How economic growth in Australia reacts to CO2 emissions, fossil fuels and renewable energy consumption

Patrícia H. Leal (Department of Management and Economics, University of Beira Interior, Covilhã, Portugal)
Antonio Cardoso Marques (NECE-UBI, Department of Management and Economics, University of Beira Interior, Covilhã, Portgual)
Jose Alberto Fuinhas (NECE-UBI, Department of Management and Economics, University of Beira Interior, Covilhã, Portgual)

International Journal of Energy Sector Management

ISSN: 1750-6220

Publication date: 5 November 2018

Abstract

Purpose

Australia is one of the ten largest emitters of greenhouse gases but stands out from the others due to its economic growth without recession for 26 consecutive years. This paper aims to focus on the energy-growth nexus and the effects of energy consumption on the environment in Australia.

Design/methodology/approach

This analysis is performed using annual data from 1965 to 2015 and the autoregressive distributed lag model.

Findings

The paper finds empirical evidence of a trade-off between economic growth and carbon dioxide (CO2) intensity. The results show that increased gross domestic product (GDP) in Australia increased investment in renewable energy sources (RESs), although the renewable technology is limited and has no impact on reducing CO2 intensity in the long run. In contrast to investment in RES, fossil fuels, coal and oil, are decreased by GDP. However, oil consumption increased renewable energy consumption, and this reflects the pervading effect of the growing economy.

Originality/value

Overall, this paper contributes to the literature by analysing the behaviour of both energy consumption and the environment on the growing Australian economy. In addition, this paper goes further by studying the impact of economic growth on renewable and non-renewable energy consumption, as well as on CO2 emissions. The study is conducted on a single country for which literature is scarce, using a recent approach and a long time period.

Keywords

Citation

Leal, P., Marques, A. and Fuinhas, J. (2018), "How economic growth in Australia reacts to CO2 emissions, fossil fuels and renewable energy consumption", International Journal of Energy Sector Management, Vol. 12 No. 4, pp. 696-713. https://doi.org/10.1108/IJESM-01-2018-0020

Download as .RIS

Publisher

:

Emerald Publishing Limited

Copyright © 2018, Emerald Publishing Limited


Acronyms list

ADF

= Augmented Dickey–Fuller;

ARDL

= Autoregressive distributed lag;

CO2

= Carbon dioxide;

ECM

= Error correction model;

EKC

= Environmental Kuznets curve;

GDP

= Gross domestic product;

GHG

= Greenhouse gases;

GT

= Gigatons;

KPSS

= Kwiatkowski Phillips Schmidt;

Mt

= Million tons;

PP

= Phillips and Perron;

RES

= Renewable energy sources;

UECM

= Unrestricted error correction model;

VIF

= Variance inflation factor; and

ZA

= Zivot and Andrews.

1. Introduction

For several decades, economic growth was considered the only tool for sustainable development, but over the years, environmental quality has been introduced as a crucial variable for sustainable development. According to the Brundtland Report or World Commission on Environment and Development, in 1987, high energy consumption will have worrying environmental consequences due to the carbon dioxide (CO2) emissions released from burning fossil fuels. In the same report, the notion of sustainable development was introduced. This concept corresponds to an approach to development in which present needs are addressed without compromising the needs of future generations. A few years later, in 1992, the Earth Summit was held, followed by the Kyoto Summit in 1997, and more attention was paid to environmental impacts and increasingly noticeable environment degradation.

In general, economic growth requires energy, and its availability puts pressure on environmental quality. This condition raises the question of whether there is always a trade-off between economic growth and environmental quality or if it is possible for economies to keep growing without causing environmental degradation. A reduction of CO2 emissions is often associated with a reduction in the consumption of fossil fuels, and in some countries, these reductions have a negative impact on economic growth. Considering that, a reduction in CO2 emissions is more significant when applied to developing countries (Ito, 2017; Narayan and Narayan, 2010). Overall, CO2 emissions from fossil fuel consumption and industrial processes doubled between 1974 and 2014, from 16.9 gigatons (Gt) to 35.5 Gt (BP, BP Statistical Review of World Energy, 2015, 2015, BP press).

This paper focuses on Australia, which has certain particularities that make the country especially interesting to study. Australia is the sixth largest country in the world and has experienced economic growth without recession for 26 consecutive years (Rank et al., 2017). Simultaneously, it is one of the ten largest emitters of greenhouse gases (GHGs). Australia has a free-market economy, with a high gross domestic product (GDP) per capita and a low poverty level. In the past decade, its economy performed consistently, with an annual economic growth rate between 1.5 and 4.5 per cent (IEA, 2012). Lim et al. (2012) analysed the behaviour of the Australian economy in 2011 and its expected behaviour in 2012.

The energy sector makes a very significant contribution to the Australian economy. In 2012, Australia ranked ninth out of the world’s largest energy producers, and the energy sector represented between 16 and 17 per cent of current GDP (IEA, 2012). Regarding the use of energy sources, Australia is a country with extensive natural resources and fossil fuel reserves. Coal, oil, natural gas, uranium and thorium are among its base resources, and accordingly, petroleum products are the main energy source mostly allocated to the transport sector; this means that the transport sector is very dependent on oil products (IEA, 2018).

With regard to renewable energy sources (RESs), solar and wind are the country’s main natural resources. Since 2012, RESs have increased through the growing renewable capacity from wind and solar energy (IEA, 2018). In terms of emissions, CO2 is the main GHG emitted. In 1990, Australia emitted 26 millions of tons (Mt) of CO2, and emissions increased continuously up to 2005, reaching 372 Mt, then rising only slightly to 374 Mt in 2014 (Rank et al., 2017). The total final energy consumption is motivated by the transport and industry sectors, which represent 40 and 35 per cent, respectively (IEA, 2018).

The main objective of this paper is to study the relationship between economic growth, CO2 emissions and energy consumption in Australia. Consequently, the central questions are: Is there a trade-off in Australia between economic growth and CO2 emissions? What is the impact of energy consumption on GDP and the environment in Australia? What is the impact of specific energy sources? To accomplish the aims of this paper, an autoregressive distributed lag (ARDL) approach was used.

