Does nuclear energy consumption mitigate carbon emissions in leading countries by nuclear power consumption? Evidence from quantile causality approach

Abstract

Nuclear energy has sparked international attention as one of the most important strategies for reducing emissions thanks to its ability to provide low-carbon power. Based on this interesting fact, the current research explores the effect of nuclear energy on CO₂ emissions in the leading countries by nuclear power consumption using a quarterly dataset from 1990 to 2019. The study employs the quantile-on-quantile (QQ) estimator, which accounts for both non-parametric and conventional analyses and enhances the provision of unbiased and consistent estimates. In addition, the Granger causality in quantiles approach is adopted to assess the causality in quantiles between the variables of investigation. The outcomes from the QQ estimator reveals that in the majority of the quantiles, nuclear energy contributes to decreased degradation of the environment in the USA, France, Russia, South Korea, Canada, Ukraine, Germany, and Sweden. Contrawise, the feedbacks from Spain and China expose that Nuclear Energy Consumption (NUC) contributes to the deterioration of the environment. Moreover, the outcomes of the causality test disclose that nuclear energy and CO₂ emissions can predict each other in the majority of the quantiles. The findings above provide profound ramifications for policymakers planning nuclear energy and CO₂-emission policies towards achieving sustainable environment in the sample countries and beyond..

Keywords

nuclear energy consumption quantile-on-quantile regression granger causality in quantile

Introduction

The global economy is resolute toward resolving the pervasive threats of global warming on the ecosystem and peaceful human coexistence. This resolution is not only premeditated but also inevitable because inactions to address the rising levels of global greenhouse gas (GHG) emissions could result in an unimaginable catastrophe for the present and future generations.¹ Consequently, nations across the globe are not hesitant to key into initiatives and treaties geared toward significantly moderating global GHG emissions.^2,3 Recently, the 2021 Conference of the Parties (COP26) addressed significant global warming issues and provided directions on actions to achieve unprecedented feats in the drive towards a sustainable environment. Specifically, COP26 established new dimensions for reaching net-zero global warming by 2050. For instance, the conference recorded commitments to cover nearly 85% of global GDP within the net-zero agreement. Additionally, approximately 153 countries bought into the idea of taking practical actions on the Nationally Determined Contributions (NDCs), which are responsible for almost 80% of the global GHG emissions, with plans to achieve a substantial reduction by 2030. To achieve the new or modified commitments, the Glasgow Climate Pact (COP26) emphasized a massive reduction in coal power, an end to deforestation, fast-tracking the transition to electric vehicles, and phasing down methane emissions. In addition, COP26 advocated for the indispensable need to reduce fossil fuel subsidies (estimated at around $5.9 trillion in 2020), contributing not less than 89% to global carbon emissions.⁴

From the theoretical and empirical viewpoints, numerous efforts have been exerted along with significant policy implications believed to help accomplish the agreements on environmental-related issues. Despite the persistent efforts, environmental challenges emanating from global warming caused by the consistent surge in carbon emissions have remained unresolved and may cause more future threats,^5.6 This unresolved issue can be associated with at least three factors. First, despite the global support for shifting from fossil fuel to renewable energy sources, the world's productive capacities remain primarily dependent on fossil fuels. One plausible argument for this continued fossil fuel consumption despite its role in increasing global GHG emissions is the lack of attention to adequate alternative energy sources, particularly nuclear energy. Second, the massive focus on renewable energy is yet to fully yield the expected results because of its limited accessibility and affordability. Hence, in the current transition stage, a clean alternative energy source is necessary. Third, nuclear energy remains the closest alternative to fossil fuels in terms of being a cheap, accessible, and affordable energy source. In addition, nuclear energy is a low carbon-intensive energy source capable of enhancing the smooth transition to clean energy consumption. Supporting this view, international energy agency (IEA)⁷ stated that the operation of nuclear reactors does not contribute to air pollution or carbon dioxide. In another study, it was suggested that nuclear energy is low-carbon that is available on a large scale for providing electricity at a marginal rate of insignificant carbon emissions equivalent to those produced by wind and solar (World Nuclear Energy Association.⁸ The fundamental role of nuclear energy in achieving the ongoing goal of carbon neutrality by 2050 is evident in its ability to reduce emissions unevenly, corresponding to a one-third decrease in transport-induced emissions globally.⁸

Based on the above narratives, it is essential to note that despite the critical need to replace fossil fuels with nuclear energy on the one hand and complement renewable energy in the drive towards the decarbonization of the global economy, on the other hand, minimal attention has been paid to it on the policy front. This can largely be attributed to the scarcity of empirical evidence supporting the adoption of nuclear energy. Consequently, there is a significant need to promote and propagate the pertinence of adopting nuclear energy from theoretical and empirical standpoints. Against this backdrop, this study primarily seeks to investigate the functional nexus between nuclear energy and carbon emissions in the top nuclear power-consuming countries. The sample countries include the top ten nuclear power consumption countries comprising the United States, China, France, Russia, South Korea, Canada, Ukraine, Germany, Spain, and Sweden. The economies mentioned above are highly ranked among the thirty-two operating nuclear power plants globally (International Atomic Energy Agency.⁹ These countries account for 84.4% of the global nuclear power plants, with the United States topping the list.⁹

The present study provides at least five novelties to the literature. First, this study is the first deliberate empirical effort to investigate the functional nexus between Nuclear Energy Consumption (NUC) and carbon emissions, specifically focusing on the top nuclear power-consuming countries. A study of this type is highly pertinent when the world is on a global mission to neutralize carbon emissions in the environment. Second, the study focuses on the top ten nuclear power-consuming economies comprising the United States, China, France, Russia, South Korea, Canada, Ukraine, Germany, Spain, and Sweden. An investigation of this type is logical because policy implications that emanate from the empirical analyses of these countries will constitute a sufficient sample valuable size for effectively utilizing nuclear energy in offsetting the environmental effects of carbon emissions. Specifically, the motivations for focusing on the top ten economies are anchored on their notable roles and contributions to the composition of global nuclear power and GHG emissions. For instance, available statistics reveal that these countries consume 84.5% of global nuclear power energy.¹⁰ Furthermore, the countries operate 71% of the global nuclear reactors, equating to 321 of the functional 448 reactors. When it comes to contributions to global GHG emissions, six of these countries (China, the United States, Russia, Germany, and Canada) are ranked among the top ten emitters. Consequently, evaluating the direction of causality between nuclear energy and carbon emissions can be best understood within the context of these countries. Third, from a methodological perspective, the employment of the quantile-on-quantile (QQ) estimator is worth lauding on four grounds. First, the QQ estimator accounts for both non-parametric and conventional quantile analyses, enhancing the provision of unbiased and consistent estimates. Second, the approach dissects the effects of nuclear energy on carbon emissions into three value points comprising the lower, middle, and upper quantiles, enabling policymakers to differentiate the environmental effects of low, moderate, and high consumption of nuclear power. Third, the NUC-carbon emissions association is estimated from both panel and time series perspectives. This enhances the recommendation of practical policies that will address environmental issues in the panel of ten countries and the individual economies. Other conventional estimators do not offer this novelty provided by the QQ estimator. Fourth, the study expands the body of knowledge in the nuclear energy-environment debates by conducting the Troster¹¹ Granger causality test, which allows for simultaneous assessment of the median as well as the tails of the distribution.

The remainder of this paper is structured as follows. Section two reviews the relevant studies, section three focuses on the methods, section four presents and explains the empirical results, and section five concludes with policy implications.

Literature review

The emerging drive towards halting the pervasive impacts of carbon emissions has attracted unprecedented attention from scholars towards promoting a sustainable environment globally. Consequently, there has been evolving support for the necessary transition from traditional energies to clean and green energy (renewable & nuclear energy). Therefore, this study reviews previous studies on the nexus between clean energy and environmental indicators.

