Analyzing the mechanism between nuclear energy consumption and carbon emissions: Fresh insights from novel bootstrap rolling-window approach

Abstract

This research utilizes a bootstrap rolling-window (BRW) causality test to explore the causal interrelationship between nuclear energy consumption (NUC) and carbon dioxide emissions (CO₂) in 6 developed countries from 1980 to 2020. When there are structural shifts in the full-sample time series, empirical research exploring causality between two-time series generates erroneous conclusions. On the other hand, the BRW method allows researchers to find potential time-varying causality between time series using sub-sample data. The outcomes of the BRW causality test disclosed the following results: (i) a unidirectional negative causality from NUC to CO₂ without feedback was found for Japan; (ii) a negative causality at sup-sample periods from NUC to CO₂ surfaced at the sub-sample period while a positive causality surfaced from NUC to CO₂ in sub-sample period for the United States of America (USA) and France; (iii) a negative feedback causality between NUC and CO₂ was found For Canada; (iv) a positive unidirectional causality surfaced from NUC to CO₂ was found for Germany, which implies that consumption of NUC worsens the environment in the sub-sampled period. The results may have policy consequences for the selected developed countries regarding NUC and CO₂ nexus.

Keywords

Nuclear energy consumption bootstrap rolling-Window causality developed countries

Introduction

The growing concentration of CO₂ in the atmosphere adds to global warming and climate uncertainty.^1,2 Unsustainable energy use threatens the economy, society, and environment as energy production and consumption contribute 70% of all emissions.³ CO₂ must be regulated to satisfy the Kyoto Protocol and the COP-26 accord; otherwise, it will double energy-linked emissions by 2050. Energy-linked emissions may be lowered by improving energy efficiency, substituting low-emissions technologies, such as nuclear and renewable energy, for traditional carbon-intensive technologies, and electrifying heat and transportation.^4–10 Multiple approaches are needed to decarbonize economies because no single alternative can accomplish the 1.5°C objectives alone. As a result, inexpensive and dependable energy sources are essential to provide energy security and reduce foreign fuel reliance, as energy promotes economic growth and development.

In the past five decades, nuclear energy has stopped 60 gigatons of CO_2.³ Nuclear power facilities do not release greenhouse gas emissions while they operate and have very minimal emissions while supplying enormous amounts of energy.¹¹ Nuclear power plants can be evaluated as relatively inexpensive, offer reliable energy supplies, and encourage economic development and growth. In fact, nuclear power facilities have high initial capital expenditures, but they have low operational expenses.^12,13 As a result, nuclear energy costs are predictable and steady over long periods as variations in nuclear energy are less noticeable due to the cost structure of nuclear reactors.¹⁴ Nuclear energy promotes human health by ensuring a green environment. Nuclear energy has been improving and becoming more versatile and efficient. It gives emerging economies access to dependable, low-cost, and carbon-free power, which helps their economic growth. Nuclear energy has a lower mineral need than other energy industries.¹⁵ The electricity generated by nuclear power reactors grew from 2563 terawatt hours in 2018 to 2657 terawatt hours in 2019. Moreover, nuclear energy's proportion in global power generation is predicted to increase to 25 by 2050%. Since nuclear energy can substitute fossil fuel energy safely, reliably, economically, and sustainably, it will play a significant part in the transition to clean energy.¹⁶

Numerous attempts have been made from both theoretical and empirical perspectives, along with significant policy ramifications, which are thought to aid in reaching agreements on environmental-linked problems. Despite ongoing efforts, ecological problems resulting from the continual increase in carbon emissions, which cause global warming, have not only stayed unaddressed but also have the possibility of presenting bigger risks in the future. There are at least three elements that are connected to this unsolved problem. First, the globe's productive capacity continues to be primarily dependent on fossil fuels despite widespread support for switching from fossil fuel to renewable forms of energy. Secondly, due to the restricted accessibility and cost of renewable energy, the intense concentration on it has not yet produced the anticipated outcomes. Consequently, a clean alternative energy source is required throughout the present transition period. Thirdly, nuclear energy is the closest substitute for fossil fuels when it comes to being a cheap, readily available, and reasonably priced energy source. Nuclear energy is also a low-carbon-intensive energy source that can facilitate the transition to sustainable energy use. The International Energy Agency¹⁷ referred to this viewpoint when it claimed that the operation of nuclear reactors did not increase carbon dioxide or air pollution. According to different research, nuclear energy is low-carbon and can create electricity on a big scale with negligible carbon emissions, on par with solar and wind energy.¹⁸ The potential of nuclear energy to cut emissions unevenly, equal to a one-third drop in transport-induced emissions globally, makes apparent the crucial role that it will play in achieving the continued aim of carbon neutrality by 2050.¹⁸

Exploring the role of nuclear energy in decreasing emissions in developed countries is essential. Likewise, the selected 6 developed countries, namely Canada, Germany, France, the United Kingdom, the United States, and Japan, are dedicated to lowering CO₂ by diversifying their energy portfolios, particularly by including renewable energy sources into their conventional energy mix. However, these countries’ CO₂ per capita (see Figure 1) is higher than that of India and China.^19–21 These countries contribute around 30% and 25% of total global energy consumption and CO₂, respectively.²² Furthermore, these countries rely heavily on imported and indigenous non-renewable energy supplies. Interestingly, the selected countries rely on non-renewable energy imports to meet their energy needs. For instance, Japan and Germany import over 96% and 64% of their total primary energy supply, respectively.²³ Consequently, these figures highlight the selected countries’ dilemma of dirty fuel reliance. These statistics demonstrate why, despite their economic success, these countries have faced environmental problems. A persistent disparity between energy supply and demand has existed for decades, resulting in significant social and economic losses. Moreover, the current power demand will be expected to double by 2025. As a result, these countries have difficulties relating to ecological deterioration and energy scarcity that require rapid policy remedies. Based on these interesting facts about the selected nations, there is a need to evaluate the association between nuclear energy and CO2 emissions.

Figure 1.

Trend of CO₂ per capita between 1980 and 2020.

This study adds to the current knowledge by accounting for time fluctuations in the causal relationships between nuclear energy and CO₂ in the selected countries. When the underlying series undergo structural shifts, empirical research investigating causation between two variables may produce erroneous conclusions. In the face of structural alteration, the linkages across series will demonstrate instability over distinct sub-periods.²⁴ We employ the Mean-F, Exp-F, and Sup-F tests initiated by Andrews²⁵ and Andrews & Ploberger²⁶ as well as the Lc test suggested by Hansen²⁷ and Nyblom²⁸ to evaluate the short-run and long-run parameter stability interconnection between CO₂ and nuclear energy in the selected countries by taking structural changes into account. Furthermore, we used the BRW test to evaluate the causal linkages between nuclear and CO₂. The BRW technique, which uses a fixed-size window to test for causality on rolling sub-samples, enables researchers to catch time fluctuations and structural changes in causal linkages. In this aspect, this study paper stands out from the previous research, which, in particular, only addresses full causality and, unlike this research, is vulnerable to erroneous findings and inferences in the face of uncertainty in parameters caused by the structural break(s). The results will aid environmentalists, development practitioners, and energy experts in developing and implementing ecologically friendly measures that will assure ecological sustainability. This research will be beneficial to economies by prioritizing nuclear energy investments.

The research is organized as follows: section 2 presents the information regarding nuclear power in the selected countries. A summary of prior studies and methodology is presented in Sections 3 and 4, respectively. Sections 5 and 6 illustrate the result and conclusion.

