System GMM-based Model for Monitoring Joint Impact of ICT-infrastructure, Financial Development, and Trade-openness on Economic-growth

This study has analyzed the joint impact of information communications technology (ICT) Infrastructure, �nancial development (FD), and trade openness (TO) on economic growth (EG). We have used the data from 85 countries (including 27 low-income and 58 high-income countries). In this sample, we have collected the data from the year 2000-2019. We have framed hypotheses for samples and applied OLS, �xed effect regression (FER), and GMM method. Our results provide evidence that (a) ICT infrastructure is a signi�cant and positive interpreter of economic growth, (b) Individual consideration of trade openness and �nancial development is insigni�cant and negative for both groups of countries, ICT infrastructure requires more �nancial development and trade openness in low-income countries (LICs) in comparison to high-income countries(HICs), (c) ICT infrastructure-nancial development-trade openness nexus differ signi�cantly for both groups of countries. Our study acquiesces that variables are grave drivers of any economy and comprehensive growth in low-income countries. In this work, we have also discussed policy implications.


Introduction
The economic growth of any country depends on the various determinants, nancial development (FD), foreign direct investment (FDI), nancial development, and trade openness (TO) are most important among these (Adom et al., 2019;Boamah, 2017;Dunne & Masiyandima, 2017).However, information and communication technology (ICT) is a modern driver of growth as compared to TO and FDI (Donou-Adonsou, 2019; Myovella et al., 2020).Furthermore, a wide range of studies is an emphasis on different channels of ICT; mainly here discussed about telecommunication infrastructure because telecommunication infrastructures contribute to EG like: reducing transaction time period and cost, increasing e ciency, growth in trade, enlarge innovation and development, increasing job opportunities, etc. (Pradhan, Arvin, Hall, & Bennett, 2018; Pradhan, Arvin, & Norman, 2015; Datta & Agarwal, 2004;Shahiduzzaman & Alam, 2014).According to United Nations, (2019), ICT Infrastructure directly affects job creation, GDP, increasing rate of return, pro tability by using skilled persons, and income in the IT sectors, especially in developing countries.Similarly, World Bank (2012) argues that ICT is a driving force in encouraging EG in developing countries.However, developing countries are gradually using more developed ICT Infrastructure.Therefore, it shows that the number of internet and mobile users is increasing signi cantly in developing countries (including low-income & middle-income countries).To improve ICT infrastructure, we are still requiring funds, and it is possible when investing in new technologies, including arti cial intelligence, robotics, and AR/VR [1].ICT is increasingly associated with more positive economic growth.Nowadays, ICT plays a vital role in all the sectors with extensive ICT applications.With the help of new ICT technologies, cost-reducing, innovative behaviour, economic reconstruction, and increased performance in all sectors make possible (Sharafat and Lehr, 2017).ICT connects the customers, suppliers, and manufacturers through the ease of doing the business.Moreover, it helps in developing new knowledge and spread through the more e cient process of easy transformation of all types of information without any geographical and demographical barriers.
Even though the ICT plays a signi cant role in economic growth, it requires more nancial support for further investment in new ICT projects and adaptation of more trade openness policies of particular countries.For the making of good ICT infrastructure required high speed of internet, costly installation & operation charges, consume more electricity, etc.The rapid growth of ICT requires more nancial development and trade openness.Previous studies regarding the nexus between ICT and EG are discussed exclusively without looking at the impact of nancial development and trade openness.Therefore, it is necessary to continue research and analysis on this issue, including these two most signi cant variables which affect EG.
The potential contributions of our study are in four aspects: First, this study empirically tested the joint impact of ICT infrastructure, nancial development, and trade openness on EG.Second, our study analyzes the different mechanisms of ICT infrastructure's in uence on the economic growth of LICs and HICs.Third, this study research has enhanced the research literature between ICT infrastructure, nancial development, and trade openness moreover provided the possibility of reference for subsequent analysis in this eld.Fourth, the variables nancial development and trade openness removes previous de ciencies to examine the collaborative relationship among ICT infrastructure, nancial development, trade openness and EG.
This study is structured in the following way: a related study discussed in section 2. How collects and chooses the data for this study is explained in section 3. Econometric results and discussion are described in section 4. The nal section 5 offers a conclusion and policy implications based on the ndings.
[1] AR-Augmented reality and VR-Virtual reality  Chavula, 2013) and they have found a positive association between ICT and EG.Most of the studies considered the developed countries for the empirical analysis of ICT infrastructure.Only a few studies (Youse , 2011) have been done on underdeveloped countries because of the lack of data availability of ICT variables.
Appiah-Otoo and Song, (2021) examined the countries that have high ICT revolution among the high, middle, and low-income countries.
They have used panel data of forty-eight high, fty-eight middle, and twenty low-income countries for this examination.Therefore, this analysis included 123 countries, and the time frame was from 2002 to 2017.They have taken the independent variable as ICT index measured by internet, mobile, and xed broadband users, and the dependent variable was EG.Finally, they concluded that the ICT index encouraged EG in these countries.