Starting from the most recent order, Kartal¹² evaluate the effects of disintegrated fossil fuel energy (oil, gas, and oil), nuclear energy, and renewable energy on carbon emissions in five top carbon-emitting economic countries from 1965 to 2019. The study relies on multivariate adaptive regression to estimate the empirical model, and the following outcomes are evident. Findings from empirical analyses show that the three energy sources (fossil fuel, nuclear, and renewable energy) are crucial to influencing the variations in carbon emissions. Fell et al.¹³ probe the role of nuclear energy in mitigating the surging trends in global warming with a specific assessment of the 442 functional nuclear power reactors. The study reveals that an unrestrained surge in greenhouse gases exacerbates the global climate. Besides, it finds empirical support to advance that nuclear power plants contribute less to the global GHG emissions. Akram et al.¹⁴ investigate the impacts of nuclear energy, renewable energy, energy efficiency, and economic growth on carbon emissions in a panel of four countries comprising Mexico, Indonesia, Nigeria, and Turkey (MINT) based on yearly data covering 1990 to 2014 subjected to estimation within the ARDL method. The empirical outcomes show that renewable energy, nuclear energy, and energy efficiency significantly reduce carbon emissions. Rehman et al.¹⁵ assess the nexuses among renewable energy, nuclear energy, fossil fuel energy, carbon emissions, economic growth, and carbon emissions in Pakistan from 1975 to 2019. The empirical evidence relies on the ARDL estimator to explore the long-run and short-run estimates in the empirical model. The results reveal that energy consumption from fossil fuel, nuclear energy, and renewable energy significantly hinder Pakistan's real growth.

Furthermore, Ozcan and Ulucak¹⁶ have examined the extent to which nuclear energy moderates the level of environmental pollution in India by utilizing time series data stretching from 1971-to 2018. The results uncover a declining trend in carbon emissions based on the moderation of nuclear energy. A similar Indian study conducted from 1978 to 2019 by Bandyopadhyay and Rej¹⁷ shows that nuclear energy significantly mitigates carbon emissions. The empirical outcomes from a quantile egression conducted in Emerging Seven economies (E7) from 1990 to 2016 by Gyamfi et al.¹⁸ show that nuclear energy reduces carbon emissions. In contrast, Azam et al.¹⁹ probed the interlock between nuclear energy and CO₂ in top emitter economies from 2000 to 2016 by relying on panel fixed effects, random effects, and pooled OLS regression estimators. The empirical outcomes reveal that nuclear energy escalates the surge in carbon emissions. That notwithstanding, the empirical study on Pakistan's economy by Mahmood et al.²⁰ from 1993 to 2017 unveils the contributions of nuclear energy to promoting environmental quality. With a specific focus on the OECD economies from 1995 to 2015, Lau et al.²¹ utilize the dual estimators comprising the generalized method of moments (GMM) and fully modified OLS (FMOLS) to justify the submission that nuclear energy promotes environmental quality. Luqman et al.²² evaluated the nexus between nuclear energy and environmental quality in Pakistan from 1990 to 2016, and they find an asymmetric effect between NEC and environmental quality.

The debate on the choice of energy between nuclear and renewable energy in resolving the continued environmental issues motivates the research interest of Jin and Kim²³ in a panel of selected 30 economies from 1990 to 2014. The empirical evidence relies on panel cointegration technique and Granger causality analysis. Findings from the study confirm longrun association between nuclear energy, carbon emissions, and renewable energy. Besides, the longrun results show that nuclear energy is insufficient to moderate the surge in carbon emissions, whereas renewable energy significantly mediates it. However, the empirical outcomes in Dong et al.²⁴ establish the dual roles of per capita nuclear energy and per capita renewable energy in mitigating carbon emissions in China based on an analysis spanning 1993 to 2016. In contrast, the study finds per capita fossil fuels and per capita GDP as positive predictors of carbon emissions, confirming their inducing role in China's environmental problem. Baek²⁵ examines the functional effects of nuclear energy and income on carbon emissions in top twelve nuclear-generating countries using the panel cointegration approach. The empirical analyses reveal that nuclear energy substantially moderates carbon emissions. Further, the EKC hypothesis is not validated for the nuclear-producing countries due to the inverse relationship between income and carbon emissions. However, the findings in Al-Mulali²⁶ reveal that nuclear energy fails to influence the variation in carbon emissions of 30 nuclear energy-consuming countries from 1990 to 2010.

The reviewed extant studies reveal the existence of substantial nexus between nuclear energy and carbon emissions from the empirical and theoretical viewpoints. However, some notables lacunas worth mentioning are as follows. First, despite the long-standing empirical verifications of the nuclear energy-carbon emissions nexus across the global and regional economies, no study has considered the most recently ranked ten consuming economies of nuclear power. This is surprising despite the ten countries consuming over 84% of global nuclear energy. Second, few existing studies have explored the novelties of the QA regression as in this study. Third, assessing the environmental effects of nuclear energy in the top nuclear power economies is scarce in the literature, thus allowing the present research to address the cavity. Other noteworthy gaps have been elucidated in the introductory section.

Data and method

The current study explores two subsections in explicating the methodological approach to gauging the functional nexus between nuclear energy and carbon emissions. The first involves empirical modelling of the relationship followed by the econometric procedures. Figure 1 presents the research flow. First, we evaluate the stationarity attributes of series in each nation. Since both the Jarque-Bera (JB) and BDS test suggest nonlinearity of the variables, we proceed to use nonlinear approaches. This is followed by assessing the quantile cointegration (QC) since using the linear cointegration tests such as the ARDL bounds test and Johansson cointegration does not consider the variables’ nonlinearity features. In the next phase, we assess the effect of nuclear energy on CO₂ emissions in each quantile. Lastly, we used the Troster¹¹ quantile causality test, which allows simultaneous assessment of the median as well as the tails of the distribution.

Figure 1.

Flow chart on empirical procedures.

Model specification

To estimate the relationship between nuclear energy and carbon emissions in a panel of 10 top consuming countries, this study adopts a single linear panel model as thus:

C O_{2 t} = β_{0} + β_{1} N U E_{t}

(1)

Where CO₂ denotes carbon emissions (metric tonnes per capita) and NUE signifies nuclear energy (exajoules). Although GHG emissions are a composition of carbon emissions (CO₂), methane, water vapour (H₂O), ozone (O₃), and nitrous oxide (N₂O), the preponderance of empirical studies focuses more on carbon emissions because they contribute most to the volume of GHG emissions.^27–31 Besides, carbon emissions have the highest contribution among the components of GHG emissions.^32–34 Theoretically, empirical studies advanced that nuclear energy constitutes the energy source that moderates the surge in carbon emissions.^35,36 To this end, we hypothesize a negative nexus between nuclear energy (NUE) and carbon emissions (CO₂) thus

θ_{1} \frac{C O 2}{N U E} < 0

. An explicit description of the dataset is provided in Table 1 below.

Table 1.

Variable source and measurement.

Variable	Sign	Measurement	Source
Carbon Emissions	CO₂	Metric Tonnes Per Capita	BP
Nuclear Energy Consumption	NUC	Exajoules	BP

Econometric strategies

Quantile cointegration test

This study extends the frontier of knowledge in the environment literature by employing advanced quantile techniques to probe the functional effects of nuclear energy on carbon emissions in the top ten nuclear-consuming countries. Drawing from Mishra et al.,³⁷ this study utilizes the QC test credited to Xiao,³⁸ which is an extended version of the traditional Engle and Granger³⁹ test The supremacy of the QC on the traditional approach lies in its ability to disentangle the error terms of the cointegration model into pure lead-lag terms and invention terms. The model elaborating the QC is stated below.

