The role of renewable energy in the energy– growth–emission nexus in the ASEAN region

Background The Association of Southeast Asian Nations (ASEAN) relies mainly on fossil fuels in their energy supply, leading to higher CO 2 emissions, pollution, and further environmental degradation. This paper uses the panel vector autoregressive and the Granger non-causality test in the heterogeneous panels, together with long-run estimation techniques, to examine the dynamic link among energy consumption, economic growth, and carbon emissions with the focus on renewable energy for the ASEAN countries in the past three decades. Results The findings from this paper indicate that carbon emissions are associated with energy consumption. In contrast, renewable energy usage reduces CO 2 emissions, improving environmental quality. Economic growth is associated with increased energy consumption and carbon emissions in the ASEAN countries. The findings also indicate that the effects of energy consumption on economic growth are more significant than those of renewable energy in ASEAN. When considered together, these findings form a vicious circle regarding the energy–growth–emission nexus for the ASEAN economies. In addition, a bidirectional Granger causality among energy consumption, economic growth, CO 2 emissions and renewable energy usage is confirmed. Conclusions Renewable energy has emerged as an important viable option for the ASEAN nations to achieve their dual objectives of enhanced economic growth, reduced CO 2 emission, leading to improved environmental quality.


Background
Countries in the Association of Southeast Asian Nations (ASEAN) region have achieved significant economic growth and social transformation in the past three decades.Together with the increased population, the overall energy supply in the region increased by approximately 80 per cent between 2000 and 2020.During this period, coal consumption climbed by six, and its proportion of the overall energy supply increased from 8 per cent to 26 per cent [1].Oil consumption has surged by more than 40 per cent since 2000, despite its proportion of the overall energy supply decreasing from 40 per cent to 32 per cent.Natural gas use increased by more than 80 per cent between 2000 and 2020, and it now accounts for almost 20 per cent of overall energy consumption [1,2].Specifically, the power and industrial sectors account for 70 per cent of total natural gas use today.Traditional biomass usage as a cooking fuel has steadily declined over the last 20 years, with overall use halving.During the period, the energy provided by current renewable types more than quadrupled.However, solar PV and wind energy have grown fast in recent years.Modern bioenergy, geothermal energy, and hydropower account for more than 98 per cent of all modern renewable energy in Southeast Asia today.Geothermal resources are mostly found in Indonesia and the Philippines; Cambodia, Lao PDR, and Myanmar have continued to develop domestic hydropower resources, leveraging their steep terrains and heavy precipitation [1] (Fig. 1).
With global worries about climate change and carbon lock-in growing, Southeast Asia has a unique possibility to progress its economy and global leadership by building a regional low-carbon electrical infrastructure.The ASEAN nations have vast solar energy resources and excellent manufacturing capabilities for battery energy storage and electric cars, making them ideal candidates to lead the worldwide transition to clean energy [2].Following the Paris Agreement, several ASEAN nations have amended their power development plans to incorporate aggressive pledges to decarbonize the power sector.ASEAN member nations have promised to produce at least 23 per cent of their power from renewable sources by 2025.Vietnam has authorized over 11 gigatons (GW) of new wind projects, while Thailand is creating 2.7 GW of floating solar [2].Following negotiations at the COP26 in 2021, five major Southeast Asian countries set carbon neutrality goals for 2050 and plans to phase out coal.These initiatives are designed to give ASEAN nations the tools to increase and divert investments toward renewables and innovative energy technology [3,4].
