Impact of Afforestation, Energy Productivity, Renewable and Nuclear Electricity Generation on CO₂ Emissions: Empirical Findings from the BRICS Countries

Abstract

Global warming and climate change have considerably enhanced worldwide environmental concerns since the 1970s. Therefore, researchers have extensively researched the nexus between renewable energy utilization and CO₂ emissions in the literature. However, the influence of afforestation and energy productivity along with renewable and nuclear electricity generation on CO₂ emissions has not been explored sufficiently in the associated literature regarding the multiple effects of these actors on the decarbonization process. Thus, this article analyzes the short- and long-term effects of afforestation, energy productivity, renewable and nuclear electricity production on CO₂ emissions in the BRICS states over the 1993–2021 term via robust bootstrap cointegration and causality tests. The findings confirm a cointegration interplay among CO₂ emissions, afforestation, energy productivity, renewable and nuclear electricity generation. Further, the cointegration coefficients demonstrate a negative influence of afforestation, energy productivity, renewable electricity generation on CO₂ emissions in most of the BRICS states in the long term, but a negative effect of nuclear electricity production only in China and the Russian Federation. The findings of causality examination also uncover that afforestation, energy productivity, and generation of renewable and nuclear electricity are effective tools in reducing CO₂ emissions, but their long-term effects are found to be relatively higher than short-term effects. These findings indicate that promotion of afforestation, along with energy productivity and electricity from renewables and nuclear sources is highly useful for curbing CO₂ emissions in the short and long term.

1. Introduction

Climate change has been among the leading problems that the globalized world has confronted for a long time and is giving rise to many economic, environmental, and health problems, such as natural disasters, loss of species, spread of infectious diseases, food insecurity, and poverty. In this context, non-renewables make the largest contribution to climate change by emitting about 68% of worldwide greenhouse gas emissions and around 90% of worldwide CO₂ emissions which are causing global warming [1]. Therefore, Sustainable Development Goals (SDGs) of climate action, affordable and clean energy, life below water, and life on land along with other SDGs within the 2030 Agenda for Sustainable Development in 2015, are crucial to combat climate change [2].

In this regard, worldwide electricity in 2024 is mainly produced from nonrenewable energy sources of coal (about 34.5% of global electricity production) and natural gas (about 21.8% of global electricity production) [3], and over 40% of worldwide CO₂ emissions grow out of the utilization of non-renewables in electricity production [4]. Therefore, utilization of renewables and nuclear sources with relatively lower CO₂ emissions in electricity production can be useful to decrease CO₂ emissions. Further, energy productivity can also decrease CO₂ emissions by means of achieving economic growth with relatively less energy [5]. Last, along with the clean energy development and high energy efficiency technologies, afforestation is a natural tool to mitigate climate change [6]. Thus, this research is implemented to uncover the influence of afforestation, energy productivity, renewable and nuclear electricity production on CO₂ emissions in the BRICS (Brazil, Russian Federation, India, China, and South Africa) states. China, India, Russia, and Brazil, the BRICS states, are amongst the top ten global CO₂ emitters probably due to the fact that these countries are main drivers of global economic growth during the past four decades [7]. However, the BRICS states have made significant progress in the capacity of renewable and nuclear energy generation and energy productivity and have significantly preserved their forest areas during this period [8,9,10,11]. For this reason, this study investigates the role of afforestation, energy productivity, generation of nuclear and renewable energies on mitigation of CO₂ emissions in the BRICS states.

Renewables and nuclear sources are the largest sources of electricity generation with low carbon [12] and the rates of renewables and nuclear sources in 2024 worldwide electricity generation are respectively 32.1% and 9.1% [3]. But the global power mix shows heterogeneity amongst the countries. For example, the electricity generation mix in 2024 is as follows in China, India, and the United States of America:

58.4%—coal, 3.2%—natural gas, 33.9%—renewables, and 4.4%—nuclear (China);

73.4%—coal, 3.3%—natural gas, 20.5%—renewables, and 2.6%—nuclear (India);

15.6%—coal, 42.6%—natural gas, 23.3% renewables, and 17.9% nuclear (United States of America). In conclusion, coal is generally the primary electricity source in the emerging and developing markets, while natural gas and renewables are the key factors in electricity production in advanced economies [3]. However, the countries should rapidly shift to energy sources with low carbon like renewable and nuclear energy to make progress in the global decarbonization process.

