International Journal of Multidisciplinary Research and Growth Evaluation www.allmultidisciplinaryjournal.

com

Infrastructure, institutions and economic growth in Sub-Saharan Africa: New evidence

using the CS-ARDL approach
Eyitayo Oyewunmi Ogbaro 1*, Bayode Aluko 2, Julius Olaniyi Adeoye 3
1
Department of Economics, Adeleke University, Ede, Osun State, Nigeria
2-3
Department of Political Science/Public Administration, Adeleke University, Ede, Osun State, Nigeria

* Corresponding Author: Eyitayo Oyewunmi Ogbaro

Article Info Abstract

This study examined the interactive effect of infrastructure and quality of institutions
on economic growth in Sub-Saharan Africa over the period 1996 and 2020 using the
ISSN (online): 2582-7138 Cross-Sectional-Autoregressive Distributed Lag estimator. This was with a view to
Volume: 04 providing additional macroeconomic evidence that is specific to the SSA region on
Issue: 02 the response of economic growth to infrastructure development contingent on the
March-April 2023 quality of the underlying institutions. Annual data on variables such as real Gross
Received: 04-03-2023; Domestic Product (GDP) per capita, fixed telephone subscriptions, mobile cellular
subscriptions, electric power consumption, improved water source, improved
Accepted: 22-03-2023 sanitation facilities, gross capital formation per capita, labour force, urbanisation ratio,
Page No: 373-382 land area and population density were obtained from the World Bank Development
Indicators (WDI) of the World Bank, 2021 edition. Also, data on institutional quality
variables, namely, control of corruption, government effectiveness, political stability
and absence of violence/terrorism, regulatory quality, rule of law, as well as voice and
accountability were sourced from the World Governance Indicators (WDI) also of the
World Bank, 2021 edition. Results showed that the per capita convergence holds true
for countries in SSA in the long run, while it does not hold in the short run. The
findings also revealed that infrastructure does not independently lead to economic
growth in SSA when the role of institutions is not considered. This implies that any
improvement in the stock of infrastructure will promote growth only when it is
complemented with strong institutions. The study concluded that countries with weak
institutions do not benefit maximally from infrastructure development policies as
weak institutions constrain the efficient use of infrastructure assets.

DOI: https://doi.org/10.54660/.IJMRGE.2023.4.2.373-382

Keywords: Infrastructure, economic growth, CS-ARDL

1. Introduction
In the last one and half decades, the literature on the nexus between infrastructure and growth in developing countries in general
and in Sub-Saharan African (SSA) countries in particular has shifted to the investigation of the factors that moderate the nexus.
This is because efforts at scaling up the stock of infrastructure in these economies have not really translated to the much desired
growth. This experience of developing countries in terms of the growth effect of infrastructure, however, contradicts that of a
country like China. The Chinese economy has been able to record significant growth in the last two decades due to massive
investment in infrastructure, among other factors. This contradiction has, therefore, prompted scholars to search for factors which
are responsible for the differences in the effects among countries. One of the factors that have been touted to be capable of
causing the differences in the returns to infrastructure is institutional quality.
The fact that the infrastructure-growth link is moderated by the quality of institutions has been established in both the theoretical
and empirical literature.

373 | P a g e
International Journal of Multidisciplinary Research and Growth Evaluation www.allmultidisciplinaryjournal.com

