Professional Documents
Culture Documents
net/publication/320868688
CITATIONS READS
0 7,765
1 author:
SEE PROFILE
All content following this page was uploaded by Zelalem Ejigu Kabeta on 29 November 2017.
DOCTOR OF PHILOSOPHY
IN
ECONOMICS
DOCTOR OF PHILOSOPHY
IN
ECONOMICS
by
Zelalem Ejigu Kabeta
Supervisor
Professor Inderjeet Singh
Supervisor’s Signature,
Page | i
DECLARATION
I, Zelalem Ejigu Kabeta, certify that the work embodied in
this Ph. D. thesis is my own bonafide work carried out by me under
the supervision of Professor Inderjeet Singh from 13/05/2014 to
10/03/2017, at Department of Economics, Punjabi University,
Patiala. The matter embodied in this Ph. D. thesis has not been
submitted for the award of any other degree/diploma. I declare that
I have faithfully acknowledged, given credit to and referred to the
research workers wherever their works have been cited in the text
and the body of the thesis. I further certify that I have not wilfully
lifted up some other’s work, paragraph, text, data, results, etc.
reported in the journals, books, magazines, reports, dissertations,
theses, etc., or available at web-sites and included them in this Ph.
D. thesis and cited as my own work. I also declare that I have
adhered to all principles of academic honesty and integrity and
have not misrepresented or fabricated or falsified any
idea/data/fact/source in my submission. I understand that any
violation of the above will be cause for disciplinary action by the
University.
Page | ii
DEDICATION
To my Parents Abaye and Ejigu, for your care, love, and support
And
And
Page | iii
ACKNOWLEDGMENTS
Primarily, I would like to express my gratitude to my
supervisor Professor Inderjeet Singh, whose guidance, enthusiasm
and encouragement are invaluable to the completion of this
study. It was the first conversation I made with him on the topic
that gave me the impetus to make this study.
I am also grateful to the Department of Economics, Punjabi
University Patiala for giving me admission and capacitating me
through a coursework program. I am also grateful to Department
heads Professor Inderjeet Singh and later Professor Lakhwinder
Singh, who were very helpful throughout the course work and the
study. In addition, I appreciate the services provided by the office
assistants and the librarians at the department. Last but not the
least; I would like to thank Wollega University and Ministry of
Education of Ethiopia for permitting the scholarship program and
financing the study.
Special thank also goes to my friend Tasew Tadesse for
enlightening me regarding the time series model in the study as
well as the insightful comments and ideas we shared. My gratitude
also goes to Ermias Engida and my dear friend Henock Adamu for
providing me important ideas and data on Ethiopian Social
Accounting Matrix.
Finally, I would like to thank my parents. My mother
Azalech Ayele (Abaye) and my father EjiguKabeta, for your care,
love, support. Abaye, this should belong to you more than it
belongs to me. My sisters Eske, Sihnu, Hawiye, thank you for your
unconditional love and support. In addition, I am thankful to my
family Engida, Kete, Alye, Bele, Adanu, Getish, Tigist, Sewdegu,
Trhas. Thank you for your encouraging words. If it were not for
the friendship of the following people, my journey would surely
have been more difficult and lonely. My friends Tamru, Tasew,
Dereje, Habtamu, Gahsaw, Teshale, and Tefera thank you for your
friendship. When times get tough, I always turned to Medi who
deserves all the credit for her continuous support and love for
keeping me sane especially at the final stages of my study. I love
you Medi. I would like also to extend my special thanks to my dear
friends Yite, Bety, Get, Heny, Bini, Shumet and Eyob. Thank you so
much for your friendship.
Page | iv
ACRONYMS
BL Backward Linkage
CGE Computable General Equilibrium
EDRI Ethiopian Development Research Institute
EEA Ethiopian Economic Association
ESSP Ethiopian Strategy Support Program
ETB Ethiopian Birr
FL Forward Linkage
GDP Gross Domestic Product
GDPPC Gross Domestic Product per Capita
GERD Grand Ethiopian Renaissance Dam
GGDC Groningen Growth and Development Center
GTP Growth and Transformation Plan
GVA Gross Value Addition
HDI Human Development Index
HDR Human Development Report
ICT Information Communication Technology
IFPRI International Food Policy Research Institute
IMF International Monetary Fund
IO Input Output
ISIC International Standard Industrial Classification
LP Labor Productivity
NBE National Bank of Ethiopia
NLFS National Labor Force Survey
PASDEP Plan for Accelerated and Sustainable Development to End Poverty
PPP Purchasing Power Parity
SAM Social Accounting Matrix
SDPRP Sustainable Development and Poverty Reduction Program
TFP Total Factor Productivity
UN United Nations
UNCTAD United Nations Conference on Trade and Development
UNDP United Nations Development Program
USD United States Dollar
WB World Bank
WTO World Trade Organization
Page | v
TABLE OF CONTENTS
CERTIFICATE .............................................................................................................. i
DECLARATION ......................................................................................................... ii
DEDICATION ............................................................................................................ iii
ACKNOWLEDGMENTS .......................................................................................... iv
ACRONYMS ............................................................................................................... v
TABLE OF CONTENTS ............................................................................................ vi
LIST OF TABLES .................................................................................................... viii
LIST OF FIGURES ..................................................................................................... x
CHAPTER - I: INTRODUCTION .............................................................................. 1
1.1 Background......................................................................................................... 1
1.2 Problem Statement ............................................................................................. 2
1.3 Objectives .......................................................................................................... 5
1.4 Justification ........................................................................................................ 6
1.5 Significance ....................................................................................................... 6
1.6 Scope ................................................................................................................. 7
1.7 Data and Methodology ...................................................................................... 7
1.8 Organization ...................................................................................................... 8
1.9 Main conclusion of the Study ............................................................................ 8
1.10 Main Limitations .............................................................................................. 9
CHAPTER-II: REVIEW OF THEORY AND EMPIRICS ........................................ 10
2.1 Theories on Service Sector Growth and Structure ........................................... 11
2.1.1 Conceptualization of Service Sector ...................................................... 11
2.1.2 Structural Change and the Growth of Service Sector ............................ 17
2.1.3 Determining Factors of the Service Sector Growth ............................... 30
2.1.4 Sustainability of Service-led Growth .................................................... 35
2.1.5 Sectoral Inter-Linkage of the Service Sector ......................................... 37
2.2 Empirical Review of Literature ....................................................................... 38
2.2.1 Growth of the service sector .................................................................. 38
2.2.2 The Sustainability and Viability of Service-Led Growth ..................... 40
2.2.3 Structural Change, Economic Growth, and the Service Sector ............ 44
2.2.4 Determining Factor of Service Sector Growth ..................................... 51
2.2.5 The Sectoral Inter-linkages of the Service Sector ................................. 58
2.2.6 Experience of Developing Countries in Structural Change and Service
Sector Growth ........................................................................................ 61
CHAPTER-III : METHODOLOGY, DATA AND MODEL ..................................... 81
3.1 Growth Decomposition Analysis ..................................................................... 82
3.1.1 Job Generation and Growth Decomposition .......................................... 82
3.1.2 Shift Share Analysis .............................................................................. 84
3.1.3 Structural Bonus and Structural Burden Hypothesis ............................. 87
Page | vi
3.2 Specification of the Autoregressive Distributed Lag Model ............................ 89
3.3 Sectoral Inter-linkage Analysis ....................................................................... 94
3.3.1 Social Accounting Matrix ...................................................................... 95
3.3.2 Input-Output Model Framework ............................................................ 98
3.3.3 Description of Ethiopian 2005/06 SAM and I-O table ......................... 100
3.3.4 Inter-industry linkage analysis of Service sector.................................. 102
3.3.5 Key Sector identification ...................................................................... 107
CHAPTER-IV: STRUCTURAL CHANGE AND PRODUCTIVITY IN ETHIOPIAN
ECONOMY ........................................................................................ 108
4.1 Growth of the Service Sector in Ethiopia ....................................................... 108
4.1.1 Service sector growth in output, employment, and productivity.......... 108
4.1.2 Growth and Composition of Service sub-sectors ................................. 122
4.1.3 Composition of Service sub Sectors in GVA and Employment ........... 130
4.1.4 The Contribution of Service sector in Output Growth and
Employment ........................................................................................ 133
4.2 Role of the Service Sector in Structural Change in Ethiopia ...................... 143
4.2.1 The Shift-share Analysis ...................................................................... 143
CHAPTER- V: DETERMINANTS OF GROWTH OF SERVICE SECTOR AND
ITS SUSTAINABILITY ..................................................................... 158
5.1 Determinants of Service Sector Growth in Ethiopia ..................................... 158
5.1.1 Descriptive Statistics ............................................................................ 159
5.1.2 Unit Root Test ..................................................................................... 160
5.1.3 Bound Test for Co-integration Analysis............................................... 164
5.1.4 Regression Results of ARDL ............................................................... 165
5.2 Sustainability of Service-Led Growth in Ethiopian........................................ 174
CHAPTER- VI: LINKAGE ANALYSIS OF SERVICE SECTOR IN ETHIOPIA . 184
6.1 Service Sector Inter-linkages in Ethiopia ....................................................... 184
6.1.1 Description of Ethiopian Economy from 2005/06 SAM ...................... 184
6.1.2 Structure of Import and Exports ........................................................... 193
6.1.3 Results from SAM and Input-Output Analysis .................................... 195
Page | vii
LIST OF TABLES
Table 3.1: Basic Structure of SAM 97
Table 3.2: Input-Output Framework of Ethiopia for 10 sectors 101
Table 4.1: GVA Share of Ethiopia across Comparator Countries (1961- 109
2011)
Table 4.2: Contribution of Sectors to Real GDPPC Growth of Ethiopia 110
(2007/08 – 2014/15)
Table 4.3: Employment Share for Ethiopia and Comparator Countries 116
(1961-2011)
Table 4.4: Labour Productivity level of Ethiopia and Comparator 119
Countries (1961-2011)
Table 4.5: GVA Share of Service Sectors, Ethiopia and Comparator 124
Countries (1961-2011)
Table 4.6: Employment Share of Service Sectors, Ethiopia and 127
Comparator Countries (1961-2011)
Table 4.7: Labour Productivity of Service Sectors, Ethiopia, and 128
Comparator Countries (1960-2011)
Table 4.8: Composition of Service Sectors GVA and Employment 132
Share in Ethiopia (1961-2011)
Table 4.9: Decomposition of Growth in PCGDP, Ethiopia (1999 - 2013) 138
Table 4.10: Contribution of Sectors to the Change in Employment rate, 139
Ethiopia (1999 - 2013)
Table 4.11: Contribution of Employment Changes to Change in 140
PCGDP, Ethiopia (1999-2013)
Table 4.12: Growth Decomposition, Contribution to Total Growth in 142
GDPPC, Ethiopia (1999-2013)
Table 4.13: Shift-share Result of Labour Productivity Growth, Ethiopia 145
(1961 - 1991)
Table 4.14: Shift-share Result of LP growth, Ethiopia (1992 - 2011) 146
Table 4.15: Shift-share of LP growth, Role of Sectors in Ethiopia 150
(1961 – 1992)
Table 4.16: Shift-share of LP growth, Role of Sectors in Ethiopia 152
(1992 – 2011)
Table 4.17: Gross Value Added, Employment, and Labor Productivity, 156
Ethiopia (1961 - 2011)
Table 5.1: Pearson Correlation for Service GVA and Employment 159
Table 5.2: Augmented Dicky-Fuller Test for Unit Root 162
Table 5.3: ADF Unit Root Test with Break 163
Table 5.4: ARDL Bound Test for Service GVA and Service 164
Employment
Table 5.5: Estimated Long run and Short run Coefficients using the
ARDL Approach 168
Page | viii
Table 5.6: Diagnostics Tests 173
Table 5.7: Productivity Level and Growth in Ethiopia (1981 – 2011) 175
Table 6.1: Aggregated Ethiopian Macro SAM (2005/06) 186
Table 6.2: Share of Sectors in Total Gross Output and Total Demand 187
Table 6.3: Share of Sectors Value Added at Factor Cost 189
Table 6.4: Households Consumption Expenditure Shares by Sector, 192
Ethiopia (2005/06)
Table 6.5: Export and Import trade Structure of Ethiopia, 2005/06 194
Table 6.6: Production Technology Coefficients, Ethiopia (2005/06) 197
Table 6.7: Backward and Forward Linkages, Ethiopia (2005/06) 200
Table 6.8: Weighted Backward and Forward Linkages, Ethiopia 203
(2005/06)
Table 6.9: Pure Backward and Forward linkages, Ethiopia (2005/06) 205
Page | ix
LIST OF FIGURES
Figure 4.1: Sectoral Shares and GDPPC Growth of Ethiopia (1961- 112
2011)
Figure 4.2: Relationship Between Sectoral Output Shares in GVA and 113
GDPC (2013)
Figure 4.3: Relationship Between Sectoral Shares in Employment and 117
GDP Per Capita (2013)
Figure 4.4: Relative Labor Productivity, Ethiopia and SSA (1961 – 121
2011)
Figure 4.5: Sectoral Labor Productivity Levels of Ethiopia (1961- 122
2011)
Figure 4.6: Labour Productivity of Services and Manufacturing in 130
Ethiopia (1961-2011)
Figure 4.7: Composition of Service Sector in GVA in Ethiopia (1961- 132
2011)
Figure 4.8: Composition of Service Sector Employment in Ethiopia 133
(1961-2011)
Figure 4.9: Contribution of Sectors to Annual GVA growth (1962- 135
2011)
Figure 4.10: Contribution of Sectors to Employment growth (1962- 136
2011)
Figure 4.11: Shift Share Decomposition results, Labor Productivity 155
Growth, Ethiopia (1961 - 2011)
Figure 5.1: Line Graph for Service GVA, Employment, and 161
Determinant Variables
Figure 5.2: Service GVA and Employment (Hodrick-Prescott Filter 164
Decomposition)
Figure 5.3: Hodrick-Prescott Decomposition of GDPPC Income, 172
Ethiopia (1981-2011)
Figure 5.4: Cumulative Sums of Recursive Residuals for Service GVA 173
Figure 5.5: Cumulative Sums of Recursive Residuals for Service 173
Employment
Figure 5.6: Annual Growths of the Service and Manufacturing Sector, 176
Ethiopia (1982-2014)
Figure 5.7: Service Convergence in Ethiopia (1981 – 2001) and (2001 177
– 2011)
Figure 5.8: Manufacturing Convergence in Ethiopia (1981 – 2001) 178
and (2001 – 2011)
Figure 5.9: Sectoral Productivity and Change in Employment Share, 180
Ethiopia
(1981–2001)
Figure 5.10: Sectoral Productivity and Change in Employment Share, 181
Ethiopia (2001 – 2011)
Figure 5.11: Share of Service Export Earnings and Manufacturer 182
Imports
in Ethiopia
Figure (2005-2012)
6.1: Relative Factors Share of Ethiopia (2005/06) 189
Figure 6.2: Sectoral Composition of Labour, Ethiopia (2005/06) 191
Page | x
Figure 6.3: Key Sectors Based on Hirschman-Rasmussen Linkages, 207
Ethiopia (2005/06)
Figure 6.4: Key Sectors based on weighted Backward and Forward 209
Linkages, Ethiopia (2005/06)
Figure 6.5: Key Sector based on Pure Backward and Forward 211
Linkages, Ethiopia (2005/06)
Figure 6.6: Economic landscape of Ethiopia (2005/06) 213
Page | xi
CHAPTER-I
INTRODUCTION
1.1 : Background
Page | 1
1.2 : Problem Statement
Page | 2
Agriculture in Ethiopia has been the center of the country’s
development policy, given the largest role and share that the sector has in the
economy. Various studies on Ethiopia have identified that agricultural growth
induces higher overall growth and faster poverty reduction and larger growth
effect than growth in non-agricultural sector (Diao, Hazell, Resnick, &
Thurlow, 2007; Mellor & Dorosh, 2010; Engida, et al., 2011). According to
Diao et al. (2007), the impact of growth on poverty is larger when the
additional growth is driven by agriculture rather than non-agriculture sector. In
addition, according to Dorosh & Thurlow (2014), in Ethiopia, there are larger
linkages between agricultural production and small towns and these shows that
redirecting urban growth toward towns rather than cities leads to broad-based
economic growth and poverty reduction. Within non-agriculture, industry-led
growth has a larger multiplier than services-led growth. Both industry and
services have weak linkages to rural areas. This suggests that, based on current
economic structures, industrialization in Ethiopia will weaken national growth
linkages and widen the rural–urban divide (Dorosh & Thurlow, 2014).
Page | 3
investments prefer investment in the service sector as the manufacturing sector
is constrained by many factors. These constraints are infrastructure, power
shortage, transportation, and customs clearing and associated services during
import and exports, shortage of foreign exchange to import intermediate
inputs, access to finance, lack of skilled workforce in local labor markets and
other related problems. Therefore, the productivity improvements and efficient
services specifically in the transportation, storage and communication
services, finance and business services, and training institutes in the education
sector can facilitate the gear for the growth of the manufacturing sector.
Various studies have shown that productivity in services plays a critical role as
a strategic driver of competitiveness in the manufacturing sector (World
Economic Forum, 2015). Moreover, opening up of the economy to
international service and further liberalization of the economy will attract FDI
that could be used to finance import bills of the intermediate goods.
Page | 4
The role of the service sector in Ethiopia, hence, requires a detailed
analysis of structure and growth of the sector that covers not only composition
and temporal dynamics but also the linkage patterns and the determinants of
tertiarization. In this context, this work is an attempt to analyze the structure
and growth of the tertiary sector in Ethiopia with special reference to
structural change, linkage patterns, and determinants.Therefore, with the
objective of transforming its economy and joining the middle-income country
group, Ethiopia is striving to industrialize its economy. However, with a
higher share of GDP contributed by the service sector and low level of
manufacturing sector growth; industrializing the economy is a daunting task.
So, evaluating the service sector growth option at this time may provide
alternative strategies to move the manufacturing sector and the economy
forward.
The main objective of the study is to analyze the structure and the
growth of the Ethiopian service sector. The specific objectives are as follows:
a) To assess the growth trend of Ethiopian service sector and its sub-
sectors.
Page | 5
1.4 : Justification of Study
This research on the service sector growth and structure focuses on one
of the important macroeconomic issues of the time in developing countries.
The service sector growth is a controversial issue in development economics
as some consider it as unproductive and cannot lead economic growth and
others view it as a dynamic sector with a capacity to be an engine of growth.
Given this controversy and the problem associated with poor performance of
the manufacturing sector and overgrowth of the service sector, this study will
be pertinent to developing countries like Ethiopia. The study has relevance to
sustainable economic development in Ethiopia and it has a multifaceted
contribution to Ethiopian economy and fills the research gaps that exist in
service sector growth and structural change literature. The thesis brings to the
attention of scholars the role of the service sector in output, employment and
Page | 6
structural change in Ethiopia. Further, it tests the various hypotheses that drive
the growth of the service sector. There is also knowledge gap in the linkage of
the Ethiopian service sectors with the manufacturing and agriculture sectors.
The study will be a stepping stone for further studies on the role and
mechanisms of the service sector in stimulating growth in the manufacturing
sector in Ethiopia.
Page | 7
sector and to identify the role of the sector in Ethiopian output and
employment growth and structural change. The Autoregressive Distributed
Lag (ARDL) model, Social Accounting Matrix (SAM)/Input-Output (I-O)
analysis and sustainability analysis are conducted in order to identify the
determinants, inter-linkage, and sustainability of the service sector
respectively.
The thesis has seven chapters, which are organized in the following
way. The first chapter is an introduction, which introduces the study by
presenting the rationale and objectives of the study. The second chapter
presents the theoretical and empirical reviews on service sector growth and
structural change along with the experience of Asian and African countries.
The third chapter focuses on the methodological issues, describing the growth
decomposition analysis, the time series analysis, and input-output analysis.
The fourth chapter presents analyses and discusses the structural change and
productivity in the Ethiopian economy. Chapter five presents results and
discussion on determinants of service sector growth and its sustainability. The
sixth chapter deals with the linkage analysis of the service sector in Ethiopia.
Finally, the last chapter, chapter seven, gives conclusions and policy
implications of the study.
Page | 8
1.10 : Main Limitations
The main limitation of this study is associated with the availability of
recent SAM or IO data for Ethiopia to identify the key sectors and other inter-
linkage analysis for the recent years. Similarly, the other limitation of the
study is that the time included in most of the structural change and time series
analyses are before 2011. Recent data were not available from the same
sources at the time of this research work.
Page | 9
CHAPTER-II
REVIEW OF THEORY AND EMPIRICS
Page | 10
2.1 : Theories on Service Sector Growth and Structure
Hill (1977) has made the earliest attempt to define services. He defined
services as a change in the condition of a person or of a good belonging to
some economic unit, performing some activity for the benefit of the other
(Hill, 1977). He also argues that ‘goods and services belong to different
logical categories’ because one cannot store services, rather services must be
consumed as they are produced. This signifies the specific characteristic
feature of services, which is non-storability. Thus, services can be also defined
Page | 11
by looking into the common or specific features, which make them different
from goods and other types of economic activities. Parasuraman, Zeithaml,
and Berry (1985) identified three characteristic features of services. These are
intangibility, heterogeneity, and inseparability. The intangibility feature
implies that services cannot be counted, tested, stored or verified prior to sale.
Heterogeneity of services arises due to variation in the production of same
services from day to day or from producer to producer as service often
depends upon unreliable labor (Parasuraman, Zeithaml, & Berry, 1985).
Services are also inseparable as production and consumption of service are
simultaneous. In addition, services are perishable, which means that ‘once the
time of potential service passes, the opportunity to sell that service perishes’
(Rust & Chung, 2006). However, recent development in the sector has eased
the problems associated with these features of services and the gap between
goods and services is narrowed. Due to fast development in the information
and technology, some services are having components of tangibility, visibility,
storability and may not require face-to-face contact between producers and
consumers. For instance, software programs and digital electronics services
have only limited tangibility but are storable and transferable. Since some of
the definitions given for services have contributed for the negative
connotations and perceptions towards services, it is important to look it by
classifying services.
The WTO (2010) classification of service into two categories may give
the broader view about services. Accordingly, there are change-effecting
services (transformation services) and margin services. The change-effecting
services are ‘activities of producers that change properties of consuming units
or the physical or/and mental conditions of an agent or good’. In this type of
services, ownership cannot be established, as they cannot be separated from
production and the change brought about may be permanent or temporal. For
instance, it includes transforming physical goods through repairing, cleaning
or transporting, or change in the mental condition of persons through services
of education and entertainment. On the other hand, the margin services
facilitate the exchange of goods or assets or services between the user of such
Page | 12
services and another party. These margin services are offered by wholesale
and retail distribution, financial institutions that provide insurance, packaging,
transportation and communication intellectual property rights and distribution
(WTO, 2010, p. 7). In addition to the transformation and margin services,
there are also knowledge-capturing products. These products are hybrid
products, with characteristics of both goods and services such as newspapers,
electronic media, and digital information. They are transferable and can be
accessed more than once. In addition, the production and consumption of these
products may not be necessarily simultaneous due to their physical existence.
However, such products have the features of services as they can effect
changes on the mental condition of users through the information and
knowledge they contain (UNCTAD, 2015).
Page | 13
goods and services. For instance, input services such as financial management
may be provided by the producer of an output, in which case they are not
measured separately as a services output. The other issue with regard to
measurement is related to the informal sector activities, which is dominated by
services activities, overlap with the unobserved economy in the national
accounts system (UNCTAD, 2015).
Page | 14
that do not require physical proximity (Bhagwati, 1984a). The categorization
is used in the WTO General Agreement on Trade in Services (GATS), which
is the first set of multilaterally negotiated and legally enforceable rules on
international trade in commercial services (excluding government services)
since 1995. The services that do not require physical proximity (the service
only crosses the border) are a cross-border trade (mode 1 in GATS). In this
case, the delivery of the service can take place through telecommunications.
On the other hand, the services that require physical proximity can be further
categorized into three different groups. These are (i) mobile provider and
immobile user – commercial presence (mode 3). This type of services includes
services such as medical services provided by a foreign-owned hospital, and
banking services supplied by a subsidiary of a foreign bank.(ii) mobile user
and immobile provider – consumption abroad (mode 2) that includes hospital
and tourism services, and (iii) mobile user and mobile provider - the presence
of natural persons (mode 4), which requires temporarily migrating of labor to
provide the services such as lectures, consultancy, and similar services.
Page | 15
According to Katouzian (1970), the growth pattern of the service
sector can be analyzed by dividing the sector into new services, old services,
and complimentary services. New services are services that have experienced
a great shift in demand due to ‘high mass consumption', which is an increasing
function of per capita income and leisure time. Because of further growth and
automation, the demand for these services will further increase. The old
services had demand that came from a less equitable distribution of income
and the relative lack of alternative employment for the producers of these
services. However, due to economic progress, these services are continuously
substituted by industrial durable consumer goods and the new services.
Finally, the demand for complementary services is mainly associated with the
growth of demand for intermediate goods due to industrial expansions,
urbanization, and unification of home and international markets. These
services have high growth potential as long as the technological revolution
continues (Katouzian, 1970).
Eichengreen and Gupta (2013) also classify service into traditional and
modern services. They argue that the aggregation of services as one broader
activity obscures two distinct ‘waves of service sector growth’. The first wave
of the service sector growth occurs in economies having relatively low level of
per capita GDP in the traditional service sectors (such as personal services).
The second wave occurs at higher per capita incomes in activities (such as
financial, communication, computer, technical, legal, advertising and business
services) that depend largely on information technology, and it is easily
tradable across borders. Even though the share of services has increased since
the 1970s in all countries irrespective of their income levels, growth in the
modern services (second wave) is observed in lower income levels in recent
periods than before. However, the change in the second wave is not equally
observed in all countries. Countries that are ‘open to trade, more democratic,
and relatively close to the global financial centers have increased their share in
modern services’ (Eichengreen & Gupta, 2013).
Page | 16
Therefore, the service sector involves economic activities that have
diverse characteristic features than the goods sector that have limited their
tradability, transferability, and storability. However, in this age due to
advancement in information and communication technology, the sector is
expanding in all economies, specifically; developing countries are flooded
with service activities ahead of their underdeveloped manufacturing and
industrial sectors. The growth of service sector is highly interlinked with
changes in the structure of economies. Based on early period observations
from the stages of growth theories to the recent observation of the waves of
service sector growth, the change in the composition of service sector growth
involves structural change. The next sub-topic discusses the growth of the
service sector and tries to put the sector in the economic growth theories and
structural change literature.
Page | 17
theory focuses on the mechanisms by which developing economies transform
their economic structures from the dominant traditional subsistence agriculture
to the modern manufacturing and service economy. The proponent of this
theory used modern economic theory and statistical analysis to identify the
process of structural change to a sustainable rapid economic growth of
developing economies. The well-known examples of this model are the ‘two-
sector surplus labor’ theoretical model of Arthur Lewis Lewis (1954), and
‘patterns of development’ empirical analysis of structural change model of
Chenery & Syrquin (1989).
Page | 18
production, physical and human capital accumulation, the change in consumer
demands from food and basic necessities to manufactured goods and services.
In addition, they observed ‘growth of cities and urban industries due to the
migration of people from farms and small towns and the decline in family size
and overall population growth, and population growth first increasing and then
decreasing in the process of development’ (Todaro & Smith, 2012). The
various schools of development economists of 1950s and 1960s ‘emphasized
the importance of structural change and saw structural differences because of
market failures’. In addition, they proposed the use of government
interventions for facilitating structural change by import substitution. Hence,
most developing countries and multilateral development institutions followed
these policy recommendations by using Keynesianism as the main intellectual
foundation. However, results were disappointing as the gap between
developing countries and the industrialized countries widened. According to
Lin (2012), the structural difference among countries at different levels of
development was ignored in the policy recommendations(Lin, 2012).
The other development theory that has followed the structural change
theory was the international dependence model, which gained prominence
during the 1970s among intellectuals of developing countries due to the
dissatisfaction with the previous stages of growth and structural change
models. The model views developing countries as affected by their
institutional, political, and economic relations; and they are caught up in a
dependence and dominance relationship with rich countries. The three major
streams of dependence mode are the neocolonial dependence model, the false-
paradigm model, and the dualistic-development thesis.
In the 1980s and 1990s, some countries of the developed world with
conservative governments started advocating a neoclassical (neoliberal)
counter-revolution economic theory and policy. The main argument of the
theory is that underdevelopment is the result of poor resource allocation due to
incorrect pricing policies and too much government intervention and
regulation of the economy by governments of developing economies. The
Page | 19
theory also favored supply-side macroeconomic policies and privatization of
public corporations in the developed countries, and free markets and the
dismantling of public ownership in developing countries. The traditional
neoclassical growth theory has also revived in the 1990s as the cornerstone of
the neoclassical free-market argument. Accordingly, the assertion is that
liberalization of national markets by the developing nations could draw
additional domestic and foreign investment, which can increase the rate of
capital accumulation. The Solow neoclassical growth model represented this
theory, which argues that there are diminishing returns to each factor of
production but constant returns to scale and the exogenous technological
change generates long-term economic growth (Todaro & Smith, 2012, p. 128).