Similarly, Ghosh (2016) used Indian states and, using data from 2001-2012, analyzed the impact of a mobile telephone on EG.The results of panel data analysis revealed a positive association between the mobile telephone and EG.Moreover, Habibi and Zabardast (2020) examined the effects of education and ICT on EG in the Middle East countries (17) and the Organization for Economic Cooperation and Development (OECD) economies (24).Further, they have also compared the impact of the variables between these selected two groups of countries.To compare the effects of ICT and education on EG, they have used two kinds of countries; most and least developed.They took 18 years of data from 2000 to 2017 and used it as panel data.They have applied OLS, xed-effect, and GMM methods.Results of panel analysis show that ICT is positively affecting the EG in both countries (the Middle East countries and OECD).
Parwantoa and Wulansari (2020) analyzed the impact of ICT on EG in Asian countries, including lower, middle, and high-income countries.They used the period 2010-2018 and applied panel data regression.The panel analysis results found that middle-income countries (MICs) used more ICT facilities than LICs and HICs countries.Similarly, Lam and Shiu (2010) used lower-middle-income countries (LMICs) (59 countries) and used panel data analysis for the causal relationship between ICT diffusion and EG (real GDP per capita).This analysis used the data from1980 to 2006, and the results show a unidirectional causality between ICT diffusion and real GDP per capita.Pohjola (2000) explored the impact of ICT on EG in the 39 countries and data used from 1980-1990 by applying an explicit model of EG.Here, it can be observed that the countries that have invested in ICT sectors/ICT infrastructure/technology encourage the EG.But they have found those countries that have not invested; they found a negative and insigni cant relationship between ICT and EG.Youse (2011) analyzed the effect of ICT capital on EG in developed and developing countries for 2000-2006.It is indicated that ICT capital played a signi cant role in the EG of upper, middle, and low-income countries.
On the other hand, studies related to nancial development and EG are well explained in the literature.De Gregorio and Guidotti (1995) examined the relationship between nancial development and economic growth.They found that nancial development has a positive relation with EG in the cross-country sample, but in the panel data (Latin American countries) found a negative impact of nancial development on EG.Arestis and Demetriades (1997) found empirical proof related to the relationship between nancial development and EG in two ways: how and to what extent the FD can contribute to upgrading EG.So, rstly, nd out the causal relationship between FD and EG.Secondly, the nancial liberal policy can stimulate investment and EG.However, Shaw (1973) Kaushal and Pathak, 2015).The study of Fetahi-Vehapi, Sadiku, and Petkovski (2015) analyzed the impact of trade openness on EG in South-Eastern European (SEE) nations.They have included time spam from 1996-to 2012 and applied the GMM Model.Outcomes indicate that they have not found any robust association between trade openness and EG.On the other hand, Chatterji, Mohan, and Dastidar (2014) proved that trade growth encouraged EG in India from 1970 to 2010 and applied the Vector Autoregression (VAR) method.Moreover, they have also concluded that there is no evidence of a signi cant relationship between trade barriers and rates.Kumari et al. (2021) examined the long-term and causal relationship among foreign direct investment (FDI) in ows, trade openness, and EG from India.They have applied the Johansen co-integration and vector autoregression (VAR) Model.They have found no longterm co-integration but present the causality between FDI in ow and EG in India for the period of 1991-2013.Further, the results of the VAR Model are bi-directional causality between FDI and EG.Still, in contrast, they found no bi-directional relationship between trade openness and EG in India.Bangake and Eggoh (2011) used panel data for the empirical analysis, and these panel countries were classi ed into low, middle & high-income countries.They concluded bidirectional long-run causality between nancial development and EG in LICs, MICs, and HICs but no found short-run causality in LICs & MICs.Only in HICs found a signi cant short-run causal relationship between nancial development and EG.Hence, the impact of ICT, nancial development, trade openness on EG in the existing studies is mixed (positive, negative, and insigni cant), a summarized information is shown in Table 1.ICT infrastructure variables data are collected from Worldwide Telecommunications Union (WTU 2019).
We have applied the generalize method of moments (system GMM approach) to estimate the framed hypotheses.This approach has various bene ts: rstly, not nd biased behaviour.Secondly, this method is capable of controlling the country or sector-speci c effects.
Thirdly, and most importantly, controls the endogeneity problem which is not possible in OLS and Fixed Effect Regression (FER) method.Therefore, our study applied the Generalized method of moments (GMM) approach.This technique is widely used when incorporating panel data includes a couple of periods with an enormous number of observations (Roodman, 2006).As suggested by Hoe er, (2002) and Das and Paul, (2011), the GMM approach is utilized in both situations where variables are lagged and differenced versions of the regressors as instruments in obtaining coe cient estimates.In this method, Arellano-bond shows that there is no serial correlation along with the hypothesis and error terms.This is supported by the selection of instrumental variables in the GMM model.In these studies, it is suggested that why the GMM approach is preferred over OLS and FER.