In the theoretical literature, scholars have argued that, based ground of its inability to capture cross-section dependence
on their quality, institutional indicators will either limit or and slope parameter homogeneity (Amin et al., 2022; Chudik
enhance the efficient use of infrastructure. For example, Hall, & Pesaran, 2015; Sarafidis et al., 2008; Saygin &
Sobel and Crowley (2010) [17] argued that countries with Iskenderoglu, 2022) [2, 6, 38]. To correct this second limitation,
well-developed institutions, which do not encourage scholars have suggested the use of the Cross-Sectional-
unproductive activities, tend to experience growth when they Autoregressive Distributed Lag (CS-ARDL) estimator.
increase their capital stock. Conversely, investing in Another important advantage of CS-ARDL is its ability to
infrastructure in countries where corruption, rent seeking and yield robust estimates even when available dataset is
other unproductive activities are prevalent may lead to zero relatively small (Amin et al., 2022) [2].
if not negative growth rates. Also, Agénor and Montiel Hence, this study was conducted to investigate the role of
(2010) [1] posited that devoting resources to infrastructure institutions in the infrastructure-growth relationship in SSA
development is a necessary but not sufficient condition for over the period 1996 to 2020 using the CS-ARDL approach.
boosting economic growth. The authors underscored the
importance of strong institutions which play the role of 2. Literature Review
catalyst in the efficiency of capital. Furthermore, Wu, Tang This section reviews the literature on the relationship
and Lin (2010) [44] have attributed the efficacy of government between infrastructure and economic growth, while paying
spending on growth to the institutional quality of the country particular attention to the role of institutions, in line with the
in question. They argued that in low-income countries, which belief among economists and other stakeholders in recent
are usually characterised by poor institutions, government years that “institutions matter”.
expenditures have the tendency to retard growth or become
ineffective. While also supporting this line of thought, Dabla- 2.1 Theoretical Review
Norris et al. (2012) [8] argued that embarking on considerable Available theories regarding the economic impact of
infrastructure development in an environment characterised infrastructure can be reviewed from three fields based on the
by weak institutions has the tendency of potentially different analytical approaches: traditional approach,
undermining its growth benefits. endogenous growth approach, and the new institutional
In SSA in particular, many countries are plagued by poor economics.
maintenance of existing facilities, coupled with wanton
vandalisation and destruction of infrastructure facilities as a A) Traditional Theory
result of high rates of corruption and terrorism. The estimates The traditional neoclassical growth theory, which is largely
arrived at by Gulati and Rao (2006) [16] on the cost of due to the work of Solow (1956) [40], assumes that capital has
corruption in infrastructure suggest that between 5% and 20% diminishing returns. This leads to the condition that there is
of construction costs are being lost to bribe payments, while an inverse relationship between per capita income growths of
as much as between 20% and 30% of electric power is being countries and their initial income levels. If economies are
stolen by consumers who collude with staff. Using similar in the sense of preferences and technology, countries
investment and maintenance estimates from Fay and Yepes that are poor in the level of capital grow faster than capital
(2003) [13], Kenny (2009) [18] found that the financial burden rich countries (Olson, Sarna & Swamy, 2000) [31]. In this
may sum up to around US$18 billion annually in low-income model, technology is assumed to be exogenously given. Thus
countries if as much as 5% of expenditure and maintenance taking the saving and population growth rates as constant, the
outlays in infrastructure are lost to institutional failure. What model predicts that countries would tend to converge into
this suggests is that “better infrastructure (that is, infrastructure similar steady state level. Diminishing returns to physical
development embedded within a sound institutional capital is thus a force that allows countries to converge into
framework), more growth” is a more accurate proposition than income equability when they reach similar steady states level
“more infrastructure, more growth”. Hence, investigating the (Barro et al., 1991; Barro & Sala-i-Martin, 1992) [5, 4].
relationship between infrastructure and growth without Despite its ability to explain international differences as a
paying attention to how institutional factors contribute to this result of conditional convergence, the neoclassical model
relationship as a complementary factor may lead to seriously cannot however explain why the income gap between rich
misleading inferences. and poor countries is still widening (Mankiw, 1995) [21].
Although many studies acknowledge the high a priori Beside this, there is no room for public decisions to have
probability that the linear relationship between infrastructure long-term effects on the economy since exogenous technical
and economic growth is moderated by institutions, only a few progress is the sole determinant of long-run growth in
have subjected this hypothesis to empirical testing in both the traditional neoclassical growth models (Getachew, 2009) [15].
developed and developing countries. Studies that have been A policy shock will have a transitory effect, influencing the
conducted to address this gap in SSA include Damijan and level of (long-run) output only. Thus, an economy’s
Padilla (2014) [9], Esfahani and Ramirez (2003) [11], Garlick institutions and infrastructure development, among others,
(2008) [14], Kodongo and Ojah (2016) [19], as well as Okoh will have no lasting effect on its output. By and large,
and Ebi (2013) [30]. These studies suffer two serious economic agents are assumed to operate almost in a vacuum.
limitations, one of which is their inability to capture recent However, the developments in post-independence Africa told
changes in the trend of infrastructure in particular, especially a different story. Economic growth at that time was
in the last decade. The other one concerns the methodology disappointing. Many of the economies were growing very
used by these studies, especially the few that employed panel slowly and some were even retrogressing. The problem then
data estimation techniques. For example, the System was not attributed to resource constraint but was linked to
Generalised Method of Moments (GMM) adopted by poor policies. Most countries had implemented a
Kodongo and Ojah (2016) [19] has been criticised on the “government led” development paradigm that restricted

374 | P a g e
International Journal of Multidisciplinary Research and Growth Evaluation www.allmultidisciplinaryjournal.com

private sector development. Heavy intervention in the growth models to address the remarkable fact of the
economy in the form of government production, control of contemporary cross-country growth disparities led a large
prices and exchange rates, over-regulation of the economy number of scholars to seek for other fundamental factors that
and the adoption of import substitution strategies that are necessary in explaining economic growth and cross-
included severe import controls and foreign exchange country income disparities. Towards this end, economists
rationing inhibited economic growth. incrementally advanced the notion that, in addition to
government policies, high institutional quality is required to
B) Endogenous Growth Theory bring about higher economic growth.
Due to the less convincing empirical relevance of the The phrase “New Institutional Economics (NIE)” was
neoclassical model, subsequent researches took a more coined by Williamson (1985) [43] in order to distinguish it
radical approach to the study of economic growth. The goal from the “old institutional economics” which was pioneered
of this approach, known as “endogenous growth theory”, is by Veblen (1899) [41] and Commons (1934) [7]. Although the
to develop models of persistent growth that give up the old institutional school regarded institutions as a key factor
assumption of exogenous technological change (Mankiw, in defining and determining economic performance, but it
1995; Weil, 2005) [21, 42]. The approach embraces diverse had little analytical rigour and lacked theoretical framework
body of theoretical and empirical works that emerged in the as a school of thought. Its operation was outside of
1980s, following the path pioneered by Arrow (1962) [3]. neoclassical economics and deriving reliable generalization
Notable contributions include, among others, Romer (1986) or making sound policy choices was impossible because there
[36]
, Lucas (1988) [20] and Robelo (1991) [36]. was no quantitative theory for doing so. Unlike neoclassical
Instead of diminishing returns to capital, endogenous growth economics, a striking feature of the theory is that the
theory assumes constant returns to capital, i.e., doubling institutional framework is not assumed to be exogenous.
capital would double output. Technological change is Instead, it is clearly treated as an object of research such that
assumed endogenous. As a result of the assumption of the way and manner any given institutional arrangements
constant returns to capital, endogenous growth model affect economic behaviour is accorded due consideration
predicts the gap between rich and poor countries remaining (Richter, 2005) [35].
the same or even widening (Lucas, 1988; Mankiw, Romer & This new thinking in economics believes that the cost of
Weil, 1992) [20, 22]. The main contribution of the theory of transacting, which is determined by institutions and
endogenous growth as far as the advancement of economic institutional arrangements, is critical to economic behaviour.
growth is concerned, according to Fanta (2011) [12], is that it It is therefore posited that a country’s institutions, such as its
helps in explaining the existence of worldwide technological legal, political, and social systems, define its economic
process, which is assumed to be exogenous by the performance. The proponents of NIE argue for the need to put
neoclassical model. In this context, infrastructure is seen to in place regulatory mechanisms and other institutional
affect productivity and aggregate output through direct and frameworks in addition to substantive policies so that the
indirect channels. efficient operation and functioning of the “hard” component
This important contribution to knowledge notwithstanding, can be facilitated. The bottom line of this argument is that it
endogenous growth model has succeeded little in explaining is the responsibility of the government to put in place
cross-country income differences because growth is a appropriate institutions that allow the positive effects of
complex phenomenon which cannot be explained by the infrastructure development to reflect on the country’s
theory of the creation of technology alone (Mankiw, 1995) economic performance. This emphasises the proposition that,
[21]
. Although policy reforms yielded some gains in terms of while it is true that inputs and good policy are important
economic growth, these gains were small and far apart. Of components of the economic growth process, it is largely the
particular concern was that for the majority of countries, the quality of institutions that determines wealth accumulation.
instituted reforms had serious adverse impacts on the
wellbeing of the majority of the citizens. Poverty rates tended 2.2 Brief Empirical Review
to increase and also the policies magnified inequalities in the The failure of the traditional economic analysis on the growth
distribution of income. As a result, there were frequent policy effect of infrastructure to capture the mediating role of
reversals resulting in stagnation. In some cases, dissatisfied institutions in the relationship led to the emergence of a new
citizens engaged in violent protests and in others, poor strand. This strand of the literature is preoccupied with the
economic outcomes and the increasing inequality were used need to assess the extent to which institutions affect the
to justify military takeovers. Evidently, many of the policies relationship between the two variables. Studies in this
instituted were not sustainable. Through the 1980s, most category can be broadly divided into two based on their
countries in SSA regressed in many of the dimensions of modelling of the relationship. The first category assumes a
welfare. linear or symmetric relationship, while the second one
assumes a non-linear or asymmetric relationship. The focus
C) New Institutional Economics of this study is, however, limited to the first category of
Following the seminal works of North and Thomas (1973) studies whose analysis is based on the use of a linear
[27]
, Williamson (1985) [43], Matthews (1986) [23], and North interaction model. The model is made up of a linear
(1990) [26], economists have emphasized the instrumentality interaction term between infrastructure and institutions.
of strong institutions to growth (Fanta, 2011) [12]. This is as a One of the studies in the first category in the context of
result of an expanding argument about the inability of countries in SSA is Esfahani and Ramirez (2003) [11]. The
markets alone to ensure economic efficiency. While earlier study examined the role of institutional factors in shaping the
works on growth take the existence of institutions as given, growth effects of infrastructure in 75 countries out of which
more recent works showed the flaw in such approach. In 19 are from SSA between 1965 and 1995. Using instrumental
particular, the failure of both the neoclassical and endogenous variable (IV)/two stage least squares (2SLS) technique, the