Overall, this paper contributes to the literature by analysing the behaviour of both energy consumption and the environment on the growing Australian economy. In addition, this paper goes further by studying the impact of economic growth on renewable and non-renewable energy consumption, as well as on CO2 emissions. The study is conducted on a single country for which literature is scarce, using a recent approach and a long time period. The main findings are in the long run, a bidirectional causality between GDP and CO2 intensity, between RES and oil consumption and between CO2 intensity and the consumption of coal and oil.

This paper is organised into six sections. Section 2 presents literature review, Sections 3 and 4 set out the data and method used and the results obtained, respectively, and Sections 5 and 6 present a discussion and the conclusions of this paper, respectively.

2. Literature review

The direct relationship between economic growth and energy consumption is traditionally verifiable through four hypotheses. The growth hypothesis represents the unidirectional causality from energy to economic growth (Menyah and Wolde-Rufael, 2010). This means that energy consumption is a determinant factor of economic growth and, consequently, economic growth is a function of energy consumption. The conservation hypothesis portrays the unidirectional causality from economic growth to energy (Mehrara, 2007). This hypothesis implies that an increase in economic growth causes an increase in energy consumption. The feedback hypothesis indicates the bidirectional causality between energy and economic growth. This means that there is a causal interdependence between economic growth and energy consumption (Eggoh et al., 2011; Fuinhas and Marques, 2012). The last is the neutrality hypothesis that expresses the non-causal relationship between energy consumption and economic growth (Menegaki, 2011). This means that any reduction in energy consumption will not affect economic growth, and vice versa. Energy consumption does not represent a significant portion of GDP (Tang and Abosedra, 2014). In addition to the aforementioned, there is another, less-conventional hypothesis, the resource curse. This hypothesis contends that energy consumption has a negative impact on economic growth.

Over the years, as economies have grown, generally speaking, environmental quality has decreased. The first research studies undertaken about the effect of energy consumption on economic growth explored the relationship between energy consumption and economic growth (Kraft and Kraft, 1978). Since then, the relationship between energy consumption and economic growth has been studied to observe the different causalities between the two variables, causality running from economic growth to energy consumption (Cheng and Lai, 1997), causality running from energy consumption to economic growth (Stern, 2000), bidirectional causality between economic growth and energy consumption (Hondroyiannis et al., 2002) and the lack of causality between economic growth and energy consumption (Altinay and Karagol, 2004). This topic was particularly important because of the role that energy consumption plays in economic growth and due to the policy implications invoked. Some years later, environmental quality began to be included in the analysis of the energy-growth nexus, combining economic growth, energy consumption and environmental pollution (Ang, 2007; Soytas et al., 2007; Acaravci and Ozturk, 2010).

Several studies about this topic can be found in the literature, for numerous individual countries or groups of countries, using various methodologies. Chen et al. (2012) analysed the relationship between economic growth and energy consumption based on the conclusions of 174 studies. A summary of the literature can be found for instance in Ozturk (2010), Omri (2014) and Tiba and Omri (2017). While previous articles provide a summary of studies conducted from 1978 to 2014, we give a summary of studies conducted from 2015 to 2018 (Table I). The table indicates the country or countries studied, as well as the method used and the results obtained.

As mentioned before, differing approaches have been used in analysing the relationship between economic growth, energy consumption and environmental pollution. However, as can be seen in Table I, the one most commonly used in recent literature is the ARDL model. This approach is also used in this study. In addition to the approaches shown in Table I, the Kaya identity (Kaya, 1990) can also be used. This methodology indicates the relationship between the main emissions generation sources, GDP per capita, population, energy intensity and carbon intensity.

Considering the specific relationship between economic growth and environmental degradation, the environmental Kuznets curve (EKC) by Grossman and Krueger (1991) was first proposed in 1991. This concept had its origin in the “Inverted-U hypothesis” developed by Kuznets (1955). The EKC explains the relationship between economic growth and CO2 emissions during two different phases. During the first phase, GDP and environmental degradation increase. The second phase begins once a turning point is reached, and environmental degradation starts to decrease, while GDP continues to increase. This concept arose to describe how a country’s pollution level is determined by its development over time (Panayotou, 1993).

The energy-growth nexus has been a central theme of energy economics literature. However, there is no consensus in terms of results. These differing results may arise for several reasons. The results depend on the country or group of countries studied, the different variables used for energy consumption and economic growth and the total combination of variables, the time periods studied, the methods used and whether the data are monthly or annual. Considering the lack of consensus and the volatility of results depending of the factors mentioned earlier, the present research contributes to fill the gap in the literature by using a long time period, using a recent approach with appropriated characteristics for the data and by studying individually a country not studied until now, with interesting and idiosyncratic characteristics previously revealed.

3. Methodology

This section is divided into three subsections. The first one presents the variables and units of measurement used in this study, as well as a summary of data sources and statistics. The second contains a preliminary analysis of the variables. The last provides an explanation of the model used and the tests subsequently applied.

3.1 Data

The time period used was from 1965 to 2015, totalling 51 years. This period was chosen because of the data available. Table II describes the variables, their units of measurement and sources.

The variables COAL and OIL were transformed into percentages of primary energy consumption, and the variable CO2 was transformed into CO2 intensity to reduce the correlation between the variables. To obtain the growth rates of the respective variables by their differenced logarithms, all variables were transformed into their natural logarithms. This transformation also reduced the phenomenon of heteroskedasticity. Table III presents the descriptive statistics.

After transforming the variables and interpreting the descriptive statistics, unit root tests were performed to determinate the integration order of the variables. The variables may be integrated of order zero or one, but cannot be integrated of order two. The results of the unit root tests are presented in next subsection.

3.2 Preliminary analysis

To determine the integration order of the variables, the traditional unit root tests were performed, namely, augmented Dickey–Fuller (ADF) (Dickey and Fuller, 1981), Phillips and Perron (PP) (Phillips and Perron, 1988) and Kwiatkowski Phillips Schmidt (KPSS) (Kwiatkowski et al., 1992). In both the ADF and PP tests, the null hypothesis is that the variable is non-stationary, i.e. there is a unit root. The opposite happens in the KPSS test, in which the null hypothesis is that the variable is stationary. Table IV shows the results of the tests.