Hypothesis framed
Based on the above literature review, we have framed the hypotheses to analyze the impact of ICT infrastructure, nancial development, and trade openness on EG in LICs and HICs.The formulated hypotheses are as follows.

Hypothesis 1
Good ICT infrastructure has a signi cantly positive impact on EG in LICs and HICs.Salahuddin and Gow (2015) applied the unit root test, Johansen, and ARDL co-integration tests to analyze the effects of internet usage, nancial development, and TO on economic growth.This study was based on the time-series data for South Africa from 1991 to 2013.They have strongly supported a positive relationship between internet usage and EG in South Africa.Similarly, they found the same results regarding nancial development and economic growth.But, in this study, TO is not disused to encourage economic growth.As we expect more TO economy attracts nancial development and improve ICT infrastructure, these three variables promote EG in LICs and HICs (Salahuddin and Gow, 2015).Hypothesis 3 (c) Joint effects of ICT infrastructure, FD, and TO encourage EG to vary between two groups of countries (LICs and HICs).
Joint effects of ICT infrastructure and FD on EG and ICT infrastructure and TO on EG encourage EG.It may also vary from country to country because past studies' results were mixed (positive, negative, and insigni cant) (Ghosh, 2016; Habibi and Zabardast, 2020; Appiah-Otoo and song, 2021; Salahuddin and Gow, 2015; Shahbaz and lean, 2012).Thus, we also expect the joint effects of ICT infrastructure, FD, and TO encourage EG, and it may vary between two groups of countries (LICs and HICs).The framed hypothesis is represented in Fig. 1 to show the impact of ICT infrastructure, nancial development, and trade openness on EG.