V_{t} = φ + π^{'} Q_{t} + \sum_{j = - k}^{k} Δ z_{t - j}^{'} β_{j} + υ_{t}

(2)

and

Y_{ω}^{V} (V_{t} | I_{t}^{V} . I_{t}^{z}) = φ (ω) + π (ω)^{^{'} Q_{t}} + \sum_{j = - k}^{k} Δ z_{t - j}^{'} β_{j} + H_{u}^{- 1} (ω)

(3)

The model is further expanded with the addition of a quadratic term of the regressors as follows:

Y_{ω}^{V} (V_{t} | I_{t}^{V} . I_{t}^{z}) = φ (ω) + π (ω)^{' Q_{t}} + ϑ (ω)^{'} z_{t}^{2} + \sum_{j = - k}^{k} Δ z_{t - j}^{'} β_{j} + \sum_{j = - k}^{k} Δ z_{t - j}^{2^{'}} ϕ_{j} H_{u}^{- 1} (ω)

(4)

Equation (4) is explored to obtain the stability of the cointegration test, which is based on the null hypothesis denoted as

H_{0} = π

for all quantiles. Following the stated null hypothesis, the present study employs the Mishra et al.'s³⁷ supremum rule in ascertaining the t-statistic diagonally for all quantiles. More so, the Monte Carlo simulations are employed for obtaining the CV of the t-statistic.

Quantile-on-Quantile regression

We explore the Sim and Zhou⁴⁰ QQ approach in the current study to estimate the association between nuclear energy and carbon emissions in the sample economies. Among many other points, the Sim & Zhou⁴⁰ QQ approach can combine non-parametric, and Quantile regression (QR) approaches such that a given indicator is estimated against the quantile of another indicator. Employing the QR technique involves two steps as thus. Firstly, the conventional method is employed to obtain the average impacts of the regressors on varying quantiles of the outcome variables. Unlike the OLS estimator, the QR approach is usually employed to ascertain the impacts of the explanatory variables at the tail and centre points of the outcome variable. Secondly, Cleveland⁴¹ and Stone⁴² local linear regression (LLR) is employed to evaluate the spatial effect of each of the regressors on the outcome variable. The LLR accounts for the “curse of dimensionality” issue of strictly non-parametric techniques. Combining these two phases enhances the employment of the nexus between outcome and explanatory variables, which provides robust outcomes. Summarily, the QQ technique is perceived to be high in examining the effect of the X quantiles on Y quantiles. Following Equation (1), the QQ technique is rooted in the QR model:

Y_{t} = β^{ϑ} (X_{t}) + μ_{t}^{θ}

(5)

where

Y_{t}

and

X_{t}

signify the outcome indicator in period t and the regressors parameter in period t.

θ

represents

θ t h

quantile of the conditional distribution of Y and is disturbance term in quantile whose θth conditional quantile is 0. The anonymous function is exemplified by

β^{ϑ} (.)

Suggesting that prior details of X and Y are unknown.

More so, the QR model does not support the assumption that X shocks are likely to stimulate the nexus between Y and X. drawing from this, the outcomes of substantial positive X shocks, for example, can change from those of small positive X shocks. Likewise, positive and negative X volatility may stimulate asymmetric variation in Y.

Therefore, to examine the relationship between $θ t h$ quantile of Y and $τ t h$ of X, signified by $X^{τ}$ , the LLR is adopted to examine $X^{τ}$ region. The unknown function represented by $β^{ϑ} (.)$ can be calculated as Taylor expansion first-order positioned on $X^{τ}$ quantile in Equation (6);

β^{θ} (X_{t}) \approx β^{θ} (X^{t}) + β^{θ ι} (X^{t}) (X^{t} - X^{t})

(6)

Where

β^{θ} (X_{t})

partial derivative concerning X is illustrated by

β^{θ ι}

represented as reaction and is analogous to the slope coefficient in an LLR model, the main characteristics of Equation (2) are that

β^{θ} (X^{τ}) and β^{θ ι} (X^{τ})

parameters are indexed twice in

τ and θ

. Assumed that

β^{θ} (X^{τ}) and β^{θ ι} (X^{τ})

are functions of

θ

and

X^{t}

and

X^{t}

is a function of

τ,

it is evident that

β^{θ ι} (X^{τ}) a n d β^{θ} (X^{τ})

are both functions of

τ and θ,

respectively. Also,

β^{θ} (X^{τ}) and β^{θ ι} (X^{τ})

can be denoted as

β_{0} (θ, τ)

and

β_{1} (θ, τ)

. Consequently, Equation (2) can be changed into Equation (7):

β^{θ} (X_{t}) \approx β_{0} (θ, τ) + β_{1} (θ, τ) (X_{t} - X^{τ})

(7)

When Equation (3) is substituted into Equation (1), Equation (4) is obtained as follows:

Y_{t} = \underset{(*)}{\underset{⏟}{β_{0} (θ, τ) + β_{1} (θ, τ) (X_{t} - X^{τ}) + μ_{t}^{θ}}}

(8)

In Equation (4),

θ t h

conditional quantile Y is depicted by (*). Equation (6) estimates that the overall dependence nexus between X and Y is built on the connection between their different distributions. Besides,

X_{t} a n d X^{τ}

is interchanged by

\hat{X_{t}} a n d \hat{^{}}

Xτ when assessing Equation (4). The parameters

b_{0} a n d b_{1}

, which are

β_{0} a n d β_{1}

estimates, are assembled by solving the minimization issue as follows:

min_{b_{0}, b_{1}} \sum_{i = 1}^{n} ρ θ [Y_{t} - b_{0} - b_{1} (\hat{X_{t}} - \hat{X^{τ}})] K (\frac{F_{n} (\hat{X_{t}}) ⊓ - τ}{h})

(9)

Where ρθ demonstrates the function of the quantile loss

_{} (u)

, and

ρ_{θ} (u) = u (θ - I (u < 0)), I

is the normal function of the variable. The kernel function and h, which is the bandwidth of the kernel parameter, are portrayed by

K (.)

The Gaussian kernel is applied to calculate the observations in the

X^{τ}

neighbourhood, being a notable kernel function in economics and finance research as a result of being accessible, computable, and reliable.

The Gaussian kernel is usually symmetric around. More so, it provides weight lower observations that are further out. These weights are in reverse proportionate to the distance between the function of the analytical distribution of $\hat{X_{t}}$ , shown by $F_{n} (\hat{X_{t}}) = \frac{1}{n} \sum_{k = 1}^{n} I ({\hat{X}}_{k} > {\hat{X}}_{t})$ , and the distribution function value that corresponds with the $X^{τ} i s$ illustrated by $τ$ . It is pertinent to note that the bandwidth is fundamental when exploring the non-parametric methods. The bandwidth influences the horizontalness of the analogous estimate because of the neighbourhood scale specified within the target point. Consequently, a broader bandwidth implies a greater risk of estimation distortion, while a lower smaller bandwidth suggests otherwise. Hence, it is pertinent to choose a bandwidth to provide a balance between variance painstakingly. Drawing from Sim and Zhou⁴⁰, a bandwidth h = 0.05 parameter is assumed appropriate for the empirical analysis.

Granger causality in quantiles

One of the main novelties of the present study is the application of the novel Granger causality in quantiles propounded by Troster.¹¹ Drawing from the conventional Granger causality, the case of no causality is determined when Xi does not predict Zi. Hypothesizing that the vector $(M_{i} = M_{i}^{Z}, M_{i}^{X})$ ∈ $ℜ^{e}, e = o + q,$ with $M_{i}^{X}$ is the initial evidence set of $X_{i}$ $M_{i}^{X} = (X_{i - 1,} \dots \dots X_{i - q,})^{^{'}}$ $\in ℜ^{q}$ . Moreover, $H_{o}$ hypothesis is depicted as follows:

H_{o}^{X 0 =\to Z} : F Z (Z | M_{i}^{Z}, M_{i}^{x}) = F Z (Z | M_{i}^{Z}) forall X ℜ,

(10)

F Z (. | M_{i}^{Z}, M_{i}^{x})

represents the conditional sprinkling function of

Z i

so far.

