Migbaru Alamirew Workneh Master of Development Evaluation and Management Supervisor: Prof. Dr. Nathalie Francken Academic year 2014-2015 UNIVERSITY OF ANTWERP INSTITUTE OF DEVELOPMENT POLICY AND MANAGEMENT Dissertation Impact of Foreign Aid on Domestic Savings in Sub-Saharan Africa (Panel Data Analysis)
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Migbaru Alamirew Workneh
Master of Development Evaluation and Management
Supervisor: Prof. Dr. Nathalie Francken
Academic year 2014-2015
UNIVERSITY OF ANTWERP INSTITUTE OF DEVELOPMENT POLICY
AND MANAGEMENT
Dissertation
Impact of Foreign Aid on Domestic Savings in Sub-Saharan Africa (Panel Data Analysis)
Migbaru Alamirew Workneh
Master of Development Evaluation and Management
Supervisor: Prof. Dr. Nathalie Francken
Academic year 2014-2015
UNIVERSITY OF ANTWERP INSTITUTE OF DEVELOPMENT POLICY
AND MANAGEMENT
Dissertation
Impact of Foreign Aid on Domestic Savings in Sub-Saharan Africa (Panel Data Analysis)
I
PREFACE
This dissertation is submitted in partial fulfillment of the requirements for the award of the
Master of Development Evaluation and Management of the Institute of Development Policy and
Management (IOB) of the University of Antwerp.
The selection of my topic was motivated by two main factors; firstly because of my personal
interest and experiences in the area of foreign aid to developing countries as I was working in
Bilateral Cooperation Department, Ministry of Finance and Economic Development of Ethiopia.
Secondly, to see the tremendous impact of the flow of huge amounts of money starting from the
last more than five decades as a form of official development assistance on gross domestic
savings of developing countries especially to Sub-Saharan Africa, which is the dominant aid
receiver region.
In doing this research and in my study here, I have received invaluable help from many people.
Above this, I have learnt a lot of things in an effort to accomplish this research. Getting
ideas/information for this study and seeing to it to achieve the objective of the study has not been
an easy task. First and for most, I am grateful to Almighty God for giving me grace, mercy and
strength in all my endeavors. My special thanks and gratitude extends to my supervisor Prof. Dr.
Nathalie Francken, from whom I get a lot, for her invaluable help and advice and also
constructive comments which helped me to bring this research to what it is. My heartfelt thanks
also to my sponsors the VLIR-UOS Scholarship for their financial support throughout the whole
my study and stay.
I am also very grateful to my beloved family; my father Mr. Alamirew Workneh, my Mother
Wubalech Admasie, my brothers and sisters, and also my girlfriend D. A. for their support and
encouragement throughout my stay away from home. Thanks to all my friends and fellow
students for your unforgettable and memorable friendship and help.
II
Table of Contents PREFACE ............................................................................................................................................................... I
List of Tables ...................................................................................................................................................... III
EXCUTIVE SUMMARY................................................................................................................................. IV
CHAPTER ONE ................................................................................................................................................. 3
THEORETICAL AND EMPIRICAL LITERATURE REVIEW ...................................................................... 3
1.1. Theoretical Literature .................................................................................................................................. 3
1.1.1. Definition of Foreign Aid ......................................................................................................................... 3
1.1.2. The Macroeconomic Rationale for Aid .................................................................................................... 4
1.1.2.1. Harrod-Domar Growth Model ............................................................................................................... 4
1.1.2.2 The Two Gap Growth Model ................................................................................................................. 6
1.2. Empirical Studies on Foreign Aid and Domestic Saving ............................................................................ 7
1.2.1. Brief Summary of some articles on Domestic Saving and Foreign Aid ................................................. 12
CHAPTER TWO .............................................................................................................................................. 15
Description of Variables, Methodology and Empirical Model Specification................................................... 15
2.1. Variables of Interest .............................................................................................................................. 15
2.3. Methodology and Empirical Model Specification ................................................................................ 19
2.4. Scope and Limitation of the Study ....................................................................................................... 24
CHAPTER THREE .......................................................................................................................................... 25
EMPIRICAL DATA ANALYSIS .................................................................................................................... 25
3.2. Interpretation of Estimation Results .......................................................................................................... 28
3.2.1. Interpretation of the Aggregate Estimation Results ................................................................................ 28
3.2.2. Interpretation of the Disaggregated Estimation Results ......................................................................... 29
CHAPTER FOUR ............................................................................................................................................ 32
Annex I .............................................................................................................................................................. 43
Annex II ............................................................................................................................................................. 52
Annex III ............................................................................................................................................................ 53
III
List of Tables
Table 1: Brief Summary of scientific articles on domestic saving and Foreign Aid ..................... 12
Table 2: Variable used in the panel data estimation analysis and their symbols ........................... 24
Table 3: Summary of Estimation Results for the impact of aggregate official development
assistance on gross domestic savings ................................................................................. 25
Table 4: Summary of Estimation Results for the impact of disaggregated official development
assistance on gross domestic savings ................................................................................. 26
Table 5: Descriptive Statistics of Variables Used .......................................................................... 52
IV
EXCUTIVE SUMMARY
This paper tried to address the impact of foreign aid on gross domestic savings in forty Sub-
Saharan African countries in aggregate and by disaggregating foreign aid into bilateral aid from
DAC member countries, including the European Union and multilateral aid from UN agencies,
World Bank, IMF and African Development Bank. Using annual panel data from 2002 to 2013
for twelve years in the sample countries, Simple Panel data analysis with fixed effects and
without fixed effects is done, and also Hausman test, Breusch-Pagan LM test and time fixed
effect tests are applied. The study seeks to determine whether the direction of the impact of
foreign aid on gross domestic savings is different based on aid modalities (bilateral and
multilateral aid). Based on the results from the random effect model estimation, and other
diagnostic tests, the impact of bilateral aid and multilateral aid is the same with the aggregate
effect of net official development assistance on gross domestic saving, even if multilateral aid is
insignificant. The absence of Good governance in Sub-Saharan Africa, as an institutional factor,
is also affected gross domestic savings negatively. The estimation result is in favor of those
researchers who claim about the negative impact of foreign aid flow on gross domestic savings
based on their research, but the result may depend on the variables and the methodology used.