The interaction among energy consumption, economic growth, and carbon emissions in different parts of the world has attracted significant attention from policymakers, practitioners, and academics in the past three decades.This concern is relevant in the ASEAN region.Over previous decades, the ASEAN region has achieved impressive economic growth and social transformation.Supporting economic growth requires energy consumption, which relies heavily on fossil fuels in the ASEAN countries.As a result, the energy demand in the region is expected to grow as much as 2.3 times (or 230 per cent) over the long-term projections to 2040.Unlike fossil fuels, renewable energy releases less CO 2 emissions, thus mitigating the negative impact on the environment.Renewable energy from solar and wind is generally considered unlimited resources.Renewable energy ensures increased energy security, sustainable economic growth, and improved environmental quality [5].The ASEAN countries have considered that renewable energy and energy efficiency bring a potential benefit to reduce the reliance on fossil fuels, ensuring targeted economic growth and achieving energy security, affordability, and sustainability.
Various studies have examined the inter-relationship between economic growth, energy consumption and renewable energy in the ASEAN region [6][7][8].However, the focus of these previous studies is different.Saboori and Sulaiman [6] examine the cointegration and causal relationship between economic growth, carbon emissions, and energy consumption for selected ASEAN countries from 1971 to 2009 using the autoregressive distributed lag (ARDL) methodology and Granger causality test based on vector error-correction model (VECM).However, renewable-a key pillar in energy policy for the ASEAN region has been ignored in their analysis.Vo et al. [7] examine the dynamic link between CO 2 emissions, energy consumption and renewable energy consumption.However, their study tests the validity of the region's long-run environmental Kuznets curve (EKC).Vo and Vo [8], focusing on renewable energy and population, examine a causal link between them.Their study appears to focus more exclusively on the region's population growth.The findings from these studies have motivated us to conduct an empirical analysis with a focus on the role of renewable energy in the energy-growth-emission nexus, in the long run, using state-of-the-art estimation techniques for the ASEAN region.Anwar et al. [9] examine the moderating role of renewable and nonrenewable energy in the environment-income nexus for ASEAN countries using the novel method of moments quantile regression for ASEAN countries.Their findings confirm that non-renewable energy consumption stimulates carbon emissions across all quantiles, and renewable energy consumption decreases CO 2 emissions across all quantiles.Yan and Uprasen [10] investigated the carbon neutrality potential of the ASEAN-5 countries with a focus on the asymmetric effects of income inequality on renewable energy consumption using a nonlinear panel autoregressive distributed lag (ARDL) model from 1990 to 2015.They found that alleviating income inequality promotes renewable energy consumption in the long run and vice versa.In addition, their findings reveal that the positive shock (worsening of inequality) of income inequality generates a larger impact on renewable energy consumption than the result from the negative shock (improvement of inequality).
The contributions of our paper to the existing literature on the energy-growth-emission nexus, particularly for the ASEAN region, are twofold.First, previous studies confirm the vital role of renewable energy in boosting economic growth and reducing carbon emissions.However, those studies mainly focus on the EU countries and the American regions.The ASEAN region has largely been underexamined in the current literature.In addition, empirical studies focusing on the ASEAN region tend to examine the relationship between carbon emissions, energy consumption, and economic growth.The role of renewable energy has largely been underexamined.Second, governments of the ASEAN countries such as Indonesia, Thailand and Vietnam have made strong commitments to relying on renewable energy usage in the national energy policy to reduce pollution, leading to improved environmental quality.As such, understanding the important role of renewable energy in the energygrowth-emissions nexus, particularly in the long run, is essential.Findings from this study provide additional evidence to support the governments in the ASEAN region to formulate and implement policies and strategies for recognizing the important role of renewable energy in the national energy strategy.
The paper is structured as follows.Following this background, the next section discusses relevant theories, empirical studies, and the causal relationship between energy consumption, economic growth, and CO 2 emission.The research methodology is then presented.The next section presents empirical results and discussions of these findings.The conclusions are then discussed in the last section of the paper.

