Content uploaded by Janet Ama Dzator
Author content
All content in this area was uploaded by Janet Ama Dzator on Aug 28, 2023
Content may be subject to copyright.
1
Leveraging Digital Technology for Development: Does ICT Contribute to
Poverty Reduction?
Janet Dzator 1, 2, Alex O. Acheampong 3,4, Isaac Appiah-Otoo5, Michael Dzator2,6
1 Newcastle Business School, University of Newcastle, Australia
2 Centre for African Research, Engagement and Partnerships (CARE-P), University of Newcastle,
Australia
3Bond Business School, Bond University, Gold Coast, Australia
4Centre for Data Analytics, Bond University, Gold Coast, Australia
5College of Management Science, Chengdu University of Technology, Chengdu, China
6SAE, Central Queensland University, Mackay, QLD, Australia
Corresponding author’s email: aacheamp@bond.edu.au
Abstract: The Sustainable Development Goal (SDG) 1 requires countries to end poverty of all
forms. At the same time, SDG 9 target 9.c calls for increasing Information and Communication
Technology (ICT) access and striving to provide universal and affordable access to the internet
in the least developed countries by 2020. While the existing studies have explored the effect of
ICT on economic growth, there is a limited empirical study on the effect of ICT on poverty
reduction. This article, therefore, investigates the effect of ICT on poverty reduction using
comprehensive panel data for 44 sub-Saharan Africa (SSA) countries from 2010 to 2019. Using
the dynamic system-generalized method of moment estimator to control endogeneity, the
findings show that telephone penetration, mobile phone penetration, and ICT goods imported
contribute to poverty reduction, while internet penetration, broadband penetration, and ICT
goods exported increase poverty rate regardless of gender and age group. The net effect
estimates from the conditional effect analysis reveal that the maximum value of most ICT
variables, economic growth, income inequality, and access to credit significantly increase the
poverty rate regardless of age group. We argue that for SSA to leverage ICT to drive economic
prosperity, policies that seek to drive ICT availability and affordability should be supported
with policies that ensure income redistribution, trickle-down development, and minimizing the
cost of accessing credit.
Keywords: Information and Communication Technology; Poverty reduction; Economic
recovery; Sustainable Development Goals; Sub-Saharan Africa
JEL Classification: I32; L86
2
1. Introduction
Sub-Saharan Africa (SSA) remains a region of concern for poverty reduction, which is critical
for a global effort to mitigate most of the current global challenges, including tackling
pandemics, climatic disturbances, and sustainable energy transitions. Globally, extreme
poverty
1
continues to decline at a plodding pace. For instance, the proportion of people living
in extreme poverty fell from 36% in 1990 to 16% in 2010 and 10% in 2015 (UN, 2019). This
suggests that for the past 25 years, approximately 1 billion people have been lifted out of
extreme poverty. Much of this progress is concentrated in Eastern and Southern Asia, while
extreme poverty in SSA remains higher (UN, 2019). In -contrast, approximately 413 million
people lived on less than $1.90 a day in 2015 in SSA, and the rate continues to rise (UN, 2019).
The COVID-19 pandemic has even exacerbated the SSA poverty rate. The African
Development Bank (ADB, 2021) report revealed that approximately 30 million Africans were
pushed into extreme poverty in 2020 and are expected to increase to 39 million at the end of
2021. A large proportion of these newly poor due to the COVID-19 pandemic comprises
women and female-headed households (ADB, 2021). The United Nations (UN) sustainable
development goal (SDG) 1 requires countries to end poverty of all forms. Specifically, SDG 1
requires countries to eradicate poverty for all people living on less than $1.25 a day and reduce
by half the proportion of men, women, and children of all ages living in poverty in all its
dimensions by 2030
2
.
Eradicating poverty in SSA requires policymakers to understand the region’s poverty factors
(Acheampong, Appiah-Otoo, Dzator, & Agyemang, 2021). With regard, the existing empirical
literature has investigated factors such as economic growth (Anyanwu & Anyanwu, 2017),
foreign aid (Mahembe & Odhiambo, 2021), financial development (Abosedra, Shahbaz, &
Nawaz, 2016), remittance (Acheampong, Appiah-Otoo, et al., 2021), institutions (Workneh,
2020), globalization (Fowowe & Shuaibu, 2014; Magombeyi & Odhiambo, 2018), government
spending (Anderson, d'Orey, Duvendack, & Esposito, 2018) and education (Anyanwu &
Anyanwu, 2017) on poverty reduction in SSA. However, despite the increasing scholarly and
policy discussions on the role of Information and Communication Technology (ICT) on
economic development (Asongu & Le Roux, 2017; Asongu, Nwachukwu, & Pyke, 2019),
broad research based on the effects of ICT on poverty reduction in developing countries
remains limited.
Does ICT matter in poverty reduction? The relevance of ICT in improving living standards has
been demonstrated during the COVID-19 pandemic. Globally, ICT usage increased during the
COVID-19 pandemic, significantly improving access to healthcare, education, social cohesion,
and business activities. ICT has also played a critical role in protecting against losses and
extending social protection coverage during the pandemic. Mora-Rivera and García-Mora
(2021) and Yang et al. (2021) argued that ICT could affect poverty status. A single country or
small sample studies have shown that farmers and small-scale traders have used ICT devices
to source markets for their produce, health care, etc., some to their benefit (Aker, Ghosh, &
Burrell, 2016; Cocosila, & Archer, 2010), but comprehensive research about the effects of ICT
on poverty reduction in developing countries and especially in SSA is limited. Currently, about
3.7 billion people do not use the internet, and these people comprise vulnerable, marginalized
groups such as females, rural, poor, elderly, and those with limited education and low literacy
(Deganis et al., 2021). From a regional perspective, SSA lags behind other regions in terms of
access to ICT (Mahler, Montes, & Newhouse, 2019). For instance, it is reported that only 1 in
5 in Sub-Saharan Africa used the internet in 2017. Also, 23% of the SSA population used the
1
Poverty rate is the percentage of persons living on less than US$1.90 per day in purchasing power parity
2
https://sdgs.un.org/goals/goal1
3
mobile internet regularly as of 2018 (Okeleke & Suardi, 2019). Factors such as lack of access
to the ICT infrastructure, high cost of internet connection and devices, a lack of digital and
literacy skills, weak awareness of the benefits of being connected, and lack of relevant content
in local languages are blamed for causing digital exclusion and low usage of ICT in developing
countries (Deganis et al., 2021). The relatively less ICT adoption in developing countries,
especially in SSA, has adverse socio-economic effects such as hampering economic growth,
social inclusion, quality of education, and health and increased income inequality and poverty
(Steele, 2018).
From the literature, the theoretical effect of ICT on poverty reduction is ambiguous. The
Amartya Sen Capability Approach offers an appropriate approach to understanding the impact
of development processes and interventions, such as ICT, on development outcomes
(Haenssgen & Ariana, 2018). For instance, Sen (2010) indicated that ICT variables such as
mobile and telephones are generally freedom-enhancing. The first strand of the literature
indicates that ICT can contribute to poverty reduction by providing opportunities that advance
social progress and foster social inclusion (Deganis, Haghian, & Tagashira, 2021). Thus, ICT
contributes to poverty reduction by facilitating economic growth, financial inclusion,
employment generation, health, education, democracy, and inclusive governance (Deganis et
al., 2021; Kelles-Viitanen, 2003). In addition, ICT can reduce poverty by enhancing firms’
productivity, improving market coordination, and strengthening social and human capital
(Galperin & Fernanda Viecens, 2017). For instance, based on the classical growth theory, ICT
leads to the efficient sharing of information and ideas, allowing firms to find innovative ways
to combine factors of production (labor and physical and human capital) and thus increasing
firms' productivity and economic growth (Aghion, & Howitt, 1992; Romer, 1990), thereby
contributing to poverty reduction (Galperin et al., 2017).
Contrarily, ICT is also argued to increase poverty through widening income inequality.
Drawing on Moll, Rachel, and Restrepo's (2022) recent theoretical framework on automation
and income and wealth inequality, technology can widen income inequality by increasing
returns on wealth. In their theoretical framework, the authors argue that new technologies can
result in wage stagnation and income stagnation at the bottom of the income distribution. In
addition, the skilled-bias argument shows that the rise of technologies is putting a premium on
the demand for skilled labor while displacing unskilled labor (Galperin et al., 2017). Thus, ICT,
a form of skilled-based technological change, has resulted in the rise of income among skilled
workers relative to non-skilled workers. For instance, Akerman, Gaarder, and Mogstad (2015)
confirmed the skilled-bias hypothesis by indicating that broadband availability in Norway
increases the income and productivity of skilled workers and displaces low-skilled workers.
Carlos (2010) further revealed that in Peru, adopters of the internet experienced faster income
growth than non-adopters. Given the positive relationship between income inequality and
poverty (Fosu, 2010), ICT contributing to income inequality further increases poverty. In
addition, ICT usage is argued to increase poverty because household spending on ICT can lead
to intra-household conflict, increase male dominance over resources, and divert household
resources away from food and other essentials (May, Waema, & Bjåstad, 2014).
The inconsistent theoretical arguments about ICT's impact on poverty reduction indicate that it
would be challenging for policymakers to design and implement sound policies for reducing
poverty in this digital age. Therefore, empirical studies are needed to understand the effect of
ICT on poverty reduction. In support of our call, May et al. (2014) argue that the limited
information regarding the impact of ICT on poverty reduction in developing countries has
raised concerns among policymakers who are being repeatedly advised to invest a large part of
the national budget in ICT infrastructure based on incomplete empirical evidence. Therefore,
4
it is crucial for researchers to empirically investigate whether investments in ICT represent a
worthwhile option for developing countries. Some studies have attempted to use micro-data to
examine the impact of the internet and mobile phones on poverty reduction in countries such
as Mexico (Mora-Rivera & García-Mora, 2021), China (Yang, Lu, Wang, & Li, 2021) and Peru
(Beuermann, McKelvey, & Vakis, 2012). The results from these studies have been conflicting.
For instance, while some studies found that ICT contributes to poverty reduction (Beuermann
et al., 2012; Mora-Rivera & García-Mora, 2021; Yang et al., 2021), other studies have revealed
that ICT plays a significant role in increasing poverty (Sujarwoto & Tampubolon, 2016; Tang
& Zhu, 2020).
Apart from the inconclusive empirical results, there are several important research questions
and issues that the earlier empirical studies have overlooked. First, the existing empirical papers
did not consider which aspect of ICT matters for poverty reduction. The empirical studies on
ICT have proxied ICT with different variables such as fixed-broadband subscriptions, fixed
telephone subscriptions, internet usage, and mobile cellular subscriptions and others (see,
Adeabah, Asongu, & Andoh, 2021; Kallal, Haddaji, & Ftiti, 2021; Pradhan, Arvin, Nair, Hall,
& Bennett, 2021). The literature suggests that each of these ICT variables has a different impact
on development outcomes; therefore, using different ICT indicators for empirical analysis
provides more room for policy implications and helps to understand which aspect of ICT
matters most for development (Asongu & Le Roux, 2017; Yang et al., 2021). Second, the
existing empirical studies have not considered the effect of ICT on gender poverty. Examining
the effect of ICT on gender poverty is essential for policy because of the digital gender divide.
For instance, Co-operation and Development (2018) indicated that 327 million fewer women
than men had smartphones and could access the mobile internet in 2017. This suggests that
ICT can have a heterogeneous effect on gender poverty, which existing studies have
overlooked.
Third, the existing empirical studies have only examined the direct effect of ICT on poverty
reduction without probing the indirect effect of ICT on poverty reduction. For instance, May
et al. (2014) argue that there may be a well-direct relationship between ICT and poverty
reduction, but the mechanisms (indirect effect) through which ICT affects poverty are not fully
understood in the literature. The theoretical literature argues that ICT can shape poverty by
conditioning countries’ economic growth, income inequality, and financial inclusion (Mushtaq
& Bruneau, 2019). Therefore, empirical studies are needed to extend the current literature by
examining how ICT conditions economic growth, income equality, and access to credit to
reduce poverty, but the existing empirical studies have not considered such a conditional effect.
Finally, there is a paucity of empirical studies examining the effect of ICT on poverty reduction
in SSA. However, as discussed above, the poverty rate in SSA remains higher, and ICT
penetration in the region is the lowest compared to other regions worldwide. Therefore, studies
evaluating the effect of ICT on poverty reduction focusing on SSA will contribute significantly
to literature and discussions on sustainable development policy.
Therefore, this study seeks to address these research gaps by examining the effect of ICT on
poverty reduction in SSA. To achieve the objective of this study and contribute to the literature
on the ICT-poverty nexus and policy discussions, we address the following four (3) key
research questions:
1. Do broadband, telephone, ICT goods exported and imported, internet usage, and
mobile cellular subscriptions contribute to poverty reduction in SSA?
2. Do broadband, telephone, ICT goods exported and imported, internet usage, and
mobile cellular subscriptions have a heterogeneous effect on gender poverty in
SSA?
5
3. Do broadband, telephone, ICT goods exported and imported, internet usage, and
mobile cellular subscriptions condition economic growth, income inequality, and
access to credit to affect poverty in SSA?
This study is novel and contributes to policy discussions and the literature on ICT investment
and economic development in the following ways. First, this study provides new empirical
evidence on the effect of different ICT variables on poverty reduction in SSA. Thus, unlike
prior studies such as Ofori et al. (2021), we use six (6) ICT variables: fixed-broadband
subscriptions, fixed telephone subscriptions, ICT goods exported, ICT goods imported, internet
usage and mobile cellular subscriptions, to capture ICT. Using these variables to capture ICT
helps analyze which aspect contributes to poverty reduction and provides more room for policy
implications. In addition, the adoption rate of these ICT variables differs in SSA, and these
variables are expected to have a heterogenous effect on poverty. Second, this study contributes
to the literature by examining the impact of ICT on different age population cohorts’ poverty
in SSA. In addition, this paper extends the literature by examining the effect of ICT on gender
poverty in SSA. The use of gender and aged dimensions of poverty would contribute to policy
discussions on equity in access to infrastructure facilities and socioeconomic development in
developing countries. Employing the gender and aged dimensions of poverty make this paper
unique from previous literature such as Ofori et al. (2021), Diga, Nwaiwu, and Plantinga
(2013), and Urean, Lacatus & Mocean (2016) that focus on poverty in general.
Third, this study further adds to knowledge by exploring if ICT conditions for economic
growth, income inequality, and access to credit affect poverty in SSA. Analyzing the
conditional effect of ICT on poverty would enable policymakers to design and implement
holistic policies for addressing the increasing poverty in the region. Although studies such as
Ofori et al. (2021) attempted to examine the interactive effect of ICT and financial development
on poverty reduction in SSA and evaluated how financial development conditions ICT to affect
poverty and not how ICT conditioned financial development to affect poverty in SSA. Form
Ofori et al. (2021) findings, it was suggested that at different values of financial development,
ICT reduces poverty in SSA. However, our findings suggest that at the maximum value of most
of the ICT variables, access to credit worsens poverty in SSA. Finally, we address our research
questions using the dynamic system-generalized method of moment estimator to control
endogeneity that may result from mismeasurements, reverse causality, and variable omission
bias. Given the reliability of the dynamic system-generalized method of the moment estimator,
the outcome of this study will contribute to sustainable development policies because
investment in ICT infrastructure can be guided by policy to achieve poverty reduction in SSA.
2. Review of related literature
2.1. Theoretical relationship between ICT and poverty reduction
Theoretically, there are conflicting views about the effect of ICT on poverty reduction. The
first school of thought argues that ICT contributes to poverty reduction. The Amartya Sen
Capability Approach (CA) views human development as enhancing people’s freedom to
achieve the capabilities to function effectively (Sen, 1999; 1980). The CA offers an appropriate
approach to understanding the impact of development processes and interventions, such as ICT,
on development outcomes (Haenssgen & Ariana, 2018). For instance, Sen (2010) indicated
that ICT variables such as mobile and telephones are generally freedom-enhancing. Sen (2010)
6
cited that ICT and accompanying technologies serve as an instrument of liberation. The
implication is that ICT could serve as an instrument for people to enhance their capability to
function, contributing to economic development. Also, based on the classical growth theory,
ICT leads to the efficient sharing of information and ideas, allowing firms to find innovative
ways to combine factors of production (labor and physical and human capital) and thus
increasing firms' productivity and economic growth (Aghion, & Howitt, 1992; Romer, 1990),
thereby contributing to poverty reduction (Galperin et al., 2017).
In support of these theories, the literature shows that ICT contributes to economic growth and
firm productivity by enhancing e-commerce, network externalities, productive efficiency,
economic diversification, business retention, cost reduction, improving human capital
development, promoting social and political stability, economic freedom, and information
dissemination (Thompson and Garbacz, 2011; Vu, 2011; Pradhan et al., 2014; Adeleye and
Eboagu, 2019; Vu, Hanafizadeh and Bohlin, 2020; Appiah-Otoo and Song, 2021b). Therefore,
ICT's economic growth and firms’ productivity impact contribute to poverty alleviation by
enhancing domestic investment, employment opportunities, and household consumption
(Anyanwu, 2014; Asongu and Odhiambo, 2020)). Similarly, ICT is argued to reduce poverty
by improving household welfare (Chang and Just, 2009; Galperin and Fernanda Viecens, 2017;
Ma et al., 2020). ICT is argued to generate employment opportunities by reducing information
asymmetry in the labor market (Yang et al., 2021). By creating employment, ICT has
stimulated household income growth leading to improved welfare and poverty reduction
(Chang and Just, 2009; Ma and Wang, 2020; Ma et al., 2020; Yang et al., 2021).
Also, ICT can contribute to poverty reduction by enhancing access to credit. ICT can improve
access to credit by reducing information asymmetry and transaction costs associated with the
traditional financial market (Cheng, Chien, and Lee, 2021; Yang et al., 2021). Access to credit
makes it easier for poor households to establish businesses and, thus, enhance their earnings
leading to poverty reduction (Yang et al., 2021). Similarly, ICT can reduce poverty by
promoting market coordination (Galperin and Fernanda Viecens, 2017). Due to credit,
logistics, and information constraints, the poor are often sidelined from modern market
exchanges (Li et al., 2019). ICTs help address these challenges at a reduced cost (Li et al.,
2019). The rapid development of ICTs has facilitated e-commerce (Freund and Weinhold,
2004; Li et al., 2019), payment and financing (Miao and Jayakar, 2016; Iman, 2018) and assist
the poor to participate in business ecosystems communities, with all these aiding to combat
poverty (Galperin and Fernanda Viecens, 2017; Li et al., 2019). Also, ICT facilitates financial
knowledge and, thus, prevents one from falling into asset poverty (Yang et al., 2021).
Besides, ICT can contribute to poverty eradication by spurring social and human capital
development (Galperin and Fernanda Viecens, 2017; Yang et al., 2021). ICT provides
employable skills via online education (Zheng and Lu, 2021). Rapid ICT development reduces
the distance, time, and space involved in the face-to-face education system (Zheng and Lu,
2021). For instance, the outbreak of the COVID-19 pandemic, with its associated lockdown
policies, facilitated the use of ICT for remote teaching and learning (Adedoyin and Soykan,
2020; Bao, 2020). In addition, ICT has enhanced health and healthcare expenditure through e-
health. Currently, virtual medical consultation is made possible by ICT, enabling one to reduce
medical spending and improve health (Yang et al., 2021). ICT provides social connections at
reduced cost leading to greater social networks and, thus, improving household satisfaction and
happiness (Wang and He, 2020; Yang et al., 2021). Moreover, ICT contributes to poverty
reduction by enhancing inclusive governance (Galperin and Fernanda Viecens, 2017).
Recently, ICTs have encouraged governance participation, reduced corruption, enhanced voice
7
and accountability, and improved transparency (Bertot, Jaeger, and Grimes, 2010; Sassi and
Ben Ali, 2017; Dhaoui. Also, ICTs have promoted e-government initiatives that have
transformed how citizens associate with their government, helping alleviate poverty (Galperin
and Fernanda Viecens, 2017).