Econometric Model
In this study, we have applied a standard estimation for the EG Model where the EG is determined by (  .FD is represents by the total credit provided to the private sectors (% of GDP).TO is represents by total sum of imports and exports divided by GDP per capita, capfor is represented by capital formation (the gross domestic investment as a share of GDP), consum is represented by general government nal consumption expenditure (%1 of GDP), popugr is represented by total population growth rate.i implies no of observations (total countries considered) and t is period.ϵ it is represented by error term.
Total ten variables included for the empirical analysis, one variable taken as a dependent, and the other nine variables are considered as explanatory variables, including three control variables.We have applied the system GMM approach on normalized values of the variables.After converting all variables into the normalized form, we have applied the system GMM method.All the framed hypotheses 1, 2 (a, b), and 3 (a, b, c) shown in Eq. ( 2) are well explained in sub-section 3.2.We have described the variables, measured proxies, and data sources in Table 2.

Results And Discussion
Tables 5 and 6 show the correlation results of LICs and HICs.There is a positive association between the secinternet and telephone in LICs with a value of 0.375, which is quite good.It speci es that people who have a telephone subscription are more likely to use the secure internet in LICs.It also shows that these countries are still facing a lack of ICT infrastructure.
Similarly, from Table 6, we can see that the correlation between the secinternet and broadband is 0.146 in HICs which is relatively high compared to LICs.So, it shows that internet and broadband services are good in HICs.Most people who do not even use broadband and phone subscription also rely on broadband.Therefore, we can say that developed countries are more updated and well-structured in ICT infrastructure than LICs.7 shows the results of OLS, FER, and system GMM.Initially, we have applied the OLS method and FE regression method, but we found that the results were biased in terms of endogenous.Therefore further, we have applied the GMM method to overcome the endogeneity problem.So, here we discussed only the results of GMM and showed 12 different models related to the framed hypothesis.In the GMM Model, we found TO, secinternet and FD negatively related to EG in LICs so these ndings did not support hypotheses 2 (a) and (b).For the HICs, we found almost all results were the same but ICT infrastructure proxy sign (+/-) changes according to the Model.In the case of HICs, TO, broadband, and FD have a negative relation with EG which is also not supported by hypotheses 2 (a) and (b).But, the majority of the ICT infrastructure variables have a positive relation with EG along with FD and TO, therefore we can conclude that ICT infrastructure has a positive relation with EG in LICs and HICs.As we thought interaction of FD and ICT infrastructure is always positive and signi cant.FD motivates economic growth through the ICT infrastructure in LICs.When FD is considered, the marginal effects of the internet, broadband, and secured internet are more potent than that of the telephone in LICs.Similarly, this nding was found in the study of Sassi and Goaied (2013), but they support more telephone use than the internet and mobile.
Table 9 shows the joint impact of ICT infrastructure, FD, and TO on EG in LICs.Jointly impact of FD*internet*TO (.638), FD* secinternet* TO (0.103), FD* broadband* TO (0.096) and FD* telephone* TO (0.031) is positive relation with EG in LICs and statistical signi cance at 1% and 10% level of signi cance, respectively.But, when we analyzed the individual impact of FD and TO on EG then found a negative relationship with EG in LICs but statistically signi cant at 5% & 1% levels of signi cance.Our study proved hypothesis 3 (b) and supported the joint effects of ICT infrastructure, FD, and TO encouraging EG in LICs (Sassi and Goaied, 2013).9. Therefore, we failed to prove hypothesis 2(a) in the case of LICs but in the case of HICs we did not reject hypothesis 2 (a) so our results are considered for the ICT infrastructure and FD more encouraged in HICs as comparison of LICs.When we check the individual impact of FD, we found it's a positive impact on EG in and we found FD has a negative relationship with EG only in model 8. So, the majority of the different Models found FD have a positive relationship with EG in HICs.Moreover, TO has a negative impact on EG in HICs.More trade openness in any country is negatively affected to the business activities because liberal legal rules and regulations increase FDI in ow which is harmful at the point of view ownership transfer (Kumari and Sharma, 2017).Apart from TO, control variables (capital formation, consumption, and population growth) positively correlate with EG in all Models.These ndings were also found in the study of (Sassi and Goaied, 2013) and (Das, Khan, and Chowdhury, 2016).As we expected, ICT infrastructure and FD have a positive relationship with EG, but in HICs jointly broadband and FD have a negative relationship with EG.Therefore, we can conclude that the marginal effect of the internet, secinternet, telephone with FD is stronger than the joint effect of broadband and FD.Our study proved hypothesis 3 (a) and supported the joint impacts of FD and TO encouraging EG in HICs (Sassi and Goaied, 2013).The impact of ICT infrastructure and its joint implications are summarized in Figures 2, 3, and 4 using tables 8, 9, 10, and 11.
Figure 2 shows the individual impact of ICT infrastructure.We can see that the internet positively affects economic growth, and its in uence is almost the same in low-income and high-income countries.Secure internet and broadband negatively in uence the GDP per capita (EG) of LICs while the positive in uence on HICs.Telephone lines are positively in uencing the EG of LICs and HICs.From gure 2, it can be concluded that secure internet and telephone lines signi cantly impact the high-income countries' EG.
Figures 3 and 4 show the joint effects of ICT infrastructure with FD and FD*TO, respectively.Observations from Fig. 3 show that the joint effect of the Internet and secure internet is favourable on both LICs and HICs.When we look at the broadband and telephone lines, both are having a reverse correlation with EG in LICs and HICs.Observations from Fig. 4 show that when we look at the joint impact of ICT infrastructure with FD and TO, EG of LICs are positively in uenced while HICs are negatively affected.Overall observation drawn from Figs. 2, 3, and 4 shows that individual ICT infrastructure and jointly ICT infrastructure along with FD can play a crucial role in improving the GDP per capita of HICs when we invest in it.However, in the case of LICs, it can be observed that ICT infrastructure, FD, and TO show a positive relationship with GDP per capita.Therefore, to improve the GDP per capita of LICs, it is recommended to invest in ICT infrastructure along with FD and TO.