M_{i}^{x}, M_{i}^{z}

in the ambit of null hypothesis is explained by Equation (9). Based on Troster’s¹¹ identification of the Dt test, which categorizes the structure of QA (·) for all π∈Γ⊂[0,1], following Granger causality null hypothesis. The model may be outlined thus:

Q A R (1) : m_{1} (M_{i}^{Z} | \partial (π)) = λ 1 (π) + λ 2 (π) Z_{i - 1} + μ t ψ_{X}^{- 1} (π)

(11)

Given that the values

\partial (π) = λ 1 (π), λ 2 (π) and μ t

are re-estimated by the probability of supreme in quantiles grid space that is equal, and

ψ_{X}^{- 1} (.)

Constitutes the reversal of a normal orthodox dispersal function. By estimating the quantile auto-covariance function (QAF) model in Equation (10) with the lagged parameter to another parameter, appropriate adjustment of the causality sign between the indicators can be obtained. Consequently, Equation (12) presents the QAR (1):

Q_{π}^{Z} = (Z_{i} | M_{i}^{Z}, M_{i}^{x}) = λ 1 (π) + λ 2 (π) Z_{i - 1} + η (π) X_{i - 1} + μ t ψ_{X}^{- 1}

(12)

Findings and discussion

Pre-Estimation outcomes

The current empirical analysis commenced by presenting brief information on each nation's variables (nuclear energy and CO₂ emission). Table 2 shows descriptive statistics of the data. Based on the results presented, Ukraine contributes the most to CO₂ emissions per capita, with 5.09% among the top nuclear power-consuming countries. Next to Ukraine are the USA (1.57%), Canada (1.44%), Russia (1.09%), and Germany (1.02%). The least contributor to CO₂ emissions per capita is China with −0.01%, followed by Sweden (0.38%), Spain (0.46%), France (0.48%), and South Korea (0.77%). Regarding the share of nuclear power, the USA ranks highest with 6.19%, followed by France (5.55%) and China (4.83). The least contributor is Spain with 3.61%. Normality test is conducted using values provided by Skewness, Kurtosis, and Jarque-Bera. The outcomes reveal the series are not normally distributed, suggesting that the variables of investigation in each country, using a linear approach, will produce misleading results. Therefore, the current study applied nonlinear methods (non-parametric causality and QQ regression approaches) to capture the effect of nuclear energy on CO₂ emissions in the leading countries by nuclear power consumption. Furthermore, the stationarity properties of the variables are checked using the ADF and PP unit root tests. The PP and ADF outcomes in Table 3 show that all the variables (nuclear energy and CO₂ emission) in each nation are I(1) variables. The BDS outcomes are presented in Table 3. The outcomes of the BDS reveal the nonlinear attributes of the variables for each nation. This is reflected in the significant level of each variable across the ten selected countries. Consequently, we can models leading to the investigation of the relationship between carbon emissions and nuclear power consumption are nonlinear. This submission is supported by the outcomes presented in Table 2 based on the values of the normality tests confirmed by the Jarque-Bera findings. Furthermore, we assess the cointegration between nuclear energy and CO₂ emission in each nation utilizing the QC test proposed by Xiao.³⁸ Table 4 reports the cointegration outcomes. This table shows the β and γ coefficients and supremum norm values, as well as their critical values (CV1, CV5, and CV10), with levels of significance of 1%, 5%, and 10%, respectively. The research results demonstrate that the coefficients β and γ are above all CV at the 1% significance level, indicating that nuclear energy and CO₂ emissions have a long-term cointegration. These findings suggest that nuclear energy and CO₂ emissions have a nonlinear connection for all the selected nations.

Table 2.

Descriptive statistics and unit root outcomes.

	Descriptive Statistics									Unit Root
	Mean	Median	Maximum	Minimum	SD	Skewness	Kurtosis	JB	Probability	$Δ$ ADF	$Δ$ PP
	Carbon Emissions (CO₂)
Canada	1.4398	1.4297	1.5447	1.3300	0.0572	0.2264	1.9160	8.0511	0.0179	−5.288*	−6.127*
China	−0.0980	−0.3266	0.5800	−0.7992	0.4975	0.1794	1.3847	15.971	0.0003	−7.741*	−6.985*
France	0.4809	0.5376	0.6321	0.2115	0.1212	−0.9178	2.3982	21.769	0.0000	−6.145*	−7.657*
Germany	1.0155	1.0129	1.2159	0.7048	0.1175	−0.0263	2.5089	1.4231	0.4909	−7.405*	−7.743*
Russia	1.0928	1.0310	1.4579	0.9020	0.1705	1.1240	2.7636	29.8031	0.0000	−4.747*	−4.740*
South Korea	0.7707	0.8614	1.1350	0.0877	0.3103	−0.8459	2.5896	17.677	0.0001	−6.173*	−7.243*
Spain	0.4641	0.4292	0.7441	0.1866	0.1577	0.1831	2.0107	6.4909	0.0390	−4.723*	−4.734*
Sweden	0.3815	0.4482	0.6254	0.0447	0.1747	−0.6747	2.1296	15.040	0.0005	−7.325*	−7.382*
Ukraine	3.9479	3.9730	4.1210	3.3720	0.1527	−2.2271	8.1392	269.80	0.0000	−3.658**	−3.598**
USA	1.5693	1.6075	1.6740	1.3645	0.0885	−0.7246	2.1034	16.941	0.0002	−6.618*	−6.745*
	Nuclear Energy Consumption (NUC)
Canada	4.0643	4.0905	4.3181	3.6509	0.1196	−0.5733	3.3716	8.4136	0.0149	−5.304*	−5.304*
China	4.8334	4.987	6.7609	1.4941	1.2021	−0.5573	3.3622	1.5456	0.4617	6.090*	7.067*
France	5.5456	5.6026	5.7171	4.9765	0.1641	−1.4496	4.5162	61.9971	0.0000	−6.549*	−10.22*
Germany	4.5004	4.6526	4.7784	3.8315	0.2993	−1.0964	2.7279	28.2788	0.0000	−7.727*	−8.237*
Russia	4.5576	4.5798	4.8728	4.1877	0.1928	−0.2021	1.9433	7.4134	0.0246	−4.954*	−4.977*
South Korea	4.1333	4.4098	4.6498	2.1437	0.5518	−1.2511	4.0754	42.9621	0.0000	−5.110*	−5.123*
Spain	3.6141	3.6479	3.7842	2.7833	0.1548	−2.7915	12.6649	721.5263	0.0000	−8.411*	−8.411*
Sweden	3.7978	3.8099	3.9886	3.5139	0.1133	−0.4266	2.4917	5.7122	0.0575	−8.367*	−9.447*
Ukraine	3.9478	3.9734	4.1210	3.3720	0.1532	−2.2170	8.0754	263.0611	0.0000	−6.154*	−7.854*
USA	6.1887	6.2710	6.3360	5.6128	0.1788	−1.6974	5.1194	92.7602	0.0000	−5.170*	−5.780*

Note: *, ** and *** represents 1%, 5% and 10% level of significance.

Table 3.

BDS test outcomes.

	China	Canada	France	Germany	Ukraine	Russia	Sweden	USA	Spain	South Korea
	Nuclear Energy (NUE)
M2	31.974*	44.641*	32.703*	37.729*	36.961*	23.571*	40.482*	24.602*	42.563*	44.547*
M3	33.086*	46.808*	34.357*	39.774*	38.723*	25.223*	42.702*	26.212*	45.025*	47.430*
M4	34.572*	49.833*	36.713*	42.614*	41.138*	27.248*	45.692*	28.157*	48.177*	51.052*
M5	36.779*	54.397*	40.404*	47.074*	44.876*	30.140*	50.284*	30.992*	52.799*	56.383*
M6	40.015*	60.630*	45.587*	53.495*	50.406*	34.080*	56.742*	34.916*	59.236*	63.711*
	Carbon Emissions (CO₂)
M2	18.819*	31.974*	26.462*	31.572*	27.261*	53.045*	37.727*	20.489*	19.655*	48.542*
M3	19.569*	33.080*	28.125*	33.211*	29.169*	55.887*	38.271*	21.397*	20.414*	51.726*
M4	20.702*	34.543*	30.214*	35.513*	31.518*	59.714*	39.062*	22.510*	21.319*	56.169*
M5	22.568*	36.775*	33.357*	39.116*	34.911*	65.527*	40.797*	24.222*	22.736*	62.476*
M6	25.234*	40.016*	37.580*	44.259*	39.540*	73.762*	43.587*	26.717*	25.092*	70.987*

Note: 1% level of significance is denoted by *.