The second school of thought argues that ICT could worsen poverty by widening income
inequality. Contrarily, ICT is also argued to increase poverty through widening income
inequality. Drawing on Moll, Rachel, and Restrepo's (2022) recent theoretical framework on
automation and income and wealth inequality, technology can widen income inequality by
increasing returns on wealth. In their theoretical framework, the authors argue that new
technologies can result in wage stagnation and income stagnation at the bottom of the income
distribution. In addition, the skilled-bias argument shows that the rise of technologies is putting
a premium on the demand for skilled labor while displacing unskilled labor (Galperin &
Fernanda Viecens, 2017). Thus, ICT, a form of skilled-based technological change, has resulted
in the rise of income among skilled workers relative to non-skilled workers. For instance,
Akerman, Gaarder, and Mogstad (2015) confirmed the skilled-bias hypothesis by indicating
that broadband availability in Norway increases the income and productivity of skilled workers
and displaces low-skilled workers. Carlos (2010) further revealed that in Peru, adopters of the
internet experienced faster income growth than non-adopters. Given the positive relationship
between income inequality and poverty (Fosu, 2010), ICT contributing to income inequality
further increases poverty. In addition, ICT usage is argued to increase poverty because
household spending on ICT can lead to intra-household conflict, increase male dominance over
resources, and divert household resources away from food and other essentials (May, Waema,
& Bjåstad, 2014). ICT is also argued to increase poverty by driving conflicts, terrorism, and
war (Bailard, 2015; Dafoe and Lyall, 2015; Weidmann, 2015; Chin, 2019; Ackermann,
Churchill, and Smyth, 2021).
2.1. Survey of empirical studies on ICT and poverty reduction
Because of the ambiguous theoretical effect of ICT on poverty, the literature on the ICT-
poverty relationship has recently gained attention among scholars. Most studies support the
argument that ICTs have been “pro-poor.” For instance, Mora-Rivera and García-Mora (2021)
found that internet access reduces poverty in Mexico, using the propensity score matching and
household survey data obtained in 2016. They also found that the rural sector has a more
significant impact on mobile access than the urban sector. In the case of China, Yang et al.
(2021) assessed the effect of mobile internet use on multidimensional poverty and found that
mobile internet use reduces poverty. They also showed that mobile internet use indirectly
reduces poverty by improving household income, promoting access to public services, and
increasing overall satisfaction and happiness. Rodríguez and Sánchez-Riofrío (2017) also
investigated the role of ICTs in poverty reduction in the case of Latin America and found that
ICTs reduce poverty. Also focusing on Peru, Beuermann, McKelvey, and Vakis (2012) found
that mobile phones reduce the incidence of poverty and extreme poverty in rural Peru by 8%
and 5.4%, respectively.
For a panel of 62 countries, Mushtaq and Bruneau (2019) showed that ICT variables such as
mobile phones, fixed telephones, and the internet significantly contribute to poverty reduction.
In collaboration with the previous findings, Ofori et al. (2021) examined the impact of ICT
access, usage, and skills to reduce the severity and intensity of poverty for a panel of 42 SSA
countries. Evidence from the dynamic system-GMM and panel corrected standard error
estimations and showed that ICT access, usage, and skills reduce severity and intensity. The
authors further show that financial development and access also condition ICT access and skills
8
to contribute to SSA's poverty reduction. Horn (2014) also investigated the nexus between ICT
and poverty in South Africa and showed that a negative correlation exists between the
geographic spread of ICT and poverty and vice versa. Also, Diga et al. (2013), using textual
analysis from relevant ICT and poverty reduction policy documents from Uganda, South
Africa, and Nigeria between 2005 and 2012, concluded that ICT policy interventions drive
economic capability among the poor. Lechman and Popowska (2022) further examined the
effect of ICT indicators on poverty eradication in 40 lower- and middle-income countries.
Evidence from the fixed effect estimator showed that imported ICT goods do not directly affect
poverty headcount, while the internet and mobile phones contribute to poverty eradication. The
authors also showed evidence that at the mean of human capital and economic growth, mobile
and internet contribute to poverty eradication. Table 1. shows a summary of the selected studies
on ICT and poverty.
In summary, the literature review suggests the existing studies have not considered which
aspect of ICT matters for poverty reduction. Also, the existing empirical studies have not
considered the effect of ICT on gender poverty. Third, the existing empirical studies have only
examined the direct effect of ICT on poverty reduction without probing the indirect effect of
ICT on poverty reduction. Finally, there is a paucity of empirical studies examining the effect
of ICT on poverty reduction in SSA. Therefore, this study seeks to address these research gaps
by investigating the effect of ICT on poverty reduction using comprehensive panel data for 44
SSA countries from 2010 to 2019. From the review of the theoretical and empirical literature,
the following hypothesis is formulated:
Hypothesis 1: Broadband, telephone, ICT goods exported and imported, internet usage, and
mobile cellular subscriptions directly contribute to poverty reduction in SSA.
Hypothesis 2: Broadband, telephone, ICT goods exported and imported, internet usage, and
mobile cellular subscriptions have a heterogenous effect on gender poverty in SSA.
Hypothesis 3: Broadband, fixed telephone, ICT goods exported and imported, internet usage,
and mobile cellular subscriptions condition GDP per capita (economic growth), income
inequality, and access to credit to reduce poverty in SSA.
[INSERT TABLE 1 HERE]
3. Methodology and Data
3.1. Empirical model
In this study, we follow the standard poverty function of Bourguignon (2004), Ravallion
(1997), and Adams Jr (2004) to model the effect of ICT on poverty reduction in SSA. Thus,
from Eq. (1), poverty is specified to be a function of economic growth , income
inequality , access to domestic credit
, information and communication technology
variables and other covariates that affect poverty.
In this study, we adopt the dynamic reduced-form modeling approach of the above equation.
Therefore, the dynamic log-linear form of the empirical equation for estimating the poverty
function is stated in Eq. (2):
9
The literature suggests that ICT can reduce poverty by increasing access to credit, reducing
income inequality, and enhancing economic growth. In other words, ICT can condition
economic growth, income inequality, and access to credit to affect poverty. To test this
assumption, we augment Eq. (2) with the interaction terms for ICT and economic growth, ICT
and income inequality, and ICT and access to credit. In Eq. (3), we estimate the interactive
effect of ICT and economic growth on poverty.
Also, in Eq. (4), we estimate the interactive effect of ICT and income inequality on poverty.
In Eq. (4), we estimate the interactive effect of ICT and access to credit on poverty.
In this study, we further estimate the marginal effects for Eq. (3), (4), and (5). The marginal
effect analysis will help to evaluate how changes in the ICT variables affect economic growth,
income inequality, and access to domestic credit to affect poverty. In other words, the marginal
effect analysis demonstrates how ICT reinforces the effect of economic growth, income
inequality, and access to domestic credit to affect poverty. We must emphasize that are many
channels through which ICT can affect poverty. However, this study is interested in moderated
variables such as economic growth, income inequality, and access to credit because, as
identified in the literature, these are the immediate variables that can be conditioned by ICT to
influence poverty. The total effect of economic growth, income inequality, and access to
domestic credit can be obtained using the first partial derivates as stated in Eq. (6), (7), and (8)
as follows:
From Eq. (6), (7), and (8), the total/net effect of economic growth, income inequality, and
access to credit on poverty are conditioned on the level of ICT penetration. The total/net effect
of economic growth, income inequality, and access to finance on poverty would be evaluated
at the minimum, mean, and maximum values of the ICT variables.
From the above equations, , is the natural log of poverty
measure, is the lagged poverty, is the natural log of economic growth
measure, is the natural log of income inequality, is the natural log of access to
domestic credit,
is the natural log of ICT variables, is the natural log of the control
10
variables and is the stochastic error term. captures the coefficients to be estimated
elasticity and are the interaction term coefficients to be estimated.
3.2. Econometric estimation strategy
In this study, the Blundell and Bond (1998) dynamic system-generalized method of moment
(System-GMM) to estimate the above equations. The two-step dynamic system-GMM is
important for addressing endogeneity issues emanating from measurement errors, variable
omission bias, and reverse causality. Besides endogeneity, the number of countries (N=44)
used in this study exceeds the years considered (T=10), which makes it imperative to apply the
Blundell and Bond (1998) system-GMM estimator to achieve reliable and unbiased estimates.
In addition, the Blundell and Bond (1998) system-GMM estimator has been shown to provide
more reliable estimates than the Arellano and Bond (1991) first difference generalized method
of moment when the number of countries (N) exceeds the time (T). For the post-estimation
statistics, we applied the Hansen test and the Arellano–Bond first-order (AR 1) and second-
order (AR 2) autocorrelation to evaluate the validity of the two-step dynamic system-GMM
models. In applying the dynamic two-step system-GMM, we imposed the collapse option to
prevent instrumental proliferation. We also conduct a cross-sectional dependency test of the
regressions residual term to present cross-sectionally independent estimates. This approach is
consistent with the literature (see, for instance, Acheampong, 2019; Amuakwa-Mensah et al.,
2018). Evidence from the post-estimation statistics indicates that our results are free from
cross-sectional dependency, and there is no issue of instrument proliferation and second other
autocorrelation.
3.3. Data description
This study uses comprehensive panel data for 44 SSA
3
countries between 2010 to 2019
4
. This
study focuses on 44 SSA countries because of data availability. The data used in this study
were obtained from the International Labour Organization (ILO), World Bank World
Development Indicators (WDI), and Standardized World Income Inequality Database
(SWIID). Except for the poverty variable, which was obtained from ILO, and the income
inequality variable obtained from Solt (2016) SWIID, the remaining variables were obtained
from WDI. All the variables were transformed into their natural logarithms for the empirical
estimation. Table 2 shows the descriptive statistics for the variables.
Dependent variable: This study relies on the International Labour Organization (ILO) working
poverty dataset. According to the ILO, the working poverty rate is the percentage of employed
persons living in poverty despite being employed using the international poverty line of
US$1.90 per day in purchasing power parity. This dataset is essential for this study because it
provides disaggregated data for different working-age populations and gender. Also, the ILO
poverty dataset covers the majority of SSA countries studied. To contribute significantly to
knowledge and policy, the poverty rate indicator used in this paper is categorized into the
3
Angola, Benin, Botswana, Burkina Faso, Burundi, Cabo Verde, Cameroon, Central African Republic, Chad,
Comoros, Congo, Dem. Rep., Congo, Rep., Cote d'Ivoire, Equatorial Guinea, Eritrea, Eswatini, Ethiopia, Gabon,
Gambia, Ghana, Guinea, Guinea-Bissau, Kenya, Lesotho, Liberia, Madagascar, Malawi, Mali, Mauritania,
Mauritius, Mozambique, Namibia, Niger, Nigeria, Rwanda, Senegal, Somalia, South Africa, Sudan, Tanzania,
Togo, Uganda, Zambia, Zimbabwe.
4
As of the time of writing this paper, the ILO poverty variable, which is the key dependent variable, was only
available between 2010 to 2019.
11
proportion of the working population below the international poverty line (%) aged 15+, 15-
24, and 24+. In this study, we referred to the poverty rate for the working population aged 15+
as youth poverty, 15-24 as middle-aged poverty, and 24+ as adult poverty. We considered the
effect of ICT on poverty across different aged groups because the adoption of ICT and poverty
rate differs among different aged groups. For instance, the older population depends on a fixed
income source and has limited access to and usage of ICT compared to the youth. Therefore,
we argue that the impact of ICT on poverty among the older population would be different
from poverty among the youthful population. Also, men and women have unequal access to
ICT in SSA (Adepetun, 2021; FAO, 2021) and expect ICT to have a disparate effect on gender
poverty.
Explanatory variables: Following the existing studies on ICT, we use six (6) variables, namely,
fixed-broadband subscriptions per 100 people, fixed telephone subscriptions per 100 people,
ICT service exports (% of service exports, BoP), ICT goods imports (% total goods imports),
Secure Internet servers per 1 million people and mobile cellular subscriptions (per 100 people),
to capture ICT. Using these variables to capture ICT helps analyze which aspect contributes to
poverty reduction and provides more room for policy implications. From the literature review,
we expect these ICT variables to have a significant negative effect on poverty (Lechman &
Popowska, 2022; Mushtaq & Bruneau, 2019; Diga et al., 2013).
Control variables: Following the poverty literature, this study also accounts for other factors
affecting poverty. The following variables are also included in the poverty model:
➢ Economic growth: This is measured as GDP per capita at constant 2010 US dollars in
the poverty equation. The trickle-down hypothesis argues that higher economic growth
is associated with poverty eradication because increasing economic growth is
associated with higher employment opportunities and provides financial resources for
implementing pro-poor policies (Dollar & Kraay, 2002). Therefore, we expect
economic growth to negatively impact poverty (Dollar & Kraay, 2002; Mastromarco et
al., 2014; Santos et al., 2019).
➢ Income inequality: This is measured with the post-tax/transfer Gini index of income
inequality and captures the extent to which income is distributed in an economy. This
study uses the post-tax/transfer Gini index from Solt's (2016) SWIID to represent
income inequality. The social unrest theory suggests that income inequality can lead to
the rise of poverty. The social unrest theory suggests income inequality encourages
disruptive socio-political activities such as riots, crime, political turnovers, etc. (Alesina
& Perotti, 1996; Barro, 2000; Ezcurra, 2007; Perotti, 1996). Due to the injustice the
poor face, they mostly get involved in social vices and political disturbances. Barro
(2000) argues that the poor engagement of the poor in disruptive social and political
activities is a direct waste of productive resource since the time spent on these
disruptive activities are not devoted to something productive. Also, political unrest due
to income inequality distorts market activities and deters investment, hampering
economic growth and leading to poverty. We expect income inequality to have a
significant positive effect on poverty (Barro, 2000; Bourguignon, 2004; Fosu, 2010).
➢ Access to credit: This is measured as domestic credit to the private sector as a
percentage of GDP and is used to the extent to which people have access to credit. We
expect access to have a significant negative effect on poverty because having access to
credit enables the poorer to build and own assets to increase their income (Jalilian and
Kirkpatrick, 2002; Appiah-Otoo and Song, 2021; Naceur and Zhang, 2016). Access to
credit contributes to poverty eradication by encouraging entrepreneurship and enabling
poorer people to invest in education, health, and technology (Koomson, Villano, &
12
Hadley, 2020). Therefore, access to credit is expected to have a negative relationship
with poverty (see Ofori et al., 2021; Inoue, 2018; Rehman and Shahbaz, 2014; Quartey,
2005).
➢ Access to electricity: This is measured as a percentage of the total population with
access to electricity. Access to electricity reduces poverty by enhancing people's
functioning ability (Acheampong, Dzator, & Shahbaz, 2021). Thus, access to electricity
contributes to human development and, for that matter, poverty reduction by improving
health, human capital, firms’ productivity, and gender empowerment (Dinkelman,
2011; Rao & Pachauri, 2017). Therefore, we expect access to electricity to have a
negative relationship with poverty (Acheampong et al., 2021; Acheampong, Erdiaw-
Kwasie, & Abunyewah, 2021).
➢ Trade openness: This is measured using total trade volume as a percentage of GDP, and
we expect trade openness to have a significant negative effect on poverty
(Acheampong, Appiah-Otoo, Dzator, & Agyemang 2021; Magombeyi & Odhiambo,
2018; Le Goff, & Singh, 2014). The Stolper–Samuelson theorem and Heckscher-Ohlin
model inform this expectation, which suggests that there will be an increase in income
for the abundant factor when a country exports goods that intensively uses that factor
of production to produce goods (Krugman, Obstfeld, & Melitz, 2017). For instance,
given that labor is more abundant in developing countries, if developing countries
export goods that intensively use labor in their production process, trade would increase
labor income. Similarly, if unskilled labor is an abundant factor in developing countries,
these theories indicate that unskilled labor in developing countries would gain from
trade (Le Goff & Singh, 2014).
➢ Urbanization: This is measured with the urban population, which captures the extent to
which the size of urbanization. Urbanization contributes to poverty because people
migrate from the rural areas characterized by low income and wages to the urban sectors
with higher income and wages. This movement initially leads to the poorest share of
the total income declining as economic growth rises while increasing the income of the
relatively small and wealthy people in the industrial and urban sectors, thus increasing
income inequality and poverty (Deininger & Squire, 1997; Barro, 2000; Robinson,
1976). We expect urbanization to have a significant positive effect on poverty
(Ravallion, 2002; Sulemana et al., 2019).
➢ Domestic capital investment: This is measured as gross capital formation. The asset-
based theory argues that asset accumulation, including physical capital, can make
owners derive returns or income from the commercialization or the sale of the assets
and thus contribute to upward mobility and transition people out of poverty (Etim &
Edet, 2014; Moser, 2006). The implication is that a country that poses more improved
land, plant, machinery, and equipment purchases; and the construction of roads,
railways, and the like, including schools, offices, hospitals, private residential
dwellings, and commercial and industrial buildings, could have less poverty rate and
vice versa. From this, we expect domestic capital investment to have a significant
negative relationship with poverty (Acheampong et al., 2021; Piachaud, 2002).
[INSERT TABLE 2 HERE]
13
3.3.1 Overview of key variables
This section shows the poverty and ICT variables trends among the sample SSA countries used
for the analysis. The values presented in the figures are average values, and any interpretation
based on the figures should be maintained as such. Figure 1 shows the trends of the proportion
of the working population below the international poverty line (US$1.90 per day) in SSA
between 2010-2019. Also, Figure 2 shows the trends of ICT variables in SSA for the period
between 2010-2019. As a note, Figure 2 has two axis values. The primary axis values are for
telephone and ICT goods exported and imported, while the secondary axis values are for
internet, broadband, and mobile phone variables.
From Figure 1, it can be observed that poverty rate among all the age groups has experienced
a decline since 2010. However, the poverty rate among people aged 15-24 (middle-aged
poverty) is higher, followed by poverty rate among people aged 15+ (youth poverty age group)
and finally poverty rate among people aged 24+ (youth poverty age group). Poverty rate is
higher among people aged 15+ and 15-24 because most people within these age brackets are
primarily in school does not actively participate in market activities that generate a regular and
permanent source of income. However, people aged 24+ are active in the labor market. These
observations show that in sub-Saharan Africa, people within the age brackets of 15+ and 15-
24 are more vulnerable to poverty than those within the age bracket of 24+. These observations
align with ILO's observations that young people are far more likely to be in poverty than
adults
5
. It must be noted that despite the decline in the poverty rate among all age groups, the
proportion of the working population below the international poverty line remains high.
[INSERT FIGURE 1 HERE]
Figure 2 also shows that ICT goods exported as a percentage of service exports from SSA have
sharply declined since 2010. Also, ICT goods imported as a percentage of total goods imports
experienced a decline from 2010 until 2015 and then fell again. However, the decline in ICT
goods imported is not sharped compared to the ICT goods exported. Figure 2 also depicts that
fixed-broadband subscriptions per 100 people have risen since 2012. Despite the rise, fixed-
broadband subscriptions per 100 people remain below 40, indicating poor broadband
penetration in SSA. Also, secure internet servers per 1 million people were very low in SSA
from 2010 to 2015. However, after 2015, internet penetration rose until 2018 it started dipping
in 2019. From these, we could argue that broadband and internet penetration in SSA is
problematic and deserves attention to speed up internet and broadband coverage in the region.
Regarding mobile phone penetration, Figure 2 shows that mobile cellular subscriptions per 100
people have increased in SSA. For instance, in 2016, mobile cellular subscriptions per 100
people on average increased from approximately 80.5 to 81.3 in 2017 and then 86.8 and 91.7
in 2018 and 2019, respectively. On the other hand, as shown in Figure 2, fixed telephone
subscriptions per 100 people experienced a sharp decline from 2010 to 2018. However,
telephone subscriptions have started rising since 2019. Given the difference in dynamics among
the ICT variables, we expect them to have different impacts on SSA's poverty reduction.
[INSERT FIGURE 2 HERE]
5
https://ilostat.ilo.org/young-people-are-far-more-likely-to-be-in-working-poverty/
14
4. Results and Discussion
This section presents and discusses the effect of ICT on poverty reduction. In this section, we
also reported the interactive effect of ICT and economic growth, income inequality, and credit
access on SSA poverty. It must be noted we deployed the fixed-effect estimator to estimate the
baseline results (see Appendix Table 1). However, due to the inability of the fixed effect
estimator to handle endogeneity, a significant issue in the ICT-poverty relationship (Ofori et
al., 2021), we focused our discussion primarily on the results from the two-step dynamic
system-GMM estimator.