And Policy Implication
Our study analyses the joint impact of ICT Infrastructure, nancial development, trade openness on EG for the sample of 85 countries (including 27 low-income and 58 high-income countries) from 2000 to 2019.Our results reveal that individual (ICT infrastructure, FD, and TO on EG), and joint impact ICT infrastructure, FD (including different models like internet*FD, secinternet*FD, broadband*FD, and telephone*FD) on EG in LICs and HICs.And, another combination has been analyzed to observe the joint impact of ICT*FD*TO on EG in LICs and HICs.
Based on the analysis of this paper, we contributed to the literature review related to ICT infrastructure and nancial development in two different samples (LICs and HICs).In HICs, FD does not have a positive relationship with economic growth; so, our hypothesis 2 (b) failed to prove.But when we estimate the combined effect of ICT infrastructure and FD, we found that internet*FD, secinternet*FD, broadband*FD have a positive relationship with economic growth.Only telephone*FD has a negative impact on EG in LICs.Therefore, nancial development works as a harmonizing for the ICT infrastructure.ICT infrastructure, nancial development, and trade openness have joint effects on both economies, but results vary in both countries' context.Thus, we have framed the number of policies and suggestions for both countries (LICs and HICs).With the help of these policies, both countries increase potential output capital gains and maintain sustained GDP per capita.
Given the above results, broadband has a positive relation with economic growth, although broadband (or high-speed) internet access is not lavish, but is a basic need to encourage human and economic development in low and high-income countries (World Bank, 2021).Till now, only around 35% of the population in low-income countries has access to the internet compared to high-income countries (85% access to the internet).It may be due to poor infrastructure in low-income countries.Broadband networks and services play a signi cant role in making smart ICT infrastructures like Intelligent Transport Systems and Smart Electric grids.Therefore, the government of developing countries should focus on access to the internet and broadband because it creates jobs in the ICT sector, develops skills, helps to reduce poverty, bridges the digital gap, and helps in making the international connections.