According to Lin (2012), the emergence of stagflation in the 1970s, the Latin
American debt crisis, and the failure of the socialist planning system in the
1980s also gave rise to the rational expectations theory and helped to refute
the structuralist theoretical foundation for the state’s role in using fiscal,
monetary, and trade policy. The multilateral institutions and development
agencies advocated this thinking and influenced economic policies of
developing countries through their conditionality programs of stabilization and
structural adjustments. The policy advice and programs promoted economic
liberalization, privatization, and the implementation of rigorous stabilization
programs, whose results are mixed and controversial.
Page | 20
Structural Change
Page | 21
(Chenery, 1986; Chenery & Syrquin, 1986; Kuznets, 1966). It is also about the
reallocation of economic activity across the three broad economic sectors of
agriculture, manufacturing, and services (Herrendorf, Rogerson, & Valentinyi,
2014). Several stylized facts characterize the outcome of structural change or
transformation process. These are: (i) a declining share of agriculture in total
output and employment; (ii) rapid urbanization; (iii) emergence a modern
industrial and service economy; and (iv) demographic transition that leads to a
spurt in population growth.
Page | 22
structural transformation is the change in sectoral proportions. An increase in
per capita income is also usually accompanied by the rise in the share of
industrial and service output, which is associated with Engel’s Law, a falling
share of food demand as income increases (Syrquin, 1988).
Page | 23
because of the large domestic market (Syrquin, 1988). However, all the
developed countries have followed a similar process of structural
transformation. See (Chenery & Syrquin, 1986; Syrquin, 1988).
Page | 24
between the traditional agricultural sector and modern manufacturing sector.
The sectoral changes in output, which represents structural change, occur as
development proceeds due to the different effect of income elasticity of
demand in each sector. The income elasticity of demand for agriculture sector
is inelastic while manufacturing goods have higher income elasticity.
However, the income elasticity of demand for services is greater than one
(Falvey & Gemmell, 1996). Besides, this implies that at a higher level of
income services will have more demand than the other sectors. Those sectors
with the strongest linkages to the growth will result in the development of a
leading sector and structural change. The balanced and unbalanced growth
debate in development economics is also related with whether the
development would proceed more effectively with the more investment in a
leading sector or in all sectors. The balanced growth approach focuses on
investment in all sectors to grow by supporting those sectors that may not
grow naturally. That means a big push might be needed by a government to
make a balanced growth in the economy. The unbalanced growth approach, as
described by Albert Hirschman, advocates unbalanced sectoral growth for the
overall economic growth (Hirschman, 1958). He suggests that a developing
economy can promote economic growth by initially investing in industries
with high backward and forward linkages.When growth is unbalanced, prices
will rise in those sectors where output growth is relatively slow, which will
call for more investment by private investors or government.
Page | 25
can ignite industrialization, speed up the process of structural change and lead
the overall economic growth of nations at early stages of development.
Page | 26
economic growth and structural change through high investment in the
manufacturing sector is the strategy most countries follow. In his recent
research, Szirmai (2011) concludes that the manufacturing sector is important
in accelerating growth and achieving catch-up in developing countries that are
at the lower level of per capita income. However, when he compares to the
past 60 years, the market service sectors will become relatively more
important as potential sources of growth and catch up than manufacturing
sector (Szirmai, 2011).
Page | 27
share of the service sector. The share of the service sector grows together with
the growth of the per capita income. For countries that have low per capita
incomes, the large proportion of the population spends the greater share of
their income on basic goods, foodstuffs, and agricultural products. But, along
with the increase in incomes the demand for manufactured products also
increases and the society will spend more of their income on services. Second,
the relative price of services when compared to the price of goods is
continuously increasing due to slower productivity growth in the service
sector. An increase in the productivity of industry results in the rise of wages
with a tendency of the leveling of wages, which in turn increases the prices in
sectors relatively lower growth in productivity. In those sectors, the rate of
wage exceeds that of productivity and leads to an increase in the prices more
than the price of those sectors with a rapid growth of productivity. As a result,
more developed countries have a higher relative price for services than
countries at a lower level of development.
Page | 28
(Fisher, 1939; Clark, 1940; Fuchs, 1968; Glasmeier & Howland, 1994;
Katouzian, 1970; Kuznets, 1973). Fisher had suggested that economic
progress would lead to the emergence of a large service sector following the
development of a primary and secondary sector. Fisher (1935) and Clark
(1940) also attributed the dominance of services of advanced economies into
their level of income. They recognized the low productivity in services as a
factor behind the faster employment growth in services than in industry.
Similarly, Fuchs (1965) also observed the same for the US economy in the
1960s. This productivity differential is the basis of the well-known ‘cost
disease’ hypothesis of services propounded by Baumol (1967). Baumol points
out that productivity improvement in services is harder to achieve than in
goods-producing industries. The higher productivity of the industry was
expected to raise wages even in services more than the productivity growth,
which leads to a tendency for costs and prices of services to increase relative
to goods. The unbalanced growth models by Baumol and Fuchs (1968) helped
in promoting the idea that service-sector activities cannot be made more
efficient through capital accumulation, innovation, or economies of scale
because of their labour-intensive nature.
Page | 29
level of services productivity is assumed to be a because of mismeasurement
of services output. Besides, factors like technological change, deregulation and
increased competition are set to raise productivity at least in selected services
(Maclean, 1996). In his later article, Baumol himself has recognized the case
of ‘progressive’ services with substantial productivity gains (Baumol,
Blackman, & Wolff, 1985).
There have been various theories and arguments regarding the cause of
the shift in the service sector in both output and employment. Here in the
literature, the causes of the shift in services are classified into the following
Page | 30
four categories. These are the ‘hierarchy of needs’ hypothesis, the ‘cost
disease’ hypothesis, the ‘exogenous demand’ hypothesis and
‘deindustrialization’ (Kim, 2006).
Page | 31
service sector, the share of the service sector increases on the current price
basis (Kim, 2006). This is because of the fall in the relative price of the
manufacturing sector as the cost and the relative price of the service sector
increased by an equivalent amount of the productivity gap. According to this
view, if the ratio of services in real value added is more or less constant
irrespective of the income level, the proportion of employment in the service
sector should rise. Therefore, the share of service sector employment is larger
in high-income countries and grows with rising income, because more labor
input is needed due to the low productivity of the sector. Clark (1940) also
puts productivity differences as a major force behind employment shifts.
According to him, labor will be reallocated from manufacturing industries that
experience high rates of productivity growth but stagnating demand, to
services that experience lower rates of productivity growth but rising demand.
Nevertheless, to Kravis, Heston, & Summers (1982), the driving force behind
the expansion of service employment is the evolution of technology rather
than the change in demand due to the rising income. In addition, they observed
that ‘productivity is lower in poor countries relative to the rich countries in
both services and commodity-producing sectors, but it is lower by a larger
margin in commodities than services’ (Kravis, Heston, & Summers, 1982).
Page | 32
participate in paid economic activities, the household's will have more money
income and the household may reallocate expenditures for additional
consumption of services since the women may have less time for production in
the home. Therefore, the change in female economic participation can increase
the share of services in the economy through increased household
consumption on services (Kim, 2006). Other than intermediate input demand
and female economic participation, other factors like economic reform (in the
1990s) have also contributed to the change in the structure of demand.
Economic reform and liberalization can be seen as causes for the growth of
service sector that is observed after the 1990s. Most of developing economies
have conducted economic reforms by deregulating their economy (Jain &
Ninan, 2010).
Page | 33
ensure that liberalization delivers the expected benefits (Cali, Ellis, & Velde,
2008). According to Cali, Ellis, & Velde (2008), the risks of liberalization in
service are the following. 1) The crowding out of domestic providers and
instead allowing foreign firms to capture the profits for themselves and taking
the money out of the country. 2) There is a risk that private providers will
cream skim the most profitable clients and cease to serve some other groups of
clients or geographical areas that are deemed likely to be unprofitable
specifically in the financial and energy sectors. 3) There will be also risks of
financial sector instability as the world financial crises are transmittable. 4)
The movement of persons to provide services can aggravate skills shortages in
developing countries and results in brain drain. 5) Finally, service
liberalization can have a risk of environmental degradation due to uncontrolled
tourism development associated with deforestation and erosion; degradation
and depletion of biological diversity; disruption of natural habitats; and over-
consumption of resources like fresh water and energy (Cali, Ellis, & Velde,
2008).
Page | 34
2.1.4 : Sustainability of Service-led Growth
Page | 35
In the recent periods, however, literature that challenges the role of
industrialization and the manufacturing sector as an engine of economic
growth are emerging. This is due to the change in the service sectors as some
of the service sectors are fulfilling Kaldor’s observation. For instance, the
business, financial, and the communication service sectors have the scope of
achieving economies of scale, high contribution to technological progress and
enhancing overall productivity growth and(Dasgupta & Singh, 2005; Dasgupta
& Singh, 2006). In addition, Hansda (2001) shows the expansionary effect of
the service sector on other sectors through its interlinkage. Banga & Goldar
(2004) also find the increasing use of the service sector has contributed to the
productivity growth in the manufacturing sector and claims that the service
sector will be ‘successful in creating its own demand, which enables the
service sector to sustain its own growth’ (Nayyar, 2012). According to Nayyar
(2012), the service-led growth could be sustained through its effect in the
international trade. In the past decades, the tradability of the service sector has
increased in the international market. Specifically, business services, transport,
communication and financial services are boosting productivity growth
through ‘learning-by-exporting’ effects, ‘scale economies’, ‘technological
transfers’ and competition (Nayyar, 2012).
Page | 36
economic growth can be sustainable. In addition, the high inter-sectoral inter-
linkage, and service being a major source of employment growth, it could be
important not only for sustainable development but also for poverty reduction
and equitable growth. Thus, in this age, the service-led growth looks more
likely than in the past, three to four decades ago. However, in this case, the
role of the manufacturing and the agriculture sector in sustaining economic
growth cannot be discredited.
Page | 37
Particularly this applies to business services that provide key intermediate
inputs such as finance, human resource recruitment, marketing, legal services,
and information technology to manufacturing and other sectors. The evidence
for the growth of such services is the increasing trend of outsourcing tasks
related to business services to specialized service providers, which makes the
manufacturing firms more cost efficient. Therefore, sectoral inter-linkages, in
both the production and consumption sides are important for growth and
growth can be enhanced and sustained by promoting those sectors with larger
inter-sectoral linkages.
A number of studies have tried to examine the service sector and its
productivity growth after the works of Kaldor (1966), who emphasized that
labor in non-manufacturing sector is less productive. In a recent study, Lee &
Mckibbin (2014) investigated the role of an expanding service sector for
structural adjustment and economic growth in Asia. An employment shift
toward the service sector is observed in 11 Asian economies since 1990. A
decomposition result shows that the service sector has made a significantly
positive contribution to aggregate labor productivity growth, both through
within and structural change effects, exceeding the net contribution of the
manufacturing sector. Despite the strong convergence of labor productivity at
both the aggregate economy level and sectoral levels, there remain significant
differences in labor productivity across economies and across sectors.
Nevertheless, modern services including the transportation, storage, and
communications and the financial and business services have higher
Page | 38
productivity growth, playing a role of a second growth engine in Asia. In
addition, the service trade in Asia, specifically inICT-based services has
increased. Ghani & Kharas (2010) also identified that the global trade in
services has grown faster than the global trade in merchandise goods. The
ratio of service trade in service output for developing countries has increased
much faster than it has for the developed countries. This implies that
developing countries in Asia are more focused on the production of services,
which can be traded, rather than used for domestic consumption.
Eichengreen and Gupta (2010) used the cross-country data from the
European Union (EU) to study the productivity levels for different groups of
services. They divided the service sector into three groups. The first groups are
traditional services, which include wholesale and retail trade, transport and
storage, public administration and defense. The second group, a hybrid of
traditional and modern services, includes hotels and restaurants, education,
healthcare and social work, and other community, social and personal
services. The third groups are modern services that consist the business
services, financial intermediation, computer services, communications, legal
and technical services. They find that the productivity growth of the modern
services is the highest and the fastest growing services are business services,
communications, and banking. The traditional services have a low elasticity
of demand and the hybrid group has a cost-disease problem, which leads to
low productivity.
Park & Shin (2012b), in their research on the Asian service sector
growth, identified that ‘the lower the per capita GDP, the greater the scope for
labor productivity growth in the service sector’. Hence, there is still a lot of
room for services productivity growth in Asia, as the income level of much of
the countries remains relatively low. According Park & Shin (2012b), if a
‘relatively underdeveloped service sector contributes significantly to growth,
then a more developed service sector will contribute even more’. Further, they
suggest that the Asian service sector could generate more jobs and growth if
Page | 39
the obstacles of excessive regulation and state monopolies, barriers to services
trade and FDI are removed.
Page | 40
the service sector the engine for growth and lead a strong and sustainable
growth in Asia. Using simulations of a multi-sectoral general equilibrium
model, they find that ‘faster productivity growth in the service sector in Asia
can significantly benefit all sectors, contributing to more balanced and
sustainable growth of Asian economies’. In Asia, particularly, in India, the
resilience of the economy is due to the resilience of the services sector.
Numerous studies since the 2000s have shown that services induced growth of
GDP of India is feasible.
Aggarwal & Kumar (2012) also raised the question whether service
oriented growth of India is sustainable or not? According to their observation,
however, the high informality of the Indian economy in general and the
service sector in particular, is considered to inhibit the sustainability of the
service led growth. Despite the increasing global economic integration of the
Indian economy, the informal economy continues in terms of its share in the
sectoral and total economy, which can constrain the growth due to the low
productivity. Besides, the service-led growth model is sustainable not only
from the economic perspective but from social and environmental perspectives
as well. According to Singh (2012b), service-led growth is sustainable as
services are tradable without any harm to the environment and they can be
easily transported via satellite without any carbon dioxide emissions.
Page | 41
In addition, Dasgupta & Singh (2006) examined the role of
manufacturing and services in sustaining growth using a Kaldorian
framework. Their findings indicate that manufacturing continues to be an
important sector in economic development, but services those connected with
ICT, also make a positive contribution in a number of developing countries
such as India. In India, apart from ICT related services, many other have a
faster rate of growth than either manufacturing or GDP. Dasgupta & Singh
(2006) concluded that services fulfill the requirements of dynamic sectors in
the Kaldor sense and thus, they could be regarded as an additional engine of
growth. Therefore, in the Indian case, according to Dasgupta & Singh (2006),
with the fast development of the country achieved in ICT, the new engine
might help India to leapfrog in technological development to catch up with
advanced countries.
Ghani (2010) has also studied the service growth pattern of the South
Asian countries and examined the sustainability of the growth. According to
him and other co-authors, the conventional knowledge that labor-intensive
manufacturing is the only sustainable way to promote rapid and sustained
growth is challenged. This is because of the global services revolution that has
altered the characteristics of services. As a result of the ICT revolution, wider
varieties and types of new products can be produced and exported at low costs,
where the South Asian countries have taken advantage of these possibilities.
Page | 42
India is leading service exporter as the country has established a clear
comparative advantage in ICT based services. There are also changes that the
service revolution has brought in most of the developing countries. The wages
in the service sector have risen faster than wages in the agriculture and
industry sectors. The service sector encourages greater labor force
participation of women. In addition, the services growth also leads to a rapid
urbanization as people join the urban labor market in search of jobs.
However, there has been a lot of debate about the capacity of the
service sector to generate employment. The change in the output structure
from agriculture to services has not been reflected by an equivalent change in
the employment. As a result, service-led growth has been jobless growth
(Banga, 2005b). This is the major concern for among researchers that limit the
viability of service-led growth in developing countries. There seems an
agreement that the service sector has limited potential for employment
opportunities due to its low employment elasticity and non-uniform growth of
disaggregated services. However, according to Singh (2012a), ‘synergies from
service-led growth pattern can be built through expansion of services in micro,
small and medium enterprises and rural non-farm economy by the promotion
of entrepreneurship’. Furthermore, liberalizing the service sub-sectors of
wholesale and retail trade, financial services and software services will boost
employment and output growth in the service sector through FDI (Mukherjee,
2013).
Page | 43
of productivity, 2) a dynamic productivity growth (high growth rates coupled
with domestic and international convergence), 3) an expansion of the sector in
terms of its use of inputs, 4) comparative advantage (alignment between
resource requirements of the sector and resource endowments of the country)
and 5) exportability. Based on these criteria Amirapu & Subramanian (2015)
compare the performance of manufacturing against specific service subsectors
of India as a case. They find that the manufacturing sector’s high productivity
and international and domestic convergence in productivity are shared by the
finance, insurance, and real estate service sectors. These subsectors, similar
with the manufacturing, are also too skill intensive and hence unaligned with
India’s comparative advantage (Amirapu & Subramanian, 2015).
Page | 44
countries. The empirical literature on structural change and structural
transformation generally focus on the shifts towards industrialization in terms
of gross value addition, employment and productivity. This section, however,
deals with empirical literature on structural change, productivity growth,
drivers of structural change and urbanization associated with the service sector
growth.
Page | 45
Besides, Aggarwal & Kumar (2012) also attempted to identify the
causality of structural change and growth in India. Accordingly, they find no
significant relationship between growth and structural change during the state-
led growth regime although the causality runs from structural change to
growth in the market driven growth regime. In addition, the finding by Cortuk
& Singh (2013)shows that structural change of the Indian economy is
significant in explaining the growth of the economy for the period of 2000 to
2006 but not vice versa, as growth does not seem to lead to structural change.
They identify a one-way causality from structural change to growth in the
period 1988–2007 but there is no evidence for this linkage before 1988.
Several studies have shown that the TFP growth in the modern services
is increasing and becoming the source of structural change. Verma (2012)
studied the structural transformation and TFP growth in services sectors in
India during 1980–2005. She found that the TFP growth was fastest in
services contributing to structural transformation.Duarte & Restuccia (2010)
Page | 46
also investigated the role of sectoral labor productivity in explaining the
process of structural transformation across countries. According to their
observation, productivity differences across countries are large in the
agriculture and services but smaller in the manufacturing sectors. Over time,
the productivity gaps have been substantially reduced in agriculture and
industry but not in services. As countries move through the process of
structural transformation, relative aggregate labor productivity can first
increase when labor moves from agriculture to industry and later stagnate or
decline as labor moves from agriculture and industry to services. The larger
productivity gaps in services relative to manufacturing across countries are
because of a tendency for services to be less competitive pressure as services
are less traded than manufacturing (Duarte & Restuccia, 2010).
Timmer & de Vries (2009) also studied structural change and growth
acceleration using a modified shift-share method that accounts the
contribution to growth from the expanding sectors for 19 countries in Asia and
Page | 47
Latin America for the period of 1950-2005. They find that growth
accelerations are explained by productivity increases within sectors, but not by
structural change (reallocation of employment to more productive sectors).
One important finding of their study that challenges the conventional wisdom
is that productivity improvement in market services is found to be more
important than productivity growth in the manufacturing sector.
On the other hand, the study by de Vries et al. (2012) on the structural
transformation and its implications for productivity growth in the BRIC
countries (Brazil, Russia, India, and China) shows that reallocation of labor
across sectors is contributing to aggregate productivity growth for China,
India, and Russia, but not in Brazil. In Brazil, however, the increasing
formalization of the economy since 2000 became growth enhancing. The
employment reallocation towards formal activities increases aggregate growth.
Whereas in India the increase in the informality after the reforms is growth
reducing. For example, in India, the informal sector accounts up to 30 percent
of the manufacturing’s value added and 80 percent share in employment,
which indicates large differences in the productivity between the formal and
informal activities.
Page | 48
the course of development and income growth, the inter-sectoral productivity
gaps, as measured by the coefficient of variation of the log of sectoral labor
productivities, tend to decline. This implies that structural change contributes
to convergence within and across countries.
Page | 49
innovation and technology adoption, capital investments especially in the
agriculture sector, changes in consumer demand, and institutional changes in
general and market development in particular.
Page | 50
and resources out of agriculture and rural areas into services and industry in
urban areas (Kolavalli, et al., 2012). Jedwab (2013) documents several stylized
facts regarding the processes of urbanization and structural transformation
using the cross and within country evidence. According to Jedwab (2013),
most developing countries, especially in Asia have urbanized with structural
transformation and resource-rich developing countries, especially in Africa
have experienced urbanization without structural transformation. In Africa,
thus, the processes of urbanization and structural transformation are
disconnected. Contrary to the standard theories of structural transformation,
Africa did not urbanize following a green revolution or an industrial
revolution, but Africa’s urbanization is because of natural resource exports
(Gollin, Jedwab, & Vollrath, 2016).
The empirical studies show that the relationship between income and
service sector growth is positive as it is postulated in the ‘hierarchy of needs’
hypothesis. In order for the view to be supported, however, there must be a
positive correlation, in a cross-sectional and time series data between income
and the proportion of the service sector. Baumol (1985), using cross-sectional
data, shows that income per capita and the proportion of the service sector
may be positively correlated on the nominal basis, but that they are not related
on the real PPP basis. According to Schettkat & Yocarini (2006), through the
analysis of input-output tables of major OECD countries, they reported that as
Page | 51
per capita income increases, the proportion of services in final demand has
increased on the constant price basis as well as on the current price basis.
Page | 52
Eichengreen & Gupta (2013) also examine the relationship between
service sector size, income, and productivity. They find a positive correlation
between output share of services and income per capita. The relationship holds
only for service activities that are a combination of traditional and modern
services consumed by households (education and health) and for modern
services consumed by both households and businesses. Additionally, their
study shows that modern services not only have the highest productivity
growth among the service activities, but their share in output increases rapidly
at high-income levels.
Gani & Clemes (2002) also studied the determinants of per capita
economic growth particularly; the contribution of services to the growth of
real GDP per capita in Asia using data from 1965 to 1994. The result confirms
a strong positive effect of growth in manufacturing and government spending
on service sector expansion. According to them, services have a spillover
effect on manufacturing and the service sector contributes directly to GDP and
per capita income through increased production of services. Further, it also
contributes to GDP and income indirectly through its impact on raising the
productivity of other sectors and enhancing the efficiency of resource
allocation.
Page | 53
state and local levels on services and for Medicare. In addition, the growth in
the private components of final demand is almost equally important with the
increase in government expenditure. Eckstein & Heien (1985) believe the
evidence that services have higher income elasticity than goods is not definite
unless government expenditures as part of the demand are not considered.
However, if government expenditures are included, the income elasticity for
services exceeds unity. In addition, ‘technological progress and the availability
of highly skilled workforce have led to the growth of services in ICT and ICT-
enabled services’ and due to this developed countries are outsourcing services
to developing countries like India, leading to a rise in demand for services
(Bhagwati, 1984a).
Page | 54
contribution of the service input to output growth in manufacturing was about
1 percent in the 1980s, and it increased to about 25 percent in the 1990s.
The determining factors to the shift in the service sector in both output
and employment are studied more comprehensively by Kim (2006). Kim
studied the factors that cause a shift towards services in Korean economy
using regression analysis. He tested the hypotheses of ‘hierarchy of needs’,
‘cost disease’ and ‘exogenous demand shock’. The result of the estimation
shows that the share of the service sector in employment is significantly
associated with inter-sectoral productivity differential and the share of
producer services. Female economic participation ratio also became
significant for the period 1981-2003. However, the share of producer services
is the most significant factor accounting for the rise in the share of the service
sector in total value addition. On the other hand, per capita real value added is
insignificant and shows negative sign in contrast to the ‘hierarchy of needs’
hypothesis. According to Kim (2006), the increase of the service sector's share
in employment is mainly caused by the productivity gap between the
manufacturing and service sectors.
Apart from the major determining factors of service sector growth that
are discussed in the previous paragraphs, economic reforms and liberalization
are other important factors that a number of empirical literature has studied.
Existing empirical literature shows that liberalization and reforms have
contributed to the growth of the service sector (Jain & Ninan, 2010).Moreover,
service liberalization will have positive effects on performance of
manufacturing sector. Thus, due to the increasing service intensity of
manufacturing, one can conclude that service trade barriers not only affect the
service sector but also manufacturing firms (Falk & Jarocinska, 2010).
Page | 55
different stages of economic development. Their estimates confirm that
telecommunication services openness is positive indicating its positive impact
on growth for low-income economies but the opposite is true for financial
services. According to El Khoury & Savvides (2006), liberalizing
telecommunication services trade by lower income economies frequently
produce FDI that contributes to raising productivity growth in the sector.
Moreover, lower income countries do not have the necessary institutional or
regulatory structures to ensure the effective functioning of the financial sector,
which cannot be solved through the greater openness of the financial sectors.
Page | 56
in GATS or in North–South trade agreements is not necessarily the best way
to improve services sector efficiency and exploit the sector’s growth and
employment potential. However, a number of studies confirm the importance
of liberalization and trade openness as one important channel for improving
services performance through adoption of new technologies and competitive
environment created by foreign suppliers. According to Hoekman & Mattoo
(2008), the competitiveness of firms in open economies is increasingly
determined by access to low-cost and high-quality producer services such as
telecommunications, transport and distribution services, financial
intermediation, etc.
On the other hand, Estrada, et al. (2013) studied the service sector
growth and factors determining the growth in low-income Asian countries.
Accordingly, service sector growth has been supported by strong industrial
growth in some countries while in others the critical factors have been
liberalization, structural reforms, government support, and foreign
investments. According to Estrada, et al. (2013), in Viet Nam, services
liberalization led to the rapid expansion of retail trade, telecommunications,
Page | 57
and transportation industries. In addition, Viet Nam had higher FDI in services
due to the country’s entry into the World Trade Organization in 2007 and
several bilateral trade agreements. On the other hand, in Papua New Guinea,
the removal of the state monopoly in the mobile telephone market, greater
competition in the aviation industry, and structural reforms in the financial
sector contributed to the growth in service output and employment. Generally,
based on the experience of the Asian low-income countries, policy reforms
that ease trade in service will help to achieve a balanced growth in which the
service and industry sectors support and reinforce each other. According to
Estrada, et al. (2013), for most of the developing Asia, services that are more
labor-intensive than industry, can foster inclusive growth and serve as an
engine of growth.
Page | 58
Hansda (2001) also used the input-output approach at a much more
disaggregated level with a goal of addressing the sustainability of services-led
growth for the Indian economy. His analysis suggests that the Indian
economy, in general, is services-intensive with the industry being the most
services-intensive sector. The manufacturing sector, construction, and utilities
are all services intensive whereas transport, storage and communication are
industry intensive. However, he finds that the level of linkages of the
agriculture sector, financing, insurance and real estate are quite low.
On the other hand, Falk & Jarocinska (2010) studied the direct and
indirect linkages between manufacturing and services both on the input side
and on the output side for the EU-25 countries and the US. They find that
direct and backward linkages between manufacturing and services increased in
the last few decades with the change more substantial for the backward
Page | 59
linkages. Similarly, there is an increase in service occupations in
manufacturing. On the output side, manufacturing firms generate an increasing
proportion of sales from services activities.
Page | 60
process. The next section discusses the empirical literature on the specific
experiences of Asian and African countries in structural change and service
sector growth.
Page | 61
strengthening human capital and greater labor market flexibility, particularly
in countries with high services shares, can have a significant positive impact
on productivity growth in the sector.
On the other hand, Timmer, de Vries, & de Vries (2014) used the
updated and extended Groningen Growth and Development Centre (GGDC)
10-Sector database to document patterns of structural change in 11 countries
in Asia, 9 in Latin America and 11 in Sub-Saharan Africa from 1950 onwards.
They find that the expansion of manufacturing activities during the early post-
World War II period was related to a growth-enhancing reallocation of
resources in most countries in Asia, Africa, and Latin America. But the
process of structural change stalled in many African and Latin American
countries during the mid-1970s and 1980s. In the 1990s, when growth
improved, workers mainly shifted to the market services specifically to
distributive trade services. However, such services have higher productivity
than much of the agriculture; they are not technologically dynamic and have
been falling behind the world frontier (Timmer, de Vries, & de Vries, 2014).
Page | 62
According to Noland, Park, & Estrada (2012), services’ contribution to
growth has been higher in the South Asian region than in other regions. In
India, the Maldives, and Sri Lanka, roughly over 60 percent of the growth in
2000-2010 was due to services. The sector also contributed to over half of the
growth in Indonesia, Malaysia, the Philippines, and Singapore. Yet, in East
Asia, particularly the Peoples Republic of China, the Republic of Korea, and
Taipei China, the industry rather than the service sector is driving the overall
growth. In most of the South Asian countries and the Philippines, where the
pace of industrialization has been slow, the service sector especially the
modern service sector has played an important role in driving the overall
growth.