4.1 Effect of ICT on poverty rate
Table 3A presents the estimates for total youth and middle-aged poverty, while Table 3B
presents the estimates for total adult poverty. The results show that the coefficient of telephone
penetration is negative and statistically significant at a 1% level. The estimate indicates that a
1% increase in telephone penetration reduces youth poverty, middle-aged poverty, and adult
poverty by 0.083%, 0.102%, and 0.092%, respectively. Also, the coefficient of mobile
penetration is negative and statistically significant at a 1% level. The estimate shows that a 1%
increase in telephone penetration reduces youth poverty, middle-aged poverty, and adult
poverty by 0.061%, 0.068%, and 0.065%, respectively. These results indicate that telephone
and mobile phone penetration contributes to poverty reduction in SSA. Telephone and mobile
phone coverage reduce poverty by enhancing social networks, reducing transport and
transaction costs, reducing information asymmetry, and enhancing production efficiency and
productivity (Sife, Kiondo, & Lyimo-Macha, 2010). Also, inferring from the Amartya Sen
Capability Approach (CA), mobile and telephones are generally freedom-enhancing (Sen,
2010). Sen (2010), and they could serve as an instrument enhancing people's capability to
function and thus contribute to poverty reduction in SSA. This result aligns with Beuermann
et al. (2012) and Mushtaq and Bruneau's (2019) findings, which suggested that telephone and
mobile phone penetration reduces poverty.
Contrarily, the results indicate that the estimated internet penetration coefficient is positive and
statistically significant at a 5% level for youth poverty and 1% for middle-aged and adult
poverty. The estimated coefficient suggests that youth poverty, middle-aged poverty, and adult
poverty increase by 0.010%, 0.012%, and 0.012%, respectively when internet penetration
increases by 1%. Similarly, the estimated coefficient on broadband penetration is positive and
statistically significant at 1%. The estimated coefficient suggests that youth poverty, middle-
aged poverty, and adult poverty increase by 0.281%, 0.259%, and 0.268%, respectively, when
broadband penetration increases by 1%. Thus, contrary to the claim that internet and broadband
penetration contribute to poverty reduction by enhancing households access to information,
using mobile financial services, reducing transaction costs, and improving production
efficiency (Yang et al., 2021), our result suggests that both internet and broadband penetration
worsens poverty in SSA. This result could reflect the region's lower internet and broadband
penetration rate (see discussion in section 3.3.1). Globally, SSA has poor internet and
broadband access (Mahler, Montes, & Newhouse, 2019). As of 2017, only 1 in 5 in Sub-
Saharan Africa used the internet, while 23% of the SSA population used the mobile internet
regularly as of 2018 (Okeleke & Suardi, 2019). Also, poor telecommunication and ancillary
infrastructures such as internet cables and an electricity grid in SSA could explain the positive
effect of internet and broadband penetration on poverty (Langmia, 2006). Finally, the positive
effect of internet and broadband penetration on poverty could also reflect the higher cost
associated with using the internet and broadband in SSA. Our result contradicts the findings of
15
earlier studies, such as Yang et al. (2021) and Mora-Rivera and García-Mora (2021), which
showed that internet penetration contributes to poverty reduction.
Focusing on trade in ICT goods, the results also suggest that the estimated coefficient of ICT
goods exported is positive and statistically significant at 1%. The estimated coefficient suggests
that youth poverty, middle-aged poverty, and adult poverty increase by 0.025%, 0.025%, and
0.025%, respectively when ICT goods exported increase by 1%. These results argue that ICT
goods exported worsen poverty in SSA. ICT goods exported increase poverty because it
exacerbates income inequality. Thus, based on the Heckscher-Ohlin model, we argue that ICT
goods exported increases poverty because few people in SSA are involved in ICT jobs, and the
ICT jobs are also skilled intensive; therefore, any ICT goods exported will increase the income
of these fewer population at the expense of the majority of the people, thereby driving income
inequality and poverty. Also, based on specific-factor theory, we argue that ICT goods exported
increase poverty because an increase in the relative price of ICT exported goods will benefit
the fewer population involved in ICT jobs in SSA, and thus, increase income inequality and
poverty. Finally, for the period of analysis, ICT-exported goods experienced a sharp decline in
SSA (see Figure 2), which also explains the inability of ICT-exported goods to contribute to
poverty reduction in SSA.
Also, the estimated coefficient of ICT goods imported is negative and statistically significant
at 1%. The estimated coefficient suggests that youth poverty, middle-aged poverty, and adult
poverty decline by 0.084%, 0.089%, and 0.085%, respectively, when ICT goods imported
increase by 1%. On the other hand, imported ICT goods reduce poverty in SSA because higher
importation of ICT goods enlarges the consumption possibilities of ICT goods and, therefore,
improves the well-being of the people. The evidence that imported ICT goods contribute to
poverty reduction is inconsistent with Lechman and Popowska's (2022) findings, which
revealed that imported ICT goods do not affect poverty in lower- and middle-income countries.
[INSERT TABLE 3A & 3B HERE]
In the baseline models (see Models 1, 8, and 15), it is observed that economic growth has an
insignificant effect on youth, middle-aged, and adult poverty. However, the estimated
coefficient of economic growth on youth, middle-aged, and adult poverty becomes positive
and statistically significant in the presence of telephone penetration and becomes negative and
statistically significant in the presence of ICT goods imported and broadband penetration. The
significantly significant effect of economic growth on poverty after the inclusion of the ICT
variables demonstrates the role ICT plays in economic growth to influence poverty. In the
literature, ICT is demonstrated to drive firm productivity and economic growth by enhancing
e-commerce, network externalities, productive efficiency, economic diversification, business
retention, cost reduction, improving human capital development, promoting social and political
stability, economic freedom, and information dissemination (Adeleye and Eboagu, 2019; Vu,
Hanafizadeh and Bohlin, 2020; Appiah-Otoo and Song, 2021b). Also, the significantly
significant effect of economic growth suggests that economic growth could be conditioned by
ICT to influence poverty reduction
6
.
Also, in the baseline models (see Models 1, 8, and 15), income inequality has a significant
positive effect on youth, middle-aged, and adult poverty; however, the estimated coefficient of
6
See the interactive effect results presented in 4.3 for more details discussion.
16
income inequality becomes negative and significant in the presence of telephone penetration.
As expected, income inequality drives poverty because higher income inequality contributes
to disruptive socio-political activities such as riots, crime, political turnovers, etc. (Alesina &
Perotti, 1996; Barro, 2000; Ezcurra, 2007; Perotti, 1996). Higher-income inequality distorts
market activities and deters investment, contributing to the rise of poverty. We expect income
inequality to have a significant positive effect on poverty (Barro, 2000; Bourguignon, 2004;
Fosu, 2010). However, it can be observed that income inequality lowers poverty when
telephone is controlled. This indicates that telephones could serve as a potential transition
channel mechanism through which income inequality can alleviate poverty in SSA.
Also, in the baseline models (see Models 1, 8, and 15), access to electricity has an insignificant
negative effect on youth, middle-aged, and adult poverty. However, the estimated coefficient
of access to electricity becomes negative and statistically significant when internet penetration
and ICT exported goods and positive when ICT imported goods are controlled. The
significantly significant effect of electricity on poverty after the ICT inclusion shows that ICT
serves as a medium through which electrification influences poverty. Access to electricity is
not an end in itself, but a means to an end (Acheampong, Dzator & Shahbaz, 2021). This
suggests that electricity is a critical input for ICT and its related technologies to function
effectively and contribute to development. Some studies have provided evidence that without
electricity, there would be no ICT for development (Armey & Hosman, 2016; Unwin & Unwin,
2009).
Further, in the baseline models (see Models 1, 8, and 15), trade openness has a statistically
insignificant negative effect on youth, middle-aged, and adult poverty. However, the impacts
become statistically significant in the presence of digitization variables. This evidence shows
that ICT promotes trade and that contributes to poverty reduction. The literature suggests that
costs relating to searching, advertising, and establishing network distribution impede trade
(Nath & Liu, 2017; Freund & Weinhold, 2004). ICT penetration is argued to lessen these
market costs, thereby increasing the volume of imports and export between countries (Nath &
Liu, 2017). Therefore, increasing trade volume due to ICT contributes to poverty reduction.
Theoretical evidence from the Heckscher-Ohlin model shows that countries gain from trade
when they export goods whose production is intensive in factors with which the countries are
abundantly endowed. SSA is a labor-endowed region, and labor income could increase
whenever there is an exportation of goods whose production requires intensive use of labor,
thereby reducing poverty. The role of trade in poverty eradication is consistent with the findings
of Acheampong, Appiah-Otoo, Dzator, and Agyemang (2021), Magombeyi and Odhiambo
(2018) and Le Goff, and Singh (2014).
The evidence also shows that access to domestic credit has a statistically insignificant positive
effect on youth, middle-aged, and adult poverty, which becomes significant only in the
presence of imported ICT goods. Consistent with the literature (see, for instance, Ravallion,
2002; Sulemana et al., 2019), our estimates further show that urbanization has a positive and
statistically significant effect on youth, middle-aged, and adult poverty. This evidence indicates
that during urbanization, people migrate from the rural areas, characterized by low income and
wages, to the urban sectors with higher income and wages. This movement initially led to the
poorest share of the total income declining while increasing the income of the relatively small
and wealthy people in the industrial and urban sectors, thus increasing income inequality and
poverty (Deininger & Squire, 1997; Barro, 2000; Robinson, 1976). Finally, the evidence shows
that physical capital has a statistically insignificant negative effect on youth, middle-aged, and
adult poverty.
17
4.2. Effect of ICT on gender poverty
It is indicated that men and women have unequal access to ICT in SSA (Adepetun, 2021; FAO,
2021). Therefore, we test the effect of ICT on gender poverty in SSA. Table 4A presents the
estimates for female youth and middle-aged poverty, while Table 4B shows the estimates for
female adult poverty. Also, Table 5A presents the estimates for male youth and middle-aged
poverty, while Table 5B presents the estimates for male adult poverty. The results from these
tables show that the coefficient of telephone penetration is negative and statistically significant
at a 5% level for female poverty measures and 1% for male poverty measures. For female
poverty measures, the estimate indicates that a 1% increase in telephone penetration reduces
female youth poverty, middle-aged poverty, and adult poverty by 0.063%, 0.059%, and
0.078%, respectively. Also, for the male poverty measures, a 1% increase in telephone
penetration reduces male youth poverty, middle-aged poverty, and adult poverty by 0.102%,
0.095%, and 0.101%, respectively. Also, the coefficient on mobile phone penetration is
negative and statistically significant at a 1% level for female and male poverty measures. For
female poverty measures, the estimate shows that a 1% increase in telephone penetration
reduces female youth poverty, middle-aged poverty, and adult poverty by 0.070%, 0.071%,
and 0.068%, respectively. Also, for the male poverty measures, a 1% increase in t mobile phone
penetration reduces male youth poverty, middle-aged poverty, and adult poverty by 0.068%,
0.063%, and 0.073%, respectively. These estimates suggest that telephone and mobile phone
penetration matters in reducing male and female poverty in SSA; however, on average, the
impact of telephone and mobile phones reducing male poverty is greater than reducing female
poverty. This observation reflects the inequality between males and females in using ICT
goods. Generally, the number of males using telephone and mobile phones in SSA exceeds that
of females. Evidence suggests there are still some 74 million unconnected women in SSA, as
13% lack access to telephone and mobile phones more than men in the region (Adepetun,
2021). This indicates women are less likely to benefit from telephone and mobile penetration
in SSA.
The results also indicate that the estimated internet penetration coefficient is positive and
statistically significant for female and male poverty measures. The estimated coefficient
suggests that female youth poverty, middle-aged poverty, and adult poverty increase by
0.008%, 0.009%, and 0.008%, respectively when internet penetration increases by 1%. Also,
for male poverty, a 1% increase in internet penetration increases male youth poverty, middle-
aged poverty, and adult poverty by 0.012%, 0.008%, and 0.013%, respectively. Also, the
estimated coefficient on broadband penetration is positive and statistically significant for
female and male poverty measures. The estimated coefficient suggests that female youth
poverty, middle-aged poverty, and adult poverty increase by 0.373%, 0.369%, and 0.325%,
respectively, when broadband penetration increases by 1%. The estimated coefficient suggests
that male youth poverty, middle-aged poverty, and adult poverty increase by 0.259%, 0.282%,
and 0.286%, respectively, when broadband penetration increases by 1%. These estimates also
suggest that, on average, the impact of internet and broadband penetration in increasing female
poverty is greater compared to increasing male poverty. This observation reflects the inequality
between males and females using internet resources in SSA.
[INSERT TABLE 4A & 4B HERE]
[INSERT TABLE 5A & 5B HERE]
18
For instance, in SSA, only 27% of women have access to the internet, and only 15% can
afford to use it (FAO, 2021). According to the Global System for Mobile Communications
GSMA (2019), the internet penetration rate for men and women in SSA was 33.8% and 22.6%,
respectively, in 2019. The internet user gender gap increased from 20.7% in 2013 to 37% in
2019 (FAO, 2021). These indicate that women are much less likely to benefit from broadband
and internet services.
The results also suggest that the estimated coefficient of ICT goods exported is positive and
statistically significant for female and male poverty measures. The estimated coefficient
indicates that female youth poverty, middle-aged poverty, and adult poverty increase by
0.028%, 0.022%, and 0.027%, respectively, when ICT goods exported increase by 1%. Also,
for male poverty, the estimated coefficient suggests that male youth poverty, middle-aged
poverty, and adult poverty increase by 0.025%, 0.021%, and 0.025%, respectively, when ICT
goods exported increase by 1%. On the other hand, the estimated coefficient of ICT goods
imported is negative and statistically significant for female and male poverty measures. For
female poverty, the estimated coefficient suggests that female youth poverty, middle-aged
poverty, and adult poverty decline by 0.073%, 0.062%, and 0.091%, respectively, when ICT
goods imported increase by 1%. Also, for male poverty, the estimated coefficient suggests that
male youth poverty, middle-aged poverty, and adult poverty decline by 0.089%, 0.069%, and
0.088%, respectively, when ICT goods imported increase by 1%. These estimates suggest that,
on average, ICT goods imported worsen female poverty more than male poverty, while ICT
goods exported have a greater impact on reducing male poverty relative to female poverty. This
finding substantiates the argument that the male population benefits substantially from
engaging in the trade of ICT goods and services compared to the female population. This
argument stems from the observation that males are more involved in trading in ICT services
in SSA than females.
4.3. Analysis of interactive effect results
4.3.1. Interactive effect of ICT variables and economic growth on poverty
Table 6A presents the interactive effect of ICT and economic growth on total poverty. Also,
based on Equation (6), we calculated the net effect of economic growth on poverty conditions
on different values of the ICT variables. Panel A of Table 6D presents the total/net effect of
economic growth on poverty conditioned on the different values of ICT penetration. From
Table 6, the results show that the interaction between telephone penetration and economic
growth has a statistically insignificant effect on youth poverty, middle-aged poverty, and adult
poverty. Also, the estimates suggest that the interaction between mobile phone penetration and
economic growth has a statistically significant positive effect on youth poverty, middle-aged
poverty, and adult poverty. Also, the estimates indicate that the interaction between internet
penetration and economic growth has a statistically significant positive effect on youth poverty,
middle-aged poverty, and adult poverty. Similarly, the estimates indicate that the interaction
between broadband penetration and economic growth has a statistically significant positive
effect on youth poverty, middle-aged poverty, and adult poverty. In the same vein, the
interaction between ICT goods exported and economic growth has a statistically significant
positive effect on youth poverty, middle-aged poverty, and adult poverty. Also, the interaction
between ICT goods imported and economic growth has a statistically insignificant impact on
youth poverty and middle-aged poverty while having a significant positive effect on adult
poverty.
19
The net effect estimates in Panel A of Table 6D show that economic growth significantly
increases youth, middle-aged, and adult poverty at the minimum, mean, and maximum values
of telephone penetration. It can be observed from the estimates that the net effect of economic
growth on the poverty indicators increases at the increasing values of telephone penetration. In
addition, at the minimum and mean values of mobile phone, internet, broadband, and ICT
imported goods, the net effect of economic growth shows a statistically significant negative
effect on youth, middle-aged and adult poverty, while at their maximum values, the net effect
of economic growth significantly increases youth, middle-aged and adult poverty. This
evidence further shows that at higher values of mobile phones, internet, broadband, and ICT
imported goods, economic growth contributes to poverty and vice versa. The net effect values
also suggest that at the minimum, mean, and maximum values of ICT imported, the net effect
of economic growth has a significantly negative relationship with youth, middle-aged and adult
poverty, indicating that ICT goods imported conditions economic growth to contribute to
poverty alleviation.
From the net effect results, at the increasing values of telephone, mobile phone, internet,
broadband, and ICT goods exported, economic growth increases poverty in SSA. This result
could be that the economic growth effect of these ICT variables is not shared equally among
people in SSA. It could also be that the economic growth effect of ICT is not trickle down to
the poorest households. Available evidence suggests that ICT boosts economic growth
(Adedoyin, Bekun, Driha, & Balsalobre-Lorente, 2020; Kallal et al., 2021); however, economic
growth does not always support poverty reduction (Škare & Družeta, 2016).
[INSERT TABLE 6A HERE]
4.3.2. Interactive effect of ICT variables and income inequality on poverty
Table 6B also presents the interactive effect of ICT and income inequality on total poverty.
Also, based on Equation (7), we calculated the net effect of income inequality on poverty
conditions on different values of the ICT variables. Panel B of Table 6D presents the total/net
effect of income inequality on poverty conditioned on the different values of ICT penetration.
The results from Table 7 show that the interaction between telephone penetration and income
inequality has a statistically insignificant effect on youth poverty, middle-aged poverty, and
adult poverty. The estimates show that the interaction between mobile phone penetration and
income inequality has a statistically significant positive effect on youth poverty, middle-aged
poverty, and adult poverty. Also, the estimates indicate that the interaction between internet
penetration and income inequality has a statistically significant negative effect on middle-aged
poverty and adult poverty while having an insignificant effect on youth poverty. Similarly, the
estimates indicate that the interaction between broadband penetration and income inequality
has a statistically significant negative effect on adult poverty while having a neutral impact on
youth poverty and middle-aged poverty. Also, the interaction between ICT goods exported and
income inequality has a statistically significant positive effect on youth poverty, middle-aged
poverty, and adult poverty. Also, the interaction between ICT goods imported and income has
a statistically insignificant impact on youth poverty, middle-aged poverty, and adult poverty.
The net effect estimates presented in Panel B of Table 6D show that at the minimum and mean
values of telephone penetration, the net effect of income inequality shows a negative effect on
youth, middle-aged and adult poverty, while at the maximum value, the net effect of income
inequality has a positive effect on youth, middle-aged and adult poverty. Also, at the minimum
value of mobile penetration, the net effect of income inequality shows a significant negative
20
effect on youth, middle-aged and adult poverty, while at the mean and maximum value, the net
effect of income inequality has a statistically significant positive effect on youth, middle-aged
and adult poverty. The estimates show that at the minimum and mean values of internet and
broadband penetration, the net effect of income inequality shows a positive effect on youth,
middle-aged and adult poverty, while at their maximum values, the net effect of income
inequality has a negative relationship with youth, middle-aged and adult poverty. The evidence
further suggests that at the minimum of ICT exported, the net effect of income inequality has
a statistically significant negative impact on youth, middle-aged and adult poverty, while at the
mean and maximum values, the net effect of income inequality has a statistically significant
positive impact on youth, middle-aged and adult poverty. Finally, at the minimum, mean and
maximum values of ICT goods imported, the net effect of income on youth, middle-aged, and
adult poverty is positive, but the impact is statistically significant only at the minimum and
means values of ICT goods imported.
These results suggest that when telephone penetration, mobile phone penetration, ICT goods
exported and imported increase (at the maximum), income inequality worsens poverty in SSA.
