1 Introduction

Many nations have made the commitment to achieve Net-Zero emissions by the year 2050 throughout the previous few decades. But because of the sharp rise in greenhouse gas (GHG) emissions, global warming has become a significant policy problem on a worldwide scale [1,2,3]. The main objective of a circular economy (CE) is to explore the avenues those leads towards a more sustainable and harmless world by utilizing the virgin sources by applying renewable methods in terms of energy found [4, 5]. The world is searching the new methods and techniques in terms of circularity to avoid any kind of worse consequences for the upcoming generations.

By converting to a circular economy, greenhouse gas emissions could be cut by 39% and demand for virgin materials could be eased by 28%, as per the latest Circularity Gap Report (2021). Researchers have shown that because they require energy to run, cyclical systems deplete resources, produce waste, and emit pollutants as per the findings of the some recent studies [see 6, 7]. As a result, a circular economy appears to have a mixed influence on carbon emissions, ranging from no impact to a positive one. There may be disagreement over the results of the circular economy due to the variety of ways it influences carbon emissions. Indicators from the circular economy were used as explanatory variables in the study stated above to assess the direct impact of the circular economy on carbon emissions. [8] found that indirect reduction of carbon emissions is another benefit to the circular economy.Using modern technology to control GHGs or making agreements to reduce emissions are two ways that nations throughout the world are attempting to combat climate change [9]. Most nations still want to reach net-zero carbon emissions by 2050, although only the governments of 55% of the world have established definite targets for further carbon emission reductions by 2030 according to [10]. Furthermore, the energy infrastructures of developing nations still heavily rely on fossil fuels. In the meanwhile, concerns with climate change and environmental pollution are brought on by the extensive use of coal and oil in the industry as well as transportation [11].

Energy utilization and production have a crucial role to promote economic activities and world's sustainability as SDGs for the 2030 goals have great significance for all countries [12]. Prior to the 1990s, a huge portion of the energy was produced by utilizing fossil fuels, which fueled global economic expansion but also increased GHG, which led to environmental damage for the current population and upcoming generations. This problem compelled global decision-makers to act to stop the pollution of the environment from getting worse and demand a swift switch to renewable energy sources. As a result, several international accords were established, including Countries of Preferences (COP)26 and (COP)27, and others upcoming series of the said accords are expected to held in future. As a result, the global energy sector has undergone a paradigm change, with renewable energy sources predicted to account for 61% of worldwide power output by 2030, up from 21% in 2021.

This study uses dataset of 28 low income countries for the period of thirty years from year 1990 to 2019. Countries use energy to initiate economic growth processes, expand their economies, and reach their full potential in terms of productivity. The world's energy consumption is constantly rising due to notwithstanding crises and the inclusion of technological innovation into methods of energy production.

For a nation to flourish, there may be pressure to use non-renewable carbon-based energy sources that causes significant pollution. In this way, the issue of the environment's ability to maintain its energy use becomes a goal of economic policy. In fact, governments must strike a balance between maintaining international agreements that encourage cooperation in the fight against global warming by lowering CO2 emissions and the need to preserve their nation's growth and prosperity. In this way, Paris Agreements and the recently finished COP27 in Egypt aim to build ambitious environmental policies for combatting climate change at the state level as well as on a global scale.

Literature is filled in the empirical estimation of the renewable and non-renewable energy implications on the environment [12]. In this study, we computed Circular Economic Index (CEI) on the basis of renewable energy sources (Hydro, Wind and Solar and Biofuels and Waste). Digitization Index is also generated on the basis of Internet Users, Fixed Telephone and mobile cellular telephone. CEI and Digitization Index are computed by using Principle component analysis (PCA) and regressed on Environment. Two different models with same regressors and by using two different regressands i.e., carbon emission and ecological footprint are estimated.

2 Review of literature

2.1 Economic growth and environment

Researchers give serious attention to the topics of economy and ecology. Empirical research on the association between environmental indicators and economic variables are widely available in literature. The inverted U-shaped correlation between carbon emissions and economic development was discovered by the [12,13,14]. In contrast from the above context, it was discovered that there is an inverted U-shape connection between carbon emissions and economic development [15,16,17,18]. According to certain research i.e., [16, 19, 20] the association between carbon emissions and economic development is constantly increasing. It was also shown in multiple studies [21, 22] that there is no meaningful connection between Carbon emission (CO2) and economic growth. Therefore, we are persuaded to identify the most current developments in low income nations as a result of the researchers conflicting findings about the relationship between environmental protection and economic growth. [23] found positive correlation between economic growth and carbon emission among other explanatory indicators. On the basis of these studies, economic growth included as  a explanatory indicator for the present study to quantify the impact on environment of the low income countries.