Hence, to see the different arguments of researchers on foreign aid and domestic savings and to
know the real impact of foreign aid in more broad and detailed concept, macro-economic policy
soundness as an institutional factor and the role of aid beyond growth, which may have a
potential influence on gross domestic savings in developing countries, may play an important role
in the statistical estimation in addition to good governance and the disaggregation of official
development assistance.
Keywords: Sub-Saharan Africa, Panel Data Analysis, Gross Domestic Savings, Foreign Aid
(Official Development Assistance), Bilateral Aid, Multilateral Aid
1
INTRODUTION
Africa has reached a turning point in 2000s and starts to play a more significant role in the global
economy since its economy has experienced high and continuous economic growth in the past
decade (UNCTAD, 2014). The economic growth of Africa during 2000s is impressive and higher
than during the 1990s and 1980s as the average gross domestic product (GDP) grows more than
double from just above 2% during the 1980s and 1990s to above 5% between 2001 and 2014
(AEO, 2015). This Economic growth varies across Africa, which reflects the factors, such as
differences in income levels, availability of natural resources, macroeconomic policies, and
political and social stability, that affects the growth of the economy which are different in
different regions of Africa. As AEO (2015) indicates the economic growth remains highest in the
East, West and Central Africa, 7%, 6% and 5.6%, respectively, and lowest in North and Southern
Africa, 1.7% and 3% respectively in 2014. In North Africa, except Mauritania, almost all
countries of the region’s experienced very low economic growth and even negative growth of
Libya’s economy due to Arab spring which results the political unrest and civil war. In sub-
Saharan Africa, the region which contains more than 47 countries of Africa, the average
economic growth was 5.2% in 2014 (AEO, 2015).
Despite the rapid economic growth in Africa, specifically in Sub-Saharan Africa in the last
decade, many countries in the region are struggling with several development challenges like
self-insufficiency in food supply (food security problem), poverty and inequality, low economic
infrastructure, environmental degradation and low regional and global economic integration
(UNCTAD, 2014). These challenges influence the investment and domestic saving in the region,
which are the main drivers of sustained and transformative economic growth. In addition, the
domestic saving and investment can be influenced by foreign aid, the growth of gross national
product per capita in each country, the productivity of agriculture measured in value added in
agriculture, and unemployment may also affect the investment and savings.
The flow of Official Development Assistance (ODA), which is the most common foreign aid
transfer, received by Sub-Saharan African countries increases for the last 53 years and in 2013 it
was around 46.77 billion USD which is 78 times more than the amount in 1960 (597 million
2
USD) (as shown in figure 1 in the Annex III). This implies that, developed countries invest their
huge amount of money in these developing countries for the last more than five decades for the
sake of economic growth (whatever the reason was behind) and their interest increases time to
time as the graph shows. But regarding the effectiveness of Official Development aid transfer,
there are opposite arguments made by researchers on the area. Some researchers like Kalyvitis
(2007) and Moyo (2009) shows a negative impact of foreign aid on gross domestic savings, while
other researchers like Balde (2011), Irandoust and Ericsson (2005) and Shields (2007) observes a
positive impact of aid. Not only at a glance, rather foreign aid based on its sources (bilateral or
multilateral aid), may affect gross domestic savings differently.
In addition to foreign aid, the Gross National Product per capita (GDP per capita), value added
agriculture, sound administration and unemployment may cause a possible impact on gross
domestic savings and investment, and on economic growth in world-wide. In developing
countries, especially in Africa, domestic saving is very low, even if it is the main source of
funding for domestic investment and economic growth (World Bank, 2015). These major foreign
aid receiver countries are less developed and hence, domestic saving in these countries may
affected by different factors, including foreign aid, per capita GDP growth, value added
agriculture and good governance. In favor of this, the main research question of this paper is:
what will be the impact of foreign aid, at a glance and also in disaggregation of bilateral and
multilateral aid, on gross domestic savings in Sub-Saharan African countries? In addition, the
impact of good governance, value added agriculture, unemployment and per capita GDP growth
on gross domestic savings in the region will statistically tested and analyzed. To answer these
questions, the paper has four Chapters and it organized as follows: Chapter one provides a review
of theoretical and empirical studies on the relationship mainly between gross domestic savings
and foreign aid in the general and in Sub-Saharan Africa in particular. Chapter two provides the
model specification, variables of interest and hypotheses. Chapter three discusses the empirical
analysis of the study, which mainly focused on fixed effect and random effect model estimations
using secondary data from World Development Indicators and World Wide Governance
Indicators. The Hausman test and F-test are discussed in this chapter. Finally, chapter four
concludes the study and provides concluding remarks.
3
CHAPTER ONE
THEORETICAL AND EMPIRICAL LITERATURE REVIEW1
1.1. Theoretical Literature
1.1.1. Definition of Foreign Aid
Foreign aid can be in-kind like physical goods, skills and technical know-how, or it can be in
cash and/or noncash financial support like grants and loans at concessional rates transferred from
donors to aid recipient developing countries. The Development Assistance Committee (DAC) of
the Organization for Economic Cooperation and Development (OECD) defines aid as Official
Development Assistance (ODA). According to the DAC, aid qualifies as ODA when the
following three criteria are met: it has given by official agencies; based on the main objective of
economic development and welfare promotion and twenty five or more percent of the aid should
be grant. Project aid, humanitarian aid including food aid, technical assistance and programme
aid (balance of payments support and budget support) are most common modalities that ODA
provided to recipient countries. Also, in addition to the official development assistance, the Non-
Governmental Organizations (NGOs) provide aid in support of poverty reduction activities and
emergency relief in developing countries.