From observing Table IV, it is possible to conclude through the employment of the traditional unit root tests, ADF, PP and KPSS that all variables are stationary in their first differences, i.e. they are I(1). However, through the visual inspection of the variables, there arises a suspicion that the variables could have structural breaks and the structural breaks could be a limitation for the traditional unit root tests. Consequently, the unit root test with structural breaks (Zivot and Andrews, 1992) (Zivot and Andrews, ZA) was performed to determine the period where structural breaks could exist (Table V).

Given the existence of structural breaks, the ZA unit root test with structural breaks provides information on the specific period in which they occur. This information is useful for determining whether to apply dummies when the models are being estimated. The characteristics of the data under consideration did not compromise the use of the ARDL approach chosen.

The variance inflation factor (VIF) test was performed and suggested the presence of multicollinearity between LGDP and LOIL. Consequently, models were estimated with both variables and without LOIL to confirm if the existence of multicollinearity would change the results. Comparing the results of these estimates, it was possible to conclude that there was no change in the signs so that multicollinearity would not be a problem for estimating the model with all the variables.

3.3 Method: autoregressive distributed lag

To analyse the short- and long-run relationship between all variables used, the approach chosen was the ARDL model, developed by Pesaran et al. (2001). Bearing in mind the characteristics of the data, a period of 51 years was studied. During such a lengthy period, it is likely that several statistically significant events will have occurred and, as such, they should be identified by testing. Considering that the ARDL model allows for dummies to be applied, thus this procedure was used for controlling for these events in the model estimated. In addition, this approach allows for the separation of short- and long-run effects, which is important for determining the dynamics of the variables. Consequently, it makes possible to confirm implicit causalities between all variables through the existence of long-run relationships of cointegration. Besides that, the ARDL model allows for the analysis of direct and indirect effects in the elasticities and provides unbiased long-run estimation (Ahmad and Du, 2017). Last but not least, the ARDL model also allows for the treatment of endogeneity.

The following equation represents the general unrestricted error correction model (UECM) equivalent to the ARDL bounds test used in the five ARDL models estimated:

(1) DYt=α0+α1TREND+α2DZt+α3p=1kYt-p+α4p=1kZt-p+εt,
where D denotes the first differences of variables, Yt represents all logarithm-dependent variables, Zt represent all logarithm-independent variables, α2i is the short-run coefficient, α3i is the error correction model (ECM), α4i is the long-run coefficient and εt is a white-noise error term.

The reverse models were estimated analysing the optimal number of lags necessary. The significance of the parameters was observed, and the residues were examined to ensure that the estimations were as parsimonious as possible. After estimation of the models, diagnostics tests were performed, namely, the Jarque–Bera normality test (including skewness, kurtosis and Jarque–Bera), the Breusch–Godfrey serial correlation LM test, the ARCH test for heteroskedasticity, the Ramsey RESET test to model specification and the stability tests of CUSUM and CUSUM squares.

The ARDL bounds test (Pesaran et al., 2001) was calculated with the null hypothesis of non-existence of cointegration, which means that there is no long-run relationship. In addition, the short-run semi-elasticities and long-run elasticities were calculated. The calculation is made as follows: semi-elasticities result directly from the coefficients of the model variables in the short run, and the elasticities were calculated through the following expression: [c(var)/c(ECM)]*(-1) = 0. This expression consists of the coefficient number of the respective long-run variable, for instance, c(6), which is divided by the coefficient number of the ECM, for instance c(5), and the ratio multiplied by −1.

4. Results

In this section, the results of estimating the ARDL models and the diagnostic tests to which they were submitted are presented. The ARDL bounds test results are then shown, along with the calculations of the semi-elasticities and elasticities.

Considering the objective of studying the relationships between all the variables, five models were estimated. The results of all the models are presented in Table VI.

After the estimation, diagnostic tests were performed that confirmed the normal behaviour of the residuals, the rejection of serial correlation of first and second order, the homoskedasticity of the residues, the correct specification of the model and the stability of the parameters during the period studied, as shown in Figure 1.

From Table VI, it is possible to conclude that the ECM of all models is within an interval between −1 and 0 and revels a good adjustment velocity.

Regarding the dummies applied in model I-LGDP, the dummy in 1983 can be explained by the liberalisation and deregulation of the economy; 1988 was the year when Australia’s economic growth fell below the average rate of the other advanced economies; and 2009 represented the worst year of economic growth in all the years of consecutive growth. In model II-LOIL, the unit root test with structural breaks reveals a break point in 1990. With respect to model IV-LCO2, on one hand, the consumption of natural gas increased in 1969 and caused an exponential increase in CO2 emissions; on the other hand, a high level of CO2 emissions occurred in 1990, and this year became the base year of the Kyoto protocol. The last model, V-LRES, has dummies in 1983, which was the year in which Australia had less production of renewable energy; 1990 was the year that the Renewable Energy Target encouraged the growth of wind capacity; a break point was detected in the ZA test in 2008; and 2009 was when the Australian Government signed a contract to accelerate energy efficiency.

Considering all the results obtained from the five models, certain results can be highlighted. On one hand, the negative impact of LCO2 on LGDP, as well as of LRES on LGDP and of LGDP on LOIL and LCOAL. On the other hand, the positive impact of LCO2 on LCOAL and LOIL, as well as of LGDP on LCO2 and LRES and of LOIL on LRES. Also of note is the absence of any impact by LRES on LCO2 in the long run (Table VII).

The ARDL bounds test was performed by an analysis of the F-statistic in the Wald test, and the aforementioned null hypothesis was rejected. This meant that there was a long-run relationship between the variables (cointegration).

As previously mentioned, the direct and indirect effects on elasticities, semi-elasticities and elasticities were calculated.

From the results shown in Table VIII, it can be concluded that in model I-LGDP, in the long run, an increase in 1 per cent in LCO2 and LRES causes decrease in LGDP of 1.82 per cent and 0.38 per cent, respectively, and an increase in 1 per cent in LOIL causes an increase of 0.66 per cent in LGDP. In the short run, in percentage points, LCO2 and LRES decrease LGDP by 0.82 and 0.06, respectively. Among the other results, the model IV-LCO2 should be highlighted, in which increases of 1 per cent in LGDP, LCOAL and LOIL create increases in LCO2 of 0.04, 0.32 and 0.18 per cent, respectively. In the short run, variations in LRES, LCOAL and LOIL lead to a decrease of 0.02 and increases of 0.26 and 0.11 in LCO2, in percentage points, respectively.