375 | P a g e
International Journal of Multidisciplinary Research and Growth Evaluation www.allmultidisciplinaryjournal.com

results showed significant growth effects of infrastructure in regards institutions as “soft” infrastructure, i.e., regulatory
the data set in which the contribution of infrastructure is mechanisms and other institutional frameworks that must be
affected by institutional factors. Also, Garlick (2008) [14] put in place to facilitate the efficient operation and
investigated the relationship among infrastructure, output and functioning of the “hard” component. In countries with good
institutional quality in South Africa using time series data institutions, investments in infrastructure are both privately
over the period 1960-2006. Results of their univariate beneficial to individuals and also create a positive return for
cointegrating autoregressive distributed lag (ARDL) and society as a whole. In countries with poor institutions,
multivariate cointegrating vector error correction model however, the higher returns to investments in rent-seeking
(VECM) estimations revealed that institutions are an activities that plunder the wealth of others, through lobbying
important factor in the infrastructure-output nexus, although and lawsuit abuse, for example—draw significant resources
the exact form of this relationship is not clear. into these privately beneficial, but socially unproductive
In addition, Okoh and Ebi (2013) [30] examined the effect of activities.
the interaction of infrastructure investment and institutional In the absence of strong institutions, the link between
quality in Nigeria using the pair-wise Granger causality infrastructure spending and growth is weakened by evidence
approach. Findings revealed that the effects of the interaction of low efficiency of public investment, and where a high
between infrastructure investment and institutional quality on degree of inefficiency and/or waste, often distorts the impact
economic growth are insignificant. Furthermore, Damijan of infrastructure investment, leading to poorly executed
and Padilla (2014) [9] investigated the impact of various types and/or ineffective projects. The prevalence of corruption in
of infrastructure investment on Gross Domestic Product such a setting reduces the quality of infrastructure investment
(GDP) per capita growth conditional on institutional and its economic benefits, in addition to raising the cost of
advancement and foreign co-financing. The study used panel infrastructure. Awarding public procurement contracts within
data on Egypt, Morocco, Tunisia, Namibia and South Africa the context of a corrupt system may lead to inferior public
for the period 1990-2010 as well as the fixed effects infrastructure and services. This is because when the
approach. Results showed that infrastructure projects are government official saddled with the responsibility of
long-term sustainable in a less corrupt environment. Finally, monitoring infrastructure projects is corrupt, he or she may
Kodongo and Ojah (2016) [19] assessed whether infrastructure allow the use of substandard materials. In weak government
availability and quality act through the mediating effects of quality conditions, therefore, new investment in
institutions to impact economic growth for a panel of 45 SSA infrastructure may respond more to political and individual
countries from 2000 to 2011. Using the system GMM, the interests than to economic and collective ones.
study found that a largely insignificant effect of infrastructure Weak institutions will not only reduce the rate of return to
access and quality on economic growth due to Africa’s new investment in infrastructure, but will also affect the rate
relatively low institutional quality which has made of return to existing stock. A common phenomenon in
infrastructure less effective as a growth catalyst. developing countries where governments are ineffective is
Some important points can be drawn from the brief empirical the poor condition of existing infrastructure (roads with
review above. One, existing empirical evidence on the role of potholes, water and sanitation facilities in bad state, buildings
institutions in mediating the effect of infrastructure on the in need of serious repairs, etc.). More often than not, new
growth of SSA countries is not conclusive. Two, the existing projects are undertaken while the existing ones are
studies did not capture the recent changes in the trend of abandoned to deteriorate as a result of poor maintenance
infrastructure in the selected countries. Finally, the system culture. Also, in cases of very high incidence of corruption,
GMM technique employed by Kodongo and Ojah (2016) [19], operation and maintenance on the physical infrastructure may
whose analysis is close to the one done by this current study, be purposely reduced so that some infrastructures may
has some limitations. The System GMM lacks the capacity to deteriorate quickly to the point that they will need to be built
to capture cross-section dependence and slope parameter again, thus affording some high-level officials the
homogeneity (Amin et al., 2022; Chudik & Pesaran, 2015; opportunity of collecting commission from the enterprise that
Sarafidis et al., 2008; Saygin & Iskenderoglu, 2022) [2, 6, 38]. will handle the project. By and large, we have a situation in
To correct this second limitation, scholars have suggested the which the deterioration in the existing infrastructure retards
use of the Cross-Sectional-Autoregressive Distributed Lag growth more than the rate at which the new ones add to it.
(CS-ARDL) estimator. Another important advantage of CS- From the NIE perspective, therefore, modelling the growth
ARDL is its ability to yield robust estimates even when effects of infrastructure without incorporating the quality of
available dataset is relatively small (Amin et al., 2022) [2]. institutions will yield inconsistent results. Hence, the analysis
This current study was therefore conducted to address the in this study is based on the assumption that the quality of
limitations in the existing empirical literature in the context institutions affects output through the effect that institutions
of countries in SSA. have on the productivity of infrastructure. This is opposed to
the usual practice of implicitly assuming an underlying set of
3. Methodology good institutions. The proposition that emanates from the
This section presents the methodology that is employed to position adopted is that countries with strong institutions
achieve the objectives of the study. Specifically, it presents benefit more from infrastructure development policies as
the theoretical framework, estimating model, techniques of strong institutions enhance the efficient use of infrastructure
analysis, as well as measurements and sources of data. assets.