This result further suggests that increasing ICT penetration may not lead to equitable income
distribution, hence driving poverty in SSA. Using the insight from the skilled-bias argument,
these findings indicate that these ICT-related technologies are putting a premium on the
demand for skilled labor while displacing unskilled labor, thereby widening the income
inequality gap (Galperin & Fernanda Viecens, 2017). Since ICT is a skilled-based
technological change, it can contribute to the rise of income among skilled workers relative to
non-skilled workers. Similarly, Moll, Rachel, and Restrepo's (2022) automation and income
inequality theoretical framework shows that new technologies, including ICT-related
technologies, can result in wage stagnation and income stagnation at the bottom of the income
distribution. Evidence has shown that ICT could worsen income inequality (Carlos, 2010;
Akerman, Gaarder, and Mogstad, 2015). Therefore, our evidence shows that some dimensions
of ICT can contribute to poverty through widening income inequality. Contrarily, when internet
and broadband penetration increase (to their maximum), income inequality reduces poverty in
SSA. The implication is that increasing internet and broadband penetration in SSA would
bridge the income gap between the rich and poor, thereby reducing poverty.
[INSERT TABLE 6B HERE]
4.3.3. Interactive effect of ICT variables and access to credit on poverty
Table 6C also presents the interactive effect of ICT and access to credit on total poverty. Also,
based on Equation (8), we calculated the net effect of access to credit on poverty conditions on
different values of the ICT variables. Panel C of Table 6D presents the total/net effect of access
to credit on poverty conditioned on the different values of ICT penetration. The results from
Table 8 show that the interaction between telephone penetration and access to domestic credit
has a statistically insignificant effect on youth poverty, middle-aged poverty, and adult poverty.
The estimates show that the interaction between mobile phone penetration and access to
domestic credit has a statistically significant positive effect on youth poverty, middle-aged
poverty, and adult poverty. Also, the estimates indicate that the interaction between internet
penetration and access to domestic credit has a statistically significant positive effect on youth
poverty, middle-aged poverty, and adult poverty. Similarly, the estimates indicate that the
interaction between broadband penetration and access to domestic credit has a statistically
significant positive effect on youth poverty, middle-aged poverty, and adult poverty. Also, the
interaction between ICT goods exported and access to domestic credit has a statistically
significant negative effect on youth poverty, middle-aged poverty, and adult poverty. Also, the
21
interaction between ICT goods imported and access to domestic credit has a statistically
positive impact on youth poverty, middle-aged poverty, and adult poverty.
The net effect estimates presented in Panel C of Table 6D show that at the minimum values of
telephone, mobile phone, internet, broadband, and ICT imported goods, the net effect of access
to credit shows a negative effect on youth, middle-aged and adult poverty, while at their mean
and maximum values, the net effect of access to credit increases youth, middle-aged and adult
poverty. On the other hand, at the minimum and mean values of ICT goods exported, the net
effect of access to credit shows a positive effect on youth, middle-aged and adult poverty, while
at the maximum value, the net effect of access to credit shows a negative effect on youth,
middle-aged and adult poverty.
From the net effect estimates, except for ICT goods exported, when telephone, mobile phone,
internet, broadband, and ICT goods imported increase, access to credit increases poverty in
SSA. It is argued that ICT can reduce poverty by reducing information asymmetry and
transaction costs associated with the traditional financial market and, thus, improving access
to credit (Cheng, Chien, and Lee, 2021; Yang et al., 2021). However, our results suggest that
telephone, mobile phone, internet, broadband, and ICT goods imported condition access to
credit would increase poverty in SSA. We argue that ICT can condition access to credit to
reduce poverty only when borrowing cost is relatively low. However, the borrowing cost in
SSA remains high despite the region experiencing the fastest growth in internet connectivity
and telephone and mobile phone penetration (FAO, 2021). On the other hand, we observed that
domestic credit access reduced SSA poverty when ICT goods exported increased.
[INSERT TABLE 6C HERE]
[INSERT TABLE 6D HERE]
5. Conclusion and policy implications
SDG 1 requires countries to end poverty of all forms, including eradicating poverty for all
people living on less than $1.25 a day and reducing by half the proportion of men, women, and
children of all ages living in poverty by 2030. In the same vein, SDG 9 calls for building
resilient infrastructure, promoting inclusive and sustainable industrialization, and fostering
innovation. Specifically, SDG 9C calls for increasing ICT access and striving to provide
universal and affordable access to the internet in the least developed countries by 2020. While
the existing studies have explored the effect of ICT on economic growth in SSA, there is a
paucity of empirical studies on the effect of ICT on poverty reduction in SSA. Therefore, this
article examines ICT's effect on poverty reduction using comprehensive data for 44 sub-
Saharan Africa (SSA) countries from 2010 to 2019 using the two-step dynamic system-GMM
estimator to control endogeneity. In achieving the objective of this paper, our study contributes
to the literature and policy discussions by providing empirical answers to the following specific
research questions (i) Do broadband, telephone, ICT goods exported and imported, internet
usage, and mobile cellular subscriptions contribute to poverty reduction in SSA? (ii) Do
broadband, telephone, ICT goods exported and imported, internet usage, and mobile cellular
subscriptions have a heterogeneous effect on gender poverty in SSA? (iii) Do broadband,
22
telephone, ICT goods exported and imported, internet usage, and mobile cellular subscriptions
condition economic growth, income inequality, and access to credit to affect poverty in SSA?
Generally, this study's findings show that the impact of ICT on poverty reduction depends on
the specific ICT variable. The evidence shows that while telephone penetration, mobile phone
penetration, and ICT goods imported contribute to poverty reduction, internet penetration,
broadband penetration, and ICT goods exported increase poverty. Our study is not the first to
show that different dimensions of ICT impact economic development outcomes differently.
For instance, Richmond and Triplett (2018) demonstrated that the impact of ICT on income
inequality depends on the variable used as a proxy for ICT and that while mobile phones and
the internet contribute to income inequality reduction, broadband widens income inequality.
We argue that the inability of internet and broadband to poverty reduction is the low adoption
rate of internet and broadband in the region coupled with higher internet and broadband costs
in the region. Globally, SSA is the region with poor access to ICT (Mahler, Montes, &
Newhouse, 2019). As of 2017, only 1 in 5 in Sub-Saharan Africa used the internet, while 23%
of the SSA population used the mobile internet regularly as of 2018 (Okeleke & Suardi, 2019).
Also, poor telecommunication and ancillary infrastructures such as internet cables and an
electricity grid in SSA could explain the positive effect of internet and broadband penetration
on poverty (Langmia, 2006).
Evidence from this study suggests that ICT has a disparate effect on gender poverty. It is
evident from this study that while telephone penetration, mobile phone penetration, and ICT
goods imported contribute to male and female poverty reduction, internet penetration,
broadband penetration, and ICT goods exported increase male and female poverty. However,
the estimated elasticity of telephone penetration, mobile phone penetration, and ICT goods
imported are higher for male poverty, while the estimated elasticity of internet penetration,
broadband penetration, and ICT goods exported are higher for female poverty. Finally, our
study shows that the impact of ICT on poverty reduction is not always direct. The findings
from the net effect estimates based on interaction/conditional effect analysis indicate that at the
maximum value of most ICT variables, economic growth, income inequality, and access to
credit significantly increase the poverty rate regardless of age group. Thus, in the context of
SSA, ICT can worsen poverty through income inequality, economic growth, and access to
credit.
These findings have significant policy implications for poverty reduction in SSA. From these
results, we argue that ICT matters for poverty alleviation in SSA, but it depends on what is
used to capture ICT in the literature. We also argue that ICT has a disparate effect on gender
poverty in SSA by benefiting men more than women. We also argue that while ICT can directly
contribute to poverty reduction, it can indirectly worsen poverty by conditioning economic
growth, income inequality, and access to credit. These findings have important implications
for policy in SSA. Our study calls for policies that facilitate ICT penetration and usage in SSA.
To ensure universal access to ICT in SSA, policymakers need to implement strategies that
minimize the cost of internet, mobile phone, and telephone usage. Evidence from the
Worldwide Mobile Data Pricing 2021 reveals that SSA is the most expensive region, with the
average price for 1 gigabyte of mobile data coming in at $6.44
7
. Given the SSA's low
socioeconomic status, most of the population cannot afford to purchase mobile data, limiting
their capability to function. We, therefore, recommended policymakers subsidize the cost of
internet and other ICT services usage in the region to ensure universal accessibility of digital
technologies in the region. In addition, ensuring universal access to ICT requires policymakers
to provide an enabling environment that supports an expansion of ICT infrastructure in the
7
https://www.cable.co.uk/mobiles/worldwide-data-pricing/#highlights
23
region. Therefore, governments in SSA implementing a regulatory framework that ensures
enabling investment climate and prohibits monopoly in the telecommunication industry are
critical for improving access to ICT and eliminating the digital divide in SSA. Further, our
findings indicate that ICT benefits males more than females reflecting the gap in ICT
accessibility between males and females in the region. Therefore, a gender-responsive and
inclusive approach is needed to bridge the ICT access gap between males and females in SSA.
Developing a gender-responsive and inclusive approach to close the ICT access gap between
males and females in SSA requires policymakers to understand the factors that
disproportionately limit ICT usage among women and girls in the region.
Despite the contributions of this paper, there are still some avenues for future research. First,
our study examined if ICT moderates the effect of economic growth, access to credit, and
income inequality on poverty. However, there are many variables, including employment,
environmental degradation, trade, and gender empowerment, institutions, among others, that
ICT can interact with to affect poverty. We, therefore, recommend future studies to consider if
ICT moderates some of these variables when investigating the ICT-poverty relationship. Also,
future studies would contribute to the literature by conducting a mediation analysis, which is
different from moderation/interaction analysis, to examine some of the potential channels
through which ICT influences poverty. Given that our study focused on SSA, future studies
can extend this study to other developing countries in Asia-Pacific, and Southern America, to
guide policymakers in developing holistic policies that concurrently drive technological
changes while achieving better social and economic development outcomes.
Declaration of Conflict of Interest: The authors declare no known competing financial
interests or personal relationships that could have appeared to influence the work reported in
this paper.
Acknowledgments
The authors sincerely thank the Editor-in-Chief and the Handling Editor for their time and
support. We are also grateful to the five anonymous reviewers for their valuable comments that
help improve this paper's quality.
APPENDIX
[INSERT APPENDIX TABLE 1 HERE]
24
References
Abosedra, S., Shahbaz, M., & Nawaz, K. (2016). Modeling Causality Between Financial
Deepening and Poverty Reduction in Egypt. Social Indicators Research, 126(3), 955-
969. doi:10.1007/s11205-015-0929-2
Acheampong, A. O. (2019). Modelling for insight: Does financial development improve
environmental quality? Energy Economics, 83, 156-179.
doi:https://doi.org/10.1016/j.eneco.2019.06.025
Acheampong, A. O., Appiah-Otoo, I., Dzator, J., & Agyemang, K. K. (2021). Remittances,
financial development and poverty reduction in Sub-Saharan Africa: Implications for
post-COVID-19 macroeconomic policies. Journal of Policy Modeling.
doi:https://doi.org/10.1016/j.jpolmod.2021.09.005
Acheampong, A. O., Dzator, J., & Shahbaz, M. (2021). Empowering the powerless: Does
access to energy improve income inequality? Energy Economics, 99, 105288.
doi:https://doi.org/10.1016/j.eneco.2021.105288
Acheampong, A. O., Erdiaw-Kwasie, M. O., & Abunyewah, M. (2021). Does energy
accessibility improve human development? Evidence from energy-poor regions.
Energy Economics, 96, 105165. doi:https://doi.org/10.1016/j.eneco.2021.105165
Ackermann, K., Churchill, S. A. and Smyth, R. (2021) ‘Mobile phone coverage and violent
conflict,’ Journal of Economic Behavior & Organization. North-Holland, 188, pp. 269–
287. doi: 10.1016/j.jebo.2021.04.038.
Adams Jr, R. H. (2004). Economic growth, inequality and poverty: estimating the growth
elasticity of poverty. World Development, 32(12), 1989-2014.
ADB. (2021). From debt resolution to growth: The Road ahead of Africa. Africa Development
Bank Group Publications.
Adeabah, D., Asongu, S., & Andoh, C. (2021). Remittances, ICT and pension income
coverage: The international evidence. Technological Forecasting and Social Change,
173, 121148. doi:https://doi.org/10.1016/j.techfore.2021.121148
Adedoyin, F. F., Bekun, F. V., Driha, O. M., & Balsalobre-Lorente, D. (2020). The effects of
air transportation, energy, ICT and FDI on economic growth in the industry 4.0 era:
Evidence from the United States. Technological Forecasting and Social Change, 160,
120297. doi:https://doi.org/10.1016/j.techfore.2020.120297
Adedoyin, O. B. and Soykan, E. (2020) ‘Covid-19 pandemic and online learning: the
challenges and opportunities’, https://doi.org/10.1080/10494820.2020.1813180.
Routledge. doi: 10.1080/10494820.2020.1813180.
Adeleye, N. and Eboagu, C. (2019) ‘Evaluation of ICT development and economic growth in
Africa’, NETNOMICS: Economic Research and Electronic Networking.
NETNOMICS: Economic Research and Electronic Networking, pp. 31–53. doi:
10.1007/s11066-019-09131-6.
Adepetun, A. (2021). 74m women unconnected as 13% lack access to telephone. The
Guardian. Retrieved from https://guardian.ng/technology/74m-women-unconnected-
as-13-lack-access-to-telephone/
Aghion, P., & Howitt, P. (1992). A model of growth through creative destruction.
Econometrica, 60(2), 323-351.
Aker, J. C., Ghosh, I., & Burrell, J. (2016). The promise (and pitfalls) of ICT for agriculture
initiatives. Agricultural Economics, 47(S1), 35-48.
Akerman, A., Gaarder, I., & Mogstad, M. (2015). The skill complementarity of broadband
internet. The Quarterly Journal of Economics, 130(4), 1781-1824.
25
Amuakwa-Mensah, F., Klege, R. A., Adom, P. K., Amoah, A., & Hagan, E. (2018). Unveiling
the energy saving role of banking performance in Sub-Sahara Africa. Energy
Economics, 74, 828-842. doi:https://doi.org/10.1016/j.eneco.2018.07.031
Anderson, E., d'Orey, M. A. J., Duvendack, M., & Esposito, L. (2018). Does Government
Spending Affect Income Poverty? A Meta-regression Analysis. World Development,
103, 60-71. doi:https://doi.org/10.1016/j.worlddev.2017.10.006
Anyanwu, J. C. (2014) ‘Factors Affecting Economic Growth in Africa: Are There any Lessons
from China?’, African Development Review, 26(September 2014), pp. 468–493. doi:
10.1111/1467-8268.12105.
Anyanwu, J. C., & Anyanwu, J. C. (2017). The key drivers of poverty in Sub-Saharan Africa
and what can be done about it to achieve the poverty sustainable development goal.
Asian Journal of Economic Modelling, 5(3), 297-317.
Appiah-Otoo, I. and Song, N. (2021a) ‘The Impact of Fintech on Poverty Reduction: Evidence
from China’, Sustainability (Switzerland), pp. 1–13. doi:
https://doi.org/10.3390/su13095225.
Appiah-Otoo, I. and Song, N. (2021b) ‘The impact of ICT on economic growth-Comparing
rich and poor countries’, Telecommunications Policy. Pergamon, 45(2), p. 102082. doi:
10.1016/j.telpol.2020.102082.
Arellano, M., & Bond, S. (1991). Some tests of specification for panel data: Monte Carlo
evidence and an application to employment equations. Review of Economic Studies,
58(2), 277–297.
Armey, L. E., & Hosman, L. (2016). The centrality of electricity to ICT use in low-income
countries. Telecommunications Policy, 40(7), 617-627.
doi:https://doi.org/10.1016/j.telpol.2015.08.005
Asongu, S. A. and Odhiambo, N. M. (2020) ‘Foreign direct investment, information technology
and economic growth dynamics in Sub-Saharan Africa’, Telecommunications Policy.
Elsevier Ltd, 44(1), p. 101838. doi: 10.1016/j.telpol.2019.101838.
Asongu, S. A., & Acha-Anyi, P. N., (2019). " The Murder Epidemic: A Global Comparative Study,"
International Criminal Justice Review, 29 issue: 2, page(s): 105-120.
Asongu, S. A., & Le Roux, S. (2017). Enhancing ICT for inclusive human development in Sub-
Saharan Africa. Technological Forecasting and Social Change, 118, 44-54.
doi:https://doi.org/10.1016/j.techfore.2017.01.026
Asongu, S. A., Le Roux, S., & Biekpe, N., (2017). "Environmental degradation, ICT and
inclusive development in Sub-Saharan Africa", Energy Policy, 111(December), pp.
353-361.
Asongu, S. A., Nwachukwu, J. C., & Pyke, C. (2019). The Comparative Economics of ICT,
Environmental Degradation and Inclusive Human Development in Sub-Saharan Africa.
Social Indicators Research, 143(3), 1271-1297. doi:10.1007/s11205-018-2009-x
Bailard, C. S. (2015) ‘Ethnic conflict goes mobile: Mobile technology’s effect on the
opportunities and motivations for violent collective action’, Journal of Peace Research.
SAGE Publications Ltd, 52(3), pp. 323–337. doi: 10.1177/0022343314556334.
Bao, W. (2020) ‘COVID-19 and online teaching in higher education: A case study of Peking
University’, Human Behavior and Emerging Technologies. John Wiley & Sons, Ltd,
2(2), pp. 113–115. doi: 10.1002/HBE2.191.
Barro, R. J. (2000). Inequality and Growth in a Panel of Countries. Journal of Economic
Growth, 5(1), 5-32.
Begum, S. S., Deng, Q. and Gustafsson, B. (2012) ‘Economic growth and child poverty
reduction in Bangladesh and China’, Journal of Asian Economics, 23(1), pp. 73–85.
doi: 10.1016/j.asieco.2011.10.001.
26
Bertot, J. C., Jaeger, P. T. and Grimes, J. M. (2010) ‘Using ICTs to create a culture of
transparency: E-government and social media as openness and anti-corruption tools for
societies’, Government Information Quarterly. Elsevier B.V., 27(3), pp. 264–271. doi:
10.1016/j.giq.2010.03.001.
Beuermann, D. W., McKelvey, C. and Vakis, R. (2012) ‘Mobile Phones and Economic
Development in Rural Peru’, Journal of Development Studies, 48(11), pp. 1617–1628.
doi: 10.1080/00220388.2012.709615.
Beuermann, D. W., McKelvey, C., & Vakis, R. (2012). Mobile Phones and Economic
Development in Rural Peru. The Journal of Development Studies, 48(11), 1617-1628.
doi:10.1080/00220388.2012.709615
Bhagwati, J., & Srinivasan, T. N. (2002). Trade and poverty in the poor countries. American
Economic Review, 92(2), 180-183.
Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel
data models. Journal of Econometrics, 87(1), 115-143.
doi:https://doi.org/10.1016/S0304-4076(98)00009-8
Bourguignon, F. (2004). The poverty-growth-inequality triangle. Working Paper, No. 125,
Indian Council for Research on International Economic Relations (ICRIER), New
Delhi
Brambor, T., Clark, W. M., & Golder, M., (2006). "Understanding Interaction Models:
Improving Empirical Analyses," Political Analysis, 14 (1), pp. 63-82.
Carlos, A. (2010). Impacto del Uso de Internet en el Bienestar de los Hogares Peruanos.
Retrieved from
Chang, H.-H. and Just, D. R. (2009) ‘Internet Access and Farm Household Income – Empirical
Evidence using a Semi-parametric Assessment in Taiwan,’ Journal of Agricultural
Economics. John Wiley & Sons, Ltd, 60(2), pp. 348–366. doi: 10.1111/J.1477-
9552.2008.00189.X.
Cheng, C., Chien, M. and Lee, C. (2021) ‘ICT diffusion , financial development, and economic
growth : An international cross-country analysis’, Economic Modelling. Elsevier Ltd,
94(February 2020), pp. 662–671. doi: 10.1016/j.econmod.2020.02.008.
Chin, W. (2019) ‘Technology, war and the state: Past, present and future’, International Affairs.
Oxford University Press, pp. 765–783. doi: 10.1093/ia/iiz106.
Cocosila, M., & Archer, N. (2010). Adoption of mobile ICT for health promotion: An empirical
investigation. Electronic Markets, 20(3), 241-250.
Co-operation, O. f. E., & Development. (2018). Bridging the digital gender divide include,
upskill, innovate: OECD, Paris, France.
Dafoe, A. and Lyall, J. (2015) ‘From cell phones to conflict? Reflections on the emerging ICT–
political conflict research agenda’, Journal of Peace Research. SAGE Publications Ltd,
52(3), pp. 401–413. doi: 10.1177/0022343314563653.