Aid can be from individual governments through a bilateral agreement and negotiations between
the donor and the recipient country, bilateral aid; from multilateral organizations like the World
Bank, the United Nations, the International Monetary Fund, and regional development banks,
including the African Development Bank and Asian Development Bank, multilateral aid; or from
non-governmental organizations (NGOs) such as World Vision, Red Cross Society, and Oxfam,
non-governmental aid. This study uses the DAC definition of foreign aid.
1 For the development of my literature review, I used the first End of Module Paper as a base.
4
1.1.2. The Macroeconomic Rationale for Aid
The macroeconomic rationale for aid, which is based on the growth model of Harrod-Domar, is
about how aid can substitute and increase domestic savings, foreign exchange and government
revenue for economic growth. In the Harrod-Domar growth model, which assumes physical
capital formation drives growth, investment rate and productivity of investment are the factors
that affect output. Tradition considering physical capital formation as a central driving force of
economic growth is not only in the Harrod-Domar model but also in the 1950s and 1960s gap
models (Hjertholm, Laursen and White, 2005). The total saving of countries is generated from
domestic sources (domestic saving) and from foreign sources (foreign savings) in an open
economy and these savings are the main sources of investment. Hjertholm, Laursen and White
(2005) argue that when countries saving from domestic sources are not sufficient to finance their
investment to attain the planned economic growth, a savings gap will occur; and trade gap or the
foreign exchange gap will occur when the revenue from exports are not enough to import the
desired level of capital and services, based on the assumption that not all goods and services are
produced domestically. This argument of Hjertholm, Laursen and White (2005) make sense and
strongly based on the idea that, there should be a clear distinction between the desired and actual
investment and domestic savings (savings gap); and also between the desired and actual import-
export (trade gap) in a given an exogenously determined planned growth rate. The difference
between these two actual and desired gaps become large, it affects the investment and economic
growth if there is no any other option like foreign aid to finance the large gap and the desired
growth rate will not be attained finally. If foreign aid is allocated in the desired and appropriate
way, it can fill both gaps simultaneously (by paying for imported capital equipment, a single aid
dollar relaxes both the savings and the foreign exchange constraint). The Harrod-Domar growth
model and The Two Gap model are discussed below:
1.1.2.1. Harrod-Domar Growth Model
An econometric growth model of Harrod-Domar, which is an influential and very handy
applicable growth model in modern aid theory, assumes that capital is the most crucial factor for
enhancing the growth rate of the economy (Pankaj, 2005). According to the Harrod-Domar
5
model, output depends on the productivity and the rate of investment in which saving (sum of
domestic and foreign savings in open economy) is the main source of finance. This model, which
explains economic growth in terms of a savings ratio and capital-output coefficient, (as cited in
Kabet, C. N., 2008: 19) is expressed as;
g = (I/Y) /μ ……….. (2.1) and
I/Y= A/Y + S/Y …………. (2.2)
where I is required investments, Y is output; g is the target GDP growth, A is aid, S is domestic
saving and μ the incremental capital-output ratio (ICOR). The ICOR, which is the ratio of
investment rate to the growth rate, gives the amount of additional capital units required to yield a
unit of additional output. When the value of the incremental capital-output ratio (ICOR), which is
mostly range between 2 and 5, is high, it is an indication of poor quality of investment which
implies, to attain a very low economic growth rate, huge amount of investment should undertake.
By using the idea of ICOR, the Harrod-Domar model was the base for the national development
plans in developing countries and even now a day it is mostly used by researchers and some
policy makers (de Silver, 1984 cited by Kabet, 2008). As of Sheilds (2007), the simple version of
the Harrod-Domar growth model is the base for the most famous models which claim that aid
induces growth when growth is determined by the saving rate where the growth rate of per capita
income (g) is given by:
g = s/v – n ……………..2.3
Where s is the marginal saving rate, v is the incremental capital-output ratio and n is the population
growth rate. In this model, saving is equal to the investment and anything which increases the
marginal saving rate (s), decreases the ICOR (v), or decreases the population growth rate (n) and if n
is less than g will increase the growth rate of per capita income (g) and aid is taken as either
augmented savings or improving technology. Hence, from the above arguments since savings is the
sum of domestic and foreign saving (like foreign aid) in an open economy, it is possible to say
that foreign aid can influence the savings and economic growth rate and can fill the saving-
investment gap to achieve a target growth rate. Despite this argument, savings, especially
domestic savings are the main source of investment and hence can play the most imperative role
in the economic growth of countries. Thus, for those aid recipient countries, to minimize their
dependence on foreign aid and also the amount of aid flows from donors may decrease due to
different factors like financial and economic crises like Ireland and Italy did for Ethiopia in 2008
6
and 2009, they need to increase their capacity of generating domestic saving, which will increase
the domestic revenue to finance investment.