In addition, Estrada, et al. (2013) also confirm that the service sector
has been a huge contributor to growth more than industry and agriculture from
2000 to 2010 across the lower-income economies of Asia including
Bangladesh, Cambodia, Nepal, and Uzbekistan, services contributed. When
compared to the 1990s, growth proceeded more rapidly from 2000 to 2010 in
these countries with the service sector as the key driver. However, a key issue
is whether growth in the service sector has led to significant job creation and
to improvements in productivity. In terms of employment, lower-income
economies are still primarily agricultural, but there is some evidence that
services have contributed substantially to employment growth in the past
decade. In Bangladesh, services contributed to about one-third of the
employment growth from 2000 to 2010, which was more than industry did.
Therefore, services will continue to make substantial contributions in lower-
income Asian countries as the share of the output rises with per capita income.
Since the income level of the developing Asia’s lower-income countries is
low, they have relatively wider possibility for labor productivity growth and
hence growth in the service sector as their income increases through time. In
addition, these countries are still in the early stages of transforming, so that the
share of agriculture in GDP is still relatively high and the shares of industry
and services are correspondingly low. Thus, the sectors have sufficient room
to grow in the future (Estrada, et al., 2013).
Page | 63
There are also country-specific studies in Asia predominantly that
focuses on India, China, and South Korea. To start with India, Ramaswamy &
Agrawal (2012)studied employment growth, structure and job quality
outcomes in urban India between 1999/2000 and 2009/10.Their result did not
indicate any acceleration in employment growth of the service sector relative
to manufacturing in the urban areas of India. Rather they find greater duality
in the services sector in terms of the incidence of informality and wage
inequality. The service sector is found to be relatively more skill demanding
than manufacturing. The service sector is also skill-biased as those with more
skills have received higher increases in real wage than the wages in the
manufacturing. Therefore, the service sector will not be the destination for the
millions of low-skilled job seekers. Thus, they suggested that India needs to
focus on manufacturing sector to provide large-scale employment. However,
according to Eichengreen & Gupta (2011), the growth in the service sector
output and employment in India depends on the continued expansion of
modern services and the application of modern information technology to the
traditional services.
Among the few studies that investigated the service sector growth and
structural change in China, Qin (2006)explores whether the growing service
sector in China leads to cost disease, which is a likely consequence of
tertiarization according to Baumol’s unbalanced growth model. According to
Page | 64
the model, tertiarization is likely to decelerate overall productivity
improvement of an economy and hence retard economic growth. According to
Qin (2006), the positive contribution of the service sector to growth in China
is largely due to shifts of labor from the primary sector into services.
However, the signs of cost disease are visible from the weak responses to price
signals in the demand for services, in wage determination and labor input
demand of the service sector.
Page | 65
economies. Wu (2007) found that the role of services in both China and India
has been rising, with China starting from a lower base. Growth has been
driven in both societies mainly by increasing specialization of production,
rising living standards and accelerated urbanization. There are also some non-
economic factors, which are difficult to measure in empirical analysis but have
played an important role in service development including biased development
strategies in China.
South Korea is the other Asian growth miracles that achieved and
sustain high output growth for many years. South Korea fits very well the
structural transformation process accompanying economic development as
described by Kuznets (Bah, 2010). According to Park & Shin (2012a), the
core problem of the country is associated with the underperforming of
productivity growth in the service sector, although the share of services in
output and employment has risen. The analysis by Park & Shin (2012a) clearly
confirms the belief that Korea’s services sector still lags the manufacturing
sector even though deindustrialization already began in the early 1990s.
Therefore, the central challenge for the South Korea in the post-industrial
phase is thus to renovate and upgrade the services sector so that a productive,
high value-added, modern services sector can become an engine of growth.
Park & Shin (2012a) argue that ‘the highly educated workforce of Korea,
which enabled the country to quickly move up the technological ladder, can in
principle also serve as a key ingredient in the leveling up of the services
sector’ (Park & Shin, 2012a).
Page | 66
sector's share in GDP has increased but it has not been able to displace the
labor from primary sector. Indian primary sector is depicting a greater
backward integration with itself, in comparison to South Korea, whereas, the
South Korean service sector is well integrated with other sectors on the
backward side. On the other hand, the ratio of service exports as a percentage
of total exports in India is not only higher than South Korea but it has also
been growing at a much faster pace from 20.2 percent in 1990 to 35.5 percent
in 2010 (Singh & Singh, 2013).
Page | 67
Vries (2013), the expansion of manufacturing activities during the early post-
independence period in Africa led to a growth enhancing reallocation of
resources. The process of structural change held up in the mid-1970s and
1980s, but in the 1990s, labor mainly shifted to the market services. In
addition, the market services activities had above-average productivity levels,
but productivity growth was low and increasingly falling behind the world
frontier. Therefore, the pattern of static gains but dynamic losses of
reallocation since 1990 are observed for many African countries. The situation
is comparable to the patterns observed in Latin America but different from
those in Asia. According to McMillan and Rodrik (2011), the structural
change did not contribute to growth in Africa, despite Africa’s high growth
performance. They also show that during the period from 1960 to 1975, Africa
took a step forward by expanding its manufacturing activities, which was
related to growth-enhancing structural change. However, after 1990 market
services activities expanded and the productivity levels in market services
were above the economy average. The overall effect was a limited role for
structural change post-1990, which compares unfavorably to Africa’s earlier
period of high growth, which is also observed in Latin America, but not in
Asia.
Page | 68
In this respect, rural-urban migration is one of the necessary
components of the structural transformation and economic development
process. The migration of labor out of agriculture into the manufacturing and
service sectors has been a feature of the growth path of every country that has
developed. According to de Brauw, Mueller, & Lee (2014), the lack of more
rapid rural–urban migration and large shares of the population living in rural
areas may hinder economic growth in Sub-Saharan Africa. The migration rates
in Sub-Saharan Africa are very low due to different factors. Some of these
factors are the explicit policies in some countries designed to inhibit rural-
urban migration, policies that indirectly limit mobility (weak property rights
over land), and lack of education, which appears to be an important
determinant of rural–urban migration (de Brauw, Mueller, & Lee, 2014).
Christiaensen & Todo (2013) attempted to unveil the driving forces for
the rapid reductions in poverty in developing countries given limited growth-
enhancing structural change and weak rural – urban migration using a cross-
country panel data from household surveys for developing countries during
1980–2004. Particularly in Africa, structural transformation and urbanization
patterns are different. In some countries, the migration is out of agriculture
into rural off-farm activities and secondary towns, whereas in other countries
there is rapid agglomeration in mega cities. The evidence from Christiaensen
& Todo (2013) suggests that migration out of agriculture into the missing
middle, which is to the rural nonfarm economy and secondary towns, is
strongly associated with poverty reduction than migration/agglomeration to
megacities.
Page | 69
effective, despite slower national growth. Therefore, poverty reduction in the
short-term of necessitates further agricultural investment. Based on these
findings, there is a need to understand the pattern of urbanization in Africa to
exploit the untapped potential of economic gain from the structural
transformation in the long-term and poverty reduction in the short-term.
Barrett, et al. (2017) argues that ending extreme poverty in Africa will require
reallocation of resources from low to high productivity sectors (structural
change). Aryeetey & Moyo (2012) also stressed the huge opportunity of
African countries for significant gains from the structural change in the future,
as the productivity growth due to structural change in 1990 – 2005 was
negative even if there was significant growth within the sectors. However, for
Ajakaiye & Page (2011), breaking into global markets for manufacturing,
agro-processing industry and tradable services are the only viable path to
accelerated structural change in Africa.
Page | 70
In addition, Africa is a marginal player in global services exports and
imports. In 2012, it accounted for only 2.2 percent of the world’s total exports
of services compared to 3.6 per cent for developing America and 24.3 per cent
for developing Asia. At the country level, Egypt and South Africa are Africa’s
two main global service exporters, accounting for 0.49 percent and 0.34
percent of world services exports. On the other hand, Angola, Nigeria, and
South Africa are the main African global service importers, accounting for
more than 0.40 percent of total world services imports. Even though Comoros,
Ethiopia, and Liberia were classified as services exports-dependent during
2009–2012, the share of services in output in these countries did not exceed 50
percent (UNCTAD, 2015).
Page | 71
economies that share at least some of the characteristics to lead a rapid and
sustained growth. According to Enache, Ghani, & O’Connell (2016), these
sectors are the wholesale and retail trade services and the construction sector.
However, both sectors suffer from a lack of tradability and exportability.
Page | 72
countries that fit the second pattern, late transitions out of agriculture and high
employment in wholesale & retail trade services, are Thailand and India.
On the other hand, Uwitonze & Heshmati (2016) attempt to study the
development of the service sector over the years in Rwanda’s economy and
empirically estimate the determinants by using an econometric methodology.
In Rwanda, services are considered as an alternative to manufacturing led
Page | 73
development as the country aims to become a service-based hub to serve
countries in the East African Community following an impressive growth of
the sector. The empirical results are based on micro-data collected from the
Rwanda Enterprise Survey of 2011 and 2014. The results show that the growth
of service firms in innovation, sales, and turnovers is driven by access to
finance, increased labor force, training personnel, ICT applications, embryonic
innovations and the tax system. The size of the loans approved by financial
institutions like banks and cooperatives has had a positive effect on the three
years total annual sales, capital used by the service firms also positively
influenced turnovers of service firms and the acquisition of fixed assets
positively affected service innovativeness.
Page | 74
Eshete & Kimuyu (2014a) suggests the active involvement of the government
in the manufacturing sector, foreign trade openness, service trade
liberalization for driving sectoral total factor productivity.
Another study by martins (2014) on the Ethiopian structural change
from employment perspective revealed that structural change in the
composition of employment in the sectors is absent in the Ethiopian economy
(Martins, 2014). Based on data disaggregated into eight sectors for the period
1996–2011, his analysis suggests that the structure of output has changed
considerably from agriculture to services. Yet, changes in the composition of
employment have lagged behind. In addition, the labor productivity growth in
the period has been strong across most sectors, though mainly driven by the
within-sector productivity improvements. Structural change is also
accelerating in its pace and relative contribution to output growth. On the
other hand, labor productivity growth has been strong, with output per worker
doubling in the past decade. In fact, GVA growth has been mainly driven by
labor productivity growth. Conversely, employment elasticities have fallen but
remain positive – suggesting that productivity growth was accompanied by
employment growth. Moreover, productivity gaps across economic sectors are
very large. While within-sector productivity growth accounts for much of the
aggregate productivity performance, structural change (i.e. between-sector
productivity growth) is playing an increasingly important role. Therefore, in
Ethiopia, economic growth has generated some employment, but the growth-
employment link has been weakened, which is intrinsically linked to
demographic trends. Not only has population growth slowed down, but the
proportion of people outside the labor force has also increased (Martins,
2014).
McMillan and Rodrik (2011) argue that the potential gains in overall
productivity growth from sectoral reallocations can be very large. Based on
their estimate, Ethiopia’s productivity could increase six-fold by assuming that
sectoral productivity levels remain constant and that the inter-sectoral
distribution of employment matches that of advanced economies. The
potential gain of Ethiopia is positive as compared with other countries in
Page | 75
Africa, only lagging behind Senegal and Malawi. The increases in labor
productivity can be accounted by several factors which include increased
efficiency in the use of labor, improved use of other inputs, and labor shifting
to more productive activities (structural change). McMillan and Rodrik (2011)
also observed a significant degree of heterogeneity across countries in the
structural change. Some countries in Africa experienced growth-enhancing
structural change, such as Ethiopia and Ghana. They find that in a sample of
38 developed and developing countries, Ethiopia has the second highest
contribution of structural change to productivity growth.
Strong economic growth may not always help in poverty reduction,
because of its failure to generate sufficient productive employment in terms of
better jobs. Martins (2013) also compares the fast-growing African countries
including Ethiopia, Ghana, Mozambique, and Tanzania. He compares the
different growth paths being pursued, and the policy choices, which might
explain the gaps in key development outcomes. According to him, the pattern
of economic growth matters for poverty reduction. In addition, from the
experience of Mozambique and Tanzania, growth that is driven by capital-
intensive sectors seems to generate limited benefits for the poor. So, to
improve inclusiveness of the growth, Martins (2013) suggests diversifying
production structures into more employment-intensive sectors. From the
experience of Ethiopia, sector specific policies play an important role in
reducing poverty. Particularly, investment in the agriculture is not only crucial
to reducing the incidence of poverty but can also accelerate the pace of
structural transformation. Therefore, according to Martins (2013),
improvement in agricultural productivity and employment in higher-
productivity employment-intensive activities, such as in the manufacturing
and modern services, will be important to sustainably raise living standards in
Africa.
According to Dorosh, Schmidt, & Shiferaw (2012), although
agriculture continues to be the foundation for economic growth and poverty
reduction, in Ethiopia government policy slowed rural-urban migration
through various regulations including prohibiting the sale of land, loss of land
Page | 76
rights for those who leave rural areas, and registration requirements for new
migrants. The modes rate of urbanization and slow development of industrial
sector inhibited structural transformation in Ethiopia. Therefore, allocation of
public investments across sectors along with policies and regulations on land
and labor mobility will be important factors for the growth of Ethiopia’s
economy and poverty reduction in the coming decade (Dorosh, Schmidt, &
Shiferaw, 2012).
An obvious change in the Ethiopian structural change is the high
growth in per capita GDP, rapid accumulation of both physical and human
capitals, and the declining contribution of the agricultural sector and rise of the
service sector. Moreover, the Ethiopian economy since 2005/06 has been
stable and resilient to drought, agricultural price and financial price shocks
implying the sustainability of the growth. (EEA, 2015). The per capita GDP of
Ethiopia in 2014/15 is 567 USD, which has increased from its low level in the
past decades. In terms of accumulation, Ethiopia is a typical poor country that
had an average rate of gross domestic saving of 5.8% during 2005/06 –
2009/10 and gross fixed investment of 26.3%, creating a resource gap of
20.9% that is financed by remittances, FDI, grants, and loans. However,
during the GTP period, the gross domestic saving rate increased to 22.5% and
the rate of gross fixed investment reached 40.3% in 2013/14 due to
government extensive resource mobilization efforts through Great Ethiopian
Renaissance Dam (GERD) bond and condominium housing schemes. The
source of Ethiopian growth from factor decomposition shows the labor has
been the major source of growth exceeding capital until 2004/05 (EEA, 2015).
Since 2004/05, however, capital is contributing more than labor and it has an
increasing trend. During the GTP I (2009/10 – 2013/14) period, the
contribution of capital and labor to the 10% growth was 6.6% and 5.7%
respectively. The residual (‘Solow’s residual), which is 2.8% is the
contribution of the TFP. The TFP to growth in Ethiopia before 2004/05 has
been erratic and mostly negative. However, after 2004/05 it is positive ranging
between 4.5 and 5.4%, contributing much to the high growth observed in the
period. From the demand side in 2010/11 – 2013/14, the decomposition of
Page | 77
GDP growth shows that expansion in the domestic demand driven by high
public investment contributes 85% of the growth as export and import
substitution accounted 5.2% and 3.4% respectively. From the service sub-
sectors, the wholesale and retail trade, the hotel and restaurant and the
business and real estate sub-sectors contribute 16.3%, 9.1% and 7.9% to the
10% GDP growth rate during 2009/10 – 2013/14 (EEA, 2015).
According to WB (2015) report on Ethiopian economy, the rapid
economic growth during the growth periods is concentrated in agriculture and
services and was driven by a considerable investment in the public
infrastructure, which is supported by a conducive external environment. At the
beginning of the take-off, the agriculture sector was the main economic sector.
But in recent years the services sector took over the role, complemented by a
construction boom. Out of an average annual growth rate of 10.9 percent in
2004-14, services contributed by 49.5 percent followed by agriculture 33
percent and industry with 15.6 percent (EEA, 2015). Private consumption
contributed to most growth on the demand side with public investment
becoming increasingly important.
Growth decompositions for Ethiopia reveal relatively high
contributions from total factor productivity and structural change. According
to WB(2015a), Ethiopia had high total factor productivity growth of 3.4
percent per year, which is greater than other non-resource rich fast-growing
Sub-Saharan African countries. Most of the labor productivity growth is due to
the within sectors productivity. The structural change also explains a quarter
of the per capita GDP growth. Using a cross-country regression model,
Ethiopia’s key drivers of economic growth in Ethiopia was driven primarily
by public infrastructure investment. Ethiopia has experienced high economic
growth and some structural change on the output side but from the
employment side, a shift from agriculture to the services is modest. Since
agricultural labor productivity is so small, the shift of labor out of agriculture
gave rise to static efficiency gains in construction and services that have
higher average value added of a worker (WB, 2015a).
Page | 78
For the 1999-2013 periods as a whole, more than 70% of growth is
attributed to within-sector labor productivity gains, especially in agriculture
and commerce. The structural change and demographic effects are noticeable
in (2005-13) period. The employment effect is negative due to a rise in the
student population. Between 1999 and 2013, labor productivity exhibited an
average annual growth rate of 4.5 percent. However, the nature of the
structural change and sectoral growth dynamics in Ethiopia differs from the
government policy and targets. Specifically, economic strategy in Ethiopia
aims to promote the kind of Lewis model-based structural change in which
workers move out of agriculture into manufacturing. World Bank has called
this model as ‘trodden path of development’ only partially as economic
activity (output and jobs) have shifted from agriculture and into construction
and services, bypassing the critical phase of industrialization (WB, 2015a).
Finally, Ferede & Kebede (2015)analyzed the structural change of
Ethiopia using decomposition tool. According to their findings, the within-
sector productivity represents the largest contribution to aggregate labor
productivity growth, accounting for about 66% of total labor productivity
growth. The structural effect, which consists of employment and the
interaction effects, contributed close to a third of the total labor productivity
growth, with the employment effect accounting for about 40% of the total
labor productivity growth. The contribution of the interaction effect to the
overall labor productivity growth is negative, supporting the structural burden
hypothesis. This indicates that sectors with fast-growing labor productivity
cannot maintain their shares in total employment. The negative effect can be
larger if sectors with high productivity growth are faced with declining
employment shares. While agriculture, manufacturing, and electricity and
water sectors exhibited declining employment shares with positive labor
productivity growth, transport and communication and other services
experienced the opposite.
Page | 79
considered as a path for structural change, sustainable growth and greater
interlinkage with the manufacturing and agriculture sectors. Theories and
empirics also show that the growth of service in output and employment is
mainly driven by income, productivity difference and changes in the structure
of demand including economic reforms, the participation of female in
economic activities and demand for intermediate service inputs.
Page | 80
CHAPTER-III
METHODOLOGY, DATA AND MODEL
Page | 81
3.1 : Growth Decomposition Analysis
𝑌 𝑌 𝐸 𝐴
= 𝐸 . 𝐴 . 𝑁; (1)
𝑁
𝑌
Where 𝐸 is total output per worker, as a measure of productivity,
𝐸
is employment rate and
𝐴
𝐴
is the share of working-age population measuring demographic changes.
𝑁
Thus, per capita GDP can be decomposed into growth associated with
changes in the size of the working age population. This means that total
change in per capita GDP is the sum of growth attributed to each of the
components 𝜔, 𝑒 𝑎𝑛𝑑 𝑎; each representing output per worker, employment
̅, 𝑒̅ and 𝑎̅as
rate and working-age population respectively. Thus, if we denote 𝜔
Page | 82
a fraction of the growth attributed to each component of 𝜔, 𝑒 𝑎𝑛𝑑 𝑎, then the
total growth rate of an economy can be expressed as;
∆𝑦 = 𝜔
̅ ∗ ∆𝑦 + 𝑒̅ ∗ ∆𝑦 + 𝑎̅ ∗ ∆𝑦 (2)
𝐸𝑖
Where ∆𝑒𝑖 = ∆ 𝐴 - is the changes in employment in sector i as a share of total
Data
Page | 83
National Labour Force Survey of the Central Statistical Agency of Ethiopia.
The analysis is made for the period (1999-2013) by classifying the period as
the low growth period of 1999-2005 and the high growth period of 2005-2013
(See (WB, 2015a). Another sort of secondary data for the descriptive analysis
is collected from the Groningen Growth and Development Center (GGDC) 10
sectors database for the periods 1961-2011 (de Vries, de Vries, Gouma, Pahl,
& Timmer, 2014).
Page | 84
(Timmer & Szirmai, 2000; de Vries, Timmer, & de Vries, 2013) to decompose
the change in aggregate productivity. The other alternative is using final
period employment shares and base period productivity levels. On the other
hand, Timmer & de Vries (2009) used period averages with the average share
of sectors in overall employment, and the average labor productivity level of
sectors. This decomposition takes a middle ground between the base and end
year choices with respect to the contribution of structural change. It can be
represented as:
Page | 85
term in the decomposition that can be written as follows: (see (Timmer & de
Vries, 2009; de Vries, Timmer, & de Vries, 2013)
Page | 86
ones based on their changes in employment shares and calculate the static
between-effect relative to the average productivity level of the shrinking
sectors and the dynamic between- effects relative to the average productivity
change of the shrinking sectors. The decomposition is modified as follows:
∆𝑃 = ∑𝐽𝑖(𝑃𝑖𝑇 − 𝑃𝑖𝑜 )𝑆̅𝑖 + ∑𝐽𝑖(𝑆𝑖𝑇 − 𝑆𝑖𝑜 )(𝑃𝑗𝑜 − 𝑝𝑜∗ ) + ∑𝐽𝑗 ((𝑃𝑗𝑇 − 𝑃𝑗𝑜 ) − (𝑝𝑇∗ − 𝑃 𝑜∗ )) (𝑆𝑗𝑇 − 𝑆𝑗𝑜 ) (7)
Where J is the set of expanding sectors and K is the set of shrinking sectors
and average labor productivity of shrinking sectors at time O and T is given
by,
∑𝐾 𝑇 𝑂 𝑂
𝑘 (𝑆𝑘 − 𝑆𝐾 )𝑃𝑘
𝑝𝑂∗ = ⁄ 𝐾 𝑇 (8)
∑𝑘 (𝑆𝑘 − 𝑆𝐾𝑂 )
∑𝐾 𝑇 𝑂 𝑇
𝑘 (𝑆𝑘 − 𝑆𝐾 )𝑃𝑘
𝑝𝑇∗ = ⁄ 𝐾 𝑇 (9)
∑𝑘 (𝑆𝑘 − 𝑆𝐾𝑂 )
This adjusted decomposition does not affect the aggregate contributions from
the within and between effects. However, equation (7) is better to examine the
contribution of sectors in accounting for productivity growth.
Page | 87
static effect, will be positive when the sectors with high productivity levels
attract more labor. It may also have a negative impact on aggregate
productivity growth if labor shifts to sectors with slower productivity growth.
On the other hand, according to Baumol (1967), the structural burden
hypothesis is valid if the value of this is negative due to a shift of labor from
manufacturing to services that have less productivity. Thus, the interaction
term can be used to capture Baumol’s hypothesis of a structural burden of
labor reallocation on aggregate growth, which predicts that employment shares
shift away from progressive industries towards those with lower growth of
labor productivity. See (Peneder, 2003).
̅𝑖 ≥ 0
∑𝑖(𝑆𝑖𝑇 − 𝑆𝑖𝑜 )𝑃 (10)
Page | 88
According to Van Ark & Timmer (2003), the methodology is ‘a labor
productivity model and as such non-labour inputs are ignored’. Economic
activities are not solely using labor inputs only, other inputs such as capital,
materials, and service inputs are used along with labor. Therefore, the sectoral
productivity measures should incorporate non-labour inputs. The other
criticism is that the decomposition into shift effects and intra-sectoral effects
depend on the selected price base year of the output series.
Data
The data for the shift-share analysis in this study is collected from
Groningen Growth and Development Center (GGDC) 10 sectors database (de
Vries, de Vries, Gouma, Pahl, & Timmer, 2014). The database covers the ten
main sectors of the economy as defined by the International Standard
Industrial Classification, Revision 3.1 (ISIC, 2002). These ten sectors are;
agriculture, manufacturing, other industry (includes construction and utilities,
distributive trade services (includes wholesale and retail trades, hotels and
restaurants), transport, storage and communication services, finance and
business services, public services (public administration and social services),
and other services (includes other community, personal and household
services). The employment data refers to ‘all persons employed’, including all
paid employees, also self-employed and family workers. The Gross value
added is measured at constant 2005 Ethiopian prices.
The ARDL approach can be applied whether the regressors are I(1)
and I(0) (Pesaran, Shin, & Smith, 2001). The Johansen co-integration
Page | 89
techniques require large data samples for validity, but the ARDL procedure
avoids the problem of biases that arise from small sample size and gives a
statistically significant result (Pesaran, Shin, & Smith, 2001); (Narayan, 2005).
In addition, a dummy variable can be included in the co-integration test
process when using the ARDL model. Therefore, the following ARDL model
is specified to identify the long run and short run relationships of service gross
value addition and service employment growths with other explanatory
variables.
Page | 90
The variables in (𝑦𝑡 ,𝑥𝑖,𝑡 ) are allowed to be purely I(0), purely I(1), or
co-integrated. The optimal lag orders p and q can be obtained by minimizing a
model selection criterion of the Akaike information criterion (AIC) or the
Schwarz Bayesian information criterion (BIC) (Kripfganz & Schneider, 2016).
Based on the theoretical framework described in the literature section, the
following log-linear type model is specified. The ‘hierarchy of needs’
hypothesis, the ‘cost disease’ hypothesis and the ‘exogenous demand shock’
hypothesis are included. The ‘deindustrialization’ hypothesis is not considered
as it is irrelevant to Ethiopia.
All the ratio variables are log transformed to reduce the problem of
heteroscedasticity as it compresses the measured scale of the variables
reducing the difference between two values (Gujarati, 2004). Thus, in the
model, all parameter coefficients represent constant elasticities.
Page | 91
Where
LnSGVAt is Natural Logarithm of Service in Gross Value Addition at time t
LnSEMPt is Natural Logarithm of Service Employment at time t
LnGDPPCt is Natural Logarithm of GDP per Capita at time t
LnPTVDt is Natural Logarithm of Productivity difference at time t
LnPSERVt is Natural Logarithm of Producers service at time t
LnFEACTt is Natural Logarithm of Female Economic Activity at time t
DREFO92 is Dummy variable for Reform Period (1992-2011)
Based on the above equations the functions becomes in the form of:
𝑞 𝑞
𝐿𝑛𝑆𝐺𝑉𝐴𝑡 = 𝛼0 + ∑𝑃𝑖=0 𝛽1 𝐿𝑛𝑆𝐺𝑉𝐴𝑡−1−𝑖 + ∑𝑖=0 𝛽2 𝐿𝑛𝐺𝐷𝑃𝑃𝐶𝑡−𝑖 + ∑𝑖=0 𝛽3 𝐿𝑛𝑃𝑇𝑉𝐷𝑡−𝑖 +
𝑞 𝑞
𝐿𝑛𝑆𝐸𝑀𝑃𝑡 = 𝛼0 + ∑𝑃𝑖=0 𝛽1 𝐿𝑛𝑆𝐸𝑀𝑃𝑡−1−𝑖 + ∑𝑖=0 𝛽2 𝐿𝑛𝐺𝐷𝑃𝑃𝐶𝑡−𝑖 + ∑𝑖=0 𝛽3 𝐿𝑛𝑃𝑇𝑉𝐷𝑡−𝑖 +
𝛾𝐸𝐶𝑀𝑡−1 + 𝜇𝑡 (17)
𝑞 𝑞
𝑑𝐿𝑛𝑆𝐸𝑀𝑃𝑡 = 𝛼0 + ∑𝑃𝑖=0 𝛽1 𝑑𝐿𝑛𝑆𝐸𝑀𝑃𝑡−𝑖 + ∑𝑖=0 𝛽2 𝑑𝐿𝑛𝐺𝐷𝑃𝑃𝐶𝑡−𝑖 + ∑𝑖=0 𝛽3 𝑑𝐿𝑛𝑃𝑇𝑉𝐷𝑡−𝑖 +
∑𝑞𝑖=0 𝛽4 𝑑𝐿𝑛𝑃𝑆𝐸𝑅𝑉𝑡−𝑖 + ∑𝑞𝑖=0 𝛽5 𝑑𝐿𝑛𝐹𝐸𝐴𝐶𝑇𝑡−𝑖 + ∑𝑞𝑖=0 𝛽6 𝑑𝐷𝑅𝐸𝐹092𝑡−𝑖 + 𝜆1 𝐿𝑛𝑆𝐸𝑀𝑃𝑡−1 +
𝜆2 𝐿𝑛𝐺𝐷𝑃𝑃𝐶𝑡−1 + 𝜆3 𝐿𝑛𝑃𝑇𝑉𝐷𝑡−1 + 𝜆4 𝐿𝑛𝑃𝑆𝐸𝑅𝑉𝑡−1 + 𝜆5 𝐿𝑛𝐹𝐸𝐴𝐶𝑇𝑡−1 + 𝜆6 𝐷𝑅𝐸𝐹092𝑡−1 +
𝛾𝐸𝐶𝑀𝑡−1 + 𝜇𝑡 (18)
Page | 92
In addition, the degree of stationarity is determined by carrying out a
unit root testing through the standard Augmented Dicky-Fuller (ADF) test.