3 .
Data Collection, Hypothesis And Methodology 3.1 Data collection and assessment issues Panel data analysis is becoming more popular among the researchers because panel data provide multiple bene ts compared to crosssection and time-series data set.In this work, we have used the panel data considering LICs and HICs and the time frame from 2000 to 2019.All the variables data is collected since 2000 because data related to the ICT variables for the underdeveloped countries is available from 2000.The variables data were collected from the World Bank Indicators (WDI, 2019) except ICT infrastructure variables.

Hypothesis 3 (
a): Joint effects of FD, and TO encourages EG in LICs and HICs.Previous studies found mixed results regarding the relationship between EG and trade openness.As a few examples, Shahbaz and lean (2012) argue that TO promotes the EG of Pakistan's economy.Similarly, Eris and Ulasan (2013) found no direct relationship betweenTO and EG in cross-country data.On the other hand,Kumari, et al., (2021) found that TO has a positive and long-run with EG in India.Thus, we also expect the joint effects of FD and TO will encourage EG in both LICs and HICs.Hypothesis 3 (b): Joint effect of ICT infrastructure, FD, and TO promote to EG in LICs and HICs.
Das, Khan, and Chowdhury 2016; Kumari et al., 2021; Myovella, Karacuka, and Haucap, 2020), nancial development, trade openness, and ICT as suggested by Barro et al. (1991).Secure internet servers, broadband servers, mobile users, and individual internet users are used as proxies of ICT infrastructure.From the following (Myovella, Karacuka, and Haucap, 2020; Sassi and Goaied, 2013) literature, we have used government consumption, capital formation, population growth as control variables.Therefore, the proposed model is framed as mentioned in Eq. (1).Where, β 0 represents the coe cient of the parameters.
EG it is represented by GDP per capita (Constant in US$), telephone is represented by the [ xed telephone subscriptions (per 100s people) [individuals using the Internet (% of the population)], secinternet is represented by [Secure Internet servers (per 1 million people)], broadband is represents by [ xed broadband subscriptions (per 100 people)]

Figure 1 Framed Hypothesis Figure 2
Figure 1 2. Literature ReviewAs we have studied in previous papers (S.H.Lee, Levendis, and Gutierrez, 2012; Deloitte, 2012; Vu, 2011; Gruber and Koutroumpis, 2011; Lam and Shiu, 2010; Antonopoulos and Sakellaris, 2009; Waverman, Meschi, and Fuss, 2005) that ICT represents as an ICT infrastructure, ICT diffusion, telecommunication, telephone penetration, mobile users, digitalization, digitization, etc. Different proxies are used to measure ICT infrastructure/ICT diffusion/ ICT expansion like the mobile telephone, internet users, broadband user's internet servers, etc. Whereas, the impact of ICT on EG is well discussed in the previous studies (Appiah-Otoo and Song, 2021; Habibi and Zabardast, 2020; Parwantoa and Wulansari, 2020; Adeleye and Eboagu, 2019; Albiman and Sulong, 2016; Bertschek and Niebel, 2016; Ghosh, 2016; Das, Khan, and Chowdhury, 2016; Ahmed and Ridzuan 2013; Dimelis and Papaioannou (2010)examined the effect of ICT and nancial development on EG in developed and developing countries from 1993-2001.In this paper, FDI was used as a proxy of nancial development and applied the GMM Model.They found various results based on the GMM Model; rstly, ICT positively affects developing countries.Secondly, FDI has contributed to the overall, but it positively affects developed countries.Thirdly, FDI does not have a signi cant impact in developing countries.AlthoughDimelis and Papaioannou (2010)and Pradhan, Arvin, and Norman (2015) studies included nancial development and ICT diffusion in empirical estimation, they were ignored the joint effects of ICT and nancial development on EG.In the same way, Pradhan, Arvin, and Norman (2015) analyzed the relationship between ICT infrastructure, nancial development, and EG in Asian countries from 2001 to 2012.They have applied panel co-integration techniques with the results of short and long-term causal relationships among the ICT infrastructure, nancial development, and EG.

