackout in weeks
per hour
Total area of the Area + Measured in m2
firm
Relative prices Rp + Domestic price/ similar imported
good price
Discount rate of the R - -
investor
Initial capital Inicap + -

28 | P a g e
 3.4 ECONOMETRIC FRAMEWORK

The standard probit or the homoskedastic probit model assumes constant variance or
homoskedasticity. But what happens if there is heteroskedasticity? What if σ is not a constant? In
the context of heterogeneous choice, σ is known or expected to vary systematically, such that σ =
σ i with i = 1…N. It should be clear, though, that if the errors are not constant (heteroskedastic),
^β
then ^β= and the parameter estimates will be biased, inconsistent, and inefficient; the standard
σ
errors will also be wrong. As a result this research focuses on the heteroskedasticity problem that
is ignored in most of the probit applications.

The rationale for focusing on heteroskedasiticity problem is related with the data especially with
the variable initial capital. There are many outlier observations in initial capital. There are
observations which deviate greatly in relation to the observations in the sample. In other words
there are observations with too maximum or minimum value. In face of such observation it will
be hard to maintain the assumption of homoskedasticity. So the only solution is to use an
estimation technique which relaxes the classical assumption of homoskedasticity, i.e.
heteroskedastic probit.

One approach would be to treat the heteroskedasticity as a nuisance factor and use robust
standard errors the same way with OLS and heteroskedasticity. However robust standard error
need asymptotically large sample which is not available in this research. An alternative approach
would be to estimate a heteroskedastic probit (or logit). We might want to do this if we are
explicitly interested in knowing how an independent variable affects the variance in the
probability of some choice.

Heteroskedasticity can cause problems such as incorrect standard errors, and biased and
inconsistent parameters. For this reason the determinants of investment in manufacturing sector
is examined using the heteroskedastic probit model. The standard probit model is as well used to
compare with the results of the heteroskedasctic probit model. The following section presents the
setup of the heteroskedastic probit model.

The binary probit model is based on the assumption that a latent variable y* is linearly related to
the observed x’s

¿
y i = xi β +ε i

29 | P a g e
Where xi is the vector of values for the ith observation β is a vector of parameters, and ε i is the
unobserved error. The relation between y* and the observed binary variable y can be expressed
as follows:

y i=
{ 1if y i∗¿ τ
0 if y i∗≤ τ

Whereτ is a threshold or cut point. The errors of yi* are assumed to follow a normal distributed,
the binary probit model expressed as follows:

Pr ( y i=1) = ϕ ( x i β)

Where ϕ is the cumulative normal distribution. In the probit model, the error term is assumed as
a homoscedastic. Probability can be written as follows:

xi β
Pr ( y i=1) = ϕ ( σ
)

When σ is a constant that equals 1, it is removed from equation to estimate β s . If the errors are
heteroskedastic the parameters estimate will be biased in consistent and inefficient. In the context
of heterogeneous choice, σ is known or expected to vary systematically such that σ =σ iwith i=
1…N.

The heteroskedastic probit model as proposed by Alvarez and Brehm (1995) can be expressed as
follows:

xi β
Pr ( y i=1) = ϕ ( σi
)

Where σ i=exp ( z i γ ) and z iis a vector of covariates of the ith observation and γ is a vector of
parameters to be estimated. If the γ ' s is equal to 0, the model is identified as a probit model.
Thus, the probit model is nested in the heteroskedastic probit model. As a result, you can use a
likelihood ratio test to determine whether you need to run the heteroskedastic version or not. The
heteroskedastic probit can be estimated by maximum likelihood like probit model. For
heteroskedastic probit, log likelihood expressed as follows:

log L=∑ ¿¿ ¿ ¿

30 | P a g e
Interpretation of coefficients obtained from probit model is complex. Coefficients are interpreted
with the marginal effects. Marginal effect in a standard probit model with respect to same x k is:

∂ Pr ( y i=1 )
=Φ ( ^
x i β ) β^k
∂ xk

For heteroskedastic probit model marginal effect is calculated as follows:

∂ Pr ( y i=1 ) xi β ^
βk
=Φ ( )( )
∂ xk exp ( z i γ ) exp ( z i γ )

Thus using the above setup for heteroskedastic probit the model specification for this research is
as follows:

invmf = β0 + β 1 edu+ β 2 loan + β 3 blackout + β 4 area+ β 5 r + β 6 inicap+ β7 rp+ ε i

Where ε i N ( 0 , σ ) ∧σ i=e
2 γ inicap
1

CHAPTER FOUR

31 | P a g e
 RESULTS AND DISCUSSION
The main aim of this section is to provide results of the study as well as to discuss each one of
them rigorously. All the variables were tested using descriptive and econometric methods of data
analysis. In addition in line with the objectives of the paper secondary data from NBE and
Mekelle investment agency was used to study the status of investment in manufacturing sector in
Ethiopia and Mekelle city respectively.

This section begins with the discussion of recent trends in Ethiopia’s manufacturing sector and
then proceeds on to the same discussion in Mekelle. It will then advance to the discussion of the
descriptive analysis of the research. Eventually this section will present the final results of the
study supported by econometric analysis.

4.1 PATTERNS OF INVESTMENT IN ETHIOPIA’S

MANUFACTURING SECTOR

This section discusses the recent trends in Ethiopia’s manufacturing sector starting from 2002/03
up to 2011/12. But let’s first start by the classification of the secondary sector, i.e. industry
sector, according to the national income account of Ethiopia.

The national income account of Ethiopia classifies the industrial sector into these sub sectors:

 manufacturing,
 mining, quarrying,
 construction,
 Water and energy supply

Mining and quarrying, construction and water and energy supply are generally termed as non-
manufacturing industry. The researcher will also use this term in the proceeding sections.

Since its foundation as a new entity the Ethiopian investment agency (EIA) has given numerous
investment grants to investors from abroad and inside the country. It has also given grants to
investors from different sectors. The sectoral breakdown of investment grants is presented in the
following table.

Table 4.1 Investment projects distribution by sector from 2002/03-2011/12

year Agriculture Service Manufacturing Non- Total Investment
industry manufacturing

32 | P a g e
 industry
No.
No. No. No.
Investme Investme of Investme No. of
of of of Investmen Investmen
nt nt proj nt projec
proje proje proje t capital* t capital*
capital* capital* ect capital* ts
ct cts cts
s
2002/
96 1401.9 618 6572.61 361 3525.96 142 1937.42 1217 13437.89
03
2003/
254 4,142.39 995 6237.98 822 8,507.00 154 2332.34 2225 21219.71
04
2004/
463 7,657.15 1329 15533.56 919 10,520.01 161 2755.93 2872 36466.65
05
2005/
731 7153.52 3441 17321.81 1331 32881.51 356 22679.44 5859 80036.28
06
2006/
1108 14479.9 3804 39621.58 1260 37382.24 300 2095.2 6472 93578.92
07
2007/
2135 36984.56 4204 62744.9 1998 64473.36 624 6176.39 8961 170379.21
08
2008/
2462 47253.95 3678 44614.06 1804 58351.11 863 89305.67 8807 239524.79
09
2009/
1350 21644 2758 28873 1433 35,583 955 10315 6496 96415
10
2010/
908 82777 3149 26221 1294 43,530 971 85803 6322 238331
11
2011/
437 23300 3243 40305 1211 45,482 758 37082 5649 146169
12
*in million Birr
Source: Ethiopian Investment Agency and own computation

Table 4.1 depicts sectoral breakdown of investment by sector in the last decade. The total
investment grant by the agency increases apart from the last year. The last year in the data
(2011/12) is shown to have lower number of investment grant which is measured by number of
projects and amount of registered capital than previous years. It is even the agency’s lowest
performance since 2005/06. This is ironically amazing since 2005/06 after a turbulent general
election was able to register better number of projects.