3.1 Theoretical Framework 3.2 Model Specification

The theoretical basis for the analysis adopted in this study is The linear interactive model that is adopted for the analysis
the New Institutional Economics hypothesis. This view carried out in this study is specified as follows:

376 | P a g e
International Journal of Multidisciplinary Research and Growth Evaluation www.allmultidisciplinaryjournal.com

ln yit   0  1 ln yit 1   2 ln kit   3 (q  ln k )it   4 ln xit   5 (q  ln x)it   6 qit   it (1)

where ln denotes natural logarithm, y is real GDP per 3.4 Definitions of Variables and Sources of Data
The data employed in this study is a panel of forty-one (41)
worker, yt 1 is lagged real GDP per worker, k is physical SSA countries over the period 1996-2020. The list of the
capital per worker, x is infrastructure per worker, q is selected countries is presented in the appendix section.
institutional quality, q  ln k  is the interaction term between Details of the measurement of variables and sources of data
institutional quality and physical capital, q  ln x  is the
used for analysis are shown in Table 1. The table shows that
economic growth, which is the dependent variable, is
interaction term between institutional quality and physical measured using real GDP per capita. Gross capital formation
capital,  the disturbance term is assumed to have two is employed to measure investment in physical capital.
orthogonal components: the fixed effects, i , and the Physical measures of infrastructure rather than monetary ones
are used. Due to data availability, the study considered four
idiosyncratic shocks, v , i denotes country, while t is time out of the five infrastructure sub-sectors (telecommunications,
series observation. electric power, clean water and improved sanitation). Using
In equation (1), emphasis is on the statistical significance of the Principal Component Analysis (PCA) method, the four
the interaction coefficient  5 . Depending on its sign, it can be sub-sectors are built into a synthetic index summarising
inferred whether infrastructure and institutions are complements different dimensions of infrastructure in line with Badalyan
or substitutes in the growth process. A negative coefficient et al (2015), Calderón et al. (2015) as well as Chakamera and
will indicate that infrastructure development is more effective Alagidede (2017).
in boosting economic growth in countries with weak Following Kurul (2017), Helliwell, Huang, Grover, and
institutions (Effiong, 2015) [10]. In other words, a negative Wang (2018), Seth (2018), as well as and Sondermann
interaction (i.e.  5  0 ) provides evidence of substitutability (2018), institutional quality was measured using the
between infrastructure and institutions. On the other hand, a indicators developed by Kaufmann, Kraay and Mastruzzi
(2010). The three authors have provided measures of national
positive interaction (i.e.  5  0 ) would imply that the growth
institutional environments through their work on the World
effects of infrastructure are enhanced in a strong institutional Bank’s Worldwide Governance Indicators (WGI). The WGI
environment, thus supporting the complementarity of have lately become one of the most commonly used
infrastructure and institutions. indicators of governance or institutional quality in empirical
studies undertaken by academics as well as policymakers.
3.3 Technique of Data Analysis The WGI dataset summarizes six dimensions of institutional
The main objective of this study is to examine the interactive quality or governance as follows: Voice and Accountability
influence of infrastructure and institutions on economic (VAC), Political Stability and Absence of Violence/Terrorism
growth in SSA. This objective is achieved by estimating (PSV), Government Effectiveness (GEF), Regulatory
equation (1) using the CS-ARDL in view of its strength over Quality (RQL), Rule of Law (ROL), as well as Control of
other panel data estimators especially the System GMM. Corruption (COC). The scores range between −2.5 and +2.5.
However, the validity of the CS-ARDL depends on the The classification by Omilola and Akanbi (2014) is adopted
existence of cross-sectional dependence and slope parameter in order to categorise institutional quality at different levels.
heterogeneity (Saygın & Iskenderoglu, 2022) [39]. In view of For the purpose of analysis, the study used an indicator of the
this, the study tested the two conditions to confirm the overall quality of institutions by combining all six
validity and reliability of the estimates obtained from the dimensions into a single index using the PCA method.
empirical analysis.