Table 1
Shaw (1973)rastructure should accelerate EG by providing the development and adaptation of innovation processes.The new EG theories suggest the growth effects of modern communication processes in countries.The EG theories suggest that the ICT arrival of internet technology may have different qualities.Good ICT infrastructure may accelerate the foster competition and new ideas to develop products, processes, business models & how it circulates from one place to another, facilitating economic growth.Therefore, we also support the previous studies (Parwanto and Wulansari, 2020; Adeleye and Eboagu, 2019; Albimanand Sulong, 2016) and expected ICT infrastructure signi cantly impacts EG in LICs and HICs.Financial development encourages EG in both countries (LICs and HICs countries).Financial development is the most prominent independent variable that highly impacts the EG of any economy.Firstly, Goldsmith (1969) andShaw (1973)included FD and EG to nd out the kind of relationship.Before, EG theory did not include nancial development as an endogenous variable in the regression model.However, a growing number of previous studies shows how allocated resources, diversi ed risks, and intermediation savings is important contributor to EG (Greenwood andJovanovic, 1990; Jibili et al.   1997).Thus, we also support these studies and expect that FD encourages EG in both countries (LICs and HICs countries).
Hypothesis(b): Jointly, ICT infrastructure and FD positively correlate with EG in both countries (LICs and HICs countries).The above hypothesis supports the ICT and FD to encourage the EG, but it is checked individually only in the previous studies.Now we expect ICT infrastructure and FD jointly more effective to encourage the EG in both sample.We support Ghosh, (2016); Habibi and Zabardast, (2020); Appiah-Otoo and song, (2021) and expect jointly ICT infrastructure and FD to positively correlate with EG in LICs and HICs.Because jointly ICT infrastructure and FD are most signi cant factors affect to the economic growth.

Table 2
Name of variables, proxies, measured and data source Descriptive statistics and correlation analysis Tables3 and 4show the descriptive analysis of LICs and HICs.The average EG for the sample of LICs is 625.856, and Std.Dev. is 371.990.The average and std.dev of internet, secinternet, broadband and telephone is 9.741, 0.558, 0.793, 1.528 and 14.468, 0.860, 1.070, 3.302, respectively.In HICs (Table4), the average and std.dev of EG is 31773.85and 21389.4,including min 1659.908,and max is 118823.6.For HICs, we have observed an average of internet, secinternet, broadband, and telephone is 60.731, 7194.66,16.007, and 37.49, respectively.

Table 7
Comparative Models impact of ICT infrastructure, FD and TO on EG Table8shows the individual and joint effect of ICT infrastructure and FD on EG.When we check the individual effect of ICT infrastructure (internet, secinternet, broadband, and telephone) on EG, we found that internet and telephone have a positive impact on EG and are statistically signi cant at the level of 5%.On the other hand, secinternet and broadband have a negative impact on EG but are statistically signi cant at 10% and 5% level.AR (1) and AR(2)prove that all estimated versions of this model meet Areellono bond criteria for the valid estimations, so do not reject the hypothesis of the GMM Model, and it supports hypothesis 1.Moreover, Table8also revealed that the joint impact of internet*FD (0.025), secinternet*FD (0.101), broadband*FD (0.056) found a positive impact on EG and signi cant at 5% and 1% level of signi cance.Only telephone*FD (-0.023) have a negative effect on EG.Our ndings support Sassi and Goaied (2013) for MENA countries.On the other hand, in the individual analysis, the internet (0.002), secinternet (0.05), and telephone (0.003) have found positive relation with EG and signi cant at 5% level of signi cance.In contrast, broadband (-0.006) has a Source: Authors' estimations, Note: Std.Dev shown in bracket and *10%, ** 5% and *** 1 level of signi cance negative relation with EG in LICs but is signi cant at a 5% level of signi cance.Hypothesis 2 (a) has been failed to prove that FD encourages EG in LICs.When we check the individual impact of FD than found that it has a negative impact on EG.But on the other hand, we proved hypothesis 2 (b) because jointly ICT infrastructure and FD positively correlate with EG in both countries.TO, capfor, and positively affect EG in the entire models.These ndings support Sassi and Goaied, (2013) and Das, Khan, and Chowdhury, (2016).