1.1.2.2 The Two Gap Growth Model
The first and standard model, even still the most influential growth model, which is used to
justify the role of foreign aid to allow countries in achieving the desired investment and economic
growth rate, was the ‘two gap model’ of Chenery and Strout (1966) (Ahmad and Ahmed, 2002;
Kabet, 2008 and Serieux, 2009). This growth model is based on two assumptions; linear and
stable relationship between investment and growth, and aid finances investment. The saving gap
and foreign exchange gap (trade gap) are the two gaps considered in this growth model. The
inflow of foreign resource from the outside world like foreign aid can enable developing
countries to fill their saving gaps and foreign exchange (trade gap) by providing the needed funds
and foreign exchange (Serieux, 2009). This filling of the two gaps by foreign aid will hold true
only if the only constraint on investment is a shortage of fund that is a liquidity problem, not
another problem like lack of incentives, and also the ‘two gap model’ supports the investment-
limited growth assumption of Harrod- Domar growth model that assumes a specific amount of
investment to increase growth (Kabet, 2008). This is in a sense that if poor incentives are the
cause for low investment, aid will not fill those gaps and not increase investments rather it will
finance consumption or other reverse flows. In addition, Easterly (2001) and Bender and
Lowenstien (2005) also criticizes the two assumptions of two-gap model as; the linearity of
investment and foreign aid relationship may not happen, i.e. the production function may allow
substitution of capital by labor and hence if non-substitutable assumption fails, the model fails to
see how the foreign aid allocate and what was the role of this resource. Foreign aid may also use
to finance consumption, and even to finance reverse flows like debt repayment as Serieux (2011)
argue. The effectiveness of foreign aid in filling the two gaps may also determine by the
productivity of the investment itself (White, 1992 cited by Kabet, 2008). In line with the
effectiveness of aid in this model, the saving gap and foreign exchange gap may not be at the
same time. As Serieux (2009) argue, the saving gap is the binding constraint in the early stages of
growth and as the economy develops, since the saving is expected to increase due to increase in
income, the saving-investment gap will be covered by domestic savings; while when the
7
economy grows, it will come with higher demand of investment and hence higher demand for
imported intermediate and capital goods, and this may exceed the revenue from exports to
finance it and thus the foreign exchange gap (trade gap) will become the binding constraint on
investment and sustained economic growth.
1.2. Empirical Studies on Foreign Aid and Domestic Saving
Even if the researchers cannot reach at the same argument in common about the impact of foreign
aid on domestic saving and economic growth, the area is widely studied and is still more
investigations are going on. From those literatures which examine the impact of foreign aid on
domestic savings and economic growth in recipient countries, some studies find evidence of a
positive effect, while other studies find evidence of a negative effect. For instance, an influential
study by Burnside and Dollar (2000) founds that foreign aid can be effective, and only increases
economic growth in developing countries when the good macroeconomic policy environment
exists, but Hansen and Trap (2001) show that foreign aid still can play an important role in the
economic growth of those recipient countries even without sound macroeconomic policy
conditionality. The necessity of sound macroeconomic policy and management for the
effectiveness of foreign aid on economic growth of developing countries by increasing domestic
saving and filling the foreign exchange gap is also highly recommended by Tassew (2011) and
Girma (2015). Moreira (2005), Hatemi-J and Irandoust (2005), Adamu (2013) and Basnet (2013)
also argue that the role of foreign aid in the economic growth of developing countries is
significant and foreign aid transfer is necessary to run out of poverty.
The impact of foreign aid on investment and economic growth can be through domestic saving,
which is the main determinant of economic growth and main source of fund for investment
(Hansen and Tarp, 2000) or through income. Foreign aid can enhance the main source of fund for
investment, savings, and also foreign aid can influence investment through an income effect (i.e.
transfer of purchasing power) (Hansen and Tarp, 2000). The importance of sound
macroeconomic policies and management for the effectiveness of foreign aid is not only essential
for the economic growth but also to accelerate the growth of domestic saving which is an
important prerequisite and determinant for capital formation (additions to capital stock) and
8
increasing aid to Sub-Saharan Africa is one way to achieve the Millennium Development Goals
as Armah and Nelson (2008) and Freytag and Voll (2013) argue. In addition to sound
macroeconomic policy environment and management, income level, levels of aid allocation and
geographical location of recipient countries also determine the positive impacts of foreign aid on
economic growth (Durbarry, Gemmel and Greenway, 1998).
The flow of foreign aid from the developed world can have a significant positive impact on
domestic saving and hence promotes investment in the major aid recipient region, Sub-Saharan
Africa, as Balde (2011) argue on his study of foreign aid and domestic saving using ordinary least
squares and instrumental variables estimation method. The results of other studies done in this
area, like “aid effectiveness in Africa” by Loxley and Sackey (2008) and “aid and investment in
least developed countries” by Gyimah-Brempong and Racine (2010) also show that the major
transmission-mechanism in the aid-growth relationship, investment rate, significantly and
positively influenced by foreign aid. This implies that, since investment is equal to saving in the
Harrod-Domar growth model theory and also based on the two-gap model which states that
foreign aid has a potential to fill the saving-investment gap and the foreign exchange gap, foreign
aid also have a positive and significant impact on domestic saving. In addition to filling the two
gaps, foreign aid can play a crucial role in the growth of the country’s economy by creating
access to modern technology and managerial skills, and by allowing easier access to foreign
markets, which have a potential to affect domestic savings directly and indirectly (Irandoust and
Ericsson, 2005).
In his study in 119 aid recipient countries, Michael P. Shields (2007) also tried to see the
crowding out effect between foreign aid and domestic saving by adding value added in
agriculture as a percentage of Gross Domestic Product and labor force as additional control
variables and he confirm that there is a positive relationship between foreign aid and domestic
saving which is in favor of the above arguments that foreign aid can increase domestic saving and
investment. Tolessa (2001), on his study of “Impact of foreign aid on domestic saving,
investment and economic growth”, argue that the influence of foreign aid on domestic saving and
investment is not only at a glance, but also its impact may depend on the type of aid modalities,
and hence, foreign grant has a negative effect while loan has a positive impact on domestic
saving and investment.