This test was undertaken to check the order of integration of the variables. It
was done for three alternative specifications. First, it is tested with constant,
trend, and then it is tested with neither constant nor trend. The null hypothesis
to test the stationarity is that the time series data have a unit root and the
alternative hypothesis is it has no unit roots. Therefore, if the critical statistic
in absolute value is greater than the t-value, we reject the null hypothesis and
if it is less we accept the null hypothesis.
Page | 93
model to check the robustness of the model. These tests are model
specification test, normality test, serial correlation test, heteroscedasticity test
and test for stability of the model. Stata 13 statistical package and the ARDL
model for Stata (Kripfganz & Schneider, 2016)is used in order to estimate the
models specified in equation 15 and equation 16 and to perform the pre-
estimation and post estimation diagnostic tests.
Data
The data for this study is collected from secondary sources from 1981
– 2011. The service sector output measured in Gross Value Addition (GVA),
Service employment (number of the labour force employed in the service
sector), the output level of producer services are collected from GGDC
(Groningen Growth and Development Centre) 10 sector Database (de Vries,
de Vries, Gouma, Pahl, & Timmer, 2014). The Productivity difference
between manufacturing and service sector is also computed using the same
database. Furthermore, female economic activity data is also collected from
GGDC database. The GVA output of service sector and producer services
output are measured in constant 2005 local currency unit (Ethiopian Birr). The
sectors are classified on the basis of the International Standard Industrial
Classification of All Economic Activities, Rev.3.1(ISIC, 2002) code. Thus, the
service sectors aggregated in this study includes distributive trade services,
transport and communication services, finance and business services, public
services and personal services. Finally, the data on GDP per capita income is
measured in constant 2005 USD and it is collected from WDI (2016).
Page | 94
production activities, 61 commodity groups, 5 primary factors, 2 household
groups, 17 tax instruments as well as aggregate accounts for trade margins,
transport margins, government, investment, and the rest of the world (Taffesse
& Ferede, 2005). The 2005/06 Ethiopian SAM is the most recent SAM and I-
O table constructed for the Ethiopian financial year of 1998 (2005/06) and it is
completed in 2009. Thus, the 2005/06 Ethiopian SAM is used to extract the I-
O data for the analysis of sectoral inter-linkage using SimSIP SAM software.
Page | 95
Table 3.1: Basic Structure of SAM
Expenditure
Saving & Rest of the
Activities Commodities Factors Households Government Total
Investment World
Domestic Activity
Activities
Supply income
Intermediate Consumption Recurrent Investment Export Total
Commodities
Demand Spending Spending Demand Earnings Demand
Total Factor
Factors Value Added
income
Factor Total
Social Foreign
Households payments to Households
Transfers remittances
Households income
Income
Page | 96
On the other hand, commodities can be either supplied domestically or
imported from the rest of the world and various economic entities purchase
commodities. Activities demand commodities as intermediate inputs for
production. A final demand for commodities also comes from household
consumption, government consumption, gross capital formation, and export
demand. These sources of demand make up the commodity row or the total
demand, which are payments by different entities for commodities. Therefore,
the values in the commodity accounts are measured at market prices. SAM
does not only trace the income and expenditure flow of activities and
commodities. It also contains complete information on different institutional
accounts including households and the government (Breisinger, Thomas, &
Thurlow, 2009). As households are the ultimate owners of the factors of
production, they receive the incomes earned by factors during the production
process. They also receive transfer payments from the government and from
the rest of the world, pay taxes directly to the government, and purchase
commodities. The household income that remained after payment of tax and
consumption expenditure can be saved. On the other hand, in the government
account, a government may receive transfer payments from the rest of the
world as foreign grants or/and development assistance. All the different tax
incomes and foreign grants and loans add up to determine the total
government revenues. It uses these revenues to pay for recurrent consumption
spending and transfers to households. There could be a fiscal surplus or deficit
based on the difference between total revenues and expenditures.
Furthermore, SAM shows the investment or gross capital formation account,
which includes changes in stocks or inventories. The total investment must be
equal to total savings (private savings and public savings). The total capital
inflows from abroad or the current account balance is the difference between
total domestic savings and total investment demand. This is equal to the
difference between foreign exchange receipts (from exports and foreign
transfers) and expenditures (for imports and government) (Breisinger,
Thomas, & Thurlow, 2009).
Page | 97
3.3.2 : Input-OutputModel Framework
SimSIP SAM stands for Simulations for Social Indicators and Poverty
through Social Accounting Matrix. The software is based on a ‘Microsoft
Excel application with MATLAB running in the background’, which is used to
analyze I-O tables and SAM (Parra & Wodon, 2010a). It was developed in
2008 by J. C. Parra and Q. Wodon of the World Bank and version 1.1 became
available in 2009 and the latest version 1.1.2 is available since 2010. The
application is useful to perform several types of analysis such as aggregation,
disaggregation (decomposition), updating and balancing of the I-O tables to
obtain descriptive and analytical results of an economy. In addition, the tool is
useful for interlinkage analysis, multiplier analysis, and different experimental
designs. Various studies have used the software including Parra and Wodon
(2009), (Parra & Wodon, 2010b) for Tanzania; (Fofana, Parra, & Wodon,
Page | 98
2009) for Senegal; and (Khanal, 2011; Khanal, Gan, & Becken, 2014) for Lao
Tourism analysis.
The initial step in using the I-O model with SimSIP SAM software is
to identify the endogenous and exogenous accounts of the economy.
According to Sadoulet and de Janvry (2003) as cited in Parra & Wodon
(2010a), ‘exogenous accounts are those for which the expenditure is set
independently of income and endogenous accounts are those that change the
level of expenditure following any change in income’. Therefore, the software
applications consider the government, capital and rest of the world (exports
and imports) accounts as exogenous accounts and all other accounts as
endogenous. In addition, each cell in the I-O table is expressed in price or
value terms (Parra & Wodon, 2010a). The SimSIP SAM software application
is divided into several analysis components/blocks suitable for I-O and SAM
analyses out of which four analysis blocks were used in this study. These are
the matrix design and balance, GDP and labor multiplier, multiplier analysis
and sectoral impacts.
Page | 99
during 2005/06 was graphed based on the imposed linkage hierarchies from
the Multiplier Product matrix of the I-O (See Annex).
For the purpose of this study, the Ethiopian I-O inter-industry flows are
extracted from the 2005/06 SAM and the economy is divided into 10 sectors.
The classification is based on ISIC Rev. 3.1. These sectors are agriculture,
manufacturing, mining and quarrying, utility, construction, distributive
service, transport and communication, business and finance, public services,
and personal and other services. The values for each cell is measured in
Ethiopian Birr (1USD = 8.681 ETB on average in 2005/06).
In Table 2, the columns of the I-O table are the selling sectors and the
rows are the purchasing sectors of the economy.According to Miller & Blair
(2009), the inter-industry sale shows that the total sales of each sector output
to itself, to another sector as intermediate inputs, and to the final demanders.
This flow of product depends on the total output of the sectors and I-O
Page | 100
coefficients. The I-O coefficients can be derived by dividing the inter-industry
flows of each sector by the total output of that sector.
Capital r1 r2 … r10 R
Gross Domestic Product
Indirect Taxes t1 t2 … t3 T
Import m1 m2 … m10 M
Total Value Added V1 V2 … V10
Total Inputs X1 X2 … X10 C G I E X
Source; adopted from (Khanal, 2011) and (EDRI, 2009) I – Investment, C – Households
Consumption, E – Exports and G - Government, W – wage, R – return to capital, T – Tax, and
M - import
Apart from the selling and buying sectors (commodities), there is value
addition block which is used in the production process. It includes labor;
capital and different inputs and these factors are the value added block of the
I-O table. They are compensation for employees as a wage for the total 10
sectors (W10); capital (R10); taxes (T10); and imports (M10). Therefore, the
total expenditures on inputs constitute the sum of all domestically produced
and imported goods and services and value added inputs. Thus, total inputs
for each sector (represented as𝑋1 , 𝑋2 and 𝑋10 ) is the sum of inter-industry sale
(intermediate inputs) and value added inputs. It is given as;
Page | 101
purchases of each sector’s outputs. The total final demand, each sector’s
outputs purchased as a final demand is given as;
One of the features of I-O table is the balancing of the sum of the total
inputs (row totals) with the sum of the total outputs (column totals). Hence,
the total output column which is the total gross output (X) in each sector is
𝑋 = 𝑋1 + 𝑋2 + ⋯ + 𝑋10 + 𝑊 + 𝑅 + 𝑇 + 𝑀, Similarly, the total input (X) in each
sector is given as 𝑋 = 𝑋1 + 𝑋2 + ⋯ + 𝑋10 + 𝐶 + 𝐺 + 𝐼 + 𝐸 . Therefore,
combining the two equations, we get; (see (EDRI, 2009)).
𝑊 + 𝑅 + 𝑇 = 𝐶 + 𝐺 + 𝐼 + (𝐸 − 𝑀) (22)
Page | 102
Chenery and Watanabe (1958) and Hirschman (1958), are extensively used to
make a comparison of different sectors of an economy in terms of the
interdependence of their production structures (Pratt, 2011).However, the I-O
model is highly useful to show the inter-industry linkages of sectors. I applied
I-O methodology is applied in the study to explore about service sector in
Ethiopia specifically to find out the level of its inter-linkage with other sectors
and its place as a key sector in the economy.
(23)
Where 𝑎𝑝𝑗 is the amount of imported input required per unit of sector j. Now,
given a final demand vector for the three sectors output (F) production and
value added in the three sectors are immediately obtained from equation (23)
and (24) as equation (25)
[𝐼 − 𝐴]𝑋 = 𝐹 (25)
𝑥1
Where, 𝑋 = [𝑥2]gives the column vector of the three sectors’ gross output x1,
𝑥3
x2, and x3 and represents the three sectors final demand, f1 f2 and f3
Page | 103
Multiplying both sides of equation (25) by [𝐼 − 𝐴]−1 we get output of the three
sectors X as a function of F equation (26)
(26)
(27)
Now, the aggregate income of the economy Y is the sum of sectoral incomes
as in equation (28)
𝑌 = ∑3𝑖=1 𝑦𝑖 (28)
Equations (27) and (28) indicate how aggregate and sectoral incomes of the
economy are given by the composition of final demand and the economy’s
input- output coefficient matrix (Chenery & Watanabe, 1958; Miller & Blair,
2009). In the I-O framework, the measurement of linkages has been made
based on the Leontief production matrix (the Matrix A) or the Leontief inverse
matrix [𝐼 − 𝐴]−1
Backward Linkages
Page | 104
with those (“upstream”) sectors from which it purchases inputs’ (Miller &
Blair, 2009, p.555). Thus, the measure of the strength of the backward linkage
of sector j (the amount by which sector j production depends on inter-industry
inputs) is given by the sum of the elements in the jth column of the direct input
coefficients matrix, namely 𝑎𝑖𝑗. This is the direct backward linkage since the
coefficients are measures of direct effects only (Miller & Blair, 2009, p.556).
∑𝑖 𝑋𝑖𝑗
𝐵𝐿𝑖 = = ∑𝑖 𝑎𝑖𝑗 (29)
𝑋𝑗
Page | 105
With ∑𝑖 𝑙𝑖𝑗 as the column sums of the Leontief inverse, sectors with
above average backward linkage reveal indices greater than one, indicating a
strong integration with the rest of the economy. Those with ‘below average’ or
weaker direct backward linkage will have indices lower than one (Miller and
Blair 2009).
Forward Linkages
On the other hand, for the forward linkage ‘the direction of causation is
in the supply-side model indicating ‘interconnection of a particular sector with
those (“downstream”) sectors to which it sells its output’(Miller & Blair, 2009,
p.555).
∑𝑗 𝑋𝑖𝑗 ∑ 𝑋
𝐹𝐿𝑖 = = ∑𝑛 ∑𝑗 𝑛𝑖𝑗𝑋 (32)
𝑋𝑖 𝑗 𝑗 𝑖𝑗
Page | 106
1
∑ 𝑔
𝑛 𝑗 𝑖𝑗
∑𝑖 𝑈𝑖𝑗 = 1 (34)
∑ 𝑔
𝑛2 𝑖 𝑖𝑗
According to (Miller & Blair, 2009, p.559), based on the backward and
forward linkages sectors can be distributed in one over a four-way
classification.
Page | 107
CHAPTER IV
STRUCTURAL CHANGE AND PRODUCTIVITY
IN ETHIOPIAN ECONOMY
The broad objective of the study is to analyze the structure and the
growth of the Ethiopian service sector. Service sector pattern and growth is
related to the structural change in an economy. Structural change refers to
change in sectoral shares and employment patterns. Structural change is also
manifested by a change in productivity, both at the aggregate and
disaggregates level. In this context, the present chapter explores the broad
structure and structural change in Ethiopian economy and its manifestation in
employment structure. Sector-wise analysis of productivity forms the next
portion of this chapter. Then, the last sub-section assesses the role of the
service sector in structural change using the shift-share analysis.
Page | 108
percent in 1960 to 44 percent in 2011. The manufacturing sector’s output
share has changed slightly from just 2 percent to 5.2 percent within fifty years
(1961-2011). This sectoral structure of the country makes Ethiopia unique
when it is compared with other countries that had the same level of per capita
income at various periods. As it is shown in Table 4.1, in 1961 Ethiopia had
the highest share of output in the agriculture sector (85 percent), which is
incomparable with the other countries' agricultural output share. The service
sector share was also one of the lowest, accounting just only 9 percent of total
output. Furthermore, the manufacturing sector had also share of only 2 percent
of the total output. These figures indicate the country’s level of development,
which is at the initial stage of development dominated by agriculture and
underdeveloped modern sector. In 2011, after 50 years, the Ethiopian
agriculture sector share to total output is still one of the highest (42 percent)
when compared with other countries output share. The manufacturing sector
output share just increased from 2 percent to 5 percent, which is lower than the
SSA average and other developing countries. However, what has changed
significantly is the share of the service sector. It increased from 9 percent to 44
percent, becoming the largest sector in output share and comparable with the
level of service output in other developing countries.
1961 1981 2001 2011 1961 1981 2001 2011 1961 1981 2001 2011
Agriculture 51 18 26 19 85 69 50 42 37 41 33 26
Industry 15 33 21 22 3 5 7 9 20 12 10 19
Manufacturing 6 11 10 9 2 4 5 5 14 13 11 9
Service 29 37 44 50 9 21 38 44 29 34 46 46
China India South Korea
Agriculture 62 40 16 9 51 39 24 16 21 11 4 3
Industry 4 12 13 15 10 13 12 13 5 11 12 10
Manufacturing 6 20 35 37 12 15 17 18 4 18 29 35
Service 28 27 37 40 27 34 47 53 70 60 54 51
Source; Author’s computation using GGDC Database(de Vries, de Vries, Gouma, Pahl, &
Timmer, 2014)
*SSA – Sub-Saharan Africa average of 11 countries. These are Ghana, Kenya, South Africa,
Nigeria, Tanzania, Malawi, Botswana, Mauritius, Ethiopia, Senegal and Zambia
Page | 109
The service sector is an important leading sector in Ethiopian economy
by contributing the largest share to the recent economic growth. As it is
depicted in Table 4.2, the sector had the highest absolute growth rate than
other sectors up until 2010/11. After 2010/11, though the sector continued
with its highest contribution to GDP growth, its growth rate was below the
industry. This could be because of the government attention to the industry
sector under the first GTP (2010/11- 2014/15). Moreover, the contribution of
industry sector to the GDP growth for the first time exceeded the agriculture
sector in 2014/15. During the period, the service sector contributed 46.1
percent as the industry sector become the second important sector by
contributing 29.4 percent to the growth in GDP; while the agriculture sector
contributed 24.5 percent.
Page | 110
in atlas method. The country is far from the lower middle-income group lower
limit of 1045 USD per capita income. Ethiopian GNI has declined from 280 in
1988 to 110 USD in 2004. During these years Ethiopia was within its longest
civil war which ended in 1991. In 1998, the country once again entered into a
war with the neighboring Eritrea, which was separated from Ethiopia in 1992.
Other than wars, the Ethiopian economy was contained by the recurrent
droughts. It was after the period (2004-2014) that Ethiopia started to grow by
an average of 10.2 percent. This growth of the GDP is accompanied by an
increase in GNI from 110 USD in 2004 to 550 USD in 2014. Nevertheless,
still, Ethiopia is one of the low-income countries in the world ranking 10th
from the bottom. Ethiopia lags behind countries that are in the low-income
group in terms of its sectoral structures. The sectoral GDP shares and
employment share of sectors in Ethiopia are different from other countries.
Page | 111
share in the manufacturing sector before 50 years. After 50 years their
manufacturing sector increased by 30 percentage points but the Ethiopian
manufacturing sector increased just only by 3 percentage points. Most of the
SSA countries including Ghana experienced a decline in the manufacturing
sector, which is explained by (Rodrik, 2014) as ‘Premature
deindustrialization’. However, what is observed in Ethiopia is the ‘stagnation
of manufacturing’ sector.
400
.8
.6
300
.4
200
.2
100
0
Agriculture Industry
Manufacturing Service
GDP Per Capita
The Ethiopian agriculture and service sectors’ GDP share has changed
by larger percentage points as compared to other countries. Agriculture
declined by 43 percentage points and service increased by 35 percentage
points from 1960 to 2010. The change in the percentage share of the
manufacturing and other non-manufacturing industry sectors is very small.
These changes in the sectoral structure of the Ethiopian economy are an
important indicator that there is some sort of sectoral growth dynamics at a
very low level of GDP per capita [See Figure 4.1]. However, there are no
significant changes occurred in the Ethiopian sectoral structure and GDP per
capita within 20 years. Countries that were at the similar level of GDP per
capita and economic structure with Ethiopia are changed. Figure 4.2 below
shows the sectoral GDP share and GDP per capita of Ethiopia and other
countries.
Page | 112
Based on the data from World Development Indicator for 200
countries, a fitted regression line is computed which shows the relationship
between the share of GDP or GVA in sectors and the natural logarithm of
GDP per capita in 1994 for Ethiopia and 2013 for the rest of the countries. As
it is depicted in Figure 4.2, Ethiopia has agriculture sector whose share is far
greater than countries with similar level of income making the country one of
the most agrarian countries in the world. With the increase in the GDP per
capita, the share of agriculture in GDP has declined in 2013 as compared to
the 1994 level. In 2013, similar to the structure in 1994, the agriculture sector
is exceptionally one of the highest in its GDP share. What makes the Ethiopian
case different is that the share of agriculture in GDP is still high, even though
it reduced by 10 percentage point from 55 percent in 1994 to 44 percent in
2013. Based on the fitted regression line, the agriculture sector share in
Ethiopia should have been around 40 percent in 1994 and 30 percent in 2013,
which is the average level at the respective income level in the periods. Even
the lower middle-income countries have agriculture GDP share around 26
percent in 1994 and around 17 percent in 2013.
Figure 4.2: Relationship between Sectoral output shares in GVA and
GDPC (2013)
Agriculture GVA share Vs GDP per capita, 2013 Industry GVA share Vs GDP per capita, 2013
100
60
ETH_1994
Agriculture GVA share (%)
20 40 60 80
ETH_2013
40
CHN
BWA KOR
20
IND
LMI LMI
IND
SSA
SSA
CHN ETH_2013
ETH_1994
BWA KOR
0
4 6 8 10 12 4 6 8 10 12
Natural Logarithm of GDP per capita (Constant 2005 USD) Natural Logarithm of GDP per capita (Constant 2005 USD)
Manufacturing GVA Vs GDP per capita, 2013 Service GVA share (%)
100
10 20 30 40 50
Manufacturing GVA share (%)
20 40 60 80
4 6 8 10 12 4 6 8 10 12
Natural Logarithm of GDP per capita (Constant 2005 USD) Natural Logarithm of GDP per capita (Constant 2005 USD)
Page | 113
On the other hand, Ethiopia also had one of the lowest GDP shares in
the industry and manufacturing sectors. The manufacturing sector share in
Ethiopia is lower than the fitted regression line implying that Ethiopia is far
below from its peer countries that are in the similar level of income in its
manufacturing development. With the change in GDP per capita, the change in
the GDP sectoral share of the manufacturing sector within 19 years is
insignificant. There was only a 1.1 percentage point change from 4.5 percent
in 1994 to 5.6 percent in 2013. Accordingly, given the income level of
Ethiopia during 1994 and 2013, the share of GDP in the manufacturing sector
should have been around 10 percent of its GDP share on the fitted regression
line. In terms of the industry share in GDP Ethiopia is again at the lower level.
Ethiopia also lags behind some low-income countries that have increased their
GDP per capita by utilizing their natural resources. In Ethiopia, the share of
the industry sector that constitutes the mineral, construction and utility sub-
sectors has also small share in GDP. It is also one of the lowest as compared to
countries with same income level with Ethiopia. In 1994 the share of the
industry sector in GDP was 8.6 percent and it increased just only by 3.3
percentage point to 11.9 percent in 2013.
Nevertheless, it is only the service sector of Ethiopia that did not
deviate from the fitted regression line in both 1994 and 2013. The Ethiopian
service sector growth possesses a similar growth pattern with the world
average for respective level of incomes. In 1994, Ethiopia had service GDP
share of 35.6 percent which was more than Ghana and China. In 2013, it
increased to 43 percent, but it was lower than the share in Ghana and China.
Therefore, from the sectoral GDP shares data, we can infer that
Ethiopia is just only in the initial stage of structural change, in which the
country lags behind the world in its sectoral structure and income level.
However, the service sector has shown the relatively similar level of growth
which maintained the average service growth path of countries at the
respective income levels. The other general issue that can be observed from
the figure is that the slope of the fitted regression line for the service sector is
steeper, which implies the share of services in output increases with income.
Page | 114
That means high-income countries have a higher share of service in their total
output than developing countries. In addition, the share of the service sector is
more than 40 percent for most of the countries. This implies that countries are
becoming more and more dominated by the service sector even at the lower
level of income. On the other hand, the regression line for the manufacturing
and industry sectors increases first with the increase in income up to a certain
level. However, after reaching the maximum level, which is not more than 15
percent for manufacturing and not more than 30 percent for the industry, their
share in total output declines. This may imply that the demand for
manufactured and industrial goods gets concentrated and decline after a
country reaches certain high level of income. In other words, there would be
deindustrialization for high-income countries. This type of relationship
between the manufacturing sector and income is also observed using time-
series and cross-sectional data. For agriculture, as income grows the share in
total output declines fast. Thus, most of the countries in the world have
agricultural share below 20 percent accounting very small amount of the total
production.
Page | 115
service employment lower than 20 percent. The share of employment in the
manufacturing sector in Ethiopia was much lower in the early periods as
compared to SSA average. Ethiopia’s labor employed in manufacturing has
increased from 1 percent in 1961 to 7 percent in 2011 as the SSA average
increased from 4 percent to 7 percent for the same period. In Ethiopia, the
sector employed 119 thousand people in 1960 and it increased to 2.8 million in
2011. Whereas, the service sector employed more than 218 thousand labor in
1960 and it increases to 6.7 million in 2011. This indicates that the shift of
employment is observed in the service sector more than the manufacturing
sector. As the Lewis-type of the dual economy model suggests in the
conventional growth model, workers that move out of subsistence or
traditional agriculture are absorbed in the modern manufacturing sector
(Lewis, 1954). However, in Ethiopia, the sectoral shift of the economy from
agriculture to service sector is opposite to the conventional Lewis model and
the structural change approach.
Table 4.3: Employment share for Ethiopia and Comparator Countries,
1961-2011
SSA average Ethiopia Ghana
Sector 1961 1981 2001 2011 1961 1981 2001 2011 1961 1981 2001 2011
Agriculture 77 61 65 58 96 89 85 73 60 58 53 39
Industry 4 4 3 4 0 0 1 3 6 3 5 6
Manufacturing 4 6 5 7 1 2 3 7 11 12 11 11
Service 15 29 27 31 2 9 11 17 23 27 32 44
China India South Korea
Agriculture 77 68 50 37 72 72 59 55 62 33 10 7
Industry 3 4 8 10 2 2 6 8 4 7 8 8
Manufacturing 8 14 14 19 10 9 12 12 8 21 20 18
Service 12 14 28 35 16 17 24 26 26 39 63 67
Page | 116
of 1994 and 2010, though it declined from 89 percent in 1994 to 73 percent in
2010. Employment in the industry sector is lower than the service sector for all
the countries observed as most of the share in employment in the sector is
between 20-40 percent. Employment share in the industry sector also has
shown a decline in 2010 because of the deindustrialization in the high-income
countries and premature deindustrialization in the low-income countries of
Africa, Asia, and Latin America.
Agriculture Employment share Vs GDP per capita, 2013 Industry Employment Share Vs GDP Per capita, 2013
Agriculture Emloyment Share (%)
100
60
Industry Employment Share (%)
ETH_1994
80
ETH_2013
40
60
IND CHN
LMI
40
CHN IND
20
LMI
KOR
20
ETH_2013
KOR ETH_1994
0
4 6 8 10 12 4 6 8 10 12
Natural Logarithm of GDP per capita (in constant 2005 USD) Natural Logarithm of GDP per capita (in constant 2005 USD)
80
KOR
60
40
LMI CHN
IND
20
ETH_2013
ETH_1994
0
4 6 8 10 12
Natural Logarithm of GDP per capita (in constant 2005 USD)
Note: The Scatter points shows other countries and the line is fitted regression line
Source: World Bank, WDI (2016)
CHN – China, ETH – Ethiopia, IND – India, KOR – Korea, LMI – Lower Middle-Income
countries, SSA – Sub- Saharan African Countries
Page | 117
employment as compared to low-income countries. However, the employment
share in the industry sector first increases as income increases, and it is
saturated at a certain level of income and later it tends to decline.
Page | 118
Table 4.4: Labour Productivity level of Ethiopia and Comparator Countries, 1961-2011
Sector SSA Average Ethiopia Ghana
1961 1981 2001 2011 1961 1981 2001 2011 1961 1981 2001 2011
Agriculture 73.3 114.2 161.7 116.4 2.6 2.3 1.6 2.4 0.9 0.7 1.0 1.4
Industry 291.1 1740.4 1538.3 593.6 10.9 14.0 7.7 5.8 3.2 1.8 2.1 3.2
Manufacturing 156.5 722.9 727.6 762.0 4.8 7.1 4.4 3.2 2.0 1.1 1.6 1.6
Service 209.6 497.3 650.4 268.1 12.1 7.0 9.2 11.3 2.0 1.4 2.3 2.1
Total 111.4 383.6 405.7 214.7 3.0 2.9 2.7 4.2 1.6 1.1 1.5 2.0
Agriculture 1.6 2.5 5.2 9.9 2038.6 3752.5 11766.9 17019.1 16.5 16.8 23.4 29.9
Industry 1.8 7.5 35.0 71.1 4569.0 11642.8 39981.1 58148.5 42.7 75.6 94.8 161.5
Manufacturing 1.5 6.2 40.2 75.9 2935.1 9775.2 39757.5 65365.6 28.6 49.3 82.6 155.4
Service 4.5 8.6 21.8 45.6 16272.2 17261.7 23258.3 25784.0 37.8 64.5 114.4 208.3
Total 2.0 4.3 16.5 39.8 6082.6 11211.0 26711.9 33761.3 23.1 31.4 57.5 101.7
Source; Author’s computation using GGDC Database
For China, South Korea and India 2010 is used instead of 2011 and for South Korea 1963 is used instead of 1961
Page | 119
in 2001 and 2011 respectively. Similarly, the manufacturing sector has
increased in its productivity per unit of labor from 4.8 in 1961 to 7.1 in 1981
but later on it declined to 4.4 in 2001 and further to 3.2 in 2011.
Page | 120
Figure 4.4: Relative Labor Productivity, Ethiopia, and SSA, (1961 – 2011)
2
1.5
SSA
1
.5
Ethiopia
0
Page | 121
Figure 4.5: Sectoral Labor Productivity Levels of Ethiopia, (1961-2011)
3
Market Service
Service
2
Non-Market Service
Manufacturing
1
Agrculture
0
Page | 122
finance, insurance, real estate and business services (aggregated as business
and finance services). The non-market services include the public service
(government service) and the community, social and personal service
(aggregated as personal service).