5. Discussion

On the whole, Australia is a country with a strong economic path, surpassing The Netherlands, in 2017, as the country with longest consecutive number of years without a recession. This makes Australia an attractive subject for investigation. This study makes a deeper analysis of the relationship in both the short and long run, between GDP, CO2 intensity, fossil fuel (coal and oil) consumption and RES consumption in Australia. Figure 2 shows the short- and long-run causalities.

Our findings prove that LCO2 intensity and LRES have caused a slowdown in economic growth, although insufficient to interrupt strong economic activity and continuous growth. This decrease can be explained by the huge investment needed to expand RES capacity and by restrictive energy consumption policies that reduce CO2 intensity and, consequently, LGDP. This effect shows that it is possible for a country to address environmental preoccupations, not only emissions reduction but also mix diversification, while continuing to experience economic growth. Regarding the effects of LGDP on LRES and on LCO2, on one hand, higher GDP leads to higher RES consumption (Saidi and Ben Mbarek, 2016) because with increased GDP, the country invests more in renewable energy. On the other hand, increasing GDP implies more energy consumption, and considering that the renewable technology is limited, the energy consumed are the fossil fuels which increase the CO2 emissions (Bilgili et al., 2016). Despite its growing GDP, Australia has a high level of CO2 intensity, and LRES only decreases it in the short run. LRES causes a decrease in CO2 intensity by avoiding the burning of fossil fuels, given that primary energy consumption remains constant. In the long run, RES has no impact, because the renewable technology used has limited and insignificant potential growth.

With regard to fossil fuels, Australia has extensive reserves. However, based on empirical evidence, this paper confirms that there is a negative relationship between LGDP and LCOAL and LOIL. This effect confirms that Australia intends to diversify its energy mix promoting substitution. With growing LGDP, primary energy consumption increases, and the mix of primary energy consumption increases LRES. In view of this, the country is investing in clean energy and measures to promote energy efficiency to achieve environmental targets. Therefore, with growing LGDP, the shares of LOIL and LCOAL are reduced. In addition to the effect of fossil fuels on the economy, they are also associated with environmental degradation. Fossil fuels are considered the main cause of the high CO2 emissions. The empirical results show that fossil fuel consumption increases LCO2, (Ito, 2017). If primary energy consumption remains constant and the consumption of fossil fuels increases in the mix, CO2 emissions increase.

Australia has defined environmental targets to reduce CO2 emissions by 26-28 per cent by the year 2030 based on 2005 values in accordance with the Paris Agreement. In view of the results obtained, one way to be successful would be to apply policies to reduce coal consumption. A variation of 1 per cent in LCOAL causes an increase of 0.32 per cent in LCO2, and this variable has the greatest impact in both the short and long run. RES can also be used to achieve environmental targets. Therefore, CO2 intensity could be reduced by influencing the fossil fuel and RES consumption, but in different ways. On one hand, LCOAL and LOIL have an impact on CO2 intensity in both the short and long run, while RES only has an impact in the short run. On the other hand, fossil fuel consumption has a positive impact on CO2 intensity, while the impact of RES is negative. Regarding the magnitude of the impact, it is large from fossil fuel consumption, as shown in Table VIII. Considering this, a restriction of coal consumption could have a more relevant reduction in CO2 emissions. However, it is necessary to understand which variables influence RES to promote their impact on the reduction of CO2 emissions. LRES is encouraged by LOIL (Saidi and Ben Mbarek, 2016). This effect could be explained as an effect of a growing economy, in other words, the Australian economy. On one hand, the economy continues to be dependent on oil, and this dependency helps economic growth, and on the other hand, the economy invests in renewable energy. This also explains the positive effect of LOIL on LGDP.

In addition, Australia should invest in energy efficiency measures, specifically tailored to certain economic sectors. For instance, in the transport sector, Australia could invest more in electric vehicle technology and charging infrastructures and in the implementation of car sharing and incentives for electric vehicle adoption. Another example is the residential sector, where Australia could wager on demand-side management measures through a good practice guide to encourage the population to save electricity, to consume larger amounts of electricity in the off peak periods and to invest in efficient home appliances. It was confirmed that Australia needs to slow down its economic growth to achieve better environmental quality and reduce CO2 emissions.

6. Conclusion

This paper analyses the relationship between economic activity through LGDP, energy consumption through LCOAL, LOIL and LRES and environmental degradation through LCO2 and focuses on Australia. With this objective, all relationships were studied, and this meant that five models were estimated with all variables as dependent variables. The ARDL methodology was used to study the dynamics of adjustment for a period from 1965 to 2015. Empirical evidence for Australia remains scarce, which leads to the main aim of this research. In fact, it is important to examine the energy-growth nexus question in a country that has had no recession for several consecutive years and increasingly experienced economic growth.

The results of this study can be organised in the relationship between energy consumption of the different sources; the relationship between energy consumption and the environment; the hypothesis of the relationship between economic growth and energy consumption; and the relationship between economic growth and environmental degradation. Considering that, the results reveal a substitution effect between the fossil fuel sources. In addition, RES has a positive effect on fossil fuel consumption, and in turn, RES is given incentive by oil consumption. Regarding the environment, CO2 intensity increases with an increase in fossil fuel consumption, while it could be reduced by increasing RES consumption.

This paper finds evidence for the feedback hypothesis between economic growth and both oil and RES consumption. Moreover, the conservation hypothesis is supported between economic growth and coal consumption. Concerning the relationship between economic growth and CO2 intensity, the results are quite different. Indeed, economic growth increases CO2 intensity, while CO2 intensity has a negative impact on GDP. In other words, in Australia, there is a trade-off between economic development and environmental quality. Overall, the finding of implicit causalities in the ARDL models revealed a strong consistency.

To achieve its environmental goals, Australia should change its energy mix, i.e. it should rethink the relative consumption of the different energy types to reduce CO2 emissions without changing the amount of primary energy consumed. This means that Australia should increase the share of renewable energy and reduce the share of fossil fuels. Other alternatives would be applying policies to restrict fossil fuels consumption, particularly coal; energy efficiency measures; and investing in RES technology, to increase RES consumption.