The main focus of this section is to examine the level of investment in manufacturing sector.
Looking at the above data investment in manufacturing sector is decreasing. The year 2007/08 is
a turning year for manufacturing sector investment. This is so because before that year
investment in manufacturing sector registered positive growth. But things changed after that and
they seem once and for all unless an effort is made by government and other agencies. After that
year it has been registering negative growth. This conflicts with the goal of the government and
the grand transformation plan (GTP). It is the government’s plan to make the industry especially
the manufacturing sector the leading sector. But this would not be successful unless there is
investment in manufacturing sector.

33 | P a g e
 70.00%

60.00%
Percentage share in total
investment by no. of

50.00%
Agriculture
40.00%
projects

Service
30.00%
Manufacturing
industry
20.00%

10.00% Non-manufactur-
ing industry
0.00%
2002/ 2003/ 2004/ 2005/ 2006/ 2007/ 2008/ 2009/ 2010/ 2011/
03 04 05 06 07 08 09 10 11 12

Year

Source: NBE Annual Report and own computation, 2012

Fig 4.1: Percentage share of sectors in total investment measured by number of projects
70.00%
vetment by investment capital
Percentage share in total in-

60.00%

50.00% Agriculture

40.00% Service

30.00% Manufactur-
ing industry
20.00% Non-manu-
facturing
10.00% industry

0.00%
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
/03 /04 /05 /06 /07 /08 /09 /10 /11 /12

Year

Source: NBE Annual Report and own computation, 2012

Fig 4.2: Percentage share of sectors in total investment measured by investment capital

In table 4.1 the pattern of investment in manufacturing sector in absolute term was discussed.
But in figure 4.1 and 4.2 the relative level of investment in manufacturing sector is elaborated in

34 | P a g e
relation to other sectors. In the first graph the percentage share of each sector’s investment is put
in terms of number of projects. The percentage share of manufacturing industry has somewhat a
downward trend. It has a peak in 2003/04 and after that it is decreasing. It has been hitting
around 20% for the last 2 years. Among the sectors the service sector has been taking the lion’s
share in investment in terms of number of projects. However agriculture sector’s share has been
declining indefinitely.

Figure 4.2 portrays the share of each sector’s investment in terms of investment capital. Contrary
to figure 4.1 where the sectors have a definite trend, in figure 4.2 it is somewhat unstable. I.e.
their path is indefinite. For instance if we observe the path of manufacturing industry investment
it interchanges between increasing and decreasing at different intervals. It reaches peak in the
year 2005/06. But after that year it decreases again until it reaches the maximum of all sectors in
the year 2011/12. The most intriguing part of figure 4.2 is the low share of service sector
investment in terms of capital. This is despite having the highest share in terms number of
projects. This in part reinforces the claim by the government that Ethiopian investors are
investing in sectors where there is need for small capital and which in turn create low
employment opportunity.

The following figure, i.e. fig 4.3, depicts the share of manufacturing and non-manufacturing in
total industry investment in terms of number of projects and investment capital. As it has been
discussed earlier the industry sector is divided into manufacturing and non-manufacturing. The
total investment in the industry sector is measured by adding the investment in non-
manufacturing and manufacturing. The dashed lines indicate investment measured by investment
capital while the continuous lines are for number of projects. As can be witnessed from the graph
the share of manufacturing investment in total industry sector investment in terms of number of
registered projects is higher than non-manufacturing. But what should be noted here is its share
is fast declining. If this trend continues the non-manufacturing sector can the dominant industry
sector in terms of investment in some years time. When we look at the investment capital section
the manufacturing sector is the larger of the two by the capital it registers. However their gap is
decreasing. Even at some years the non-manufacturing sector was able to overtake the
manufacturing sector. But both sectors have registered unstable investment capital. There is
much fluctuation on the capital both sectors register on year to year basis.

What we understand from the above analysis is there is a declining trend in manufacturing sector
investment both in absolute and relative terms. Though it is early to judge whether this is

35 | P a g e
 Percentage share of manufacturing 100.00%
and non-manufacturing in total in- 90.00%
Manufacturing
industry No. of
80.00% projects
70.00%
dustry investment

Manufacturing
60.00% industry In-
vestment capi-
50.00% tal

40.00% Non-manufac-
turing industry
30.00% No. of projects

20.00% Non-manufac-
turing industry
10.00% Investment cap-
ital
0.00%
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
/03 /04 /05 /06 /07 /08 /09 /10 /11 /12

Year

Source: NBE Annual Report and own computation, 2012

Fig 4.3: Percentage share of manufacturing and non-manufacturing in total industry
investment in terms of number of projects and investment capital

something to worry very much, it will surely have a negative implications for the long term
prospects of Ethiopian economy. At least it diverges from the plans of the government in making
the manufacturing sector the dominant sector in the economy.

4.2 PATTERNS OF INVESTMENT IN MANUFACTURING

SECTOR OF MEKELLE CITY

In earlier sections the pattern of investment by sectors was discussed at the national level.
Deductive approach is used in this paper to investigate the status of investment in manufacturing
sector. This means this section will focus on the intensity of investment in manufacturing sector
in Mekelle city. With that it will be also possible to elaborate the status of investment in other
sectors too.

Analogous to fig 4.1 and 4.2 the following two graphs depict the level of investment of each
sector in percentage measured by number of registered projects and registered initial capital.

36 | P a g e
 90.00%
80.00%
Percentage share in total

Agriculture
70.00%
investment by no. of

60.00%
Service
50.00%
projects

40.00%
Manufacturing
30.00% industry

20.00%
Non-manufactur-
10.00% ing industry
0.00%
2003/ 2004/ 2005/ 2006/ 2007/ 2008/ 2009/ 2010/ 2011/ 2012/
04 05 06 07 08 09 10 11 12 13

Year

Source: Mekelle Investment Agency and own computation, 2013

Fig 4.4: Percentage share of each sectors’ investment in total investment registered in
Mekelle city measured by number of projects
90.00%
Percentage share in total in-
vetment by investment cap-

80.00%
70.00% Agriculture
60.00%
Service
50.00%
Manufactur-
ital

40.00%
ing industry
30.00%
20.00% Non-manu-
facturing
10.00% industry
0.00%
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
/04 /05 /06 /07 /08 /09 /10 /11 /12 /13

Year

Source: Mekelle Investment Agency and own computation, 2013

Fig 4.5: Percentage share of each sectors’ investment in total investment registered in
Mekelle city measured by investment capital

37 | P a g e
When we look at both figures they seem similar to fig 4.1 and 4.2 which depict the level of
investment in each sector albeit at national level. What makes them similar is the downward
trend of investment in manufacturing sector from previous highest peaks. However the
difference is in the national case it is the service sector which is the most attractive sector
whereas in Mekelle city it is the non-manufacturing industry like construction which is most
attractive to investors.