9
In the estimation of foreign aid and domestic saving, important variables like investment rate,
which is the main transmission mechanism to the economic growth, are often didn’t used and
hence, the estimated coefficients of aid generated from the statistical estimations may suffer from
omitted variable bias. Through this transmission mechanism, investment, which is equal to saving
under Harrod-Domar growth theory, foreign aid has been beneficial to African countries’
economic growth through that saving and investment even if more investigations are necessary to
ensure that these benefits lead to sustainable growth since economic growth is a result of growth
of different indicators (Girma, Gomannee and Morrissey, 2005). Based on their statistical
estimation using ordinary least square regression with an autoregressive model, Eregha and
Irugha (2009) argue that the role of foreign aid for the growth of aggregate domestic savings was
very important in the long run and short run in Nigeria even if debt service payment have a
negative impact.
In sharp contrast with the argument about foreign aid effectiveness and its role in increasing
domestic saving and promoting growth, whether under sound macroeconomic policy and
management or not, some researchers found negative impact of foreign aid flow on domestic
saving which influence economic growth. Easterly, Levine, and Roodman (2003) found that there
is no real evidence to support the argument given by Burnside and Dollar (2000) since the results
obtained are not robust when different measures of foreign aid, policies, and growth are used.
Kalyvitis (2007) also strengthens their idea that foreign aid may become the main source of fund
for those rent seeker governments of developing countries and will hurt economic growth by
distorting individual incentives and reducing domestic savings since those rent seeker
governments are incapable and irresponsible to mobilize domestic resources. Rather than playing
an important role in the economic growth of recipient countries, the money around $1 trillion
transferred from developed countries to developing countries for the last more than 60 years to
finance development related activities has trapped many African nations in corruption and it
slows down the economic growth, and cutting of the aid flows would be more beneficial than
continuing its flow (Moyo, 2009).
In particular for bilateral aid, Moyo (2009) is totally against it since government to government
aid only makes the developing country's government not to be responsible for their citizens,
10
rather being loyal for donors, and leads to stagnant poor economic performance and aid
dependent. This implies that those governments will not generate domestic resources rather being
dependent on external resources, and hence the domestic saving (particularly public saving) will
decrease. Moyo (2009) also argues, only humanitarian aid should continue and the aid for NGOs
should be for a short period of time and in a strong control for specific objectives. In addition to
corruption, effectiveness of foreign aid in Africa is also determined by conflict, fractionalized
society and dependence on primary commodities (Collier, 2006). He recommended those donors
that in addition to increasing their aid flows, they should also focus on security, good governance,
temporary trade preferences, like AGOA (African growth and Opportunity Act), and conditioning
aid on good governance rather than policies.
The ineffectiveness of foreign aid in Sub-Saharan African countries for the last more than thirty
years is because of diversion of the aid flows to reverse flows (debt service payment, finance
capital flight, accumulates reserves), and this makes foreign aid flow ineffective and lack
appropriate response for the desired investment and economic growth by filling the domestic
savings-investment gap (Serieux, 2011). As Serieux (2011) argue, from 1980-2006 nearly 50 %
of the aid flows spent to finance reverse flows in that undeveloped region. This unrecognized and
unacceptable way of foreign aid spending limits the impact of the incremental foreign aid flows
on domestic saving and investment (Serieux, 2009). The study done by Boyce and Ndikumana
(2012) in the thirty three Sub-Saharan African countries supports the arguments made by Serieux
(2009) that Sub-Saharan Africa is the source of largest capital flight even now during relatively
high economic growth and for the last forty years (from 1970 to 2010) thirty three countries from
the region lost 814 billion USD which exceeds the external liabilities of this group of countries.
Dutch disease, which is the appreciation of real exchange rate due to the flow of foreign aid, is
the other problem faced by those aid recipient countries, and this decreases country's
competitiveness in the international market (Rajan and Subramanian, 2011). In his study of
foreign aid, domestic saving and growth in South Asia, Basnet (2013) argue that even if foreign
aid has positive impact on growth during the study period (1960-2008), in the very long run, it
has a negative impact on domestic saving and hence it offsets the positive impact on growth in
the study period. Depending on theory, which says the investment capacity of developing
countries is limited by the entrepreneurial stock, Taslim and Weliwita (2000) argue that even if
11
there is a huge amount of aid flow to developing countries, since those countries lack sufficient
entrepreneurial skill to invest, that huge aid flow will not spend in the right way for the right
purpose, and hence the relationship between foreign aid and domestic saving is inverse.
The effectiveness of foreign aid, in its role in increasing and promoting domestic saving and
investment, may also depend on the aid composition and source, whether from bilateral sources
or multilateral. As of the cross-country estimation on foreign aid (bilateral and multilateral aid)
and domestic saving done by Nushiwat (2007), the impact of bilateral aid on domestic saving is
positive and significant, but there is a negative impact from multilateral aid. As Nushiwat (2007)
argue, in most cases, multilateral organizations come to deliver their aid during poor economic
and political conditions, natural disasters, civil wars, and low saving, and at that stage, economic
and saving growth is not expected. The argument of McGillivray (2009) is also in favor of
Nushiwat (2007) that, the multilateral aid is more sensitive to be fungible and results in a
decrease of domestic tax generation and public sector savings since the aid recipient government,
especially in least developing countries like Sub-Saharan Africa, may depend on external
sources and will not concern about its citizens. The inflow of foreign aid to developing countries,
specifically in Sub-Saharan Africa, where the study was done, can be a substitute for domestic
saving rather than being an addition when the government lack fiscal discipline and use
international resources as a source of revenue and expenditure. This decreases the ability of the
government to generate domestic revenue and even contribute to fiscal deficit and decreases the
domestic saving (Mallik, 2008). Foreign aid has a significant positive impact for the economic
growth of Pakistan by increasing saving and investment in a sound macro-economic policies and
institutions and when disaggregated the aid in two bilateral and bilateral, bilateral aid is
significantly positive in the short run and multilateral aid is insignificant (Javid and Qayyum,
2011).