The environmental Kuznets curve hypothesis
The relationship between economic growth and environmental degradation has recently gained significant attention among scholars, focusing on testing the environmental Kuznets curve (EKC) hypothesis.The hypothesis considers that economic growth initially leads to environmental damage and improves environmental quality after a specific threshold of economic growth.However, the validity of the hypothesis has been challenged.
Saboori et al. [11] used the ARDL to investigate the EKC hypothesis for Malaysia from 1980 to 2008.The study confirms an inverted U-shaped relationship between income and CO 2 emission in the long run.Heidari et al. [12] also examine the relationship between economic growth and CO 2 emission in the five ASEAN countries (including Indonesia, Malaysia, the Philippines, Singapore, and Thailand) using panel smooth transition regression (PSTR).The authors conclude a nonlinear relationship between CO 2 emissions, energy consumption, and economic growth.Furthermore, the study finds that CO 2 emissions increased with economic growth and decreased if the per capita GDP was above USD 4686.Aslan et al. [13] used the bootstrap rolling window estimation of the VAR model for the US from 1966 to 2013.Their findings confirm that the effect of economic growth on CO 2 emissions increased from 1982 to 1996 and decreased from 1996 to 2013.As such, their findings confirm the validity of the U-shaped EKC hypothesis.
On the other hand, other empirical studies provide evidence against the EKC hypothesis.For example, Al-Mulali et al. [14] employ the ARDL to investigate the causal relationship between CO 2 emission and Vietnam's economic growth from 1981 to 2011.They argue that economic growth is positively related to pollution in the short and long run, suggesting rejecting the EKC hypothesis.While previous studies that find evidence for the ECK hypothesis assume a quadratic relationship between economic growth and CO 2 emission, Fakih and Marrouch [15] show that a non-parametric model might confirm a positive relationship between CO 2 emission and GDP without a turning point.The study uses data from the Middle East and North African countries from 1980 to 2010 and finds evidence to reject the ECK hypothesis.Zambrano-Monserrate et al. [16] use an ARDL framework to compare the long-run and short-run elasticities of CO 2 emission to economic growth in Peru from 1980 to 2011.The findings show that the impact of economic growth is positive and more robust in the long run than in the short run, rejecting the validity of the EKC hypothesis.
In other studies, mixed evidence is found regarding the EKC hypothesis.For example, Saboori and Sulaiman [6] investigated the EKC hypothesis for selected countries in the Association of Southeast Asian Nations (ASEAN) over the period 1971-2009 using the autoregressive distributed lag (ARDL) approach and Granger causality test based on vector error-correction model (VECM).The study confirms an inverted U-shaped relationship between income and CO 2 emissions when energy consumption is disaggregated based on various sources such as gas, oil, coal, and electricity.However, the evidence for the EKC hypothesis can only be found in Singapore and Thailand, while the results for Indonesia, Malaysia, and the Philippines do not support the hypothesis.The author argues that the outcomes are expected because the selected countries have different economic development stages.Similarly, Le and Quah [17] examine the CO 2 emission-growth nexus for 14 selected countries in the Asia Pacific region from 1984 to 2012 using the fully modified OLS estimators (FMOLS).The study finds evidence for the ECK hypothesis for the highincome economies (including Hong Kong and South Korea) and evidence against the inverted U-shaped between CO 2 emission and economic growth for middle-income countries (Thailand, China, Indonesia, and Pakistan).Malaysia and the Philippines are the only middle-income countries whose evidence supports the ECK hypothesis.
Some studies even support an N-shaped relationship between CO 2 emission and economic growth.Churchill et al. [18] argue that investigating panel data and individual countries may yield different outcomes.The authors find supportive evidence for the ECK hypothesis using the mean group estimators for the panel data.However, only 9 of 20 OECD countries confirm the ECK hypothesis when the authors investigate individual countries.Among these nine countries, six of them show a second turning point.The resurging of CO 2 emission, when income reaches the second turning point suggests an N-shaped rather than a U-shaped relationship between CO 2 emission and economic growth.Shahbaz, Haouas, and Hoang [19] test the possibility of an N-shaped curve using the cube term of economic growth, ARDL bound testing, and VECM Granger causality test for Vietnam over the 1974-2016 period.The existence of the N-shaped curve is confirmed in the long run but not in the short run.