Table 1: Definition of Variables and Sources of Data

Variable Identifier Definition Source
Gross Domestic Product
y it Real GDP per capita (constant 2010 US$) and is expressed in logarithmic form WDI, 2021
(GDP)
Real GDP per capita from previous period (constant 2010 US$) and is in logarithmic
Initial real GDP y it 1 WDI, 2021
form
Constructed as the principal component of fixed telephone subscriptions or main lines
Index of Infrastructure (per 100 inhabitants), mobile cellular subscriptions or mobile phones (per 100 people), Author’s
Stock
xit electric power consumption (kWh per capita), improved water source (% of population Computation
with access) and improved sanitation facilities (% of population with access)
These refer to the sum of active number of analogue fixed telephone lines, voice-over-IP
Fixed Telephone (VoIP) subscriptions, fixed wireless local loop (WLL) subscriptions, ISDN voice-channel
Subscriptions
FTS equivalents and fixed public payphones. Measurement is per 100 inhabitants and is
WDI, 2021
expressed in logarithmic form
These measure subscriptions to a public mobile telephone service that provide access to
Mobile Cellular
Subscriptions
MCS the PSTN using cellular technology. Measurement is per 100 people and is expressed in WDI, 2021
logarithmic form
This measures the production of power plants and combined heat and power plants less
Electric Power
Consumption
EPC transmission, distribution, and transformation losses and own use by heat and power WDI, 2021
plants. Measurement is in kWh per capita and is expressed in logarithmic form
Access to an improved water source refers to the percentage of the population using an
Improved Water Source IWS improved drinking water source. Measurement is % of population with access and is
WDI, 2021

377 | P a g e
International Journal of Multidisciplinary Research and Growth Evaluation www.allmultidisciplinaryjournal.com

expressed in logarithmic form

Access to improved sanitation facilities refers to the percentage of the population using
Improved Sanitation
Facilities ISF improved sanitation facilities. Measurement is % of population with access and is WDI, 2021
expressed in logarithmic form
Computed as the log of the principal component of control of corruption, government
Index of Institutional Author’s
Quality
q it effectiveness, political stability and absence of violence/terrorism, regulatory quality, rule
Computation
of law, and voice and accountability
This captures perceptions of the extent to which public power is exercised for private
Control of Corruption COC gain, including both petty and grand forms of corruption, as well as “capture” of the state WGI, 2021
by elites and private interests
This captures perceptions of the quality of public services, the quality of the civil service
and the degree of its independence from political pressures, the quality of policy
Government Effectiveness GEF formulation and implementation, and the credibility of the government’s commitment to
WGI, 2021
such policies
Political Stability and
This measures perceptions of the likelihood of political instability and/or politically-
Absence of PSV motivated violence, including terrorism
WGI, 2021
Violence/Terrorism
This indicator captures perceptions of the ability of the government to formulate and
Regulatory Quality RQL implement sound policies and regulations that permit and promote private sector WGI, 2021
development
This measure captures perceptions of the extent to which agents have confidence in and
Rule of Law ROL abide by the rules of society, and in particular the quality of contract enforcement, WGI, 2021
property rights, the police, and the courts, as well as the likelihood of crime and violence
This measure captures perceptions of the extent to which a country’s citizens are able to
Voice and Accountability VAC participate in selecting their government, as well as freedom of expression, freedom of WGI, 2021
association, and a free media.
Gross capital formation per capita (constant 2010 US$) and expressed in logarithmic
Physical Capital PCP form
WGI, 2021
Labour Force LAB Total Labour Force WDI, 2021
Urbanization Ratio UBR Percentage of urban population in the total WDI, 2021
Land Area LAR Land area in sq. km. WDI, 2021
Population Density POD People per sq. km of land area WDI, 2021
Notes: WDI denotes World Development Indicators, while WGI denotes World Governance Indicators.

4. Results and Discussion This is a characteristic of right skewed distributions. The

This section presents empirical evidence, by means of mean value of the infrastructure index (-0.335328) is similar
statistical and econometric analysis, on the interactive to the ones obtained by Ogbaro (2019) [28] as well as Ogbaro
influence of infrastructure and quality of institutions on and Oladeji (2021) [29].
growth. The summary statistics of the index of institutional quality is
very close to the one obtained by Ogbaro (2019) [28]. On a
4.1 Descriptive Statistics of Data scale of 0 (lowest) to 100 (highest), the study reported weak
Before proceeding to the econometric analysis, this study institutional quality with an average score of 39. This study
examined the descriptive statistics of the main data used for recorded a mean of 0.38 (38 on a scale of 0 to 100) which still
analysis. This is necessary as it provides information on the suggests weak institutional quality across the region.
characteristics of the variables used in the study and hence, The results of the skewness statistic show that real GDP per
the credibility and quality of the data. Table 2 displays the capita, initial real GDP per capita, physical capital and index
descriptive statistics of the main variables used in the of institutional quality are positively skewed (their skewness
analysis. values been greater than zero). On the other hand, the index
From the table, it is observed that all the mean and median of infrastructure stock is negatively skewed (its skewness
values are positive except for the index of infrastructure value been greater than zero).
stock. The table reports that while the average real GDP per Results of the kurtosis statistic show that while real GDP per
capita is US$2,303.45, the median real GDP per capita is capita, initial real GDP per capita and physical capital are
US$930.57. This means that half the countries recorded real leptokurtic, the indices of infrastructure stock and
GDP per capita less than US$945.82 and half recorded more. institutional quality are platykurtic.
However, mean GDP is significantly higher than the median As for the Jarque-Bera test statistic, the results in Table 2
GDP and so, most of the countries recorded less than the reveal that the calculated probability values of the statistic for
average. The same also goes for each of initial real GDP per all the variables except index of infrastructure stock are less
capita, physical capital as well as the index of infrastructure than the 5% significance level which suggests the rejection of
stock. The index of institutional quality is the only variable the null hypothesis of a normal distribution for the variables.
that has a mean value that is very close to the median value.