In fig 4.4 it is obvious that the level of investment in manufacturing sector is decreasing. It
reaches its minimum in the year 2012/2013. This means the level of investment in manufacturing
sector measured by number of projects has been decreasing relative to other sectors. This has
been despite the effort of the city council in setting up industry zones and other special
incentives.

The situation is same when viewed from fig 4.5. Manufacturing sector investors have been
registering lower capital when compared from previous years. Low initial capital means few jobs
are created for the massive youth section of the city which accounts for 60% of the city’s
population. Leaving this portion of the population from the development process is certain to
have huge negative implication on the city’s development. But saying this I don’t mean that the
manufacturing industry sector is the only sector that can involve the youth. Yet it is
manufacturing sector that can absorb the youth in mass.

Table 4.2: Investment projects distribution by sectors 1996-2005 in Mekelle city

Year Sector No. of projects Investment capital* Job opportunities’

created
2003/04 Agriculture 70 301.00 10141
- Service 218 3,204.00 17531
2012/13
Manufacturing industry 494 7,095.50 37713

Non-manufacturing 1,007 4,259.30 51503

industry
Total 1,789 14,859.80 116888
*in million birr
Source: Mekelle Investment Agency, 2013

The above table presents the intensity of investment in each sector from 2003/04-2012/13. As
can be seen from the above table the non-manufacturing industry especially the construction
sector has been the main employer. But in terms of registered investment capital manufacturing
industry leads the way. In the registered projects meanwhile it is headed by the non-

38 | P a g e
manufacturing industry. In conclusion while relative to other sectors the investment in
manufacturing sector is decreasing in absolute terms it does not have a definite trend.

The analysis both in the city and national level reveal one thing. The manufacturing sector is
losing its steam at least in terms of investment. The manufacturing sector is trailing behind other
sectors in term of attractiveness. Even worse it is losing its attractiveness to new investors.
However this reality contradicts to the plan of the government who is determined to make the
manufacturing sector the main sector. If there is no investment in manufacturing sector it will be
like a fantasy to make the manufacturing sector the dominant sector.

4.3 DESCRIPTIVE ANALYSIS OF THE SURVEY

This section of the research paper consists of the descriptive analysis of the survey done by the
researcher. This section is reliant on the questionnaire distributed to different of investors. As
mentioned on chapter 1 of this paper 94 samples are used to assess the determinants of
investment in manufacturing sector in Mekelle city. Respondents were asked a different set of
questions. The questions gravitate around which sector the investors put their money in. They are
classified into two, i.e. investors in manufacturing sector and non-manufacturing sector. The
non-manufacturing sector consist economic activities which are not included under
manufacturing sector. However this should not be confusing with non-manufacturing industry.
The non-manufacturing industry is included under non-manufacturing sector in this paper.

In this section we will first start our analysis by discussion of personal information of
respondents and then proceed into analysis of non-personal questions asked in the questionnaire.

Respondents were asked 4 questions about their personal information. These personal
information questions are valuable in studying the personal behavior of economic agents. In
accordance with the above statement the respondents were asked about their gender, marital
status, age and education level.

As can be witnessed in the following table most of the respondents are married male who are
mostly in their late 40s. Though males dominate investment in Mekelle city females also have
fundamental role in increasing investment in the city. It is expected also that their role in
investment is going to increase which will the city and the country to more egalitarian society.
Another aspect of investment, age, has fair distribution compared with other personal variables
in the questionnaire. The age group above 45 is the dominant section in this category comprising
about 52% of the total sample. A summary of the results is presented in the following table.

39 | P a g e
Table 4.3: Personal information of respondents
Variables Frequency Percentage (%)
Sex Male 71 75.5
Female 23 24.5
Married 53 56.4
Single 35 37.2
Marital Status
Divorced 5 5.3
Separated 1 1.1
18-30 11 11.71
Age 30-45 34 36.17
>45 49 52.12
Source: Own survey in Mekelle, 2014

Another personal factor which is of most important to this paper is education level of
respondents. Education level is measured by school year the respondent has taken so far. The
following table depicts the educational attainment of respondents.

Table 4.4: Educational level of respondents

Education level of the Percentage

Frequency Percentile
investor (%)

0-5 21 22.35 25% 6

6-10 50 53.18 50% 8.5
11-15 23 24.47 75% 10
Source: Own survey in Mekelle, 2014

The table 4.4 shows that most of the respondent investors have barely finished their high school.
Some have attained degree but there was no one who was above the degree threshold. Out of the
total sample only 2 have degree. In addition the percentile section shows that a quarter of the
samples have below six years of educational attainment. While half of the population even didn’t
enroll into high school. Furthermore a quarter of the population was able to reach 10 years of
school. Overall it can be said that most of the investors have low educational attainment.

Until know we have analyzed the personal questions. This paper will next proceed into the non-
personal questions and will try to examine the responses by the sample investors.

40 | P a g e
A different set of questions were asked to investors about the general status of their business and
how they fared in the market. One of the questions was initial capital and its source. The
following table illustrates the sample investor’s initial capital.

Table 4.5: Initial capital of investors

Initial capital in
Frequency Percentage (%) Percentile
thousands
100-2500 52 55.32 10% 600
2500-11000 23 24.37 25% 1150
11000-75000 11 11.7 50% 2400
>75000 8 8.51 75% 9500
Source: Own survey in Mekelle, 2014

In table 4.5 sample investors were classified under 4 groups by the amount of initial capital. It is
observable from the table that investors with smallest initial capital have the lion’s share among
the sample investors. Analogously investors with huge capital are few comprising only 8.5% of
the sample investors. As a result we can conclude that as the amount of initial capital increases
the number of investors in that initial capital group decreases. In the table it is revealed that only
a quarter of the population claim more than 9.5 million birr initial capital.

The sample investors were also asked about their perception on interest on the prevailing interest
rate. Their response is presented in the following table with the sector they invested.

Table 4.6: Perception on the prevailing interest rate

Perception on interest rate
Invested in manufacturing Total
High Modest Low
0 14 34 19 67
1 9 10 8 27
Total 23 44 27 94
Source: Own survey in Mekelle, 2014

In the questionnaire the respondents were asked if the prevailing interest rate was high, modest
or low. The numbers 0 and 1 indicate investment in non-manufacturing sector and manufacturing
sector respectively. Most of the sample respondents believe that the prevailing interest rate is
modest with the rest fairly distributed among high and low. But the ratio of manufacturing
investors who believe that the prevailing interest rate is “high” is greater than the non-

41 | P a g e
manufacturing sector investors. This leads us to the conclusion that the manufacturing sector
investors more of believe that the prevailing manufacturing sector is higher.

To measure access to finance of investors in Mekelle collateral requirement as percent of

principal of last loan was used. This variable is important in measuring access to finance. High
ratio will lead to investors refraining from credit market.

Table 4.7: Collateral requirement as percent of principal of last loan

Collateral requirement as percent of principal of
Invested in
last loan Total
manufacturing
35-78 80-100 101-135 140-175 200-300
0 30 19 15 2 1 67
1 2 4 6 10 5 27
Total 32 23 21 12 6 94
Source: Own survey in Mekelle, 2014

Table 4.7 reveals one striking point. That is compared to other sectors the manufacturing sector
is asked higher collateral requirement as percent of principal loan. This may be one of the causes
that are driving investors away from the manufacturing sector. Most of the manufacturing sector
investors are asked about 140 to 175 percent of principal. This is higher than the loan given to
the investors. Relatively participants in other sectors are asked lower collateral requirements.
Even if the manufacturer wants to expand his business he/she is expected to be short of finances.
Consequently we can conclude that the manufacturing sector is the sector in which finances are
hard to get by.