Nuclear energy demand has begun to increase in the world owing to environmental and energy security concerns as a result of increasing geopolitical tensions and conflicts, especially in the regions where energy is produced. Nearly 9% of global electricity is already produced from about 440 nuclear energy power reactors, and this corresponds to over 20% of the global low-carbon electricity [13]. IAEA [12] also predicts that the global nuclear operational capacity of 377 GW in 2024 would increase to 561 GW (low case) and 992 GW (high case) by 2050. No CO₂ emissions are generated during nuclear electricity production owing to the non-use of fossil fuels at this stage. But CO₂ emissions are released at other stages of the nuclear electricity generation cycle, such as construction of nuclear power plants, extraction and conversion of raw materials, depending on reactor types and technology choice [14,15]. Therefore, the influence of nuclear electricity generation on CO₂ emissions can vary among the countries. Thus, Lee et al. [16], Petruška et al. [17], and Petach [18] uncovered a negative impact of nuclear energy on CO₂ emissions, while Bozkaya et al. [19] revealed a positive effect of nuclear energy on CO₂ emissions. Furthermore, Mahmood [20] and Soto and Martinez-Cobas [21] obtained mixed results from the research on the nexus between nuclear energy and CO₂ emissions.

The major renewables utilized in electricity generation are hydropower, solar sources, and wind. The renewable electricity production was about 9900 TW in 2024, and the rate of hydropower, wind and solar sources were respectively 45.12%, 25.45%, and 21.40% [22]. Furthermore, IEA [3] predicts that renewable electricity generation would increase to 16,200 TWh in 2030 from 9900 TWh in 2024. In this regard, renewables such as hydropower, wind, hydro, solar, and geothermal sources considerably decrease greenhouse gas emissions when compared with non-renewables of coal, oil, and natural gas. But electricity generation process from renewables may also have negative environmental implications, such as habitat disruption, land use, and ecosystem degradation during their life cycles [23]. Therefore, the net impact of electricity generation from renewables on CO₂ emissions is unclear. Thus, Ng et al. [24], Suri et al. [25], and Silva et al. [26] unraveled a negative impact of renewable electricity generation on CO₂ emissions, while Maslyuk and Dharmaratna [27] and Suh and Joo [28] uncovered both negative and positive effects of renewable electricity production over CO₂ emissions.

Additionally, forests are one of the significant tools in combat with climate change and global warming. In this regard, trees remove CO₂ emissions from atmosphere and absorb CO₂ in forest soil and release oxygen through the photosynthesis process [29,30,31]. Furthermore, trees can cool the air by releasing moisture. Therefore, deforestation can also cause more intense heat waves [32]. Also, forests can contribute to climate change by emitting significant amounts of CO₂ emissions from the stored carbon in case of forest fires, deforestation, and wood harvest [33]. Consequently, the net impact of afforestation over CO₂ emissions can differ. But Mighri et al. [30] and Kocoglu et al. [34] have identified a negative impact of afforestation over CO₂ emissions.

In the associated literature, academics have usually explored the connection between overall renewable energy utilization and CO₂ emissions, and a small number of academics have examined the environmental influence of renewable and nuclear electricity generation along with afforestation and energy productivity. Furthermore, a consensus on the effects of afforestation, renewable and nuclear electricity production over CO₂ emissions has not been attained yet after these empirical studies. Therefore, this research is evaluated to contribute to the relevant literature in the light of these aforementioned considerations. The next part of this research summarizes the empirical studies on the association among afforestation, nuclear and renewable electricity generation, energy productivity, and CO₂ emissions. Section 3 defines the dataset and methods. Section 4 executes the econometric applications and evaluates the consequences, and Section 5 finalizes the article.

2. Literature Review

Climate change became a fundamental threat to our world through air and water pollution, hotter temperatures, natural disasters, drought, food insecurity, and health risks. Energy-related CO₂ emissions and deforestation have been documented as the significant factors behind climate change. For this reason, this study empirically questions the influence of afforestation, renewable, and nuclear electricity generation along with energy productivity over CO₂ emissions.

In the associated literature, Mighri et al. [30] researched the effect of afforestation on CO₂ emissions in China through spatial methods and discovered that forest investments negatively impacted CO₂ emissions in 30 Chinese provinces. Sheng et al. [35] also uncovered a limited short-term effect of afforestation on CO₂ emissions in 30 Chinese provinces. Similarly, Kocoglu et al. [34] researched the influence of forest area over CO₂ emissions in 181 countries for the years of 1990–2022 by regression and unraveled a negative impact of afforestation over CO₂ emissions.

Furthermore, some researchers researched the effects of deforestation on CO₂ emissions and unveiled a positive relation between CO₂ emissions and deforestation. Assis et al. [36] examined the forest degradation on CO₂ emissions in the Brazilian Amazon between 2006 and 2016 and revealed a positive influence of forest degradation on CO₂ emissions. Ranjan and Gorai [37] also examined the environmental effects of mining-based deforestation between 2000 and 2019 and revealed a significant contribution of deforestation to CO₂ emissions, especially in Indonesia, Brazil, and Canada. Pata et al. [38] researched the influence of forest load capacity factor on CO₂ emissions in India for the years 1990–2021 and identified a negative influence of forest load capacity factor on CO₂ emissions. Similarly, Pata et al. [39] uncovered a negative influence of forest load capacity over CO₂ emissions in 10 states with the largest forests.

In a similar manner, a limited number of researchers have analyzed the association between renewable electricity production and CO₂ emissions, but they have uncovered mixed results. The outcomes of Ng et al. [24], Suri et al. [25], and Silva et al. [26] pointed out the mitigation influence of renewable electricity generation over CO₂ emissions, while Maslyuk and Dharmaratna [27] and Suh and Joo [28] uncovered both positive and negative effects of electricity production from renewables over CO₂ emissions.

Ng et al. [24] examined the effect of electricity generation from fossil fuels and renewables over CO₂ emissions in OECD states during the 1990–2013 years through FMOLS, DOLS, and PMG estimators and D-H causality test. Their outcomes uncovered a negative influence of renewable electricity production over CO₂ emissions, but a positive influence of electricity generation from non-renewables on CO₂ emissions. Also, they identified a bidirectional causal nexus among CO₂ emissions and electricity generation from fossil fuels and renewables. Suri et al. [25] also researched the emissions resulting from wind and solar electricity generation in California and Texas and identified the reductions in CO₂ emissions under normal operating conditions. Silva et al. [26] explored the nexus between renewable electricity and CO₂ emissions in the EU-27 members and the United Kingdom during the years 2007–2022 by regression models and uncovered a mitigation effect of renewable electricity production on greenhouse gas emissions.

Maslyuk and Dharmaratna [27] analyzed the interplay amongst CO₂ emissions and electricity generation from renewables in Asian states with a middle-income level over the 1980–2010 years through the SVAR method. Their analysis indicated that renewable electricity production increased CO₂ emissions in all Asian states except China and Malaysia, but decreased CO₂ emissions in Malaysia and China. Similarly, Suh and Joo [28] analyzed the influence of renewable electricity production over CO₂ emissions in the United States between 1995 and 2023 by means of regression and unveiled mixed outcomes. Their outcomes demonstrated that solar energy could decrease fossil fuel demand, while biomass energy could increase coal demand.

Additionally, most researchers have concentrated on the nexus between renewable energy use and CO₂ emissions. These empirical studies, presented in

Table 1. Recent empirical studies on the association between RNWEN utilization and CO₂ emissions.

Empirical Study	Sample; Period	Methodology	Nexus Between RNWEN Use and CO2 Emissions
Huang et al. [40]	Leading renewable energy-using economies; 2000–2015	Regression	Negative
Uğurlu [41]	Visegrad countries; 2000–2018	FMOLS	Negative
Hao [42]	China; 1990–2020	Cointegration and causality tests	Negative; bidirectional causality
Ofori-Sasu et al. [43]	138 developing countries; 1990–2020	Regression	U-shaped interplay
Aliani et al. [44]	G7 countries; 2000–2019	Regression	Negative
Jie and Rabnawaz [45]	Developing and developed economies; 1970–2022	Regression	Negative
Justice et al. [46]	Ghana; 1990–2020	Regression	Negative
Deng et al. [47]	Developing and developed countries; 2000–2019	Regression	Negative
Almulhim et al. [48]	BRICS countries; 1996–2020	Regression	Negative
Gür et al. [49]	EU transition states; 2000–2021	Cointegration and causality tests	Negative; bidirectional causality
Lorente-de-Las-Casas and Marrero [50]	OECD members; 1990–2019	Event study	Negative
Lojanica et al. [51]	EU-15 members; 1980–2022	PMG-ARDL approaches	Negative
Kara [52]	Türkiye; 1990–2023	ARDL	Negative
Yang and Xu [53]	China	Two-way fixed effects model	Negative
Addis [54]	Middle Eastern and BRICS countries; 1995–2020	Westerlund panel cointegration test, DOLS estimator, and D-H causality test	Negative; bidirectional causality in Middle Eastern countries and one-way effect from RNWEN use to CO2 emissions in BRICS countries

for country samples with different socio-economic characteristics, have confirmed the negative influence of renewable energy employment on CO₂ emissions, despite differences in samples and methodologies.

Furthermore, a very small number of researchers have explored the connection between CO₂ emissions and nuclear energy through different estimators, but the nexus between these two variables has stayed inconclusive based on different socio-economic characteristics of country samples or environmental indicators. On one hand, Lee et al. [16], Petruška et al. [17], and Petach [18] uncovered a negative influence of nuclear energy on CO₂ emissions while Bozkaya et al. [19] specified the positive influence of nuclear energy on CO₂ emissions. Further, Mahmood [20] and Soto and Martinez-Cobas [21] suggested that the relation between CO₂ emissions and nuclear energy differs based on diverse socio-economic characteristics and environmental indicators.

Lee et al. [16] explored the effects of nuclear power proportion over CO₂ emissions in 18 countries, which had 95% of the global nuclear reactors, through regression methodology and uncovered the negative influence of nuclear energy over CO₂ emissions. Petruška et al. [17] also questioned the nexus among CO₂ emissions, renewables, and nuclear energy in the European states between 1992 and 2019 through FMOLS and DOLS estimators and determined a negative effect of nuclear and renewable energies over CO₂ emissions. Additionally, Petach [18] explored the nexus between CO₂ emissions and closures of nuclear power plants in the US between 1993 and 2022 through the difference-in-difference method and regression and found a positive association between nuclear power plant closures and CO₂ emissions.

Bozkaya et al. [19] interrogated the connection amidst CO₂ emissions and nuclear energy in 27 nuclear energy-using countries between 2000 and 2020 through regression and causality methodologies and discovered a positive effect of nuclear energy over CO₂ emissions and a one-way significant causal nexus from CO₂ emissions to the utilization of nuclear energy. Mahmood [20] questioned the effect of nuclear energy over CO₂ emissions in 28 countries for the years of 1996–2019 through cointegration and regression methods and unraveled a negative effect of nuclear energy over CO₂ emissions in upper-middle and high-income states and an insignificant nexus between CO₂ emissions and nuclear energy in lower-middle-income states.

Soto and Martinez-Cobas [21] explored the influence of nuclear sources over CO₂ emissions in the EU members during the 1990–2022 years by means of FMOLS and CCE estimators and D-H causality test, and their findings unveiled that the effect of nuclear sources over CO₂ emissions and ecological footprint was positive and negative, respectively. Furthermore, they identified a unilateral causal nexus from nuclear energy to CO₂ emissions and a one-way significant effect from ecological footprint to nuclear energy. Lastly, Zhang et al. [5], Wahab et al. [55], Safi et al. [56], and Altın [57] respectively disclosed a negative effect of energy productivity over CO₂ emissions in Morocco, G7 countries, E7 countries, and G7 countries.

3. Data and Methods

This paper analyzes the influence of afforestation, renewable and nuclear electricity production, and energy productivity over CO₂ emissions in the BRICS countries for the years of 1993–2021 by way of panel econometrics. The series utilized in the analysis part are demonstrated in

Table 2. Dataset of the study.

Variables	Explanation	Data Source
COEMS	CO2 emissions (tCO2e per capita)	[7]
FOREST	Forest area (% of land area)	[8]
RNWEN	Electricity generation from renewables, excluding hydroelectric	[9]
NUCLEN	Electricity generation from nuclear sources	[10]
ENPRD	GDP per unit of energy utilization (constant 2021 PPP $ per kg of oil equivalent)	[11]

. In this regard, CO₂ emissions were proxied by CO₂ emissions (per capita tCO₂e) and obtained from Climate Watch [7]. Climate Watch [7] uses the data from the International Energy Agency, Food and Agriculture Organization of the United Nations, and US Environmental Protection Agency to calculate country-level CO₂ emissions and provide production-based emissions data (please see the technical note [58] for detailed information). On the other hand, afforestation is proxied by forest area as a percentage of total land area) and sourced from World Bank [8]. Forest area is the rate of land area which is under planted stands of trees or natural vegetation of at least 5 m in situ, and leaves out tree stands in agricultural production and trees in gardens and urban parks [8].

Lastly, electricity production from renewables and nuclear sources is represented by generation of renewable electricity and nuclear electricity as a percent of overall electricity generation, and both series are procured from the World Bank [9,10]. Electricity generation from renewables includes biomass, biofuels, geothermal, solar, wind, and tides. Nuclear electricity generation reflects the electricity produced by nuclear power plants [9,10]. Last, energy productivity is proxied by PPP (purchasing power parity) GDP per unit of energy use. PPP GDP is based on 2021 constant international dollars using PPP rates, and then it was divided by the per kilogram of oil equivalent of energy utilization to obtain energy productivity [11].

The study’s sample comprises BRICS countries. The presence of nuclear energy production as of 1993 for all BRICS countries causes us to identify the beginning of the study period as 1993. Further, the RNWEN series ends in 2021. Therefore, the study’s duration is between 1993 and 2021. In the applied part of this paper, tests of cross-sectional dependence (CSD), homogeneity, unit root, causality test, and AMG (augmented mean group) estimation were executed through Stata 17.0, while the cointegration test was implemented through Gauss 12.0 statistical software.

The summary figures of COEMS, FOREST, RNWEN, NUCLEN, and ENPRD are indicated in

Table 3. Dataset’s summary statistics.

Variables	Mean Value	Standard Deviation	Minimum	Maximum
COEMS	5.288	3.790	0.71	12.65
FOREST	34.211	19.009	14.025	69.101
RNWEN	24.284	17.139	3.2	51.5
NUCLEN	6.550	4.522	1.46	15.44
ENPRD	7.809	3.152	2.885	13.567

. The average figures of COEMS, NUCLEN, RNWEN, FOREST, and ENPRD are respectively 5.288 tCO₂e per capita, 6.550%, 24.284%, 34.211%, and 7.809 PPP $ per kg of oil equivalent. However, the series of RNWEN and FOREST display a considerable change amongst BRICS states, while COEMS, NUCLEN, and ENPRD demonstrate a moderate change amongst BRICS states.

The association amongst afforestation, energy productivity, renewable, and nuclear electricity generation, and CO₂ emissions is respectively examined by Westerlund and Edgerton [59], cointegration test and causality test of Dumitrescu and Hurlin (D-H) [60], seeing the existence of heterogeneity and CSD amongst COEMS, FOREST, RNWEN, NUCLEN, and ENPRD. The bootstrap cointegration test [59] takes into account heterogeneity and CSD and also generates efficient results in small datasets. The test is produced from Equation (1).

$$y_{it}={\alpha }_i+x_{it}{\beta }_{it}+Z_{it}$$

(1)

where $y_{it}=\mathit{COEMS}$; $x_{it}=\mathit{FOREST},\mathrm{NUCLEN},\mathrm{RNWEN},\mathrm{and}\mathrm{ENPRD}$

$$Z_{it}={\mu }_{it}+V_{it}={\sum }_{J=1}^t{ŋ}_{ij}$$

(2)

t and i respectively symbolize the years of 1993–2021 and the BRICS states. $Z_{it}$ is the disturbance term.

The AMG estimator of Eberhardt and Bond [61] was employed regarding its high ability of robust estimation under the existence of CSD and heterogeneity. Further, the AMG estimator enables us to predict country- and panel-level coefficients, unlike regression-based approaches. The estimator uses Equation (3) and takes notice of CSD by attaching a common dynamic effect derived from the period dummy coefficients of a pooled regression with the first differences to the group regressions [61].

$$y_{it}={\beta }_i^{\prime }{x}_{it}+u_{it} u_{it}={\alpha }_i+{\lambda }_i^{\prime }{f}_t+{\epsilon }_{it}$$

(3)

where $u_{it}$ are generated from a composition of ${\alpha }_i$ (group-specific effects), ${\lambda }_i^{\prime }$ (group-specific factor loadings) and $f_t$ (common factors). Last, the causal nexus amongst COEMS, FOREST, RNWEN, NUCLEN, and ENPRD was analyzed through D-H causality test that is developed for heterogeneous panels and also generates reliable consequences in the presence of CSD.

4. Results and Discussion

In the analysis part of this research, LM CSD and homogeneity tests of delta are executed to make a choice amongst current causality and cointegration tests and cointegration estimators. The consequences of LM CSD tests are displayed in

Table 4. CSD and homogeneity pre-tests.

Test	Test Statistic	Test	Test Statistic
LM	20.67 ***	Delta	14.483 ***
LM adj	6.031 ***	Adjusted delta	16.262 ***
LM CD	4.264 ***

*** significant at 1%.

. The H0 hypothesis of CSD independence is declined and the CSD’s presence amongst COEMS, FOREST, RNWEN, NUCLEN, and ENPRD is identified. The entity of heterogeneity is analyzed by way of delta tilde tests, and their consequences are demonstrated in Table 4. The H0 hypothesis of homogeneity’s presence is declined and the existence of heterogeneity is inferred. As a consequence, it is evaluated that the utilization of econometric tests sensitive to CSD and heterogeneity will enhance the reliability of this research’s findings.

The unit roots of COEMS, FOREST, RNWEN, NUCLEN, and ENPRD are examined by Pesaran [62] using the CIPS unit root test owing to the entity of CSD, and their consequences are demonstrated in

Table 5. CIPS test.

Variables	Constant	Constant + Trend
COEMS	0.663	1.142
d(COEMS)	−3.750 ***	−3.263 ***
FOREST	0.327	0.499
d(FOREST)	−4.622 ***	−3.234 ***
RNWEN	−0.759	0.733
d(RNWEN)	−2.918 ***	−4.336 ***
NUCLEN	−0.390	0.217
d(NUCLEN)	−5.961 ***	−2.452 ***
ENPRD	0.444	0.943
d(ENPRD)	−5.246 ***	−4.464 ***

*** significant at 1%.

. All series under consideration include a unit root at level values. But the first-differenced values of COEMS, FOREST, RNWEN, NUCLEN, and ENPRD do not have unit roots.

The cointegration relation amongst COEMS, FOREST, RNWEN, NUCLEN, and ENPRD is questioned by bootstrap cointegration test. The test statistics along with asymptotic and bootstrap probability figures introduced in

Table 6. LM bootstrap cointegration test.

Constant			Constant and Trend
Test Statistic	Asymptotic p-Value	Bootstrap p-Value	Test Statistic	Asymptotic p-Value	Bootstrap p-Value
1.876	0.016	0.546	2.642	0.054	0.724

demonstrate that the null hypothesis of significant cointegration is confirmed for bootstrap probability values but rejected for asymptotic probability values. However, bootstrap probability values are regarded owing to CSD presence, and a significant cointegration interplay amongst COEMS, FOREST, RNWEN, NUCLEN, and ENPRD is concluded.

The long-run coefficients of FOREST, RNWEN, NUCLEN, and ENPRD are predicted by the AMG estimator after identification of significant cointegration amongst COEMS, FOREST, NUCLEN, RNWEN, and ENPRD, and demonstrated in

Table 7. Outcomes of AMG estimation.

BRICS States	FOREST	RNWEN	NUCLEN	ENPRD
Brazil	−1.090 **	−1.166 ***	−0.236	−0.037
China	−0.421 ***	−0.366 **	−0.089 **	−0.735 ***
India	−0.234 ***	−1.296 ***	−0.018	−0.146 **
Russian Federation	−1.246 **	0.934	−0.520 **	−0.374 **
South Africa	−1.246	−0.0451 **	−0.053	−0.324 **
Panel	−0.524 **	−0.451 **	−0.086 ***	−0.172 **

*** and ** are significant at 1% and 5%, respectively.

. The panel coefficients unveil a significant negative influence of afforestation, renewable, and nuclear electricity production, and energy productivity over CO₂ emissions. Further, the coefficients of BRICS states also uncover that afforestation has a negative influence over CO₂ emissions in Brazil, China, India, and the Russian Federation. Similarly, electricity production from renewables has a negative effect on CO₂ emissions in Brazil, China, India, and South Africa, while nuclear electricity production has a negative effect over CO₂ emissions only in China and the Russian Federation. Lastly, energy productivity negatively impacts CO₂ emissions in India, China, South Africa, and the Russian Federation.

The estimation of the error correction model (ECM) is presented in

Table 8. ECM estimation.

Dependent Variable: ΔCOEMSt	β	Sd.	t-Stat	p Value
ΔFORESTt	−0.832	0.076	−10.947	0.000
ΔRNWENt	−0.374	0.043	−8.698	0.000
ΔNUCLENt	−0.093	0.025	−3.720	0.000
ΔENPRDt	−0.244	0.032	−4.642	0.000
ΔECTt−1	−0.318	0.067	−4.746	0.012
Constant	1.109	0.092	12.054	0.005

R² = 0.759, Durbin–Watson = 2.19, Jarque–Bera = 0.273, Harvey test (p-value) = 0.228.

. The coefficient of the error correction term is found to be negative and statistically significant. Therefore, the error correction mechanism of the model is working. In this case, 31.8% of the short-term deviations between long-term convergent series disappear, and the series converge back to their long-term equilibrium values. Furthermore, afforestation, nuclear and renewable electricity generation also negatively impact CO₂ emissions in the short term.

Forests can contribute to the decarbonization process simultaneously through multiple channels by removing CO₂ emissions from the atmosphere, absorbing CO₂ in forest soil, and releasing oxygen through the photosynthesis process. Furthermore, deforestation, wood harvest, and forest fires can also emit CO₂ from the stored carbon. Thus, Assis et al. [36] and Ranjan and Gorai [37] identified a positive effect of deforestation over CO₂ emissions. On the contrary, Mighri et al. [30] and Kocoglu et al. [34] identified a negative influence of afforestation over CO₂ emissions. Therefore, the associated theoretical and empirical results confirm our negative influence of afforestation on CO₂ emissions in the BRICS states.

Electricity production from non-renewables is one of the common factors of worldwide CO₂ emissions. Therefore, generation of electricity from nuclear and renewable sources is expected to support the low-carbon transition. In this connection, renewable sources such as hydropower, solar, and wind are frequently utilized in the electricity production process due to their low greenhouse gas emissions. However, the life cycle of renewable electricity production may also have negatively impact on the environment through habitat disruption, land use, and ecosystem degradation. In the literature, most researchers have examined the nexus between renewable energy consumption and CO₂ emissions, and only a few scholars have already analyzed the nexus between renewable electricity generation and CO₂ emissions and have found mixed consequences. The results of Ng et al. [24], Suri et al. [25], and Silva et al. [26] pointed out the negative influence of renewable electricity generation on CO₂ emissions, while Maslyuk and Dharmaratna [27] and Suh and Joo [28] uncovered both positive and negative effects of renewable electricity production on CO₂ emissions. However, our outcomes show that renewable electricity negatively impacts CO₂ emissions in BRICS countries except the Russian Federation in harmony with the outcomes of Ng et al. [24], Suri et al. [25], and Silva et al. [26].

No CO₂ emissions are produced during nuclear electricity generation, but CO₂ emissions can emerge at the stages of construction of nuclear power plants, extraction and conversion of raw materials, depending on reactor types and technology choice. Therefore, the nexus between CO₂ emissions and generation of nuclear electricity can differ among the countries. The associated empirical studies have uncovered mixed consequences in harmony with these theoretical considerations. On the one hand, Lee et al. [16], Petruška et al. [17], and Petach [18] uncovered a negative influence of nuclear energy on CO₂ emissions, while Bozkaya et al. [19] identified a positive impact of nuclear energy on CO₂ emissions. On the other hand, Mahmood [20] and Soto and Martinez-Cobas [21] have attained mixed outcomes. Our findings demonstrate a negative influence of nuclear electricity generation on CO₂ emissions in China and the Russian Federation and are supported by the results of Lee et al. [16], Petruška et al. [17], and Petach [18].

Last, increases in energy productivity are expected to negatively impact CO₂ emissions due to more output with less energy use depending on share of non-renewables in overall energy consumption of a country. However, Zhang et al. [5], Wahab et al. [55], Safi et al. [56], and Altın [57] discovered a negative influence of energy productivity on CO₂ emissions in Morocco, G7 countries and E7 countries while Karim et al. [63] uncovered a positive effect of energy productivity on CO₂ emissions in Malaysia. Therefore, our negative effect of energy productivity on CO₂ emissions in most of the BRICS states is supported by majority of the related empirical literature.

The causal interplay amongst COEMS, FOREST, RNWEN, NUCLEN, and ENPRD is analyzed by the D-H causality test, and the results are demonstrated in

Table 9. D-H causality test.

Null Hypothesis	W-Bar	Z-Bar	Z-Bar Tilde
FOREST ⇏ COEMS	4.3671 ***	5.3238 ***	4.5262 ***
COEMS ⇏ FOREST	0.8642	0.7246	0.2418
RNWEN ⇏ COEMS	4.7466 ***	5.9240 ***	4.9797 ***
COEMS ⇏ RNWEN	5.5612 ***	7.2119 ***	6.0881 ***
NUCLEN ⇏ COEMS	2.4984 **	2.3691 **	1.9205 *
COEMS ⇏ NUCLEN	2.4761 **	2.3340 **	1.8903 *
ENPRD ⇏ COEMS	3.001 ***	3.165 ***	2.605 **
COEMS ⇏ NUCLEN	2.595 **	2.522 **	2.052 **

***, **, and * are significant at 1%, 5%, and 10%, respectively.

. The outcomes uncover a bidirectional causal nexus between NUCLEN, RNWEN, ENPRD, and COEMS, but a one-way causal nexus from FOREST to COEMS. Therefore, a mutual interplay between NUCLEN, RNWEN, ENPRD, and COEMS exists and FOREST also has a significant influence on COEMS.

A two-way empirical analysis among COEMS, FOREST, ENPRD, NUCLEN, and RNWEN has been carried out by a few researchers. In this connection, Soto and Martinez-Cobas [21] identified a unilateral causal nexus from nuclear energy to CO₂ emissions, but a unidirectional causality from ecological footprint to nuclear energy in the EU members. Bozkaya et al. [19] also disclosed a unilateral causality running from CO₂ emissions to consumption of nuclear energy in 27 nuclear energy-using countries. However, outcomes display significant feedback between nuclear electricity generation and CO₂ emissions in the BRICS states and demonstrate that increasing CO₂ emissions also turn these countries towards nuclear sources. On the other hand, Ng et al. [24] identified a bidirectional causal nexus between renewable electricity generation and CO₂ emissions, which is similar to the findings of our study. Lastly, our one-way causality from afforestation to CO₂ emissions supports the significant effect of afforestation on CO₂ emissions.

5. Conclusions

Climate change is amongst the primary global challenges that our world has faced nowadays and can give rise to many negative environmental, economic, and health-related problems. Therefore, the UN’s sustainable development goals include multiple SDGs related to climate change, such as climate action, affordable and clean energy, life below water, and life on land. In this context, CO₂ emissions resulting from electricity production and deforestation play a significant role in climate change. This study aims to analyze the influence of afforestation, energy productivity, nuclear and renewable electricity generation on CO₂ emissions in the BRICS states.

Our study consists of the following restrictions:

The study period is identified as 1993–2021, as data on renewable electricity generation is present until 2021.

The study employs country-level data due to the absence of provincial-level data for all countries and in turn disregards regional heterogeneity in the countries.

The empirical literature has put forward numerous drivers of CO₂ emissions, but this study has focused on the influence of afforestation, energy productivity, nuclear and renewable electricity generation on CO₂ emissions, employing appropriate tests of cointegration and causality tests sensitive to the characteristics of the research’s dataset.

The results of AMG estimation demonstrate that especially afforestation, energy productivity, and renewable electricity generation have a negative influence in most of the BRICS states, but nuclear electricity generation is an effective tool in decreasing CO₂ emissions only in the Russian Federation and China. On the other hand, the causality analysis suggests a feedback interplay between energy productivity, nuclear and renewable electricity generation and CO₂ emissions, but a one-way effect from afforestation to CO₂ emissions.

Based on our findings, the following policies would be suggested for BRICS states to make progress in low-carbon transition:

BRICS countries usually have experienced increases in forest area, but forest area in terms of overall land area is already low, especially in South Africa, China, and India. Therefore, countries should support afforestation through financial and regulatory incentives considering that afforestation is a complementary and natural tool to decrease CO₂ emissions.

Energy productivity is found to be a significant instrument to decrease CO₂ emissions. Therefore, human capital investments in development of energy-efficient technologies should be incentivized especially by Brazil and South Africa.

The share of nuclear electricity generation in total electricity generation is very low in all BRICS states probably due to financial and technological constraints. Therefore, infrastructure of nuclear power generation should be financially supported, but harmful environmental effects of nuclear energy generation should be minimized by regulations.

Furthermore, renewable electricity generation seems to be an effective tool to decrease CO₂ emissions, but further investments are required to increase the share of renewable electricity consumption in total electricity consumption.

The empirical literature indicates that environmental effects of both nuclear and renewable electricity generation and afforestation have been explored by few researchers and their consequences have stayed inconclusive. For this reason, future studies can be focused on influence of both afforestation and green electricity generation on CO₂ emissions through provincial-level data. Further, future studies can analyze the nexus between renewable energy and CO₂ emissions by functionally decomposing the renewable energy.