Dai, X. and Sun, Z. (2021) ‘Does firm innovation improve aggregate industry productivity?
Evidence from Chinese manufacturing firms’, Structural Change and Economic
Dynamics. Elsevier B.V., 56, pp. 1–9. doi: 10.1016/j.strueco.2020.09.005.
Deganis, I., Haghian, P. Z., & Tagashira, M. (2021). Leveraging digital technologies for social
inclusion. Retrieved from https://www.un.org/development/desa/dspd/2021/02/digital-
technologies-for-social-inclusion/
Deininger, K., & Squire, L. (1997). Economic growth and income inequality: reexamining the
links. Finance and development, 34, 38-41.
Dhaoui, I. (2021) ‘E-Government for Sustainable Development: Evidence from MENA
Countries’, Journal of the Knowledge Economy, (0123456789). doi: 10.1007/s13132-
021-00791-0.
27
Diga, K., Nwaiwu, F., & Plantinga, P. (2013). ICT policy and poverty reduction in Africa. info,
15(5), 114-127. doi:10.1108/info-05-2013-0032
Dollar, D. and Kraay, A. (2004) ‘Growth is Good for the Poor’, Growth, Inequality, and
Poverty: Prospects for Pro-poor Economic Development, pp. 195–225. doi:
10.1093/0199268657.003.0002.
Dollar, D., & Kraay, A. (2002). Growth is Good for the Poor. Journal of Economic Growth,
7(3), 195-225. doi:10.1023/A:1020139631000
Etim, N. A. A., & Edet, G. E. (2014). Does Asset Ownership Reduce Chronic Poverty?.
European Scientific Journal, 10(1).
FAO. (2021). Gender-responsive digitalization: A critical component of the COVID-19
response in Africa. doi:https://doi.org/10.4060/cb5055en
Fosu, A. K. (2010). Inequality, Income, and Poverty: Comparative Global Evidence*. Social
science quarterly, 91(5), 1432-1446. doi:https://doi.org/10.1111/j.1540-
6237.2010.00739.x
Fowowe, B., & Shuaibu, M. I. (2014). Is foreign direct investment good for the poor? New
evidence from African countries. Economic Change and Restructuring, 47(4), 321-339.
doi:10.1007/s10644-014-9152-4
Freund, C. L. and Weinhold, D. (2004) ‘The effect of the Internet on international trade, Journal
of International Economics, 62, pp. 171–189. doi: 10.1016/S0022-1996(03)00059-X.
Freund, C. L., & Weinhold, D. (2004). The effect of the Internet on international trade. Journal
of International Economics, 62(1), 171-189.
Galperin, H. and Fernanda Viecens, M. (2017) ‘Connected for Development? Theory and
evidence about the impact of Internet technologies on poverty alleviation’,
Development Policy Review. Blackwell Publishing Ltd, 35(3), pp. 315–336. doi:
10.1111/DPR.12210.
GSMA. (2019). The mobile gender gap report 2019. Retrieved from London:
https://www.gsma.com/mobilefordevelopment/wp-content/uploads/2019/03/GSMA-
Connected-Women-The-Mobile-Gender-Gap-Report-2019.pdf.
Haenssgen, M. J., & Ariana, P. (2018). The place of technology in the Capability Approach.
Oxford Development Studies, 46(1), 98-112. doi:10.1080/13600818.2017.1325456
Horn, A. (2014). The relationship between access to Information and Communications
Technology (ICT) and poverty in South Africa. Africanus, 44(1), 21-33.
Iman, N. (2018) ‘Is mobile payment still relevant in the fintech era?’, Electronic Commerce
Research and Applications, 30(May), pp. 72–82. doi: 10.1016/j.elerap.2018.05.009.
Iniguez-Montiel, A. J. (2014) ‘Growth with equity for the development of Mexico: Poverty,
inequality, and economic growth (1992-2008)’, World Development. Elsevier Ltd, 59,
pp. 313–326. doi: 10.1016/j.worlddev.2014.01.011.
Inoue, T. (2018) Financial development, remittances, and poverty reduction: Empirical
evidence from a macroeconomic viewpoint. Journal of Economics and Business.
Elsevier Inc., 96, pp. 59–68. doi: 10.1016/j.jeconbus.2017.12.001.
Judson, R. A., & Owen, A. L. (1999). Estimating dynamic panel data models: a guide for
macroeconomists. Economics Letters, 65(1), 9-15. doi:https://doi.org/10.1016/S0165-
1765(99)00130-5
Kallal, R., Haddaji, A., & Ftiti, Z. (2021). ICT diffusion and economic growth: Evidence from
the sectorial analysis of a periphery country. Technological Forecasting and Social
Change, 162, 120403. doi:https://doi.org/10.1016/j.techfore.2020.120403
Kelles-Viitanen, A. (2003). The role of ICT in poverty reduction. The Finnish Economy and
Society, 1, 82-94.
Koomson, I., Villano, R. A., & Hadley, D. (2020). Effect of Financial Inclusion on Poverty and
Vulnerability to Poverty: Evidence Using a Multidimensional Measure of Financial
28
Inclusion. Social Indicators Research, 149(2), 613-639. doi:10.1007/s11205-019-
02263-0
Krugman, P., Obstfeld, M., & Melitz, M. (2017). International Economics: Theory and Policy,
the latest edition: Addison-Wesley.
Langmia, K. (2006). The role of ICTs in the economic development of Africa: The case of
South Africa. International Journal of education and Development using ICT, 2(4).
Le Goff, M., & Singh, R. J. (2014). Does trade reduce poverty? A view from Africa. Journal
of African Trade, 1(1), 5-14. doi:https://doi.org/10.1016/j.joat.2014.06.001
Lechman, E., & Popowska, M. (2022). Harnessing digital technologies for poverty reduction.
Evidence for low-income and lower-middle income countries. Telecommunications
Policy, 46(6), 102313. doi:https://doi.org/10.1016/j.telpol.2022.102313
Li, L. et al. (2019) Poverty alleviation through government-led e-commerce development in
rural China: An activity theory perspective, Information Systems Journal. doi:
10.1111/isj.12199.
Ma, W. and Wang, X. (2020) ‘Internet Use, Sustainable Agricultural Practices and Rural
Incomes: Evidence from China’, Australian Journal of Agricultural and Resource
Economics. Blackwell Publishing Ltd, 64(4), pp. 1087–1112. doi: 10.1111/1467-
8489.12390.
Ma, W. et al. (2020) ‘Impact of Internet use on economic well-being of rural households:
Evidence from China’, Review of Development Economics. Blackwell Publishing Ltd,
24(2), pp. 503–523. doi: 10.1111/RODE.12645.
Magombeyi, M. T., & Odhiambo, N. M. (2018). Dynamic impact of fdi inflows on poverty
reduction: Empirical evidence from South Africa. Sustainable Cities and Society, 39,
519-526. doi:https://doi.org/10.1016/j.scs.2018.03.020
Mahembe, E., & Odhiambo, N. M. (2021). Does foreign aid reduce poverty? A dynamic panel
data analysis for sub-Saharan African countries. The Journal of Economic Inequality.
doi:10.1007/s10888-021-09496-5
Mahler, D. G., Montes, J., & Newhouse, D. L. (2019). Internet access in Sub-Saharan
Africa(Issue i). Retrieved from
http://documents.worldbank.org/curated/en/518261552658319590/Internet-Access-in-
Sub-Saharan-Africa
May, J., Waema, T. M., & Bjåstad, E. (2014). Introduction: The ICT/poverty nexus in Africa.
ICT pathways to poverty reduction: Empirical evidence from East and Southern Africa,
1-32.
Miao, M. and Jayakar, K. (2016) ‘Mobile payments in Japan, South Korea and China: Cross-
border convergence or divergence of business models?’, Telecommunications Policy.
Elsevier Ltd, 40(2–3), pp. 182–196. doi: 10.1016/j.telpol.2015.11.011.
Moll, B., Rachel, L., & Restrepo, P. (2022). Uneven growth: automation's impact on income
and wealth inequality. Econometrica, 90(6), 2645-2683.
Mora-Rivera, J. and García-Mora, F. (2021) ‘Internet access and poverty reduction: Evidence
from rural and urban Mexico,’ Telecommunications Policy. Pergamon, 45(2), p.
102076. doi: 10.1016/J.TELPOL.2020.102076.
Mora-Rivera, J., & García-Mora, F. (2021). Internet access and poverty reduction: Evidence
from rural and urban Mexico. Telecommunications Policy, 45(2), 102076.
doi:https://doi.org/10.1016/j.telpol.2020.102076
Moser, C. (2006). Asset-based approaches to poverty reduction in a globalized context. Global
Economy and Development Working Paper, (01).
Mushtaq, R. and Bruneau, C. (2019) ‘Microfinance, financial inclusion and ICT: Implications
for poverty and inequality’, Technology in Society. Elsevier Ltd, 59. doi:
10.1016/J.TECHSOC.2019.101154.
29
Mushtaq, R., & Bruneau, C. (2019). Microfinance, financial inclusion and ICT: Implications
for poverty and inequality. Technology in Society, 59, 101154.
doi:https://doi.org/10.1016/j.techsoc.2019.101154
Nath, H. K., & Liu, L. (2017). Information and communications technology (ICT) and services
trade. Information Economics and Policy, 41, 81-87.
doi:https://doi.org/10.1016/j.infoecopol.2017.06.003
North, D. C. (1990) Institutions, institutional change and economic performance. doi:
10.1017/cbo9781139175302.016.
Ofori, I. K., Armah, M. K., Taale, F., & Ofori, P. E. (2021). Addressing the severity and
intensity of poverty in Sub-Saharan Africa: how relevant is the ICT and financial
development pathway? Heliyon, 7(10), e08156.
doi:https://doi.org/10.1016/j.heliyon.2021.e08156
Ofori, I. K., Armah, M. K., Taale, F., & Ofori, P. E. (2021). Addressing the severity and
intensity of poverty in Sub-Saharan Africa: how relevant is the ICT and financial
development pathway?. Heliyon, 7(10), e08156.
Okeleke, K., & Suardi, S. (2019). The mobile economy sub-Saharan Africa 2019. GSMA
Intelligence Report. Retrieved from https://data.gsmaintelligence.com/api-
web/v2/research-file-download?id=45121567&file=2794-160719-ME-SSA.pdf
Olaniyi, C. O. and Oladeji, S. I. (2020) Moderating the effect of institutional quality on the
finance–growth nexus: insights from West African countries, Economic Change and
Restructuring. Springer US. doi: 10.1007/s10644-020-09275-8.
Perera, L. D. H. and Lee, G. H. Y. (2013) ‘Have economic growth and institutional quality
contributed to poverty and inequality reduction in Asia?’, Journal of Asian Economics.
Elsevier Inc., 27, pp. 71–86. doi: 10.1016/j.asieco.2013.06.002.
Piachaud, D. (2002). Capital and the determinants of poverty and social exclusion. LSE
STICERD Research Paper No. CASE060.
Pradhan, R. P. et al. (2014) ‘Economic growth and the development of telecommunications
infrastructure in the G-20 countries: A panel-VAR approach’, Telecommunications
Policy. Elsevier, 38(7), pp. 634–649. doi: 10.1016/j.telpol.2014.03.001.
Pradhan, R. P., Arvin, M. B., Nair, M. S., Hall, J. H., & Bennett, S. E. (2021). Sustainable
economic development in India: The dynamics between financial inclusion, ICT
development, and economic growth. Technological Forecasting and Social Change,
169, 120758. doi:https://doi.org/10.1016/j.techfore.2021.120758
Quartey, P. (2005) Financial sector development, savings mobilisation and poverty reduction
in Ghana. The Open Access Publication Server of the ZBW – Leibniz Information
Centre for Economics. doi: 10.1057/9780230594029_5.
Ravallion, M. (1997). Can high-inequality developing countries escape absolute poverty?
Economics Letters, 56(1), 51-57. doi:https://doi.org/10.1016/S0165-1765(97)00117-1
Ravallion, M. (2002). On the urbanization of poverty. Journal of Development Economics,
68(2), 435-442. doi:https://doi.org/10.1016/S0304-3878(02)00021-4
Rehman, I. U. & Shahbaz, M. (2014) Multivariate-based Granger causality between financial
deepening and poverty: the case of Pakistan. Quality and Quantity. Kluwer Academic
Publishers, 48(6), pp. 3221–3241. doi: 10.1007/s11135-013-9952-z.
Rice, R. E., & Katz, J. E. (2003). Comparing internet and mobile phone usage: digital divides
of usage, adoption, and dropouts. Telecommunications Policy, 27(8), 597-623.
doi:https://doi.org/10.1016/S0308-5961(03)00068-5
Richmond, K., & Triplett, R. E. (2018). ICT and income inequality: a cross-national
perspective. International Review of Applied Economics, 32(2), 195-214.
Robinson, S. (1976). A note on the U hypothesis relating income inequality and economic
development. The American economic review, 66(3), 437-440.
30
Rodríguez, J. G. and Sánchez-Riofrío, A. (2017) ‘ICTs and Poverty in Latin America’, Íconos
- Revista de Ciencias Sociales. Facultad Latinoamericana de Ciencias Sociales,
Ecuador (FLACSO), (57), p. 141. doi: 10.17141/ICONOS.57.2017.2095.
Romer, P. M. (1990). Endogenous technological change. Journal of Political Economy, 98(5,
Part 2), S71-S102.
Sassi, S. and Ben Ali, M. S. (2017) ‘Corruption in Africa: What role does ICT diffusion play’,
Telecommunications Policy. Elsevier Ltd, 41(7–8), pp. 662–669. doi:
10.1016/j.telpol.2017.05.002.
Sen, A. (1980). Equality of what? The Tanner lecture on human values, 1, 197-220.
Sen, A. (2010). The mobile and the world. Information Technologies & International
Development, 6(SE), pp-1.
Sen, A. K. (1999). Development as freedom. Oxford: Oxford University Press.
Sife, A. S., Kiondo, E., & Lyimo-Macha, J. G. (2010). Contribution of Mobile Phones to Rural
Livelihoods and Poverty Reduction in Morogoro Region, Tanzania. The Electronic
Journal of Information Systems in Developing Countries, 42(1), 1-15.
doi:https://doi.org/10.1002/j.1681-4835.2010.tb00299.x
Škare, M., & Družeta, R. P. (2016). Poverty and economic growth: a review. Technological
and Economic Development of Economy, 22(1), 156-175.
doi:10.3846/20294913.2015.1125965
Solt, F. (2016). The standardized world income inequality database. Social science quarterly,
97(5), 1267-1281.
Steele, C. (2018). The Impacts of Digital Divide. Retrieved from
http://www.digitaldividecouncil.com/the-impacts-of-digital-divide/
Sujarwoto, S., & Tampubolon, G. (2016). Spatial inequality and the Internet divide in Indonesia
2010–2012. Telecommunications Policy, 40(7), 602-616.
Sulemana, I., Nketiah-Amponsah, E., Codjoe, E. A., & Andoh, J. A. N. (2019). Urbanization
and income inequality in Sub-Saharan Africa. Sustainable Cities and Society, 48,
101544. doi:https://doi.org/10.1016/j.scs.2019.101544
Tang, W., & Zhu, J. (2020). Informality and rural industry: Rethinking the impacts of E-
Commerce on rural development in China. Journal of Rural Studies, 75, 20-29.
Tchamyou, V. S., &Asongu, S. A., (2017)." Information Sharing and Financial Sector Development in
Africa," Journal of African Business, 18(7), pp. 24-49.
Tchamyou, V. S., (2019)."The Role of Information Sharing in Modulating the Effect of Financial
Access on Inequality.” Journal of African Business, 20(3), pp. 317-338.
Tchamyou, V.S., Erreygers, G., &Cassimon, D., (2019). "Inequality, ICT and Financial Access in
Africa," Technological Forecasting and Social Change,139(February), pp. 169- 184.
Thompson, H. G. and Garbacz, C. (2011) ‘Economic impacts of mobile versus fixed
broadband’, Telecommunications Policy. Elsevier, 35(11), pp. 999–1009. doi:
10.1016/j.telpol.2011.07.004.
UN. (2019). The Sustainable Development Goals Report 2019. Retrieved from
https://unstats.un.org/sdgs/report/2019/
Unwin, P., & Unwin, T. (2009). ICT4D: Information and communication technology for
development: Cambridge University Press.
Urean, C. A., Lacatus, V. D., & Mocean, L. (2016). Information And Communications
Technology As A Poverty Reduction Tool. Annals-Economy Series, 1, 107-113.
Vu, K. M. (2011) ‘ICT as a source of economic growth in the information age: Empirical
evidence from the 19962005 period’, Telecommunications Policy. Elsevier, 35(4), pp.
357–372. doi: 10.1016/j.telpol.2011.02.008.
31
Vu, K., Hanafizadeh, P. and Bohlin, E. (2020) ‘ICT as a driver of economic growth: A survey
of the literature and directions for future research’, Telecommunications Policy.
Elsevier Ltd, 44(2), p. 101922. doi: 10.1016/j.telpol.2020.101922.
Wang, X. and He, G. (2020) ‘Digital financial inclusion and farmers’ vulnerability to poverty:
Evidence from rural China’, Sustainability (Switzerland), 12(4). doi:
10.3390/su12041668.
Weidmann, N. B. (2015) ‘Communication networks and the transnational spread of ethnic
conflict,’ Journal of Peace Research. SAGE Publications Ltd, 52(3), pp. 285–296. doi:
10.1177/0022343314554670.
Workneh, M. A. (2020). Gender Inequality, Governance, and Poverty in Sub-Saharan Africa.
Poverty & Public Policy, 12(2), 150-174. doi:https://doi.org/10.1002/pop4.278
Yang, L. et al. (2021) ‘Mobile Internet Use and Multidimensional Poverty: Evidence from A
Household Survey in Rural China’, Social Indicators Research. Springer Science and
Business Media B.V. doi: 10.1007/S11205-021-02736-1.
Yang, L., Lu, H., Wang, S., & Li, M. (2021). Mobile Internet Use and Multidimensional
Poverty: Evidence from A Household Survey in Rural China. Social Indicators
Research, 158(3), 1065-1086. doi:10.1007/s11205-021-02736-1
Zhang, X. et al. (2020) ‘The Trickle-down Effect of Fintech Development: From the
Perspective of Urbanization’, China and World Economy, 28(1), pp. 23–40. doi:
10.1111/cwe.12310.
Zheng, X. and Lu, H. (2021) ‘Does ICT change household decision-making power of the left-
behind women? A Case from China’, Technological Forecasting and Social Change.
North-Holland, 166, p. 120604. doi: 10.1016/J.TECHFORE.2021.120604.
32
List of Figures
Figure 1: Trend of poverty among the different age group in SSA
Figure 2: Trends of ICT variables in SSA
0
5
10
15
20
25
30
35
40
45
50
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Proportion of the population below the
international poverty line
Years
youth poverty middle-aged poverty adult poverty
0
100
200
300
400
500
600
0
2
4
6
8
10
12
14
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Secondary axis
Primary axis
Years
Telephone ICT good exported ICT goods imported
Mobile phone Internet Broadband
33
List of Tables
Table 1. Summary of empirical literature on ICT and poverty reduction
Author (s)
Study Area/
Period
Model (s)
Findings
Mora-Rivera and García-
Mora (2021)
Mexico
2016
Propensity Score Matching
Internet access reduces poverty. The rural sector
has a higher significant impact than the urban
sector.
Yang et al. (2021)
Rural China
2019
endogenous switching regression
Mobile internet use reduces poverty
Mushtaq and Bruneau
(2019)
62 countries
OLS
FE
RE
2SLS
ICTs reduce poverty
Beuermann, McKelvey
and Vakis (2012)
Peru
[2004-2009]
FE
Mobile phones reduce poverty and extreme poverty
Ofori et al. (2021)
1980-2019
dynamic system GMM and the
panel corrected standard error
ICT access, usage and sills reduce severity and
intensity of poverty
Horn (2014)
ICT reduces poverty
Diga et al. (2013
2005-2012
Textual analysis
ICT reduces poverty
Lechman and Popowska
(2022)
FE
ICT reduces poverty
Abbreviations: OLS, ordinary least squares; POLS, pooled OLS, FE, fixed effect; RE, random effect; 2SLS; two-
stage least squares.
Table 2: Variable descriptive statistics.
Variables
Definition
Mean
Std. Dev.