378 | P a g e
International Journal of Multidisciplinary Research and Growth Evaluation www.allmultidisciplinaryjournal.com

Table 2: Descriptive Statistics

GDPPC LAGGDPPC PCP INF IIQ
Mean 2303.452 2175.885 797.9548 -0.335328 0.373395
Median 945.817 911.2192 173.9937 -0.280144 0.372747
Maximum 21443.98 21276.51 18123.49 2.942351 0.561594
Minimum 197.7725 170.5817 3.123834 -5.767763 0.154917
Standard Deviation 3262.968 3215.426 1831.973 1.632676 0.104179
Skewness 2.711868 2.668671 4.381622 -0.159120 0.352229
Kurtosis 11.59217 11.28744 25.79509 2.863532 2.6749489
Jarque-Bera (p-value) 0.000000 0.000000 0.000000 0.162752 0.000108
Observations 1025 1025 1025 1025 1025
Source: Author’s computations (2022) based on WDI and WGI of the World Bank (2021).
Notes: GDPPC represents real GDP per capita; LAGGDPPC represents real GDP per capita from previous period; PCP represents physical
capital per capita; INF represents index of infrastructure stock; and IIQ represents index of institutional quality.

4.2 Cross-Sectional Dependence, Heterogeneity and under cross-sectional dependence. Based on the results in
Order of Integration Table 3, the Cross-Sectional Augmented Im–Pesaran–Shin
Prior to estimating equation (1) using the CS-ARDL (CIPS) test proposed by Pesaran (2007) was employed for
estimator, this study carried out some important tests for the this purpose. The results of the test in Table 5 shows that the
purpose of obtaining estimates that are reliable. The first variables are of mixed orders, i.e., they are integrated of
among these tests is the cross-section dependence (CD) test, orders zero and one. Thus, these results as well as the findings
which also helps to choose the unit root test that is appropriate of the presence of CD and homogeneity tests confirm that the
for determining the orders of integration of the series used for CS-ARDL approach is the appropriate tool for estimating
analysis. In lie with extant studies such as Nathaniel (2021), equation (1).
the CD tests proposed by Pesaran (2004) was employed for
this purpose. The results of the test, which are reported in Table 5: Panel Unit Roots Test Results
Table 3, show the existence of cross-sectional dependence in Variable CIPS (Level) CIPS (1st Difference)
each series. lnGDPPC -1.489 -5.327***
lnLAGGDPPC 0.091 7.318***
Table 3: Cross-Sectional Dependence Test Results lnPCP 1.539 5.184***
Breusch–Pagan Pesaran scaled Pesaran lnINF 2.103** -
Variable IIQ 1.047 10.824***
LM LM CD
lnGDPPC 559.8701*** 50.4327*** 20.5748*** Source: Authors’ computations (2022) based on WDI and WGI of
lnLAGGDPPC 474.0912*** 41.8361*** 16.9754*** the World Bank (2021).
lnPCP 855.8923*** 70.5290*** 7.8934*** Notes: ln denotes natural logarithm, GDPPC represents real GDP
lnINF 1421.6267*** 122.6984*** 35.4512*** per capita; LAGGDPPC represents real GDP per capita from
IIQ 505.6634*** 67.9523*** 24.9030*** previous period; PCP represents physical capital per capita; INF
represents index of infrastructure stock; and IIQ represents index of
Source: Authors’ computations (2022) based on WDI and WGI of
the World Bank (2021). institutional quality. *** and ** denote statistical significance at 1%
and 5% levels, respectively.
Notes: ln denotes natural logarithm, GDPPC represents real GDP
per capita; LAGGDPPC represents real GDP per capita from
previous period; PCP represents physical capital per capita; INF 4.3 Interactive Influence of Infrastructure and
represents index of infrastructure stock; and IIQ represents index of Institutions on Growth
institutional quality. *** denotes statistical significance at 1%. Table 6 presents the results obtained from the estimation of
equation (1) using the CS-ARDL estimator. The table shows
The second test is the homogeneity test which is required for that the error correction term (ECT) is negative and
determining the appropriate unit roots and cointegration significant. This implies that the system returns to
estimator. Following Sanyin and Indes (2022), the Delta Test equilibrium in case of a shock that causes a disequilibrium. It
proposed by Pesaran and Yamagata (2008) was employed for also implies that a stable cointegrating relationship exists
this purpose. The results of the test, which are presented in among the variables in the long-run. More specifically, the
Table 4, show that economic growth is homogenous among coefficient implies that the short-run speed of adjustment
the selected countries. towards the long-run equilibrium is 67.2% per year, which is
equivalent to about one and half years.
Table 4: Delta Test for Slope Heterogeneity The coefficient of LAGGDPPC is negative and significant
Coefficient p -value in the long run, implying that per capita convergence holds
true in SSA. This finding is consistent with the prediction of
 1.768 0.127
Mankiw et al. (1992) [22] that the SSA region will form a
 adj. 1.8562 0.116
homogenous set of countries in many respects particularly in
 (HAC) -1.372 0.193 the long run. In contrast, the coefficient of the series is
 adj. -1.438 0.188 positive and significant in the short run, implying that per
Source: Authors’ computations (2022) capita convergence does not hold in SSA. This finding is not
surprising because, contrary to the prediction of Mankiw et
The third test is the panel unit roots test which was conducted al. (1992) [22], the SSA does not form a homogenous set of
to investigate the stationarity of homogenous panel series countries in many respects. Countries in the region vary