The causality relationship among energy consumption, economic growth, carbon emission, and renewable energy usage
Ang [22] investigates the relationship between economic growth, CO 2 emission, and energy use in Malaysia from 1971 to 1999 using ECM-based causality tests.The author finds a unidirectional causality from economic growth to energy consumption.The study also finds weak evidence for the feedback effect in the long run.Azlina and Mustapha [23] present a unidirectional causality from CO 2 emission to energy consumption in Malaysia from 1970 to 2010.The results are robust for both the long run and short run.Hwang and Yoo [24] use the ECM to test the causality between CO 2 emission and energy consumption in Indonesia.They show that there is a bidirectional causality between these variables.Finally, Saboori and Sulaiman [6] employed the VECM to detect the causality link between CO 2 emission and energy consumption for five ASEAN countries (including Indonesia, Malaysia, the Philippines, Singapore, and Thailand) from 1971 to 2009.Long-run causality from energy consumption to CO 2 emission is found in all selected ASEAN countries except Indonesia.In contrast, the short-run causality is found in all except Thailand at a 5 per cent significance level.They conclude that a bidirectional causality between these variables is found in the short run for only Indonesia, Thailand, and Singapore.In contrast, the causality for all these five selected countries is found in the long run.
Shahzad et al. [25] confirm an inverted U-shaped relationship between energy consumption and CO 2 emission in Pakistan from 1971 to 2011, using an ARDL bound tests framework.The authors capture the threshold effect of energy consumption where the impact of energy consumption on CO 2 emission is positive below a certain threshold and becomes negative when energy use passes a turning point.When energy consumption is below the threshold, the economy expands and consumes more energy, leading to more CO 2 emissions (scale effect).On the other hand, when economic development and energy consumption are above the threshold, more efficient technologies are applied, leading to less pollution (technology effect).
The sources of energy consumption play an essential role in the relationship between energy consumption and CO 2 emission.Al-Mulali et al. [14] find a significantly positive relationship between fossil fuel energy consumption and CO 2 emission in Vietnam.The authors show that renewable energy consumption and CO 2 emission are negatively correlated.Similar results are found by [26] for Pakistan from 1970 to 2016, using an ARDL approach.In contrast, Raza and Shah [27] found a significantly negative link between CO 2 emission and renewable energy for the G7 countries (including Canada, France, Germany, Italy, Japan, the UK, and the US) for the 1991-2016 period using the DOLS, FMOLS, and the fixed effects technique.Solarin, Al-Mulali, and Ozturk [28] confirm a negative impact of hydroelectricity on CO 2 emission in China and India from 1965 to 2013, utilizing the ARDL bound testing approach.
Research also emphasizes the causal relationship between CO 2 emission and economic growth.A seminal work is by [22], who found causality from CO 2 emission to economic growth in the long run, without feedback effects, for Malaysia from 1971 to 1999.After that, various studies have been conducted to determine how CO 2 emission and economic growth interact.Bekhet and Othman [29] show a bidirectional causality relationship between these two variables in the short-run in Malaysia over the 1971-2015 period using the VECM Granger causality test.Chandran and Tang [30] conducted a comprehensive investigation to examine the effect of income on CO 2 emission for each of the 5 ASEAN countries-Malaysia, the Philippines, Singapore, Thailand, and Indonesia.Their findings found a bidirectional causality between economic growth and CO 2 emission for Indonesia and Thailand in the long run.A unidirectional causality from economic growth to CO 2 emission is also found in the short run.Moreover, they confirm a unidirectional relationship is from economic growth to CO 2 emission for Malaysia and the reversed effect for the Philippines in the short run.Singapore has a bidirectional causality between economic growth and CO 2 emission.Also, within the VECM framework, Saboori and Sulaiman [6] find evidence for long-run bidirectional Granger causality relationships between economic growth and CO 2 emission for Indonesia, Malaysia, and the Philippines.The same study found short-run bidirectional relationships for Indonesia, Singapore, and Thailand.Sulaiman and Abdul-Rahim [31] confirm the important role of economic growth and energy consumption in increasing CO 2 emission in Malaysia from 1975 to 2015, employing the ARDL approach and VECM Granger causality framework.

Method
Our main objective is to investigate the important role of renewable energy in the energy-growth-emission nexus in the long run and their causality relationship for the ASEAN countries in the past three decades.Previous studies have been considered in forming our regression model [21,[32][33][34][35][36][37][38][39][40][41][42][43][44].Our regression specification is expressed as follows: where i and t represent the number of the country and the period, respectively.PCCO 2it represents the per capita CO 2 emissions and PCGDPR it and PCGDPR 2  it are the real per capita gross domestic product (GDP) and its squared term.PCEC it denotes the per capita energy consumption.PCREC it represents per capita renewable energy usage.POP it is the population.Finally, ε it is the error term, and Ln denotes the nature of the logarithm.All variables are expressed in the form of the logarithm.As such, the estimated coefficients can be interpreted in terms of the elasticity, showing the per cent of CO 2 emissions per capita changes with a 1 per cent change in the independent variables.The descriptive statistics for the variables used in our analysis are presented in Table 1 below.
Table 2 presents the correlation matrix between variables used in our analysis.The results indicate a weak correlation between real per capita GDP, the square of real per capita GDP, and per capita CO 2 emissions at the first differences.The estimated approach used in this study relies on the first differences.As such, multicollinearity does not matter.