Figures

Diagnostic tests

Figure 1.

Diagnostic tests

Short- and long-run causalities

Figure 2.

Short- and long-run causalities

Summary of empirical studies on the energy-growth nexus

Authors and year Country(ies) Period Methodology Main findings
Alshehry and Belloumi (2015) Saudi Arabia 1980-2011 VECM Unidirectional causality from EC to economic growth and CO2 emissions and bidirectional causality between CO2 emissions and economic growth in the LR. Unidirectional causality from CO2 emissions to EC and economic output in SR
Arvin et al. (2015) G-20 countries 1961-2012 Panel VAR On a sample of developing countries in the G-20: unidirectional causality from GDP to CO2, in the long-run. On a sample of developed countries in the G-20: unidirectional causality from CO2 to GDP, in the long-run. On a sample of all G-20 countries: unidirectional causality from GDP to CO2, in the long-run
Jammazi and Aloui (2015) Six countries of the Gulf Cooperation Council 1980-2013 WWCC Bidirectional causality between EC and GDP. Unidirectional causality from EC to CO2 emissions
Saidi and Hammami (2015) 58 countries 1990-2012 GMM CO2 emissions and GDP have a positive impact on EC
Vidyarthi (2015) India, Pakistan, Bangladesh, Sri Lanka and Nepal 1971-2010 VECM Unidirectional causality and bidirectional causality from energy consumption pc to GDP pc in the SR and LR, respectively
Bouznit and Pablo-Romero (2016) Algeria 1970-2010 ARDL EC increases CO2 emissions.
EKC hypothesis is verified
Kais and Sami (2016) 58 countries 1990-2012 GMM GDP has a positive impact on CO2 emissions
EKC hypothesis is verified
Saidi and Ben Mbarek (2016) Nine developed countries 1990-2013 FMOLS Unidirectional causality in the SR and bidirectional causality in the LR from RES to real GDP per capita. Unidirectional causality from GDP to CO2 emissions, in the LR
Streimikiene and Kasperowicz (2016) 18 European Union countries 1995-2012 Panel cointegration Test, FMOLS and DOLS Positive relationship between energy consumption and economic growth
Wang et al. (2016) China 1995-2012 FMOLS Bidirectional causality between GDP and EC, and between EC and CO2 emissions. Unidirectional causality from economic growth to CO2 emissions
Ahmad and Du (2017) Iran 1971-2011 ARDL Energy production has a positive effect on GDP. CO2 emissions have a positive effect on GDP
Antonakakis et al. (2017) 106 countries 1971-2011 Panel VAR Bidirectional causality between GDP and EC.
EKC hypothesis is not verified
Bekhet et al. (2017) Countries of the Gulf Cooperation Council 1980-2011 ARDL Long-run and causal relationship between CO2 emissions, GDP and EC in all GCC countries except United Arab Emirates (UAE). And long-run unidirectional causality from CO2 emissions to EC in the case of Saudi Arabia, UAE, and Qatar
Destek and Aslan (2017) 17 emerging economies 1980-2012 Bootstrap panel Granger causality Growth hypothesis for Peru. Conservation hypothesis for Colombia and Thailand. Feedback hypothesis for Greece and South Korea. Neutrality hypothesis for the other 12 economies
Ito (2017) 42 developing countries 2002-2011 GMM NRE consumption leads to a negative impact on GDP for developing countries. RES consumption positively contributes to GDP in the long-run
Mirza and Kanwal (2017) Pakistan 1971-2009 ARDL Bidirectional causality between EC, GDP and CO2 emissions
Appiah (2018) Ghana 1960-2015 TY and Granger Causality test Feedback Granger causality between CO2 emissions and EC
Balsalobre-Lorente et al. (2018) Five European Union countries 1985-2016 PLS N-shaped EKC relationship between economic growth and CO2 emissions
Cai et al. (2018) G7 countries 1965-2015 ARDL Clean EC causes real GDP pc for Canada, Germany and the US and CO2 emissions provoke clean EC for Germany.
Feedbacks between clean EC and CO2 emissions for Germany, and unidirectional causality from clean EC to CO2 emissions for the USA
Gozgor et al. (2018) 29 OECD countries 1990-2013 PQR RES and NRE consumption affect positively the economic growth
Magazzino (2018) Italy 1960-2014 ARDL EC is affected by real GDP, an increase in the real GDP has a significant impact on EC in the LR
Shahbaz et al. (2018) China, USA, Russia, India, Japan, Canada, Germany, Brazil, France and South Korea 1960Q1-2015Q4 QQ Relationship between economic
growth and EC mostly positive for all countries
Tugcu and Topcu (2018) G7 countries 1980-2014 NARDL Asymmetric relationship between EC and economic growth in the LR
Wang et al. (2018) 170 countries 1980-2011 VECM On the global panel, bidirectional causality between CO2 emissions and EC, CO2 emissions and economic growth, EC and economic growth in the SR and LR

Notes: (a) DOLS = dynamic ordinary least squares; (b) EC = energy consumption; (c) FMOLS = fully modified ordinary least squares; (d) GMM = generalised method of moments; (e) LR = long-run; (f) NARDL = nonlinear autoregressive distributed lag; (g) NRE = non-renewable energy; (h) pc = per capita; (i) PLS = panel least squares; (j) PQR = panel quantile regression; (l) QQ = quantile-on-quantile; (m) SR = short-run; (n) TY = Toda-Yamamoto; (o) VAR = vector autoregressive; (p) VECM = vector error correction model; (q) WWCC = wavelet window cross correlation

Variables

Variable description Description Source
GDP Gross domestic product(constant LCU) World Development Indicators
RES Renewable energy (Mtoe) BP Statistical Review of World Energy, 2016
CO2 CO2 emissions (Mt) BP Statistical Review of World Energy, 2016
OIL Oil consumption (Mt) BP Statistical Review of World Energy, 2016
COAL Coal consumption (Mtoe) BP Statistical Review of World Energy, 2016

Notes: (a) Mtoe: millions of tons in oil equivalent; (b) Mt: million tons; (c) constant LCU: local currency unit; and (d) L: natural logarithm