Min
Max
lnpovb
Youth poverty
3.159
1.255
-2.996
4.547
lnpovm
Middle-aged poverty
3.311
1.177
-2.659
4.566
lnpov25p
Adult poverty
3.118
1.267
-3.219
4.540
lnrgdpc
Economic growth
7.204
0.994
5.338
9.812
lngini
Income inequality
3.780
0.157
3.497
4.109
lnelec
Access to electricity
3.535
0.688
1.411
4.605
lntra
Trade openness
4.172
0.434
2.781
5.012
lndcp
Access to credit
2.868
0.678
1.315
4.945
lnurb
Urbanisation
15.119
1.359
12.170
18.448
lngfcf
Physical capital
3.074
0.392
1.738
4.395
lntele
Telephone penetration
-0.088
1.439
-4.022
3.587
lncellu
Mobile phone penetration
4.171
0.577
1.683
5.110
lninter
Internet penetration
1.335
2.110
-3.375
9.572
lnictsex
ICT good exported
1.646
1.110
-3.258
4.118
lnbroad
Broadband penetration
3.361
0.493
2.199
5.100
lnict
ICT goods imported
1.110
0.573
-0.614
2.383
34
Table 3A: Effect of ICT on total youth and middle-aged poverty rate (Two-step dynamic system-GMM results)
Variables
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Youth poverty
Middle-aged poverty
L.lnpov
0.990***
1.003***
0.969***
0.979***
0.978***
0.937***
0.997***
0.989***
1.006***
0.974***
0.976***
0.981***
0.935***
1.000***
(0.024)
(0.024)
(0.016)
(0.016)
(0.035)
(0.024)
(0.019)
(0.024)
(0.025)
(0.016)
(0.016)
(0.033)
(0.025)
(0.018)
lnrgdpc
-0.017
0.164***
0.004
-0.003
0.002
-0.163***
-0.065***
0.000
0.200***
0.004
0.007
0.009
-0.166***
-0.061**
(0.033)
(0.045)
(0.025)
(0.015)
(0.046)
(0.020)
(0.024)
(0.033)
(0.047)
(0.022)
(0.018)
(0.045)
(0.024)
(0.025)
lngini
0.126***
-0.128***
0.196***
0.121***
0.319***
0.446***
0.158***
0.101***
-0.196***
0.177***
0.098***
0.304***
0.473***
0.154***
(0.026)
(0.044)
(0.014)
(0.026)
(0.024)
(0.061)
(0.042)
(0.027)
(0.047)
(0.014)
(0.029)
(0.024)
(0.071)
(0.039)
lnelec
-0.014
-0.021
-0.034
-0.049***
-0.041***
0.047***
-0.009
-0.023
-0.015
-0.027
-0.054***
-0.042***
0.048**
-0.008
(0.034)
(0.027)
(0.023)
(0.017)
(0.015)
(0.018)
(0.030)
(0.034)
(0.028)
(0.022)
(0.018)
(0.015)
(0.020)
(0.031)
lntra
-0.020
-0.111***
-0.049***
-0.027***
-0.035**
-0.146***
-0.074***
-0.028
-0.139***
-0.047***
-0.026**
-0.033*
-0.157***
-0.069***
(0.019)
(0.027)
(0.017)
(0.009)
(0.017)
(0.030)
(0.016)
(0.019)
(0.030)
(0.017)
(0.010)
(0.018)
(0.038)
(0.017)
lndcp
0.032
0.035
0.022
0.003
0.015
0.093***
-0.077
0.026
0.039
0.028
-0.004
0.018
0.094***
-0.067
(0.031)
(0.027)
(0.020)
(0.018)
(0.013)
(0.016)
(0.048)
(0.031)
(0.028)
(0.017)
(0.018)
(0.013)
(0.017)
(0.050)
lnurb
0.012*
-0.010
0.017***
0.016***
0.021***
0.031***
0.024***
0.013*
-0.017**
0.017***
0.017***
0.020***
0.031***
0.024***
(0.007)
(0.008)
(0.004)
(0.006)
(0.007)
(0.008)
(0.008)
(0.007)
(0.008)
(0.004)
(0.006)
(0.007)
(0.009)
(0.008)
lngfcf
-0.021
-0.061***
-0.019
-0.011
-0.012
0.027
-0.013
-0.025
-0.065***
-0.019
-0.016
-0.015
0.027
-0.013
(0.025)
(0.021)
(0.015)
(0.013)
(0.016)
(0.029)
(0.016)
(0.026)
(0.023)
(0.016)
(0.014)
(0.016)
(0.030)
(0.017)
lntele
-0.083***
-0.102***
(0.027)
(0.029)
lncellu
-0.061***
-0.068***
(0.017)
(0.018)
lninter
0.010**
0.012***
(0.004)
(0.004)
lnictsex
0.025***
0.025***
(0.008)
(0.008)
lnict
-0.084***
-0.089***
(0.019)
(0.021)
lnbroad
0.281***
0.259***
(0.065)
(0.071)
Constant
-0.447***
0.069
-0.411***
-0.352***
-1.285***
-0.693*
-0.888***
-0.385***
0.263
-0.367***
-0.286***
-1.264***
-0.733*
-0.877***
(0.140)
(0.205)
(0.130)
(0.098)
(0.457)
(0.363)
(0.304)
(0.138)
(0.224)
(0.137)
(0.085)
(0.438)
(0.394)
(0.300)
Observations
92
92
92
91
82
84
92
92
92
92
91
82
84
92
Hansen
14.044
15.138
18.921
18.270
17.721
12.875
11.975
14.866
14.899
19.345
18.099
17.755
13.006
11.180
P(Hansen)
0.231
0.515
0.273
0.308
0.340
0.682
0.746
0.189
0.532
0.251
0.318
0.338
0.672
0.798
Instruments for levels
Hansen excluding group
0.186
0.619
0.313
0.358
0.618
0.519
0.650
0.143
0.730
0.286
0.383
0.586
0.513
0.673
Diff (null H= exogenous)
0.465
0.237
0.261
0.256
0.078
0.856
0.696
0.497
0.148
0.262
0.238
0.091
0.842
0.818
IV(lists of instruments)
Hansen excluding group
0.567
0.488
0.175
0.801
0.839
0.536
0.735
0.604
0.675
0.177
0.823
0.824
0.560
0.684
Diff (null H= exogenous)
0.135
0.464
0.518
0.075
0.078
0.673
0.548
0.096
0.310
0.466
0.073
0.081
0.629
0.705
AR(1)
0.289
0.272
0.300
0.296
0.270
0.247
0.310
0.291
0.268
0.302
0.296
0.270
0.246
0.310
AR(2)
0.210
0.694
0.431
0.443
0.396
0.812
0.661
0.284
0.675
0.369
0.491
0.402
0.820
0.628
No. of countries
44
44
44
44
44
44
44
44
44
44
44
44
44
44
No. of instruments
20
26
26
26
26
26
26
20
26
26
26
26
26
26
35
CD- test (p-value)
0.269
0.164
0.776
0.215
0.220
0.959
0.423
0.213
0.168
0.877
0.241
0.199
0.862
0.372
Wald statistics
984246.889
600222.738
4071211.048
2926017.102
1877221.445
15744928.946
8352620.866
851400.977
575049.801
3910323.343
2334073.444
1599756.716
4187838.434
25882103.647
Standard errors in parentheses. Diff: Difference. Hansen-test refers to the over-identification test for the restrictions in system-GMM estimation. The AR (1) and AR (2) tests
are the Arellano–Bond tests for first and second-order autocorrelation in first differences. Note: There is no instrument proliferation since the number of instruments is less than
the number of countries. The p-value for CD is the probability value for cross-sectional dependency test. The CD test suggests that residual from all the models are cross-
sectional independent. The Wald statistics indicates that coefficients of all the predictors are statistically different from zero. * p < 0.10, ** p < 0.05, *** p < 0.01.
36
Table 3B: Effect of ICT on total adult poverty rate (Two-step dynamic system-GMM results)
Variables
1
2
3
4
5
6
7
Adult poverty
L.lnpov
0.988***
1.000***
0.969***
0.975***
0.981***
0.939***
1.001***
(0.024)
(0.024)
(0.017)
(0.016)
(0.034)
(0.025)
(0.018)
lnrgdpc
-0.007
0.179***
-0.003
-0.004
0.010
-0.160***
-0.059**
(0.032)
(0.045)
(0.024)
(0.016)
(0.046)
(0.022)
(0.025)
lngini
0.119***
-0.161***
0.193***
0.113***
0.316***
0.449***
0.142***
(0.028)
(0.045)
(0.016)
(0.028)
(0.023)
(0.063)
(0.046)
lnelec
-0.019
-0.019
-0.028
-0.050***
-0.043***
0.045**
-0.009
(0.034)
(0.028)
(0.023)
(0.019)
(0.015)
(0.019)
(0.032)
lntra
-0.025
-0.125***
-0.048***
-0.026**
-0.034**
-0.151***
-0.070***
(0.019)
(0.031)
(0.017)
(0.011)
(0.017)
(0.034)
(0.018)
lndcp
0.028
0.037
0.027
0.000
0.017
0.094***
-0.070
(0.030)
(0.026)
(0.019)
(0.018)
(0.013)
(0.017)
(0.052)
lnurb
0.013*
-0.014
0.018***
0.017***
0.020***
0.032***
0.025***
(0.007)
(0.009)
(0.005)
(0.006)
(0.007)
(0.009)
(0.008)
lngfcf
-0.025
-0.056**
-0.020
-0.012
-0.011
0.028
-0.007
(0.026)
(0.022)
(0.015)
(0.014)
(0.016)
(0.030)
(0.017)
lntele
-0.092***
(0.028)
lncellu
-0.065***
(0.016)
lninter
0.012***
(0.004)
lnictsex
0.025***
(0.008)
lnict
-0.085***
(0.020)
lnbroad
0.268***
(0.072)
Constant
-0.433***
0.181
-0.391***
-0.311***
-1.324***
-0.724**
-0.895***
(0.139)
(0.215)
(0.135)
(0.095)
(0.426)
(0.362)
(0.301)
Observations
92
92
92
91
82
84
92
Hansen
13.881
14.425
18.694
17.607
17.272
13.310
11.347
P(Hansen)
0.240
0.567
0.285
0.347
0.368
0.650
0.788
Instruments for levels
Hansen excluding group
0.177
0.706
0.308
0.385
0.588
0.499
0.629
Diff (null H= exogenous)
0.551
0.206
0.296
0.289
0.112
0.812
0.903
IV (lists of instruments)
Hansen excluding group
0.622
0.690
0.174
0.836
0.848
0.512
0.733
Diff (null H= exogenous)
0.128
0.339
0.547
0.082
0.087
0.649
0.626
AR(1)
0.284
0.265
0.294
0.290
0.262
0.241
0.302
AR(2)
0.195
0.730
0.295
0.388
0.356
0.935
0.527
No. of countries
44
44
44
44
44
44
44
No. of instruments
20
26
26
26
26
26
26
CD- test (p-value)
0.253
0.160
0.800
0.238
0.228
0.984
0.376
Wald statistics
967617.168
857870.942
3979877.065
2321856.045
1609395.829
16985154.866
9527093.852
Standard errors in parentheses. Diff: Difference. Hansen-test refers to the over-identification test for the
restrictions in system-GMM estimation. The AR (1) and AR (2) tests are the Arellano–Bond tests for first and
second-order autocorrelation in first differences. Note: There is no instrument proliferation since the number of
instruments is less than the number of countries. The p-value for CD is the probability value for cross-sectional
dependency test. The CD test suggests that residual from all the models are cross-sectional independent. The Wald
statistics indicates that coefficients of all the predictors are statistically different from zero. * p < 0.10, ** p <
0.05, *** p < 0.01.
37
Table 4A: Effect of ICT on female youth and middle-aged poverty rate (Two-step dynamic system-GMM results)
Variables
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Youth Poverty
Middle-aged poverty
L.lnpov
0.985***
0.982***
0.958***
0.975***
0.931***
0.940***
1.000***
0.985***
0.986***
0.973***
0.973***
0.948***
0.939***
1.013***
(0.027)
(0.026)
(0.016)
(0.013)
(0.011)
(0.029)
(0.023)
(0.025)
(0.023)
(0.013)
(0.012)
(0.014)
(0.030)
(0.025)
lnrgdpc
-0.056
0.119**
-0.007
-0.042**
-0.061***
-0.167***
-0.081***
-0.063*
0.094**
-0.019
-0.050***
-0.043**
-0.165***
-0.093***
(0.043)
(0.051)
(0.029)
(0.021)
(0.015)
(0.027)
(0.026)
(0.038)
(0.047)
(0.023)
(0.019)
(0.022)
(0.026)
(0.025)
lngini
0.198***
-0.041
0.205***
0.162***
0.344***
0.492***
0.239***
0.217***
0.001
0.190***
0.175***
0.304***
0.514***
0.227***
(0.022)
(0.044)
(0.017)
(0.027)
(0.012)
(0.042)
(0.033)
(0.020)
(0.037)
(0.012)
(0.029)
(0.050)
(0.036)
(0.029)
lnelec
0.003
-0.041
-0.032
-0.026
-0.035**
0.044**
-0.009
0.006
-0.022
-0.013
-0.021*
-0.035***
0.039
0.006
(0.042)
(0.027)
(0.025)
(0.018)
(0.015)
(0.021)
(0.031)
(0.034)
(0.021)
(0.021)
(0.013)
(0.012)
(0.027)
(0.030)
lntra
-0.008
-0.073***
-0.041***
-0.013
-0.031**
-0.132***
-0.089***
-0.003
-0.063***
-0.026**
-0.005
-0.032**
-0.104***
-0.076***
(0.023)
(0.024)
(0.014)
(0.013)
(0.013)
(0.032)
(0.017)
(0.020)
(0.021)
(0.013)
(0.014)
(0.015)
(0.022)
(0.019)
lndcp
0.043
0.025
0.025
0.022
0.012
0.079***
-0.121***
0.040
0.024
0.036*
0.017
0.014
0.071***
-0.108**
(0.036)
(0.027)
(0.022)
(0.016)
(0.014)
(0.024)
(0.047)
(0.030)
(0.024)
(0.019)
(0.011)
(0.011)
(0.024)
(0.048)
lnurb
0.009
-0.001
0.018***
0.013***
0.030***
0.032***
0.027***
0.008
-0.001
0.013***
0.012***
0.022***
0.031***
0.024***
(0.006)
(0.008)
(0.005)
(0.004)
(0.004)
(0.007)
(0.008)
(0.005)
(0.008)
(0.005)
(0.003)
(0.005)
(0.005)
(0.007)
lngfcf
-0.007
-0.051***
-0.015
0.004
-0.026
0.029
-0.004
-0.004
-0.042***
-0.008
-0.000
-0.017
0.018
0.001
(0.024)
(0.020)
(0.014)
(0.011)
(0.017)
(0.018)
(0.017)
(0.021)
(0.015)
(0.013)
(0.009)
(0.015)
(0.031)
(0.017)
lntele
-0.063**
-0.059**
(0.027)
(0.026)
lncellu
-0.070***
-0.071***
(0.019)
(0.020)
lninter
0.008*
0.009**
(0.004)
(0.004)
lnictsex
0.028***
0.022***
(0.008)
(0.007)
lnict
-0.073***
-0.062**
(0.023)
(0.026)
lnbroad
0.373***
0.369***
(0.059)
(0.065)
Constant
-0.552***
-0.108
-0.364***
-0.419***
-0.896***
-0.882**
-1.266***
-0.598***
-0.233
-0.379***
-0.407***
-0.818***
-1.001***
-1.279***
(0.153)
(0.183)
(0.130)
(0.101)
(0.131)
(0.394)
(0.283)
(0.150)
(0.161)
(0.133)
(0.106)
(0.168)
(0.319)
(0.301)
Observations
92
92
92
91
82
84
92
92
92
92
91
82
84
92
Hansen
16.607
17.782
18.627
18.353
17.022
13.972
17.574
17.071
18.088
19.416
18.442
18.940
14.109
18.172
P(Hansen)
0.120
0.337
0.288
0.304
0.384
0.601
0.349
0.106
0.319
0.248
0.299
0.272
0.591
0.314
Instruments for levels
Hansen excluding group
0.195
0.349
0.347
0.337
0.605
0.525
0.444
0.208
0.361
0.361
0.351
0.612
0.528
0.515
Diff (null H= exogenous)
0.118
0.331
0.237
0.282
0.114
0.584
0.210
0.083
0.272
0.156
0.250
0.047
0.550
0.111
IV (lists of instruments)
Hansen excluding group
0.377
0.136
0.205
0.399
0.399
0.377
0.652
0.517
0.135
0.217
0.332
0.388
0.404
0.659
Diff (null H= exogenous)
0.089
0.763
0.484
0.258
0.369
0.746
0.151
0.054
0.730
0.381
0.314
0.226
0.692
0.123
AR(1)
0.286
0.277
0.296
0.292
0.273
0.258
0.321
0.292
0.281
0.300
0.297
0.274
0.266
0.325
AR(2)
0.268
0.774
0.545
0.371
0.491
0.970
0.788
0.307
0.749
0.485
0.487
0.549
0.820
0.942
No. of countries
44
44
44
44
44
44
44
44
44
44
44
44
44
44
No. of instruments
20
26
26
26
26
26
26
20
26
26
26
26
26
26
CD- test (p-value)
0.340
0.163
0.428
0.339
0.228
0.701
0.329
0.389
0.166
0.422
0.538
0.182
0.540
0.382
Wald statistics
943765.646
1259328.523
2690086.824
4108709.585
5177251.395
16799403.613
9238889.258
2233397.096
1498022.562
5562839.394
5886782.841
11243172.142
22544673.647
4018938.657
Standard errors in parentheses. Diff: Difference. Hansen-test refers to the over-identification test for the restrictions in system-GMM estimation. The AR (1) and AR (2) tests are the
Arellano–Bond tests for first and second-order autocorrelation in first differences. There is no instrument proliferation since the number of instruments is less than the number of
countries. The p-value for CD is the probability value for cross-sectional dependency test. The CD test suggests that residual from all the models are cross-sectional independent.
The Wald statistics indicates that coefficients of all the predictors are statistically different from zero. * p < 0.10, ** p < 0.05, *** p < 0.01.
38
Table 4B: Effect of ICT on female adult poverty rate (Two-step dynamic system-GMM results)
Variables
1
2
3
4
5
6
7
Adult poverty
L.lnpov
0.973***
0.977***
0.952***
0.967***
0.917***
0.956***
0.989***
(0.027)
(0.028)
(0.017)
(0.019)
(0.045)
(0.029)
(0.021)
lnrgdpc
-0.050
0.136**
-0.004
-0.036**
-0.068
-0.174***
-0.079***
(0.039)
(0.056)
(0.030)
(0.014)
(0.051)
(0.024)
(0.024)
lngini
0.172***
-0.073
0.201***
0.176***
0.339***
0.489***
0.172***
(0.029)
(0.051)
(0.019)
(0.026)
(0.039)
(0.054)
(0.048)
lnelec
-0.003
-0.030
-0.036
-0.038**
-0.040***
0.073**
-0.008
(0.040)
(0.030)
(0.027)
(0.019)
(0.015)
(0.030)
(0.032)
lntra
-0.007
-0.092***
-0.046***
-0.021***
-0.044*
-0.145***
-0.080***
(0.021)
(0.027)
(0.017)
(0.008)
(0.025)
(0.031)
(0.018)
lndcp
0.035
0.028
0.021
0.013
0.010
0.090***
-0.095*
(0.034)
(0.030)
(0.024)
(0.019)
(0.014)
(0.026)
(0.051)
lnurb
0.014*
-0.005
0.019***
0.016***
0.032***
0.032***
0.027***
(0.007)
(0.008)
(0.006)
(0.006)
(0.008)
(0.008)
(0.008)
lngfcf
-0.016
-0.058***
-0.019
-0.002
-0.020
0.061**
-0.010
(0.025)
(0.021)
(0.017)
(0.014)
(0.016)
(0.027)
(0.017)
lntele
-0.078**
(0.030)
lncellu
-0.068***
(0.020)
lninter
0.008*
(0.005)
lnictsex
0.027***
(0.008)
lnict
-0.091***
(0.028)
lnbroad
0.325***
(0.067)
Constant
-0.472***
0.023
-0.317**
-0.416***
-0.745
-1.026***
-0.936***
(0.153)
(0.205)
(0.140)
(0.116)
(0.640)
(0.336)
(0.319)
Observations
92
92
92
91
82
84
92
Hansen
13.987
16.153
18.365
18.404
18.583
12.104
13.565
P(Hansen)
0.234
0.442
0.303
0.301
0.291
0.737
0.631
Instruments for levels
Hansen excluding group
0.235
0.546
0.330
0.312
0.621
0.534
0.592
Diff (null H= exogenous)
0.307
0.224
0.293
0.324
0.052
0.980
0.505
IV (lists of instruments)
Hansen excluding group
0.569
0.404
0.165
0.721
0.864
0.496
0.782
Diff (null H= exogenous)
0.136
0.451
0.608
0.094
0.052
0.811
0.333
AR(1)
0.282
0.271
0.292
0.289
0.274
0.258
0.306
AR(2)
0.178
0.888
0.359
0.285
0.355
0.754
0.517
No. of countries
44
44
44
44
44
44
44
No. of instruments
20
26
26
26
26
26
26
CD- test (p-value)
0.266
0.146
0.472
0.190
0.202
0.612
0.441
Wald statistics
1304619.929
950409.610
3870598.660
4155307.319
2491103.316
43098902.150
1394634.330
Standard errors in parentheses. Diff: Difference. Hansen-test refers to the over-identification test for the restrictions
in system-GMM estimation. The AR (1) and AR (2) tests are the Arellano–Bond tests for first and second-order
autocorrelation in first differences. There is no instrument proliferation since the number of instruments is less than
the number of countries. The p-value for CD is the probability value for cross-sectional dependency test. The CD test
suggests that residual from all the models are cross-sectional independent. The Wald statistics indicates that
coefficients of all the predictors are statistically different from zero. * p < 0.10, ** p < 0.05, *** p < 0.01.