Nuclear power in the selected countries

Due to a long-standing strategy predicated on energy security, France gets around 70% of its electricity from nuclear power. The goal of government policy, established in 2014 under a previous administration, was to decrease nuclear power's contribution to energy generation to 50% by 2025.²⁹ This objective was pushed back from 2019 to 2035. France revealed plans to construct six nuclear reactors in 2022 and said it would explore building an additional eight. Due to its extremely cheap cost of production, France is the most significant net exporter of electricity in the world, earning over €3 billion annually. The nation has made great strides in the development of nuclear technology. Reactors, particularly fuel-related goods and services, have been major exports.²⁹ French power is made from nuclear fuel reprocessing to 17%. Figure 2(a) presents the Operable nuclear power capacity in France.

Figure 2.

(a) Operable nuclear power capacity in France. (b) Operable nuclear power capacity in Germany. (c). Operable nuclear power capacity in Japan. (d) Operable nuclear power capacity in the United Kingdom. (e) Operable nuclear power capacity in the United States. (f) Operable nuclear power capacity in Canada.

Prior to March 2011, Germany used 17 nuclear reactors to generate one-fourth of its power. Germany plans to phase out nuclear energy gradually. The phase-out of nuclear energy was a component of the policies of a coalition government established following the 1998 federal elections.²⁹ The phase-out was abandoned in 2009 after a change of administration, but it was reinstated in 2011 in response to the Fukushima catastrophe in Japan, which resulted in the temporary shutdown of eight reactors. In Germany, there is still strong public opposition to nuclear power and hardly any support for constructing additional reactors.²⁹ As a result of its policy, Germany has some of the lowest wholesale power rates in all of Europe and some of the highest retail prices. The price of home power is more than half covered by taxes and fees. Figure 2(b) presents the operable nuclear power capacity in Japan.

Japan needs to import about 90% of its energy requirements. The country's first commercial nuclear power reactor began to operate in the middle of 1966, and nuclear energy has been a high national strategic priority since 1973. This was considered after the Fukushima disaster in 2011, but it has since been proven.²⁹ Up until 2011, Japan obtained roughly 30% of its energy from reactors; by 2017, it was anticipated that this percentage would reach at least 40%. By 2030, it is currently expected to replace at least 20% of the fleet. Since the first two reactors were revived in August and October of 2015, eight more have been done so.²⁹ The restart of 16 reactors is currently being authorized. Figure 2(c) presents the operable nuclear power capacity in Japan.

With a nuclear capacity of around 6.5 GW, the UK produces 15% of its power. Most of the current capacity is scheduled for retirement by the decade's end. However, the first nuclear plant of a new generation is now being built. Up to 24 GW of new nuclear capacity is required under government plans by 2050 to produce around 25% of the nation's power.²⁹ The UK has implemented a comprehensive evaluation procedure for new reactor designs and their location. Due to the UK's privatization of power generation and liberalization of its energy market, substantial capital investments are troublesome.²⁹ Figure 2(d) presents the operable nuclear power capacity in the United Kingdom.

With more than 30% of all nuclear electricity generated globally, the USA is the world's biggest nuclear power generator. In 2019, the nation's nuclear power plants generated 843 billion kWh, or 19% of all electricity production.²⁹ After a 30-year era in which few new reactors were constructed, it is anticipated that two more new units will go online shortly after 2020 as a consequence of 16 license applications submitted since mid-2007 for the construction of 24 new nuclear reactors.²⁹ The financial sustainability of certain current reactors and potential projects is in doubt due to some governments’ liberalization of wholesale electricity markets. This has made it challenging to finance capital-intensive power projects. Lower gas prices since 2009 have also contributed to this. Figure 2(e) shows the operable nuclear power capacity in the United States.

Nuclear energy generates around 15% of Canada's electricity, with 19 reactors, most of which are in Ontario, with a combined capacity of 13.6 Gwe.²⁹ Canada had intended to build two more new reactors during the following ten years to increase its nuclear capacity, but these plans have been postponed.²⁹ Canada has long been a global leader in nuclear research and technology, exporting reactor systems developed in Canada and a sizable share of the radioisotopes utilized in medical diagnosis and cancer treatment. Figure 2(f) shows the operable nuclear power capacity in Canada.

Literature review

The SDGs are inextricably intertwined. At a global level, synergies in decision-making are essential to achieve sustainability. As the sustainable development theory highlights, it is necessary to utilize resources in such a way that subsequent generations will be able to meet their demands. Since the energy sector primarily relies on fossil fuel extraction, which is constrained in quantity and diminishing owing to increased extraction to satisfy rising demand, future generations’ capacity to use this resource is endangered.^20,30 As a result, resources that assure long-term viability are essential; in this sense, nuclear energy based on uranium is a viable choice.

Nuclear power plants use uranium widely disseminated across five continents, whereas fossil fuels are seen in vulnerable places that make them less susceptible to interruptions and fuel supply. Nuclear plant facilities can be built so that they are less susceptible to climate change and can assist in developing synthetic fuels for the transportation sector, significantly reducing emissions.^2,31–35 Nuclear energy differs from hydrocarbon energy in two ways: first, it generates heat without emitting greenhouse gases, and second, it keeps dangerous materials inside the fuel throughout nuclear fusion, decreasing its total footprint. Moreover, nuclear plant emissions are caused by the use of fossil fuels during their life cycle, which can be decreased by substituting non-emitting fuels.³⁶ It is also known that nuclear plant emissions are comparable to renewable energy sources over their whole life cycle, which implies that it is a source aiding in decarbonization and minimizes the influence of other energy sectors on human health.

Moreover, it can help with medicinal and industrial uses in addition to reducing emissions. Nuclear medicine is helping a rising number of individuals, with around 30 million people, who have benefited each year. Moreover, nuclear power averted 1.8 million emission-linked fatalities worldwide from 1971 to 2009.³⁷ Nuclear energy preserves food by subjecting it to gamma rays, which extend its shelf life and destroy microorganisms. It also aids in the conservation and management of water resources as well as the discovery of new sources and water desalination. Nuclear fuel has a higher energy density than other fuels and a more negligible ecological impact. Therefore, nuclear plants consume less area and have fewer resource requirements than other sources, which results in minor global damage to biodiversity.¹⁴

Nuclear power plants produce highly radioactive waste that must be shielded appropriately and disposed of because it is highly hazardous. All energy industries create toxic waste, which, if not properly managed, can pollute the environment; hence management is essential. Nuclear waste is produced in tiny quantities, is carefully controlled and disposed of, is not distributed in the ecosystem, and is recyclable. Nuclear power stations can impact the aquatic ecosystem and cause water scarcity since they require water for cooling and 66% of that water is sent directly into the environment as heat.³⁸ On the other hand, ecosystem stress can be alleviated via proper planting that decreases aquatic and water species stress. Nuclear power facilities with cogeneration have higher thermal efficiency and less environmental impact. In addition, while the fatality rate connected with nuclear energy is modest compared to other conventional sources, there is high displacement.

Studies examining the relationship between NUC and CO₂ can be divided into two categories based on their findings. Numerous researches has looked at whether or not using nuclear energy reduces emissions. For instance, using Granger causality and dataset from 1960 to 2007, Menyah & Wolde-Rufael³⁹ examined the NUC-emissions nexus in the USA. The findings revealed that while NUC can assist in curbing CO₂, renewable energy usage has yet to reach a point where it can significantly contribute to reducing emissions. Similarly, Saidi & Omri⁴⁰ studied the OECD countries and reported that the reduction in CO₂ was caused by a surge in NUC between 1990 and 2018 using VECM and FMOLS approaches. Likewise, Dong et al.⁴¹ study the connection between NUC and CO₂ using data from 1993 to 2016 in China and reported that the utilization of nuclear energy triggers the reduction in China's emissions level. The study research of Hassan et al.⁴² using CUP-BC and CUP-FM in the BRICS economies between 1993 and 2017 reported that the fall in BRICS countries’ emissions levels is due to an increase in consumption of nuclear energy. Nathaniel et al.⁴³ study on the role of NUC towards carbon neutrality in G7 countries reported that NUC curbs CO₂. Recently, a Danish et al.⁴⁴ study in OECD countries reported that NUC contributes to the quality of the environment.