Page | 123
Table 4.5: Gross Value Added (GVA) Share of Service Sectors, Ethiopia and ComparatorCountries, (1961-2011)
SSA Average Ethiopia Ghana
Service Sub sectors 1961 1981 2001 2011 1961 1981 2001 2011 1961 1981 2001 2011
Services 29 37 44 50 9 21 38 44 29 34 46 46
Market Services 21 24 33 38 7 16 24 31 23 22 33 33
Distributive Trade 16 15 20 20 5 12 14 19 12 9 11 11
Transport & communication 4 5 6 10 1 3 5 6 9 10 16 16
Finance & business 2 5 8 7 0 2 4 7 1 2 6 6
Non-Market 7 13 11 13 3 5 14 13 6 13 14 14
Government 7 12 9 12 2 3 11 11 5 11 9 9
Personal 1 1 1 1 1 2 3 3 1 2 4 4
China India South Korea
Services 12 14 28 35 27 34 47 53 70 60 54 51
Market Services 6 7 12 15 18 22 32 39 22 25 32 29
Distributive Trade 3 4 8 10 11 13 16 18 9 12 14 12
Transport & communication 2 2 3 4 4 5 8 9 1 5 8 8
Finance & business 1 1 1 1 4 4 8 11 12 9 10 9
Non-Market 6 7 16 19 8 12 15 14 48 34 22 22
Government 5 4 5 6 6 10 12 12
Personal 2 2 11 14 3 2 2 2
Source; Author’s computation using GGDC Database
For China, South Korea and India 2010 is used instead of 2011 and for South Korea 1963 is used instead of 1961
Page | 124
2 percent in 1961 to 14 percent in 2011. From the service sub-sectors, the
personal service sub-sector constitutes the major share of China’s total output.
But in Ethiopia, the personal service sub-sector constitutes the least share in
the total output in 2011.
On the other hand, the distributive service sector has grown rapidly
during the Imperial government from 1961-1974. During the Military
government, the distributive sector growth stagnated because of the
discouraging business environment in the country because of nationalization
of business and private properties. After the change in government in 1991,
the sector is increased gradually from 10 percent share in total output in 1991
to 19 percent in 2011. The finance and business, the transport and
communication and personal services are also increasing their share in total
output. However, the growth in these sub-sectors is very slow and the share
for finance and business and transport and communication remained around 6
percent as the share for personal service is still stagnating on 3 percent of the
total output of the country. Therefore, the Ethiopian service sector output
growth can be characterized by the higher and increasing share of the
Page | 125
distributive trade and public service sub-sectors and low share of personal
services.
Page | 126
for the distributive service sector has more than doubled within 7 years
increasing from 5 percent in 2004 to 11 percent in 2011. The distributive
service sector constitutes the largest share of market service and much of the
shift in employment occurs in the distributive service since 1961. From 1961
to 2011, the employment shares in the distributive sector increased by 13 folds
from 0.8 percent to 10 percent of the employment in the service sector. The
output in the distributive sector has also increased by from 5 percent in 1961
to 19 percent in 2011.
Personal 2 2 11 14 2 3 3 3
Page | 127
Table 4.7: Labour Productivity of Service Sub-Sectors, Ethiopia, and Comparator Countries (1960-2011)
SSA average Ethiopia Ghana
Service Sub-sectors 1961 1981 2001 2011 1961 1981 2001 2011 1961 1981 2001 2011
Services 209.6 497.3 650.4 268.1 12.1 7.0 9.2 11.3 2.0 1.4 2.3 2.1
Market Services 222.6 579.0 822.3 317.8 18.5 12.0 11.4 11.0 2.1 1.3 2.3 2.1
Distributive Trade 205.2 437.8 636.7 277.3 20.0 9.9 7.8 7.3 1.3 0.7 1.0 0.9
Transport & comm.. 266.3 757.5 1055.9 531.8 18.8 18.3 32.6 45.9 5.5 5.2 8.2 9.3
Finance & business 386.8 2393.6 2058.4 341.5 9.8 65.8 70.0 54.1 7.7 5.1 5.9 4.9
Non-Market 179.7 393.9 394.5 178.7 6.3 2.9 6.9 12.1 1.7 1.4 2.2 2.1
Public service 348.8 793.0 667.7 288.2 6.7 3.8 10.1 18.7 2.1 1.6 2.8 3.0
Personal service 40.6 60.3 92.4 55.0 5.5 2.0 3.4 5.0 0.7 0.9 1.5 1.3
Total Economy 111.4 383.6 405.7 214.7 3.0 2.9 2.7 4.2 1.6 1.1 1.5 2.0
China India South Korea
Services 4.5 8.6 21.8 45.6 37.8 64.5 114.4 208.3 16272.2 17261.7 23258.3 25784.0
Market Services 6.3 11.6 35.4 74.1 61.5 84.5 120.5 213.5 8349.8 10342.0 19567.1 21927.4
Distributive Trade 6.5 12.4 22.4 49.2 51.6 74.2 92.9 162.5 4252.4 6618.1 13921.9 17632.1
Transport & comm.. 5.2 9.5 40.2 79.2 44.6 66.9 112.1 199.6 3035.7 12099.4 35514.1 39599.7
Finance & business 8.3 13.1 110.3 229.4 374.5 343.3 358.2 507.9 71261.5 34354.8 24841.6 21072.5
Non-Market 2.7 5.4 11.3 23.5 20.6 45.4 103.1 195.0 28789.0 33933.4 32022.7 33476.6
Public service 2.8 6.6 28.4 62.2 18.2 54.3 146.4 284.9
Personal service 2.2 3.4 3.4 7.3 29.4 27.5 40.6 75.8 28789.0 33933.4 32022.7 33476.6
Total Economy 2.0 4.3 16.5 39.8 23.1 31.4 57.5 101.7 6082.6 11211.0 26711.9 33761.3
Source; Author’s computation using GGDC Database
For China, South Korea and India 2010 is used instead of 2011 and for South Korea 1963 is used instead of 1961
Page | 128
2011. Similarly, for China, India, and South Korea the service productivity
level has risen in 2001 and 2011. However, for the SSA average and Ghana
the service productivity level has declined in 2011 as compared to 2001.
The labor productivity levels of the market service sectors have been
over the manufacturing sector in all the periods [see Figure 4.6]. However, the
level of productivity in the market services is slightly declining. This situation
is similar to other Sub-Saharan African countries, where workers relocated
mainly to the market services industries. According to de Vries et al (2013),
‘market services activities had above-average productivity levels, but the
productivity growth was low and increasingly falling behind the world
Page | 129
frontier’(de Vries, Timmer, & de Vries, 2013). On the other hand, the non-
market service productivity level has been declining even though it was
greater than the manufacturing sector in the early periods. It started to rise in
1984 as the manufacturing sector productivity level starts to decline. In 2010
and 2011, the non-market service productivity level has surpassed the market
service productivity level due to the higher productivity level observed in the
public service. The aggregate economy-wide productivity level has also shown
improvement since 2004 as a result of high productivity observed in the non-
market services.
Figure 4.6: Labour Productivity of Services and Manufacturing in
Ethiopia (1961-2011)
3.2
2.8
Productivity level (in natural logarithm)
Market Service
2.4 Service
2.0
Manufacturing
1.6
Non-market Service
1.2
0.4
1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
Year
Source; Author’s computation using GGDC Database
Page | 130
However, the share of the public service in the output has increased since 1967
as the share of the distributive service decreased. The share of public service
increased from its lowest share of 10 percent in the early 1970s to 24 percent
in 2011. On the other hand, the share of personal services that includes
community and social services has shown a decline from its highest level of
10 percent in the late 1990s to 6 percent in 2011. The transport and
communication service share in GVA has also shown little change increasing
from 10 percent in 1961 to its highest level of 15 percent in 1992 and then it
declined to 13 percent in 2011. On the other hand, the finance and business
service share has increased steadily from its lowest level of 5 percent in 1961
to 15 percent in 2011.
70%
60%
50%
40%
30%
20%
10%
0%
1979
1961
1963
1965
1967
1969
1971
1973
1975
1977
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
2011
Year
From the employment side, the distributive trade service is again the
dominant contributor to the total service employment as shown in Table 4.8
and Figure 4.8. However, the share of the distributive trade service sector in
employment was below 50 percent for most of the period up until 2004 and
exceptionally it was 50 percent in 1970 and 1971. Its contribution to
employment share increased from 44 percent in 1990 to 65 percent in 2011 as
its contribution to GVA share has declined from 50 percent to 42 percent in
Page | 131
the same period. On the other hand, the employment shares in the public
service and personal service have declined since 2000. The public sector has
exhibited a larger decline in its share of employment within the service sector.
It has constituted 35 percent of employment in services in 1961 and declined
steadily to 25 percent in 1990 and 15 percent in 2011. This indicates that the
government limited the public service employment in the country even though
government sizes increases due to its huge public investment and increase in
output. On the other hand, the decline in employment share for the public
service could be associated with the shift of labor from public services to
personal services in the distributive trade services and growing employment
opportunity in the non-governmental organizations (NGOs). In addition, it is
also associated with a change in the economic system from high government
control in the socialist system to a market economy in 1991. Further,
following the economic reforms, privatization of public enterprises may have
contributed to the decline of public service employment share. In recent
periods, there is also growing tendency among the urban population to engage
in self-employment activities, specifically in the distributive trade service.
Market Service 72.6 77.2 62.5 69.8 47.4 45.0 50.2 71.7
Distributive Trade 57.4 55.8 38.5 42.0 34.6 39.4 45.1 65.4
Transport and communication 10.0 11.8 12.7 12.6 6.4 4.5 3.6 3.1
Finance and business 5.1 9.7 11.4 15.2 6.4 1.0 1.5 3.2
Non Market Service 27.4 22.8 37.5 30.2 52.6 55.0 49.8 28.3
Public services 19.3 14.5 28.9 24.1 34.5 26.6 26.4 14.5
Personal Services 8.2 8.3 8.6 6.1 18.0 28.4 23.4 13.8
Page | 132
communication and finance and business services has declined. The
employment share of transport and business service constitutes 6.4 percent of
the service employment in 1961. This level of share has declined continuously
and reached 4.7 percent in 1992 and 3.1 percent in 2011. Similarly, the share
of the finance and business service declined from 6.4 percent in 1961 to 0.8
percent in 1984 and started to rise to 1.1 percent in 1984 and 3.2 percent in
2011. The employment share of the finance and business, and transportation
and communication sub-sectors is very small as both sectors involve highly
skilled labor force and pay higher wages than the other service sub-sectors.
However, they have an output share that is relatively higher, which contributes
to the higher productivity observed in the sectors.
80%
60%
40%
20%
0% 2001
2009
1961
1963
1965
1967
1969
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2003
2005
2007
2011
Year
Page | 133
measures. At this time when there is no developed industrial sector (including
mining, construction, and manufacturing), the role taken by the service sector
may be used to fill the gap by contributing to GDP and employment growth
and to stimulate growth in the other sectors. Therefore, this section will try to
highlight the contribution of the service sector in GDP (GVA) and
employment for the years 1962-2011 using descriptive statistics and Shapely
growth decomposition.
Page | 134
was 3.9 percent with the minimum -4.4 percent and maximum 11.9 percent.
Thus, the larger contribution for the rise in the GVA growth was attributed
more to the service sector.
The growth in the economy from 2004-2011 was stable and the
average growth was 11.2 percent with a minimum of 9.9 percent in 2010 and
the maximum of 12.8 percent in 2005. The contribution of agriculture to the
total economy growth rate has declined from 68 percent in 2004 to 35 percent
in 2011 and the period average was 41 percent. The manufacturing sector’s
contribution to the output growth in Ethiopia was 5 percent on average during
2004 -2011 and it increased from 3 percent in 2004 to 6 percent in 2011.
Whereas the service sector contribution has increased from 20 percent in 2004
to 62 percent in 2009 and 45 percent in 2011 with the period (2004-2011)
average of 47 percent, which is more than the agriculture sector. During this
period, the larger contribution from the service sub-sectors is attributed to the
distributive service sector. Therefore, the leading role of agriculture in driving
growth in Ethiopia was overtaken by the service sector since 2007.
80%
Percentage contribution to GVA growth
10%
60%
40%
5% GVA growth (%)
20%
0% 0%
2000
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2002
2004
2006
2008
2010
-20%
-5%
-40%
-60%
-10%
-80%
-100% -15%
Year
Agriculture Manufacturing Services Total Economy Growth
Page | 135
Therefore, we can observe that the role of the service sector to the total
economy growth is increasing from time to time. The market service sector
specifically the distributive service sector takes the largest share of the
contribution to total output growth. After 2004, the service sector leading role
could be attributed to the change within the service sector itself as its output
share and employment share has increased. The other reasons for the growth
of service in Ethiopia, which is associated with income, productivity
difference, and change in demand structure, are studied in chapter 6.
80%
5.0%
60%
20% 3.0%
0%
2.0%
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
-20%
1.0%
-40%
-60% 0.0%
Year
Agriculture Manufacturing Services EMP Growth
Page | 136
to zero employment growth was observed when Eritrea separated from
Ethiopia and become an independent country in 1993. From 1990 to 1993
employment growth declined below 2 percent as a result of civil war and
succession of Eritrea from Ethiopia. After 1994 the employment growth has
increased though it declined in 1998 and 2002/3 because of Ethio-Eritrean war
and drought respectively. However, since 2004 the employment growth is
increasing along with the GDP growth. The agriculture sector has been
contributing the largest share to the employment growth in all the periods prior
to 2006. In 2006 and 2007, the contribution of agriculture to employment
growth was overtaken by the service sector. Its contribution to employment
growth was only around 40 percent. The dominance of the agriculture sector
contribution in employment growth was reduced since 2006/07 as both the
service sector contribution increased to 40 percent and the manufacturing
sector contribution also increased to about 20 percent.
Therefore, from the descriptive analysis, we can observe that the role
of the service sector to the total economy growth in GVA and employment is
increasing. The leading role of the service sector could be attributed to the
change within the service sector itself as its output share and employment
share has increased along with a rise in its productivity level. The
decomposition analysis can further indicate the sources of this growth.
Page | 137
of 4.8 percent and high growth period (2004-2014) with average growth rate
of 11 percent.
Page | 138
Table 4.10: Contribution of Sectors to the Change in Employment Rate,
Ethiopia (1999 - 2013)
1999-2005 2005-2013
Contribution of
Change in total Change in total Contribution of
the sector to
Sector employment employment the sector to
total
rate (percent rate (percent total e growth
employment
points) points) (percent)
growth (percent)
Agriculture 4.87 85.16 -0.50 -21.43
Mining 0.16 2.79 0.19 8.27
Manufacturing 0.59 10.24 0.00 -0.20
Utilities 0.00 0.01 0.10 4.34
Construction 0.45 7.84 0.56 24.24
Service -0.35 -6.05 1.97 84.77
Market Service -0.81 -14.19 0.34 14.55
Distributive Trade -0.85 -14.91 -0.25 -10.56
Transport & Comm. 0.00 0.08 0.41 17.44
Finance & business 0.04 0.64 0.18 7.67
Non Market Service 0.47 8.14 1.63 70.22
Public service 0.13 2.29 0.64 27.61
Personal services 0.33 5.85 0.99 42.61
Total employment rate 5.7 100.0 2.3 100.0
Source: Author’s computation using decomposition tool
Nevertheless, the service sector contribution to the employment rate
has changed in the high growth periods, accounting 84.7 percent of the total
employment rate growth or job generation. The agriculture sector contributed
negatively (-21.43 percent) to employment rate growth witnessing the shift to
the service sectors. The largest contribution in this period is observed in the
government and personal services (non-market services) accounting 70
percent of the change in the employment rate. The other important
contributors are the construction, and transport and communication sectors
accounting 24.4 percent and 17.4 percent respectively. The manufacturing, the
agriculture, and the distributive trade experienced employment rate reducing
growth. Therefore, the contribution of agriculture to job creation is overtaken
by the service sector. Despite the fact, the government of Ethiopia had
introduced employment generation policies focusing on micro and small scale
manufacturing enterprises, the decomposition result indicates that the
manufacturing and industry sector it is not contributing to job creation, rather
it is the service sector.
Page | 139
Table 4.115: Contribution of Employment Changes to Change in PCGDP,
Ethiopia, (1999 - 2013)
1999-2005 2005-2013
Percentage Percentage
Contribution Contribution
Sector of total of total
to Change in to change in
change in change in
PCGDP PCGDP
PCGDP PCGDP
Agriculture 217.22 41.20 -33.81 -0.95
Mining 7.12 1.35 13.05 0.37
Manufacturing 26.12 4.95 -0.31 -0.01
Utilities 0.02 0.00 6.85 0.19
Construction 20.01 3.79 38.26 1.07
Service -15.42 -2.93 133.76 3.75
Market Services -36.19 -6.86 22.96 0.64
Distributive trade -38.04 -7.21 -16.67 -0.47
Transport & Comm. 0.21 0.04 27.52 0.77
Finance and business 1.64 0.31 12.11 0.34
Non Market Service 20.76 3.94 110.80 3.10
Public service 5.83 1.11 43.57 1.22
Personal Service 14.93 2.83 67.23 1.88
Page | 140
growth by 38.04 birr. This undermines the contribution of service sector at the
aggregate level, reducing per capita GDP by 15.42 birr. The positive and high
contribution of the non-market service (20.76 birr) could not offset the
negative contribution that the distributive service sector experienced during
1999-2005. In addition, during 2005-2013, the share of agriculture in
employment rate growth is overtaken by the service sector, specifically by the
non-market service accounting 3.1 percent of 4.4 percent. On the other hand,
agriculture, manufacturing, and distributive service had a negative
contribution to the change in per capita GDP. This decomposition analysis of
employment changes on per capita GDP identifies the importance of the non-
market service sector for GDP per capita growth through employment
changes.
Page | 141
Table 4.12: Growth Decomposition, Contribution to Total Growth in GDPPC, Ethiopia (1999-2013)
1999 -2005 2005-2013
Percentage Output per Worker Total Percent Change in Output per Worker Total (%)
Change in Percentage Inter-sectoral (Percent) Employment Percent within Inter-sectoral
Sector
Employment within Sector Shifts (%) Sector Change Shifts (%)
Change in w* in w*
Agriculture 41.2 3.9 -0.8 44.3 -0.9 30.2 1.9 31.2
Mining 1.4 -11.8 10.7 0.2 0.4 0.5 0.7 1.6
Manufacturing 5.0 -2.2 -0.3 2.5 0.0 5.4 0.0 5.4
Utilities 0.0 1.2 -0.5 0.7 0.2 -0.6 1.2 0.8
Construction 3.8 -1.3 6.4 8.8 1.1 6.7 2.1 9.9
Service -2.9 18.2 9.6 24.9 3.7 34.1 18.4 56.2
Market Service -6.9 27.7 9.3 30.1 0.6 26.3 16.8 43.7
Distributive -7.2 27.1 -9.6 10.3 -0.5 28.5 -1.3 26.8
Transport 0.0 9.7 -1.4 8.3 0.8 0.0 4.1 4.9
Finance 0.3 -9.1 20.3 11.5 0.3 -2.3 14.0 12.1
Non-Market 3.9 -9.5 0.3 -5.3 3.1 7.8 1.6 12.5
Public 1.1 -7.4 0.0 -6.3 1.2 6.9 1.8 10.0
Personal 2.8 -2.1 0.4 1.1 1.9 0.9 -0.3 2.5
Subtotals 48.4 8.0 25.0 81.4 4.4 76.4 24.3 105.1
Demographic component 18.6 -5.1
Total 100 100
Total % change in GDP per capita 16.7 97.0
*w - refers to productivity
Source: Own Computation using decomposition tool
Page | 142
The inter-sectoral shifts with positive contribution refer that on average
labor moved from lower than average productivity sectors to above average
productivity sectors. Hence, in the period 2005-2013 except for distributive
trade and personal service sector, all the other sectors has a positive
contribution implying that on average labor has moved from low productivity
sectors of agriculture and manufacturing sectors to high productivity sectors of
the service sector.
Page | 143
are subdivided into two periods based on the economic growth performance of
the country (WB, 2015a). These are the low growth period (1992-2004) and
high growth periods (2004-2011). Structural change in this study refers to the
shifts in the sectoral composition of economic activity from sectors of low
productivity to sectors of high productivity. In other words, it can be expressed
as the reallocation of labor from low-productivity sectors to higher-
productivity sectors. Therefore, the structural change process in Ethiopia is
first tested using the structural bonus and structural burden hypothesis for the
different periods. Then, results and discussions on productivity growth
decomposition into structural change effects, within-sector productivity effect
and the role of the service sector are presented.
Page | 144
On the other hand, the dynamic or interaction effect shows structural
burden hypothesis, which is calculated as the sum of the interactions between
the changes in the weight of employment and the changes in the labor
productivity of each branch of activity. The more shifts that take place towards
highly productive sectors, the greater will this effect. This effect can be used
to check the hypothesis of the structural burden proposed by Baumol (1967).
The hypothesis is valid if the value of this is negative due to a shift of labor
from manufacturing to services that have less productivity. Thus, the
interaction term can be used to capture Baumol’s hypothesis of a structural
burden of labor reallocation on aggregate growth, which predicts that
employment shares shift away from progressive industries towards those with
lower growth of labor productivity (see Peneder (2003)).
Aggregate LP Growth -0.70 0.67 -0.29 -0.33 -0.46 -0.06 -0.02 -0.54
Percentage
Contribution to 212 -201 89 100 84 12 4 100
Aggregate LP
Source; Author’s computation using GGDC database
Page | 145
period, the average productivity level of the manufacturing sector was one of
the highest. From the earlier analysis of the productivity growth in Ethiopia,
the manufacturing sector has been increasing until 1984 (see Figure 4.6). On
the other hand, the employment growth for the manufacturing sector was more
than the service sector but it was less than agricultural employment growth.
Thus, the shift effect in 1981 – 1992 for the agriculture sector was positive and
it was negative for all the other sectors. This means there was a shift of labor
from the manufacturing sector and other service sectors to the agriculture
sector. However, in the early industrialization period (1961- 1981), and post-
reform periods (1992 – 2004) and (2004 – 2011) the static shift effect for the
manufacturing sector was positive implying that workers moved out of the low
productive sectors of the agriculture and manufacturing sectors, and joined the
service sectors. Therefore, the structural bonus hypothesis was confirmed in
Ethiopia in all the periods considered except in 1981-1992.
This result is also consistent with the finding by de Vries, Timmer, &
de Vries (2013), which states that most of the African countries have de-
industrialized in the 1970s and 1980s due to a decline in manufacturing share
in aggregate GDP. In addition, they confirmed that Africa had static structural
gain (structural bonus) and dynamic losses (structural burden), which is
comparable to the situation in Latin America but different in Asian countries.
Timmer & Szirmai (2000) also found no evidence on structural-bonus
hypothesis for the manufacturing sector in four Asian countries in 1963–1993.
Page | 146
Structural burden, a negative between effects for the service sector,
was detected in 1961-1981 and 1982 – 1991, while no structural burden was
observed in the post-reform periods and growth periods (1992–2004) and
(2004 – 2011). In 1961-1981, the between effect was -0.34 percentage point
and in 1981 – 1992 it was -0.01 percentage point. According to the structural
burden hypothesis, the negative between effects is due to a shift of labor from
the manufacturing sector to the services sector that has less productivity. That
means the employment shares shift away from progressive industries towards
those with lower growth of labor productivity (Peneder, 2003). Thus, during
(1961-1981) and (1981 – 1992), the negative interaction term could be due to
the shift of labor from the manufacturing sector to the service, whose
productivity level was declining. We have observed that the productivity level
in the service sector has declined from 1961 to 1984 while the productivity
level of the manufacturing sector increased until 1984.
In the periods 1992 – 2004 and 2004 – 2011, the productivity level of
the manufacturing sector has declined, though labor has continued to move
into the sector. On the other hand, labor productivity has increased in the
service sector as more labor also shifted into the sector from the agriculture
sector. Thus, no structural burden was observed in the periods due to a higher
productivity level of the service sector than the manufacturing sector. Based
on Baumol’s ‘Cost Disease Hypotheses’; an increasing share of services
results in a productivity slowdown due to the inherent labor-intensive nature
of service production. However, recently high productivity in the service
sectors was observed because of improvements in the productivity of market
service sectors such as the financial sector and transport and communication
services. The high productivity level for most of the market services in
developing countries tend to reduce or eliminate the structural change burden
as it was the case in India. Hence, developing countries are expected to ‘suffer
from a structural change burden at early stages of development’ (Szirmai,
2011). However, the within productivity level of the market service in
Ethiopia for the periods 1992–2004 and 2004–2011 was negative. Rather the
non-market services had a high level of within productivity level (see Table
Page | 147
4.14). From the early descriptive analysis in Figure 4.6, the non-market
service productivity level was increasing due to high productivity level
observed in the public service sector, while the market service productivity
level has slightly declined in recent periods. Therefore, even though Ethiopia
had high growth and expansion of service sector, the structural burden effect
of the service sector was eliminated due to productivity growth in the non-
market services as oppose to the case in other developing countries. This could
be due to the growth in the size of the Ethiopian public service output in GVA
and decline of employment share in the public sector.
In addition to the structural bonus and structural burden observed in
Table 4.13 and 4.14, the aggregate labor productivity growth and the
contribution of the within productivity effect and structural change effect can
be identified. The aggregate labor productivity growth for Ethiopia in the early
industrialization period (1961–1981) and pre-reform period (1982–1991) was
negative, (-0.33) and (-0.54) respectively. It becomes positive (0.29) and
(1.39) in 1992-2003 and 2004-2011 respectively. The positive labor
productivity growth (0.29 percentage point) in 1992-2003 was mainly
attributed to the structural shift effect accounting 143 percent of the total
change, which originated from the market service sector. The within
productivity growth effect had productivity reducing effects due to small
productivity level of the agriculture sector. On the other hand, the highest
productivity growth (1.39) was observed in 2004-2011 as 58 percent of the
growth is contributed by the shift effect because of the labor movement to the
market services. The within productivity effect has also contributed 49 percent
to the growth. In this case, the source for the within productivity growth
emanates from the high labor productivity observed in the agriculture sector
(0.65 percentage point) and non-market services (0.20 percentage points).
Further analysis, focusing on service sectors on the movement of labor to
above-average productivity level sectors and productivity growth level for
expanding and shrinking sectors, can be conducted by considering the
shrinking and expanding sector in the decomposition tool.
Page | 148
4.2.1.2 : Role of the Service Sector in Structural Change in Ethiopia
The contribution of sectors in the structural change effect can be either
positive or negative, depending on whether a sector is expanding or shrinking.
The earlier decomposition methods or the ‘traditional decomposition methods’
are not good enough to measure the contribution of sectors to productivity
growth (de Vries, Timmer, & de Vries, 2013). By dividing sectors into
expanding and shrinking sectors based on the changes in employment shares,
they calculated the static between-effect relative to the average productivity
level of the shrinking sectors and the dynamic between- effects relative to the
average productivity change of the shrinking sectors. Here the structural
change or the reallocation effect is split into two terms: static reallocation
effect and dynamic reallocation effect. The static reallocation effect captures if
workers move to above-average productivity level sectors and the dynamic
reallocation effect identifies if productivity growth is higher in sectors that
expand in terms of employment shares (see (de Vries, Timmer, & de Vries,
2013); (Timmer & de Vries, 2009).
The output of the decomposition in Table 4.15 and 4.16 show the role
of sectors in a structural change in Ethiopia during (1961 – 1992) and in (1992
– 2011) respectively. According to result in Table 4.15, the aggregate labor
productivity growth in 1961 – 1981 was -0.04 and in 1981 – 1992 it was -0.52.
The negative productivity growth implies the slowdown of the growth in the
country. However, in 1992 – 2004, the aggregate labor productivity growth
improved and became 0.3. It further increased in 2004 – 2011 to 1.5
contributing to the high growth observed in the period. Looking at the effects,
in all the periods except in 1992–2004, the within effect dominates the
between effects of structural changes. As shown in Table 4.16, in (1992–2004)
the between effect dominates the within effect, which means, the within
sectors productivity growth have little contribution to the improvement in the
productivity growth as compared to the structural change effect. However, in
the later period (2004–2011) the within sectors productivity growth become
the source of the high productivity growth observed in the country.