39
Table 5A: Effect of ICT on male youth and middle-aged poverty rate (Two-step dynamic system-GMM results)
Variables
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Youth poverty
Middle-aged poverty
L.lnpov
0.989***
1.006***
0.974***
0.976***
0.981***
0.935***
1.000***
0.975***
0.982***
0.971***
0.967***
0.944***
0.901***
0.992***
(0.024)
(0.025)
(0.016)
(0.016)
(0.033)
(0.025)
(0.018)
(0.023)
(0.026)
(0.013)
(0.016)
(0.009)
(0.025)
(0.020)
lnrgdpc
0.000
0.200***
0.004
0.007
0.009
-0.166***
-0.061**
-0.035
0.173***
-0.008
-0.017
-0.021
-0.190***
-0.077***
(0.033)
(0.047)
(0.022)
(0.018)
(0.045)
(0.024)
(0.025)
(0.035)
(0.055)
(0.021)
(0.014)
(0.015)
(0.019)
(0.022)
lngini
0.101***
-0.196***
0.177***
0.098***
0.304***
0.473***
0.154***
0.143***
-0.132**
0.167***
0.135***
0.267***
0.509***
0.170***
(0.027)
(0.047)
(0.014)
(0.029)
(0.024)
(0.071)
(0.039)
(0.024)
(0.052)
(0.013)
(0.025)
(0.022)
(0.055)
(0.036)
lnelec
-0.023
-0.015
-0.027
-0.054***
-0.042***
0.048**
-0.008
-0.010
-0.026
-0.019
-0.042***
-0.041***
0.042**
-0.003
(0.034)
(0.028)
(0.022)
(0.018)
(0.015)
(0.020)
(0.031)
(0.033)
(0.024)
(0.018)
(0.014)
(0.012)
(0.020)
(0.027)
lntra
-0.028
-0.139***
-0.047***
-0.026**
-0.033*
-0.157***
-0.069***
-0.007
-0.101***
-0.035**
-0.020***
-0.064***
-0.133***
-0.066***
(0.019)
(0.030)
(0.017)
(0.010)
(0.018)
(0.038)
(0.017)
(0.018)
(0.024)
(0.016)
(0.006)
(0.017)
(0.034)
(0.014)
lndcp
0.026
0.039
0.028
-0.004
0.018
0.094***
-0.067
0.028
0.024
0.029*
0.005
0.012
0.077***
-0.079*
(0.031)
(0.028)
(0.017)
(0.018)
(0.013)
(0.017)
(0.050)
(0.030)
(0.027)
(0.015)
(0.016)
(0.012)
(0.019)
(0.044)
lnurb
0.013*
-0.017**
0.017***
0.017***
0.020***
0.031***
0.024***
0.011**
-0.010
0.013***
0.014***
0.018***
0.034***
0.023***
(0.007)
(0.008)
(0.004)
(0.006)
(0.007)
(0.009)
(0.008)
(0.005)
(0.007)
(0.004)
(0.004)
(0.004)
(0.006)
(0.007)
lngfcf
-0.025
-0.065***
-0.019
-0.016
-0.015
0.027
-0.013
-0.014
-0.058***
-0.016
-0.014
-0.019
0.003
-0.016
(0.026)
(0.023)
(0.016)
(0.014)
(0.016)
(0.030)
(0.017)
(0.022)
(0.022)
(0.014)
(0.013)
(0.013)
(0.031)
(0.017)
lntele
-0.102***
-0.095***
(0.029)
(0.029)
lncellu
-0.068***
-0.063***
(0.018)
(0.018)
lninter
0.012***
0.008**
(0.004)
(0.004)
lnictsex
0.025***
0.021***
(0.008)
(0.007)
lnict
-0.089***
-0.069***
(0.021)
(0.025)
lnbroad
0.259***
0.282***
(0.071)
(0.060)
Constant
-0.385***
0.263
-0.367***
-0.286***
-1.264***
-0.733*
-0.877***
-0.385***
0.094
-0.285**
-0.269**
-0.602***
-0.582
-0.841***
(0.138)
(0.224)
(0.137)
(0.085)
(0.438)
(0.394)
(0.300)
(0.148)
(0.200)
(0.134)
(0.112)
(0.142)
(0.415)
(0.298)
Observations
92
92
92
91
82
84
92
92
92
92
91
82
84
92
Hansen
14.866
14.899
19.345
18.099
17.755
13.006
11.180
15.840
17.177
19.439
19.038
13.329
12.846
13.757
P(Hansen)
0.189
0.532
0.251
0.318
0.338
0.672
0.798
0.147
0.374
0.247
0.267
0.649
0.684
0.617
Instruments for levels
Hansen excluding group
0.143
0.730
0.286
0.383
0.586
0.513
0.673
0.165
0.490
0.308
0.265
0.635
0.510
0.698
Diff (null H= exogenous)
0.497
0.148
0.262
0.238
0.091
0.842
0.818
0.234
0.194
0.216
0.344
0.454
0.889
0.284
IV (lists of instruments)
Hansen excluding group
0.604
0.675
0.177
0.823
0.824
0.560
0.684
0.634
0.303
0.175
0.731
0.801
0.531
0.669
Diff (null H= exogenous)
0.096
0.310
0.466
0.073
0.081
0.629
0.705
0.066
0.476
0.461
0.073
0.337
0.683
0.422
AR(1)
0.291
0.268
0.302
0.296
0.270
0.246
0.310
0.321
0.299
0.328
0.325
0.300
0.285
0.341
AR(2)
0.284
0.675
0.369
0.491
0.402
0.820
0.628
0.151
0.463
0.311
0.395
0.391
0.547
0.658
No. of countries
44
44
44
44
44
44
44
44
44
44
44
44
44
44
No. of instruments
20
26
26
26
26
26
26
20
26
26
26
26
26
26
CD- test (p-value)
0.213
0.168
0.877
0.241
0.199
0.862
0.372
0.230
0.179
0.844
0.183
0.164
0.523
0.485
Wald statistics
851400.977
575049.801
3910323.343
2334073.444
1599756.716
4187838.434
25882103.647
2365406.014
2386787.314
8056539.986
6287575.321
7192459.656
20344989.857
20989938.645
Standard errors in parentheses. Diff: Difference. Hansen-test refers to the over-identification test for the restrictions in system-GMM estimation. The AR (1) and AR (2) tests are the
Arellano–Bond tests for first and second-order autocorrelation in first differences. There is no instrument proliferation since the number of instruments is less than the number of
countries. The p-value for CD is the probability value for cross-sectional dependency test. The CD test suggests that residual from all the models are cross-sectional independent.
The Wald statistics indicates that coefficients of all the predictors are statistically different from zero. * p < 0.10, ** p < 0.05, *** p < 0.01.
40
Table 5B: Effect of ICT on male adult poverty rate (Two-step dynamic system-GMM results)
Variables
1
2
3
4
5
6
7
Adult poverty
L.lnpov
0.990***
1.012***
0.971***
0.977***
0.980***
0.939***
0.996***
(0.024)
(0.027)
(0.016)
(0.016)
(0.034)
(0.024)
(0.019)
lnrgdpc
0.011
0.219***
0.022
0.022
0.019
-0.163***
-0.064**
(0.034)
(0.045)
(0.025)
(0.018)
(0.049)
(0.024)
(0.026)
lngini
0.092***
-0.222***
0.185***
0.084***
0.323***
0.448***
0.159***
(0.030)
(0.048)
(0.013)
(0.031)
(0.026)
(0.070)
(0.042)
lnelec
-0.029
-0.022
-0.041*
-0.065***
-0.052***
0.051***
-0.010
(0.036)
(0.032)
(0.023)
(0.020)
(0.017)
(0.019)
(0.032)
lntra
-0.038*
-0.151***
-0.058***
-0.032**
-0.040**
-0.157***
-0.082***
(0.021)
(0.029)
(0.019)
(0.012)
(0.020)
(0.032)
(0.017)
lndcp
0.026
0.038
0.020
-0.008
0.013
0.100***
-0.076
(0.032)
(0.029)
(0.018)
(0.019)
(0.014)
(0.017)
(0.052)
lnurb
0.011
-0.020**
0.018***
0.016**
0.021***
0.030***
0.024***
(0.008)
(0.008)
(0.004)
(0.007)
(0.008)
(0.010)
(0.009)
lngfcf
-0.022
-0.062***
-0.022
-0.015
-0.012
0.032
-0.011
(0.028)
(0.024)
(0.017)
(0.013)
(0.018)
(0.029)
(0.017)
lntele
-0.101***
(0.027)
lncellu
-0.073***
(0.018)
lninter
0.013***
(0.005)
lnictsex
0.025***
(0.008)
lnict
-0.088***
(0.020)
lnbroad
0.286***
(0.072)
Constant
-0.354**
0.317
-0.390***
-0.269***
-1.357***
-0.697*
-0.881***
(0.153)
(0.244)
(0.130)
(0.101)
(0.448)
(0.383)
(0.325)
Observations
92
92
92
91
82
84
92
Hansen
14.434
13.429
19.587
17.938
17.488
12.708
11.506
P(Hansen)
0.210
0.641
0.239
0.328
0.355
0.694
0.777
Instruments for levels
Hansen excluding group
0.149
0.719
0.269
0.398
0.607
0.516
0.700
Diff (null H= exogenous)
0.568
0.291
0.268
0.233
0.092
0.904
0.663
IV (lists of instruments)
Hansen excluding group
0.558
0.589
0.185
0.785
0.840
0.551
0.723
Diff (null H= exogenous)
0.121
0.544
0.423
0.088
0.084
0.675
0.619
AR(1)
0.282
0.261
0.299
0.289
0.264
0.239
0.304
AR(2)
0.399
0.531
0.669
0.697
0.566
0.732
0.845
No. of countries
44
44
44
44
44
44
44
No. of instruments
20
26
26
26
26
26
26
CD- test (p-value)
0.180
0.169
0.997
0.226
0.217
0.980
0.351
Wald statistics
687282.357
518164.006
4162460.961
2315346.750
1283322.408
5432775.124
26765651.197
Standard errors in parentheses. Diff: Difference. Hansen-test refers to the over-identification test for the restrictions
in system-GMM estimation. The AR (1) and AR (2) tests are the Arellano–Bond tests for first and second-order
autocorrelation in first differences. There is no instrument proliferation since the number of instruments is less than
the number of countries. The p-value for CD is the probability value for cross-sectional dependency test. The CD test
suggests that residual from all the models are cross-sectional independent. The Wald statistics indicates that
coefficients of all the predictors are statistically different from zero.* p < 0.10, ** p < 0.05, *** p < 0.01.
41
Table 6A: Interactive effect of ICT and economic growth on total poverty (Two-step dynamic system-GMM results)
Variables
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Youth poverty
Middle-aged poverty
Adult poverty
L.lnpov
1.079***
0.987***
0.980***
0.900***
0.955***
1.031***
1.029***
0.989***
0.983***
0.868***
0.944***
1.033***
1.022***
0.985***
0.977***
0.894***
0.946***
1.033***
(0.035)
(0.012)
(0.023)
(0.015)
(0.025)
(0.015)
(0.040)
(0.012)
(0.022)
(0.022)
(0.024)
(0.015)
(0.042)
(0.012)
(0.025)
(0.017)
(0.022)
(0.014)
lnrgdpc
0.504***
-0.377**
-0.060*
-0.224***
-0.244***
-0.366**
0.462***
-0.450**
-0.058
-0.259***
-0.298***
-0.353**
0.464***
-0.454*
-0.064*
-0.229***
-0.328***
-0.334**
(0.077)
(0.168)
(0.033)
(0.012)
(0.075)
(0.162)
(0.064)
(0.228)
(0.036)
(0.030)
(0.097)
(0.162)
(0.067)
(0.242)
(0.038)
(0.018)
(0.092)
(0.159)
lngini
-0.625***
0.175***
0.219***
0.541***
0.436***
0.106***
-0.583***
0.175***
0.212***
0.518***
0.466***
0.110***
-0.561***
0.174***
0.225***
0.549***
0.455***
0.094***
(0.040)
(0.021)
(0.065)
(0.043)
(0.062)
(0.030)
(0.060)
(0.020)
(0.069)
(0.064)
(0.074)
(0.028)
(0.062)
(0.024)
(0.076)
(0.046)
(0.067)
(0.032)
lndcp
-0.013
-0.006
-0.025
0.034**
0.089***
-0.049
-0.000
-0.011
-0.030*
0.042***
0.047
-0.047
-0.005
-0.012
-0.036
0.036**
0.032
-0.054
(0.037)
(0.022)
(0.019)
(0.014)
(0.026)
(0.044)
(0.038)
(0.021)
(0.016)
(0.009)
(0.032)
(0.045)
(0.040)
(0.021)
(0.022)
(0.015)
(0.031)
(0.045)
lntele
-0.260
-0.203
-0.207
(0.196)
(0.195)
(0.209)
lntele × lnrgdpc
0.014
0.005
0.005
(0.027)
(0.027)
(0.029)
lncellu
-0.596**
-0.706**
-0.706**
(0.252)
(0.335)
(0.352)
lncellu × lnrgdpc
0.085**
0.102**
0.103**
(0.035)
(0.047)
(0.050)
lninter
-0.143***
-0.153***
-0.160***
(0.043)
(0.045)
(0.046)
lninter × lnrgdpc
0.021***
0.022***
0.023***
(0.006)
(0.006)
(0.006)
lnictsex
-0.625***
-0.588***
-0.643***
(0.071)
(0.106)
(0.080)
lnictsex × lnrgdpc
0.106***
0.100***
0.108***
(0.011)
(0.017)
(0.013)
lnict
-0.348
-0.601
-0.749**
(0.271)
(0.385)
(0.365)
lnict × lnrgdpc
0.039
0.083
0.106*
(0.039)
(0.059)
(0.056)
lnbroad
-0.420
-0.407
-0.359
(0.319)
(0.318)
(0.311)
lnbroad × lnrgdpc
0.087**
0.084**
0.080**
(0.038)
(0.038)
(0.038)
Controls
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
Constant
0.919**
2.038*
-0.409***
-0.568***
-0.144
1.466
1.359***
2.553*
-0.404***
-0.176
-0.002
1.378
1.308***
2.597
-0.417***
-0.559***
0.159
1.255
(0.457)
(1.129)
(0.137)
(0.121)
(0.451)
(1.036)
(0.453)
(1.549)
(0.146)
(0.331)
(0.588)
(1.023)
(0.452)
(1.624)
(0.155)
(0.140)
(0.577)
(0.989)
Observations
92
92
91
82
84
92
92
92
91
82
84
92
92
92
91
82
84
92
Hansen
11.659
14.417
15.611
14.189
8.151
10.608
9.213
13.945
15.848
12.448
11.824
10.027
10.128
14.265
15.709
13.942
11.961
10.249
P(Hansen)
0.705
0.494
0.408
0.511
0.918
0.780
0.866
0.530
0.392
0.645
0.692
0.818
0.812
0.505
0.402
0.530
0.682
0.804
Instruments for levels
Hansen excluding group
0.649
0.481
0.270
0.590
0.657
0.638
0.801
0.470
0.260
0.606
0.661
0.658
0.759
0.444
0.255
0.581
0.648
0.619
Diff (null H= exogenous)
0.565
0.415
0.775
0.274
1.000
0.837
0.700
0.524
0.759
0.506
0.506
0.917
0.616
0.524
0.814
0.315
0.508
0.963
IV (lists of instruments)
Hansen excluding group
0.314
0.166
0.698
0.847
0.437
0.658
0.548
0.161
0.754
0.807
0.442
0.628
0.545
0.148
0.769
0.821
0.435
0.679
Diff (null H= exogenous)
0.941
0.908
0.185
0.184
1.000
0.696
0.940
0.950
0.147
0.340
0.790
0.796
0.865
0.949
0.147
0.215
0.782
0.716
AR(1)
0.306
0.307
0.308
0.261
0.255
0.311
0.317
0.316
0.310
0.265
0.275
0.314
0.313
0.308
0.304
0.254
0.269
0.307
AR(2)
0.335
0.188
0.438
0.431
0.715
0.148
0.308
0.236
0.411
0.568
0.294
0.205
0.250
0.183
0.372
0.428
0.213
0.152
No. of countries
44
44
44
44
44
44
44
44
44
44
44
44
44
44
44
44
44
44
No. of instruments
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
CD- test (p-value)
0.266
0.423
0.324
0.804
0.515
0.311
0.223
0.262
0.295
0.667
0.689
0.320
0.272
0.380
0.331
0.789
0.444
0.262
Wald statistics
271749.289
7803752.161
3382925.546
2361874.690
3900640.820
14480137.144
194043.608
2734187.280
2391091.861
1514209.833
8119201.211
70962878.897
307830.494
3022153.790
2554960.387
1608397.381
3276285.078
42524708.369
Standard errors in parentheses. Diff: Difference. Hansen-test refers to the over-identification test for the restrictions in system-GMM estimation. The AR (1) and
AR (2) tests are the Arellano–Bond tests for first and second-order autocorrelation in first differences. There is no instrument proliferation since the number of
instruments is less than the number of countries. The p-value for CD is the probability value for cross-sectional dependency test. The CD test suggests that residual
from all the models are cross-sectional independent. The Wald statistics indicates that coefficients of all the predictors are statistically different from zero. * p <
0.10, ** p < 0.05, *** p < 0.01.