On the contrary, Alvi and Senbeta (2012) argue that, since mostly the flow of multilateral aid is
less vulnerable to political pressures and focused on poverty reduction developmental strategies
and goals, multilateral aid and grant aid do better and more effective than bilateral aid and loan
to increase domestic saving and investment and hence to reduce poverty in developing countries.
McGillivray et al. (2004) also argue since multilateral aid has a greater focus on the property than
12
bilateral aid and if the situations that makes foreign aid fungible reduced through monitoring or
other controlling mechanisms, multilateral aid is more effective than bilateral aid for increasing
domestic saving and hence to increase investment. This shows that the inconsistency of results
and disagreement between scholars about the impact of foreign aid on domestic savings goes to
not only foreign aid at a glance but also on the source and composition of foreign aid.
1.2.1. Brief Summary of some articles on Domestic Saving and Foreign Aid
Table 2.1 below shows the authors, the research topic, the methodology they used, the area of the
research and the results obtained by the researchers to summarize the above mentioned scientific
articles on the relationship between foreign aid and domestic saving in aid recipient countries
since 2000.
Table 1: Brief Summary of scientific articles on domestic saving and Foreign Aid
No. Author (year) Research Topic Methodology Area of Research Result
1 Basnet (2013) Foreign aid, Domestic
savings and Economic
Growth
Simultaneous
Equation System
(Growth and Saving
Equations)
South Asia
(Bangladesh, India,
Nepal, Pakistan and
Sri Lanka)
Foreign Aid Affects
Economic growth Positively
but Domestic Saving
negatively
2 Alvi and Senbeta (2012) Does Foreign Aid Reduce
Poverty?
Dynamic Panel
Data Estimation
techniques
100 developing
countries
Multilateral aid is more
significant for domestic
saving and economic
growth since it face less
political pressure and it
mostly focuses on Poverty
reduction strategies than
bilateral aid
3 Balde (2011) The Impact of Remittances
and Foreign aid on
Savings/Investment
Ordinary Least
Square (OLS) and
Instrumental
Variables (2SLS)
Sub-Saharan Africa Foreign Aid has Positive
and Significant impact on
Saving and Investment
4 Javid and Qayyum (2011) Foreign Aid and Growth
Nexus in Pakistan: The Role
of Macroeconomic Policies
ARDL
cointegration
Approach,
Pakistan Foreign aid has a positive
impact for growth through
investment and saving
under sound
macroeconomic policies
5 Serieux (2011) Aid and Resource
Mobilizations: The Role of
reverse flows
Pooled Mean Group
(PMG) estimator
Sub-Saharan Africa Aid flow spent for financing
of reverse flow (debt
service payment, finance
capital flight and
accumulate reserves)
6 Gyimah-Brempong and
Racine (2010)
Aid and Investment in
LDCS: A Robust Approach
Panel Data and
Local Linear Kernel
Estimator (LLKE)
Least Developed
Countries
Foreign aid has a positive
impact on physical
investment
7 Eregha and Irugha(2009) An empirical Analysis of Time series Both in short run and long
13
the long run and short run
impacts of Aid on Domestic
Saving
Analysis (OLS with
an autoregressive
model)
Nigeria run foreign aid affects
domestic saving positively
8 McGillivray (2009) Aid, Economic Reform, and
Public Sector Fiscal
Behavior in Developing
Countries
Fiscal response
model
Philippines Multilateral aid is more
sensitive to be fungible and
has no significant impact
rather bilateral aid is better
9 Serieux (2009) Aid and Savings in Sub-
Saharan Africa: should we
worry about rising aid
levels?
Panel Data Analysis 29 Sub-Saharan
Africa countries
Aid flow spent for the
finance of reverse flow and
consumption and hence
decrease saving
10 Loxley and Sackey (2008) Aid Effectiveness in Africa Panel (Fixed Effect
growth model
estimation) data
analysis
40 AU member
countries
Aid increases the major
transmission mechanism in
aid-growth relationship-
Investment
11 Mallik (2008) Foreign Aid and Economic
Growth: A Cointegration
Analysis of the Six Poorest
African Countries
A Cointegration
Analysis
Six Poorest African
Countries
Foreign aid can be
substituted for domestic
saving rather than
increasing it and reduces the
ability of domestic resource
mobilization.
12 Nushiwat (2007) Foreign Aid to Developing
Countries: Does it crowd
out the recipient countries
Domestic Saving?
Multivariate
regression
Developing countries Impact of Aid may depend
on its sources and hence
bilateral aid has a positive
impact while multilateral
aid has negative impact on
domestic saving
13 Shields (2007)
Foreign Aid and Domestic
Saving: Crowding-out
Effect
Ordinary Least
Squares regression
119 aid recipient
countries
Foreign aid is beneficial for
domestic saving and
investment, and crowding
out effect does not appear
as a common problem
14 Girma, Gomannee and
Morrissey (2005)
Aid and Growth in Sub-
Saharan Africa: Accounting
for Transmission
Mechanisms
Panel Data Analysis 25 Sub-Saharan Foreign aid increases
economic growth through
transmission mechanism
(Investment)
15 Irandoust and Ericsson
(2005)
Foreign Aid, Domestic
Saving and Growth in LDCs
Likelihood based
Panel Co-
integration
African Countries Foreign aid can supplement
domestic saving and fill the
exchange gap, to foster
economic growth
16 McGillivray, Feeny, and
White (2004)
Multilateral Development
Assistance: Good, Bad and
Just Plain Ugly
Statistical
Description about
Multilateral Aid
Developing countries Under sound monitoring
and control to reduce
fungibility, Multilateral aid
is more significant for
developing countries
17 Tolessa (2001) Impact of Foreign aid on
domestic saving, investment
and growth
Times series
analysis
Ethiopia Loan has positive impacts
and the grant has negative
impact on domestic saving
18 Taslim and Weliwita (2000) The inverse relation
between saving and aid: An
Alternative Explanation
Co-integration
Analysis using time
series data
Bangladesh Inverse relation between aid
and domestic saving
14
The reviewed literatures on the topic verify that there is disagreement among the researchers on
how the impact of foreign aid on domestic saving and economic growth looks like and there is
also inconsistency of results, as one can see from the above table. The political, economical and
social difference between sample countries, availability of data and difference in the use of
control variables (like per-capita income, agricultural value added, dependency ratio, financial
development) are may be the sources of the difference in the arguments of researchers in the area.