Page | 149
Table 4.15: Shift-share of LP growth, Role ofSectors in Ethiopia (1961 – 1992)
Economic Sectors 1961-1981 1981-1992
Within Between effect Average LP Within Between effect Change Average LP
effect Static Dynamic effect Static Dynamic
Agriculture -0.34 0.03 0.00 -0.30 -0.46 -0.62 -0.05 -0.37
Industry 0.05 0.02 0.04 0.10 -0.06 -0.02 0.00 -0.02
Mining 0.00 -0.02 0.02 0.00 0.00 -0.08 0.01 -0.02
Manufacturing 0.03 0.01 0.01 0.05 -0.02 0.06 0.00 0.01
Other Industry 0.00 0.00 0.05 0.06 -0.03 0.00 0.00 -0.01
Utilities 0.00 -0.01 0.03 0.02 0.00 0.00 0.01 0.00
Construction 0.00 0.00 0.03 0.03 -0.04 0.01 -0.01 -0.01
Service -0.12 -0.34 0.62 0.16 0.08 0.61 -0.02 0.23
Market Service -0.07 -0.19 0.46 0.20 0.00 0.17 -0.03 0.05
Distributive Trade -0.08 -0.28 0.47 0.11 -0.04 0.15 -0.01 0.03
Transport & Comm. 0.00 0.00 0.04 0.04 0.02 0.03 -0.01 0.01
Finance and business 0.08 -0.03 0.00 0.05 0.01 0.00 0.00 0.01
Non Market Service -0.04 -0.13 0.13 -0.04 0.05 0.45 0.03 0.18
Public services -0.02 -0.05 0.06 -0.01 0.04 0.18 0.01 0.07
Personal Services -0.01 -0.07 0.06 -0.03 0.01 0.27 0.02 0.10
Aggregate LP growth -0.41 -0.29 0.67 -0.04 -0.43 -0.02 -0.06 -0.52
Percentage Contribution to
11.25 8.12 -18.37 1.00 84 4 12 100
Aggregate LP growth
Source; Author’s computation using GGDC database
Page | 150
Now, let us see the contribution of the service sector and other sectors
in the structural change process of the country in the respective periods.
During 1961-1981, the negative growth of the aggregate labor productivity (-
0.04) was mainly due to the negative within sectors productivity level of the
agriculture and service sectors (distributive service, public service, and
personal services). However, the manufacturing sector had a positive within
productivity level, which implies that the country was in the early
industrialization process due to favorable growth in the manufacturing sector.
Even though, the manufacturing sector had a higher productivity growth it
could not offset the decline in the productivity of the service and agriculture
sector, which contributed to the negative productivity growth observed in the
period. The static effect was also negative contributing to the decline in
aggregate labor productivity growth. The negative static between effects
implies that the shift of labor in the period was not towards sectors that have
above average productivity level. Rather labor shifted from below average
productivity sectors to above average productivity sectors. That means the
average productivity level of the shrinking sectors was more than the
productivity level of the expanding sectors. In table 4.17, we can observe that
the agriculture sector was the shrinking sector as all other sectors expand in
terms of their employment share. In addition, productivity has declined in the
agriculture and service sectors. Therefore, labor has moved out of the
agriculture but the productivity level of the agriculture was more than other
sectors, resulting in negative structural change effect. On the other hand, the
dynamic between effects is positively contributing to the overall structural
change effect because of the labor movement to sectors that have above the
average productivity growth. All the sectors had a positive dynamic effect;
however, the larger effect is generated from the service sector, specifically
from the distributive service trade. Therefore, in the early industrialization
period in Ethiopia, even though aggregate productivity growth was negative,
there was productivity gain from the structural change effect due to positive
dynamic effects.
Page | 151
Table 4.16 Shift-share of LP growth, role of sectors in Ethiopia (1992 – 2011)
1992-2004 2004-2011
Within Between effect Change Average Between effect Average
Economic Sectors Within effect
effect Static Dynamic LP Static Dynamic LP
Agriculture -0.23 0.00 0.00 -0.07 0.75 0.00 0.00 0.25
Industry -0.06 -0.08 0.26 0.04 -0.09 -0.13 0.30 0.03
Mining 0.00 0.00 0.01 0.00 0.00 -0.01 0.02 0.00
Manufacturing -0.02 -0.03 0.08 0.01 -0.03 -0.05 0.08 0.00
Other Industry -0.05 -0.12 0.25 0.03 -0.07 -0.10 0.25 0.02
Utilities 0.00 0.00 0.01 0.00 0.00 0.00 0.02 0.01
Construction -0.03 -0.09 0.19 0.02 -0.05 -0.08 0.19 0.02
Service 0.19 0.07 0.15 0.14 0.14 0.03 0.51 0.22
Market Service -0.02 -0.01 0.27 0.08 -0.04 -0.09 0.60 0.16
Distributive Trade -0.02 -0.01 0.16 0.04 -0.02 -0.07 0.34 0.08
Transport &Comm 0.04 0.00 0.01 0.02 0.04 0.01 0.03 0.03
Finance and business -0.02 -0.02 0.09 0.02 -0.02 -0.03 0.19 0.05
Non Market Service 0.16 0.00 0.00 0.06 0.19 0.00 0.00 0.07
Public services 0.12 0.01 0.01 0.05 0.15 0.01 0.02 0.06
Personal Services 0.03 0.00 0.00 0.01 0.03 0.00 0.00 0.01
Aggregate LP growth -0.10 -0.01 0.41 0.30 0.80 -0.11 0.81 1.50
Percentage Contribution to
-33 -4 137 100 53 -7 54 100
Aggregate LP growth
Source; Author’s computation using GGDC database
Page | 152
In 1981-1992, as table 4.15 shows, all the within effect, static and
dynamic effects were negative leading the country into a sluggish productivity
growth. About 84 percent of the negative productivity growth was contributed
by the decline in the within sectors productivity growth mainly caused by the
decline in the productivity of the agriculture sector. In addition, the sector has
contributed to the negative static and dynamic effects. Therefore, in the pre-
reform period in Ethiopia, productivity growth was stalled because of negative
within sectors productivity growth coupled with negative structural change
effects.
Page | 153
points). Productivity in the agriculture sector has improved in Ethiopia since
2004 as the use of improved agricultural inputs increased. The public service
sector productivity has also increased mainly due to huge public investment in
the public sector. As it is shown in table 4.17, productivity has improved in in
the public service sector, agriculture and transport and communication
services in (2004-2011). However, the within productivity growth of the
manufacturing sector, construction, distributive service trade, and finance and
business services were negative, limiting further growth in the aggregate
productivity growth. By their nature, these sectors are expected to show high
productivity growth and drive the growth as they have high potential to
growth. Thus, this implies that the country has much potential for high
aggregate productivity growth if policies are laid to improve the productivity
of these sectors.
Page | 154
Figure 4.11: Shift Share Decomposition Results, Labor Productivity
Growth, Ethiopia (1961 - 2011)
2004- 2011
1992-2004
1981-1992
1961-1981
Page | 155
Table 4.17: Gross Value Added, Employment, and Labor Productivity, Ethiopia (1961 - 2011)
Gross Value Added Share (%) Employment Share (%) Labour Productivity Level
Economic Sectors 1961 1981 1992 2004 2011 1961 1981 1992 2004 2011 1961 1981 1992 2004 2011
Agriculture 85.1 69.2 64.7 47.1 41.8 96.2 89.1 90.3 84.4 73.4 2.6 2.3 1.6 1.5 2.4
Industry 5.4 9.5 9.2 14.1 13.8 1.5 2.0 1.8 5.0 10.0 10.9 14.0 10.9 7.6 5.8
Mining 0.1 0.1 0.5 0.6 0.8 0.0 0.0 0.2 0.3 0.7 72.3 11.1 6.2 5.2 4.7
Manufacturing 2.0 4.0 3.8 5.4 5.2 1.3 1.7 1.4 3.6 6.9 4.8 7.1 5.8 4.0 3.2
Other Industry 3.2 5.4 5.0 8.2 7.8 0.2 0.3 0.3 1.1 2.4 50.0 52.3 41.0 20.2 13.5
Utilities 0.4 1.3 2.2 2.3 2.0 0.0 0.1 0.1 0.1 0.1 77.5 62.6 64.1 69.7 73.2
Construction 2.8 4.1 2.7 5.9 5.8 0.2 0.2 0.2 1.0 2.3 47.6 49.8 31.6 15.9 10.6
Services 9.5 21.4 26.1 38.8 44.4 2.3 9.0 7.8 10.5 16.5 12.1 7.0 7.3 10.0 11.3
Market Service 6.9 16.5 17.9 25.4 31.0 1.1 4.0 3.2 5.9 11.8 18.5 12.0 12.2 11.6 11.0
Distributive Trade 5.4 11.9 10.5 15.0 18.7 0.8 3.5 2.7 5.3 10.8 20.0 9.9 8.4 7.7 7.3
Transport &comm 1.0 2.5 4.0 5.6 5.6 0.2 0.4 0.4 0.4 0.5 18.8 18.3 23.9 36.1 45.9
Finance and business 0.5 2.1 3.3 4.9 6.7 0.1 0.1 0.1 0.2 0.5 9.8 65.8 82.7 63.1 54.1
Non Market Services 2.6 4.9 8.3 13.3 13.4 1.2 4.9 4.6 4.6 4.7 6.3 2.9 3.9 7.8 12.1
Public services 1.8 3.1 5.5 10.0 10.7 0.8 2.4 2.1 2.2 2.4 6.7 3.8 5.8 12.1 18.7
Personal Services 0.8 1.8 2.8 3.3 2.7 0.4 2.5 2.5 2.4 2.3 5.5 2.0 2.4 3.8 5.0
Total Economy 100 100 100 100 100 100 100 100 100 100 3.0 2.9 2.2 2.7 4.2
Page | 156
puts the country in similar patterns of structural change and service sector
growth.
The other section describes and discusses the results of the inter-
sectoral linkage effect of the service sector with other sector estimated using
input-output analysis. The results and discussion on service sector growth
determinant estimated using Autoregressive Distributed Lag Model are also
presented. Finally, the sustainability of service-led growth is evaluated using
scorecards.
Page | 157
CHAPTER-V
DETERMINANTS OF GROWTH OF SERVICE
SECTOR AND ITS SUSTAINABILITY
Page | 158
policy and economic changes in the country as result of the government
change and liberalization policy reform since 1991. The coefficient for the
variables in natural logarithm represents elasticity and the coefficient for the
dummy variable is transformed to exponential values to represent percentage
changes.
Based on the theoretical and empirical findings the study identifies the
determinant of service sector growth. Hence, the descriptive statistics,
Augmented Dicky Fuller (ADF) test, bounds Test for cointegration, regression
results, and diagnostic tests are presented as follows. The model is applied to
annual data for the period of 1981 to 2011 to explore the existence of a long-
run relationship between service growth (in gross value addition and in
employment) with its determinants.
Table 5.1: Pearson Correlation Matrix for Service GVA and Employment
Variable LnSGVA LnSEMP LnGDPPC LnPTVDC LnPSERV LnFMECA
LnSGVA 1.000
LnSEMP 0.996 1.000
LnGDPPC 0.693 0.714 1.000
LnPTVDC -0.989 -0.980 -0.725 1.000
LnPSERV 0.995 0.990 0.677 -0.991 1.000
LnFMECA 0.976 0.967 0.557 -0.961 0.980 1.000
Source, Authors’ Computation
The correlation result indicates that both the service sector GVA and
employment have positive and strong relationships with all other variables
except LnPTVDC. The positive correlation between income and service GVA
and service employment supports the ‘hierarchy of needs’ hypothesis. Driven
by the income growth, the share of service in GVA and employment may
increase to meet the increasing demand for service by channeling resources
Page | 159
into the sector. In addition, the positive correlation of service GVA and
employment with LnPSERV and LnFMECA is also in line with the
‘exogenous demand shock’ hypothesis. Whereas, the Labour Productivity
difference has a negative correlation with service gross value addition and
employment. This indicates that productivity difference in Ethiopia diverges
from the expected relationship.For instance, Fuchs (1980) presumes a positive
correlation between the productivity gap of manufacturing and service sectors
and the service sector share.
Figure 5.1 also shows the time series trend and direction of the
relationship between service growth and determinant variables over the years
(1981-2011).It shows that service gross value addition and employment have
an increasing trend over the periods. Similarly, producer service and female
economic participation variables have increasing trends over 1981-2011.
However, the per capita GDP variable has declined from 1981 to 1992. After
1992 it has an increasing trend with fluctuation but starting from 2004 it is
increasing constantly. On the other hand, productivity differential variable
declined from 1986 onwards. This implies that since 1986 service productivity
is higher that manufacturing productivity. This is a unique experience to
Ethiopia as the manufacturing sector stagnated and has shown a lower
productivity than other sectors.
Page | 160
Figure 5.1: Line Graphs for Service GVA, Employment, and Determinant
Variables
lnsgva lnsemp
11.6 9.2
11.2
8.8
10.8
8.4
10.4
8.0
10.0
7.6
9.6
9.2 7.2
82 84 86 88 90 92 94 96 98 00 02 04 06 08 10 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10
lnptvdc
2.50
2.25
2.00
1.75
1.50
1.25
1.00
0.75
0.50
82 84 86 88 90 92 94 96 98 00 02 04 06 08 10
lnpserv lnfmeca
9.5 9.8
9.0 9.6
8.5 9.4
8.0 9.2
7.5 9.0
7.0 8.8
6.5 8.6
82 84 86 88 90 92 94 96 98 00 02 04 06 08 10 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10
The results of this test in Table 5.2 show that except LnPTVD and
LnPSERV with constant, all others variables are not stationary at their levels.
Page | 161
In their first differences, some of the variables are stationary at least for one of
the specifications. The results indicate that some of the variables are I (1) and
others are I (0). LnSEMP is not stationary in its level and it becomes
stationary in its first difference only with no constant and trend. Such results
of stationarity test would not allow us to apply the Johansen approach of co-
integration approach. This is one of the main justifications for using the
ARDL approach (bounds test approach of cointegration) developed by
(Pesaran, Shin, & Smith, 2001).
Table 5.3 for ADF unit root test for break shows, the structural breaks
year for each variable is different. All the break years are in the post-reform
period on 1992 and after 1992. In Ethiopia, some of the structural adjustment
Page | 162
policies are adopted stepwise since 1991 following the downfall of Derg
regime and seize power of the new government. The structural break year for
LnSGVA is 1992 that coincides with the reform period. LnGDPPC and
LnPSERV have breaks during 2003, on the year Ethiopia faced a severe
drought. All other variables break at different periods of time but during the
structural reform periods. The test points out that LnSGVA, LnGDPPC, and
LnPSERV variables are stationary with the existence of one structural break as
all the rest variables are non-stationary at the respective break periods.
Therefore, based on this result and with the intention of identifying the effect
of reform on service growth, structural break period 1992 is included in the
model as a dummy variable taking 1 for the reform periods (1992-2010) and 0
otherwise (1981-1991).
Page | 163
Therefore, we can observe that service sector growth is also affected by
different shocks observed in the periods (1981-2011).
Page | 164
ARDL bound tests the null hypothesis of no co-integration against the
alternative using the F-test with critical values. The null hypothesis will be
rejected if the calculated F-statistic is greater than the upper bound critical
value. If the computed F-statistics is less than the lower bound critical value,
then we cannot reject the null hypothesis. The result is inconclusive if the
computed F-statistic falls within the lower and upper bound critical values.
The bound test for Service GVA and Service Employment are presented in
Table5.4. The calculated F- statistics for Model 1 is 21.1, which is greater than
the upper and lower bounds at 1 percent significance level. Therefore, there is
long run co-integration between service GVA and its determinant variables at
1 percent significance level. In Model 2 for service employment, the F-
statistics is 8.8 is again significant at 1 percent significance level. Therefore,
the null hypothesis of no long-run co-integration between service
employments and determining variables is rejected. Hence, we can say that
there is co-integration among the set of (I(0) & I(1)) variables and there can be
at least long run or short run relation among these variables. The estimated
ARDL model and the regression result for both LnSGVA and LnSEMP
models are presented below.
Page | 165
shift consumption from goods to services. Services have an income elasticity
of demand greater than one, but that of demand for goods is less than one
(Engel’s Law). So, it is expected to have a positive correlation between service
and income. This finding hence confirms the ‘hierarchy of needs’ hypothesis
and the result for Ethiopia is similar to other empirical findings(Singh & Kaur,
2014; Kim, 2006).
Page | 166
the demand for intermediate service inputs. However, this is not the case in
Ethiopia as the variable is not significant. Similarly, the productivity
differential, which is the difference of manufacturing and service productivity,
is insignificant and negative. This implies that a service GVA change in
Ethiopia is not only significant but also in contrast to the Cost Disease
hypothesis. The hypothesis is that the shift towards services is due to resource
transfer from manufacturing to services because of the productivity gap
between the two sectors. According to Baumol’s ‘Cost Disease hypothesis’,
‘the service sector rises on the current price basis if there is a manufacturing
sector with high productivity growth and a stagnant service sector with low
productivity’ (Baumol, 1967). However, in Ethiopia, the service sector has
higher productivity level than the manufacturing sector resulting in a negative
productivity gap. Hence, the negative sign is due to the negative productivity
gap observed. Therefore, there could be no transfer of resources from the
manufacturing to the service sector as the Cost Disease Hypotheses argue.
Therefore, we cannot say that service growth in Ethiopia is related with
productivity difference between manufacturing and service sector. Rather,
growth in service GVA is due to the change in income (hierarchy of income
hypothesis) and change in the structure of demand (specifically due to more
female labor force participation). Female labor force participation in the
economy has a greater effect on service GVA than income and reform effects.
Page | 167
Table 5.5: Estimated Long Run and Short Run Coefficients Using the
ARDL Approach
Model 1 Service GVA Model 2 Service Employment
Variables
ARDL(1 0 0 0 0 1) ARDL(1 2 0 01 1)
Coef. t-ratio Coef t-ratio
Long Run
Dependent Variable LnSGVA LnSEMP
LnGDPPC 0.853*** 4.92 1.208*** 6.27
LnPTVD -0.199 -1.07 0.427* 2.03
LnPSERV 0.068 0.35 0.157 0.78
LnFEACT 0.875*** 3.14 1.073*** 3.31
DRFOR92 0.336*** 4.66 0.342*** 4.51
C -1.338 -1.02 -5.495** -2.66
Short Run
D(LNGDPPC) 0.416*** 5.41 0.513*** 5.83
D(LNGDPPC(-1)) -0.220** -2.32
D(LNPTVDC) -0.097 -1.06 0.226* 1.96
D(LNPSERV) 0.033 0.35 0.083 0.79
D(LNFMECA) 0.427** 2.79 1.320** 2.19
DRFOR92 0.031 0.99 0.078** 2.33
CointEq(-1) -0.487*** -6.18 -0.529** -5.31
R-squared 0.961 0.890
Adjusted R-squared 0.858 0.829
Log likelihood 84.79 77.40
F-statistic 9.31(0.0028) 14.63 (0.000)
Durbin-Watson stat 1.574 1.67
Page | 168
hypotheses, the productivity gap in Ethiopia is negative. The negative
productivity gap between manufacturing and service, due to higher
productivity level of service than manufacturing, contrasts the hypothesis.
Therefore, the growth of employment in the service sector is not due to the
channel of more labor to the service sector due to its productivity level, which
is expected to be lower than manufacturing. Thus, the reason for a significant
and positive relationship between service employment and a negative
productivity gap in Ethiopia requires further investigation by considering the
relative price of services and manufacturing.
Page | 169
5.1.4.2 : Short run Error Correction Estimates:
In Model 1, both GDP per Capita and female economic activity are
significant at 1 percent and 10 percent significance level respectively. The
directions of the relationship for all the significant variables are positive as it
is hypothesized. The coefficient for GDP per capita is 0.467 and significant at
1 percent significance level. For a one-percentage change in GDP per capita,
service GVA changes by 0.46 percent. LnFMECA has a coefficient of 0.753,
which is significant at 10 percent significance level. For a percentage change
in the female economic activity, service GVA changes by 0.75 percent. In the
short run, the post-reform period dummy failed to be significant unlike its
significant impact in the long run. The R-squared value for service GVA
explains the 96 percent of the variation in the service GVA. This variation is
hence explained by the explanatory variables that determine service sector
growth.
Page | 170
percent significance level. Finally, reform period has also a significant effect
on service employment as it results in 8.1 percent higher service employment
than the pre-reform period. The R-squared value for service employment
model is 89percent, referring 89 percent of the variation in service
employment is explained by the service determining variables.
Page | 171
Therefore, the determining factors for service growth are income and changes
in the structure of demand due to changes in the female economic
participation. However, the cost disease hypothesis does not hold in Ethiopia
because of the negative productivity gap.
5.4
5.2
5.0
.10
4.8
.05
4.6
.00
-.05
-.10
-.15
82 84 86 88 90 92 94 96 98 00 02 04 06 08 10
Page | 172
Table 5.6: Diagnostics Tests
Model 1 Model 2
Tests
Service GVA Service Employment
Chi2 Prob. Chi2 Prob.
Serial Correlation LM test (Breusch Godfrey) 1.302 0.253 0.759 0.383
Heteroscedasticity Test (Breusch Pagan) 0.39 0.532 3.64 0.056
Ramsey RESET F(3,19) = 0.33 0.800 F(3,15) = 1.02 0.409
ARCH LM 0.031 0.800 0.933 0.334
Stability Test* CUMSUM Stable Stable
CUMSUMSQ Stable Stable
10
1.2
5
0.8
0
0.4
-5
0.0
-10
-15 -0.4
1994 1996 1998 2000 2002 2004 2006 2008 2010 1994 1996 1998 2000 2002 2004 2006 2008 2010
Page | 173
Generally, the determinants of the shift in the service sector in Ethiopia
are estimated using ARDL model. Based on the estimation result, the Service
sector growth in GVA is co-integrated in the long run with GDP per capita,
female economic participation and reform period. Service sector growth in
employment is also co-integrated in the long run with GDP per capita, female
economic participation and reform period and productivity difference. These
findings are consistent with the hypothesis of ‘hierarchy of needs’ and
partially with ‘exogenous demand shock’ hypothesis. However, service sector
growth in Ethiopia is not related to changes in the demand for intermediate
service inputs from the producer service sector. Therefore, the determining
factors for service growth in both employment and output are per capita
income, changes in the structure of demand due to changes in the female
economic participation and economic reforms. Thus, further study is
recommended on the negative productivity gap observed which could have an
implication on manufacturing sector growth.
5.2 : Sustainability of Service-Led Growth in Ethiopian
In this section, the sustainability of the recent growth in Ethiopian that is
driven by the service sector is evaluated. Even though there are arguments
against the viability of service sector led growth, most of the literature from
South Asian and India’s experience indicates the viability of service-led
growth for developing countries as most of these countries have a weak
manufacturing sector. The analysis of the sustainability of Ethiopian service
sector is analyzed borrowing the methodological approach of Amirapu &
Subramanian (2015), who used multi-sector growth framework to establish
five important criteria that any sector must exhibit in order to lead an economy
to rapid, sustained and inclusive development. According to Amirapu &
Subramanian (2015), the five criteria are the existence of high level of
productivity, dynamic productivity growth, expansion of the sector in terms of
its use of inputs, comparative advantage, and exportability. Thus, the
Ethiopian service sector is evaluated based on these scorecards.
Page | 174
High Level of Productivity
From the service sectors, the market services have a slight decline in
the productivity level as compared to the non-market services. As shown in
Table 5.7, productivity level in the non-market services increased from 2.9 to
6.9 and furthermore to 12.1 in 1981, 2001 and 2011 respectively. Productivity
growth in the public service sector is the highest growth witnessed in both
1981-2001 and 2001-2011 periods. It increased from 3.8 in 1981 to 10.1 in
2001 with a growth rate of 1.64. In 2001 to 2011, productivity in the public
sector increased by 86 percent from 10.1 to 18.7. However, from the market
Page | 175
services, productivity has declined in distributive trade service and finance and
business services, while the transport and communication service sub-sector
witness an increase in productivity level. Therefore, in Ethiopia, high service
productivity is observed in the non-market services than the market services.
Even though productivity of the market services has a declining tendency, the
level of productivity is higher than the manufacturing sector, which has a
declining productivity level. Therefore, due to a weak manufacturing sector,
the service sector in Ethiopia has a high level of productivity score, which
enables it to lead and sustain the economic growth of the country.
20
10
0
-10
-20
-30
-40
Year
Page | 176
manufacturing sector continues to be negative, as the service’s productivity
growth remains positive. If we compare the productivity growth of the
manufacturing and the service sector as shown in Figure 5.6, both have a
similar trend of rise and fall, though the service sector has a little bit higher
productivity growth than the manufacturing sector. However, in the recent
years, 2012-2014, the productivity growth was reversed as the manufacturing
sector witnessed a higher productivity growth than the services. The change in
the productivity growth of the manufacturing sector is due to high investment
in the manufacturing sector following the GTP plan, which gives due
emphasis to the sector.
.4
BWA
CHN
.4
CHN
log of change in service productivity (2001-2011)
IND
.2
IND
BWA
GHA
.2
ETH
SKR SKR
ETH SSA
BRZ KNY
0
GHA
TNZ
0
-.2
BRZ TNZ
-.2
SSA
-.4
KNY
0 1 2 3 4 0 1 2 3 4
log of service productivity 2001 log of service productivity 2011
Page | 177
sector and manufacturing sector respectively. The productivity level of 9
countries is included in the analysis including China, India, South Korea,
Brazil, Ghana, Botswana, Tanzania, Kenya, Ethiopia and SSA average. In
Figure 5.7, the fitted values for the change in service productivity and
logarithm value of service productivity is downward sloping. It implies that
countries experiencing a high change in service productivity during the
periods 1981-2001 and 2001-2011 have low service productivity level. During
1981-2001, Ethiopia’s value is below the fitted values of the countries
considered. However, in 2001-2011, Ethiopia witnessed a faster catch up and
growth in service labor productivity. This shows that Ethiopia has more rooms
to catch up more quickly to the countries above the fitted line. In addition, this
indicates the convergence of Ethiopia in the service sector to countries
enjoying high productivity levels, which has an important implication for
sustainability of service-led economic growth.
CHN
IND
SKR
.2
log of change in manufacturing productivity (2001-2011)
CHN
BWA
SKR
GHA
0
BRZ
.5
KNY
TNZ
ETH
IND
-.2
GHA
BRZ
SSA
0
TNZ
-.4
ETH
SSA
BWA
-.5
-.6
KNY
0 1 2 3 4 5 0 1 2 3 4 5
log of manufacturing productivity 2001 log of manufacturing productivity 2011
Page | 178
On the other hand, in Figure 5.8, the fitted values for the
manufacturing sector productivity level and change in the productivity has an
upward sloping. It shows that countries with high productivity level in the
manufacturing sector have the larger change in the productivity growth.
Unfortunately, Ethiopia is below the fitted line in both (1981-2001) and
(2001-2011) periods indicating a sluggish progress in the manufacturing
sector. Further, the productivity growth of Ethiopia in the manufacturing
sector is negative. The negative productivity growth of the manufacturing
sector limits the capacity of the sector in growth convergence and sustaining
economic growth. Therefore, the convergence in the manufacturing sector in
Ethiopia requires much more productivity growth in the sector. Thus, with this
level, it has limited capacity to sustain the growth. Similar to this result,
(Ghani & O’Connell, 2014) also find that late comers to the development such
as Ethiopia have a more strong catch-up and growth acceleration in the service
sector than the manufacturing sector.
Expansion of Services
Page | 179
productivity and labor shift in Ethiopia as labor is shifting from low to high
productive sectors. In addition, the service sub-sectors are growing fast in
productivity growth and absorbing labor fast resulting in structural change,
which can sustain the service, led growth in Ethiopia.
construction
public service
.5
agriculture
-.5
Page | 180
Figure 5.10: Sectoral Productivity and Change in Employment Share,
Ethiopia (2001 – 2011)
1.5
utility
finance and business
transport and communication
1 public service
.5
manufacturing
agriculture
-.5
Exportability
Page | 181
generated by the Ethiopian Airline transportation service. Out of total import
in same period 5.5 percent is travel import and 65.5 percent transportation
import and 29 percent in other services. Since Ethiopia is a landlocked
country, it is a net importer of transportation services and other services too.
From the other services, Ethiopia exports constitute 44.5 percent of
government service (Peace and security in the region, as Ethiopia, is one of the
highest contributors of personnel to the UN peacekeeping missions), 24.9
percent communication service export and 24.2 other business exports. In
addition, from other imports Ethiopian service import, 28 percent is other
business service and 13 percent insurance service (UNCTAD, 2015).
Service export/manufacturer
y =9.5 40%
log of service export and
0.0115x + 0.3003
9 20%
import (%)
8.5 0%
2005 2006 2007 2008 2009 2010 2011 2012
Year
Service exports
Manufacturer imports
service exports/manufacturer imports
Linear (service exports/manufacturer imports)
Page | 182
Therefore, the service sectors’ high productivity level and the high
productivity growth, its potential in growth convergence and greater expansion
of the sector in employment indicates the possibility of the sustainability of
service-led growth in Ethiopia. However, the service sector in Ethiopia lacks a
comparative advantage over other sectors as it is capital intensive and requires
skilled labor. Similarly, Amirapu & Subramanian (2015) found that finance
and business service in India is skill intensive similar with the manufacturing
sector. Thus, it is not aligned with India’s comparative advantage. On the
other hand, a study by Enache, Ghani, & O’Connell (2016) used the
sustainability scorecard to identify sectors that can bring structural change and
rapid sustained growth. Accordingly, they find the distributive service trade
and construction sector fulfilling at least some of the characteristics except
tradability and exportability.