Results
We use the panel unit root test by [45], including the inverse Chi-square, inverse normal, inverse logit and modified inverse Chi-square test, to examine the stationarity of the variables in our sample.The results are presented in Table 3, indicating that all variables are integrated I(1).
(1)  The cointegration test is now performed to examine a long-run relationship.We use two tests Westerlund [46] and Pedroni [47] proposed.Results are reported in Table 4.No cointegration relationship is found among the selected variables.Thus, we conclude that no longrun relationship exists among the variables of interest.
All variables used in our model are I(1).We now examine the dynamic effects among these variables using the PVAR framework.The procedure begins with the selection of optimal lag length.In Table 5, three criteria of MBIC, MAIC and MQIC indicate that the first lag is optimal consistently.We use one lag for all variables in the PVAR model.
Figure 2 presents the plotted eigenvalues within the unit circles, implying that the stability of the PVAR estimation is valid.In addition, Table 6 presents our empirical results on variance decomposition.
Figure 3 from a to d illustrates a response of a shock of CO 2 emissions, renewable energy consumption, energy consumption, and economic growth to the shock of itself and other variables.For example, Fig. 3a illustrates that CO 2 emissions respond positively to a shock of energy consumption and are negatively affected by a shock of renewable energy consumption and economic growth.This means increased energy usage goes hand in hand with CO 2 emissions.
Figure 3b shows that its shock to other variables and from other variables is statistically significant for the response of renewable energy consumption.

Table 5 PVAR's optimal lags length
J statistic and corresponding p-value is based on [48], and other moment model selection criteria (MMSC) are developed by [49].Bayesian information criterion (MBIC), MMSC-Akaike information criterion (MAIC), and MMSC-Hannan and Quinn information criterion (MQIC)   Specifically, renewable energy responds positively to a shock from CO 2 emissions and adverse to a shock from economic growth and population growth.As shown in Fig. 3c, the energy consumption response revealed that the responsive patterns are entirely different regarding the energy consumption response.Energy consumption experiences a significantly negative response to renewable energy consumption and population growth shocks.A shock of economic growth affects energy consumption, while CO 2 emissions are observed to be positive and statistically significant.As shown in Fig. 3d, economic growth has a statistically positive response to population growth.It responds negatively to a shock of energy consumption rather than Fig. 3 The results of impulse response functions (IRFs).The responses of the selected variables are represented by the solid line, while the two standard error bands are covered with the two shaped lines.The response is statistically significant when the bands do not cross the zero line renewable energy consumption.This finding indicates that the impact of energy consumption on economic growth seems to overtake that of renewable energy.Table 7 presents our results regarding the forecast error variance decomposition, which provides us with the degree of the responses of one variable to the other remaining variables.We find that a change in economic growth in the ASEAN countries significantly causes variations in CO 2 emissions.Particularly, economic growth explains approximately 19 per cent of the variation in CO 2 emission after four periods.This degree reaches approximately 24 per cent after 12 years.Interestingly, our results indicate that renewable energy contributes approximately 8 per cent and 1 per cent to the variation of carbon emissions and economic growth.All these findings highlight the marginal contributions of renewable energy in supporting economic growth in the ASEAN countries.Findings also indicate that renewable energy usage contributes approximately 8 per cent to the variation of carbon emissions.
Our study also uses two long-run estimation techniques, including the fully modified OLS (FMOLS) and the dynamic OLS (DOLS), to investigate the effects of renewable energy on carbon emissions in the long run for the ASEAN countries.However, before these estimate techniques are implemented, we have conducted the preliminary tests, including the cross-sectional dependence and the slope homogeneity tests, to ensure that using the FMOLS and the DOLS techniques is appropriate for our analysis.
After the stationarity test is conducted, as presented in Table 2 above, the next step is to examine the crosssectional dependence in the sample used in our analysis.Table 7 below presents our empirical findings regarding this test.Variables used in our analysis are highly significant at a 1 per cent level.These results confirm the cross-sectional dependence among the variables used in our analysis.These findings also confirm that the residuals are auto-correlated.Based on these findings, using the FMOLS and the DOLS techniques is appropriate for our analysis.
The next step to support the use of the long-term estimation techniques, including the FMOLS and the DOLS, is to investigate the slope homogeneity and a long-run relationship among variables using Pesaran and Yamagata [50] test (the slope homogeneity) and the Westerlund and Edgerton [51] test (for the cointegration test).Empirical results from these two tests are presented in Table 8 below.We find that slope heterogeneity and long-run cointegration exist in the sample used in our analysis.In addition, we also find that a long-run relationship among the variables also exists in our analysis.These findings reconfirm our view that using the FMOLS and the DOLS techniques is appropriate for our analysis.
Table 9 presents the empirical results on the long-run effects of renewable energy and economic growth on carbon emissions in the ASEAN countries.