Table 4.

Quantile cointegration test outcomes.

Model	Coefficient	$S u p_{τ} \| V_{π} (τ) \|$	CV1%	CV5%	CV10%
Canada $C O_{2 t} Vs NU C_{t}$	Β	6191	4976	3234	2013
Canada $C O_{2 t} Vs NU C_{t}$	$γ$	605	437	305	194
China $C O_{2 t} Vs NU C_{t}$	Β	4592	3171	1936	930
China $C O_{2 t} Vs NU C_{t}$	$γ$	527	411	297	125
France $C O_{2 t} Vs NU C_{t}$	Β	5698	4213	2637	1795
France $C O_{2 t} Vs NU C_{t}$	$γ$	559	422	287	133
Germany $C O_{2 t} Vs NU C_{t}$	Β	7551	4718	3364	2340
Germany $C O_{2 t} Vs NU C_{t}$	$γ$	743	529	393	267
Russia $C O_{2 t} Vs NU C_{t}$	Β	1539	1220	932	651
Russia $C O_{2 t} Vs NU C_{t}$	$γ$	146	97	54	32
South Korea $C O_{2 t} Vs NU C_{t}$	Β	3527	2722	1302	857
South Korea $C O_{2 t} Vs NU C_{t}$	$γ$	352	223	109	75
Spain $C O_{2 t} Vs NU C_{t}$	Β	8094	6229	4302	2868
Spain $C O_{2 t} Vs NU C_{t}$	$γ$	858	624	429	318
Sweden $C O_{2 t} Vs NU C_{t}$	Β	5117	3026	1728	894
Sweden $C O_{2 t} Vs NU C_{t}$	$γ$	495	328	212	125
Ukraine $C O_{2 t} Vs NU C_{t}$	Β	2431	1326	867	452
Ukraine $C O_{2 t} Vs NU C_{t}$	$γ$	298	186	119	76
USA $C O_{2 t} Vs NU C_{t}$	Β	6595	4361	2920	1823
USA $C O_{2 t} Vs NU C_{t}$	$γ$	564	407	284	147

Quantile-on-Quantile outcomes

The influence of nuclear energy on CO₂ emissions in the leading countries by nuclear power consumption is investigated using the QQR approach. In Figure 2, the state-wise 3D graphs are demonstrated using the QQR slope estimates $β_{1} (θ, τ)$ . The first column shows the influence of nuclear energy (x-axis) on CO₂ emissions (y-axis). The empirical results of the QQR estimate confirm that the relationship between CO₂ and nuclear energy varies in each country. The slope estimates $β_{1} (θ, τ)$ may differ with $τ$ depending on $θ$ .

Figure 2.

Effect of nuclear energy on CO₂ emissions. a. USA. b. China. c. France. d. Russia. e. South Korea. f. Canada. g. Ukraine. h. Germany. i. Spain. j. Sweden.

Figure 2(a) presents the effect of nuclear energy (NUE) on the CO₂ emissions level in the USA. The outcomes disclosed that the impact of NUE on CO₂ emissions varies in each quantile. Furthermore, all NUE quantiles negatively affect the lower tails (0.1–0.25) of CO₂ emissions (0.1–0.95). This implies that NUC remains an effective medium of moderating the adverse effects of CO₂ on the environment at a low level of carbon emissions. Nevertheless, the middle tails (0.35–0.60) of CO₂ are impacted positively by the middle tails (.35–0.60) of NUE, suggesting that at equilibrium, there is a co-movement in the volume of carbon emissions and nuclear energy. In addition, the higher tails (0.65–0.90) of CO₂ emissions are negatively affected by the lower and middle tails (0.-0.70) of NUE. In summary, CO₂ is impacted negatively by NUE in the majority of the quantiles. Therefore, we can infer that, on average, NUE curbs CO₂ emissions in the USA. For China, the effect of nuclear energy (NUE) on CO₂ emissions is illustrated in Figure 2(b). Although the impact of NUE on CO₂ emissions in China is positive, the outcomes vary in each quantile. For instance, in the lower and higher tails (0.1–0.35) and (0.65–0.95) of CO₂ and all quantiles (0.1–0.95) of NUE, we observed a weak and positive effect of NUE on CO₂. However, a positive and robust impact of NUE of CO₂ is noticed in the middle tails (0.40–0.55) of CO₂ and all NUE tails (0.1–0.95). In summary, CO₂ is impacted positively by NUE in most quantiles. Therefore, NUE contributes to a surge in China's CO₂ emissions.

For France, the effect of nuclear energy (NUE) on CO₂ emissions is illustrated in Figure 2(c). We observed a negative and weak effect of NUE on the emissions of CO₂ in the lower and middle tails (0.1–0.75) of CO₂ and each quantile (0.1–0.95) of NUE. However, the higher tails (0.70–0.95) of CO₂ emissions are positively impacted by the lower tails (0.1–0.35) of NUE. In summary, the CO₂ is impacted negatively by NUE in most quantiles. Therefore, NUE enhances the quality of the environment in France. Figure 2(d) presents the effect of nuclear energy (NUE) on Russia's CO₂ emissions level. The outcomes disclosed that the impact of NUE on CO₂ emissions varies in each quantile. For example, the lower tails (0.1–0.20) of CO₂ emissions are positively affected by the middle seats (0.1–0.65) of NUE. Nevertheless, the lower, middle and higher tails (0.10–0.95) of CO₂ are impacted negatively by each NUE tail (0.1–0.95). In summary, the CO₂ is impacted negatively by NUE in the majority of the quantiles in Russia. Therefore, NUE curbs Russia's CO₂ emissions.

For South Korea, the effect of nuclear energy (NUE) on the CO₂ emissions level is illustrated in Figure 2(e). We observed a negative and weak effect of NUE on the emissions of CO₂ in each quantile (0.1–0.95) of NUE and CO₂ emissions, respectively. Therefore, we established the emissions-reducing effect of NUE in South Korea. Figure 2(f) presents the impact of nuclear energy (NUE) on Canada's CO₂ emissions. The outcomes disclosed that the impact of NUE on CO₂ emissions varies in each quantile. Furthermore, the lower tails (0.1–0.35) of CO₂ emissions are negatively affected by all NUE quantiles (0.1–0.95). Nevertheless, the middle tails (0.35–0.50) of CO₂ are impacted negatively by all the tails (0.1–0.95) of NUE. Furthermore, the higher tails (70–95) of CO₂ emissions are influenced negatively by the higher tails (0.65–0.95) of NUE. Therefore, NUE plays a crucial role in alleviating the damaging effect of CO₂ emissions.

Figure 2(g) presents the effect of nuclear energy (NUE) on Ukraine's CO₂ emissions level. The outcomes disclosed that the impact of NUE on CO₂ emissions varies in each quantile. In most of the quantiles, the negative effect of NUE on CO₂ emissions dominates the positive effect of NUE on CO₂ emissions. In summary, the CO₂ is impacted negatively by NUE in the majority of the quantiles in Ukraine. These outcomes supported the emissions-reducing effect of NUE in Ukraine. Figure 2(h) presents the impact of nuclear energy (NUE) on Germany's CO₂ emissions level. In addition, the lower tails (0.1–0.25) of CO₂ are impacted negatively by the middle and higher tails (0.40–0.85) of NUE. Furthermore, the middle and upper tails (0.30–0.85) of CO₂ are negatively impacted by NUE in each tail (0.1–0.95). In summary, the CO₂ is negatively affected by NUE in Germany in the majority of the quantiles. These outcomes support the emissions-reducing effect of NUE in Germany.