42
Table 6B: Interactive effect of ICT and income inequality on total poverty (Two-step dynamic system-GMM results)
Variables
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Youth poverty
Middle-aged poverty
Adult poverty
L.lnpov
1.007***
0.981***
0.979***
0.931***
0.951***
0.980***
1.012***
0.985***
0.980***
0.933***
0.948***
0.983***
1.008***
0.980***
0.979***
0.929***
0.953***
0.980***
(0.029)
(0.016)
(0.016)
(0.022)
(0.041)
(0.029)
(0.029)
(0.015)
(0.018)
(0.025)
(0.042)
(0.029)
(0.030)
(0.017)
(0.018)
(0.023)
(0.040)
(0.029)
lnrgdpc
0.125**
-0.014
0.067*
-0.009
-0.165***
-0.058**
0.166***
-0.010
0.073**
0.003
-0.167***
-0.055*
0.143**
-0.015
0.069**
-0.003
-0.160***
-0.055*
(0.059)
(0.038)
(0.035)
(0.035)
(0.037)
(0.029)
(0.057)
(0.037)
(0.032)
(0.038)
(0.039)
(0.029)
(0.057)
(0.038)
(0.034)
(0.037)
(0.039)
(0.030)
lngini
-0.145
-1.630**
0.082
-0.130
0.688**
1.381*
-0.226**
-1.691**
0.069
-0.232*
0.705**
1.422*
-0.185*
-1.581**
0.093
-0.179
0.662*
1.476*
(0.108)
(0.683)
(0.087)
(0.117)
(0.348)
(0.805)
(0.113)
(0.673)
(0.089)
(0.122)
(0.354)
(0.833)
(0.109)
(0.692)
(0.087)
(0.127)
(0.346)
(0.835)
lndcp
0.040
0.013
-0.017
0.008
0.131***
-0.087*
0.043*
0.014
-0.013
0.013
0.136***
-0.079
0.040*
0.013
-0.007
0.012
0.133***
-0.081
(0.025)
(0.019)
(0.032)
(0.009)
(0.025)
(0.048)
(0.026)
(0.019)
(0.031)
(0.009)
(0.025)
(0.051)
(0.024)
(0.018)
(0.033)
(0.009)
(0.025)
(0.052)
lntele
-0.486*
-0.512*
-0.504*
(0.270)
(0.304)
(0.274)
Lntele × lngini
0.108
0.111
0.111
(0.070)
(0.080)
(0.072)
lncellu
-1.702***
-1.750***
-1.662***
(0.547)
(0.538)
(0.555)
lncellu × lngini
0.429***
0.442***
0.419***
(0.142)
(0.141)
(0.144)
lninter
0.140
0.181**
0.192**
(0.089)
(0.088)
(0.089)
Lninter × lngini
-0.034
-0.045*
-0.048*
(0.025)
(0.025)
(0.025)
lnictsex
-0.921***
-1.072***
-1.002***
(0.262)
(0.253)
(0.273)
Lnictsex × lngini
0.260***
0.302***
0.283***
(0.073)
(0.070)
(0.076)
lnict
0.705
0.718
0.677
(0.739)
(0.754)
(0.725)
lnict × lngini
-0.216
-0.222
-0.209
(0.202)
(0.206)
(0.199)
lnbroad
1.475**
1.516*
1.572**
(0.745)
(0.780)
(0.776)
lnbroad × lngini
-0.316
-0.331
-0.344*
(0.194)
(0.202)
(0.201)
Controls
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
Constant
0.180
6.505***
-0.321
0.367
-1.492
-5.325*
0.396
6.696***
-0.296
0.672*
-1.501
-5.495*
0.287
6.312**
-0.345
0.502
-1.409
-5.713*
(0.355)
(2.491)
(0.338)
(0.358)
(1.185)
(2.812)
(0.379)
(2.465)
(0.350)
(0.375)
(1.215)
(2.936)
(0.367)
(2.531)
(0.353)
(0.393)
(1.192)
(2.929)
Observations
92
92
91
82
84
92
92
92
91
82
84
92
92
92
91
82
84
92
Hansen
16.340
13.927
20.775
20.930
11.625
10.257
15.660
14.001
20.396
20.909
11.899
9.650
15.244
13.912
21.006
20.746
11.746
9.578
P(Hansen)
0.360
0.531
0.144
0.139
0.707
0.803
0.405
0.525
0.157
0.140
0.687
0.841
0.434
0.532
0.137
0.145
0.698
0.845
Instruments for levels
Hansen excluding group
0.655
0.483
0.171
0.561
0.405
0.757
0.803
0.486
0.187
0.519
0.334
0.762
0.778
0.473
0.174
0.536
0.370
0.760
Diff (null H= exogenous)
0.079
0.496
0.230
0.016
1.000
0.593
0.048
0.478
0.230
0.020
1.000
0.714
0.067
0.523
0.203
0.020
1.000
0.737
IV (lists of instruments)
Hansen excluding group
0.381
0.192
0.673
0.918
0.726
0.651
0.565
0.188
0.731
0.921
0.854
0.601
0.553
0.192
0.743
0.918
0.804
0.651
Diff (null H= exogenous)
0.352
0.907
0.036
0.013
0.501
0.746
0.258
0.907
0.034
0.013
0.345
0.862
0.299
0.909
0.026
0.014
0.409
0.824
AR(1)
0.280
0.300
0.292
0.247
0.236
0.322
0.278
0.303
0.291
0.238
0.236
0.324
0.275
0.293
0.284
0.233
0.230
0.318
AR(2)
0.976
0.089
0.620
0.448
0.456
0.926
0.923
0.093
0.597
0.506
0.438
0.983
0.972
0.074
0.554
0.437
0.504
0.961
No. of countries
44
44
44
44
44
44
44
44
44
44
44
44
44
44
44
44
44
44
No. of instruments
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
CD- test (p-value)
0.256
0.532
0.211
0.532
0.532
0.357
0.314
0.380
0.201
0.485
0.549
0.264
0.238
0.515
0.216
0.575
0.633
0.335
Wald statistics
950463.811
1322338.013
2.370e+08
819906.390
3347340.969
9096797.901
920281.961
1151180.940
5.879e+08
2071585.669
1602777.588
4687675.651
828940.188
1061367.774
1.269e+08
744116.563
4077393.361
7810042.824
Standard errors in parentheses. Diff: Difference. Hansen-test refers to the over-identification test for the restrictions in system-GMM estimation. The AR (1) and
AR (2) tests are the Arellano–Bond tests for first and second-order autocorrelation in first differences. There is no instrument proliferation since the number of
instruments is less than the number of countries. The p-value for CD is the probability value for cross-sectional dependency test. The CD test suggests that residual
from all the models are cross-sectional independent. The Wald statistics indicates that coefficients of all the predictors are statistically different from zero.* p <
0.10, ** p < 0.05, *** p < 0.01.
43
Table 6C: Interactive effect of ICT and access to credit on total poverty (Two-step dynamic system-GMM results)
Variables
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Youth poverty
Middle-aged poverty
Adult poverty
L.lnpov
1.050***
1.008***
0.969***
0.960***
0.886***
1.023***
1.019***
1.012***
0.971***
0.975***
0.889***
1.025***
1.013***
1.007***
0.967***
0.955***
0.891***
1.024***
(0.032)
(0.018)
(0.027)
(0.011)
(0.031)
(0.013)
(0.035)
(0.022)
(0.026)
(0.024)
(0.039)
(0.013)
(0.037)
(0.021)
(0.028)
(0.011)
(0.038)
(0.013)
lnrgdpc
0.363***
-0.015
-0.064**
-0.025**
-0.227***
-0.055*
0.363***
-0.011
-0.056
-0.008
-0.228***
-0.051*
0.368***
-0.005
-0.058
-0.023*
-0.225***
-0.044
(0.063)
(0.017)
(0.032)
(0.012)
(0.055)
(0.028)
(0.055)
(0.018)
(0.034)
(0.032)
(0.060)
(0.028)
(0.057)
(0.020)
(0.037)
(0.012)
(0.063)
(0.027)
lngini
-0.466***
0.018
0.217***
0.191***
0.601***
0.087***
-0.475***
0.034
0.206***
0.203***
0.622***
0.086***
-0.473***
0.032
0.210***
0.191***
0.618***
0.073**
(0.030)
(0.017)
(0.056)
(0.026)
(0.103)
(0.028)
(0.039)
(0.023)
(0.057)
(0.029)
(0.111)
(0.028)
(0.039)
(0.020)
(0.066)
(0.028)
(0.105)
(0.031)
lndcp
0.018
-0.599***
-0.025
0.086***
-0.141
-0.343***
0.025
-0.683***
-0.029*
0.093***
-0.168
-0.310***
0.021
-0.688***
-0.032
0.079***
-0.147
-0.292***
(0.030)
(0.148)
(0.020)
(0.019)
(0.116)
(0.094)
(0.030)
(0.170)
(0.017)
(0.022)
(0.126)
(0.095)
(0.031)
(0.177)
(0.021)
(0.021)
(0.123)
(0.097)
lntele
-0.215***
-0.216***
-0.215***
(0.076)
(0.077)
(0.080)
lntele × lndcp
0.027
0.023
0.023
(0.021)
(0.021)
(0.023)
lncellu
-0.426***
-0.463***
-0.455***
(0.097)
(0.105)
(0.110)
lncellu × lndcp
0.148***
0.166***
0.166***
(0.032)
(0.037)
(0.038)
lninter
-0.064***
-0.063***
-0.064***
(0.016)
(0.014)
(0.014)
lninter × lndcp
0.022***
0.023***
0.023***
(0.004)
(0.004)
(0.004)
lnictsex
0.127***
0.131***
0.112***
(0.020)
(0.028)
(0.026)
lnictsex × lndcp
-0.043***
-0.045***
-0.037***
(0.008)
(0.010)
(0.010)
lnict
-0.449**
-0.501**
-0.454**
(0.214)
(0.230)
(0.230)
lnict × lndcp
0.128*
0.147*
0.131*
(0.074)
(0.079)
(0.078)
lnbroad
-0.100
-0.082
-0.050
(0.097)
(0.091)
(0.088)
lnbroad × lndcp
0.090***
0.082***
0.075***
(0.026)
(0.025)
(0.025)
Controls
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
Constant
1.064***
2.033***
-0.207
-0.671***
-0.276
0.596
1.346***
2.167***
-0.190
-0.901***
-0.297
0.488
1.349***
2.203***
-0.185
-0.643***
-0.387
0.414
(0.394)
(0.465)
(0.184)
(0.135)
(0.425)
(0.384)
(0.373)
(0.529)
(0.196)
(0.323)
(0.517)
(0.343)
(0.373)
(0.549)
(0.205)
(0.145)
(0.522)
(0.329)
Observations
92
92
91
82
84
92
92
92
91
82
84
92
92
92
91
82
84
92
Hansen
11.176
16.546
15.188
12.444
13.235
9.881
9.505
16.868
15.462
13.867
13.525
9.276
10.262
15.635
15.313
12.361
13.776
9.677
P(Hansen)
0.740
0.347
0.438
0.645
0.584
0.827
0.850
0.327
0.419
0.536
0.562
0.863
0.803
0.407
0.429
0.652
0.543
0.840
Instruments for levels
Hansen excluding group
0.596
0.435
0.288
0.715
0.818
0.673
0.762
0.425
0.272
0.694
0.801
0.699
0.719
0.412
0.269
0.665
0.799
0.667
Diff (null H= exogenous)
0.813
0.220
0.806
0.309
0.129
0.912
0.751
0.203
0.804
0.190
0.126
0.972
0.694
0.360
0.851
0.404
0.114
0.968
IV (lists of instruments)
Hansen excluding group
0.284
0.174
0.656
0.762
0.773
0.646
0.572
0.164
0.736
0.734
0.830
0.645
0.597
0.163
0.754
0.767
0.777
0.679
Diff (null H= exogenous)
0.984
0.657
0.234
0.381
0.300
0.795
0.901
0.643
0.174
0.279
0.237
0.860
0.801
0.792
0.173
0.384
0.255
0.784
AR(1)
0.291
0.312
0.306
0.270
0.224
0.294
0.300
0.322
0.307
0.275
0.220
0.298
0.298
0.316
0.300
0.264
0.209
0.290
AR(2)
0.405
0.216
0.352
0.632
0.502
0.231
0.406
0.288
0.324
0.537
0.443
0.285
0.333
0.248
0.307
0.533
0.455
0.227
No. of countries
44
44
44
44
44
44
44
44
44
44
44
44
44
44
44
44
44
44
No. of instruments
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
CD- test (p-value)
0.218
0.309
0.333
0.119
0.290
0.183
0.233
0.234
0.280
0.076
0.182
0.147
0.270
0.255
0.321
0.125
0.273
0.173
Wald statistics
269948.405
17296680.442
12420579.214
4121545.929
2229273.377
4131623.048
197927.588
13764053.883
32493233.214
1992526.355
2210535.584
7673595.549
176552.521
7995816.263
47348141.056
2290579.420
963749.664
7857646.184
Standard errors in parentheses. Diff: Difference. Hansen-test refers to the over-identification test for the restrictions in system-GMM estimation. The AR (1) and
AR (2) tests are the Arellano–Bond tests for first and second-order autocorrelation in first differences. There is no instrument proliferation since the number of
instruments is less than the number of countries. The p-value for CD is the probability value for cross-sectional dependency test. The CD test suggests that residual
from all the models are cross-sectional independent. The Wald statistics indicates that coefficients of all the predictors are statistically different from zero. * p <
0.10, ** p < 0.05, *** p < 0.01.
44
Table 6D: Net effect of economic growth, income inequality and access to domestic credit on poverty at different values of digitization indicators
Youth Poverty
Middle-aged Poverty
Adult Poverty
ICT variables
Minimum
Mean
Maximum
Minimum
Mean
Maximum
Minimum
Mean
Maximum
Panel A: Estimated net effect of economic growth
Telephone penetration
0.449***
0.503***
0.553***
0.441***
0.462***
0.481***
0.442***
0.463***
0.483***
Mobile phone penetration
-0.233**
-0.021
0.059*
-0.278*
-0.024
0.072**
-0.281*
-0.025
0.071**
Internet penetration
-0.130***
-0.032
0.139***
-0.134***
-0.029
0.155***
-0.143***
-0.033
0.159***
Broadband penetration
-0.174**
-0.072*
0.080**
-0.168**
-0.070*
0.077*
-0.158**
-0.064*
0.075*
ICT exported goods
-0.568***
-0.051**
0.210***
-0.584***
-0.095**
0.152**
-0.583***
-0.051**
0.217***
ICT imported goods
-0.268***
-0.201***
-0.152***
-0.350***
-0.206***
-0.100**
-0.393***
-0.210***
-0.074
Panel B: Estimated net effect of income inequality
Telephone penetration
-0.579
-0.155
0.241
-0.671
-0.236**
0.170
-0.632
-0.195*
0.214
Mobile phone penetration
-0.907**
0.161*
0.5640***
-0.948**
0.151*
0.565***
-0.875*
0.167*
0.560***
Internet penetration
0.196
0.037
-0.242
0.220
0.010
-0.357**
0.255
0.029
-0.366**
Broadband penetration
0.686*
0.318**
-0.231
0.695*
0.310**
-0.265
0.718*
0.318**
-0.281
ICT exported goods
-0.977***
0.298***
0.940***
-1.215***
0.264***
1.010***
-1.099***
0.286***
0.985***
ICT imported goods
0.821*
0.448***
0.172
0.842*
0.459***
0.177
0.791*
0.430***
0.164
Panel C: Estimated net effect of access to domestic credit
Telephone penetration
-0.091
0.016
0.116
-0.070
0.023
0.109
-0.071
0.019
0.102
Mobile phone penetration
-0.349***
0.019
0.158***
-0.404***
0.009
0.164***
-0.409***
0.003
0.159
Internet penetration
-0.101***
0.005
0.190***
-0.106***
0.001
0.187***
-0.109***
-0.001
0.188***
Broadband penetration
-0.145***
-0.040
0.117**
-0.130***
-0.034
0.108**
-0.127**
-0.039
0.0091*
ICT exported goods
0.226***
0.016
-0.090***
0.238***
0.019
-0.091***
0.199***
0.018
-0.073***
ICT imported goods
-0.220
0.002
0.165**
-0.258
-0.004
0.183***
-0.228
-0.002
0.165**
* p < 0.10, ** p < 0.05, *** p < 0.01. Note: All the net effect values are generated using STATA command
45
Appendix
Appendix Table 1: Effect of ICT on total youth and middle-aged poverty rate (Two-step dynamic system-GMM results)
Variables
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Youth poverty
Middle-aged poverty
Adult poverty
lnrgdpc
-0.455
-0.405
-0.903**
0.126
-0.190
-0.683*
-0.318
-0.260
-0.730**
0.197
-0.075
-0.528
-0.449
-0.425
-0.926**
0.161
-0.179
-0.680*
(0.376)
(0.400)
(0.371)
(0.372)
(0.394)
(0.382)
(0.348)
(0.371)
(0.344)
(0.346)
(0.369)
(0.355)
(0.380)
(0.406)
(0.374)
(0.378)
(0.397)
(0.388)
lngini
-0.463
-0.465
-0.151
-0.685*
-0.562
-0.407
-0.502
-0.505
-0.213
-0.703*
-0.583
-0.450
-0.477
-0.475
-0.151
-0.702*
-0.581
-0.421
(0.450)
(0.450)
(0.434)
(0.386)
(0.445)
(0.443)
(0.417)
(0.417)
(0.403)
(0.359)
(0.416)
(0.411)
(0.455)
(0.456)
(0.438)
(0.392)
(0.448)
(0.450)
lnelec
-0.042
-0.038
-0.034
0.047
0.014
-0.081
-0.029
-0.025
-0.021
0.040
0.010
-0.066
-0.053
-0.050
-0.045
0.043
0.009
-0.090
(0.088)
(0.088)
(0.084)
(0.100)
(0.113)
(0.089)
(0.082)
(0.082)
(0.078)
(0.092)
(0.106)
(0.083)
(0.089)
(0.089)
(0.084)
(0.101)
(0.114)
(0.091)
lntra
-0.108
-0.098
-0.103
0.007
-0.082
-0.062
-0.095
-0.084
-0.089
0.006
-0.079
-0.051
-0.108
-0.103
-0.105
0.014
-0.081
-0.066
(0.108)
(0.112)
(0.102)
(0.105)
(0.114)
(0.110)
(0.101)
(0.104)
(0.095)
(0.097)
(0.107)
(0.102)
(0.110)
(0.113)
(0.103)
(0.106)
(0.115)
(0.112)
lndcp
0.141
0.151
0.120
-0.066
0.171
0.264*
0.136
0.146
0.116
-0.047
0.165
0.251*
0.144
0.153
0.123
-0.071
0.177
0.261*
(0.119)
(0.120)
(0.114)
(0.115)
(0.135)
(0.137)
(0.110)
(0.112)
(0.106)
(0.107)
(0.126)
(0.127)
(0.120)
(0.122)
(0.115)
(0.116)
(0.136)
(0.139)
lnurb
-1.008***
-0.897***
-1.339***
-0.982***
-1.122***
-0.765***
-0.973***
-0.867***
-1.285***
-0.940***
-1.068***
-0.749***
-1.025***
-0.914***
-1.355***
-1.004***
-1.135***
-0.777***
(0.273)
(0.305)
(0.269)
(0.243)
(0.269)
(0.286)
(0.253)
(0.283)
(0.250)
(0.226)
(0.252)
(0.265)
(0.276)
(0.309)
(0.271)
(0.247)
(0.271)
(0.291)
lngfcf
0.015
0.009
0.024
0.044
-0.018
0.012
0.008
0.001
0.016
0.038
-0.024
0.004
0.008
0.004
0.018
0.034
-0.027
0.006
(0.072)
(0.074)
(0.068)
(0.069)
(0.079)
(0.071)
(0.067)
(0.069)
(0.063)
(0.064)
(0.074)
(0.066)
(0.073)
(0.075)
(0.069)
(0.070)
(0.080)
(0.072)
lntele
-0.031
-0.024
-0.044
(0.070)
(0.065)
(0.070)
lncellu
-0.054
-0.055
-0.044
(0.105)
(0.098)
(0.107)
lninter
0.075***
0.070***
0.078***
(0.021)
(0.020)
(0.022)
lnictsex
0.019
0.019
0.018
(0.016)
(0.015)
(0.016)
lnict
0.090**
0.086**
0.092**
(0.042)
(0.039)
(0.042)
lnbroad
-0.294*
-0.276*
-0.280
(0.171)
(0.159)
(0.174)
Constant
23.540***
21.663***
30.533***
19.644***
23.280***
21.816***
22.272***
20.431***
28.807***
18.693***
21.916***
20.699***
23.816***
22.084***
30.947***
19.773***
23.449***
22.026***
(4.106)
(5.079)
(4.344)
(3.700)
(4.131)
(4.062)
(3.810)
(4.710)
(4.029)
(3.436)
(3.860)
(3.765)
(4.153)
(5.147)
(4.381)
(3.752)
(4.153)
(4.122)
Observations
120
121
117
106
108
121
120
121
117
106
108
121
120
121
117
106
108
121
R2
0.390
0.390
0.474
0.349
0.402
0.410
0.372
0.373
0.459
0.340
0.391
0.394
0.396
0.395
0.483
0.351
0.408
0.412
Standard errors in parentheses.* p < 0.10, ** p < 0.05, *** p < 0.01