Discussion
As presented in Table 9, our empirical results indicate that the effects are highly significant at a 1 per cent confidence level.The results are consistent across different estimation techniques, including the FMOLS and the DOLS.Economic growth and population are associated with carbon emissions in the ASEAN countries in the long run.These findings indicate that supporting economic growth in the region results in higher CO 2 emissions in the long run.Population growth in the region will also increase carbon emissions in the ASEAN region.Interestingly, our empirical results confirm that renewable energy will contribute to reducing carbon emissions in the long run.As such, encouraging renewable energy usage in the ASEAN countries appears to be a viable option for the countries in the region to achieve their dual objectives of supporting economic growth and limiting higher CO 2 emissions in the long term.
Results from Table 10 confirm the causality relationship between CO 2 emissions, renewable energy usage, energy consumption, and economic growth in the short run.Moreover, the results confirm the bidirectional Granger causality among each pair of selected variables, except for a unidirectional causality from CO 2 emissions to population.Interestingly, there is a bidirectional Granger causality among CO 2 emissions, economic growth, energy consumption and renewable energy.
The empirical results from this study are relatively different from previous studies examining the energygrowth-emissions nexus for the ASEAN region using panel data.Our studies confirm the effects of energy consumption, economic growth, and renewable energy usage on carbon emissions in the ASEAN countries in the past three decades.Previous studies only confirm the effects of energy consumption and economic growth on carbon emissions [19,20].The differences can be explained by the data period and the estimation techniques.Our study focuses more on the long-term effects to provide additional policy implications for the governments of the ASEAN countries.In addition, our causality analysis confirms a vicious circle between energy consumption, economic growth, and renewable energy.We find a bidirectional relationship between these important variables, highlighting the importance of renewable energy in the energy mix for the ASEAN countries to achieve their dual objectives of continuing to support economic growth and limiting CO 2 emission.Previous studies for the ASEAN countries indicate a bidirectional causality relationship between economic growth and energy consumption (the feedback hypothesis).However, they only confirm the unidirectional causality from economic growth to CO 2 emissions (the conservation hypothesis).

Conclusions
Renewable energy has emerged as an important energy source for the ASEAN countries to consider in achieving their dual objectives of supporting economic growth and limiting CO 2 emission.Previous studies have widely investigated the dynamic relationship between carbon emissions, economic growth, and energy consumption.However, the focus on the important role of renewable energy in the energy-growth-emissions nexus for the ASEAN region in the past three decades has been underexamined.As such, this study is conducted to revisit this important relationship with a focus on renewable energy for the region.
The findings from this paper indicate that carbon emissions are associated with energy consumption.In contrast, carbon emissions are negatively related to renewable energy.These findings confirm that using energy from fossil fuel sources will further deteriorate the environmental quality as it will increase carbon emissions in the ASEAN region.In contrast, our findings indicate that renewable energy reduces CO 2 emissions, improving environmental quality.Economic growth is associated with increased energy consumption and carbon emissions in the ASEAN countries.These findings, when considered together, form a vicious circle for the ASEAN economies.Economic growth requires energy consumption.Increased energy consumption leads to an increase in energy demand from fossil fuel sources, leading to higher CO 2 emissions.In this circle, only renewable energy usage in the energy mix of the ASEAN countries can support continued economic growth and reduce the negative effects on environmental quality.The causality tests confirm a bidirectional Granger causality among CO 2 emissions, economic growth, energy consumption and renewable energy usage.These findings from the causality tests further confirm the vicious circle among them.These findings from the bidirectional causality have provided vitally important policy implications for the governments of the ASEAN countries and other emerging markets.First, an interactive impact among carbon emissions, economic growth, energy consumption and renewable energy usage implies that any economic plans or strategies on economic growth will lead to the trade-offs between carbon emissions and energy consumption.Energy consumption is required to support economic growth, leading to CO 2 emissions.As such, renewable energy may supplement fossil fuel energy sources in the energy mix in the transitional period.It then becomes the substitute in response to higher demand for energy consumption, resulting from economic growth.This energy source contributes to economic growth in the ASEAN region.Yet, there is still an environment-related concern regarding economic growth and energy consumption.Recent targets and commitments regarding renewable energy usage among the ASEAN members are on the right track.Efforts to increase the share of renewable energy in the total primary energy supply should be encouraged.Alternative solutions towards taking technological advances in adopting and using renewable energy could be considered to reduce carbon emissions.Second, the governments should continue encouraging firms to produce non-energy intensive and environmentally friendly goods and adopt strict environmental regulations) to discourage firms from using less-friendly-with-theenvironment technology.All in all, the governments of the ASEAN countries may need better coordination to implement policies and strategies focusing on the important role of renewable energy in combating climate change.The intergovernmental bodies such as the ASEAN Center for Energy, an intergovernmental organization within the ASEAN structure representing the 10 ASEAN Member States' interests in the energy sector, should be tasked with policy agenda regarding extending renewable energy use in the energy mix of all ASEAN countries, considering the current standing of each ASEAN country in their particular energy mix, economic growth level, budget constraints and their social characteristics.These policy briefs ensure that proposed policies can be implemented with positive spillover effects across all ten countries in the region.