The political, economical and social situation of Sub-Saharan African countries is very different
and even the features of those countries may differ from other countries in another region like
Asian or Caribbean countries. The quality of macroeconomic policy and economic institutions,
which may have a potential to influence the effectiveness of foreign aid in increasing the
domestic saving and investment to promote economic growth, also differ from country to country
and from region to region. The use of different estimation methodologies and different additional
control variables may be the second possible reason for the inconsistency of results and
disagreement. The estimation results from Ordinary Least Squares regression may not be the
same with other more advanced econometric models OLS takes into account different
assumptions like; no hetroschedasticity and endogeniety problem, and even the time frame that
researchers used may also affect the result since as the time frame increases, the quality of the
estimation may also increase especially in time series data analysis.
15
CHAPTER TWO
Description of Variables, Methodology and Empirical Model
Specification
The data set includes 40 Sub-Saharan African countries over the period 2002-2013. The sample
countries are selected based on the availability of consistent data within the period for the
variables of interest. The data set starts from 2002 since the sample countries have full data in
each variable starting from 2002, and some countries in the region (including Eritrea, South
Sudan, Somalia, Djibouti, Zambia, and Equatorial Guinea) are excluded because of data
inconsistency.
2.1. Variables of Interest
The dependent variable is the share of gross domestic savings to Gross National Product (GDP),
for which the data are taken from the World Development Indicators (WDI) (World Bank, 2015).
Foreign aid, which is considered as one of the determinants of domestic saving (Serieux, 2011),
can have a positive or negative impact on gross domestic savings. The existing literature points
out that the effect of foreign aid on gross domestic savings is inconsistent. Foreign aid may have
a positive impact on domestic saving and promotes investment in Sub-Saharan Africa, the major
aid recipient region (Balde, 2011; Loxley and Sackey, 2008). On the negative side, aid can be
fungible when it is misused/ misallocated and may generate wasteful rent seeking activities by
empowering irresponsible politicians. Hence, the impact of foreign aid can be negative or
positive, and the basic objectives of this paper is to see the impact of foreign aid, at a glance and
by disaggregating foreign aid into bilateral and multilateral aid, on domestic saving in the region
(Sub-Saharan Africa). The data for net official development assistance (Net ODA) as a share of
GDP (both at a glance, total net ODA, and in disagregation, bilateral and multilateral aid) are
taken from the World Development Indicators (WDI) (World Bank, 2015). The data for
multilateral aid is calculated by summing up the aid from UN agencies, World Bank, IMF and
African Development Bank in the sample period and for sample countries.
16
Since the growth of Sub-Saharan Africa gross domestic savings is not only determined by official
development assistance, some additional explanatory variables, which can have a potential
impact on gross domestic saving in the region, are taken into consideration in the statistical
estimation. Out of those factors, some of them can be: agricultural value added, the growth rate of
per capita GDP, good governance and unemployment. Most of Sub-Saharan African countries
economy is mostly depend on agriculture, and as Shields (2007) and Tiffin and Irz (2006) also
emphasized the importance of agriculture and being as engine for economic growth in developing
countries; Value added agriculture is necessary to move the economy forward since it enables to
maintain food security, and it provides raw material, capital and foreign exchange (Tiffin and Irz,
2006), and also it tend to enhance domestic saving and hence increase investment and economic
growth (Shields, 2007). Hence, to see the impact of value added agriculture, which is the net
output of agricultural outputs from all sub-sectors of agriculture including natural resources
(World Bank, 2015), on domestic savings in Sub-Saharan Africa, it is taken as an additional
explanatory variable in the estimation. The data for share of value added agriculture to GDP are
taken from the world development indicators (WDI) (World Bank, 2015).
The quality of governance (good governance) and unemployment are also other possible
determinants of gross domestic savings considered in this paper. The quality of good governance
can be measured by six broad dimensions of governance indicators; government effectiveness,
corruption, rule of law, political stability and absence of violence and regulation quality (WGI,
2015), and these indicators may influence the growth of domestic saving and investment. In most
of developing countries, especially in Sub-Saharan Africa and Middle East, governments are
ineffective, arbitrary, irresponsible and autocratic, and such kind of political underdevelopment is
a major cause of low level of domestic resource mobilization (domestic savings) and poverty in
those countries (Moore, 2001). Weak fiscal and financial policies, macroeconomic instability,
low level of financial infrastructure development, corruption, weak institutional capacity,
including ineffective and incapable tax administration, lack of property rights, and also capital
flight, which are the main features of lack of good governance, are among the obstacles which
adversely affect domestic savings (domestic resource mobilization) in Sub-Saharan Africa (UN,
2005; Mubiru, 2010; Culpeper, 2010). This lack of good governance, which may exist due to
Irresponsible and unaccountable aid dependent governments in Sub-Saharan Africa countries,
17
may reduce domestic savings (particularly public saving) and hence appropriate public services
and investments will not delivered to not only for the current generation but also for next
generation since good governance is essential for sustainable economic growth and development
by enabling the countries to mobilize their own domestic resources (Clark, 2012). The data for
good governance is calculated as the average of the six broad indicators of good governance
based on the data form Worldwide Governance Indicators (WGI) (World Bank, 2015).