379 | P a g e
International Journal of Multidisciplinary Research and Growth Evaluation www.allmultidisciplinaryjournal.com

strongly in terms of natural endowments, educational system, In countries where such behaviour is not constrained by
macroeconomic targets, population structure, political institutional frameworks, there exists the possibility of
administration etc. All these factors tend to affect their per rendering fixed capital formation an unreliable determinant
capita income growth. As a result, the convergence of growth. For instance, the prevalence of corruption may
hypothesis as predicted by Mankiw et al. (1992) [22] may not bias commercial and public sector decision-making
apply to SSA in the short run. Overall, these findings indicate processes, leading to investment decisions that are relatively
that, while countries in SSA can show individual differences unproductive.
in the short run, they tend to converge to form a bloc in the The coefficients of ln INF and IIQ  ln INF are both
long run. positive, but only the latter is significant in the long run. This
The coefficients of ln PCP and IIQ  ln PCP are positive and implies that infrastructure has no significant effect on growth,
negative, respectively and significant in the long run. This and institutional quality lacks the necessary stimulus to make
implies that physical capital exerts a positive effect on it significant. Similarly, the two series are positive, and only
growth, but institutional quality acts as a drag that leaks out the latter is significant in the short run. This means that
the growth benefits of capital on growth. In contrast, the institutions and infrastructure are complements and any
coefficients of the two series are both negative and weakly improvement in the stock of infrastructure will promote
significant in the short run. This implies that physical capital institutional quality and vice versa. As a result of this, one
constitutes a drag to the growth process, while institutional can say that infrastructure is as important in explaining per-
framework aggravates this negative effect. These results are capita and growth differentials as institutions.
not surprising given that only better institutions increase the The coefficient of institutional quality is positive in both the
contribution of physical capital to output. Hall et al. (2010) long and short run, and the effect is significant in the short
[17]
explained that increases in capital retard growth in run only. This implies that the variable does not affect growth
countries characterised by weak institutions because in the long run. Rather, the effect of institutions on output in
additions to the capital stock tend to be devoted to rent- the long run is entirely captured by its effect on the
seeking and other activities that are not socially productive. productivity of infrastructure.

Table 6: CS-ARDL Estimation Results

Variable Breusch–Pagan LM
Long-Run Result:
lnLAGGDPPC -0.995** (0.003)
lnPCP 0.023** (0.006)
IIQ*lnPCP -0.063** (0.011)
lnINF 0.106 (0.026)
IIQ*lnINF 0.244** (0.055)
IIQ 0.004 (0.029)
Short-Run Results:
lnLAGGDPPC 0.991** (0.005)
lnPCP -0.082* (0.020)
IIQ*lnPCP -0.182* (0.038)
lnINF 0.143 (0.061)
IIQ*lnINF 0.249** (0.124)
IIQ 0.286** (0.123)
Constant 0.148** (0.059)
ECT(-1) -0.672***(-4.24)
Source: Authors’ computations (2022) based on WDI and WGI of the World Bank (2021)
Notes: ln denotes natural logarithm, GDPPC represents real GDP per capita; LAGGDPPC represents real
GDP per capita from previous period; PCP represents physical capital per capita; INF represents index
of infrastructure stock; and IIQ represents index of institutional quality. Figures in parenthesis represent
robust standard errors. ***, ** and * denote statistical significance at 1%, 5% and 10% levels.

5. Conclusion outcomes in terms of the growth effect of infrastructure

This study examined the interactive effect of infrastructure development requires institutional reforms that are more
and quality of institutions on economic growth in SSA over fundamental than simply designing infrastructure projects
the period 1996-2020. This was with a view to providing and spending money on them. Hence, this study recommends
additional macroeconomic evidence that is specific to the that strategies aimed at massive infrastructure development
SSA region on the response of economic growth to must be complemented by measures to improve the quality of
infrastructure contingent on the quality of the underlying institutions in the countries of the region. Specifically,
institutions in the region. Results showed that infrastructure governments in the region need to develop a set of policies
does not independently lead to economic growth in SSA that are focused on the improvement of institutional
when the role of institutions is not considered. The study efficiency. This can be achieved by pursuing good
therefore concludes that countries with weak institutions do governance through a more stable socio-economic and
not benefit maximally from infrastructure development political environment, corrupt-free society, an effective
policies as weak institutions constrain the efficient use of public service, good regulatory environment, and a
infrastructure assets. transparent leadership structure. Good governance should be
The findings of this study show that achieving better allowed to rule over all other economic objectives.

380 | P a g e
International Journal of Multidisciplinary Research and Growth Evaluation www.allmultidisciplinaryjournal.com