Page | 183
CHAPTER-VI
LINKAGE ANALYSIS OF SERVICE SECTOR
IN ETHIOPIA
Page | 184
commodities 42.9 percent is household consumption demand, 11.9 percent
investment demand, 5.9 percent government consumption demand and 6.3
percent are export earnings. From this, we can observe huge trade deficit or
negative trade balance of Ethiopia, which is 16.1 percent import and 6.3
percent export. On the other hand, the 187.3 billion birr in the activity account
represents the gross output from expenditure side and activity income in the
domestic supply. It is composed 34. percent of commodity input, 32.2 percent
of labor and 33 percent of capital inputs.
The factor income from labor and capital is 60.3 billion birr and 62.4
billion birr respectively. From the payment side, out of the total 62.4 billion
birr factor payment of capital, 89 percent is paid to households, 10.7 percent is
payment to public enterprises and the rest 0.3 percent is to the rest of the
world. In addition, from the total household income of 133.1 billion birr, the
larger proportion (45.3 percent) is generated from wage and 41.7 percent from
capital, 1.12 percent from the government transfer, and 11.8 percent from
remittance.From the expenditure side, commodity consumption expenditure
accounts (114.8 billion birr) or 86.3 percent of household total expenditure. In
addition, 15.5 billion birr or 11.64 percent goes to household savings and 2.7
billion birr or 2.02 percent to tax payment.
Page | 185
Table 6.1: Aggregated Ethiopian Macro SAM (2005/06)
Code Receipts \Payments 1 2 3 4 5 6 7 8 9 10 11 12
12 Total 187.3 267.5 23.1 60.3 62.4 6.7 133.1 23.3 14.1 31.9 47.7
Source; Ethiopian Macro SAM (EDRI, 2009)
Billion Ethiopian Birr, 2005/06
Page | 186
6.1.1.1 : Sectoral Contribution to GDP
The sectoral structure of Ethiopian economy during 2005/06 can be
generated from the Social Accounting Matrix. The contribution of each sector
in total output, total demand and total value added at factor cost is computed
in the following way. The total output generated by each sector in 2005/06, as
shown in Table 6.2, is 187.3 billion Ethiopian birr. The total output
incorporates the intermediate demand and value added costs. Out of the total
output produced, the service sector constitutes the largest share, which is 42.2
percent. The largest contribution within the service sector also comes from the
distributive service sector, accounting 18 percent followed by the public
service sub-sector with 9.1 percent of the total output. Here, the share of
agriculture in total output is lower than service sector, accounting 34.7
percent, and the manufacturing sector accounts only 9.6 percent. On the other
hand, the total demand refers to the intermediate demand and the final demand
of the commodities. The value of the total demand during 2005/06 was 267.5
billion Ethiopian birr. Here also the service sector has the largest share in total
demand along with the manufacturing sector. The demand for service and
manufacturing sector accounts for 34.3 percent each; which is more than the
agriculture sector (21.9 percent). Specifically, the huge demand for services
comes from the distributive trade service-sub sector (13 percent), which
incorporates all the demands for wholesale and retail trade services and trade
margins (transaction costs), repair services, and hotel and restaurants.
Table 6.2: Share of Sectors in Total Gross Output and Total Demand
Gross Output/GDP Total DD
Sectors
ETB Percent ETB Percent
Agriculture 65.0 34.7 58.5 21.9
Manufacturing 18.0 9.6 91.9 34.3
Other Industries 25.2 13.5 25.5 9.5
Mining 0.7 0.4 1.1 0.4
Utility 3.3 1.7 3.1 1.2
Construction 21.2 11.3 21.3 7.9
Service 79.1 42.2 91.6 34.3
Distributive Service 33.8 18.0 34.8 13.0
Transport and Com 10.3 5.5 18.9 7.0
Business & Finance 14.3 7.6 17.5 6.5
Public Service 17.1 9.1 17.2 6.4
Personal and Other Service 3.6 1.9 3.2 1.2
Total 187.3 100.0 267.5 100.0
Source, Ethiopian Macro SAM (EDRI, 2009)
ETB - Billion Ethiopian Birr, 2005/06
Page | 187
The value added share of the sectors is presented in Table 6.3. During
2005/06, Ethiopia’s largest factor cost is spent on the agriculture sector
amounting 48.1 percent of the total GDP at factor cost. Next to agriculture,
40.4 percent of the total value at factor cost went to the service sector. The
distributive service sub-sector constitutes the largest proportion (13.4 percent)
of all the sub-sectors. However, the contribution of manufacturing to GDP at
factor cost is 4.7 percent, which is lower than all the service sub-sectors
contribution except the personal services. In addition to the sectoral
contribution to GDP, sectoral factor intensity can be also compared. A sector
with a larger share of labor than capital is labor-intensive sector and a sector
that has larger capital than labor is a capital-intensive sector. Thus, the most
labor-intensive sector is agriculture with 73.5 percent of the sectors’ value-
added is paid to labor. In addition, agriculture is also the least capital-intensive
sector taking only 23.4 percent of the sector’s value-added to capital. On the
other hand, services are the most capital intensive sectors which paid 61.4
billion birr of the sectors’ value added to capital and only 18.9 percent is paid
to labor. Among the service sub-sectors, distributive service; and business and
finance services are the most capital-intensive sectors. However, public
services and distributive services are more labor intensive than other service
sub-sectors. Here, we can observe that services are more capital-intensive
sector at the aggregate level but more labor intensive than the manufacturing
and industry sectors.
Page | 188
Table 6.3: Share of Sector’s Value Added at Factor Cost
Labour Capital Total Value at factor Cost
Sector
ETB % ETB % ETB %
Agriculture 44.3 73.5 14.46 23.4 58.78 48.1
Manufacturing 2.38 4.0 3.38 5.5 5.77 4.7
Other Industries 2.21 3.7 6.06 9.8 8.28 6.8
Mining 0.18 0.3 0.47 0.8 0.67 0.5
Utility 0.66 1.1 1.63 2.6 2.29 1.9
Construction 1.36 2.3 3.95 6.4 5.32 4.4
Service 11.37 18.9 38.01 61.4 49.39 40.4
Distributive Service 4.12 6.8 12.23 19.7 16.35 13.4
Transport &Comm 0.61 1.0 5.74 9.3 6.35 5.2
Business & Finance 0.39 0.7 11.54 18.6 11.94 9.8
Public Service 4.39 7.3 6.92 11.2 11.31 9.3
Personal & Other Service 1.85 3.1 1.57 2.5 3.43 2.8
Total 60.29 100.0 61.92 100.0 122.22 100.0
ETB – in billions birr (2005/06)
Source; Authors computation from Ethiopian (2005/06) SAM
In addition, the relative factor shares of each sector at aggregated and
disaggregated level are depicted inFigure6.1. In 2005/06, for Ethiopia, 50.7
percent of GDP is generated by capital, while 49.3 percent is generated by
labor. Sector-wise, there are significant differences between agriculture and
service sector in their factor use. The graph clearly indicates that agriculture is
labor intensive while service is capital intensive. Within the service sector,
distributive service, business and finance, and transport and communication
services are largely capital-intensive sectors.
60.0%
50.0%
40.0%
30.0%
20.0%
10.0%
0.0%
Labour
Sectors
Source, Authors computation using 2005/06 SAM
Page | 189
6.1.1.2 : Sectoral Composition of Labour
The factor account of the 2005/06 Ethiopian SAM can also inform the
type of labor force employed in each sector. In this study, the labor factor
account is aggregated into four types of workers. These are agricultural
worker, administrative and professional worker, skilled workers and unskilled
workers. According to the 2005/06 Ethiopian SAM, administrative workers
refer to legislation, senior officials and managers; professional workers
include technical and associate professionals. Unskilled workers refer to
workers who are engaged in jobs which only require little skill or lowest level
of education (workers engaged in elementary occupation). Skilled workers are
workers that have some skill obtained through formal education and training,
experience and informal training (EDRI, 2009).
Page | 190
Figure 6.2: Sectoral Composition of Labour, Ethiopia (2005/06)
100.0%
90.0%
80.0%
70.0%
60.0%
50.0%
40.0%
30.0%
20.0%
10.0%
0.0%
Page | 191
Table 6.4: Households Consumption Expenditure Shares by sector,
Ethiopia (2005/06)
Poor Households Non-poor Households Total
Economic Sectors ETB Percent ETB Percent ETB Percent
Agriculture 11.9 46.8 31.0 34.7 42.9 37.4
Manufacturing 8.4 33.1 36.0 40.2 44.4 38.7
Other Industries 0.3 1.2 1.4 1.6 1.7 1.5
Mining 0.0 0.0 0.3 0.3 0.3 0.2
Utility 0.3 1.2 1.1 1.3 1.4 1.2
Construction 0.0 0.0 0.0 0.0 0.0 0.0
Service 4.8 18.9 21.0 23.5 25.8 22.5
Distributive Service 1.6 6.3 7.3 8.2 8.9 7.8
Transport &Comm. 0.2 0.9 2.5 2.8 2.8 2.4
Business & Finance 2.4 9.5 7.9 8.8 10.3 9.0
Public Service 0.2 0.9 0.8 0.9 1.0 0.9
Personal & Other Service 0.3 1.3 2.5 2.7 2.8 2.4
Total 25.4 100.0 89.4 100.0 114.8 100.0
Source, Authors estimation from 2005/06 Ethiopian SAM
Page | 192
6.1.2 : Structure of Import and Exports
The SAM of Ethiopia for the year 2005/06 can indicate the structure of
imports and exports. As it is indicated earlier, Ethiopia’s export is well below
the total imports of the country leaving the country in a huge trade deficit. The
total import value of Ethiopia in 2005/06 is 47 billion birr whereas the export
value is only 16.8 billion birr. When we see what constitutes the total exports,
the export of agricultural commodities takes 40.9 percent of the total export
share. This is similar to several developing countries; where the majority of
their export relies on agricultural commodities. However, one important
feature in the Ethiopian export structure is the share of the service sector,
which is almost equal to the agriculture sector accounting 40.1 percent of the
total export. This implies that service sector export is as crucial as the
agriculture sector for export earnings. The source of the increasing role of the
service sector in export is the transport and communication service sub-sector,
which originates from the dominant transportation service of the government-
owned Ethiopian Airlines and Shipping Line. It accounts 30.7 percent of the
total exports of the country in 2005/06. The export of the manufacturing sector
is only 18.7 percent of the total exports.
Page | 193
in the value of total demands (ratio of total import to total demand), and export
intensity refers to the share of exports in the value of gross output (ratio of
exports to gross output). The import penetration ratio of Ethiopia in 2005/06,
which is shown in table 6.5 indicates that the Ethiopian manufacturing sector
faces the high import competition as 34.7 percent of total demand is imported.
The transport and communication service sub-sector is also one of the most
tradable service activity with 44.5 percent of the service demand is supplied
by foreigners. This is due to the port services import of the country from
Djibouti. According to UNCTAD (2015), the transportation, storage and
communication service sub-sector accounted for more than 25 percent of
output in Djibouti, reflecting the economic importance of its port services to
the landlocked country Ethiopia.
Personal & Other Service 0.1 0.4 2.1 0.0 0.0 0.1
On the other hand, the export intensity shows that the transport and
communication sector is the most export-intensive sector than the agriculture
sector. About 49.8 percent of the transport service output is sold abroad, while
only 10.6 percent of the agricultural outputs are exported. Therefore, in
2005/06 the service sector in general, and the transport sector, in particular, is
the most dynamic service sector with high intensity of import and export
Page | 194
trades. This means, Ethiopia exports transport services through the state-
owned Ethiopian Airlines and gets transport services (port services) for its
foreign trade from Djibouti and Somaliland. This made the country to be one
of the few large and land-locked countries in the world that exports more
services than goods (WB, 2012b).
The total factor inputs for agriculture, mining and personal service
account over 90 percent of the total input spending implying their reliance on
factor inputs than commodity inputs. About 68.1 percent of the total factor
input spending (90.5 percent) in agriculture is spent on agricultural workers
and only 21.5 percent is spent on agricultural capital. The mining sector input
requirement is also composed of 65.3 percent of input of the capital and 16.3
percent of skilled labor and 5.2 percent from the utility sector. Whereas for
the personal service sector, out of its total factor input requirement
Page | 195
(95.1percent), 43.7 percent is non-agricultural capital input and 51.5 percent is
labor input.
The input requirement for the construction sector also shows that 75
percent of the input is commodity input and 25 percent is factor input. From
the commodity inputs, the sector consumes huge share (52.9 percent) from the
manufacturing sector, 7.3 percent from business and finance service, and 3
percent from transport and communication services. Here also the business
and finance service, and the transport and communication service sectors are
more interlinked than the other service sub-sectors. The distributive trade
service sector input requirement is also satisfied by the commodity input (51.6
percent) and from factor inputs (48.4 percent). This service sub-sector is
highly dependent on transport and communication services because 26.1
percent of the commodity input is covered by the transport and service inputs.
It also purchases 13.1 percent of its inputs from the manufacturing sector.
Page | 196
Table 6.6: Production Technology Coefficients, Ethiopia (2005/06)
Commodity and Factor inputs Sectoral Production Technology Coefficients (%)
1 2 3 4 5 6 7 8 9 10 11 12 13
1 Agriculture 4.3 16.2 4.7 0.0 0.0 5.5 1.7 3.8 0.0 0.0 0.2 0.0 4.4
2 Manufacturing 3.7 38.9 47.5 1.7 22.8 52.9 12.8 13.1 26.1 2.4 15.3 1.3 16.8
3 Other Industries 0.0 3.6 6.0 5.2 3.9 6.3 3.9 2.5 0.6 9.0 5.1 0.5 2.8
4 Mining 0.0 1.5 4.2 0.0 0.0 5.0 0.0 0.0 0.0 0.0 0.0 0.0 0.7
5 Utility 0.0 2.1 1.6 5.2 2.8 1.3 1.1 1.4 0.6 0.3 1.7 0.2 0.9
6 Construction 0.0 0.0 0.1 0.0 1.1 0.0 2.8 1.1 0.0 8.7 3.4 0.3 1.2
7 Service 1.5 9.1 9.0 1.9 2.9 10.2 19.2 32.2 11.8 5.1 13.1 3.0 10.7
8 Distributive Service 0.8 0.8 0.0 0.0 0.0 0.0 1.7 0.2 7.4 0.0 3.2 0.0 1.1
9 Transport and Communication 0.4 1.9 2.6 1.1 0.4 3.0 12.2 26.1 2.2 1.4 2.2 0.8 5.8
10 Business & Finance 0.2 6.4 6.2 0.8 1.1 7.2 4.5 5.9 1.2 2.5 6.2 1.3 3.4
11 Public Service 0.0 0.0 0.0 0.0 0.0 0.0 0.3 0.0 0.0 0.0 1.5 0.2 0.2
12 Personal and Other Service 0.0 0.0 0.2 0.0 1.4 0.0 0.4 0.0 1.0 1.2 0.0 0.7 0.2
13 Total Commodity inputs 9.5 67.9 67.2 8.8 29.6 75.0 37.6 51.6 38.5 16.6 33.8 4.9 34.7
Agricultural workers 68.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 23.6
Admin.& Professional labor 0.1 0.4 0.9 2.0 1.3 0.8 5.7 0.9 1.1 1.3 19.8 15.3 2.6
Unskilled workers 0.0 1.4 3.3 7.5 12.3 1.8 1.2 0.2 0.2 0.2 0.9 19.8 1.1
Skilled workers 0.0 11.4 4.6 16.3 6.7 3.8 7.4 11.2 4.7 1.2 5.0 16.4 4.9
Agricultural Capital 21.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 7.4
Non-agricultural Capital 0.8 18.8 24.0 65.3 50.1 18.6 48.1 36.2 55.6 80.7 40.5 43.7 25.6
Factor inputs 90.5 32.1 32.8 91.2 70.4 25.0 62.4 48.4 61.5 83.4 66.2 95.1 65.3
Total input= Gross Output 100 100 100 100 100 100 100 100 100 100 100 100 100
Page | 197
The intermediate input demand is only 34.7 percent while the factor
input accounts 65.3 percent of the total input requirement for Ethiopian
economy in 2005/06. From the intermediate inputs, the manufactured goods
are the most important as it accounts 16.8 percent of the total inputs. The
service sector at the aggregate level is the next important source of input for
Ethiopian economy by contributing 10.7 percent of the intermediate input
requirement. From the specific service sub-sectors, the transport and
communication sector takes the largest share as it is used widely by all sectors
for the purpose of transporting and storing goods (transportation margin as
transaction costs) in the process of marketing. The business and service sector
follows the transportation and communication services in its role as an
intermediate input in the production process.
Page | 198
generate an increase in output above the economy’s average if the backward
linkage is greater than one (or 100 percent in percentage terms). The sector
with the higher value in backward linkages will have the greater effect on the
demand for domestic production (Drejer, 2002). Therefore, the backward
linkage can be termed as the output multiplier (Reis & Rua, 2006).
Page | 199
service results in a 1.15 unit change increase in the demand for all the sectors.
In the backward linkage analysis, we found that the manufacturing sector is
the least backward oriented sector followed by the transportation and
communication sector and personal service sectors. For the manufacturing
sector, the backward linkage value of 0.672 indicates that for a unit change in
the output of the manufacturing sector, changes the demand for all the sectors
by a lesser proportion of 0.672 amounts. This indicates that the sector’s weak
interconnection in stimulating production in the other sectors. That means the
manufacturing sector in Ethiopia is underdeveloped. Similar to the
manufacturing sector, the service sub-sectors, transport and communication,
and personal services have weaker interconnection and the small multiplier
effect on output growth in other sectors.
On the other hand, the agriculture sector is also again the most forward
oriented sector followed by the manufacturing, distributive trade service and
business and finance service sectors. The forward linkage index for the
agriculture sector is 2.84. If the final demand in each sector were to increase
by one unit, the agriculture sector could be affected by 2.84 units for a one-
Page | 200
unit expansion in all other sectors. Whereas, the mining and the construction
sectors are the least ‘forward-oriented linkage’ sectors. A sector with weak
forward linkage implies their weak interconnection with other sectors in
supplying their output as a final demand for other sectors.
The total linkages show that the agriculture sector is ranked first
(4.304); distributive trade service ranked second (3.445); manufacturing sector
ranked third (3.044); and the business and finance, public service and transport
and communication service sector ranked from fourth to six respectively. The
personal service and other services ranked last showing its weak inter-linkage
with other sectors. All the service sectors, except the personal services, have
strong inter-linkage with other sectors as compare to the mining, utility and
construction sectors. In addition, the result shows a stronger inter-linkage of
the distributive trade service more than the dynamic manufacturing sector,
which implies the importance of the sector in the Ethiopian economy. Since
the value of the inter-industry linkage shows the strength of the sectors among
all the economic sectors, the strongest sector is agriculture followed by the
distributive service trade and manufacturing sector.
Page | 201
respectively. This shows that the weight of the two sectors does not alter their
rank, which was computed based on the Hirschman-Rasmussen backward and
forward linkages. Note that the weighted total linkage indices for the
agriculture sector, distributive trade service, and the manufacturing sector are
greater than the values for the Hirschman-Rasmussen backward and forward
linkages.
Page | 202
Table 6.8: Weighted Backward and Forward Linkages, Ethiopia
(2005/06)
Backward Forward Total Linkage
Economic Sectors Linkages
WBL Rank Linkages
WFL Rank WTL Rank
1 Agriculture 1.744 1 3.253 1 4.997 1
2 Mining 0.928 6 0.035 10 0.964 9
3 Manufacturing 0.819 8 2.609 3 3.428 3
4 Utility 0.899 7 0.113 9 1.012 8
5 Construction 1.040 3 0.201 7 1.241 7
6 Distributive Trade services 1.227 2 2.698 2 3.925 2
7 Transport & Comm. 0.625 10 0.675 5 1.300 6
8 Business & Finance 1.011 4 0.718 4 1.728 4
9 Public Service 1.008 5 0.391 6 1.398 5
10 Personal & other Service 0.771 9 0.127 8 0.897 10
Maximum 1.744 3.253 4.997
Minimum 0.625 0.035 0.897
Source, Ethiopian 2005/06 SAM using SimSIP SAM Software
*The rank is based on values for each type of linkages; the highest value sector is ranked 1
and the least as 10.
WBL = Weighted Backward linkages
WFL = Weighted Forward linkages
WTL = Weighted Total linkage (sum of Weighted backward and forward linkage)
The weighted forward linkage value for agriculture is again very high,
which is 3.253, implying the sector’s high production multiplier effect. That
means, for a one birr increase in the agriculture as primary inputs, the total
output of each sector increases by 3.253 birr. The distributive trade service
sector is second, next to agriculture in its forward orientation. This indicates
the importance of the distributive service sector in Ethiopia, as it has higher
interconnection with other sectors inducing production in other sectors. For a
birr increment in the distributive trade services, the total output of all the
sector (including its self) increases by 2.698 birr. This means the distributive
trade service is highly interconnected with other sectors by providing its
service as an intermediate input. This includes all the transaction costs
(margins) and wholesale and retail trade services that can be used as an input
to other sectors, stimulating production in other sectors. With regard to the
manufacturing sector, it has a weighted forward linkage value of 2.609, which
indicates that the sector has strong forward linkage but weaker backward
linkage (0.819). The total output of all the sectors in the economy would
increase by 2.609 birr if the manufacturing sector as an intermediate input
Page | 203
increased by one birr. This indicates that the manufacturing sector is an
important supplier of input for other sectors. More of the manufacturing sector
outputs do not go to the final consumptions rather it is mainly used as an
intermediate input to other sectors. On the other hand, most of the economic
sectors in Ethiopia during 2005/06 are meant for final consumption.
According to Table 6.8, the mining, utility, personal services, construction,
public service, transport and communication sector, and business and finance
service sectors have a weighted forward linkage indices lower than one. These
sectors are weakly interlinked with other sectors, as the output of all these
sectors goes mainly for final consumption.
Page | 204
Table 6.9: Pure Backward and Forward Linkages, Ethiopia (2005/06)
Pure Backward Pure Forward Pure Total
Rank
Economic Sectors Linkages Linkages Linkage
Page | 205
backward linkage effects. On the other hand, the pure forward linkage effect
of the manufacturing sector shows that production changes in the other sectors
have a 246 billion birr pure impact on the manufacturing sector.
Page | 206
and high forward linkage coefficient greater than one. While the last group
consists of low forward as well as backward linkages, that is, both backward
and forward linkage coefficients are less than one. Thus, in this section, the
key economic sectors in Ethiopia during 2005/06 are identified from the
backward and forward linkages of the Hirschman-Rasmussen approach, the
weighted inter-linkages, and pure linkages. Figure 6.3 shows the key sectors
identified using the Hirschman-Rasmussen approach while Figure 6.4 shows
the key sectors based on the weighted linkages. Finally, Figure 6.5 depicts the
key sectors based on pure linkages.
The upper right quadrant in Figure 6.3 shows that the agriculture and
the distributive service sectors were the key sectors in 2005/06. This means
both sectors have backward and forward linkages greater than one. The other
service sub-sectors, business and finance and the public sectors, exhibited
strong backward linkages along with the mining, utility and construction
sectors. The sectors’ high backward linkage indices show the interconnection
of the various sectors from which these sectors purchased greater amounts of
intermediate inputs from the domestic economy.
1.56
Forward Linkages
1.06
7 8
0.56 9
452 Backward Oriented
Weak Sectors 10
0.06
0.57 0.77 0.97 1.17 1.37
Backward Linkages
Page | 207
The manufacturing sector is the only forward oriented sector with
forward linkages greater than one and backward linkages less than one. The
sector’s higher forward linkages indicate the interconnections of the
manufacturing sector as a direct and indirect output provider for all other
sectors. However, the transport and communication sector is the weakest
sector in the economy with lower indices of both backward and forward
linkage. This implies that the sector has weak backward and forward linkages,
poorly integrated into the economy, and heavily dependent on imports. The
transport and the communication sector in Ethiopia had a high export intensity
and high import dependency. Furthermore, the personal service sector is just
on the line between the weaker sector and backward oriented sector, as it has
backward linkage coefficient is near to one (0.995). The distributive trade
service sector, along with agriculture, is the key sector. They have strong
backward and forward linkage implying their importance in the economy and
stimulating more production of outputs in the economy. Finally, we can
observe that six of the economic sectors in Ethiopia during 2005/06 are
concentrated just on the right half of the backward oriented quadrant. This
indicates that most of the sectors in Ethiopia in 2005/06 as measured by the
Hirschman-Rasmussen method are backward oriented. That means the
economy is more of demand-oriented as most of these sectors provide most of
their output as an intermediate input to other sectors.
Page | 208
Figure 6.4: Key Sectors Based on Weighted Backward and Forward
Linkages, Ethiopia (2005/06)
1
3.00 Forward Oriented Key
3 6
2.50
2.00
Forward Linkages
1.50
1.00
7 8
0.50 Weak Sectors 9 Backward Oriented
10 42 5
0.00
0.53 0.73 0.93 1.13 1.33 1.53 1.73
Backward Linkages
Sector Sector
Economic Sectors Economic Sectors
Number Number
1 Agriculture 6 Distributive Trade service
2 Mining 7 Transport & Communication Service
3 Manufacturing 8 Business & Finance Service
4 Utility 9 Public Service
5 Construction 10 Personal & other Service
Source, Ethiopian 2005/06 SAM using SimSIP SAM Software
Figure 6.4 shows the key sector identification using weighted
backward and forward linkages. Here, the linkages for each sector are
computed taking into account their weight or share in total aggregate
production. The agriculture and the distributive service sectors are again the
key sectors in Ethiopia during 2005/06. The weighted backward and forward
linkage values for both agriculture and distributive trade service are greater
than one. Similarly, the manufacturing sector remained a forward oriented
sector. The difference here in the weighted inter-linkage as compared to the
Hirschman- Rasmussen approach is observed in the mining and utility sectors.
These sectors have been a backward oriented sector in Hirschman- Rasmussen
inter-linkage, but they become weak sectors when their share in total
production is accounted. The backward linkage coefficient for the other
sectors, which are construction, business and finance, public service and
personal services, has also decreased due to their smaller share in total
production as compared to the rest sectors. Therefore, in the weighted key
sector analysis, we observed that sectors that were concentrated in the
Page | 209
backward oriented quadrant have shifted to the weak sector. This shows the
weak inter-linkage for most of the services and utility and mining sector in
Ethiopia in 2005/06.
Finally, the key sectors using the pure backward and forward linkage
analysis is computed and presented in Figure 6.5. According to (Parra &
Wodon, 2010a), pure linkages can be used for a direct comparison of the
relative importance of the economic sectors measured in monetary values. The
sectors were defined as a key, backward oriented, forward oriented and weak
sectors based on their linkage strengths in generating pure linkage effects in
the country’s economy. The average values of the economic sectors for each
backward and forward linkage are used to categorize the sectors into the four
categories. Thus, the average effect of the pure backward linkage and the
forward linkages are 125 and 109 billion birr respectively. If a sector has
above average pure forward and backward linkage, it will be a key sector. The
backward oriented sectors have above average pure backward linkage and
below average pure forward linkage, whereas forward oriented sectors have
above average pure forward linkage and below average pure backward
linkage. The weak sector has below average backward and forward linkages.
Page | 210
Figure 6.5: Key Sector based on Pure Backward and Forward Linkages,
Ethiopia (2005/06)
350 Forward Oriented
Key
1
300
250 3
Pure Forward Linkage
200 6
150
5
100
Weak sector 9
50 8
4
7
0 Backward Oriented
0 2 1050 100 150 200 250 300 350
Apart from the three key sectors, the other important sector in the key
sector analysis is the backward oriented sector. As it is displayed in Figure 6.5,
the construction sector has above average pure backward linkage, which
makes it the only backward oriented sector. All the rest sectors are
concentrated in the weaker sector quadrant, having below average pure
backward and forward linkages. This implies that when the pure inter-sectoral
linkage is computed by discounting the inter-linkage of sectors with
themselves, most of the sectors will have weak inter-linkage with other
sectors. Specifically, except for the distributive service trade, all the service
sectors are considered to have small inter-linkage with other sectors.
Page | 211
Specifically, the personal service and transport and communication have very
weak pure forward linkages.
Therefore, the results show that the distributive trade service, business
and finance services, transport and communication service and the public
service sectors ranked from third to sixth respectively. In the economic
landscape, the agriculture sector leads as it has high backward and forward
linkages. Then, the manufacturing sector follows the agriculture sector due to
its greater forward linkages as compared to the service sectors. The remaining
sectors, which are the utility, construction, personal services, and mining
sectors have very weak backward and forward linkages. Therefore, the
economic landscape of Ethiopia is skewed as most of the economic sectors
have weak inter-linkage with other sectors.