Descriptive statistics

Variables Mean Maximum Minimum SD JB Obs.
LCOAL_P 3.7003 3.8835 3.5390 0.0831 2.4551 51
LCO2_INT 1.1490 1.1924 1.1020 0.0186 1.7052 51
LRES 1.2779 2.0935 0.5460 0.3442 0.6421 51
LGDP 27.337 28.1135 26.456 0.4827 2.7379 51
LOIL_P 3.6834 3.9699 3.4726 0.1643 5.7805 51

Notes: (a) Max.: maximum; (b) Min.: minimum; (c) Std. Dev.: standard deviation; (d) JB: Jarque–Bera; and (e) Obs: observations

Results of unit root tests

Variables ADF PP KPSS
I T N I T N I T
LGDP −1.7990 −2.6421 14.2031 −1.6744 −2.6421 12.0914 0.9611*** 0.0673
DLGDP −5.4008*** −5.6800*** −1.3466 −5.4156*** −5.7010*** −1.5646 0.2166 0.0920
LCO2_INT −1.6513 −1.6358 −1.7915* −1.9183 −1.9621 −1.4008 0.2311 0.1204*
DLCO2_INT −5.3548*** −1.0798 −5.1663*** −5.4972*** −5.4541*** −5.3255*** 0.1696 0.1650**
LOIL_P −0.6118 −1.0854 −1.5696 −0.7778 −1.7416 −1.1608 0.8291*** 0.1773**
DLOIL_P −6.0662*** −6.1142*** −5.6723*** −6.0669*** −6.1512*** −5.7174*** 0.1742 0.1673**
LCOAL_P −1.7393 −1.7607 −0.6837 −2.2237 −2.2424 −1.1364 0.1285 0.1252*
DLCOAL_P −4.5911*** −4.5390*** −4.6059*** −4.6362*** −4.5207*** −4.6311*** 0.2031 0.1834**
LRES −0.5169 −1.7920 2.4179 −0.5594 −2.0052 2.3338 0.8751*** 0.1027
DLRES −6.7163*** −6.6413*** −5.9224*** −6.7188*** −6.6446*** −6.0472*** 0.1107 0.1064
Notes:

(a) I = Intercept; (b) T = Trend and Intercept; (c) N = None; (d)

***

= 1%; (e)

**

= 5%; (f)

*

= 10%; (g) D = first differences; (h) ADF = augmented Dickey-Fuller; (i) PP = Phillips–Perron; (j) KPSS = Kwiatkowski–Phillips–Schmidt–Shin

Results of Zivot and Andrews unit root tests (four lags)

Variables I Break point T Break point B Break point
LGDP −4.5734 1998 −3.9416 1993 −4.5681 1998
LCO2_INT −4.0289 2007 −5.0246*** 2006 −4.8569* 2004
LOIL_P −3.8606 1980 −5.2880*** 1990 −4.9039* 1991
LCOAL_P −3.7892 2007 −4.1738* 2003 −4.0073 2002
LRES −3.6329 1987 −4.6758** 2008 −4.7460 2008
Notes:

(a) I = Intercept; (b) T = Trend; (c) B = Both; (d)

***

= 1%; (e)

**

= 5%; (f)

*

= 10%

ARDL estimation

Variables Model I
(LGDP)
Model II(LOIL) Model III(LCOAL) Model IV(LCO2) Model V(LRES)
D(LCO2_INT) −1.1625*** 4.2904*** 3.1891*** −6.4745***
D(LRES) −0.0923*** 0.0572* 0.0488** −0.0191***
D(LCOAL_P) −1.3965*** 0.2580***
D(LOIL_P) −0.4796*** 0.1114***
LGDP(-1) −0.2565*** −0.3239*** −0.0761*** 0.0194*** 0.4917***
LCO2_INT(-1) −0.6006*** 2.9657*** 1.7466*** −0.5040*** −3.5083***
LRES(-1) −0.1189*** −0.6280***
LCOAL_P(-1) −0.8863*** −0.5642*** 0.1603***
LOIL_P(-1) 0.2083*** −0.4799*** −0.3440*** 0.0919*** 0.3847***
C 6.7896*** 10.2912*** 3.4236*** −0.8821*** −10.0016***
@TREND 0.0130*** 0.0074***
ECM −0.2565*** −0.4799*** −0.5642*** −0.5040*** −0.6280***
Dummies
DU_1969 −0.0067***
DU_1983 −0.0750*** −0.1631***
DU_1988 0.0311***
DU_1990 0.0379*** −0.0057***
DU_1991 0.0644**
DU_2008 −0.1080**
DU_2009 −0.1625***
SD_2009 −0.0384***
Notes:

(a)

***

= 1%; (b)

**

= 5%; (c) DU: impulse dummy; and (d) SD: stability dummy

ARDL bounds test

Models F-statistic k Value
Bottom Top
Model I (LGDP) 9.7786*** 3 5.17 6.36
Model II (LOIL) 8.1340*** 3 5.17 6.36
Model III (LCOAL) 8.0068*** 3 4.29 5.61
Model IV (LCO2) 10.453*** 3 4.29 5.61
Model V (LRES) 11.4589*** 3 4.29 5.61
Notes:

(a)

***

= 1%; (b) critical values of Pesaran et al. (2001); (c) K: number of long-run variables

Semi-elasticities and elasticities

Semi-elasticities andElasticities Model I(LGDP) Model II(LOIL) Model III(LCOAL) Model IV(LCO2) Model V(LRES)
Semi-Elasticities
LCO2_INT −0.8213** 4.2904*** 3.1891*** −6.4745***
LRES −0.0570* 0.0572* 0.0488** −0.0191***
LCOAL_P −1.3965*** 0.2580***
LOIL_P −0.4796*** 0.1114***
Elasticities
LGDP −0.6748*** −0.1350*** 0.038545*** 0.7830***
LCO2_INT −1.8189* 6.1793*** 3.0960*** −5.5865***
LRES −0.3812**
LCOAL_P −1.8466*** 0.3180***
LOIL_P 0.6622*** −0.6096*** 0.1823*** 0.6126***
Notes:

(a)

***

= 1%; (b)

**

= 5%; (c)

*

= 10%

References

Acaravci, A. and Ozturk, I. (2010), “On the relationship between energy consumption, CO2 emissions and economic growth in Europe”, Energy, Vol. 35 No. 12, pp. 5412-5420.