The other variable which was asked by the researcher to sample investors’ was average power
blackout in weeks per hour. This variable as mentioned in previous chapters is used as a negative
proxy to measure access to electricity of sample investors. It measures how many times the
sample investors face power outage. The following table unveils the results of the study.

Table 4.8: Average power blackout in weeks per hour of sample investors
Average power
blackout in weeks Frequency Percentage (%) Percentile
per hour
0-3 29 30.85 25% 3
3-6 52 55.32 50% 4
6-9 13 13.83 75% 5
Source: Own survey in Mekelle, 2014

42 | P a g e
In table 4.8 we can notice that there is imminent power outage in Mekelle city. Most of the
investors don’t get 3 to 6 hours of power per week. This is significant in terms of hampering
their capacity efficiently. It is one of the reasons for manufacturing investors forcing them not to
use their full capacity. The city government as a result should not solely focus on the provision
of power but on the reliability of the provision too.

To measure the personal preference of investors’ discount rate was employed as a variable.
Discount rate is the rate which relates the present and future benefits. Someone with high
discount rate attaches small value to the future benefit and is called “present-oriented”. If
someone values future income more than present income he/she has low discount rate and is
called “future-oriented”. A set of questions rooted on the choice-based method were asked to the
investors to determine their discount rate. The results are presented in the following table.

Table 4.9: Discount rate of individual investors in Mekelle

Invested in Discount rate of individual investors
manufacturin Total
7 8 9 11
g
0 9 20 25 13 67
1 11 10 4 2 27
Total 20 30 29 15 94
Source: Own survey in Mekelle, 2014

The sample investors were given four choices of discount rate to select from. From the table it
can be observed that manufacturing sector investor have in average lower discount rate. This
means they are more future-oriented than other sector investors. Most of them are in the lowest
discount rate level of this research which is 7 and 8. Though the profit margins in manufacturing
sector are in average higher than other sector the returns take longer time to come. This means
only individuals with lower discount rate could have the patience to wait for the return and most
of them are found in the manufacturing sector. On the contrary other sector investors have higher
discount rate in average.

A special set of questions were asked to manufacturing sector investors to identify the challenges
they face while participating in this sector. They were provided with close-ended questions
related with the constraints they encounter. Identifying the challenges is half the way of solving
the problem, so the challenges should be clearly looked at to formulate the solution.

43 | P a g e
 20
18
16
14
12
10
8
6
4
2
0
ts re d ts es n n ity ce n m
r ain ctu lan p or r at atio ptio ain an atio ste
o r t n l y
st tru ss
t im Ta
x ist or
ru ce
r
ve
r u ls
con r as ce ap in C n go r eg ega
it nf c e dm u s L
ed fi fa Ch xa lic
y
Ba
d
in
es
Cr ck
o
ck
o
Ta P o s
La La Bu

Source: Own survey in Mekelle, 2014

Fig 4.6: Top business constraints faced by sample investors in Mekelle city

Credit constraint is the main challenge of investors as witnessed in the above graph followed by
lack of infrastructure. Lack of quality and reliable infrastructure was identified as the second
most constraint faced by investors in Mekelle city. Tax rates were also recognized as major
constraint encountered by many investors. Policy uncertainty was identified as the least affecting
factor in doing business. The reason for this is that there is no major change in the policy of the
government in the last decade. Governance related constraint like bad governance; corruption
and business regulation also have been identified as a major impediment to the business
environment in Mekelle city. Cheap imports especially from china also have been identified as a
significant obstruction to manufacturing sector investors in Mekelle city. In conclusion all the
major constraints identified by the investors require a quick solution from the concerned bodies
so that the business environment could improve. An improved business environment will attract
more investors and with it more growth and prosperity to this city.

44 | P a g e
 4.4 ECONOMETRIC ANALYSIS

In this study an attempt was made to determine the factors influencing the probability of
investment in manufacturing sector in Mekelle city. For this purpose both the heteroskedastic
probit and probit model were evaluated.

Following the theoretical foundations laid by Jorgenson’s model of optimal capital accumulation
and other variables added, which are deemed relevant for this particular research, the probability
of investment in manufacturing sector was estimated. In the following sections the results and
post estimation of the model is presented.

4.4.1 THE HETEROSKEDASTIC PROBIT MODEL ESTIMATION

RESULT

The binary dependent variable invested in manufacturing sector was regressed with seven
independent variables and with one variance function in the heteroskedastic probit model. The
dependent variable was also regressed using the standard probit model to compare the results
with the heteroskedastic probit model.

According to table 4.10 education, relative prices, blackout and initial capital are statistically
significant at 5 percent level of significance. The variance function initial capital is also
significant. When we look at the probit section only education, relative prices and blackout are
significant. However, the maximum likelihood estimators of the probit model will be biased and
inconsistent if the disturbances are either non-normal or heteroskedastic. This means the probit
model will be only used for comparison basis in this paper.

The probit model reports that education, relative prices and blackout are significant at 1 percent
level of significance while the rest are insignificant. The model is overall significant having log
likelihood of -33.47. The Hosmer-Lemeshow chi2 for goodness of fit reveals that the probit
model fits reasonably well. This test is a test of the observed against expected number of
responses using cells defined by the covariate patterns. To do the test we regrouped the data into
a group of 8 based on the predicted probabilities. But as it has been discussed earlier we cannot
use the probit model based on the goodness of fit since the coefficients will be biased if we use
the probit model. In addition the diagnostic tests of the heteroskedastic probit reject the use of
standard probit model.

45 | P a g e
Table 4.10: Results of Probit and Heteroscedastic Probit Model

Dependent Variable: INVMF

Heteroskedastic probit model
Variance Probit model
Mean function Variance function
0.1899*** 0.3382***
EDU
(0.0659) (0.1222)
-0.0427 -0.0089
LOAN
(0.0487) (0.0592)
-0.2903*** -0.3727**
BLACKOUT
(0.1047) (0.1535)
0.0009 0.0003
AREA
(0.0007) (0.0011)
-0.0360 -0.0058
R
(0.0393) (0.0488)
2.1575*** 2.8493**
RP
(0.8144) (1.1330)
2.46e-06 0.0002** 0.0000325***
INICAP
(4.04e-06) (0.00006) (9.82e-06)
-3.0859** -5.9283***
Constant
(1.2080) (2.067)
Observations 94 94
Pseudo R2 0.4063
Log-likelihood -33.4652*** -27.15997
Prob>chi2 0.0000
Hosmer-Lemeshow
chi2 for goodness of 6.95
fit
Prob>chi2 0.3253
Wald 19.27***
Prob > chi2 0.0074
Lnsigma2 12.61***
Prob > chi2 0.0004
(i) *, **, *** indicate significance at the level 10%, 5% and 1%, respectively.
(ii) Numbers in parentheses are standard errors.

As it has been discussed earlier the coefficients of the standard probit model are biased. To
resolve this problem the heteroskedastic probit model is applied. It is clear that some estimated

46 | P a g e
coefficients of the heteroscedastic probit model are different from the probit model. All the
variables are considered in the mean equation, and initial capital was estimated in the variance
equation. As well acknowledged, the coefficients cannot be interpreted directly in the
heteroscedastic probit models. In such a case, the marginal effects can be computed as a
nonlinear combination of the regression coefficient.