2.2 Financial development and environment

Environmental sustainability along with financial development has been carefully examined by the researchers over the past ten years and is essential for economic growth in all aspects. The results of the existing research conflict with each other, as some concluded that financial development has a good influence on the environment while other found negative impact. Studies like [24,25,26,27,28] argue that financial growth results in a better environment. In contrast, other research [29,30,31,32,33] asserted that financial development had a detrimental impact on the environment. The current study used financial development as a key indicator that has a significant influence on environmental policies for lower income nations since policies are crucially needed for the poor segments of society in terms of financial development.

2.3 Energy use and environment

Earlier studies by [16, 26, 34] attempted to investigate how overall energy consumption—the sum of all renewable and nonrenewable energy—affects environmental quality and indicated that carbon emissions are favourably impacted. However, [35] found that beneficial impact of energy use on environmental deterioration only holds true when the entire amount of energy consumed is divided between renewable and non-renewable energy sources. In addition, a more reliable indicator of environmental deterioration has recently been introduced in the literature in place of the conventional focus on carbon emissions. Now recent studies are using ecological footprint as a proxy for environment degradation as it is more complete indicator than the frequently utilised carbon emission [35, 36].

There has been an increased focus on non-renewable resource in research and empirical studies have also included aggregate and de-aggregated energy usage, single-country analysis, panel analysis, and more. For instance [37], reveals that Turkey's increasing ecological footprint is due to increased energy consumption and gross domestic product. In contrast to the conclusions given above, study by [38] for the Organization for Economic Cooperation and Development (OECD) countries estimated that the level of environmental performance can be ascribed to patterns of energy use. Following the use of various environmental quality indicators, such as carbon emissions, ecological footprints, and environmental performance levels of the countries [12, 38], in particular, come to the conclusion that the OECD countries are beginning to align with various environmental policies for environmental safety. Natural resources including energy have significant implications for the developing countries as found by [39]. Since low-income countries are focusing on energy consumption without taking any care about the use of renewable and non-renewable resources, there is an urgent need to change the direction of energy production in these countries. As energy use has become a crucial element for the environmental sustainability therefore present analysis take it as an important variable.

2.4 Circular economy and environment

The sustainability-related economic concept employed in the CE model aims to keep the value of goods, materials, and resources in the economy for as long as is practicable while also minimising the quantity of energy produced. According to [40] this paradigm represents a convergence of environmental, economic, and social factors, is the antithesis of the linear model. The CE system provides a societal model that optimises the stocks and flows of materials, energy, and waste, with its primary goal to ensure the effective use of resources. Due to this and changing raw material prices, the CE promotes supply security and the reindustrialization of the national area. According to [41, 42] it successfully converts trash into raw resources, acting as a model for a future system. It is impossible to overlook the significance of circularity for the future of the whole globe, and low-income nations must fill a significant policy vacuum in order to protect the environment for future generations. Energy efficiency leads towards the more sustainable outcomes for the high income economies found by [20].

2.5 Digitization and environment

A study by [18] quantified the impact of ICT and other controlled variables on Carbon emissions in Morocco for a period from 1985 to 2020. The effects of digitalization on the environmental health have been observed through a variety of techniques in the past literature. By applying digitization significant environmental concerns including the disposal of solid waste, food waste and agricultural waste are being addressed. Many authors have studied these concerns including [43,44,45,46]. There are lot of ways in which digital technologies may improve biodiversity and shorten its impact on the ecological systems. Digitization may also be utilised to develop business strategies those support the conservation of biodiversity as per the research of [71]. In a similar line, inadequate levels of digitization are a problem for low-income countries, and this has a negative impact on the environment. Recently [47] has investigated the relationship between digitization and the environment for East Asia nations.

2.6 Trade and environment

International trade is advantageous for both the economy and the environment [48, 49] and may help to increase environmental efficiency. The increase in national revenue brought about by trade liberalisation has been acknowledged by several writers, and this has increased public demand for environmental protection. Similar to this, increased global trade encourages spending on environmentally friendly technologies. The pollution halo hypothesis, supports these claims, contends that internationalisation enables the transfer of cutting-edge ecological technology and environmental management systems (EMS) from developed to underdeveloped nations [50]. Trade promotes large-scale industrial production, which raises energy consumption and the usage of hazardous chemicals, and worsening the environment's condition. International businesses are also drawn to some countries with laxer environmental regulations [51]. Another study estimated that trade has a bad impact on the environment in the case of Pakistan. Results found the worse effect of agricultural export on environment [52].

3 Need for the study

The specific literature focuses on how economic growth, energy use, financial development, trade, digitization and circularity impacting the environment for different regions but there is the sufficient space to estimate energy based circularity and ICT based Digitization those have influence on the environment. Given the theoretical and empirical insights, we focused the low income economies (those have huge impact in terms of energy consumption and environment, list of the selected countries attached at Appendix Table 12) to elaborate the circularity implications in terms of environment.

4 Methods and materials

4.1 Theoretical framework

This study estimated impact of carbon emission and ecological footprint on circularity, economic growth, digitization, financial development, trade and energy use. Two variables (circular economy and digitization) out of six are assumed to get negatively correlated with the environment and the other four variables (economic growth, financial development and energy use) are assumed to have positively correlation with the environment proxies. Figure. 1 shows the theoretical framework of the model. Second generation econometric techniques are employed for data analysis and Bias-corrected linear dynamic panel data estimation has been applied for long run estimates. Brief description of the variables is listed below in the Table 1.

Fig. 1
figure 1

Theoretical framework

Full size image
Table 1 Description of the variables
Full size table

4.2 Data framework

The current study explores the impact of circular economy (CE), economic growth (lnGDP), financial development (lnFD), trade (lnTRADE), digitization (DIGIT) and energy use (lnEU) on carbon emission (lnCO2) and ecological footprint (lnEFP) for a period from 1990 to 2019 for twenty-eight low income countries. Selected countries have a very little issues of data cleaning, missing values and outliers. So, the selection process of the countries was very crucial as there were more than 28 countries in that list but due to missing data, the only 28 countries were selected for the analysis. These variables have considerable impact on environment [53,54,55,56,57,58,59,60,61]. To avoid data sharpness and heteroscedasticity, the dataset (variables) for this study were transformed into a natural logarithm except CE and digitization, as these are in index form and do not have larger numeric values. Equation (1) expresses the initial form of the Model-I:

$$CO2i,t = f ({GDP}_{i,t,};{TRADE}_{i,t,};{FD}_{i,t,};{EU}_{i,t,};{CE}_{i,t,};{DIGIT}_{i,t})$$
(1)

In Eq. (1) CO2, GDP, TRADE, FD, EU, CE and DIGIT stands for carbon dioxide emission, gross domestic product, trade, financial development, energy use, circular economy and digitization, respectively. These indicators are the key impacting factors towards the environment as earlier research found [16, 23, 59, 60, 62,63,64]. Initial model Eq. (1) transformation into the estimated panel model as follows:

$$ln ({CO2}_{it}) = {\gamma }_{0} + {\gamma }_{1} ln ({GDP}_{it})+ {\gamma }_{2} ln ({TRADE}_{it})+ {\gamma }_{3} ln ({FD}_{it})+ {\gamma }_{4} ln ({EU}_{it})+ {\gamma }_{5} ({CE}_{it})+ {\gamma }_{6} ({DIGIT}_{it})+ {\varepsilon }_{it}$$
(2)

In the preceding Eq. (2), i stands for cross sections (countries), t for time periods. The stochastic error factor is shown by the symbol µ in Eq. (3). All under-examined variables' natural logarithms are denoted by the symbol ln. The parameters 1, 2, 3, 4, 5, and 6 represent, respectively, the elasticity of the GDP, TRADE, FD, EU, CE, and DIGIT variables. The second model’s initial form is presented in Eq. 3 as follows:

$${EFP}_{2i,t} = f ({GDP}_{i,t,};{TRADE}_{i,t,};{FD}_{i,t,};{EU}_{i,t,};{CE}_{i,t,};{DIGIT}_{i,t})$$
(3)

Indicators given in the above equation are the key impacting factors towards the ecological footprint as earlier research found [23, 43, 65,66,67,68,69,70,71]. Second initial model Eq. (3) transformation into the estimated panel model is given as follows:

$$ln ({EFP}_{2it}) = {\theta }_{0} + {\theta }_{1} ln ({GDP}_{it})+ {\theta }_{2} ln ({TRADE}_{it})+ {\theta }_{3} ln ({FD}_{it})+ {\theta }_{4} ln ({EU}_{it})+ {\theta }_{5} ({CE}_{it})+ {\theta }_{6} ({DIGIT}_{it})+ \mu_{it}$$
(4)

Here t, i indicates the time periods and cross-sections for the ecological footprint model. θ0 shows the intercept and θ1, θ2, θ3, θ4, θ5, and θ6 shows the elasticities of GDP, TRADE, FD, EU, CE and DIGIT, respectively.

4.3 Empirical Framework

Graphical framework of the econometric methods is implied in the following Fig. 2. Panel Data for the 28 lower income countries were extracted from the different sources i.e. World Bank, International Energy Agency, Global Footprint and World Economic Data Base. After refinement of the data set, four Cross Sectional Dependency tests employed to the data sets. In the third step, two panel unit root tests i.e., Cross-sectional Augmented Dickey-Fuller (CADF) and Cross-sectional Augmented Panel Unit Root test (CIPS) were applied. To estimate the long run result of the models, Biased Corrected linear dynamic approach used in the fourth step. Dumitrescu-Hurlin granger causality was applied in the last. Conclusion and policies were formulated on the basis of above analysis for the selected sample countries.

Fig. 2
figure 2

Econometric methods

Full size image

4.3.1 Cross sectional dependency (CSD).

In order to assess cross-national interdependence, this study used Pesaran scaled LM, Pesaran CSD, Breusch-Pagan LM, and Bias-corrected scaled LM test. Prior to doing the panel unit root testing, it is critical to test potential CSD among all study variables. If CSD is included in the dataset, the standard panel stationary tests may fail to correctly identify the integration order. Mathematical form of the CSD is given in Eq. (5)

$${\text{CSD}}=\sqrt{\frac{2{\text{T}}}{{\text{S}}\left({\text{S}}-1\right){\text{S}}}(\sum_{{\text{i}}=1}^{{\text{S}}-1}\sum_{{\text{K}}={\text{i}}+1}^{{\text{S}}}{\widehat{{\text{Corr}}}}_{{\text{i}},{\text{t}}})}$$
(5)

4.3.2 Panel unit root.

In order to determine the appropriate level of cointegration order in panel data in the presence of CSD, [72] recommended both parametric and non-parametric techniques. The first generation of panel unit root tests, according to [73] do not take into account the effects of CSD, heterogeneity, or over rejection of the null hypothesis (H0). The study makes use of the second-generation CADF and CIPS panel unit root tests to get around this problem. Equation for CADF and CIPS are given in Eqs. (6–9):

$$\Delta {{\text{Z}}}_{{\text{t}}-1}= {\mathrm{\alpha }}_{{\text{i}}}+{\upbeta }_{{\text{i}}}{{\text{Z}}}_{{\text{i}},{\text{t}}-1}+ {\upgamma }_{{\text{i}}}{\overline{{\text{Z}}} }_{{\text{t}}-1}+{\updelta }_{{\text{i}}} \Delta {\overline{{\text{Z}}} }_{{\text{i}},{\text{t}}}+{\upvarepsilon }_{{\text{it}}}$$
(6)
$${\overline{{\text{Z}}} }_{{\text{t}}-1}= \frac{1}{{\text{P}}}\sum_{{\text{i}}=1}^{{\text{P}}}{{\text{Z}}}_{{\text{i}},{\text{t}}-1}$$
(7)
$$\Delta {\overline{{\text{Z}}} }_{{\text{i}},{\text{t}}}= \frac{1}{{\text{P}}}\sum_{{\text{i}}=1}^{{\text{P}}}{\Delta \mathrm{ Z}}_{{\text{i}},{\text{t}}}$$
(8)
$${\text{CIPS}}= \frac{1}{{\text{P}}}\sum_{{\text{i}}=1}^{{\text{P}}}{{\text{CADF}}}_{{\text{t}}}$$
(9)

4.3.3 Westerlund cointegration test.

This paper used a modified version of [74] test called the Pesaran and Yamagata test [75] to identify slope heterogeneity in the data stream. In contrast to the alternative hypothesis, which implies that the slope parameters are variable, the null hypothesis makes this assumption. Nevertheless, notwithstanding [76] it is difficult to assess the best estimates and show the quantity of data when CSD is made the problem's solution, is present. Slope heterogeneity disruption, structural discontinuities, and non-stationarity issues are all taken into account by [77]. A solution to these problems is offered by [78], who analyse both homogeneous and heterogeneous concerns with probable structural fractures and correlation errors. The cointegration analysis is performed by [77] utilising the CCEMG technique between the variables, providing an accurate assessment of cross-sectional dependency independent of the strength of the statistics or the stationarity/non-stationarity issues.

4.3.4 Bias-corrected method.

The dynamic panel, "bias-corrected method of moments," with a fixed estimator (BC-MM), newly created by [79] was employed in the long run. The benefit of applying this approach is that it analytically corrects the fixed effects estimator's first-order constraint. Higher-order autoregressive models, heteroscedastic errors, fixed and random assumptions may all be accommodated by this technique. It is also suggested to use panel-corrected standard errors, which enable reliable inference in dynamic models with cross-sectionally correlated errors.

4.3.5 D.H granger causality test

The Dumitrescu-Hurlin [94] panel causality test is used in the current study to look at the causal linkages between the variables. Decision-makers may use the information from the test to develop logical and practical policies. The Dumitrescu-Hurlin causality test is advantageous in that it addresses cross-sectional dependency and is suitable for both N > T and N < T. The alternative hypothesis in the Dumitrescu-Hurlin panel causality test supports the existence of a causal link, in contrast to the null hypothesis, which asserts that there is no causal relationship in the panel data. The next section presents the analyses' findings.

5 Results and discussion

5.1 Descriptive and correlation results

Before proceeding with any modeling, it is important to ensure that the data being used is normally distributed. In this study, the descriptive statistics of the data were examined to assess normality. The findings are presented in Table 2, which shows that the mean values of LNCO2, LNEFP, LNGDP, LNTRADE, LNFD, LNEU, CE, and DIGIT are within the normal range, indicating the absence of outliers in the data. The standard deviation values suggested an acceptable level of variation in all variables. Skewness values for all variables, except LNEU and DIGIT, are within the range of −1 to + 1. The Jarque–Bera test indicates that LNCO2, LNEFP, LNGDP, LNTRADE, LNFD, CE, and DIGIT reject the null hypothesis of normal distribution, implying that these variables are not normally distributed. The correlation matrix results show that LNTRADE is negatively correlated with both CO2 and EFP, and the same correlation was found for CE and EFP. However, LNGDP and LNEU, CE, LNFD and DIGIT are positively correlated with ecological footprint and CO2 emissions.

Table 2 Descriptive statistics
Full size table

Through the calculation of Tolerance and Variance Inflation Factors (VIF) for each independent variable (LNGDP, LNTRADE, LNFD, LNEU, CE, and DIGIT), the examination of multicollinearity among the independent variables was carried out. The outcomes of the multicollinearity test employing tolerance and VIF are shown in Table 3. The findings show that the explanatory variables are not multicollinear since the VIF values are less than 5 and the Tolerance values are not less than 0.2. This implies that the factors are separate from one another and may be seen as independent variables that affect CO2 emissions and ecological impact. Some other assumptions are testified by applying Heteroscedasticity [80] Omitted variable bias [81] and Serial correlation in panel-data models [82]. All the tests validate the data set in terms of error variance, serial correlation and the omitted variables bias.

Table 3 Results of VIF test, Heteroscedasticity, Omitted variable bias and Serial correlation
Full size table

5.2 Results of CSD tests

This study employed the Breusch-Pagan LM, Pesaran Scaled LM, Bias-corrected Scaled LM, and Pesaran CD cross-sectional dependence (CSD) tests. The findings provided in Table 4 demonstrates that the cross-sectional units in the seven variables are interdependent and that there is inadequate evidence to justify cross-sectional independence. As a result, the alternative interdependence hypothesis is acknowledged.

Table 4 Results of CSD tests
Full size table

5.3 Results of unit root tests

The results of the CIPS and CADF tests are shown in Table 5. The CIPS unit root test shows that few of the variables LNGDP, LNTRADE, LNFD and CE of the variables are stationary at 1% and 5% of the level. At a 1% level of significance, the variables become stable after taking into account the initial difference. The CADF test data further support the fact that the two variables LNGDP and LNFD are stationary at the level but become stationary at the 1% significance level following the first difference. As a result, the unit root null hypothesis is disproved. These findings show that all variables for the low-income nations under examination are part of the first differentiable stationary process I (1).

Table 5 Results of unit root tests
Full size table

5.4 Results of cointegration test

The results of the Westerlund cointegration test are shown in Table 6 and show that the CO2 and EFP models, denoted as specifications 1 and 2, respectively, are cointegrated in the long run. This indicates that when the variables LNCO2, LNGDP, LNTRADE, LNFD, LNEU, and DIGIT were investigated for cointegration, the null hypothesis that there is no cointegration is rejected based on the four test statistics. Since there is enough information to reject the null hypothesis that there is no long-term link in favour of the alternative hypothesis that there is a long-term correlation at a significance level of 1%, long-run estimates may now be utilised.

Table 6 Westerlund cointegration test
Full size table

5.5 Discussion of the main models results

Following the initial panel data analysis, the Bias-corrected (BC) linear dynamic panel data estimation technique was utilized in the subsequent stage to predict the long-run impact of the model's underlying variables. The coefficients obtained from this analysis are presented in Table 7, representing the long-term effects of GDP, TRADE, FD, EU, CE, and DIGIT (in Model 1) on CO2 emissions and ecological footprint (in Model 2). The BC results reveal that all the long-run coefficients are statistically significant. Specifically, the GDP coefficient is positively significant, indicating that GDP has a positive impact on both CO2 emissions and ecological footprint in low-income countries. A 1% increase in GDP will result in a 0.016 and 0.018 percentage point increase in CO2 emissions and ecological footprint, respectively. However, the magnitudes of the coefficient values for GDP differ between CO2 emissions and ecological footprint, indicating that GDP has more impact on ecological footprint than on CO2 emissions in the sampled countries.

Table 7 Biased corrected linear dynamic approach
Full size table

According to [83] the employment of traditional technologies, which are energy inefficient and eventually cause environmental deterioration, is responsible for the detrimental effects of GDP on environmental quality. This implies that environmental deterioration may worsen as economies grow, especially if there is a lack of ecological consciousness. This ignorance could make it harder to push for the adoption of improved technology and clean energy, to help protect the local and larger ecosystem. Additionally, throughout the research period, multiple businesses and facilities that significantly contribute to environmental degradation continued to operate [83]. Our findings therefore coincide with those of [65] for China and [83] for the E5 countries.

Our results support the generally known belief that energy usage greatly raises CO2 emissions as well as ecological impact. We found that for a LIC, an upsurge in energy consumption led to increases in CO2 emissions and ecological footprint of 0.728 percent and 0.226 percent, respectively, despite a lower fraction of renewables. The importance of energy consumption to economic growth in these countries and how detrimental it is to the environment are both emphasised in this conclusion. It is not unexpected that the conclusions reached by [84] for the United States and [85] for the Austrailia line-up with our results.

In contrast, our research shows that trade in low-income nations has a detrimental effect on both CO2 emissions and the ecological footprint. The results of earlier studies by [86] for OECD nations and [87] for OIC nations support this conclusion. The scale of trade's influence on ecological footprint, however, is greater than that on CO2 emissions. In particular, a 1 percent increase in trade increases ecological footprint and CO2 emissions by 0.017 and 0.048 percent, respectively. According to [88], one reason for this negative link is the technical impact's supremacy over the size effect and composition effect. The results support the hypothesis that trade little promotes cross-border technology transfer in low income countries, providing nations with access to more environmentally friendly technologies that can reduce ecological footprint levels and improve environmental quality.

The pollution haven hypothesis, which holds that foreign companies use trade to bring new and cleaner technology to host economies, and subsequently improve the environmental quality, is also consistent with the detrimental effects of trade on environmental indicators in low-income economies [58]. The fact that low-income countries are more dedicated to environmental sustainability and only accept the transfer of ecologically beneficial technologies supports the argument that commerce has a negative impact on environmental degradation.

Additionally, our data demonstrates that the expansion of the financial sector contributes significantly to rising CO2 emissions as well as ecological impact. In the examined countries, a 1% rise in FD specifically results in reductions of 0.005 and 0.001 percent in the carbon emissions and EFP, respectively. The positive coefficient values imply that the financial sector in low-income nations is vital for funding green initiatives like investments in renewable energy. This can lessen the reliance on non-renewable energy sources and lower their proportion in the generation of all energy. Additionally, by encouraging the import of consumer durables and fixed capital expenditures made in less energy-intensive industries, the financial sector may lower the reliance on energy derived from imported non-renewables like oil. These findings suggest that low-income nations might encourage environmental sustainability by making changes to their financial growth strategies.

The advent of digitalization has rendered outdated conventional production methods. The results of the current study demonstrate a negative relationship between digitization and the deterioration of the environment as measured by CO2 emissions and ecological footprint models. The study finds that digitalization has a significant negative impact on ecological footprint and CO2 emissions. Particularly, an increase of 1% in digitalization in the chosen countries is associated with decreases of 0.009 and 0.037 percent in CO2 emissions and ecological footprint, respectively. This suggests that digitalization is necessary to stop environmental degradation in a variety of ways. For instance, it advances e-commerce and enhances sustainability, accessibility, and safety (through intelligent GPS). Additionally, digitalization affects manufacturing processes through its substitutional influence, boosting energy efficiency and reducing the intensity of energy usage. These results can be linked to digital technology's ability to lessen stress on biodiversity and the environment, which are measures of ecosystem health.

For the effective implementation and supervision of environmental regulations, governments can gain from the precise information supplied by digitization. These conclusions are supported by research done by [10] for BRICS states, and [89] for resource-intensive nations. [90] claim that through encouraging ecological governance, the expansion of the internet considerably aids China's attempts to lessen environmental pollution. According to [91] in the early stages of development, digitalization and environmental degradation have an inverted U-shaped relationship because of high consumption and cost-effects that lead to high environmental pollution. The circular economic index also shows that the CO2 emissions and ecological footprint in low-income countries have a negative and significant coefficient value. The study found that for every 1% increase in CE, CO2 emissions and ecological effect are reduced by 0.002 and 0.009 percent, respectively. According to this, adopting CE practices is essential for achieving sustainable resource management, waste reduction, biodiversity preservation, environmental quality maintenance, and economic sustainability. Similar to this, corporations are showing increasing interest in the circular economy idea since it promises to increase commercial prospects while reducing environmental impact, according to [92] These findings concur with those of [93] who examined data from 131 different nations. According to their study concentrating on the CE is also a cutting-edge strategy to accomplish sustainable development by lowering greenhouse gas emissions, energy use, and waste and safeguarding the environment.

In addition, to check the long run estimates study checked the robustness of coefficients. Table 8 demonstrate the results of robustness check applied for both models with FMOLS technique. Results showed that coefficients have the same signs and are significant as provided by the biased corrected model.

Table 8 Robustness check (FMOLS)
Full size table

5.6 Results of D-H granger causality test

We employed the panel causality test to look into potential heterogeneity problems in panel data. In this method, W-bar and Z-bar statistics are used. The results of the D-H panel causality test are shown in Table 9. The findings showed a bidirectional causal link between commerce and CO2 emissions, an association between financial development and CO2 emissions, and a connection between ecological footprint and financial development. In addition, the D-H results demonstrated a two-way causal association between CE and CO2 emissions, as opposed to the CE and ecological footprint results, which only showed a one-way causal relationship. But there was a direct causal relationship in both directions between energy usage and CO2 emissions as well as between energy use and ecological effect.

Table 9 D-H causality test
Full size table

The findings of the study also indicate a causal link between GDP and ecological footprint in both directions. This is in line with study by [95] for South Asian nations, which demonstrates that while ecological footprint decreases as GDP increases, higher ecological footprint levels lead to an increase in GDP demand in low-income nations. Additionally, the analysis shows a two-way causal relationship between GDP and CO2.

6 Conclusion and Policy Implications

6.1 Conclusion

Study in hand and evaluates the relationship between CEI and environmental footprints via creative experimentation. Results showed the positive externalities of circular practises in terms of environmental health from a low-income countries group.

The results of this study have implications for many different types of professions, such as policy makers, researchers, and sustainability practitioners. In particular, this work contributes in a unique way and can help in the management of sustainable development and global emissions objectives. The knowledge gathered from this study may also be used to optimise the integration of clean energy practises in the energy sector, resulting in better implementation and administration of clean energy strategies.

The Sustainable Development Goals (SDGs) of the United Nations cannot be attained without a healthy environment. The world's economies have just seven years to cut pollution, stop biodiversity loss, and stop global warming in order to meet their 2030 goals. This study offers a novel approach to understanding how environmental deterioration in low-income nations is affected by digitization and the CE. Other variables in the research, such as energy use, financial development, economic growth, and trade, were included to combat the bias brought on by aspects that were left out. Using yearly data from 28 low-income countries for the period 1990–2019, we suggested two empirical models to investigate the environmental impact of circularity, digitization, financial development, trade, economic growth and energy use. The CSD, CADF, CIPS, and Westerlund cointegration tests were carried out before figuring out the long-run association between the variables. The evaluation of the long-term relationships between the relevant variables was conducted using bias-corrected (BC) linear dynamic panel data estimation. The empirical findings demonstrated a negative correlation between digitalization and the circular economic index, and CO2 emissions and ecological footprint. The results have practical implications for the selected sample countries as well as the rest of the developing countries.

In the 28 low income nations in particular, environmental degradation is decreased with each increase in the CE and digitalization indices. Additionally, it becomes clear from the study's findings that trade significantly worsens the state of the ecosystem by increasing CO2 emissions and ecological footprint. It's noteworthy to find that GDP, FD, and energy consumption all have a positive influence on the environment through higher CO2 emissions and ecological footprint, which makes it harder for low-income nations to accomplish sustainable development and environmental goals.

6.2 Policy implications

In order to encourage environmental sustainability in low income countries, this paper makes a number of policy proposals. First, the governments of these sampled countries must strike a balance between their plans for environmental protection and economic development, keeping in mind that rising GDP results in environmental deterioration. This is only feasible if the future and present economic growth strategies of the chosen nations include consideration of the issue of environmental welfare. This is crucial since the earlier growth strategies of these countries did not encourage environmental protection. Without a major and faster adoption of renewable energy technology and infrastructure that is implemented with regard for the long-term life-cycle implications of their generation, use, recovery, and reuse, the green growth goals of these countries will fail.

Indeed, the results of this study imply that these governments should prioritise increasing digitization while broadly implementing CE principles in order to establish a more sustainable economy. Financial resources should be used to support CE and digitization practices, with an emphasis on green initiatives that could bring more benefit into the environment in the presence of negative externality too. These initiatives should support energy efficiency, pollution control, carbon reduction, and cleaner production in important industries and fields. Fourth, the sampled nations must increase the proportion of alternative energy sources that are substantially cleaner in the energy mix in light of the findings on the unfavourable environmental effects of energy use.

Governments should look for strong options for limiting the use of current fossil-fuel power production and try to shift towards the renewables sources, which are more environmentally friendly than the non-renewables. An immediate switch to 100% renewable energy is not feasible for any country, so along with accelerated investment in renewables, governments should also seek robust options for doing so. Governments should, for instance, concentrate more emphasis on boosting energy efficiency across economies to prevent the unnecessary and damaging use of fossil fuels and the resulting greenhouse gas emissions. Governments should formulate regulations to opt the renewable energy sources for their upcoming plans to avoid severe circumstances due to climatic reactions.

6.3 Limitations

Though, this study takes important factors influencing CO2 emissions and ecological footprint in low income economies, even it is not immune to limitations. Several crucial aspects of this investigation are yet unknown. For example, this study utilizes data until 2019. To explore the present relationships, future research can get more recent data for these variables. Moreover, other proxies of environmental degradation might be considered such as GHGs, load capacity factor, NO2 etc. that might explain the relationship in a better way. Therefore, future study should take into account these proxies as well.