ICT Adoption and Environmental Sustainability: Thresholds From the Lower and Lower-Middle-Income Countries

This paper aims to study the relationship between ICT readiness, use, and intensity, and environmental sustainability factors in the lower and middle lower-income countries from 2012-2018. ICT readiness, use, and intensity are measured with the impact of ICT on access to basic services, phone penetration, and internet penetration, while CO2 emissions per capita, fossil fuel energy consumption, and methane emissions are used as indicators for air pollution. To achieve this goal, the system two steps GMM estimation was performed which thresholds are computed contingent on the validity of tested hypotheses. The results demonstrate that increasing ICT readiness, use, and intensity in lower and lower-middle-income countries enhance environmental sustainability by decreasing CO2 emissions and energy consumption. forward the role played by on in the lower-income the adopted the Development. these international can guide to the to socio-economic environmental risks resilience and the


Introduction
Sustainable development is an increasingly important concern in the world. The public authorities have been proposed many measures to encourage it to be taken into account by society and government. ICTs have become an essential component of modern life in developed and LMICs alike. Technological innovation reduces the resource of air pollution by promoting cleaner production methods which can enhance economic growth (Belloumi., 2009;Sarkodieb, 2019; Tarek, 2020; Solomon & Klyton, 2020). Government authorities suggest for rms to employ integrated production technologies to reduce the negative environmental effects of CO2 emissions. They demonstrate that replacing the traditional production system with cleaner technologies can diminish air pollution. Despite the international efforts, the implementation of integrated production technologies still hampered by di culties linked to problems of cost, coordination, and inertia of skills and the productive organization of companies (Balsalobre, 2015); (Khurshid, 2019). Air pollution, climate change, and globalization are three major concerns characterizing the current world economy (Atici, 2012;Omri et al., 2015). Controlling for air pollution represents a very important challenge to achieve the goal of a sustainable environment as presented by Sustainable Development Goals to create a green economy; (Akinyemi, 2018; Asongu & Odhiambo, 2019). Recently, further studies provide evidence that net outputs from green economic growth are not su cient to moderate the growing climate pollution. Thus, the consequences of FDI in ows on increasing carbon emissions become a worldwide subject of great importance (Hao & Liu, 2015;Demena & Afesorgbor, 2020).
The argument of this research on enhancing information technology for a sustainable environment in lower and middle lowerincome countries is inspired by important factors in the scholarly and policymaking debate.
Firstly, recently, there have been growing opportunities for inhabitants of low and middle low-income countries to use it. These technologies have been experiencing a growing diffusion, whose speed varies signi cantly in different countries and for diverse tools. Regarding the case of mobile phones, it has been unexpectedly fast and socially widespread. Jointly with the emerging of locally-developed ICT-based applications, this phenomenon has lately renewed the attention of researchers towards ICTs and their potential impact on environmental quality by reducing air pollution.
Secondly, the attendant literature on environment quality has fundamentally focused on: (i) the relationship between CO2 emissions and ICT penetration in general. Within this framework: (Usman, Ozturk, & Hassan, 2020) are concerned with how enhancing ICT penetration in sub-Saharan Africa (SSA) can be a key solution towards environmental sustainability by diminishing CO2 emissions; Omri and Tarek (2020) demonstrated that good governance and technological innovation match the foreign direct investment to moderate carbon emissions in twenty-three emerging economies for the period 1996-2014.
Thirdly, the purpose of this study is to complement this current literature by examining how developing ICT adoption and especially the access to the internet in the basic services can affect air pollution. Hence, this paper provides new evidence that sheds light on the role that technology adoption plays in mediating the in uence of growth, remittances, education quality on reducing air pollution. To increase the policy relevance of the research, we explore whether there exist technology adoption thresholds. This relationship may be contingent on ICT indicators, where the technological environment promotes the reduction of air pollution after ICT access in basic services; mobile phone and internet penetration exceed a certain threshold level.
As we have noted, the objective of this study is to explore the conditional effects of good governance and ICT adoption on mitigating CO2 and carbon dioxide emissions to afford opportunities for a green economy in developing countries. Overall, this research aims to answer the following question: What minimum thresholds are needed for ICT adoption in the low and lower midincome countries to establish environmental sustainability?
The theoretical underpinning supporting the key role of ICT in improving environmental quality is consistent with the environmental Kuznets curve that supports the hypothesis that there is an inverted U -shaped association between ICT adoption and CO2 emissions. These emissions are estimated to increase air pollution but at a developed stage of ICT readiness and use, CO2 emissions may decrease (Akbostanci and Turut-Asi Tunc, 2009; Diao et al., 2009; He and Richard, 2010; (Ahmed & Ozturk, 2018) (Usman, Ozturk, & Hassan, 2020) ). Accordingly, empirical ndings of the previous papers recommended the need to establish exactly thresholds at which macroeconomic factors can either positively or negatively in uence environmental sustainability. Therefore, while this research supports the EKC hypothesis, it also gives more clari cation to informs policymakers on thresholds of ICT adoption and use in the basic services.
Our paper is structured as follows. The second section outlines the theoretical framework and presents related literature. The objective of the third section is to explain the methodology. In the fourth section, we present the empirical results and discussions. They are aware of reducing poverty must be through the economic growth channel and by many others social protection policies including inclusive education, health for all people, social protection, job opportunities, and environmental protection (Tchamyou et al., 2019;Omri, 2020). Nevertheless, with the emergence of the environmental quality challenge as a sustainable goal, the focus of the recent papers has progressively switched to think about how technology adoption and use must consider the harmonization of the economy and environment.
According to the previous studies, ICT value-added suggests that ICT in uences productivity for the developed countries but not for developing countries. Recently, researchers provide strong evidence that the developing countries have continued investing in their ICT infrastructure which proposes that ICT development enhanced productivity. Also, authors relying on multi-country surveys found that ICT help in reducing emissions by building smart cities and digital production systems. Overall, the results of these analyses show that digital transformation help in creating an environmentally sustainable society.
Additionally, the literature related to the new theory of growth support that technological change is of critical signi cance to deal with the polluted environment (Weitzman, 1997). As we can see, the existing literature has been focused on this association by exhibiting that the cost of controlling greenhouse gas (GHG) emissions cannot be kept at a rational level without relying on a collection of technologies that go far beyond what is presently offered. In this context, it is argued that there is a range of reasons throughout we can appreciate the importance of technological innovation in dropping GHG, explicitly: exploiting more and more energy-e cient production technologies, changes in the fuel mix, and the installation of end-of-pipe technology which is considered as the most considerable of the three (Bruyn and Sander, 1997). Correspondingly, Jones (2002) argues that investments in R&D and technological transformation can be a solution to moderate CO2 emissions and ght against air pollution.