High and persistent unemployment is a negative phenomenon in any human society since it
affects the economy and society in different dimensions and directions (Al-habees and Abu
Rumman, 2012), and hence it will decrease the saving and investment since consumption will
increase more than the income generation due to high unemployment. The negative consequence
of high unemployment rate is not only in the economic wellbeing of individuals but also on the
federal budget of the government and hence, it has a potential to affect the level of public savings
and also investment (Levine, 2013). The data for unemployment (as percentage of total labor
force) are taken from the world development indicators (WDI) (World Bank, 2015).
The other variable of interest as an explanatory variable is per capita GDP growth. As of Mohan
(2006) and Mousavi and Monjazeb (2014) the growth of per capita GDP is one of the
determinants of gross domestic savings and it has a positive impact on gross domestic savings
and increases the growth of investment in developing countries. This is in a sense that, when the
economy of countries grows, their GDP also grows and hence the per capita GDP growth rate
also increases if the midyear population doesn’t change or the increase is less than the increase of
GDP. Hence the income of citizens will also increase, which is the main source for the domestic
saving. The data for Gross National Product Per capita growth (annual growth rate) is taken from
World Development indicators (WDI) (World Bank, 2015).
2.2. Hypotheses
Since there is no consistent implication on the impact of official development assistance, as a
general and in disaggregation, impact of multilateral and bilateral foreign aid on gross domestic
savings, as existing literatures imply, there are three main hypotheses to be statistically tested
based on the given data and given methodology;
18
Hypothesis 1: The role of net official development assistance (ODA) for the growth of gross
domestic savings in the aid recipient countries is significant and crucial to alleviate poverty, as of
Shields (2007), Loxley and Sackey (2008), Balde (2011) and other pro-aid researchers’ argument.
In contrast with this argument, some researchers like Moyo (2009), Serieux (2009) and Serieux
(2011), flow of official development assistance to developing countries, particularly to Sub-
Saharan Africa countries, has been spent for financing reverse flows and creates irresponsible and
unaccountable governments for their citizens, and hence, it slows the growth of their economy
rather than promoting development and being a catalyst for growth by enable countries to fulfill
their saving-investment and foreign exchange gap. This implies that flow of official development
assistance will decrease gross domestic savings by deteriorating domestic resource mobilization
capacity of recipient governments. Thus, the null and alternative hypotheses in this case are:
H0: β >0; and Ha: β < 0, where β is the coefficient of total net official development assistance
received as a share of GDP.
Hypothesis 2: Net Official Development Assistance from DAC member countries and the
European Union (Bilateral aid) affects the gross domestic savings positively in Sub-Saharan
Africa countries based on the argument of Javid and Qayyum (2011), McGillivray (2009) and
Nushiwat (2007) that bilateral aid is more effective and has a positive significant impact on
domestic saving and hence increase investment. Hence, the null hypothesis is:
H0: β >0; where β is the coefficient of total net official development assistance from bilateral
sources as a share of GDP.
The alternative hypothesis is the opposite of the null hypothesis that bilateral aid may have a
negative or insignificant impact on gross domestic saving since it is less focused on property and
poverty oriented developmental strategies and highly determined by the political situation of
countries (Alvi and Senbeta, 2012; Mallik, 2008). The alternative hypothesis is:
Ha: β < 0
Hypothesis 3: The impact of multilateral aid from international organizations (mostly from
United Nation Agencies, World Bank, International Monetary Fund and African Development
Bank) on gross domestic savings can be positive since it is more focused on poverty reduction
19
strategies and has a greater focus on property and less political pressure than bilateral aid, as Alvi
and Senbeta (2012) argue. McGillivray et al. (2004) also support the positive and more
significant impact of multilateral aid on domestic saving than bilateral aid under a condition of
controlling and monitoring the implementation of projects and programmes, and existence of
sound economic policies.
On the contrary to the argument about the positive impact of multilateral aid on gross domestic
savings, Nushiwat (2007) and McGillivray (2009) argue that multilateral aid is less effective and
may have even negative impact since it is more sensitive and vulnerable to fungbility and
corruption than bilateral aid since the political pressure and control in developing countries is
less. Hence, the null hypothesis for the impact of multilateral aid on gross domestic saving is:
H0: γ > 0; and the alternative hypothesis is; Ha: γ < 0, where γ is the coefficient of total net
official development assistance from multilateral sources as a share of GDP.
2.3. Methodology and Empirical Model Specification
To see the impact of net official development assistance (ODA) as a total and by disaggregated
into bilateral and multilateral aid, value added agriculture, good governance, unemployment and
per capita GDP growth in forty countries of Sub-Saharan Africa, based on the data obtained from
World development indicators (WDI) and Worldwide Governance Indicators (WGI) (World
Bank, 2015), Simple Panel data analysis is used. The statistical estimation test is undertaken with
fixed effects and without fixed effects (i.e. random effects) model specifications. Fixed effect
estimation, which can be entity fixed effect or time fixed effect or both, is used to investigate the
relationship between the dependent variable and the explanatory variables within an entity
(country, person, etc.) (Torres-Reyna, 2007). This estimation technique assumes two basic things,
as (Torres-Reyna, 2007) argue; the first assumption is that time-invariant variables for each
individual are unique and are not correlated with other characteristics of individuals, and the
second one is non-correlation of error term and the constant (which captures the individual
characteristics) with others. Fixed effect model estimation used to control the omitted variable
bias and also to control the effects of time-invariant variables with time-invariant effects
(Williams, 2015). Since the variables may differ from country to country or from time to time,
20
the fixed effect can be due to the time-invariant country specific fixed effects or can be from
time-invariant time fixed effects (over time effects). To see these country specific and time
specific fixed effects, there are two separate model specifications; country fixed effects, and
country and time fixed effects models. The equation for the country fixed effects model becomes: