Carbon Neutrality Challenge: Analyse the Role of Energy Productivity, Renewable Energy, and Collaboration in Climate Mitigation Technology in OECD Economies

Abstract

Carbon neutrality has been widely acknowledged as a challenge to environmental mitigation and global climate change policy. The current study examines the association between collaboration in climate change mitigation technologies (CMTs), energy productivity (EP), natural resources rent (NRR), renewable energy consumption (REC), and environmentally related tax (ET) on CO₂ emissions for a panel dataset of 30 OECD economies from 1990 to 2020. This paper employs panel data econometric techniques such as AMG, CCEMG, and CS-ARDL. The empirical outcomes show that CMTs, EP, REC, and ET have a negative effect on CO₂ emissions, indicating that their increase will bring about the reduction of carbon emissions, whereas NRR has a positive impact on CO₂ emissions, suggesting that its increase will raise CO₂ emissions. Most interestingly, REC and EP play a leading role in all selected variables by decarbonizing and effectively converting conventional energy into clean, green energy in the process of energy production and utilization. Finally, the OECD countries are anticipated to transition their energy from conventional resources to renewable sources, which will be validated by the increase in energy productivity and the adoption of clean and green technology in the short term.

1. Introduction

Climate change mitigation has been a worldwide subject in the context of constantly increasing levels of CO₂ emissions. Seven consecutive years after the Paris Agreement in 2015, the level of carbon emissions keeps rising, and degradation timelines for the 1.5 °C and 2 °C environmental achievements have become ever more rigorous. For many countries, mitigating carbon emissions to decrease the global temperature is a fundamental preference. We focus on the OECD economies’ two methods for achieving degradation in carbon emissions levels. First, EP and RE suggest decreasing total carbon dioxide emissions from production activities. Second, climate change mitigation technologies collaborate to achieve the goal of lowering CO₂ emissions. In 2015, the United Nations Framework Convention for Climate Change (UNFCCC) developed a consensus to minimize the growing level of the world temperature to 2 °C by 2020 [1].

On the other hand, many countries agreed on 2 °C, and the initiative’s efforts to additionally mitigate the change to 1.5 °C were vigorous [2]. These efforts can be explicitly obtained by decreasing global annual greenhouse gases (GHG), widely considered the primary factors of climate change [3]. Achieving these objectives requires strict determination by countries in giving their utmost preference to decreasing CO₂ emissions. For this purpose, the OECD countries are setting up a collaboration plan in climate change mitigation technologies. This plan will help developed and developing countries nourish the adoption of green technology, thus abating climate change. Technology transfer from developed to developing countries has become an important issue for climate debaters, with the concept encompassing technology’s requirement of enabling activities, assessments, technology information, and capacity building. The Kyoto Protocol mechanism will also contribute to the transfer and dissemination of climate change technologies related to clean energy. To help mitigate climate change, many international institutions such as the United Nations request that developed countries assist with transferring technology to developing countries [4,5] and participate in international collaboration on the development of climate change mitigation technologies to tackle climate change with combined efforts [6,7]. Investment in ICT technologies can significantly reduce CO₂ emissions [8,9]. Realizing the importance of creating climate change mitigation technologies is not a new study area. Many studies have investigated the success and progress of different organisations and government initiatives developed to assist the diffusion and establishment of climate abating technologies. On the other hand, some research has associated environmental concerns with productivity and economic growth [10,11], or dealt with it from the perspectives of international trade and the FDI nexus [12,13]. Climate change significantly reduces agriculture trade exports [14].

Moreover, carbon emissions are the primary target for developed economies and international organisations to effectively limit energy utilisation, consequently assisting in tackling conventional-energy-related environmental pollution. Figure 1 compares generated CO₂ emissions between OECD economies and other regions worldwide. From 1960 to 2020, there was a positive trend in CO₂ emissions. This positive trend significantly intensified from 1990 to 2020. The figure shows that the OECD and higher-income economies are primarily responsible for increasing environmental pollution compared to other groups.

Renewable energy and energy productivity, in addition to technological innovation, are substantial strategies for mitigating the environmental destruction of the earth. In the view of the EKC theory, technological innovation is acknowledged as a key element in acquiring the last stair of the EKC proposition [15]. Nevertheless, with energy poverty and less-developed economies, it is expensive to bear climate change mitigation technologies in developing economies. The transformation from carbon-intensive energy to green, renewable energy is bearable, fast, and uncomplicated in advanced economies compared to low-income countries. Specifically, the efficient advancement of environmentally friendly technology in the energy industry abates environmental degradation by enhancing the renewable energy ratio and energy productivity [16].

This paper aims to investigate the impact of all factors, including collaboration in climate mitigation technology (CMT), energy productivity (EP), natural resources rent (NRR), renewable energy consumption (REC)), and environmental tax (ET), on CO₂ emissions for OECD economies. Establishing the magnitude and direction of these relationships will provide insight for appropriate policymaking. This analysis employs panel data econometric techniques such as the augmented mean group (AMG) technique, while the fully-modified ordinary least square (FMOLS) and dynamic ordinary least square (DOLS) are estimated for robustness purposes. Through this study, developing economies can use lessons to design a framework for international collaboration on CMTs.

This research investigates the following issues. First, do CMTs help limit the adverse effect of air contamination in OECD economies? The present study will provide key insights into whether advanced economies with the latest technologies achieve adequate advantages in energy productivity and energy transaction. Second, does a huge dependence on the consumption of natural resources affect environmental quality? Third, what role can renewable energy sources play in the decarbonisation process in OECD economies? The primary motivation for exploring these questions emerges from the global issue of environmental destruction and climate change, a challenge faced by governments and policymakers. In addition, it is clear that a considerable energy ratio is utilised in developed countries and that their role in environmental deterioration is unavoidable. Therefore, this study explores a panel of 30 OECD, high-income countries that have higher production and consumption of clean, green energy, a higher income per capita, and global innovation in advanced technology.

The rest of the paper is organised as follows: the literature review highlights previous work on CO₂ emissions, CMTs, EP, NRR, REC, and ET. The econometric strategy section includes the theoretical study framework, data source, variable description, and model specifications. The empirical results and discussion section illustrates the detailed findings obtained from AMG, CCEMG, and CS-ARDL. The last section is a conclusion and proposes policy recommendations and limitations.

2. Literature Review

The concept of technology transfer, which includes technology-needs assessments, technological knowledge, enabling activities, and capacity building, has been a top priority for climate negotiators. The Global Environmental Facility, which acts as the Convention’s funding source, enables initiatives and funds various concrete projects in these areas. Some of these initiatives specifically attempt to lower the costs of environmentally friendly technology by boosting their market shares. Three new funds were established by the Marrakech Accords to address various issues, including technology transfer for adoption and mitigation. These tools were specifically created to foster favourable circumstances and to provide leverage for private funding.

The latest study proposed by [17] explored the fact that, despite the production of renewable energy increasing year by year, countries are not moving effectively and accelerating sufficiently to elude global warming of 2 °C. The authors of [18] analysed the effects of REC on carbon emissions by utilising the FMOLS and DOLS techniques; the findings of their study revealed that REC is helpful to the environment in the G-7 group. Likewise, [19] suggested the same conclusions for 42 developed nations. A negative influence of REC on carbon emissions was found in [20], while a forward influence of non-renewable energy was found in South American economies. In addition, [21] explored the impact of renewable and non-renewable energy on environmental contamination in 22 African nations, employing ARDL and PMG econometric techniques. Their study found that non-renewable energy raises the degree of pollution, and renewable energy mitigates it as expected.

Moreover, a recent study proposed by [22] explored the relationship between trade, resource rents, energy intensity, and carbon emissions for 93 countries from 1995 to 2017. Their findings showed that NRR has a notably positive impact on carbon emissions from production and consumption in low-quality government countries. Meanwhile, they found an inverse association between NRR and production-based CO₂ emissions. This suggests that NRR can decrease carbon emissions from production. At the same time, there is an insignificant relationship between NRR and carbon emissions from consumption for upper- to high-quality government countries. Climate-friendly technologies could play an important part in slowing warming [23]. To slow the increasing global warming, achieving carbon neutrality by 2050 is the most urgent undertaking on the earth [24]. Carbon neutrality indicates a net zero emission of CO₂, which means that the amount of CO₂ emitted by human activities and the amount absorbed by human activities reach a balance within a certain period of time. CO₂ emitted by human activities includes the combustion of fossil fuels, industrial processes, agricultural and land use activities, etc. CO₂ absorbed by human activities includes afforestation to increase carbon absorption, carbon capture through carbon sink technology, etc. [24].

The influence of NRR, non-renewable energy, trade openness, environmental regulation, and the financial development of BRICS countries from 1996–2018 was examined in [25]. They found that the left variables, except for non-renewable energy and trade openness, positively and significantly impact CO₂ emissions. This is because the OECD economies are rich in resources; they are therefore less dependent on energy imports and can assist in tackling environmental mitigation better than those economies that lack natural resources [26]. This means that NRR is conducive to increasing CO₂ emissions in the BRICS. Negative effects of natural resources on environmental deterioration were found in [27]. Similarly, [25] revealed that, in the case of Pakistan, natural resources help increase the ecological footprint. In addition, the latest study proposed by [27] explored whether NRR decreases environmental quality in European regions. The authors of [28] found that natural resources adversely affect carbon emissions for BRICS economies by evaluating the importance of environmental mitigation. Furthermore, [29] explored the importance of natural resources in the ecological footprint of China; they found that a rise in NRR can enlarge the ecological footprint.

Furthermore, the present study highlights the critical significance of ET in climate mitigation policy. The author of [30] found that ET can mitigate CO₂ emissions and enhance energy efficiency, environmentally friendly technological innovation, and environmental quality. The authors of [31] revealed that ET could achieve a series of goals such as sustainable economic growth, a clean environment, and decreasing unemployment. Moreover the double function of ET, assumed by the “dual dividend” [32], enhances the quality of the environment, i.e., the green dividend, and also acquires less deformity tax, the “blue dividend” [33,34]. On the other hand, ET can raise industries’ production costs and deteriorate their competition in the international market [35]. Additionally, manufacturing enterprises will move the higher cost of ET to users, affecting low-paid people and aggravating the income gap [36,37]. If manufacturing firms move the raised cost of ET to consumers, it is of a higher probability to undermine the strategy against environmental mitigation [38,39].

There are piles of literatures about the association between CO₂ emissions and innovation, which are reviewed in

Table 1. Literature review of the association between CO₂ emissions and innovation hypotheses.

Authors	Region/Period	Method	Association Between CO2 Emissions and Innovation	Contribution
[40]	BRICS, 1980–2016	CCEMG	Positive	Empirical test of the linear relationship between innovation and CO2 emissions in the BRICS.
[41]	OECD, 1996–2015	PQR	Negative	Studied the influence of technological innovation on CO2 emissions in OECD economies.
[42]	OECD, 1990–2015	ARDL	Negative	Explored the innovation–emission relationship to find the trademark’s role in the carbon mitigation.
[43]	Developing countries, 1990–2016	FMOLS	Negative	Studied the positive and negative impacts of innovation on pollution in developing countries.
[44]	BRICS, 1990–2017	AMG	Negative	Explored the present literature by visualising the nexus between technological innovation and CO2 emissions.
[45]	OECD, 1993–2014	Differenced GMM	Positive	Studied FDI, innovation, and the energy–environment–growth nexus for OECD countries.
[46]	62 economies, 2003–2018	MEM	Negative	Explored the connection between CO2 emissions and financial risk for 62 countiries.
[47]	OECD, 1990–2014	LDM	Negative	Studied how technology progress plays a crucial part in environmental improvement.

.

This study focuses on the connections between energy production and CO₂ emissions in OECD economies. Only a few studies have highlighted the significance of the energy productivity–emissions relationship in the energy–environment literature. The latest study proposed by [48] found that EP, innovation, and exports have a negative impact on carbon emissions from consumption in G7 economies by adopting the CS-ARDL model between 1996 and 2017. The critical characteristic of EP is that it enhances the environmental quality by reducing energy costs and improving energy efficiency [49]. The importance of EP for acquiring sustainable development has remained an utmost preference [50]. In addition, many industrialised economies, such the OECD, G7, and higher-income economies, have utilised EP and RE to overcome economic and environmental concerns [51,52]. Despite the pivotal role of EP, we find that empirical research on the association between CO₂ emissions and EP is mainly extinct. Further, the previous empirical studies urged researchers to estimate CO₂ emissions by employing two individual proxies, i.e., CO₂ emissions from consumption and CO₂ emissions from production [53,54,55].

3. Theoretical Framework

We exploit a few econometric analyses tools, including the unit root test, co-integration test, AMG, CCEMG, and CS-ARDL models, to explore long-term relationships between the variables. Moreover, this paper employs FMOLS and DOLS models to cross-check the robustness of the performed models. The econometric strategy is employed to explore the linkage between CO₂ emissions, natural resources, energy productivity, and environment taxes for OECD economies.

Proposition 1. 

Climate mitigation technology is a tool for abating CO₂ emissions in the OECD countries.

Figure 2 illustrates the final energy consumption by various sectors in OECD member states. In this section, we propose an environmental mitigation strategy framework with various energy production and consumption indicators. The design of the theoretical framework, which combines two substantial contributing factors to CMT, is illustrated in the literature review section. It precisely validates the positive and robust association between CMT and carbon emissions in OECD economies. Due to the growing environmental concerns, i.e., increasing temperature levels, climate change, and CO₂ emissions, it is presumably essential to consider environmental protection from international collaboration and CO₂ emission processes. Moreover, the framework of this study demonstrates important concepts by illustrating connections and overlapping links between the three substantial areas.

It is widely acknowledged that the effective use of energy from renewable sources, such as collaboration and innovation in green technology, will mitigate environmental pollution. Converting vehicles to hybrid or electric will decrease the proportion of energy needed, and the pollution generated will substantially decline. Likewise, the consumption and production of electricity has increased productivity through technological innovation, and renewable sources decrease the amount of energy indispensable to development, reducing carbon emissions.

Proposition 2. 

Energy productivity is a source of low carbon emissions.

Conventional and non-conventional energy sources substantially influence environmental quality and sustainability. The OECD states produce and consume energy from various sources as shown in Figure 3. This figure indicates that oil is the most significant source for energy generation, making it the largest generation source in the OECD states. It accounts for 51% of the total energy production and is followed by natural gas, which accounts for 19% of the total energy generation; electricity, which shares a ratio of 18%; wind and solar, which account for 7%; biofuels and waste, which accounts for 3%; and heat, which accounts for 2%. The utmost responsibility of the OECD countries is to efficiently utilise such energy sources to reduce the essence of environmental quality by adopting energy productivity.

Proposition 3. 

Renewable energy is a driving force of environmental degradation.

According to [56], nearly all the OECD states signed the historical climate mitigation accord known as the Kyoto Protocol to mitigate their aggregate greenhouse gas (GHG) emissions by an average of almost 5.2% beneath the levels of the year 1990 in five years after 2008. Thus, to tackle environmental pollution and diversity and to secure energy flow, there has been an increased interest in adopting renewable energy consumption sources in recent years in OECD economies. This promising interest has been encouraged by several government-friendly policies such as tax reductions, incentives for the innovation of renewable technologies, and feed-in tariffs. From Figure 4, we can see that hydro contributes to a significant portion of clean energy generation. It is followed by wind, which is the second largest renewable energy generator. In addition, solar PV is the third most significant contributor to clean, green energy production.

4. Econometric Strategy

In the next section, we comprehensively discuss the model specifications and econometric analysis, such as data diagnostic, panel unit root, slop heterogeneity, cross-sectional dependence, and Westerlund co-integration tests. We also employ AMG, CCEMG, and CS-ARDL models for the short-term and long-term relationships between the study variables. In addition, we adopt FMOLS and DOLS econometric estimators to verify the model robustness.

4.1. Model Specification

We developed a model to empirically explore the association between CMT, NRR, ET, REC, EP, and CO₂ emissions in the OECD countries. The model was designed as follows:

$${\mathrm{LnCO}}_{2\mathrm{it}}={\alpha }_0+{\alpha }_1{\mathrm{LnCMT}}_{\mathrm{it}}+{\alpha }_2{\mathrm{LnNRR}}_{\mathrm{it}}+{\alpha }_3{\mathrm{LnET}}_{\mathrm{it}}+{\alpha }_4{\mathrm{LnREC}}_{\mathrm{it}}+{\alpha }_5{\mathrm{LnEP}}_{\mathrm{it}}+{\epsilon }_{it}$$

(1)

where CO₂ is the dependent variable; ${\alpha }_1\dots {\alpha }_N$ denote the coefficients of the independent variables; CMT denotes international collaboration in climate change mitigation technologies; NRR represents natural resource rents; ET represents environmentally related taxes; REC is the renewable energy consumption; EP describes energy productivity; Ln represents the natural log; ${\alpha }_0$ donates a constant term; and ${\epsilon }_t$ represents the error term.

4.2. Data and Descriptive Statistics

Table 2. Countries included in the sample.

Australia	Ireland	Slovak Republic
Austria	Italy	Spain
Belgium	Japan	Sweden
Czech Republic	Korea	Switzerland
Denmark	Luxembourg	Turkey
Finland	Mexico	United Kingdom
France	Netherland	USA
Germany	New Zealand	Canada
Greece	Norway	Israel
Hungary	Poland	Portugal

presents the OECD countries included in the sample.

Table 3. Variable definition and sources.

Variable	Definition	Source	Period	Expected Sign
CO2	CO2 emissions (metric tons per capita)	OECD	1990–2020	—
CMT	International collaboration on climate change mitigation technologies (in percentage)	OECD	1990–2020	Negative
NRR	Total natural resources rents (% of GDP)	OECD	1990–2020	Positive
EP	Energy productivity per unit of GDP (USD, 2015)	OECD	1990–2020	Negative
ET	Environmentally related taxes (tax revenue per capita) (USD 2015)	OECD	1990–2020	Negative
REC	Renewable energy consumption (% of total final energy consumption)	OECD	1990–2020	Negative

describes all variables. We utilised panel data of 30 OECD economies from 1990 to 2020 for 902 observations [57]. In addition, the data selection was based on the latest data availability. The descriptive statistics of all variables after the logarithm are shown in

Table 4. Descriptive statistics.

	LnCO2	LnCMT	LnNRR	LnEP	LnET	LnREC
Mean	9.042	0.554	−17.87	9.149	−14.10	−15.23
Median	8.036	0.589	−17.42	9.181	−14.41	−14.99
Max	30.43	0.693	−12.75	10.420	0.000	−12.00
Min	2.581	−0.507	−23.15	8.182	−20.98	−17.73
Standard Deviation.	4.364	0.131	2.166	0.375	3.177	1.208
Skewness	1.539	−2.886	−0.181	−0.094	2.593	0.042
Kurtosis	5.946	15.934	2.565	2.828	13.08	2.734
Observed	902	902	902	902	902	902

.

4.3. Panel Unit Root Tests

We conducted a unit root test for all variables, utilising the test methods of ADF, PP [58], and LLC [59].

$$∆x_{it}={\alpha }_i+{\beta }_i{X}_{i,t-1}+{\gamma }_i{\overline{X}}_{t-1}+{\displaystyle \sum }_{k=0}^m{d}_{ik}∆{\overline{X}}_{t-k}+{\displaystyle \sum }_{k=1}^m{\delta }_{ik}∆X_{i,t-k}+{\epsilon }_{it}$$

(2)

where ${\overline{X}}_t$ denotes the mean time (T). According to the CADF statistic, the CIPS unit root can be evaluated as follows:

$$CI\widehat{PS}=N^{-1}{\displaystyle \sum }_{i=0}^{n}CAD{F}_i$$

(3)

where CADF denotes t-statistics.

4.4. Cross-Sectional Dependence Test

We employed the methods of bias-corrected, scaled LM estimation [60], CD estimation [61], and LM estimation [62] to explore the residual CD in selected variables. The equations were constructed as follows:

$$LM=T {\displaystyle \sum }_{i=1}^{N-1}{\displaystyle \sum }_{j=i+1}^N{\widehat{P_{ij}}}^2$$

(4)

$$C{D}_1=\sqrt{(\frac{1}{N\left(N-1\right)}} {\displaystyle \sum }_{i=1}^{N-1}{\displaystyle \sum }_{j=i+1}^N{\widehat{(T{P}_{ij}}}^2-1)$$

(5)

4.5. Panel Co-Integration Test

We adopted three different estimation techniques for the panel co-integration as follows:

(a)

The unit root test [63], which is an improved estimation differentiated as the Padroni co-integration technique [64];

(b)

Another co-integration technique, illustrated as the Kao co-integration method [65];

(c)

An error-correction-based panel co-integration technique [66].

In addition, it is essential to note that, as a longitudinal data series, it tackles the concern of cross-sectional dependency (CSD) and supplies consolidated outcomes. To ascertain the problem of cross-sectional dependence (CSD) for the variables, the authors employed the [66] (Westerlund 2007) co-integration method so as to acquire more coherent outputs.

$$GR={\displaystyle \sum }_{i=1}^N{\displaystyle \sum }_{t=1}^T{\widehat{F}}_{it}^2\left({\displaystyle \sum }_{i=1}^N\widehat{R_i}\right)$$

(6)

where GR represents the group mean variance-ration statistics, ${\widehat{F}}_{it}={\displaystyle \sum }_{k=1}^t{\widehat{F}}_{ik}, \widehat{R_i} {\displaystyle \sum }_{t=1}^T{\widehat{F}}_{ik}^2,$ and ${\widehat{F}}_{ik}$ denotes the residuals from the panel regression.

4.6. Panel Causality Test

The panel causality method was proposed by [67] to check the causality among variables, presenting a revised form of the non-causality approach [65]. It is a widely used econometric technique [68] since it is (a) more adaptive, giving consistent results whether T > N or T < N and (b) consistent for heterogeneous or unbalanced data. Based on Z-bar and W-bar statistics, we constructed the following equation:

$$Z_{i,t}={\alpha }_i+{\displaystyle \sum }_{j=1}^p{\gamma }_t^j{Z}_{i,t-j}+{\displaystyle \sum }_j^p{\gamma }_t^j{T}_{i,t-j}$$

(7)

where ${\gamma }_t^j$ indicates the auto-regressive parameters and the lag length is donated by j.

$$G_t=\frac{1}{N}{\displaystyle \sum }_{i-1}^N\frac{{\acute{\alpha }}_i}{SE\left({\acute{\alpha }}_i\right)}$$

(8)

$$G_a=\frac{1}{N}{\displaystyle \sum }_{i-1}^N\frac{T{\acute{\alpha }}_i}{{\acute{\alpha }}_i\left(1\right)}$$

(9)

$$P_t=\frac{\acute{\alpha }}{SE\left(\acute{\alpha }\right)}$$

(10)

$$P_a=T\acute{\alpha }$$

(11)

where ά indicates the error correction (EC), acquired by combining the values of T and P_a in the above-mentioned equations.

4.7. Long-Run Panel Estimations

The present study revealed that there is a long-term connection between three various estimators. First, an AMG estimator for long-run associations was developed by [69] by carrying the production function. From the beginning, AMG estimator was introduced as the alternative to the CCEMG estimation by [70]. The foremost dominance of employing the AMG technique is that it assists in remedying the outcomes in multifactor error terms and panel heterogeneity. Apart from that, the AMG technique is a long-term co-integration estimator introduced for an average period and number of cross-sections that provides a consolidated output [71]. Moreover, the AMG technique includes time-variant fixed effects and comprises a standard dynamic effect parameter. We divided the AMG estimator into a two-step method:

Step 1:

$$∆y_{it}={\alpha }_i+{\beta }_i∆x_{it}+{\gamma }_i{f}_t+{\displaystyle \sum }_{t=2}^T{d}_i∆D_t+{\epsilon }_{it}$$

(12)

Step 2:

$${\widehat{\beta }}_{AMG}=N^{-1}{\displaystyle \sum }_{i=1}^N\widehat{{\beta }_i}$$

(13)

where ∆ represents the change in the response variable; $y_{it}$ and $x_{it}$ describe observables; β denotes coefficients of country-specific estimations; $f_t$ illustrates insignificant common factors having heterogeneous factors; $D_t$ signifies the time dummies coefficient and dynamic process; ${\beta }_{AMG}$ represents the AMG estimator; and ${\alpha }_i \mathrm{and} {\epsilon }_{it}$ represent the intercept and error terms, respectively.

Moreover, the CCEMG estimation is applicable for cross-sectional dependence [72]. This technique provides comprehensive outcomes with the condition that the data have multifactor error terms and panel heterogeneity. The CCEMG technique can be estimated by employing the following procedure:

$$m_{it}={\varnothing }_{1i}+{\tau }_1{z}_{it}+{\vartheta }_i{p}_t+{\beta }_i{\overline{m}}_{it}+{\gamma }_i{\overline{z}}_{it}+{\mu }_{it}$$

(14)

where $m_{it}$ and $z_{it}$ are defined as observable variables in the regression; $p_t$ is defined as an observable common variable with heterogeneous coefficients; ${\tau }_1$ represents the country-specific effects; ${\vartheta }_i \mathrm{represents} \mathrm{the} \mathrm{heterogeneity} \mathrm{effect};$ and ${\varnothing }_{1i}$ and ${\mu }_{it}$ are defined as the intercept term and error term in the model, respectively.

To verify the relationship among variables, this paper utilised the cross-sectional autoregressive distributive lag CS-ARDL model [73]. The structure of the CS-ARDL carries both long- and short-run parameters, including error correction terms and a long- and short-term cross-sectional mean of every single concerned, selected variable. The CS-ARDL technique is superior to other econometric methods for a number of reasons. Firstly, the CS-ARDL technique provides consolidated estimates even when the variables are merged in a different order. Secondly, the CS-ARDL technique can provide precise outcomes for circumstances of both long- and short-term CSDs [73]. Thirdly, this method is a mean group (MG) estimation with heterogeneous slope coefficients. Based on the mean group, the CS-ARDL method is a more advanced form of the ARDL estimation, depending on the values of each cross-section and the cross-section averages. This covers the lags as proxies and unobserved common terms [74]. Finally, this method is an effective technique in the case of an exogeneity that shows due to the explanatory variable lagging in the model. Furthermore, previous studies have claimed that the endogeneity issue is controlled by adding the cross-sectional lagged averages in the specified model. The baseline regression model was designed as follows:

$$∆y_{it}=C_i+{\lambda }_i\left(y_{it-1}-{\beta }_i{X}_{it-1}-{\phi }_{1i}{\overline{y}}_{t-1}-{\phi }_2{\overline{X}}_{t-1}\right)+{\displaystyle \sum }_{j=1}^{p-1}{\theta }_{ij}\mathsf{\Delta }{y}_{it-1}+{\displaystyle \sum }_{j=0}^{q-1}{\zeta }_{ij}{X}_{it-j}+{\eta }_{1i}\mathsf{\Delta }{\overline{y}}_t+{\eta }_{2i}\mathsf{\Delta }{\overline{X}}_t+{\epsilon }_{it}$$

(15)

where $∆y_{it}$ represents the response variable determined by other explanatory variables; $X_{it}$ denotes the long-run association; ${\overline{y}}_{t-1}$ and ${\overline{X}}_{t-1}$ are the means of response and explanatory variables in the long term, respectively; $\mathsf{\Delta }{y}_{it-j}$ indicates the response variable in the short term; $\mathsf{\Delta }{\overline{y}}_t$ and $\mathsf{\Delta }{\overline{X}}_t$ are the means of the response and explanatory variables in the short run, respectively; and ${\epsilon }_{it}$ defines as the error terms in the model. Furthermore, ${\beta }_i$ denotes the coefficients of explanatory variables; t denotes time; ${\zeta }_{ij}$ and ${\theta }_{ij}$ denote the response and explanatory variables’ short coefficients, respectively; ${\eta }_{1i}$ describes the response variable mean; and and ${\eta }_{2i}$ represents the explanatory variable mean in the short run.

5. Empirical Results and Discussions

By sharing cost, effort, and information, international collaboration on climate mitigation technologies could facilitate technical alteration and expedite the direction of more clean and green technologies. Collaboration on climate change technologies between countries, not coerced contests between firms, may prompt governments to accelerate their efforts to support fundamental research in innovation and development. In addition, growing cooperation among countries on climate change mitigation technologies is helping prompt more countries to participate and take concrete actions to reduce their greenhouse gas emissions. The current globalization of innovation, investment, and trade should be seen as an opportunity to capitalize on the innovation and diffusion of green climate technologies. To tackle CO₂ emissions, governance and national policies are also essential as they contribute to an environment that is favourable to the extensive dissemination of clean and green climate technologies. Since 1990, a significant, increasing trend has been demonstrated in the quantity of CMTs in the OECD economies.

In addition, the present study adopted the cross-sectional dependence test, unit root test, and co-integration tests. It validated that the variables used in the panel have cross-sectional correlations. According to the CSD, CD-LM, and CD tests [75], the null hypothesis of no cross-sectional dependence is refused. The results of the CSD test are illustrated in

Table 5. Cross-sectional dependence test.

Test	Statistic	Prob.
Breusch–Pagan LM	1675.7 ***	0.000
Pesaran scaled LM	47.197 ***	0.000
Pesaran CD	6.6313 ***	0.000

Note: *** are for the significance levels of 1%.

.

The outputs of the slope homogeneity test are presented in

Table 6. Slope heterogeneity test.

Statistics	Test Value	p-Value
$\widetilde{∆}$	27.44 ***	0.000
${\widetilde{∆}}_{adjusted}$	30.56 ***	0.000

Note: *** are for the significance levels of 1%.

. The outcomes of the $\widetilde{∆}$ and ${\widetilde{∆}}_{adjusted}$ tests suggest a rejection of slope homogeneity. This indicates that any positive or negative variations in CO₂ emissions, CMT, REC, ET, and EP in each OECD economy would have an intensively positive or negative influence on other OECD economies. Furthermore, the following section verified the variables’ stationarity by employing the LLC and IPS unit root tests.

The outcomes of the panel unit root tests are tabulated in

Table 7. Panel unit root test.

	Levin–Lin–Chu		Im–Pasaran–Shin
	Level	First Difference	Level	First Difference
LnCO2	0.012	−8.879 ***	1.939	−13.98 ***
LnCMT	−5.825 ***	−27.91 ***	−5.821	−32.89 ***
LnREC	−3.566	−9.754 ***	1.745	−26.90 ***
LnNRR	−5.615 ***	−24.87 ***	−6.161 ***	−26.11 ***
LnET	−12.29 ***	−8.976 ***	−14.97 ***	−8.793 ***
LnEP	−2.568 ***	−21.36 ***	9.364	−24.97 ***

Note: *** are for the significance levels of 1%.

. To validate the existence of a unit root for each variable, we employed a series of unit root tests such as the LLC, ADF, and PP. The findings confirmed that all the variables are level at a first difference order 1(1) apart from NRR, which has a unit root at level 1(0). Thus, the outcomes of the following three estimations validate different econometric techniques.

The results of Westerlund panel co-integration tests are tabulated in

Table 8. Westerlund co-integration test.

Statistics	Value	Z-Statistics	p-Value
Gt	−2.118 **	−0.492	0.031
Ga	−6.695	2.122	0.983
Pt	−13.139 ***	−3.637	0.000
Pa	−10.105 ***	−4.153	0.000

Note: *** and ** is for the significance level of 1% and 5%, respectively.

. The outputs from WPC estimations illustrated a long-term connection among the variables. Further the outcomes of the unit-root tests and the Westerlund panel co-integration tests confirmed for employing the CS-ARDL, AMG and CCMEG methods.

The outcomes of the panel co-integration test are reported in

Table 9. Results of panel co-integration.

H0: No Co-Integration	Panel Tests	Statistics	p-Value(s)
Pedroni test within the dimension	Panel specific tests
	Panel PP statistic	−12.70 ***	0.008
	Panel ADF statistic	−14.31 **	0.028
Between dimension	Group-specific tests
	Group PP statistic	−3.657 ***	0.000
	Group ADF statistic	−3.938 ***	0.000
Kao Test	ADF t-statistics	−10.18 ***	0.000

Note: *** and ** is for the significance level of 1% and 5%, respectively.

. The key role of these two tests was to verify the robustness of the outcomes. The co-integration test revealed that the panel Philip–Perron (PP) statistics were notable at a 10% significance level. The panel-augmented Dickey–Fuller (ADF) statistics were significant at 5%. Furthermore, among dimensions, the group Philip–Perron (PP) statistics, group augmented Dickey–Fuller (ADF) statistics, and the Kao statistics were all strongly significant at 10%. The estimation results prove that panel co-integration exists between the variables; meanwhile, the Kao co-integration test also provided identical results; that is, the co-integration exists for a possible long-run association.

Table 10. Results from AMG and CCMEG estimations.

		AMG (LnCO2)			CCMEG (LnCO2)
	Coefficients	t-Values	Std.Error	Coefficients	t-Values	Std. Error
LnCMT	−2.082 ***	−3.80	0.547	−5.285 ***	−6.43	0.822
LnREC	−7.610 ***	−49.94	0.152	−1.945 **	−2.26	0.859
LnNRR	0.146 ***	11.24	0.013	0.276 ***	4.73	0.058
LnET	−0.123 ***	−8.58	0.014	−0.095	−1.05	0.091
LnEP	−2.596 ***	−34.49	0.075	−2.815 ***	−10.23	0.275
Constant	−75.197 ***	−31.31	2.402	0.588	0.47	1.246
Obs.	900			900

Note: *** and ** is for the significance level of 1% and 5%, respectively.

presents the outputs of the augmented mean group (AMG) and common correlated effect mean group (CCEMG) estimation techniques. The outcomes between CO₂ emissions and CMT showed a notably negative relationship at the 10% significance level, indicating that CMT is detrimental to carbon emissions degradation in the OECD countries. There is a notably negative association between REC and CO₂ emissions. REC made a valuable contribution to CO₂ mitigation in Morocco by adopting the dynamic ARDL model from 1985 to 2020 [9]. The ET coefficient showed a negative connection with CO₂ emissions, and there was a notably negative association between EP and CO₂ emissions. In contrast, the coefficient of NRR showed a notably positive linkage with CO₂ emissions.

The outputs of the CS-ARDL model are presented in

Table 11. Results of CS-ARDL estimation.

	Coefficient	t-Values	Standard Error
Short-Run Coefficients
ΔECT(-1)	−1.994 ***	−4.33	0.004
ΔLnCMT	−2.576 ***	−6.88	0.757
ΔLnREC	−1.391 ***	−3.26	0.856
ΔLnNRR	0.072 **	2.55	0.056
ΔLnET	−0.077 **	−1.58	0.097
ΔLnEP	−1.333 ***	−8.6	0.31
Long-Run coefficients
LnCMT	−5.165 ***	−6.82	0.374
LnREC	−2.797 ***	3.27	0.427
LnNRR	0.144 **	2.56	0.028
LnET	−0.15 **	−1.55	0.049
LnEP	−2.664 ***	−8.59	0.155

Note: *** and ** is for the significance level of 1% and 5%, respectively.

. The coefficient of CMT was notably negative, with a 5% significance level in both the short and long term, indicating that a 1% raise in improvement in CMT helps to decrease CO₂ emissions by 2.576% in the short term and 5.165% in the long term in the prosperous OECD economies. The output infers that an increase in CMT improves environmental pollution and sustainability in the selected advanced OECD economies. Similar findings were supported by the studies [76,77]. The significance of CMT in clean, green innovation is a primary factor. Therefore, the paper highlights the critical facts about CMT by performing a series of empirical analyses.

Likewise, the coefficient regression of REC was notably negative, indicating that a 1% increment in REC diminishes the carbon emissions by 1.391% in the short term and 2.797% in the long term, indicating that REC assists in abating environmental degradation in the OECD countries. The inverse association between carbon emissions and REC is based on the fact that, as affluent economies, the OECD countries have the capital and capacity to adopt clean energy sources. The studies [78,79] support the hypothesis of a long- and short-term dynamic connection between REC and carbon emissions in the OECD countries. Another study by [80] found similar results, suggesting that a growth in REC will improve environmental pollution in both the short and long term.

Moreover, NNR exerts a positive and significant influence on CO₂ emissions. A 1% increment in the utilisation of natural resources will increase CO₂ emissions by 0.072% in the short term and 0.144% in the long term. In addition, the extensive consumption of natural resources through mining, deforesting, and farming can have an adverse effect on the environment. The authors of [81] identified similar findings in the case of the BRICS economies between 1990 and 2015. The effect of natural resources on increasing the carbon emissions of the OECD economies is linked to the advancement of industrialised countries expediting unsustainable extraction and the utilisation of natural resources, raising the dependency of OECD countries on the use of natural gas and the import of fossil fuels.

ET is one of the key decisive factors of environmental protection. The coefficient value of ET was notably negative, suggesting that a 1% growth in ET improves environmental quality by 0.077% in the short term and 0.15% in the long term in the OECD economies. These results are consistent with the preceding works, such as [82], which studied seven emerging economies by employing the AMG and panel granger causality tests between 1994 and 2015. Moreover, the utmost goal of an ET is to bring significant revenue, to bring about behavioural alterations in industries to utilise clean and green environmental technologies, and to create change in consumers’ behaviour such that they will consume fewer pollutant commodities. Carbon-related taxes can alter the consumption and production structure for clean and green energy-relevant goods [83,84]. Additionally, [85] found that a 1% growth in ET per capita can mitigate carbon emissions by 0.033% in OECD economies. Likewise, the latest study by [86] revealed that ET is an essential source of low carbon emissions in OECD countries.

Finally, the connection between EP and CO₂ emissions was notably negative. EP is the total economic output per unit of production inputs [87]. This outcome indicates that a 1% growth in EP reduces carbon emissions by 1.333% in the short term and 2.664% in the long term. The authors of [48] found an inverse association between EP and carbon emissions from the consumption for G7 countries by employing a long- and short-term co-integration estimation of the CS-ARDL model from 1996 to 2017. Furthermore, EP growth produces a larger outcome at the price of less energy utilisation. EP tackles CO₂ emissions in three different ways. First, a raise in EP creates a decrease in energy per unit utilised for manufacturing. Second, EP assists to decrease the energy costs at the minimum level. Third, EP helps to abate oil dependency, which helps to reduce CO₂ emissions.

The present study was furthered by employing a robustness check using the FMOLS and DOLS models introduced by [88,89]. The output of the robustness tests is tabulated in

Table 12. Robustness check.

	FMOLS	t-Values	Standard Error	DOLS	t-Values	Standard Error
LnCMT	−2.986 ***	−3.389	0.880	−0.807	−0.692	1.167
LnREC	−8.175 ***	−5.602	1.459	−1.738	−1.031	1.685
LnNRR	0.186 *	1.676	0.111	0.094	0.903	0.104
LnRT	−0.157	−1.52	0.103	−0.553 ***	−3.423	0.162
LnEP	−2.290 ***	−4.744	0.483	−3.620 ***	−6.574	0.551

Note: *** and * are for the significance levels of 1% and 10%, respectively.

. The robustness tests validated that the results from the AMG, CCEMG, and CS-ARDL models are authentic, and the coefficients of the selected variables, such as CMT, REC, NRR, ET, and EP, had no change. In addition, the FMOLS and DOLS tests confirmed a similar relationship between the response variable CO₂ emission and the explanatory variables CMT, REC, NRR, ET, and EP.

Table 13. Granger causality test.

	Observed	F-Statistics	Probability	Causality
LnCMT→LnCO2	900	9.737	0.001	One-way
LnCO2 ≠ LnCMT		6.526	0.108
LnNRR→LnCO2	900	3.456	0.063	One-way
LnCO2 ≠ LnNRR		1.102	0.293
LnEP→LnCO2	900	5.633	0.017	One-way
LnCO2 ≠ LnEP		1.640	0.200
LnET→LnCO2	900	3.989	0.046	One-way
LnCO2 ≠ LnET		0.056	0.812
LnREC→LnCO2	900	5.124	0.023	One-way
LnCO2 ≠ LnREC		24.44	0.907

examines the causal connections among the variables, including CO₂, CMT, REC, NRR, ET, and EP, as per [65]. The results confirmed a one-way causality from CMT to carbon emissions, indicating that CMT effectively mitigates environmental pollution in OECD countries. Additionally, there was a one-way causality between NRR and CO₂ emissions, suggesting that NRR is an origin of increasing CO₂ emissions. These results are confirmed by [90], indicating that natural resources have a notably positive effect on CO₂ emissions in Belt and Road countries. Similarly, there was a one-way Granger causality between EP and CO₂ emissions, suggesting that higher EP positively affects CO₂ emissions.

In contrast, lower EP can negatively affect CO₂ emissions, which assists in decarbonisation. Furthermore, there was a one-way Granger causality between ET and CO₂ emissions, confirming that, as a critical factor, ET is an effective way to tackle environmental pollution. Lastly, a one-way causality between RE and CO2 emissions indicates that energy consumption is an origin of decreasing CO₂ emissions in OECD countries.

6. Conclusions and Policy Recommendations

The authors examined the effects of CMT, REC, ET, ER, and NRR on the CO₂ emissions of OECD economies. Based on an investigation of the stationarity level in the data sequences by employing the unit root test, the existence of long-term connections among the variables was verified by applying the Westerlund panel co-integration, finding that CMT, REC, and EP have a forward influence on the mitigation of CO₂ emissions in the OECD countries during the years 1990–2020 by econometric analysis. In contrast, NRR has an inverse effect on CO₂ emissions in both the short and long term. The present study identifies a number of in-depth insights into the CO₂ abatement methods in the OECD economies: EP, REC, CMT, and ET can decrease energy consumption, enhance energy efficiency, and promote the advancement of REC. While rational and scientific environmental taxes are pivotal for OECD countries, ETs have an adverse effect as they will reduce the performance of social welfare, economic growth, and carbon overspill.

This paper proposes the following key insight recommendations from a policy implications perspective. First, the OECD governments and organisations must design policies encouraging CMT, EP, ET, and REC. In addition, targeting CMT, clean, green technology and energy production can assist decision-makers in designing strategies to enhance environmental quality and abate environmental pollution in the OECD countries. For future advancement in CMTs and EP, the OECD economies may be required to encourage public–private organisations to participate in the CMT and the increase of EP.

Second, the present outcomes demonstrate that OECD economies would select environmentally friendly technologies that are cost-effective and efficient, allowing for a sophisticated conversion to renewable energy. By adopting environmentally friendly technologies, the OECD countries can decrease the inverse influence of CO₂ emissions on environmental quality and economic development. Further, the OECD economies should promote collaboration in eco-friendly technological advancement and the transition of the manufacturing sector towards renewable energy utilisation. By adopting these policy recommendations, the OECD states will improve environmental quality and benefit the economy by mitigating carbon emissions.

Third, the adverse impact of natural resource consumption on CO₂ emissions urges for increased concentration on environmental enhancement by encouraging the adoption of renewable energy and the efficient use of natural resources in the OECD countries. The most recent study proposed by [91] suggested that carbon neutrality is one of the biggest challenges as the level of emissions rises day by day. Their study proposed that governments and international organizations should work on the significance of energy efficiency and energy intensity.

There are some limitations in this paper, which will be improved in future studies. Firstly, we only took into account the OECD countries. Future work can highlight the outcomes of various areas, such as the G20, RCEP, BRICS, African countries, and the EU. Secondly, EP, REC, NRR, and ET were employed as control variables to explore the association between CO₂ emissions and CMT using the AMG technique. It is possible to utilise different econometric techniques for further investigation. It is also significant to compare the OECD economies with higher-income level economies for CMT, and EKC or STIRPAT theoretical models can be used to explore the cyclical effects of CMT on innovation. Lastly, this study investigated the link between CMT and CO₂ emissions by adopting a short- and long-run estimation; future studies can examine the causative and simultaneous association between CMT and CO₂ emissions by utilising an accompanying equation-modelling framework.