Figure 2(i) presents the effect of nuclear energy (NUE) on Spain's CO₂ emissions level. The outcomes revealed that the impact of NUE on CO₂ emissions varies in each quantile. In the majority of the quantiles, the positive effect of NUE on CO₂ emissions is pronounced in each quantile (0.1–0.95). In summary, the CO₂ is affected positively by NUE in all quantiles of NUE (0.1–0.90) and CO₂ (0.1–0.95) in Ukraine. These outcomes disclosed the increasing emissions effect of NUE in Spain. Figure 2(j) displays the impact of nuclear energy (NUE) on Sweden's CO₂ emissions. The results reveal that the impact of NUE on CO₂ emissions varies in each quantile in Sweden. In most of the quantiles, the negative effect of NUE on CO₂ emissions is more pronounced than the positive effect of NUE on CO₂ emissions. These outcomes support the emissions-reducing effect of NUE in Sweden.

Granger quantile causality outcomes

Equation (11) is used in the present paper to apply Granger causality in quantiles. Table 5 shows the outcomes of Granger causality in quantiles, including the DT test value of significance for log series. We use the DT test to an eleven-quantile equivalent grid, i.e. (0.05–0.95).

Table 5.

Troster ¹¹ quantile causality test outcomes.

		0.5–0.95	0.05	0.10	0.20	0.30	0.40	0.500	0.60	0.70	0.80	0.90	0.95
Canada	CO₂→NUC	0.010	0.010	0.010	0.020	0.010	0.363	0.667	0.069	0.010	0.010	0.039	0.029
	NUC→CO₂	0.010	0.510	0.010	0.010	0.010	0.010	0.010	0.235	0.020	0.010	0.500	0.500
China	CO₂→NUC	0.667	0.353	0.353	0.353	0.657	0.657	0.676	0.676	0.676	0.676	0.676	0.676
	NUC→CO₂	0.412	0.069	0.392	0.392	0.637	0.667	0.676	0.324	0.676	0.676	0.676	0.676
France	CO₂→NUC	0.010	0.020	0.010	0.010	0.010	0.010	0.402	0.059	0.010	0.010	0.137	0.500
	NUC→CO₂	0.010	0.010	0.010	0.520	0.520	0.039	0.833	0.167	0.010	0.059	0.627	0.657
Germany	CO₂→NUC	0.010	0.422	0.049	0.059	0.010	0.010	0.147	0.098	0.010	0.010	0.010	0.588
	NUC→CO₂	0.010	0.137	0.010	0.010	0.010	0.010	0.098	0.225	0.010	0.010	0.010	0.049
Russia	CO₂→NUC	0.010	0.137	0.010	0.010	0.010	0.010	0.010	0.255	0.010	0.010	0.088	0.569
	NUC→CO₂	0.010	0.049	0.010	0.010	0.010	0.010	0.892	0.010	0.010	0.010	0.039	0.873
South Korea	CO₂→NUC	0.010	0.010	0.010	0.010	0.010	0.010	0.059	0.010	0.010	0.010	0.010	0.059
	NUC→CO₂	0.010	0.059	0.078	0.010	0.010	0.147	0.069	0.049	0.059	0.010	0.206	0.147
Spain	CO₂→NUC	0.010	1.000	0.431	0.010	0.059	0.196	0.814	0.010	0.010	0.039	0.382	0.127
	NUC→CO₂	0.010	1.000	0.049	0.010	0.010	0.010	0.078	0.029	0.010	0.010	0.010	0.196
Sweden	CO₂→NUC	0.010	0.520	0.010	0.010	0.010	0.029	0.853	0.010	0.010	0.010	0.588	0.010
	NUC→CO₂	0.010	0.598	0.127	0.010	0.010	0.010	0.500	0.010	0.010	0.010	0.010	0.275
Ukraine	CO₂→NUC	0.010	0.206	0.010	0.010	0.010	0.010	0.108	0.010	0.010	0.010	0.010	0.010
	NUC→CO₂	0.010	0.657	0.010	0.010	0.010	0.010	0.020	0.010	0.010	0.010	0.833	0.422
USA	CO₂→NUC	0.010	0.775	0.314	0.118	0.029	0.010	0.078	0.010	0.010	0.010	0.010	0.304
	NUC→CO₂	0.010	0.020	0.010	0.010	0.010	0.059	0.010	0.010	0.010	0.010	0.020	0.490

Note: Table 5 unveils the sub-sampling pvalues of the DT test The log-difference of NUC in a quarter is depicted by ΔNUCt, and the log difference of CO₂ is illustrated by ΔNUCt. The dismissal of the null hypothesis at the 5% significance level is illustrated by thick bold numbers.

For Canada, Table 5 outcomes revealed that nuclear energy fluctuation Granger-cause CO₂ emissions at a significance level of 5% in most of the distribution. On the flip side, at a significance level of 5%, the outcomes disclosed that CO₂ emissions Granger cause nuclear energy. In summary, the findings show a feedback causal interconnectedness between nuclear energy and CO₂ emissions in all quantile distributions. For China, Table 5 outcomes revealed that nuclear energy fluctuation does not Granger-cause CO₂ emissions at a significance level of 5% in most of the distribution. On the flip side, at a significance level of 5%, the outcomes disclosed that CO₂ emissions do not Granger cause nuclear energy. In summary, the findings show no evidence of causal interconnectedness between nuclear energy and CO₂ emissions in all quantile distributions.

For France, nuclear energy fluctuation Granger-cause CO₂ emissions at a significant level of 5% in most of the distribution. On the other side, at a significance level of 5%, the outcomes disclosed that CO₂ emissions Granger cause nuclear energy. In summary, the findings show a feedback causal interconnectedness between nuclear energy and CO₂ emissions in all quantile distributions. Therefore both nuclear energy and CO₂ emissions can predict each other. Furthermore, we found feedback causality between nuclear energy and CO₂ emissions in most quantiles in Germany. In summary, the outcomes reveal a two-way causal interconnectedness between nuclear energy and CO₂ emissions in all quantile distributions.

Likewise, for Russia, a feedback causality between nuclear energy and CO₂ emissions surfaced in most of the quantiles, implying that both nuclear energy and CO₂ can predict each other in most of the quantiles at 5% significance. For South Korea, bidirectional causality between CO₂ emissions and nuclear energy is evident in most of the quantiles at 5% significance. For Spain, fluctuation in nuclear energy Granger-cause CO₂ emissions at a significant level of 5% in most of the distribution. On the other side, at a significance level of 5%, the outcomes disclosed that CO₂ emissions Granger cause nuclear energy. In summary, the findings show a feedback causal interconnectedness between nuclear energy and CO₂ emissions in all quantile distributions.

Moreover, at 5% significance, bidirectional causality between nuclear energy and CO₂ emissions is evident for Sweden in most quantiles distribution. This demonstrates that both nuclear energy and CO₂ can predict each other. Furthermore, bidirectional causality between nuclear energy and CO₂ emissions is seen in Ukraine in the majority of the quantiles, which implies that both nuclear energy and CO₂ can predict each other in the majority of the quantiles at 5% significance. Lastly, for the USA, a feedback causality is observed between nuclear energy and CO₂ emissions in the majority of the quantiles, which implies that both nuclear energy and CO₂ can predict each other in the majority of the quantiles at 5% level of significance.

Discussion of results

This section of the study presents a summary of the findings obtained above. In most of the quantiles, the present research outcomes revealed that nuclear energy impacts CO₂ emissions negatively in the USA, France, Russia, South Korea, Canada, Ukraine, Germany, and Sweden. By implication, we can infer that the impacts of nuclear energy on carbon emissions are not homogeneously determined but rather heterogeneous across the different phases of quantiles and countries. This demonstrates that using nuclear energy contributes to environmental quality in the USA, France, Russia, South Korea, Canada, Ukraine, Germany, and Sweden. These findings are unsurprising, considering that nuclear energy results in a significant reduction in CO₂ emissions. As a result, it may be a viable substitute for traditional energy sources for improving the quality of the environment.¹⁶ Consequently, transitioning to clean energy sources such as nuclear power for electricity generation is also critical for lowering GHG emissions. Furthermore, nuclear energy has the potential to help designated countries upgrade their energy sectors. Nuclear energy will minimize the reliance on imported energy and fossil fuel use. As a result, CO₂ emissions can be eliminated using this strategy.