6. References 21. Mankiw GN. The Growth of Nations. Brookings Papers

1. Agénor PR, Montiel PJ. Development Macroeconomics. on Economic Activity. 1995; 1:275-326.
3rd Edition, Princeton University Press, 2010. 22. Mankiw GN, Romer D, Weil DN. A Contribution to the
2. Amin SB, Amin YN, Khandaker MK, Khan F, Rahman Empirics of Economic Growth. The Quarterly Journal of
FM. Unfolding FDI, renewable energy consumption, and Economics. 1992; 107(2):407-437.
income inequality Nexus: Heterogeneous panel analysis. 23. Matthews RCO. The Economics of Institutions and
Energies. 2022; 15(14):5160. Sources of Growth. The Economic Journal. 1986;
https://doi.org/10.3390/en15145160. 96(384):903-918.
3. Arrow KJ. The Economic Implication of Learning by 24. Miao M, Qamruzzaman M. Dose remittances matter for
Doing. Review of Economic Studies. 1962; 29:155-173. openness and financial stability: Evidence from least
4. Barro RJ, Sala-i-Martin X. Convergence. Journal of developed economies. Frontiers in psychology, 2021,
Political Economy. 1992; 100(2):223-251. 12.
5. Barro RJ, Sala-i-Martin X, Blanchard OJ, Hall RE. 25. Nathaniel SP. Ecological footprint and human well-
Convergence across states and regions. Brookings being nexus: accounting for broad-based financial
papers on economic activity, 1991, 107-182. development, globalization, and natural resources in the
6. Chudik A, Pesaran MH. Common correlated effects Next-11 countries. Future Business Journal. 2021;
estimation of heterogeneous dynamic panel data models 7(1):1-18.
with weakly exogenous regressors. Journal of 26. North DC. Institutions, Institutional Change and
econometrics. 2015; 188(2):393-420. Economic Performance. New York: Cambridge
7. Commons JR. Institutional Economics: Its place in University Press, 1990.
political economy. New York: Macmillan, 1934. 27. North DC, Thomas RP. The rise of the western world: A
8. Dabla-Norris E, Brumby J, Kyobe A, Mills Z, new economic history. Cambridge University Press,
Papageorgiou C. Investing in Public Investment: An 1973.
Index of Public Investment Efficiency. Journal of 28. Ogbaro EO. Threshold Effects of Institutional Quality in
Economic Growth. 2012; 17(3):235-266. the Infrastructure-Growth Nexus. Journal of
9. Damijan S, Padilla SB. Effectiveness of Infrastructure Quantitative Methods. 2019; 3(2):45-61.
Project Investments in Africa1. In Paper dipresentasikan 29. Ogbaro EO, Oladeji SI. Infrastructure-Growth Link and
di UNU-WIDER Conference, 2014. the Threshold Effects of Sub-Indices of Institutions.
10. Effiong E. Financial Development, Institutions and Asian Journal of Business Environment. 2021; 11(1):17-
Economic Growth: Evidence from Sub-Saharan Africa. 25.
MPRA, 2015, 66085. 30. Okoh AS, Ebi BO. Infrastructure investment,
11. Esfahani HS, Ramıŕ ez MT. Institutions, infrastructure, institutional quality, and economic growth in Nigeria: An
and economic growth. Journal of Development interactive approach. European Journal of Humanities and
Economics. 2003; 70(2):443-477. Social Sciences. 2013; 26(1):1343-1358.
12. Fanta EG. Institutional Quality, Export Performance and 31. Olson M, Sarna N, Swamy AV. Governance and growth:
Income (Doctoral dissertation, Institute of Development A simple hypothesis explaining cross-country differences
Research and Development Policy, Ruhr University in productivity growth. Public Choice. 2000; 102(3-4),
Bochum), 2011 341-364.
13. Fay M, Yepes T. Investing in infrastructure: What is 32. Pesaran HM. General diagnostic tests for cross-section
needed from 2000 to 2010? Policy Research Working dependence in panels. Fac Econ, 2004.
Papers, 2003. https://doi.org/10.1596/1813-9450-3102 https://doi.org/10.17863/CAM.5113
14. Garlick RJ. Infrastructure, Institutions and Economic 33. Pesaran MH. A simple panel unit root test in the presence
Growth: a Time-series Study of the South African of cross-section dependence. Journal of Applied
Economy (Doctoral dissertation, University of Cape Econometrics. 2007; 22(2):265-312.
Town), 2008. https://doi.org/10.1002/jae.951
15. Getachew Y. The role of Public Capital on Economic 34. Pesaran MH, Yamagata T. Testing slope homogeneity in
Development, 2009. large panels. Journal of Econometrics. 2008; 142:50-93.
https://doi.org/10.26481/dis.20091015yg. https://doi.org/10.1016/j.jeconom.2007.05.010
16. Gulati M, Rao M. Checking corruption in the electricity 35. Richter R. The New Institutional Economics: Its Start, its
sector. World Bank, mimeo. Washington, D.C.: World Meaning, its Prospects. European Business Organization
Bank, 2006. Law Review. 2005; 6(2):161-200.
17. Hall JC, Sobel RS, Crowley GR. Institutions, capital, and doi:10.1017/s1566752905001618
growth. Southern Economic Journal. 2010; 77(2):385- 36. Robelo S. Long-Run Policy and Long-Run Growth,
405. Journal of Political Economy, University of Chicago
18. Kenny C. Measuring corruption in infrastructure: Press. 1991; 99(3):500-521.
Evidence from transition and developing countries. The 37. Romer PM. Increasing Returns and Long-Run Growth.
Journal of Development Studies. 2009; 45(3):314-332. Journal of Political Economy. 1986; 94(5):1002-1037.
19. Kodongo O, Ojah K. Does infrastructure really explain doi:10.1086/261420
economic growth in Sub-Saharan Africa? Review of 38. Sarafidis V, Yamagata T, Robertson D. A test of cross
Development Finance. 2016; 6(2):105-125. section dependence for a linear dynamic panel model
doi:10.1016/j.rdf.2016.12.001 with regressors. Journal of econometrics. 2008;
20. Lucas RE. On the Mechanics of Economic 148(2):149-161.
Development. Journal of Monetary Economics. 1988; 39. Saygın O, İskenderoğlu Ö. Does the level of financial
22(1):3-42. doi:10.1016/0304-3932(88)90168-7. development affect renewable energy? Evidence from

381 | P a g e
International Journal of Multidisciplinary Research and Growth Evaluation www.allmultidisciplinaryjournal.com

developed countries with system generalized method of

moments (System‐GMM) and Cross‐Sectionally
Augmented Autoregressive Distributed Lag (CSARDL).
Sustainable Development, 2022.
https://doi.org/10.1002/sd.2319
40. Solow RM. A Contribution to the Theory of Economic
Growth. The Quarterly Journal of Economics. 1956;
70(1):65. Doi: 10.2307/1884513.
41. Veblen T. Theory of the Leisure Class. New York:
Modern Library, 1899.
42. Weil DN. Economic Growth, First Edition. Pearson
Education Inc. Publishing as Addison Wesley, 2005.
43. Williamson OE. The economic institutions of capitalism.
New York: The Free Press, 1985.
44. Wu SY, Tang JH, Lin ES. The impact of government
expenditure on economic growth: How sensitive to the
level of development?. Journal of Policy Modeling.
2010; 32(6):804-817.

382 | P a g e