Fig. 2
Fig. 2 Roots of the companion matrix.The PVAR models are stable as all the AR roots lie inside the unit circle, indicating variables are covariance stationary

Table 1
The descriptive statistics PCCO 2 : per capita CO 2 emissions; PCGDPR: per capita real GDP; PCGDPR 2 : the square of per capita real GDP; POP: population; REC: renewable energy consumption; EC: energy consumption."Ln" represents the logarithm.Fossil fuels include oil, coal, and gas.Renewable energy includes nuclear, hydro, wind, solar, geothermal, biomass in power and other renewable sources

Table 2
Correlation matrixPCCO 2 : per capita CO 2 emissions; PCGDPR: per capita real GDP; PCGDPR 2 : the square of per capita real GDP; POP: population; REC: renewable energy consumption; EC: energy consumption."Δ" denotes the variable in terms of the first differences.Fossil fuels include oil, coal, and gas.Renewable energy includes nuclear, hydro, wind, solar, geothermal, biomass in power and other renewable sources

Table 3
[45]empirical results for the unit-root tests based on Choi[45] ***, ** and * denote the rejection of containing unit roots at the significance level of 1, 5 and 10 per cent.The test includes an intercept and a trend PCCO 2 : per capita CO 2 emissions; PCGDPR: per capita real GDP; PCGDPR 2 : the square of per capita real GDP; POP: population; REC: renewable energy consumption; EC: energy consumption."Ln" represents the logarithm

Table 4
Cointegration tests *** denotes the rejection of containing unit roots at the significance level of 1 per cent

Table 6
Variance decomposition of LnPCCO 2 PCCO 2 : per capita CO 2 emissions; PCGDPR: per capita real GDP; PCGDPR 2 : the square of per capita real GDP; POP: population; REC: renewable energyconsumption; EC: energy consumption.Ln represents the logarithm.Fossil fuels include oil, coal, and gas.Renewable energy includes nuclear, hydro, wind,solar, geothermal, biomass in power and other renewable sources

Table 7
Results from Pesaran's test for cross-sectional dependenceAll variables are natural logarithms.The symbol *** denotes the significance at the 1 per cent level.CD test represents Pesaran's test statistics for cross-sectional dependence with the null hypothesis of cross-sectional independence

Table 8
Results from tests for slope homogeneity and cointegration

Table 9
The long-run effects of renewable energy and economic growth on carbon emissions in the ASEAN countries All variables are in the form of natural logarithms.FMOLS stands for fully modified OLS; DOLS stands for dynamic OLS.Standard errors are reported in parentheses.Symbol *** represents significant level at 1 per cent

Table 10
Empirical results from the Granger causality tests PCCO 2 : per capita CO 2 emissions; PCGDPR: per capita real GDP; PCGDPR 2 : the square of per capita real GDP; POP: population; REC: renewable energyconsumption; EC: energy consumption.Ln represents the logarithm.Fossil fuels include oil, coal, and gas.Renewable energy includes nuclear, hydro, wind, solar, geothermal, biomass in power and other renewable sources