Page | 212
Figure 6.6: Economic Landscape of Ethiopia (2005/06)
3.5
3
1
2.5
3
2 6
1.5 8
7
1
9
0.5 4
0 10
1
1
3 9
6 5
6 2
8
7 5
8 2
9 4
4
10 10
7
Row hierarchy of forward linkages 5 2 3
Column hierarchy of backward linkages
To Sum up, the service sector linkage analysis shows that even though
the sector is important in the production of manufacturing and other sectors, it
is in general not strongly interlinked with the other sectors. When the sub-
sectors are analyzed, particularly, the distributive trade service is an important
sector that has high backward and forward inter-linkage. This implies that the
distributive trade service is a key sector along with the agriculture sector.
However, other service sub-sectors have weak forward and backward inter-
linkages with other sectors.
Page | 213
CHAPTER-VII
MAIN FINDINGS AND POLICY IMPLICATIONS
Page | 214
The main objective of the study has been to analyze the structure and
the growth of the Ethiopian service sector. The specific objectives have been
as follows: (a)to assess the growth trend of Ethiopian service sector and sub-
sectors; (b) to delineate the structural change in Ethiopia with special
reference to tertiarization; (c) to determine the factors that explains the growth
of the Ethiopian service sector; (d) to analyse the service sector inter-linkage
with agriculture and industry sectors; (e) to examine the viability and
sustainability of service sector growth in Ethiopia; and (f) to prescribe the
policy framework for better functioning and sustainability of the Ethiopian
economy
This study investigates the structure and growth of the service sector in
Ethiopia. It attempts to identify the growth trend and composition of the
service sector, the service sector role in structural change, its determinants,
sectoral inter-linkage of the sector, and sustainability of the service led
growth. The study uses secondary data from the GGDC 10 sectors database,
World Development Indicators, and statistical information from Ethiopian
government official publications as well as the 2005/06 SAM of Ethiopia.
Main Findings
Based on the analysis, the study argues that the service sector is
contributing to the growth in output and employment and structural change in
Page | 215
Ethiopia with a potential to sustain the growth and stimulate the growth of the
manufacturing sector. In the light of above defined objectives and using the
methodology outlined, following are the main findings of the study:
Page | 216
c) Analysis of the contribution of each sector to output and employment
growth has been done since the 1960s. During 1962-1973 growth
periods, on an average 39 percent of the growth in output was
contributed by the service sector, which is lower than agriculture. In
1974-1985, even though there was very weak economic growth, the
service sector has contributed 64 percent to the average total output
growth. In 1993-2003, the larger contribution for the rise in the output
growth relative to the previous period is attributed to the service sector.
In the recent high growth periods of 2004-2011, the service sector
contribution has increased from 20 percent in 2004 to 45 percent in
2011 with the period average of 47 percent, which is also more than
the agriculture sector. On the employment side, the total employment
growth in Ethiopia is largely affected by the dominant employing
sector, agriculture. Apart from agriculture, the service sector
specifically the non-market service is contributing positively to the
growth in employment since 2004.
e) The role of the service sector in structural change has been analyzed.
Results from the shift share decomposition show the presence of
structural bonus and structural burden in Ethiopia at various periods.
The structural bonus, the presence of positive shift effect for the
Page | 217
manufacturing sector is observed in all the periods except in 1981-
1992. In the early industrialization period (1961-1981) and post-reform
periods (1992-2004) and (2004-2011), the static shift effect for the
manufacturing sector was positive implying that workers moved out of
the low productive sectors of the agriculture, and joined the
manufacturing sectors and mainly the service sectors, as the sectors
have high productivity level than the agriculture sector. On the other
hand, structural burden, a negative between effects for the service
sector, was detected in 1961-1981 and 1981-1992. However, no
structural burden was observed in the post-reform periods (1992-2004)
and (2004-2011) due to a higher productivity level of the service sector
than the manufacturing sector. The high productivity level for most of
the market services in developing countries tend to reduce or eliminate
the structural change burden. However, the within productivity level of
the market service in Ethiopia for the periods 1992 – 2004 and 2004 –
2011 was negative.
Page | 218
g) Analysis of determinants of the service sector growth has been carried
out using the ARDL model. Based on the estimation result, the service
sector growth in GVA is co-integrated in the long run with GDP per
capita, female economic participation and reform period. Service
sector growth in employment is also co-integrated in the long run with
GDP per capita, female economic participation, reform period and
productivity difference. For a percentage increment in GDP per capita,
service GVA will increase by 0.85 percent confirming the ‘hierarchy
of needs’ hypothesis for Ethiopia. As it assumed, when income
increases, there is the consumption from goods to services. In addition,
more participation of female in economic activity in Ethiopia will
increase the service sector output by increasing household
consumption of services. For a one-percentage change in female
participation, service output will increase by 0.87 percent.
Furthermore, 39.9 percent of the growth in the service sector is due to
the changes in the post-reform period as compared to the pre-reform
period. Hence, economic reform and participation of female in the
labor force cause a change in the demand structure in the economy
confirming the ‘exogenous demand shock’ hypothesis.
Page | 219
deviation and it will take about 2 years for the co-integration equation
to come to the equilibrium. For service employment, the result shows
that the service sector growth in employment is driven by the rise in
income and growth in female employment. Employment in the service
sector increases by 1.2 percent for a 1 percent growth in income. There
could be also a 1.07 percent growth in service employment because of
a 1 percent increase in female labor in the economy. In addition, a 0.42
percent change in the service employment can be observed for a 1
percent change in the productivity gap of the manufacturing and
services. Service employment in Ethiopia has also increased by 40.7
percent as compared to the pre-reform period. In the service
employment equation, the deviation of the co-integration equation
from the long run equilibrium due to certain shocks is adjusted by 53
percent over the next year and it will take about 1.9 years for the
complete adjustment of the long run equilibrium.
Page | 220
finance service sub-sector employed more of administrative and
professional workers than the skilled workers. The largest proportions
of administrative and professional workers are also employed in the
public service sector than other sectors.
l) Furthermore, the descriptive result from SAM also shows that the
service sector consumption expenditure share accounts 22.5 percent of
total household consumption expenditure. From the service sub-
sectors, households spending on business service takes the largest
share followed by the distributive service sector while the least
consumed service by households is the public service sector. Besides,
the non-poor households spend more of their income on consumption
of service as compared to poor households and relative to the poor
households the non-poor households also consume more of the
distributive service, transport and communication services, and
personal services.
Page | 221
communication sector is the most export-intensive sector as half of the
transport service output is sold abroad. Therefore, in 2005/06 the
service sector in general, and the transport sector, in particular, is the
most dynamic service sector with high intensity of import and export
trades.
Page | 222
o) The sectoral linkages analysis using Hirschman-Rasmussen, weighted
and pure inter-linkages show that agriculture has the highest backward
and forward linkages. All the service sub-sectors, except the
distributive trade service sector, have relatively stronger backward
linkage than the forward linkages indicating the inter-linkage of the
service sectors from the demand side. In the backward linkages, the
agriculture sector stood first followed by the distributive service trade
and the public service sector. The manufacturing sector is the least
backward oriented sector followed by the transportation and
communication sector and personal service sectors indicating the
sector’s weak interconnection in stimulating production in the other
sectors. On the other hand, the agriculture sector is the most forward
oriented sector followed by the manufacturing, distributive trade
service and business and finance service sectors. A sector with weak
forward linkage implies their weak interconnection with other sectors
in supplying their output as a final demand for other sectors. All the
service sectors, except the personal services, have strong inter-linkage
with other sectors as compared to the mining, utility and construction
sectors. In addition, the result shows a stronger inter-linkage of the
distributive trade service more than the dynamic manufacturing sector,
which implies the importance of the sector in the Ethiopian economy.
The distributive trade service sector is second, next to agriculture in its
forward orientation. This indicates the importance of the distributive
service sector in Ethiopia, as it has higher interconnection with other
sectors inducing production in other sectors.
Page | 223
service sub-sectors, business and finance and the public sectors,
exhibited strong backward linkages along with the mining, utility and
construction sectors. The sectors’ high backward linkage indices show
the interconnection of the various sectors from which these sectors
purchased greater amounts of intermediate inputs from the domestic
economy. The result further shows that apart from the key sectors, the
other important issue in the key sector analysis is the backward
oriented sector. The construction sector has above average pure
backward linkage, which makes it the only backward oriented sector.
All the rest sectors are concentrated in the weaker sector quadrant,
having below average pure backward and forward linkages.
Page | 224
features that enable it to sustain economic growth. The sector has high
productivity level and high productivity growth as compared to the
manufacturing and the agriculture sector. There is also growth
convergence for Ethiopia in services than the manufacturing sector.
Additionally, Ethiopia has high export potential in the transportation
services. Nevertheless, the resources endowment for Ethiopia and the
comparative advantage of the country is not aligned with service
sector.
The overall conclusion that emerges from the study is that in Ethiopia
the service sector has witnessed high growth and contributed to the growth in
output and employment during the recent growth periods. The sector is also
the source of growth-enhancing structural change. In addition, it has a strong
inter-sectoral linkage through the distributive trade service. Furthermore, the
service sector has also the potential to sustain the high economic growth and
stimulate growth in other sectors.
Policy Implications
Adequate policy and strategic focus needs to be given for the service
sector in the Ethiopian development plans as the contribution of the
service sector to the total output growth or GDP growth and to
employment growth are increasing and the sector is becoming an
important sector to drive growth. In addition, the development policies
in Ethiopia should not undermine the role played by the service sector.
The current growth of the service sector is seen as untimely and
unwanted growth from Ethiopian policies perspective. Nevertheless,
the growth in GDP and employment in Ethiopia is the result of the
growth in the service sector, especially from the employment growth
of the non-market service sector and productivity growth of the market
service sector. Without the massive growth in the service sectors, the
much-required high GDP growth would not be achieved.
Page | 225
To avoid any structural burden to the economy, productivity level in
the market services needs to be improved through the application of
new technologies and opening up the most of the trade restrictions in
the market services. Even though Ethiopia had high growth and
expansion of service sector, the structural burden effect of the service
sector was eliminated due to productivity growth in the non-market
services as opposed to the case in other developing countries. This
could be due to the growth in the size of the Ethiopian public service
output in GVA and decline of employment share in the public sector.
By their nature, these sectors are expected to show high productivity
growth and drive the growth as they have high potential. This implies
that the country has much potential for high aggregate productivity
growth if works policies are laid to improve the productivity of these
sectors.
Page | 226
do not induce growth in the other sectors. Whereas, the high weighted
background linkage values for the agriculture, distributive trade
service, and construction service sector indicates that these sectors
have a higher dependency on the domestic intermediate inputs.
The overall policy brief that emerges from the study is that
development plans and sector specific policies in Ethiopia needs to give due
considerations on strategies to make use of the service sector growth.
Specifically, strategies on economic reforms in the service sector, niche export
markets for service export and specialized services for the manufacturing
sector should be identified and implemented for a balanced and sustainable
economic growth of Ethiopia.
Page | 227
Ethiopia a comparison of the changes in the service inter-linkage with the
2005/06 SAM can be conducted. Furthermore, the negative productivity gap
between service sector and manufacturing sector requires further investigation
by considering the relative price of services and manufacturing. Further
analysis of the structural change process and role of the service sector in
Ethiopia needs to be studied in the most recent periods after, especially after
2011.
Page | 228
Bibliography
Aggarwal, A., & Kumar, N. (2012, November). Structural transformation,
industrialization and poverty reduction: The case of India. United
Nations Economic and Social Commission for Asia and the Pacific, pp.
1-75.
Badibanga, T., Diao, X., Roe, T., & Somwaru, A. (2009, April). Dynamics of
Structural Transformation: An Empirical Characterization in the Case
of China, Malaysia, and Ghana. International Food Policy Research
Institute Discussion Paper 00856, pp. 1-40.
Page | 229
Banga, R. (2005a, March). Role of Services in the Growth Process. Indian
Council for Research on International Economic Relations Working
Paper No. 159, pp. 1-36.
Banga, R., & Goldar, B. (2004). Contribution of services to output growth and
productivity in Indian manufacturing: Pre and post-reforms. Indian
Council for Research on International Economic Relations Working
paper (139), pp. 1-33.
Bhalla, G. S., Chadha, G. K., Kashyap, S. P., & Sharma, R. K. (1990, August).
Agricultural growth and structural changes in the Punjab economy: an
input-output analysis. International Food Policy Research Institute
Research Report 82.
Bhattacharya, T., & Rajeev, M. (2014). Identifying the High Linked Sectors
for India: An Application of Import-Adjusted Domestic Input-Output
Matrix. The Institute for Social and Economic Change Working Paper
329.
Page | 230
Bosworth, B. P., & Triplett, J. E. (2003). Services Productivity in the United
States: Griliches Services Volume Revisited. In E. R. Berndt, & C. R.
Hulten, Hard-to-Measure Goods and Services: Essays in Honor of Zvi
Griliches (pp. 413-447). University of Chicago Press.
Breisinger, C., Diao, X., & Thurlow, J. (2015). Modeling growth options and
structural change to reach middle-income country status: The case of
Ghana. Economic Modelling, 26(2), 514-525.
doi:10.1016/j.econmod.2008.10.007
Breisinger, C., Thomas, M., & Thurlow, J. (2009). Social accounting matrices
and multiplier analysis: An introduction with exercises. Food Security
in Practice technical guide 5, International Food Policy Research
Institute. doi:10.2499/9780896297838fsp5
Cali, M., Ellis, K., & Velde, D. W. (2008, December). The contribution of
services to development and the role of trade liberalization and
regulation. Overseas Development Institute Working Paper 298.
Chen, S., Jefferson, G. H., & Zhang, J. (2011, March ). Structural change,
productivity growth and industrial transformation in China. China
Economic Review, 22(1), 133–150. doi:10.1016/j.chieco.2010.10.003
Page | 231
Chen, S., Jefferson, G. H., & Zhang, J. (2011). Structural change, productivity
growth and industrial transformation in China. China Economic
Review, 22(1), 133-150.
Christiaensen, L., & Todo, Y. (2013, May). Poverty Reduction during the
Rural-Urban Transformation: The Role of the Missing Middle. The
World Bank Policy Research Working Paper 6445.
Cortuk, O., & Singh, N. (2013). Analyzing the structural change and growth
relationship in India: State-level evidence. UC Santa Cruz Economics
Department Working Papers No 712, pp. 1-21.
CSA . (1999). The 1999 National Labour Force Survey Report. Addis Ababa:
Central Statistical Agency.
CSA . (2014). The 2013 National Labour Force Survey Report. Addis Ababa:
Central Statistical Agency.
CSA. (2005). The 2005 National Labour Force Survey Report. Addis Ababa,
Ethiopia.: Central Statistical Agency.
Page | 232
Cuello, F. A., Mansouri, F., & Hewings, G. J. (1992). The Identification of
Structure at the Sectoral Level: a Reformulation of the Hirschman–
Rasmussen Key Sector Indices. Economic Systems Research, 4(4),
285-296.
Dabla-norris, E., Thomas, A., Garcia-verdu, R., & Chen, Y. (2013, July).
Benchmarking Structural Transformation across the World.
International Monetary Fund Working Paper WP/13/176, pp. 1-44.
Das, K. D., Erumban, A. A., Aggarwal, S., & Sengupta, S. (2013). Revisiting
the Service-led Growth in India: Understanding India’s service sector
productivity growth. IARIW-UNSW Conference on Productivity:
Measurement, Drivers, and Trends, (pp. 1-34). Sydney.
Dasgupta, S., & Singh, A. (2005). Will Services be the New Engine of Indian
Economic Growth? Development and Change, 36(6), 1035-1057.
de Brauw, A., Mueller, V., & Lee, H. L. (2014). The Role of Rural–Urban
Migration in the Structural Transformation of Sub-Saharan Africa.
World Development, 63, 33-42. doi:10.1016/j.worlddev.2013.10.013
de Vries, G. J., Erumban, A. A., Timmer, M. P., Voskoboynikov, I., & Wu, H.
X. (2012). Deconstructing the BRICs: Structural transformation and
aggregate productivity growth. Journal of Comparative Economics,
40(2), 211-227.
de Vries, G., de Vries, K., Gouma, R., Pahl, S., & Timmer, M. ( 2014). GGDC
10-Sector Database: Contents, Sources, and Methods. Groningen
Growth and Development Centre.
Page | 233
Debnath, A., & Roy, N. (2012, February 11). Structural Change and
Intersectoral Linkages: The Case of North-east India. Economic &
Political Weekly, XLVII (6), 72-76.
Diao, X., Hazell, P., Resnick, D., & Thurlow, J. (2007). The Role of
Agriculture in Development: Implications for Sub-Saharan Africa.
IFPRI Research Report 153.
Dorosh, P., & Thurlow, J. (2014). Can Cities or Towns Drive African
Development? Economywide Analysis for Ethiopia and Uganda.
World Development, 63, 113–123.
doi:http://dx.doi.org/10.1016/j.worlddev.2013.10.014
Dorosh, P., Schmidt, E., & Shiferaw, A. (2012, Fall). Economic Growth
without Structural Transformation: The Case of Ethiopia. Journal of
African Development, 14(2), 7-40.
Duarte, M., & Restuccia, D. (2009, February). The Role of the Structural
Transformation in Aggregate Productivity. University of Toronto,
Department of Economics Working paper 348, pp. 1-57.
Easterly, W., & Levine, R. (2001). It's Not Factor Accumulation: Stylized
Facts and Growth Models. The World Bank Economic Review, 15(2),
177-219. Retrieved from http://www.jstor.org/stable/3990260
Eckstein, A. J., & Heien, D. M. (1985). The U.S. Experience Causes and
Consequences of Service Sector Growth. Growth and Change, 12-17.
Page | 234
EEA. (2015). Report on the Ethiopian Economy. Addis Ababa, Ethiopia:
Ethiopian Economic Association.
Eichengreen, B., & Gupta, P. (2011, February). The Service Sector as India’s
Road to Economic Growth? National Bureau of Economic Research
Working Paper 16757, pp. 1-51.
Eichengreen, B., & Gupta, P. (2013). The two waves of service-sector growth.
Oxford Economic papers, 65(1), 96-123. doi:doi:10.1093/oep/gpr059
Engida, E., Tamru, S., Tsehaye, E., Debowicz, D., Dorosh, P., & Robinson, S.
(2011, November ). Ethiopia’s Growth and Transformation Plan: A
CGE Analysis of Alternative Financing Options. ESSP Working Paper
30.
Page | 235
Estrada, G., Acharya, Y., Batten, A., Brimble, P., Doung, P., Gulamov, M. I., .
. . Rahman, S. (2013, April). The Service Sector in Lower-Income
Asian Economies. Asian Development Bank Economics Working
Paper Series 347, pp. 1-25.
Falk, M., & Jarocinska, E. (2010, November). Linkages between Services and
Manufacturing in EU countries. SERVICEGAP Review Paper 1, pp. 1-
29.
Fan, S., Zhang, X., & Robinson, S. (2003). Structural Change and Economic
Growth in China. Review of Development Economics, 7(3), 360-377.
Ferede, T., & Kebede, S. (2015). Economic growth and employment patterns,
dominant sector, and firm profiles in Ethiopia: Opportunities,
challenges, and Prospects. R4D Working Paper 2015/2.
Fofana, I., Parra, J. C., & Wodon, Q. (2009). Exports and labor income by
gender: a social accounting matrix analysis for Senegal. Munich
Personal RePEc Archive Paper No. 28473.
Fuchs, V. R. (1968). The Service Economy. New York and London: National
Bureau of Economic Research. Retrieved from
http://www.nber.org/books/fuch68-1
Page | 236
Fuchs, V. R. (1980). Economic Growth and the Rise of Service Employment.
National Bureau of Economic Research Working Paper 486, pp. 1-32.
Gani, A., & Clemes, M. D. (2002). Services and Economic Growth in ASEAN
Economies. ASEAN Economic Bulletin, 19(2), 155-169. Retrieved
from http://www.jstor.org/stable/25773718 .
Gemmell, N., Lloyd, T., & Mathew, M. (1998). Dynamic sectoral linkages and
structural change in a developing economy. Centre for Research in
Economic Development and International Trade, Department of
Economics, University of Nottingham, pp. 1-25.
Ghani, E., & Kharas, H. (2010). The Service Revolution in South Asia: An
Overview. In E. Ghani, The Service Revolution in South Asia (pp. 1-
34). New Delhi: Oxford University Press.
Gollin, D., Jedwab, R., & Vollrath, D. (2016, march). Urbanization with and
without industrialization. Journal of Economic Growth, 21(1), 35-70.
doi:10.1007/s10887-015-9121-4
Page | 237
Hartwig, J. (2010). Structural change, aggregate demand and employment
dynamics in the OECD, 1970-2010. Zurich: KOF Swiss Economic
Institute ETH.
Herrendorf, B., Rogerson, R., & Valentinyi, A. (2014). Growth and Structural
Transformation. In Handbook of Economic Growth (Vol. 2B, pp. 855-
941). North Holland, Amsterdam and NewYork: Elsevier B.V.
Hill, T. (1977). On Goods and Services. Review of Income Wealth, 23(4), 315-
338.
Hoekman, B., & Mattoo, A. (2008). Services Trade and Growth. The World
Bank Development Research Group Policy Research Working Paper
4461.
Jain, S., & Ninan, T. N. (2010). Servicing India’s GDP growth. In A. Shankar,
& M. Rakesh, India’s economy performance and challenges, Essays in
honor of Montek Singh Ahluwalia (pp. 328–365). New Delhi: Oxford
University Press.
Page | 238
Khanal, B. R. (2011). An Economic Analysis of the Lao PDR Tourism
Industry. Ph.D. Dissertation, Lincoln University.
Khanal, B. R., Gan, C., & Becken, S. (2014). Tourism inter-industry linkages
in the Lao PDR economy: An input-output analysis. Tourism
Economics, 20(1), 171-194.
Kim, H.-J. (2006). The shift to the service economy: causes and effects.
Institute for Monetary and Economic Research, the Bank of Korea.
Kolavalli, S., Robinson, E., Diao, X., Alpuerto, V., Folledo, R., Slavova, M., .
. . Asante, F. (2012, Fall). Economic Transformation in Ghana: Where
Will the Path Lead? Journal of African Development, 14(2), 41-78.
Kongsamut, P., Rebelo, S., & Xie, D. (2001). Beyond Balanced Growth. IMF
Working Paper WP/01/85.
Kravis, I., Heston, A., & Summers, R. (1982). The share of services in
economic growth. In F. G. Adams, & B. Hickman, In Global
Econometrics: Essays in Honor of Lawrence R. Klein (pp. 188-218).
Cambridge: MIT Press.
Lee, J.-w., & Mckibbin, W. J. (2014, July). Service Sector Productivity and
Economic Growth in Asia. Asian Development Bank Institute Working
Paper Series No. 490, pp. 1-49.
Page | 239
Lewis, A. W. (1954). Economic Development with Unlimited Supplies of
Labor. Manchester School of Economic and Social Studies, 22, 139–
91.
McLachlan, R., Clark, C., & Monday, I. (2002). Australia’s Service Sector: A
Study in Diversity. Productivity Commission Staff Research Paper.
Page | 240
Mcmillan, M., & Headey, D. (2014). Introduction – Understanding Structural
Transformation in Africa. World Development, 63, 1-10.
doi:10.1016/j.worlddev.2014.02.007
Miller, R. E., & Blair, P. D. (2009). Input – Output Analysis Foundations and
Extensions (Second ed.). Cambridge University Press.
Narayan, P. K. (2005). The saving, and investment nexus for China: evidence
from cointegration tests. Applied Economics, 37(17), 1979–1990.
Nassif, A., Feijó, C., & Araújo, E. (2013). Structural change and economic
development: Is Brazil catching up or falling behind? UNCTAD
Discussion Papers No. 211.
Page | 241
Ngai, L. R., & Pissarides, C. A. (2007). Structural change in a multisector
model of growth. American Economic Review, 97(1), 429-443.
Noland, M., Park, D., & Estrada, G. B. (2012, November). Developing the
Service Sector as Engine of Growth for Asia: An Overview. Asian
Development Bank Economics Working Paper Series No. 320, pp. 1-
48.
Park, D., & Shin, K. (2012a, October). Performance of the Services Sector in
Korea: An Empirical Investigation. Asian Development Bank WP 12-
20, pp. 1-34.
Park, D., & Shin, K. (2012b, December). The service sector in Asia: Is it an
engine of growth? Asian Development Bank Working No. 322, pp. 1-
37.
Parra, J. C., & Wodon, Q. (2010a, October). SimSIP SAM: A Tool for the
Analysis of Input-Output Tables and Social Accounting Matrices
Version 1.1.2. Washington, DC: The World Bank.
Parra, J. C., & Wodon, Q. (2010b). How Does Growth Affect Labor Income
by Gender? A Structural Path Analysis for Tanzania. In J. S. Arbache,
A. Kolev, & E. Filipiak, Gender Disparities in Africa's Labor Market
(pp. 273-296). Washington DC: The World Bank.
Page | 242
Pesaran, M. H., Shin, Y., & Smith, R. J. (2001). Bounds testing approaches to
the analysis of level relationships. Journal of Applied Econometrics,
16(3), 289-326.
Pratt, S. (2011). Economic linkages and impacts across the TALC. Annals of
Tourism Research, 38(2), 630-650. doi:10.1016/j.annals.2010.11.014
Rodrik, D. (2014, October 13). Are Services the New Manufactures? Project
Syndicate. Retrieved from https://www.project-
syndicate.org/commentary/are-services-the-new- manufactures-by-
dani-rodrik-2014-10
Page | 243
Seid, Y., Taffesse, A. S., & Ali, S. N. (2015). Ethiopia — an agrarian
economy in transition. WIDER Working Paper 2015/154.
Singh, M., & Kaur, K. (2014). India’s services sector and its determinants : an
empirical investigation. Journal of Economics and Development
Studies, 2(2), 385–406.
Smith, A. (2007). An Inquiry into the Nature and Causes of the Wealth of
Nations. In S. Soares. MetaLibri Digital Library.
Page | 244
Sonnis, M., Guilhoto, J. J., Hewings, G. J., & Martins, E. B. (1995,
September). Linkages, key sectors, and structural change: some new
perspectives. The Developing Economies, 33(3), 233-270.
Taffesse, A. S., & Ferede, T. (2005, January). The Structure of the Ethiopian
Economy - A SAM-based Characterisation. A Country Economic
Memorandum, the World Bank background paper Report No. 29383-
ET.
Page | 245
Triplett, J. E., & Bosworth, B. P. (2003, September). Productivity
Measurement Issues in Services Industries: “Baumol’s Disease” Has
Been Cured. Federal Reserve Bank of New York provides Economic
Policy Review.
Van Ark, B., & Timmer, M. (2003). Asia’s Productivity Performance and
Potential: The Contribution of Sectors and Structural Change.
University of Groningen & Conference Board.
Verma, R. (2012). Can total factor productivity explain value added growth in
services? Journal of Development Economics, 99(1), 163-177.
WB. (2015a). Ethiopia’s Great Run: The Growth Acceleration and How to
Pace It Report No. 99399-ET. World Bank Group.
Page | 246
1. Annexure-1
Page | 247
Annexure-2
2 Activities 0.6445
3 Labor 0.3219
5 Enterprises 0.1073
Page | 248
Annexure-3
Activities 1 2 3 4 5 6 7 8 9 10
1 Agriculture 3.14 2.33 1.45 2.29 2.32 2.47 1.49 2.32 2.39 2.14
2 Mining 0.18 0.13 0.08 0.13 0.13 0.14 0.08 0.13 0.13 0.12
3 Manufacturing 2.62 1.95 1.21 1.91 1.94 2.06 1.25 1.94 1.99 1.79
4 Utility 0.27 0.20 0.13 0.20 0.20 0.21 0.13 0.20 0.21 0.19
5 Construction 0.26 0.19 0.12 0.19 0.19 0.21 0.12 0.19 0.20 0.18
6 Distributive trade 2.54 1.88 1.17 1.85 1.87 2.00 1.21 1.88 1.93 1.73
7 Transport &commun. 0.72 0.53 0.33 0.52 0.53 0.57 0.34 0.53 0.55 0.49
8 Business and Finance 0.83 0.61 0.38 0.60 0.61 0.65 0.39 0.61 0.63 0.56
9 Public Service 0.58 0.43 0.27 0.42 0.43 0.45 0.27 0.43 0.44 0.39
10 Personal Service 0.26 0.20 0.12 0.19 0.19 0.21 0.13 0.20 0.20 0.18
Page | 249
Page | 1