Ahmad, N. and Du, L. (2017), “Effects of energy production and CO2 emissions on economic growth in Iran: ARDL approach”, Energy, Vol. 123, pp. 521-537.

Alshehry, A.S. and Belloumi, M. (2015), “Energy consumption, carbon dioxide emissions and economic growth: the case of Saudi Arabia”, Renewable and Sustainable Energy Reviews, Vol. 41, pp. 237-247.

Altinay, G. and Karagol, E. (2004), “Structural break, unit root, and the causality between energy consumption and GDP in Turkey”, Energy Economics, Vol. 26 No. 6, pp. 985-994.

Ang, J.B. (2007), “CO2 emissions, energy consumption, and output in France”, Energy Policy, Vol. 35 No. 10, pp. 4772-4778.

Antonakakis, N., Chatziantoniou, I. and Filis, G. (2017), “Energy consumption, CO2 emissions, and economic growth: an ethical dilemma”, Renewable and Sustainable Energy Reviews, Vol. 68, pp. 808-824.

Appiah, M.O. (2018), “Investigating the multivariate granger causality between energy consumption, economic growth and CO2 emissions in Ghana”, Energy Policy, Vol. 112, pp. 198-208.

Arvin, M.B., Pradhan, R.P. and Norman, N.R. (2015), “Transportation intensity, urbanization, economic growth, and CO2 emissions in the G-20 countries”, Utilities Policy, Vol. 35, pp. 50-66.

Balsalobre-Lorente, D., Shahbaz, M., Roubaud, D. and Farhani, S. (2018), “How economic growth, renewable electricity and natural resources contribute to CO2 emissions?”, Energy Policy, Vol. 113, pp. 356-367.

Bekhet, H.A., Matar, A. and Yasmin, T. (2017), “CO2 emissions, energy consumption, economic growth, and financial development in GCC countries: dynamic simultaneous equation models”, Renewable and Sustainable Energy Reviews, Vol. 70, pp. 117-132.

Bilgili, F., Koçak, E. and Bulut, Ú. (2016), “The dynamic impact of renewable energy consumption on CO2 emissions: a revisited environmental Kuznets curve approach”, Renewable and Sustainable Energy Reviews, Vol. 54, pp. 838-845.

Bouznit, M. and Pablo-Romero, M.D.P. (2016), “CO2 emission and economic growth in Algeria”, Energy Policy, Vol. 96, pp. 93-104.

Cai, Y., Sam, C.Y. and Chang, T. (2018), “Nexus between clean energy consumption, economic growth and CO2 emissions”, Journal of Cleaner Production, Vol. 182, pp. 1001-1011.

Chen, P.Y., Chen, S.T. and Chen, C.C. (2012), “Energy consumption and economic growth-New evidence from Meta analysis”, Energy Policy, Vol. 44, pp. 245-255.

Cheng, B.S. and Lai, T.W. (1997), “An investigation of co-integration and causality between energy consumption and economic activity in Taiwan”, Energy Economics, Vol. 19 No. 4, pp. 435-444.

Destek, M.A. and Aslan, A. (2017), “Renewable and non-renewable energy consumption and economic growth in emerging economies: evidence from bootstrap panel causality”, Renewable Energy, Vol. 111, pp. 757-763.

Dickey, D. and Fuller, W. (1981), “Likelihood ratio statistics for autoregressive time series with a unit root”, Econometrica, Vol. 49 No. 4, pp. 1057-1072.

Eggoh, J.C., Bangake, C. and Rault, C. (2011), “Energy consumption and economic growth revisited in African countries”, Energy Policy, Vol. 39 No. 11, pp. 7408-7421.

Fuinhas, J.A. and Marques, A.C. (2012), “Energy consumption and economic growth nexus in Portugal, Italy, Greece, Spain and Turkey: an ARDL bounds test approach (1965-2009)”, Energy Economics, Vol. 34 No. 2, pp. 511-517.

Gozgor, G., Lau, C.K.M. and Lu, Z. (2018), “Energy consumption and economic growth: new evidence from the OECD countries”, Energy, Vol. 153, pp. 27-34.

Grossman, G.M. and Krueger, A.B. (1991), “Environmental impacts of a North American free trade agreement”, National Bureau of Economic Research Working Paper Series, Vol. 3914 No. 3914, pp. 1-57.

Hondroyiannis, G., Lolos, S. and Papapetrou, E. (2002), “Energy consumption and economic growth: assessing the evidence from Greece”, Energy Economics, Vol. 24 No. 4, pp. 319-336.

IEA (2012), “Energy policies of IEA countries - Australia, 2012 review”, available at: https://doi.org/10.1787/9789264170841-en.

International Energy Agency (IEA) (2018), “Energy policies of IEA countries: Australia, 2018 review”, p. 244.

Ito, K. (2017), “CO2 emissions, renewable and non-renewable energy consumption, and economic growth: evidence from panel data for developing countries”, International Economics, Vol. 151 No. 40, pp. 1-6.

Jammazi, R. and Aloui, C. (2015), “On the interplay between energy consumption, economic growth and CO2 emission nexus in the GCC countries: a comparative analysis through wavelet approaches”, Renewable and Sustainable Energy Reviews, Vol. 51, pp. 1737-1751.

Kais, S. and Sami, H. (2016), “An econometric study of the impact of economic growth and energy use on carbon emissions: panel data evidence from fifty eight countries”, Renewable and Sustainable Energy Reviews, Vol. 59, pp. 1101-1110.

Kaya, Y. (1990), “Impact of carbon dioxide emission control on GNP growth: interpretation of proposed scenarios”, paper Presented to the IPCC Energy and Industry Subgroup, Response Strategies Working Group

Kraft, J. and Kraft, A. (1978), “On the relationship between energy and GNP”, Journal of Energy and Development, Vol. 3 No. 2, pp. 401-403.

Kuznets, S. (1955), “Economic growth and income inequality”, The American Economic Review, Vol. 45 No. 1, pp. 1-28.