According to heteroskedastic probit’s output the Wald test of the full model versus the constant
2
only model is significant with χ ( 1 )=19.27 . Likewise, the likelihood-ratio test of
heteroskedasticity, which tests the full model with heteroskedasticity against the full model
without, is significant with χ 2 ( 1 )=12.61 . The first thing to notice is that if all the elements of
γ=0 , then exp (0) =1 and the model is just our standard probit model. Thus, the probit model is
nested in the heteroskedastic probit model. On the contrary the above model reveals that the
variance function is significantly different from zero. i.e. lnsigma2 is significant at 1 percent. We
can also use the likelihood ratio test to determine whether you need to run the heteroskedastic
version or not. The likelihood ratio test reported rejects a model without heteroscedasticity. This
means we completely reject the standard probit model.

The heteroskedastic probit estimation result reports that while education is significant at 1
percent; blackout, relative prices and initial capital are significant at 5 percent. The variance
function (Lnsigma2) is also significant at 1 percent and having the same sign with the mean
function initial capital. All the variables have the expected sign except for loan. Tough loan was
expected to have positive effect on the probability of investment in manufacturing but according
to the heteroskedastic model it has a negative effect. The explanation for this can be the
prevailing finance structure in Mekelle in particular and Ethiopia as a whole. Manufacturing
sector has much less access to finance due bank policy. The financial sector in Ethiopia at the
moment embraces only short-term financing. The manufacturing sector, which requires relatively
long-term financing, suffers from lack of such finance scheme. This scheme applies to Mekelle
city investors especially the ones who are participating in the manufacturing sector. They have
been taking fewer loans from bank compared with other sector investors.

Another variable which is insignificant is discount rate of individuals. The result shows that with
increasing discount rate, i.e. moving towards present oriented persons, the probability of
investment in manufacturing sector decreases. However I am short of deciding it has an
insignificant effect on probability of investment in manufacturing sector. The reason for the
insignificance could be the method (choice based method) I used. I asked only four discount
rates for which the investor to choose from. Another advanced discount rate determining
methods shall be used to assess its effect on probability of investment in manufacturing sector. In

47 | P a g e
conclusion the researcher recommends for further study on the effect of discount rate since it
greatly affects decisions which involve longer periods.
1 .5
1
.5
0

0 1000000 2000000 3000000 4000000

initial capital in thousands

Fitted values predicted invmf

Fig 4.7: The relationship between initial capital and investment in manufacturing sector

The above graph demonstrates the quadratic relationship between investment in manufacturing
sector and initial capital. As we can observe from the plot investment in manufacturing sector
increases at increasing rate at lower levels of initial capital while it increases at decreasing at
higher levels of initial capital.

4.4.2 MARGINAL EFFECTS OF INDEPENDENT VARIABLE

If we are interested in understanding how the independent variables affect the unobservable
latent variable y∗, then the probit and logit coefficients can be interpreted in exactly the same
way as OLS coefficients i.e. the coefficient tell you how much y∗¿ changes with a one unit
increase in the independent variables. Of course, it is almost never the case that you will be
interested in y∗. Instead, you want to know the effect of your independent variables on y i.e.
the probability of getting a 1 or 0.

48 | P a g e
Table 4.11: Marginal effects of Heteroshedastic Probit Model

Variables dy/dx
0.0083*
EDU
( 0.0045)
-0.00022
LOAN
( 0.0015)
-0.0092
BLACKOUT
(0.0056)
8.38e-06
AREA
(0.00003)
-0.000143
R
(0.0012)
0.0701*
RP
(0.04123)
-3.89e-06*
INICAP
(0.0000)
(i) *, **, *** indicate significance at the level 10%, 5% and 1%, respectively.
(ii) Numbers in parentheses are standard errors.

The marginal effects result reports that education, relative prices and initial capital have
significant marginal effect on investment initial capital. The coefficients in table 4.10 reveal the
sign of the variables. But they are not helpful in estimating the exact effect of the independent
variable on the dependent variable. We mostly use marginal effects to interpret the coefficients
of the each independent variable. These marginal effect estimates are simply the estimates of the
change in probability of choice which we expect conditioned on a change in the value of the
particular independent variable.

According to the above table a 1 percent increase in relative prices is transferred as 0.07 percent
increase in probability of investment in manufacturing sector. The likelihood of investing in
manufacturing sector also increases by 0.009 percent if there is an increase in education by one
school year. Other variables marginal effects are interpreted in the same way. The most
captivating part of this result is that there are differing results in the marginal effects and
heteroskedastic probit model results. The heteroskedastic probit model reports positive
coefficient while the marginal effects reveals that there is negative relationship between the
dependent variable and initial capital. This is due to marginal effects being reported at the mean
values of the independent variable. At its mean value it may have negative effect on the

49 | P a g e
likelihood of investment in manufacturing sector. But to know the effect of initial capital we take
all the values so we conclude that it has positive effect on investment in manufacturing sector.

4.4.3 A MEASSURE OF ELASTICITIES OF INDEPENDENT

VARIABLE

In the above section we have computed the marginal effects of each variable with respect to the
predicted value of the dependent variable. This section will discuss other forms of marginal
effects albeit with different interpretation. In other words we will obtain margins of derivatives
of responses in different forms. Derivatives are of interest because they are an informative way
of summarizing fitted results. The change in a response for a change in the covariate is easy to
understand and to explain.

The derivatives can be expressed as elasticity of the independent variables. By doing this we get
three types of elasticity; namely eyex, eydx and dyex. These elasiticities are computed at
observational level. The formulas are as follows:

ey /ex =dy /dx ×( x / y )

ey /dx =dy /dx ×(1/ y )

dy /ex =dy /dx ×(x )

In the following sections the results of each independent variable will be presented with respect
to the probability of investment in manufacturing sector.

Table 4.12: Average marginal effects (ey/ex) of independent variables

Variables ey/ex
EDU 3.5759**
LOAN -0.0564
BLACKOUT -3.0172**
AREA 0.0742
R -0.1020
RP 4.3837**
INICAP 0.7517**
(i) *, **, *** indicate significance at the level 10%, 5% and 1%, respectively.

50 | P a g e
The coefficients are interpreted proportionally. This means for a proportional change in the
independent variable there is a proportional change in the dependent variable. Probability of
investment in manufacturing increases with initial capital at a rate such that, if the rate were
constant, probability of investment in manufacturing would increase by 0.7517 percent if initial
capital had increased by 1000.

In the same way for education probability of investment in manufacturing sector would increase
by 3.6 percent had there been a one percent increase in school year of an investor. A unit of
school year can be substituted with less than one year of entrepreneurship skill training. This
means there is a space for the concerned government bodies to maneuver entrepreneurship skills
to boost investment in manufacturing sector. From table 4.12 one can deduct that education the
most elastic of all. This means for one percent increase in supply of education investors’
responsiveness will more than one percent.

Table 4.13: Average marginal effects (ey/dx) of independent variables

Variables ey/dx
EDU 0.5392**
LOAN -0.0141
BLACKOUT -0.5942**
AREA 0.0005
R -0.0092
RP 4.5426**
INICAP 0.0004**
(i) *, **, *** indicate significance at the level 10%, 5% and 1%, respectively.