2-2 Economic growth and CO2 emissions: environmental Kuznets curve
The empirical literature has given con icting results on the association of ICT and growth. For example,(Amari., Baklouti, & Mouakhar, 2020) found that technology adoption is positively associated with economic growth(Usman, Ozturk, & Hassan, 2020) measured by GDP growth rate in international comparative evidence. The environmental Kuznets Curve (EKC) hypothesis proposes that the pollution level rstly is completely and positively associated with economic growth. Besides, emissions intensity is connected to economic growth. Additionally, technology development guided the economy to change the traditional manufacturing production system to the smarter one (Amri, 2018). Thus, this change to industry 4.0 will mitigate the CO2 emissions intensity and as a result, increase energy e ciency. It was considered the year 2019 as the second warmest year on record and the end of the warmest decade (2010-2019) ever recorded. Carbon dioxide (CO2) levels and other greenhouse gases in the atmosphere rose to new records in 2019. Hence, previous studies have con rmed that climate change is touching the majority of countries around the world. It is negatively linked to economic growth and the well-being of people. While it is expected that greenhouse gas emissions will drop about 6 percent in 2020 as a result of the COVID-19 pandemic due to travel bans and economic slowdowns, this result is only provisional. Climate change is not on pause. Once the global economy begins to recover from the pandemic, emissions are expected to return to higher levels.
Sustainable development and environmental protection require more penetration of a good communication tool that enables to diminish the informational knowledge lacks that are linked with environmental change. CO2 emissions contribute to environmental degradation or air pollution. Besides in this paper, the air pollution caused by CO2 emissions, fossil fuel energy consumption, and methane emissions will be employed as proxies for environmental degradation. This study aims to determine how increasing information and communication technology adoption can affect the emission such as CO2.
Previous researchers have found that monitoring environmental governance, environmental management needs a good communication tool. Also, the latest ICT literature is reliable with the ndings that compared to other more advanced economies in the world (Asia, Europe & North America) where ICT is attaining high levels, there is more room for ICT diffusion in SSA. To achieve sustainable development goals (SDGs), policymakers have to deal with sobering policy challenges for sustainable development, like environmental pollution and climate change. In their study (Asongu et al., 2018) provided evidence that the governance of energy use in SSA is a very urgent objective to attain economic growth because of the persistent energy crisis in the sub-region and the non-effective energy management. Besides, (Anyanwu & Erhijakpor, 2014) con rm that in most African countries, the management of energy is characterized by ine ciency. For example, in Nigeria (the most populous country in Africa), a high number of government resources are using fossil fuels, as a substitute for investing in renewable energy. The use of power generators has contributed to the restoration of energy supply faults and shortfalls, which comprise about three-quarters of the