The second strand of studies reported an insignificant interrelationship between nuclear energy and CO₂. For instance, from 1990 to 2013, Saidi & Ben Mbarek³⁰ study reported positive but insignificant emissions-nuclear energy connections using advanced countries. Likewise, an insignificant association between NUC and CO₂ surfaced in the study of Jin & Kim⁴⁵ for selected 30 countries utilizing a dataset from 1990 to 2014. The third strand of literature reported positive nuclear energy-emissions nexus. For instance, using South Africa and a dataset from 1971–2017, Sarkodie & Adams⁴⁶ examine the nexus between CO₂ and nuclear energy using ARDL. The findings disclosed that environmental deterioration is worsened by the consumption of nuclear energy in South Africa. Likewise, Syed et al.⁴⁷ study on the influence of NUC on pollution in India using the NARDL from 1975 to 2018 reported that a positive (negative) shift in nuclear energy increased (decreased) pollution in India. Majeed et al.⁴⁸ research in Pakistan using data between 1974 and 2019 disclosed insignificant nuclear energy-emissions connections.

Based on the summarized studies above, it is evident that several studies have been conducted on the association between CO₂ and nuclear energy. Nevertheless, their results are mixed based on the techniques deployed, countries in focus, and timeframe. Moreover, most of these studies are premised on using conventional techniques such as Granger causality, VECM, Toda Yamamoto (TY) causality, autoregressive distributed lag, and ordinary least squares. Therefore, this study fills the gap in the literature by deploying a sophisticated technique known as the BRW approach. Unlike conventional causality tests, this test for causality on rolling sub-samples enables researchers to catch time fluctuations and structural changes in causal linkages. In this aspect, this study stands out from the previous research, which, in particular, only addresses full-sample causality.

Methodology and data

Econometric model

In this study, we use Granger causality tests to determine the causal link between CO₂ and NUC using a bivariate VAR model. The Wald statistics and Likelihood ratio are the test statistics utilized in the current paper on Granger causality tests. When the sample size is small, traditional causality tests in VARs have the drawback of being non-asymptotic. As a result, as an improved Granger causality test, Toda & Yamamoto⁴⁹ is employed, which can yield valid asymptotic critical values for any order of integration, whether cointegrated or non-cointegrated.

We employ bootstrapping Granger (BG) causality tests, which are resilient to pre-testing prejudice and small sample size, to solve sample size difficulties. Moreover, Balcilar et al.²⁴ proposed the residual-based bootstrap (RB) approach, which is applied in this empirical analysis. The RB technique delivers a strong CV when Granger causality is tested.^24,50 As a result, we use the Toda and Yamamoto adjusted edition Granger causality test to perform the bootstrap approach. We employ PP and ADF tests to check unit roots’ existence. To check for unit root, we follow the specification in Equation 1:

Δ y_{t} = β_{0} + δ y_{t - 1} + \sum_{t - 1}^{k} α_{i} Δ y_{t - 1} + ε_{t}

(1)

The technique used in this study is based on research by Balcilar et al.,²⁴ Dlamini et al.,⁵¹, and Adebayo et al.²¹ The structural changes/regime transitions, small sample size, and how various studies interpret the trending features of data are all attributed to the highly divergent and inconsistent causal linkages in this research. Balcilar et al.²⁴ employ rolling window estimate strategies and bootstrap testing to solve these difficulties. The rolling window subsamples and full sample, where causality is assessed over distinct periods, are both subjected to bootstrap Granger causality tests. We frequently presume that parameter estimations are steady over time in time series analysis. Nevertheless, policy shifts and the economy's constant evolution may make this stability claim incorrect.

We must consider the model's instability or stability characteristics to acquire trustworthy findings. Therefore, parameter constancy was utilized to verify this. In the sub-sample period, evidence of causality may surface, whether bidirectional or unidirectional. Stability tests on the VAR were conducted to see whether structural fractures had happened or if the predicted coefficients were consistent throughout our period. The Sup-F, Mean-F, Nyblom-Hansen Lc, and Exp-F tests assess structural shift and parameter instability. Suppose any of the stability tests reveal the existence of nonstable parameters. In that case, it is feasible to determine when and where the instability happened and if any plausible phenomena triggered it. This is accomplished by combining granger causality with rolling regression tests. The rolling estimation approach generates many similar parameter estimations for the system to be stable. Suppose the parameter estimates of a time series change significantly from those of others, i.e., strong volatility. In that case, the model is considered unstable, and the estimated outcomes will be incorrect. Consequently, we use RWG causality tests to identify structural shifts, which may provide varied causality findings across rolling data subsamples. The following VAR bivariate model is considered:

y_{t} = φ_{0} + φ_{1} γ_{t - 1} + \dots . . φ_{p} γ_{t - p} + ε_{t} t = 1, 2 \dots . ., T

(2)

where

ε_{t} = (ε_{1 t}, ε_{2 t})^{ι}

denotes the white noise with covariance matrix ∑ and zero mean. Furthermore, the AIC is utilized to capture the order of lag, which is denoted by p. The sub-vectors

y_{1 t} and y_{2 t}

are partitioned and represent NUC and CO₂, respectively. Equation 2 is restructured in the following compact form as follows:

[\begin{matrix} y_{1 t} \\ y_{2 t} \end{matrix}] = [\begin{matrix} φ_{10} \\ φ_{20} \end{matrix}] + [\begin{matrix} φ_{11} (L) & φ_{12} (L) \\ φ_{21} (L) & φ_{22} (L) \end{matrix}] [\begin{matrix} y_{1 t} \\ y_{2 t} \end{matrix}] + [\begin{matrix} ε_{1 t} \\ ε_{2 t} \end{matrix}]

(3)

where

φ_{ij} (L) = \sum_{k - 1}^{p} φ_{ij}, k^{L^{k}}, i, j = 1, 2, and

nd the lag operator is defined as

L_{x_{t}}^{k} = x_{i = k}

The reason for utilizing this approach rather than a conventional Granger causality test is that typical granger causality tests assume parameters remain constant across time, leading to an inaccurate conclusion. We can analyze structural break(s) in the model using RWG causality tests relying on the bootstrap approach. The causality between variables can be identified as it evolve over time utilizing rolling estimation, and we can identify structural changes and causal relationships for rolling various subsamples.

Data

We utilize yearly data to examine the causality between NUC and CO₂ for the period from 1980 to 2020 (41 observations) in the selected 6 developed countries. The dataset for NUC was assembled from the British Petroleum⁵² database and measured in exajoules. In contrast, the dataset for CO₂ is measured as metric tonnes per Capita and obtained from the BP⁵² database. The natural log of the variables used is taken in this empirical investigation. Table 1 presents the variables’ statistics.

Table 1.

Descriptive statistics.

	Canada	France	Germany	Japan	UK	USA
	CO₂
Mean	16.82632	6.669291	11.31337	9.121257	8.875139	19.12944
Median	16.71053	6.953469	11.05368	9.444917	9.599748	19.74997
Maximum	18.70544	9.398065	14.05234	10.24598	10.62260	21.29157
Minimum	13.02700	4.358200	7.872900	7.282000	5.247000	13.68000
Std. Dev.	1.112425	0.994679	1.533011	0.972016	1.590314	1.790668
Skewness	−0.640395	−0.079298	−0.071317	−0.804012	−1.082367	−0.950045
Kurtosis	4.583188	3.464361	2.233132	2.125044	2.756939	3.329601
Jarque-Bera	7.084299	0.411339	1.039402	5.725122	8.106308	6.353258
Probability	0.028951	0.814102	0.594698	0.057122	0.017368	0.041726
	NUC
Mean	221.1436	952.9703	350.7189	510.7687	190.8773	1842.169
Median	234.7078	1074.926	392.7346	540.4762	180.7708	2109.762
Maximum	297.4564	1214.557	473.1337	905.4837	276.3501	2254.808
Minimum	105.6445	170.1418	154.4140	11.44568	102.8417	734.2567
Std. Dev.	45.84240	274.3363	107.3871	297.5353	46.19983	476.3217
Skewness	−1.175744	−1.467225	−0.488064	−0.325655	0.166404	−1.195846
Kurtosis	3.810052	4.175779	1.622738	1.739129	2.161635	3.057820
Jarque-Bera	10.56720	17.07214	4.868200	3.440586	1.389928	9.777709
Probability	0.005074	0.000196	0.087677	0.179014	0.499092	0.007530

Table 2.

Stationarity test without break outcomes.

		Canada	France	Germany	Japan	UK	USA
ADF	CO₂	−0.3162	−2.6144	−2.4635	0.7284	−0.5868	0.2123
ADF	$Δ$ CO₂	−4.6273*	−5.6315*	−7.0011*	−4.3180*	−7.8198*	−5.3612*
PP	CO₂	−0.3885	−2.8960	−2.4635	0.5895	−0.4118	0.2123
PP	$Δ$ CO₂	−3.2577**	−5.6495*	−7.0764*	−4.0951*	−8.0208*	−4.2224*
ADF	NUC	−2.2673	−2.6440	−2.3172	−1.4335	−1.6429	−1.3291
ADF	$Δ$ NUC	−5.5869*	−6.3765*	−6.9545*	−4.7302*	−6.7404*	−4.9823*
PP	NUC	−2.3160	−2.4872	−2.2050	−1.5111	−1.5960	−1.3276
PP	$Δ$ NUC	−5.5836*	−6.3827*	−6.9623*	−4.2118*	−6.7816*	−5.212*

Note: *P < 10% and **P < 5%.

Discussion

Pre-estimation outcomes

Before the main analysis, it is crucial to identify the integration order of NUC and CO₂ in each country. As a result, we employ PP and ADF tests to check for the existence of unit roots. Table 2 presents the ADF and PP for each country and Table 3 presents the ZA results

Table 3.

Stationarity test with break outcomes.

	Canada	France	Germany	Japan	UK	USA
CO₂	−4.792	−1.752	−4.193	−3.553	−2.569	−4.028
BD	2010	2004	2009	2003	2005	2009
$Δ$ CO₂	−5.537**	−6.281*	−5.989*	−6.836*	−5.895*	−5.483**
BD	2009	2004	2009	2008	2005	2009
NUC	−4.119	−3.922	−4.193	−3.032	−4.294	−3.843
BD	2000	2002	1999	1995	1995	2006
$Δ$ NUC	−6.823*	−5.378**	−5.849*	−6.028*	−5.359*	−5.872*
BD	2000	2002	2006	2004	1997	2006

Note: *P < 10% and **P < 5%.

At level, NUC and CO₂ are non-stationary; however, after differencing, both NUC and CO₂ are stationary in each country. Furthermore, we used the⁵³ test to catch the stationarity feature of the variables in the presence of a single break. Table 3 presents the ZA results with results confirming stationarity at the first difference. We employ VAR to test the causal interrelationship between NUC and CO_2.

Full sample Granger causality outcomes

We undertake a full sample Granger causality test in each country after confirming that our series are I(1). Instead of utilizing a VECM, we utilize the bivariate VAR to evaluate causality. For the period 1988 to 2020, a full sample Granger (FSG) causality test is developed using the Wald test statistics and bootstrap Likelihood ratio. The FSG test results are exemplified in Tables 4 and 5.

Table 4.

Full sample Granger causality tests outcomes.

	H0: NUC does not cause CO₂		H0: CO₂ does not cause NUC
	LR-Statistics	Bootstrap p-value	LR-Statistics	Bootstrap p-value
Canada	1.0013	0.760	1.0346	0.620
France	8.2836***	0.0729	3.1405	0.160
Germany	1.2997	0.580	13.555*	0.000
Japan	2.8432	0.230	0.8539	0.680
UK	9.0105*	0.010	5.0109	0.110
USA	0.8364	0.650	1.1322	0.520

Notes: *P < 10% and ***P < 1%. Causality tests are based on a VAR model, with the length of lag determined by SIC. H0 hypothesis is "no existence of causal interrelationship.

Table 5.

Results of short-run and long-run parameter stability tests.

		NUC Equation		CO₂ Equation		VAR System
		Statistics	BootstrapP-values	Statistics	BootstrapP-values	Statistics	BootstrapP-values
Canada	Sup-LR	55.246	0.000*	63.254	0.000*	87.471	0.000*
	Exp-LR	33.227	0.000*	48.072	0.000*	48.305	0.000*
	Mean-LR	39.930	0.000*	31.892	0.000*	33.727	0.000*
	$L_{c}^{b}$					3.908	0.000*
France	Sup-LR	84.254	0.000*	41.206	0.000*	22.081	0.000**
	Exp-LR	49.052	0.000*	20.753	0.000*	13.229	0.009**
	Mean-LR	22.973	0.000*	17.437	0.258	11.449	0.029**
	$L_{c}^{b}$					1.998	0.019***
Germany	Sup-LR	60.246	0.000*	79.640	0.000*	40.684	0.000*
	Exp-LR	26.046	0.000*	56.101	0.000*	22.754	0.000*
	Mean-LR	18.327	0.000*	33.246	0.000*	17.162	0.000*
	$L_{c}^{b}$					4.99	0.000*
Japan	Sup-LR	86.085	0.000*	40.936	0.000*	28.763	0.000*
	Exp-LR	54.325	0.000*	22.437	0.000*	17.543	0.000*
	Mean-LR	20.473	0.000*	16.073	0.000*	9.947	0.053**
	$L_{c}^{b}$					0.629	0.565
United Kingdom	Sup-LR	102.028	0.000*	39.906	0.000*	28.095	0.000*
	Exp-LR	75.382	0.000*	23.754	0.000*	17.986	0.000*
	Mean-LR	30.271	0.000*	15.863	0.000*	13.454	0.013**
	$L_{c}^{b}$					1.254	0.259
United States	Sup-LR	91.607	0.000*	40.863	0.000*	33.652	0.000*
	Exp-LR	76.874	0.000*	23.765	0.000*	19.896	0.000*
	Mean-LR	29.753	0.000*	15.753	0.000*	11.258	0.021**
	$L_{c}^{b}$					3.074	0.000**

Notes: 10,000 bootstrap repetitions are used to calculate P-value. *P < 10%, ** < 5% and ***P < 1%.

For France and UK, the outcomes reveal that NUC Granger causes CO₂, while in Germany, CO₂ Granger causes NUC. No causality exists between CO₂ and NUC for the USA, Japan, and Canada.

Parameter stability tests outcomes

Furthermore, validating the reliability of our full sample causality outcomes requires analyzing the temporal coefficient stability of our computed VAR model. We argue that our FSG causality conclusions are legitimate if the parameter estimates are consistent across the whole sample period. On the other hand, if parameter estimates are discovered to be temporally unstable, resulting in erroneous full sample Granger causality conclusions, further analysis of the times in the sample when instability emerges is required. Changes in parameters may result from structural breaks and the trend of causal interrelationships may change.²⁴ Consequently, various research using different sample periods may provide contradictory causation outcomes. As a result, we must test for parameter stability and the likely reasons for such parameter changes. We used mean-F, Exp-F, and Sup-F tests to verify the parameter stability in the short-term. The mean-F, Exp-F, and Sup-F tests liken the H0 hypothesis to “parameter stable” against the Ha hypothesis “parameter instability.” The Sup-F test ascertains if a swift regime shift happened, while the Mean-F and Exp-F tests evaluate if the models stayed constant over time.^51,54

Since the parameters of investigation are I(1) variables, we perform Nyblom²⁸ and Hansen²⁷ Lc tests for the parameter stability of all the system parameters. To ascertain whether or not equations are cointegrated, we used these tests. These stability tests benefit from not requiring previous knowledge of the structural break's timing. Outcomes of the stability parameter reveal structural fractures in our sample for each country, meaning that our model is unstable (See Table 4). Moreover, with the exemption of the United Kingdom, Nyblom and Hansen Lc tests dismissed the H0 of “parameter stability for the system” for France, Canada, Japan, the United States, and Germany. Thus, structural breaks evidence surfaced, and the model displays joint instability for France, Canada, Japan, the USA, and Germany with the exemption of the United Kingdom, which exhibit joint stability. Each country's Mean-F, Sup-F, and Exp-F tests reveal some model uncertainty. The H0 hypothesis is dismissed at a 1% significance level in the Mean-F, Exp-F, and Sup-F test statistics that indicate parameter uncertainty in each country's NUC equation and CO₂ equation. Based on the outcomes of the stability tests, we can deduce that each country model has parameters instability across the period, which means that the findings of the full sample cannot be trusted and are thus unacceptable.

Bootstrap Rowling window outcomes

The time-varying pattern of the causal relationships between CO₂ and NUC is explored using the bootstrap-based TY Granger causality test in a rolling window estimate paradigm, centered on the parameter stability test outcomes. The adoption of the rolling window regression approach is justified for two reasons. Firstly, the rolling window estimation permits the connection between variables to change over time. Secondly, the RWG causality method considers the possibility of structural discontinuities and regime changes in the causal interrelationship between NUC and CO₂. The VAR model lag length is set utilizing the AIC in each stage of the rolling window estimation. The causality test is conducted utilizing the bootstrap technique on each sub-sample. The window size, which defines the number of observations handled in each subsample and the overall number of rolling estimates, is an imperative decision parameter in rolling estimation.

More crucially, the window size affects the subsample estimates’ representativeness and accuracy. Moreover, larger window sizes offer more exact estimates, but they may lower representativeness, mainly when heterogeneity exists. On the flip side, a small window size minimizes heterogeneity and enhances parameter representativeness, but it may also raise the SE of estimates that lowers parameter precision. As a consequence, the size of the window should be adjusted to strike a compromise between representativeness and accuracy. There is no rigorous requirement for setting the size of the window in the estimation of the rolling window. Moreover, Pesaran & Timmermann⁵⁵ utilized Monte Carlo simulations to disclose that the break size and persistence determine the ideal size of the window. Furthermore, they suggested that when there are numerous interruptions, the minimum window size should be 20. This window size does not include the lag observations. Therefore, it is the actual amount of VAR model observations. Considering that the small size of the window may result in imprecise estimates, each sub-sample estimation is subjected to bootstrap procedures to achieve more accurate parameter estimates and tests.

For each of the selected countries, Figures 3–8 show the p-values of the bootstrap rolling causality test statistics and the cumulative extent of the influence of one series on another. The null hypothesis of “no causal association” between the NUC and CO₂ cannot be discarded at the 10% significance level when the purple lines, which indicate estimated p-values, are above the horizontal line (red line), indicating the significance level of 10%. Panels c and d show the sum of the rolling bootstrap estimates coefficients that measure the influence of NUC on CO₂ and vice versa.

Figure 3.

Rolling window estimates for Canada.

Figure 4.

Rolling window estimates for France.

Figure 5.

Rolling window estimates for Germany.

Figure 6.

Rolling window estimates for Japan.

Figure 7.

Rolling window estimates for the UK.

Figure 8.

Rolling window estimates for the USA.

Figure 3 presents the bootstrap rolling causality test outcomes for Canada. Throughout the research period, it can be observed that the causal association between NUC and CO₂ is relatively weak. The Ho hypothesis of “no causality from NUC to CO₂” is dismissed at a significance level of 10% (Figure 3 panel a) in the sub-periods from 1988 to 1991, 2009 to 2014, and 2020. Moreover,Figure 3 panel c discloses negative and significant causality from NUC to CO₂ only in the sub-periods from 1988 to 1991 and from 2009 to 2014, which implies that NUC plays a crucial role in mitigating emissions. Furthermore, the H0 hypothesis of “no causality from CO₂ to NUC (Figure 3 panel b) is dismissed at a significant level of 10% from sub-sample periods from 2002 to 2003 and 2017 to 2018. In addition, panel d discloses negative and significant causality from CO₂ to NUC emissions only in the sub-periods from 2002 to 2003 and from 2017 to 2018 (see Figure 3 panel c). These results are not surprising given that Canada ranks sixth among countries regarding nuclear power generation. Nuclear energy produces no greenhouse gas emissions during its operation. Thanks to Canada's enormous uranium supply, it is easily extensible and does not take up nearly as much space as wind farms, solar arrays, or hydroelectric power. Canada expects nuclear power to be a key component of its clean-energy mix, which will help to reduce carbon emissions drastically. Nuclear power generates around 15% of Canada's electricity¹. Canada is a global leader in nuclear protection and energy. Its home market for the safe and responsible development of small modular reactor (SMR) technology is one of the most attractive globally. SMRs can generate significant economic advantages for Canada's economy while also assisting in the country's efforts to attain net-zero greenhouse gas emissions by 2050.

Figure 4 presents bootstrap rolling causality test outcomes for France. Panels a and b in Figure 3 also disclose a comparatively significant causality between CO₂ and NUC. The Ho hypothesis of ”no causality is dismissed from 1992–1993 and 2002. Moreover,Figure 4 panel c discloses negative and significant causality from NUC to CO₂ only in the sub-periods from 1992–1993 and 2002. This demonstrates that NUC lessens CO₂ in the sub-sample periods. Furthermore, the Ho hypothesis of “no causality from CO₂ to NUC (Figure 4 panel d) is dismissed at a significant level of 10% from sub-sample periods from 1994–2000 and 2002–2004. Moreover, panel d discloses positive and significant causality from CO₂ to NUC emissions only in the sub-periods from 1994–2000 and 2002–2004. These outcomes affirmed the feedback hypothesis between CO₂ and NUC. The negative effect of nuclear energy on emissions is not surprising, given that France generates roughly 70% of its electricity from nuclear energy due to its long-standing measure centered on energy security. France boasts the globe's largest nuclear power plant in terms of population. Indeed, as per 2019 data, nuclear energy provides 72% of electricity, renewable energy provides 20%, and fossil fuels provide 8%. Nuclear energy allows France to be 50% energy self-sufficient while still enabling it to profit via the export of electricity. The Nuclear Power Development Strategy is linked to the Multiyear Energy Plan's aims (MEP) and Energy Transition for Green Growth Act (ETGGA)².

Figure 5 reveals the bootstrap rolling causality test outcomes for Germany. We observed that the causal association between NUC and CO₂ was weak during the study period. The results refute the Ho hypothesis of “no causality from NUC to CO₂” at a 10% level of significance (Figure 5 panel a) in the sub-periods 1990, 1998, 2013, and 2019. In addition, Figure 5 panel c discloses significant and positive causality from NUC to CO₂ in Germany only in the sub-periods from 1990, 1998, 2013, and 2019, demonstrating that NUC intensifies environmental damage in Germany. Besides, the Ho hypothesis of ”no causality from CO₂ to NUC (Figure 5 panel b) is refuted at a significant level of 10% from sub-sample periods 1991, 1996, and 2002–2005. In addition, a negative and significant causality from CO₂ to NUC emissions in Germany surfaced only in the sub-periods 1991, 1996, and 2002–2005 (see Figure 5 panel d). These results support the feedback hypothesis in Germany. These results align with what is anticipated, given Germany's policy of eradicating nuclear energy. Germany used 17 reactors to generate 25% of its electricity from nuclear energy up to March 2011. After the oil price shock of 1974, German advocacy for nuclear energy was high in the 1970s, and there was a feeling of insecurity concerning energy sources, similar to France. After the Chornobyl disaster in 1986, this approach collapsed, and the final new nuclear power plant was installed in 1989. In December 2022, Germany's final nuclear power facility will shut down³. The 2011 Nuclear Energy Act (Atomgesetz) removed the authorization to run nuclear reactors for electricity generation on a phase-out timeline. Nuclear power's contribution to total energy output declined from 22.2% in 2010 to 11% in 2020. Renewables, including solar PV, biogas, and wind, on the flip side, will account for roughly 45% of total power generation in 2020. Nonetheless, it paints a very dubious image if a country that prides itself on being a climate action pioneer still has some of Europe's biggest power station polluters and would have probably failed its 2020 carbon reduction targets if not for the COVID-19 lockdown⁴.

Figure 6 discloses the bootstrap rolling causality test outcomes for Japan. We noticed that the causal interconnection between NUC and CO₂ was weak in the study period. The study results dismissed the Ho hypothesis of “no causality from NUC to CO₂” at a 10% level of significance (Figure 6 panel a) in the sub-periods 1988, 2006, 2008, 2012–2013, and 2019. In addition, panel c discloses significant and negative causality from NUC to CO₂ in Japan only in the sub-periods 1988, 2006, and 2008, demonstrating that NUC intensification enhances the environmental quality; however, the causality from sub-sample periods from 2012–2013 and 2019 is positive and significant. Moreover, the Ho hypothesis of “no causality” from CO₂ to NUC (Figure 6 panel b) is dismissed at a significance level of 10%. These results support the nuclear energy lead emissions hypothesis for the case of Japan. The findings might be explained by the fact that nuclear power reactors have been a national strategic goal since 1973, before the Fukushima disaster in 2011. The Japanese government indicated in March 2002 that it would rely significantly on nuclear energy to meet the Kyoto Protocol's GHGs emissions reduction targets. Cabinet approved a 10-year energy plan submitted to the Minister of Economy, Trade, and Industry (METI) in July 2001. This strategy projected a 30% growth in nuclear power capacity (13,000 MWe), with utilities expecting to have up to 12 additional nuclear units operational by 2011⁵. Nonetheless, after the Fukushima accident, this strategy was reviewed. The possible reason for the positive effect of nuclear energy on CO₂ is due to the Fukushima accident, which causes a decrease in nuclear energy generation.

Figure 7 reveals the bootstrap rolling causality test outcomes for the United Kingdom. During the investigation, we noticed weak causality between NUC and CO₂. There is causality from NUC to CO₂ at a 10% significance level (Figure 7 panel a) in the sub-periods 1992 and 2019. Moreover,Figure 7 panel c discloses significant and positive causality from NUC to CO₂ only in the sub-periods 1993, demonstrating that NUC intensified environmental damage in 1993. However, in the sub-sample period 2019, the Ho hypothesis of “no causality from NUC to CO₂ is dismissed. The negative causality suggests that NUC intensifies the quality of the environment. On the flip side, there is causality from CO₂ to NUC in sub-sample periods 1998, 2009, and 2013–2014 (Figure 7 panel b). In addition, the causality in the sub-sample periods 1998, 2009, and 2013–2014 are positive and significant (Figure 7 panel d). These results back up the NUC and CO₂ feedback hypothesis in the United Kingdom. These results are not surprising because the quantity of electricity generated by nuclear power in the United Kingdom has decreased since the 1990s. Nuclear power offered nearly 25% of the United Kingdom's electricity in the late 1990s⁶. Since then, a lot of installations have been officially closed, while others are now closing more frequently for repair since they are old. Nuclear power facilities will provide 16% of the UK's electricity by 2020. Since the capacity of reactors scheduled for dismantling soon surpasses the number of new reactors authorized for construction, it is anticipated that the UK's nuclear power capacity will decline in the short and medium term. Over half of the United Kingdom's present nuclear capacity will be retired by 2025. The government recently said that additional nuclear power is key to reaching the government's carbon reduction targets. The government has set aside funds to help achieve its nuclear objectives. In 2021, the UK spent £1.7 billion to facilitate a final investment decision for a large-scale nuclear plant. Additionally, the government launched the £120 million Future Nuclear Enabling Fund to remove entry hurdles to the sector.

Figure 8 reveals the bootstrap rolling causality test outcomes for the United States. During the investigation, we noticed weak causality between NUC and CO₂. The results refute the Ho hypothesis of “no causality from NUC to CO₂” at a 10% level of significance (Figure 8 panel a) in the sub-periods 2016. Moreover, there is significant and negative causality from NUC to CO₂ only in the sub-periods 2016, demonstrating that NUC intensified environmental quality in 2016 (Figure 8 panel c). On the flip side, there is causality from CO₂ to NUC in sub-sample periods 1997, 2005, and 2010 (Figure 8 panel b). In addition, the causality in the sub-sample periods 1997, 2005, and 2010 are positive and significant (panel d). These results back up the NUC and CO₂ feedback hypothesis in the United States. These results back up the NUC and CO₂ feedback hypothesis in the United States. The United States is the globe's largest generator of nuclear power, responsible for more than 30% of global nuclear electricity output⁷. Nuclear power presently accounts for more than half of the country's carbon-free electricity, and the untimely shutdown of existing nuclear power facilities would put the United States well behind schedule in fulfilling its decarbonization targets. In 2021, the US Department of Energy (DOE) spent a record $1.3 billion on nuclear energy research.

Summary of results

In summary, the current analyses disclosed that in Japan, the United States, France, and Canada, consumption of nuclear energy curbs the emissions of CO₂ in the sub-sample period. As a result, the empirical results of this study support the idea that nuclear energy is needed to prevent the negative effects of CO₂ that lead to climate change and global warming. Nuclear energy is seen as a better alternative source of energy in Japan, the United States, France, and Canada because its development and use are both environmentally and economically beneficial. These outcomes comply with the study of Menyah & Wolde-Rufael,³⁹ who examined the NUC-emissions nexus in the USA. The findings revealed that nuclear energy could assist in reducing CO₂.

Similarly, our findings comply with the study of Saidi & Omri⁴⁰ in OECD countries. According to the results, deploying nuclear technologies will reduce the usage and reliance on oil and coal, and CO₂ will be eradicated. On the flip side, we found that in the sub-sample period in Germany and United Kingdom, consumption of nuclear energy contributed to ecological damage. Moreover, although nuclear energy reduces carbon emissions, nuclear power plants always present a serious risk due to significant cross-country political, social, and economic variations. The findings of this study, which show that using nuclear energy causes the environment to deteriorate, were observed in the examples of Germany and the United Kingdom. These results align with the studies of Sarkodie & Adams⁴⁶ in South Africa. Furthermore, the study also reported that the utilization of nuclear energy contributes to the deterioration of the environment. Similarly, the research of Syed et al.⁴⁷ for Pakistan testified that an intensification in NUC causes the damage of the environment. These outcomes present significant policy intuition for policymakers in the selected developed countries.

Based on these results, it is clear that the lead-lag connection between NUC and CO₂ cannot be analyzed using linear models; instead, nonlinear models, which can specifically account for regime shifts, should be considered. The results show that the lead-lag connection between NUC and CO₂ is unstable. Furthermore, volatile periods behave very differently, and the connection is probably nonlinear, time-varying, and asymmetric, implying that time-series assessments of NUC and CO₂ that do not take these characteristics into account may be inaccurate. Additionally, linear models’ estimations that presume a constant association may be seriously deceptive.

Conclusions and policy recommendations

Conclusions

Nuclear energy is portrayed as low-carbon transitional energy that helps maintain CO₂ levels in the atmosphere, reducing global warming and climate change. Therefore, this study assesses the nexus between CO₂ and NUC utilizing the rolling-window causality and bootstrap full-sample Granger (BFSG) causality tests for selected 6 developed countries (i.e., Japan, Canada, France, United Kingdom, United States, and Germany) using a yearly dataset from 1980 to 2020. The FSG test results show that NUC Granger causes CO₂ in France and UK, whereas CO₂ Granger causes NUC in Germany. Furthermore, there is no causality between CO₂ and NUC in the United States, Japan, and Canada. However, we analyze the predicted model parameters’ stability, taking structural changes into account. The results show that both the short-run and long-run connections between the two series are unpredictable across the sample period. As a result, misleading outcomes would surface using the full-sample causality test Therefore, the study evaluates the causal association between CO₂ and nuclear energy. The current paper utilized the bootstrap rolling window technique. Instead of asserting that permanent causation exists in every time period, the rolling window technique permits the causal relationships to be time-varying. The rolling p-values of the observed LR statistics and the extent of one variable's influence on the other are calculated using the residual-based bootstrap and a rolling window approach. To the authors’ awareness, this is the first empirical analysis to explore the causal linkage between NUC and CO₂ using time-varying techniques. Therefore, this study fills the gap in the literature. The outcomes of the rolling-window causality test disclosed the following: we found empirical proof of unidirectional negative causality from NUC to CO₂ without feedback, suggesting that nuclear energy consumption can aid in mitigating CO₂ for Japan. In contrast, a negative causality in sub-sample periods from NUC to CO₂ surfaced during the sub-sample period, while a positive causality surfaced from nuclear energy to CO₂ in the United States and France sub-sample period. The results indicate that the United States and France can reduce CO₂ by increasing NUC. Furthermore, we found negative feedback causality between NUC and CO_2, implying that nuclear energy can curb emissions and vice versa for Canada. The evidence of a positive unidirectional causality surfaced from NUC to CO_2, implying that NUC's consumption worsened the environment in the sub-sampled period in Germany.

Policy suggestions

According to the results, deploying nuclear technologies will reduce the usage and reliance on oil and coal, and CO₂ will be eradicated. Furthermore, the advancement of the nuclear energy industry aids in modernizing the energy industry. Japan, Canada, the United States, and France should diversify their energy supply to include nuclear energy to minimize emissions. Nevertheless, generating power from nuclear sources necessitates a high level of safety consideration. To avoid unintended incidents with negative health and environmental consequences, management of radioactive waste and nuclear plant construction must be handled with care. Nuclear energy has the prospect of being a viable substitute for traditional energy. Still, it is contingent on the economic categorization and socioeconomic elements that aid in using energy storage for sustainable development.

Nuclear energy has vast market potential and is also very inexpensive. Furthermore, to lessen the danger (radiation) of nuclear energy, the government can impose strict laws and controls to keep radiation exposures for employees and public members below prescribed levels. The governments of Japan, Canada, the United States, and France could support using clean energy (nuclear) to meet their countries’ energy needs. Furthermore, eco-friendly measures can lead to economic development. For Germany and the United Kingdom, a positive causal effect from NUC to CO₂ surfaced, implying that the surge in CO₂ in Germany and United Kingdom is attributed to an increase in nuclear energy. This means that nuclear energy increase emissions due to likely reasons, including improper management practices and nuclear waste. Based on these results, policymakers and governments in both countries should re-strategize their policies regarding nuclear energy by increasing their share in their energy mix.

Footnotes

Abbreviations

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 authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by the National Natural Science Foundation of China (71973011).

ORCID iDs

Muhammad Irfan

Jinyang Cai

Notes

Muhammad Irfan works at the School of Management and Economics, Beijing Institute of Technology. He is serving as a Guest Editor in several International Journals (i.e. Journal of Environmental Planning and Management, Frontiers in Environmental Science, Frontiers in Energy Research, and Sustainability).

Tomiwa Sunday Adebayo currently works at the Faculty of Economics, Administrative and Social science, Department of Business Administration, Cyprus International University. Tomiwa does research in Environmental Economics, International Economics, Macroeconomics, and Energy Economics

Jinyang Cai works at the School of Management and Economics, Beijing Institute of Technology. His area of interest includes but not limited to, environmental management and sustainability.

Hazar Dördüncü works at the Department of International Trade and Logistics, Turkey. He has published several research papers in reputable journals.

Farrukh Shahzad is an Associate Professor in the Faculty of Economics and Management at the Guangdong University of Petrochemical Technology. His area of interest is green finance, corporate finance, environmental economics, energy economics, and behavioral finance. He published in reputed SSCI/SCI journals such as Sustainable Development, Resources Policy, Environmental Research, International Review of Finance and Economics, and an International Review of Financial Analysis.

References

Tang

Irfan

Razzaq

, et al. Natural resources and financial development: Role of business regulations in testing the resource-curse hypothesis in ASEAN countries. Resources Policy 2022; 76: 102612

Islam

Irfan

Shahbaz

, et al. Renewable and non-renewable energy consumption in Bangladesh: the relative influencing profiles of economic factors, urbanization, physical infrastructure and institutional quality. Renewable Energy 2022; 184: 1130–1149.

IEA. (International Energy Agency) CO2 emissions from fuel combustion. 2022. https://www.iea.org/countries. Assessed September 2 2022.

Kirikkaleli

Adebayo

Khan

, et al. Does globalization matter for ecological footprint in Turkey? Evidence from dual adjustment approach. Environmental Science and Pollution Research 2021; 28: 14009–14017

Adebayo

. Environmental consequences of fossil fuel in Spain amidst renewable energy consumption: A new insights from the wavelet-based Granger causality approach. International Journal of Sustainable Development & World Ecology 2022; 1: 1–14

Awosusi

Adebayo

Altuntaş

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

Kartal

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

Kartal

Kılıç Depren

Ayhan

, et al. Impact of renewable and fossil fuel energy consumption on environmental degradation: Evidence from USA by nonlinear approaches. International Journal of Sustainable Development & World Ecology 2022a: 1–18.

Kartal

Ali

Nurgazina

. Asymmetric effect of electricity consumption on CO₂ emissions in the USA: Analysis of end-user electricity consumption by nonlinear quantile approaches. Environmental Science and Pollution Research 2022b: 1–15.

10.

Dogan

Madaleno

Taskin

, et al. Investigating the spillovers and connectedness between green finance and renewable energy sources. Renewable Energy 2022a; 197: 709–722.

11.

IEA. International Energy Association. 2021. https://www.iea.org/countries. Assessed September 2 2022.

12.

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

13.

Dogan

Madaleno

Inglesi-Lotz

, et al. Race and energy poverty: Evidence from African-American households. Energy Economics 2022b; 108: 105908.

14.

Azam

Rafiq

Shafique

, et al. An empirical analysis of the nonlinear 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

15.

UNECE. Application of the United Nations Framework Classification for Resources and the United Nations Resource Management System: Use of Nuclear Fuel Resources for Sustainable Development -Entry Pathways United Nations Economic Commission for Europe, Geneva (2021). 2021.

16.

Pan

Ibrahim

Al-Faryan

MAS

. Does nuclear energy consumption mitigate carbon emissions in leading countries by nuclear power consumption? Evidence from quantile causality approach. Energy & Environment 2022: 0958305X221112910. https://doi.org/10.1177/0958305X221112910

17.

IEA. Nuclear energy explained: Nuclear power and the environment. Available at. 2020. https://www.eia.gov/energyexplained/nuclear/nuclear-power-and-the-environment.php# Accessed on March 9, 2022.

18.

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

19.

Ahmad

Jiang

Murshed

, et al.

Modelling the dynamic linkages between eco-innovation, urbanization, economic growth and ecological footprints for G7 countries: does financial globalization matter?

Sustainable Cities and Society 2021a; 70: 102881

20.

Dogan

Majeed

Luni

. Revisiting the nexus of ecological footprint, unemployment, and renewable and non-renewable energy for South Asian economies: Evidence from novel research methods. Renewable Energy. 2022c.

21.

Adebayo

Akadiri

Akpan

, et al. Asymmetric effect of financial globalization on carbon emissions in G7 countries: fresh insight from quantile-on-quantile regression. Energy & Environment 2022: 0958305X221084290. https://doi.org/10.1177/0958305X221084290

22.

Ahmad

Jiang

Murshed

, et al.

Modelling the dynamic linkages between eco-innovation, urbanization, economic growth and ecological footprints for G7 countries: does financial globalization matter?

Sustainable Cities and Society 2021b; 70: 102881

23.

World Bank. World Development Indicators. 2021. http://data.worldbank.org/country. Accessed on March 9, 2022.

24.

Balcilar

Ozdemir

Arslanturk

. Economic growth and energy consumption causal nexus viewed through a bootstrap rolling window. Energy Economics 2010; 32: 1398–1410

25.

Andrews

DWK

. Tests for parameter instability and structural change with unknown change point. Econometrica 1993; 61: 821–856

26.

Andrews

DWK

Ploberger

. Optimal tests when a nuisance parameter is present only under the alternative. Econometrica 1994; 62: 1383–1414

27.

Hansen

. Testing for parameter instability in linear models. Journal of Policy Modeling 1992; 14: 517–533

28.

Nyblom

. Testing for the constancy of parameters over time. J Am Stat Assoc 1989; 84: 223–230

29.

World Nuclear. World Nuclear Association. 2022. https://world-nuclear.org/information-library/country-profiles.aspx. Assessed 12 August 2022.

30.

Saidi

Ben Mbarek

. Nuclear energy, renewable energy, CO₂ emissions, and economic growth for nine developed countries: Evidence from panel Granger causality tests. Prog Nucl Energy 2016; 88: 364–374

31.

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

32.

Dogan

Chishti

Alavijeh

, et al. The roles of technology and Kyoto protocol in energy transition towards COP26 targets: Evidence from the novel GMM-PVAR approach for G-7 countries. Technol Forecast Soc Change 2022d; 181: 121756.

33.

Liang

Irfan

Ikram

, et al. Evaluating natural resources volatility in an emerging economy: The influence of solar energy development barriers. Resources Policy 2022; 78: 102858.

34.

Akram

Chen

Khalid

, et al. Heterogeneous effects of energy efficiency and renewable energy on economic growth of BRICS countries: A fixed effect panel quantile regression analysis. Energy 2021; 215: 119019

35.

Isik

Ongan

Ozdemir

, et al. The increases and decreases of the environment Kuznets curve (EKC) for 8 OECD countries. Environmental Science and Pollution Research 2021; 28: 28535–28543

36.

Mahmood

Danish

Zhang

. The role of nuclear energy in the correction of environmental pollution: Evidence from Pakistan. Nuclear Engineering and Technology 2020; 52: 1327–1333

37.

Danish

Ulucak

. An empirical investigation of nuclear energy consumption and carbon dioxide (CO₂) emission in India: Bridging IPAT and EKC hypotheses. Nuclear Engineering and Technology 2021; 53: 2056–2065

38.

Bandyopadhyay

Rej

. Can nuclear energy fuel an environmentally sustainable economic growth? Revisiting the EKC hypothesis for India. Environmental Science and Pollution Research 2021. https://doi.org/10.1007/s11356-021-15220-7.

39.

Menyah

Wolde-Rufael

. CO₂ Emissions, nuclear energy, renewable energy and economic growth in the US. Energy Policy 2010; 38: 2911–2915

40.

Saidi

Omri

. Reducing CO₂ emissions in OECD countries: do renewable and nuclear energy matter? Prog Nucl Energy 2020; 126: 103425

41.

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

42.

Hassan

Danish

S-U-Dk

Awais Baloch

, et al. Is nuclear energy a better alternative for mitigating CO₂ emissions in BRICS countries? An empirical analysis. Nuclear Engineering and Technology 2020; 52: 2969–2974

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. Environmental Science and Pollution Research 2021. https://doi.org/10.1007/s11356-021-13728-6

44.

Danish

Erdogan

. 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: 1312–1320

45.

Jin

Kim

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

46.

Sarkodie

Adams

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

47.

Syed

Kamal

Tripathi

. An empirical investigation of nuclear energy and environmental pollution nexus in India: fresh evidence using NARDL approach. Environmental Science and Pollution Research 2021; 28: 54744–54755

48.

Majeed

Ozturk

Samreen

, et al. Evaluating the asymmetric effects of nuclear energy on carbon emissions in Pakistan. Nuclear Engineering and Technology 2022; 54: 1664–1673

49.

Toda

Yamamoto

. Statistical inference in vector autoregressions with possibly integrated processes. J Econom 1995; 66: 225–250

50.

Shukur

Mantalos

. A simple investigation of the Granger-causality test in integrated-cointegrated VAR systems. J Appl Stat 2000; 27: 1021–1031

51.

Dlamini

Balcilar

Gupta

, et al. Revisiting the causality between electricity consumption and economic growth in South Africa: A bootstrap rolling-window approach. International Journal of Economic Policy in Emerging Economies 2015; 8: 169–190

52.

BP. Energy Data. 2022. https://www.bp.com/en/global/corporate/energy-economics/statistical-review-of-world-energy/downloads.html, July 5, 2022.

53.

Zivot

Andrews

DWK

. Further evidence on the great crash, the oil-price shock, and the unit-root hypothesis. J Bus Econ Stat 2002; 20: 25–44

54.

Akadiri

Adebayo

Haouas

, et al. A time-varying analysis between financial development and carbon emissions: Evidence from the MINT countries. Energy & Environment 2022: 0958305X221082092. https://doi.org/10.1177/0958305X221082092

55.

Pesaran

Timmermann

. Small sample properties of forecasts from autoregressive models under structural breaks. J Econom 2005; 129: 183–217

56.

Akhmat

Zaman

Shukui

, et al. The challenges of reducing greenhouse gas emissions and air pollution through energy sources: Evidence from a panel of developed countries. Environmental Science and Pollution Research 2014; 21: 7425–7435

57.

Anwar

Siddique

Dogan

, et al. The moderating role of renewable and non-renewable energy in environment-income nexus for ASEAN countries: Evidence from method of moments quantile regression. Renewable Energy 2021; 164: 956–967

58.

Baek

Pride

. On the income-nuclear energy-CO₂ emissions nexus revisited. Energy Economics 2014; 43: 6–10

59.

Goldemberg

. Nuclear energy in developing countries. Daedalus 2009; 138: 71–80

60.

Iwata

Okada

Samreth

. Empirical study on the environmental kuznets curve for CO₂ in France: The role of nuclear energy. Energy Policy 2010; 38: 4057–4063

61.

Kartal

. Production-based disaggregated analysis of energy consumption and CO₂ emissions nexus: Evidence from the USA by novel dynamic ARDL simulations approach. Environmental Science and Pollution Research 2022b. https://doi.org/10.1007/s11356-022-22714-5

62.

Lee

Kim

Lee

. Analyzing the impact of nuclear power on CO₂ emissions. Sustainability 2017; 9: 1428

63.

Murshed

Ahmed

Alam

, et al. Reinvigorating the role of clean energy transition for achieving a low-carbon economy: Evidence from Bangladesh. Environmental Science and Pollution Research 2021; 28: 67689–67710

64.

Rehman

Radulescu

, et al. The energy mix dilemma and environmental sustainability: Interaction among greenhouse gas emissions, nuclear energy. Urban Agglomeration, and Economic Growth. Energies 2021; 14: 7703

65.

Kang

Yuan

. The prospective of nuclear power in China. Sustainability 2018; 10: 2086