The above type of elasticity is interpreted as for a change in the independent variable there is a
proportional change in the dependent variable. If the rate was constant an increase in one unit of
relative prices there will be a proportional 4.54 percent increase in the likelihood of investment
in manufacturing sector. The coefficient in the education also indicates a one year increase in
school year increases the likelihood of investment in manufacturing by 0.54 percent
proportionally.

This elasticity is similar with the prior one. According to table 4.13 the most significant variable
is blackout. This means a one hour decrease in power outage per week will have more than
proportional positive effect on investment in manufacturing sector. But we can’t say that
blackout is the most elastic because the interpretation is not done on the same unit.

51 | P a g e
Table 4.14: Average marginal effects (dy/ex) of independent variables

Variables dy/ex
EDU 0.3387***
LOAN -0.0045
BLACKOUT -0.1613***
AREA 0.0079
R -0.0055
RP 0.3342***
INICAP 0.0525***
(i) *, **, *** indicate significance at the level 10%, 5% and 1%, respectively.

Analogous to the above elasticities dy/ex is interpreted as for a proportional change in the
independent variable there will be a change in the dependent variable. For instance if we take
initial capital, if it had doubled the likelihood of investment in manufacturing sector would have
increased by 0.05 percent. Correspondingly the probability of investment in manufacturing
sector decreases by 0.16 percent if the power blackout doubles. These variables are interpreted
holding the assumption that the rates are constant at every observation.

Table 4.14 unveils that with the same effort on all significant variables, education has more
productive output. It is the most powerful tool of increasing investment in manufacturing sector
at micro level. Apart from education initial capital and a decrease in power blackout collectively
play a strong role in developing manufacturing.

4.4.4 ASSESSING HETEROSKEDASTIC PROBIT MODEL FIT

Analysts often look for a one-number summary of model fit. In linear regression, we have the R 2.
In binary response models, we don’t even have an R 2. STATA will report a ‘pseudo- R2’ but as
its name indicates it is even worse than an R 2. There is no distribution and hence no way of
knowing whether one pseudo- R2 is significantly different from another pseudo- R2. There are
number of different way of evaluating model fit in binary response model. PCP (percent
correctly predicted), PRE (percentage reduction in error) and ePCP (expected percent correctly
predicted) are the most used methods of calculating model fit. All of them will be discussed in
the succeeding sections.

PCP (percent correctly predicted) is a technique of computing model fit where we answer the
question of how many percent of the observation is correctly predicted. We usually use the
threshold level of 0.5 to determine the predicted probability. A predicted probability greater than

52 | P a g e
or equal to 0.5 should be classified as a 1 and any observation with a predicted probability less
than 0.5 should be classified as a 0. The formula is as follows:

100 ×( Number of correct predictions)

PCP=
N

One alternative to PCP is known as the percentage reduction in error (PRE). PRE is based on a
comparison of PCP and PMC, where PMC is the percentage of observations in the modal
category of the observed data. In this paper it would be invested in non-manufacturing sector (0).

PCP−PMC
PRE=
1−PMC

PRE seeks to compare the information provided by probit fitted categories with the classification
errors a researcher would make if he/she naively assigned all fitted categories to the modal
category.

Another model fit testing statistical tool is expected percent correctly predicted (ePCP) proposed
by Herron (1999). This statistic essentially provides the expected percentage of correct
predictions and helps avoid the problem of treating an observation with ^p=0.51 the same as an
observation with ^p=0.99. ePCP is calculated as:

1
ePCP=
N ( ∑ ^p+ ∑ (1− ^p))
y i =1 yi =0

According to the above definitions of model fit the goodness of the heteroskedastic probit model
was estimated using the above three methods. The following table presents the results:

Table 4.15: Goodness of fit of the heteroskedastic probit model

Model fit technique Value

PCP 90.43%
PRE 66.67%
ePCP 82.24%

The above table unveils that the heteroskedastic probit model has a good explaining power.
Using the available techniques of testing model fit the model has more than average explaining

53 | P a g e
power. Especially for the PCP it can explain 90.43 percent of the change in the dependent
variable. For ePCP, considered the most rigorous of all, it has an explaining power of 82.24
percent. In other words looking at PCP only 9.57 percent of the change in the likelihood of
investment in manufacturing sector is explained by variables ignored by this research paper. All
in all looking at the above results it can be concluded that the model has very good explaining
power.

The result in table 4.15 also shows how robust the model specification is. This means most of the
variables which affect investment in manufacturing sector in Mekelle city at micro level are
included in this paper’s model specification. This in turn to some extent solidifies the theoretical
robustness of neoclassical model of optimal capital accumulation (Jorgenson) which this paper
applies. With good model specification and a promising goodness of fit we can infer from the
samples to generalize it into the whole population.

CHAPTER FIVE

54 | P a g e
 CONCLUSION AND RECOMMENDATION

5.1 CONCLUSION

This study analyzed the possible factors that determine investment in manufacturing sector in
Mekelle city. The study used a total of 94 respondents to assess determining factors of
investment in manufacturing sector. To assess the determinants the researcher a used a
heteroskedastic probit model which was deemed appropriate because if standard probit model
has been used the coefficients will be inconsistent and biased. Furthermore this study was able to
assess the constraints the manufacturing sector investors face. It also used a nationwide data to
investigate the level of investment in manufacturing sector in terms of number of projects and
investment capital.

The main driving force in doing this research had been the recurrent claim by the government
that there is no ample investment in manufacturing sector to structurally transform the country
from agricultural development led industrialization (ADLI) into an industrialization led
development. In other words it means there is no sufficient investment in the industrial sector for
it to be a leading sector of the economy. To verify this claim by the government the researcher
evaluated recent investment in manufacturing sector in Ethiopia and Mekelle city. The finding
goes in line with government’s claim. The relative share of manufacturing sector is in declining
trend. More and more investors are shifting their funds to the burgeoning service sector leaving
few funds available for the industrial sector.

The main focus of the research was to determine the probability of investment in manufacturing
sector. In accordance with the firm level neoclassical model of optimal capital accumulation
(Jorgenson) and adding some variables which are relevant to Mekelle the researcher was able to
estimate the likelihood of investment in manufacturing sector. The researcher used a
combination of economical and behavioral variables to decide the likelihood of investment in
manufacturing sector, viz. education, loan, area, blackout, discount rate, relative price and initial
capital. Initial capital was used as a variance function due to the existence of outliers.

The heteroskedastic probit model result indicate that while education, relative prices and initial
capital have a positive significant effect on likelihood of investment in manufacturing sector
blackout has a negative significant effect. This means with an increase in education level, higher
prices for domestic goods and with more initial capital an investor is more likely to participate in
manufacturing sector. However if the supply of electricity is unreliable it will offset the investors
willingness to participate in manufacturing sector. The most intriguing part of the result was the

55 | P a g e
negative coefficient in the variable loan. Though insignificant, loan had a negative effect on the
willingness of investors to participate in manufacturing sector. The reason for this could be
manufacturers are less likely to take loans from commercial banks due to bank policy that
prevails in Ethiopia in general and Mekelle in particular which prioritizes short term financing.
Another variable, discount rate, has positive insignificant effect on the dependent variable.
However since discount rate is a hidden force which is behind every decision of any investor its
effect on investment in manufacturing need to be properly assessed.