Kwiatkowski, D., Phillips, P.C.B., Schmidt, P. and Shin, Y. (1992), “Testing the null hypothesis of stationarity against the alternative of a unit root”, Journal of Econometrics, Vol. 54 Nos 1/3, pp. 159-178.

Lim, G.C., Chua, C.L., Claus, E. and Nguyen, V.H. (2012), “Review of the Australian economy, 2011-12: a case of déjà vu”, Australian Economic Review, Vol. 45 No. 1, pp. 1-13.

Magazzino, C. (2018), “GDP, energy consumption and financial development in Italy”, International Journal of Energy Sector Management, Vol. 12 No. 1, pp. 28-43.

Mehrara, M. (2007), “Energy consumption and economic growth: the case of oil exporting countries”, Energy Policy, Vol. 35 No. 5, pp. 2939-2945.

Menegaki, A.N. (2011), “Growth and renewable energy in Europe: a random effect model with evidence for neutrality hypothesis”, Energy Economics, Vol. 33 No. 2, pp. 257-263.

Menyah, K. and Wolde-Rufael, Y. (2010), “Energy consumption, pollutant emissions and economic growth in South Africa”, Energy Economics, Vol. 32 No. 6, pp. 1374-1382.

Mirza, F.M. and Kanwal, A. (2017), “Energy consumption, carbon emissions and economic growth in Pakistan: dynamic causality analysis”, Renewable and Sustainable Energy Reviews, Vol. 72, pp. 1233-1240.

Narayan, P.K. and Narayan, S. (2010), “Carbon dioxide emissions and economic growth: panel data evidence from developing countries”, Energy Policy, Vol. 38 No. 1, pp. 661-666.

Omri, A. (2014), “An international literature survey on energy-economic growth nexus: evidence from country-specific studies”, Renewable and Sustainable Energy Reviews, Vol. 38, pp. 951-959.

Ozturk, I. (2010), “A literature survey on energy-growth nexus”, Energy Policy, Vol. 38 No. 1, pp. 340-349.

Panayotou, T. (1993), Empirical Tests and Policy Analysis of Environmental Degradation at Different Stages of Economic Development, Technology Environment and Employment Geneva International Labour Office, Geneva.

Pesaran, M.H., Shin, Y. and Smith, R.J. (2001), “Bounds testing approaches to the analysis of long run relationships”, Journal of Applied Econometrics, Vol. 16 No. 3, pp. 289-326.

Phillips, P.C.B. and Perron, P. (1988), “Testing for a unit root in time series regression”, Biometrika, Vol. 75 No. 2, pp. 335-346.

Rank, W. Rank, R. and Status, E.F. (2017), “81.0 (▲)”, Pp. 112-113.

Saidi, K. and Hammami, S. (2015), “The impact of CO2 emissions and economic growth on energy consumption in 58 countries”, Energy Reports, Vol. 1, pp. 62-70.

Saidi, K. and Ben Mbarek, M. (2016), “Nuclear energy, renewable energy, CO2 emissions, and economic growth for nine developed countries: evidence from panel granger causality tests”, Progress in Nuclear Energy, Vol. 88, pp. 364-374.

Shahbaz, M., Zakaria, M., Shahzad, S.J.H. and Mahalik, M.K. (2018), “The energy consumption and economic growth nexus in top ten energy-consuming countries: Fresh evidence from using the quantile-on-quantile approach”, Energy Economics, Vol. 71, pp. 282-301.

Soytas, U., Sari, R. and Ewing, B.T. (2007), “Energy consumption, income, and carbon emissions in the United States”, Ecological Economics, Vol. 62 Nos 3/4, pp. 482-489.

Stern, D.I. (2000), “A multivariate cointegration analysis of the role of energy in the US macroeconomy”, Energy Economics, Vol. 22 No. 2, pp. 267-283.

Streimikiene, D. and Kasperowicz, R. (2016), “Review of economic growth and energy consumption: a panel cointegration analysis for EU countries”, Renewable and Sustainable Energy Reviews, Vol. 59, pp. 1545-1549.

Tang, C.F. and Abosedra, S. (2014), “The impacts of tourism, energy consumption and political instability on economic growth in the MENA countries”, Energy Policy, Vol. 68, pp. 458-464.

Tiba, S. and Omri, A. (2017), “Literature survey on the relationships between energy, environment and economic growth”, Renewable and Sustainable Energy Reviews, Vol. 69, pp. 1129-1146.

Tugcu, C.T. and Topcu, M. (2018), “Total, renewable and non-renewable energy consumption and economic growth: revisiting the issue with an asymmetric point of view”, Energy, Vol. 152, pp. 64-74.

Vidyarthi, H. (2015), “Energy consumption and growth in South Asia: evidence from a panel error correction model”, International Journal of Energy Sector Management, Vol. 9 No. 3, pp. 295-310.

Wang, S., Li, G. and Fang, C. (2018), “Urbanization, economic growth, energy consumption, and CO2 emissions: empirical evidence from countries with different income levels”, Renewable and Sustainable Energy Reviews, Vol. 81, pp. 2144-2159.

Wang, S., Zhou, C., Li, G. and Feng, K. (2016), “CO2, economic growth, and energy consumption in china’s provinces: investigating the spatiotemporal and econometric characteristics of china’s CO2 emissions”, Ecological Indicators, Vol. 69, pp. 184-195.

Zivot, E. and Andrews, D.W.K. (1992), “Further evidence on the great crash, the oil price shock, and the unit root hypothesis”, Journal of Business and Economic Statistics, Vol. 10 No. 3, pp. 251-270.

Acknowledgements

The authors would like to express their thanks for the opportunity to present a previous version of this paper at the 2nd AIEEE Energy Symposium and at the 1st International Conference on Energy, Finance and the Macroeconomy, 2017 (ICEFM, 2017). The comments and suggestions received were very useful. The authors gratefully acknowledge the generous financial support of the NECE-UBI – Research Unit in Business Science and Economics, Project no. UID/GES/04630/2013 – sponsored by the Portuguese Foundation for the Development of Science and Technology. The authors are also grateful for the anonymous reviewers for their helpful and invaluable comments and suggestions.

Corresponding author

Antonio Cardoso Marques can be contacted at: amarques@ubi.pt