Table 8
Individual and jointly effect of ICT infrastructure and FD on EG (LICs) Source: Authors, estimations based on data set.Note: Std.Dev shown in bracket and *10%, ** 5% and *** 1 level of signi cance Source: Authors' estimations based on the dataset Note: Std.dev shown in bracket and *10%, ** 5% and *** 1 level of signi cance

Table 10
shows the individual and joint impact of ICT infrastructure and FD on EG in HICs.The results of the GMM Model shows that AR (1) and AR (2) proved that there is no serial correlation under the predetermined variable and error terms of the different period.This helps to choose the set of lags that is used as instrumental variables in the GMM method.Moreover, Table10also revealed that the joint impact of internet*FD (0.003), secinternet*FD (0.015), telephone*FD (0.083) found a positive impact on EG and signi cant at 5% and 1% level of signi cance.Only broadband*FD (-0.035) has a negative relation with EG.On the other hand, in the individual analysis, internet (0.002), secineternet (0.040), broadband (0.001), telephone (0.034), and FD (0.001) have found positive relation with EG and signi cant at 5% level of signi cance in HICs.In contrast, broadband (-0.006) and secinternet (-0.037) have a negative relationship with EG in LICs but signi cant at 5% and 1% levels of signi cance, respectively as shown in Table

Table 11
Lee, Gholami, and Tong, 2005)ICT infrastructure, FD, and TO on EG in HICs.Jointly impact of FD*internet*TO (-0.116),FD*secinternet*TO (-0.225),FD*broadband*TO (-0.160) and FD* telephone* TO (-0.194) are negative relations with EG in HICs and statistical signi cance at 1% and 10% level of signi cance, respectively.But, the Joint impact of FD*Internet*TO (0.638), FD*secinternet*TO (0.103), FD* broadband* TO (0.096) and FD* telephone* TO (0.031) are positive relations with EG in LICs and statistical signi cance at 1% and 10% level of signi cance, respectively as shown in Table10.Therefore, we can conclude that ICT infrastructure plays a signi cant role in improving the EG of LICs.Youse (2011) study argues that there is no signi cant effect of ICT infrastructure on EG in LMICs.It may be due to the improper use of available human capital or labour; maybe they feel an extra burden with new technologies.Thus, people need higher salaries with higher ICT use and require a stronger nancial sector.Without further enhancement in ICT uses, the country's overall development is not possible.Most of the nancial intermediaries (including banks and stock markets) adopted ICT facilities like AI, machine learning, data mining, etc.If the banking sector and the stock market do not accept ICT, the number of changes to increase labour productivity will be minimal.Based on a comparison of LICs and HICs ICT infrastructure, we found that LIC will immediately require vital support of R&D, ICT infrastructure, IT skill enhancement program, human resource, more trade openness, reconstructive manual process, and new IT business Model (S.-Y.T.Lee, Gholami, and Tong, 2005).Although, few LICs adopted mobilization due to the down price in mobile but still internet and broadband penetration facilities remain unequal between developed and developing countries.
On the other hand, LMICs like Bolivia and Ghana have taken IT projects for the developed ICT infrastructure in their education sector, nancial institutions (including bank and stock/share market), and industries that help to encourage EG (Das, Khan, and Chowdhury, 2016; Rodas and Lopez, 2007).Similarly, with the help of ICT facilities, developed countries bene t from high-speed internet, free Wi-Fi (selected places), developed e-Commerce channels, sales distribution channels, etc.It also helps to create new job opportunities for the people.In contrast, LICs are still struggling for ICT infrastructure, FD, and more TO because these countries only enjoy ICT

Table 10
Individual and Joint effect of ICT infrastructure and FD on EG (HICs) Source: Authors' estimations based on dataset Note: Std.Dev shown in bracket and *10%, ** 5% and *** 1 level of signi cance.