Apart from determining the likelihood in the sector the study examined the constraints
manufacturers in the city face. They outline economic, bureaucratic, legal and infrastructural
hindrances which limit the firm’s growth. Credit constraint was identified as the main constraint.
This goes in line with the coefficient in the variable loan. Lack of reliable infrastructure and lack
of access to land were also recognized as a major impediment to the growth of manufacturing
sector in Mekelle city. Making matters worse bureaucratic red tape and bad governance from the
government side is annihilating free enterprise.

5.2 RECOMMENDATION

The main aim of this study is looking for ways on how to increase investment in manufacturing
sector. This paper by investigating the determinants of investment in manufacturing sector looks
into ways of boosting investment in that sector. Apart from boosting new investment in the
sector it should also improve the working environment of the existing investor. Therefore to
boost new investment and improve the functioning environment for existing investors the
government and other stakeholder should accomplish the following.

 Provide incentives for manufacturing sector investor which will in turn spur growth in
investment in manufacturing sector. These incentives are necessary because the
peculiarities the manufacturing sector encounter internally and externally. The internal
peculiarities include longer pay back period (it takes many years to realize profits), high
fixed costs and high variable costs. The external peculiarities include lack of credit
constraint and huge infrastructure demand by the sector. To mitigate these peculiarities
government should design a special set of incentives so as to make the sector attractive to
investors.
 One of the factors forcing investors away from manufacturing is the unreliability of
electricity supply. Though it needs a nationwide effort to alleviate this problem the local
government in cooperation with this regions’ electricity branch can do something to
minimize the problem. The local government according to its capabilities can search for
other energy sources or it can work on the distribution to make it more efficient. In the

56 | P a g e
 face of scarcity the distribution shall not be to what everybody needs but to what the
society needs.
 Many entrepreneurs when asked about their future intention of investing in manufacturing
sector claim that they don’t possess necessary business and entrepreneurial skills. As a
result to augment manufacturing investment these entrepreneurs shall have the necessary
education. The sole responsibility of providing this type of good rests in the government.
The government should improvise such skill development programmes which will
enhance the efficiency and productivity of firms’.
 The government should slash the bureaucratic red tape which is hampering the growth of
existing firms and new ones. Unnecessary producers should be cut from the bureaucracy
to improve the efficiency of public sector which in turn creates good governance. A
business friendly government starts its job by cleaning its house. Consequently the
government so as to create a smoothly functioning environment for business shall improve
its working by eliminating red tapes there by closing the door for corruption.
 Last but not least the banking sector should open its door to manufacturing sector. In
addition to short term loans it should also embrace long term loans which earn higher
interest profits to the bank itself. The bureaucracy getting loans approved shall also be
easy, fair and efficient.

57 | P a g e
 References
 Access Capital (2010), “Ethiopia: Macroeconomic handbook 2010”, Access capital research
 Access Capital (2012), “Ethiopia: Macroeconomic handbook 2011/2012”, Access capital
research
 Admasu Shiferaw (2013), “Which firms invest less under uncertainity? Evidence from
Ethiopian manufacturing”, Courant research centre
 Alan Harding, Mans Soderbom and Francis Teal (2004), “Survival and Success among African
Manufacturing Firms”, Centre for Study of African Economies Department of Economics
University of Oxford, UK.
 Alan K. Severn (1972), “Investment and financial behavior of American direct investors in
manufacturing”, Federal reserve board
 Amsalu Woldie (2008), “Sustainability of Ethiopian Economy: A Comprehensive Wealth
Perspective”, Debre Berhan University
 Antonio Ruiz-Porras and Celina Lopez-Mateo (2011), “Corporate governance, market
competition and investment decisions in Mexican manufacturing firms”, Universidad de
Guadalajara, CUCEA

 Ayalneh Yeshaw (1986), “The Structural Analysis of the Manufacturing Sector in

Ethiopia”, Paper submitted to the First National Symposium on Industrial Development in
Ethiopia, Addis Ababa
 Berihu Assefa (2010), “Why the service sector has expanded so fast relative to the
manufacturing sector”, blog on economics and political economics
 Bryant Cannon (2009), “Investment opportunities in Mekelle, Tigray state, Ethiopia”, MCI
and VCC working paper series on investment in the millennium cities
 Bulti Terfassa (1989), “Recent trends In The Development of manufacturing industries in
Ethiopia”, Paper submitted to the First National Symposium on Industrial Development in
Ethiopia, Addis Ababa
 Centre for the Study of African Economies (1997),“Investment in Africa’s Manufacturing
Sector: a Four Country Panel Data Analysis”, University of Oxford
 Dale W. Jorgenson (1966), ”Capital theory and investment behavior”, University of
California, Berkeley

 Dale W. Jorgenson (1967), “The theory of investment behavior”, University of California,

Berkeley
 Ebru ÇAĞLAYAN (2012),”Heteroscedastic probit model: an application of home ownership
in Turkey” , UN

 Eddy L. Ladue, Lynn H. Miller and Joseph H. Kwiatkwoski (1989), “ An analysis of alternate
micro level models of investment behavior”, Cornell University

i|Page
 El-hadj and M. Bahy (2009), “A Three sector model of structural transformation and
economic development”, University of Auckland
 Esayas G/kidan (2013), “Socio-economic Factors affecting urban private investment :The
case of Mekelle city”, Senior essay submitted to economics department
 Gilberto Libanio and Sueli Moro (2009), “Manufacturing industry and economic growth in
Latin America: A Kaldorian Approach”, CEDEPLAR / Federal University of Minas Gerais
 Henry G. Aubrey (1955), “Investment decision in underdeveloped countries”, Federal
Reserve Bank New York
 James G. MacKinnon and R. Davidson (1985),”Heteroskedasticity-Robust Tests in
Regression Directions”, Queen’s University working paper

 Jason P. Brown, Raymond J.G.M. Florax and Kevin T. McNamara (2009), “Determinants of
investment flows in US manufacturing”, Purdue university
 K. A. Bohr (1953), “Investment criteria for manufacturing industries in underdeveloped
countries”, International Bank for Reconstruction and Development
 Keith D. Brouthers and Lance E. Brouthers (2003), “Why Service and Manufacturing Entry
Mode Choices Differ: The Influence of Transaction Cost Factors, Risk and Trust”,
University of Texas El Paso
 Laura Alfaro (2003), “Foreign Direct Investment and Growth: Does the Sector Matter?”,
Harvard Business School
 Marion J. Levy, JR. (1955), “Some social obstacles to capital formation in underdeveloped
areas”, Princeton University
 Michael P. Todaro and Stephen C. Smith (2012), “Economic Development, eleventh edition”,
Addison Wesley
 N.Gregory Mankiw (2005), “Macroeconomics, seventh edition”, Harvard University press
 Robert D. Hormats (2010), “Importance of Investment in the Global Economy”, World
investment forum
 Robert Rowthorn and Ramana Ramaswamy (1997), “Deindustrialization–Its Causes and
Implications”, International Monetary Fund, Publication Services
 Ronald Schettkat and Lara Yocarini, “ The Shift to Services: A Review of the Literature”,
IZA discussion paper
 Rudiger Dornbusch and Stanley Fisher (1977), “Macroeconomics, sixth edtion”, Mcgraw-Hill
 Santosh Kumar Das (2003), “An analysis of investment in the registered manufacturing
sector in India”, Paper for The Indian Econometric Society Conference
 Shakeeb Khan (2012), “Distribution free estimation of heteroskedastic binary response
models using probit/logit criterion function”, Duke univeristy
 The World Bank (2004), “Determinants of Private Sector Growth in Ethiopia’s Urban
Industry: The Role of Investment Climate”, Washington D.C.
 The World Bank (2009), “Toward the Competitive Frontier Strategies for Improving
Ethiopia’s Investment Climate”, Washington D.C.