Data collection
The data used in this paper consist of 57 countries in lower and lower-middle-income countries [1]  As a result, we obtained a balanced panel with 285 observations. To assess the relationship between these variables, a dynamic generalized method of moments (GMM) was used. The different sources of data and a short de nition of all the variables used in this paper are disclosed in Table 1 WGI Before the hypotheses were tested, we examine the descriptive statistics of the variables listed in Table 2. Summary statistics and the correlation matrix have been respectively presented in Tables 2-3. The rst argument for adopting a GMM estimation is the higher number of periods for each country in our sample. The crosssections are exceeding the period. The estimation was conducted for ve periods from 57 countries. We employ a balanced panel dataset of annual periodicity over the period 2012-2016. To give reassurance of the goodness of t information, the explanatory capacity of the models was veri ed. The model often has a dynamic effect. It is closely connected to its earlier value. Thus, it has been reported in the recently empirical studies that researchers should consider the dynamic effect when conducting times series estimation. This is usually established by adding a lagged dependent variable as an explanatory in the model. Hence, the correlation matrix shows that the environmental indicators are persistent. Thus, the above equations are deduced as follow: Where, CO I,t is "a CO2 emissions indicator of country i at period t"; σ 0 is a constant, IC represents ICT readiness, use, and intensity ( ICT access in basic services mobile phone penetration and internet penetration); ICIC is the quadratic interaction of ICT variables between identical ICT variables (e.g. "ICTBASIC × ICTBASIC"; "MOBILE PHONES × MOBILE PHONES" or "INTERNET × INTERNET"); W represents the vector of control variables namely: trade, GDP growth, population growth, education system quality, and regulation quality; τ represents the coe cient of auto-regression, which is one for the speci cation, because of limited degrees of freedom; ξ t is the time-speci c constant; η i is the country-speci c effects (or factors that are particular to each country in the sample), υ i , t is the two-way the disturbance term and ε i , t is the error term.
The GMM empirical estimation adopted within the framework of this analysis is based on Roodman (2009a) which is an extension of the difference GMM methodology of Arellano and Bover (1995). The adopted process of estimation is the two-step estimator which is more robust when compared to the one-step process that exclusively accounts for homoscedasticity, compared to the

Empirical Results And Discussion
The empirical ndings are disclosed in Tables 4-6. Table 4 presents results about linkages between 'ICT access in basic services', 'internet penetration', 'mobile phone', and air pollution measured by 'CEO 2 emissions', 'fossil fuel energy consumption', 'methane emissions', Table 5 focuses on nexuses between ICT access in basic services', 'internet penetration', 'mobile phone' and 'fossil fuel energy consumption', while According to the non-interactive regressions, results con rm that ICT readiness, use, and intensity (i.e. ICT access in basic services and mobile phones) signi cantly affects CO2 emissions, and only internet penetration doses did not affect the CO2 emission. Our ndings suggest that increasing ICT access in basic services is positively and signi cantly associated with CO2 emissions per capita. The corresponding marginal effect was negative. This negative marginal effect implies that at a determined threshold of ICT access in basic services, the net impact can be transformed from positive to negative. Therefore, thresholds must be within the range of corresponding data to present economic sense and policy signi cance. people. To sum up the ndings, the mobile penetration, mobile phone, and ICT access in basic services range beyond the threshold's values computed in Table 5, it is necessary to increase the impact of ICT readiness and use to moderate the fossil fuel energy consumption. In consequence, the intended thresholds values of ICT readiness and use are the minimum levels to achieve the objective of a sustainable environment through increasing fossil fuel energy consumption in the lower and lower-middleincome countries.
Policymakers should have been given more effort on this topic and focused more on developing the technology sector. They . Therefore, governments in the lower and lower-middle-income countries necessitate to take on a series of environmental policy to concentrate on environmental challenges. Generally, government laws have a strong connection with the overall regional ecological performance; that is, strict regulatory regulation might guide rms and businesses to transform their production system and advance regional ecological e ciency by implementing the smart system in their

Conclusion And Policy Implications
This paper aims to study how increasing ICT readiness, use, and intensity in the lower and lower-middle-income countries can contribute to achieving an SDG such as environmental sustainability by reducing CO2 and methane emissions. This study suggested empirical evidence based on the Generalized Method of Moments estimations and fty-seven countries over the period 2012-2016. ICT is measured with the impact of ICT on access to basic services, phone penetration, and internet penetration, while CO2 emissions per capita, fossil fuel energy consumption, and methane emissions are used as indicators for air pollution. To sum up the ndings, the non-interactive regressions con rm that increasing ICT readiness and use (i.e. mobile phones and the internet) signi cantly in uence CO2, fossil fuel energy consumption, and methane emissions. Taken together, we conclude that increasing ICT has a positive net effect on CO2 and methane emissions per capita while increasing the impact of ICT access in basic services has a net negative effect on CO2, fossil fuel energy consumption, and methane emissions. To explore this issue further, the negative net effects suggest that ICT needs to be further developed beyond the determined thresholds, to attain the required negative net effect on fossil fuel energy consumption. We have determined these required threshold values and found them to be respectively: 244.28 per 100 people for mobile phone penetration and 10.35 for the impact of ICT access in basic services. While the former is not within practical range only for internet penetration and thus has no practical implication for policy, the latter is within achievable range (56.80) and is of policy relevance. As we have noted that an internet penetration range of above 54.92 for every 100 people helps to create a sustainable environment through moderating CO2 and fossil fuel energy consumption. The fact that the threshold at which this is possible exclusively make economic sense for mobile phone and ICT in basic services is interesting because the internet can be used on many devices, such as the mobile phone. Even though, intensifying ICT access in basic services is unthinkingly more pertinent than the mobile phone, though both are also complementary.