Consequently, the empirical results of this study support the notion that nuclear energy is required to prevent the harmful effects of CO₂ emissions that cause climate change and global warming and that it can be a better alternative source of energy in the USA, France, Russia, South Korea, Canada, Ukraine, Germany, and Sweden because nuclear energy growth and use are both environmentally and economically advantageous. This suggests that nuclear energy should have a greater quantitative share in these countries’ energy mix. The study outcomes comply with the studies of Azam et al.³⁵ and Danish et al.³⁶, Bandyopadhyay and Rej,¹⁷ and Nathaniel et al.⁴³ When the energy sector transitions from fossil fuels to nuclear power, the outcomes undoubtedly verify the advantage that nuclear plant operation does not cause CO₂ emissions, which might aid in reducing GHG emissions. Furthermore, while nuclear energy decreases carbon emissions, nuclear power facilities always pose significant hazards due to broad cross-country economic, political, and social changes. This can be found in the cases of China and Spain, as revealed by the study outcomes indicating that NUC contributes to the degradation of the environment. This outcome is in line with the studies of Dong et al.²⁴ and Sarkodie and Adams⁴⁴ for South Africa. This implies that nuclear power contributes to contamination for various reasons, including management practices and inappropriate nuclear waste disposal. Similarly, nuclear energy's role in inducing carbon emissions is evident in China and Spain, which is consistent with Anser et al.⁴⁵

While a significant relationship between two or more interacting variables does not imply causal effects, the outcomes from the Granger causality suggest the existence of feedback effects between nuclear energy and carbon emissions for most countries. In contrast, bidirectional and no causal impacts are evident in a few economies. Conclusively, we can claim that policy measures implemented to promote nuclear energy would significantly reduce the level of carbon emissions. In contrast, policy measures implemented toward reducing carbon emissions could enhance the promotion of nuclear energy, which is believed to be environmentally friendly. Moreover, the results indicate that besides renewable energy, the path to zero emissions could be achieved through conscientious efforts geared towards promoting NUC.

Conclusion and policy insights

Conclusion

Nuclear energy has sparked international attention as one of the essential strategies for reducing emissions, thanks to its ability to provide low-carbon power.¹⁷ Since 1990, nuclear power plants have prevented 1.5–2 billion tonnes of GHG emissions yearly (IEA, 2020). Furthermore, the IEA (2020) predicted that total energy-related and electricity generation emissions would have been approximately 6% and 20% higher without nuclear power between 1971 and 2018, respectively.¹⁶ Based on these interesting facts about the role of nuclear energy in reaching the carbon neutrality target, the current research explores the effect of nuclear energy on CO₂ emissions in the leading countries by nuclear power consumption using a quarterly dataset from 1990 to 2019.

We applied the QA estimator, which is worth lauding on four grounds. First, the QQ estimator accounts for both non-parametric and conventional quantile analyses, enhancing the provision of unbiased and consistent estimates. Second, the approach dissects the effects of nuclear energy on carbon emissions into three value points comprising the lower, middle, and upper quantiles, enabling policymakers to differentiate the environmental effects of low, moderate, and high consumption of nuclear power. Third, the NUC-carbon emissions association is estimated from both panel and time series perspectives. This facilitates the recommendation of practical policies that will address environmental issues in the panel of ten countries and the individual economies. This novelty provided by the QQ estimator is not available in other conventional estimators. Fourth, the study extended the frontier of knowledge in the nuclear energy-environment debates by conducting the Troster¹¹ Granger causality test, which allows for instantaneous assessment of the median as well as the tails of the distribution. The outcomes from the QQ estimator revealed that in the majority of the quantiles, nuclear energy contributes to a decrease in the degradation of the environment in the USA, France, Russia, South Korea, Canada, Ukraine, Germany, and Sweden, while in Spain and China, NUC contributes to the deterioration of the environment. Moreover, the outcomes of the causality test disclosed that nuclear energy and CO₂ emissions could predict each other in most of the quantiles.

Policy insights

The findings above may have profound ramifications for policymakers planning nuclear energy and CO₂-emission policies in the leading countries by nuclear power consumption. Policy analysts should consider the influence of nuclear energy on carbon emissions. It is apparent that nuclear energy plays a positive role in reducing emissions. Therefore, policymakers in the selected nations should focus on generating as much power as possible from nuclear energy sources as it can enhance the quality of the environment. Nuclear energy is a greener energy source that can assist with meeting growing energy needs while reducing reliance on imported energy. Furthermore, nuclear energy has the potential to aid in the achievement of SDGs and the creation of enhanced ecological strategies. Moreover, the leading nations in terms of NUC require additional reforms and investment in nuclear energy. Nuclear energy-associated technologies will undoubtedly preserve the nations’ prestige while promoting economic progress and environmental and social betterment. Nuclear power generation provides lower prices, ensuring energy security and reducing traditional energy production emissions. Furthermore, these nations and other nations should promote foreign and local investments in nuclear energy supply while taking safety and security precautions into account. Furthermore, nuclear power-generated electricity will assist in minimizing the reliance on imported energy. Policymakers in these nations should emphasize globalization since it amplifies nuclear energy's positive influences in reducing carbon emissions. They can boost the proportion of nuclear energy mix due to globalization and enhance foreign direct investment, international trade, knowledge transfer, and ecological consciousness.

Limitations of the study

This research has certain drawbacks. Firstly, the nuclear energy-emissions nexus was investigated for the leading countries by nuclear power consumption. This opens up the possibility for future research on other emerging and developed economies, using both time series and panel data. Secondly, this study uses CO₂ emissions as an indicator of environmental deterioration. As a result, other environmental indicators such as ecological footprint and load capacity factors should be considered in future studies. Lastly, bivariate analytical approaches were used to assess the effect of nuclear energy on CO₂ emissions, which may be limited due to the magnitude of the issue being tackled.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Data availability

The data for the analysis in this study is available upon request

ORCID iDs

Mamdouh Abdulaziz Saleh Al-Faryan

Ridwan Lanre Ibrahim

Bohuang Pan is a researcher from Macau University and Technology, the institute for Social and Cultural Research, Macau, China. Bohuang specializes in development, social, and environment related studies that add significant contributions to advancement in scientific research.

Tomiwa Sunday Adebayo is a prolific writer and professional researcher from Cyprus International University. Adebayo research interest covers environment, energy, development, and econometric modeling.

Ridwan Lanre Ibrahim is a potential research and excellent writer from University of Lagos while research interest focuses on energy economics, environment, international trade, and climate change conferences (COP26).

Mamdouh Abdulaziz Saleh Al-Faryan is an international researcher with special interest in corporate governance and finance, development, environment, and finance.

Nomenclature

References

Adebayo

. Renewable energy consumption and environmental sustainability in Canada: does political stability make a difference? Environ Sci Pollut Res 2022a; 4: 1–16.

Wang

Guo

. Official development assistance and carbon emissions of recipient countries: a dynamic panel threshold analysis for low-and lower-middle-income countries. Sustain Prod Consum 2022; 29: 158–170.

Umar

Kirikkaleli

, et al. Carbon neutrality target in Turkey: measuring the impact of technological innovation and structural change. Gondwana Res 2022; 4: 33–45.

Daley

. The Fossil Fuelled 5: comparing rhetoric with reality on fossil fuels and climate change. Available at https://apo.org.au/node/315102. Accessed on March 10, 2022. 2021.

Stoddard

Anderson

Capstick

, et al.

Three decades of climate mitigation: why haven’t we bent the global emissions curve?

Annu Rev Environ Resour 2021; 46: 653–689.

Awosusi

Adebayo

Kirikkaleli

, et al. Role of technological innovation and globalization in BRICS economies: policy towards environmental sustainability. Int J of Sustain Dev World Ecol 2022a: 1–18.