ii | P a g e
 Thomas cornelißen (2005), “Standard errors of marginal effects in the heteroskedastic
probit model”, University of Hannover
 UNCTAD (2007),”Globalization and the Least Developed Countries: Issues in Trade and
Investment”, United Nations Ministerial Conference of the Least Developed Countries
 UNCTAD (2013), “The least developed countries report 2013: Growth with employment for
inclusive and sustainable development”, United nation publication
 Urgaia Rissa (2007), “ The growth of industrial manufacturing in Ethiopia and its
contribution to GDP”, Masters of Science thesis to Addis Ababa University
 William H. Greene (2002), “ Econometric analysis, fifth edition”, New York University
 Zerihun Gudeta (1999), “Agriculture versus non-agriculture in the Ethiopian economy: A
study of comparative performance since the 1960s”, University of the Free State, South Africa

Other Government Publications

 NBE Annual Report: 2004-2012

iii | P a g e
 APPENDIX I
Table A-1: Results of the heteroskedastc probit model

Heteroskedastic probit model Number of obs = 94

Zero outcomes = 67
Nonzero outcomes = 27

Wald chi2(7) = 19.27

Log likelihood = -27.15997 Prob > chi2 = 0.0074

invmf Coef. Std. Err. z P>|z| [95% Conf. Interval]

invmf
edu .3381811 .1221557 2.77 0.006 .0987603 .5776018
loan -.0088554 .0591506 -0.15 0.881 -.1247885 .1070777
blackout -.3727001 .1535424 -2.43 0.015 -.6736377 -.0717626
area .0003407 .0010866 0.31 0.754 -.0017891 .0024704
r -.0057999 .0488454 -0.12 0.905 -.1015352 .0899355
rp 2.84928 1.133011 2.51 0.012 .6286204 5.06994
inicap .0001561 .0000643 2.43 0.015 .00003 .0002822
_cons -5.928256 2.067039 -2.87 0.004 -9.979578 -1.876933

lnsigma2
inicap .0000325 9.82e-06 3.31 0.001 .0000133 .0000517

Likelihood-ratio test of lnsigma2=0: chi2(1) = 12.61 Prob > chi2 = 0.0004

Table A-2: Results of the probit model

Probit regression Number of obs = 94

LR chi2(7) = 45.80
Prob > chi2 = 0.0000
Log likelihood = -33.465244 Pseudo R2 = 0.4063

invmf Coef. Std. Err. z P>|z| [95% Conf. Interval]

edu .1898892 .0658878 2.88 0.004 .0607514 .3190269

loan -.042669 .048654 -0.88 0.380 -.1380291 .0526912
blackout -.2902623 .1047026 -2.77 0.006 -.4954756 -.0850491
area .0009297 .0007464 1.25 0.213 -.0005332 .0023927
r -.0359802 .0392848 -0.92 0.360 -.112977 .0410166
rp 2.157537 .8143672 2.65 0.008 .5614061 3.753667
inicap 2.46e-06 4.04e-06 0.61 0.542 -5.46e-06 .0000104
_cons -3.08596 1.208045 -2.55 0.011 -5.453686 -.7182353

iv | P a g e
 Table A-3: Marginal effects of independent variables

Marginal effects after hetprob

y = Pr(invmf) (predict)
= .80810867

variable dy/dx Std. Err. z P>|z| [ 95% C.I. ] X

edu .0083227 .00445 1.87 0.061 -.000393 .017039 8.14894

loan -.0002179 .00147 -0.15 0.882 -.0031 .002664 4.01064
blackout -.0091722 .0056 -1.64 0.101 -.020146 .001802 4.38298
area 8.38e-06 .00003 0.30 0.763 -.000046 .000063 224.021
r -.0001427 .00121 -0.12 0.906 -.002517 .002232 9.38298
rp .0701211 .04123 1.70 0.089 -.010694 .150936 1.02602
inicap -3.89e-06 .00000 -1.87 0.062 -8.0e-06 2.0e-07 74054.1

v|Page
 Fig A-1: Average marginal effects (ey/ex, ey/dx and dy/ex respectively)

Average Marginal Effects with level(95)% CIs

10
Effects on Pr(Invmf)
0 -5 5

edu loan blackout area r rp inicap

Effects with Respect to

vi | P a g e
 Average Marginal Effects with level(95)% CIs
.6 .4
Effects on Pr(Invmf)
0 .2
-.2

edu loan blackout area r rp inicap

Effects with Respect to

Average Marginal Effects with level(95)% CIs

8 6
Effects on Pr(Invmf)
2 0
-24

edu loan blackout area r rp inicap

Effects with Respect to

Appendix II

Questionnaire
Mekelle University
College of Business and Economics
Department of Economics
Dear respondents
This questionnaire is prepared by third year economics department student. The objective
of this questionnaire is to assess the determinants of investment in manufacturing sector in
Mekelle city. This study is only intended for academicals purpose. Further I want to assure
you that your identity will never be disclosed and your response will be kept confidential.
Personal Information

vii | P a g e
1. Sex Male Female
2. Age 18-30 30-45 >45
3. Marital status Married Single Divorced
Separated
4. Education level
General Information
5. Which sector are you participating in?
Manufacturing Agriculture Service
6. Why do you choose the sector you are participating in?

7. Initial capital
8. What are the sources of your initial capital?
Own money
Loan from bank
Loan from other person
Other sources
9. How do you evaluate the prevailing interest rate in Ethiopia?
High Modest Low
10. When you ask loan from banks how do you evaluate the collateral requirement of banks
compared to the principal money you request (express in terms percentage)?

11. How many times have you got loan from banks in the last 10 years?

12. How many employees does your company have (permanent and temporary)?

13. How do you assess the current provision of electricity?

Good Modest Poor
14. How many times expressed in hours do you face power blackout per week in your area?

15. Who is the owner of the land where your company is based on?
The company’s Government owned Rent
16. How many hectares of land where you provided by government?

viii | P a g e
17. How do you compare the price of internally produced and imported goods and services
that you are currently producing or providing?

Price of internally produced goods

Price of imported goods

18. The following questions will be used the preference of money of investors in different
time periods. So you are kindly requested to select from the given alternatives. Do you
choose
A. 10000 birr now or
10700 birr next year
B. 10000 birr now or
10800 birr next year
C. 10000 birr now or
10900 birr next year
D. 10000 birr now or
11000 birr next year

Questions related to investors in manufacturing sector

19. What are the main challenges you faced while investing in manufacturing sector? You
can choose more than one option.
Lack of access to finance
Lack of access to reliable infrastructure (electricity, water, communication and
transport)
Cheap imports
Lack of access to land
Tax rates
Tax administration
Corruption
Policy uncertainty
Bad governance
ix | P a g e
Business regulation
Legal system

Thanks in advance for your cooperation

x|Page