International Energy Agency. Nuclear energy explained: nuclear power and the environment. 2021. Available at https://www.eia.gov/energyexplained/nuclear/nuclear-power-and-the-environment.php#:∼:text=Nuclear%20power%20reactors%20do%20not,require%20large%20amounts%20of%20energy. Accessed on March 9, 2022.

World Nuclear Energy Association. How can nuclear energy combat climate change? 2022. Available https://world-nuclear.org/nuclear-essentials/how-can-nuclear-combat-climate-change.aspx. Accessed on March 9, 2022.

International Atomic Energy Agency. Nuclear Share of Electricity Generation in 2020. 2022. Available at https://pris.iaea.org/PRIS/WorldStatistics/NuclearShareofElectricityGeneration.aspx. Accessed on March 9, 2022.

10.

Govind

. Ranked: Nuclear Power Production, by Country. 2022. https://www.visualcapitalistcom/ranked-nuclear-power-production-by-country/ .

11.

Troster

. Testing for Granger-causality in quantiles. Econom Rev 2018; 37: 850–866.

12.

Kartal

. The role of consumption of energy, fossil sources, nuclear energy, and renewable energy on environmental degradation in top-five carbon producing countries. Renew Energy 2022; 184: 871–880.

13.

Fell

Gilbert

Jenkins

, et al. Nuclear power and renewable energy are both associated with national decarbonization. Nat Energy 2022; 7: 25–29.

14.

Akram

Umar

Xiaoli

, et al. Dynamic linkages between energy efficiency, renewable energy along with economic growth and carbon emission. A case of MINT countries an asymmetric analysis. Energy Rep 2022; 8: 2119–2130.

15.

Rehman

Ozturk

, et al. Revealing the dynamic effects of fossil fuel energy, nuclear energy, renewable energy, and carbon emissions on Pakistan’s economic growth. Environ Sci Pollut Res 2022: 1–11.

16.

Ozcan

Ulucak

. An empirical investigation of nuclear energy consumption and carbon dioxide (CO₂) emission in India: bridging IPAT and EKC hypotheses. Nucl Eng Technol 2021; 53: 2056–2065.

17.

Bandyopadhyay

Rej

. Can nuclear energy fuel an environmentally sustainable economic growth? Revisiting the EKC hypothesis for India. Environ Sci Pollut Res 2021: 1–22.

18.

Gyamfi

Ozturk

Bein

, et al. An investigation into the anthropogenic effect of biomass energy utilization and economic sustainability on environmental degradation in E7 economies. Biofuels Bioprod Biorefin 2021; 15: 840–851.

19.

Azam

Rafiq

Shafique

, et al. An empirical analysis of the non-linear effects of natural gas, nuclear energy, renewable energy and ICT-trade in leading CO₂ emitter countries: policy towards CO₂ mitigation and economic sustainability. J Environ Manag 2021; 286: 112232.

20.

Mahmood

Wang

Zhang

. The role of nuclear energy in the correction of environmental pollution: evidence from Pakistan. Nucl Eng Technol 2020; 52: 1327–1333.

21.

Lau

L-S

Choong

C-K

C-F

, et al. Is nuclear energy clean? Revisit of environmental kuznets curve hypothesis in OECD countries. Econ Model 2019; 77: 12–20.

22.

Luqman

Ahmad

Bakhsh

. Nuclear energy, renewable energy and economic growth in Pakistan: evidence from non-linear autoregressive distributed lag model. Renew Energy 2019; 139: 1299–1309.

23.

Jin

Kim

. What is better for mitigating carbon emissions–renewable energy or nuclear energy? A panel data analysis. Renew Sustain Energy Rev 2018; 91: 464–471.

24.

Dong

Sun

Jiang

, et al.

CO₂ emissions, economic growth, and the environmental Kuznets curve in China: what roles can nuclear energy and renewable energy play?

J Cleaner Prod 2018; 196: 51–63.

25.

Baek

. Do nuclear and renewable energy improve the environment? Empirical evidence from the United States. Ecol Indic 2016; 66: 352–356.

26.

Al-Mulali

. Investigating the impact of nuclear energy consumption on GDP growth and CO₂ emission: a panel data analysis. Prog Nucl Energy 2014; 73: 172–178.

27.

Alharthi

Hanif

Alamoudi

. Impact of environmental pollution on human health and financial status of households in MENA countries: future of using renewable energy to eliminate the environmental pollution. Renew Energy 2022; 190: 338–346.

28.

Mughal

Arif

Jain

, et al. The role of technological innovation in environmental pollution, energy consumption and sustainable economic growth: evidence from south Asian economies. Energy Strategy Rev 2022; 39: 100745.

29.

Ramzan

Iqbal

Usman

, et al. Environmental pollution and agricultural productivity in Pakistan: new insights from ARDL and wavelet coherence approaches. Environ Sci Pollut Res 2022: 1–20.

30.

Awosusi

Adebayo

Altuntaş

, et al. The dynamic impact of biomass and natural resources on ecological footprint in BRICS economies: a quantile regression evidence. Energy Rep 2022b; 8: 1979–1994.

31.

Adebayo

. Environmental consequences of fossil fuel in Spain amidst renewable energy consumption: a new insights from the wavelet-based Granger causality approach. Int J Sustain Dev World Ecol 2022b; 2: 1–14.

32.

Chen

Shi

Zhang

. Structural path and sensitivity analysis of the CO₂ emissions in the construction industry. Environ Impact Assess Rev 2022; 92: 106679.

33.

Kafy

Al Rakib

Fattah

, et al. Impact of vegetation cover loss on surface temperature and carbon emission in a fastest-growing city, cumilla, Bangladesh. Build Environ 2022; 208: 108573.

34.

Qin

Horvath

. What contributes more to life-cycle greenhouse gas emissions of farm produce: production, transportation, packaging, or food loss? Resour Conserv Recycl 2022; 176: 105945.

35.

Azam

Rafiq

Shafique

, et al. Analyzing the effect of natural gas, nuclear energy and renewable energy on GDP and carbon emissions: a multi-variate panel data analysis. Energy 2022; 219: 119592.

36.

Danish, Ulucak R and Erdogan S. The effect of nuclear energy on the environment in the context of globalization: Consumption vs production-based CO₂ emissions. Nuclear Engineering and Technology 2022; 54(4): 1312–1320.

37.

Mishra

Pandey

Singh

, et al. Descriptive statistics and normality tests for statistical data. Ann Card Anaesth 2019; 22: 67.

38.

Xiao

. Quantile cointegrating regression. J Econom 2009; 150: 248–260.

39.

Engle

Granger

. Co-integration and error correction: representation, estimation, and testing. Econom: J Econom Soc 1987; 55: 251–276.

40.

Sim

Zhou

. Oil prices, US stock return, and the dependence between their quantiles. Journal of Banking & Finance 2015; 55: 1–8. https://doi.org/10.1016/j.jbankfin.2015.01.013

41.

Cleveland

. Robust locally weighted regression and smoothing scatterplots. J Am Stat Assoc 1979; 74: 829–836.

42.

Stone

. Consistent nonparametric regression. Ann Stat 1977: 595–620.

43.

Nathaniel

Alam

Murshed

, et al. The roles of nuclear energy, renewable energy, and economic growth in the abatement of carbon dioxide emissions in the G7 countries. Environ Sci Pollut Res 2021; 28: 47957–47972.

44.

Sarkodie

Adams

. Renewable energy, nuclear energy, and environmental pollution: accounting for political institutional quality in South Africa. Sci Total Environ 2018; 643: 1590–1601.

45.

Anser

Ahmad

Khan

, et al. Progress in nuclear energy with carbon pricing to achieve environmental sustainability agenda: on the edge of one’s seat. Environ Sci Pollut Res 2021; 28: 34328–34343.

46.

International Enegy Agency, I. E. A. Tracking Nuclear Power 2020 – Analysis. IEA, 2020 (June). https://www.iea.org/reports/tracking-nuclear-power-2020