Advertisement

Could alternative energy sources in the transport sector decarbonise the economy without compromising economic growth?

  • Sónia Almeida Neves
  • António Cardoso Marques
  • José Alberto Fuinhas
Article

Abstract

The transition towards a low-carbon transport sector (TS) plays a fundamental role on the decarbonisation of economies. The effects of both conventional (fossil fuels) and alternative (renewable fuels and electricity) energy consumption in the transport sector, economic growth, and carbon dioxide emissions were analysed by using a panel vector autoregressive of 21 high-income Organization for Economic Co-operation and Development countries from 1990 to 2014. The results support the feedback hypothesis between both conventional and alternative TS energy sources and economic growth. In other words, electricity use on TS has enlarged the economic growth, while consumption of renewable fuels is actually hampering it. Additionally, TS fossil fuels consumption is contributing to economic growth. With reference to the environmental impacts of TS energy use, this paper highlights the harmful effect of conventional energy sources on the environment. However, there is no evidence wherein TS alternative energy sources are directly linked with a reduction of carbon dioxide emissions. Accordingly, the promotion of alternative TS energy sources should deserve further attention. On the one hand, there is evidence that the use of renewable fuels is obstructing economic growth. On the other hand, the use of both TS electricity and renewable fuels is not reducing carbon dioxide emissions.

Keywords

Energy consumption Transport sector Conventional sources Electricity Renewable fuels CO2 emissions 

1 Introduction

Reducing the environmental impacts associated with energy use has concerned not only the literature but also policymakers. The renewable deployment within the electricity mix has been pursued bearing this objective in mind. However, sectors that are highly powered by fossil fuels, such as the transport sector (hereinafter TS) have led to inertia on the shift towards low-carbon economies. Therefore, intervention in this sector is required. On the one hand, TS is a crucial sector for the entire dynamics of the economy. On the other hand, this sector is intensive in terms of internal combustion engines powered by fossil fuels, namely oil, the latter being highly harmful to the environment. The historical data, disclosed by the World Energy Council (2011), show that in 2010, TS was responsible for 19% of the global energy consumption, with 96% coming from oil. Moreover, this sector is also responsible for 60% of the global oil used, and 23% of the global carbon dioxide (CO2) emissions. Additionally, the European Commission (2016) indicates that in 2014, among the European Union (EU) countries, TS is responsible for 33% of the final energy consumption, with 94% from petroleum products. Furthermore, this sector is responsible for 25.5% of the European Union (EU) greenhouses gases (GHG) emissions.

Over the last decades, the interactions between energy consumption and economic growth (energy–growth nexus) have attracted particular attention from the literature (Payne 2010; Omri 2014; Tiba and Omri 2017). The results of the traditional energy–growth nexus can differ from the aggregate level to sectoral level (Abid and Sebri 2012). In both these levels, the energy consumption is a critical variable to explain the growth (Camarero et al. 2015). Accordingly, the sectoral energy consumption, namely TS, has caught the attention of specialised literature, namely regarding their effects on both economic growth and CO2 emissions (Costantini and Martini 2010; Tang and Shahbaz 2013; Burke and Csereklyei 2016). Although the literature is quite consensual on the harmful effects of TS energy consumption on the environment, the effects on the economic growth are not so harmonious (see, e.g. Costantini and Martini 2010; Liddle and Lung 2013; Saboori et al. 2014; Ibrahiem 2017).

The transition to low-carbon economies remains entirely dependent on the abatement of the fossil fuels used in TS. Recently, a technological upgrade on the internal combustion engines has been designed to reduce emissions of the pollutant gases. This upgrade included the improvement of the injection systems, modification of gases circulation, combustion chamber, as well as piston head design. Similarly, a survey of this technological upgrade on the internal combustion engines can also be found, for instance, in Abdul-Wahhab et al. (2017). Furthermore, the literature has proven that energy efficiency policies are efficient on TS decarbonisation (Xu and Lin 2015a; Shafiei et al. 2017; Talbi 2017). Currently, the exigency of CO2 abatement is growing. The fulfilment of carbon standards had encouraged the vehicle’s manufacturer to go further. As a consequence, some manufacturers have announced that they would stop producing new vehicles with internal combustion engines.

Although the improvement in terms of efficiency of internal combustion engines has been pursued, the designed policy intervention was aimed at promoting the use of alternative TS energy sources. In fact, the penetration of renewable fuels and electricity is mandatory so as to reduce both the dependence on fossil fuels and GHG emissions. The literature has not been consensual on the effects of the alternative sources on the environment. As to the renewable fuels, the literature supports that the use of biofuels on TS could reduce CO2 emissions (Zhang and Chen 2015). However, Månsson (2016) argues that biofuels could be ineffective on environmental protection if the competition for this kind of sources increases. The use of renewable fuel still faces several technical and social challenges that actually hinder their penetration (Bae and Kim 2017). The efficiency of renewable fuels increases with the increment of octanes in the fuel. For instance, ethanol improves the number of octane. However, it keeps producing lower heating value, and therefore, it needs to be mixed with an additive, for instance, gasoline so as to increase engine efficiency (Bae and Kim 2017).

With reference to electricity use on TS, it is also faced with several challenges. Therefore, it will only contribute to reducing CO2 emissions if the electricity is being generated from renewable sources (Ajanovic and Haas 2016). The share of TS energy consumption achieved through electricity remains low, mainly occurring on railways. The deployment of large amounts of electric vehicles on the road transport continues quite dependent upon a technological upgrade so as to achieve higher capacity and therefore enhance the lifecycle batteries of electric vehicles at a lower cost. Electric mobility also remains dependent on the social acceptance, and the improvement on the charging infrastructures (Mahmoudzadeh Andwari et al. 2017). It is expected that high penetration of electric vehicles will decrease their costs. This occurs due to the economies of scale and the increase in the learning curves. Therefore, batteries of electric vehicles are expected to be more competitive than internal combustion engines by 2030 (Mahmoudzadeh Andwari et al. 2017).

Although the literature has ascertained the relationships between TS energy consumption, economic growth, and CO2 emissions (Chandran and Tang 2013; Saboori et al. 2014), the empirical literature has not considered, on an individual basis, the role played by conventional and alternative TS energy sources. Moreover, the literature has identified some factors that are hindering the transition to the alternative sources. However, the literature has not yet proved what this transition really implies for the economic growth and TS decarbonisation. Therefore, the novelty of this paper for the literature is twofold. First, it simultaneously analyses the role played both by conventional and by alternative TS energy sources on the economic growth and CO2 emissions. Second, this approach also allows to check whether the conventional sources had been replaced by the alternative, which is a desirable effect within the scope of the shift in TS energy paradigm. To sum up, in order to fill these gaps identified in the literature, this paper aims to answer the following central questions: (1) what is the role played by both TS conventional and alternative energy sources on the economic growth and CO2 emissions, and (2) are TS alternative energy sources replacing TS conventional energy sources?

2 An overview of the energy consumption in the transport sector

With reference to the reduction of environmental impacts associated with energy use, TS has deserved much attention from the literature not only for the technical specificities that hinder the transition for the low-carbon sector but also because of their importance for the economy. In fact, the relationship between TS energy consumption, economic growth and CO2 emissions is frequently found on the literature. Table 1 shows a brief survey of the literature that analyses these relationships as well as their conclusions. Although the harmful effects on the environment are well known, the relationship with the economic growth is not so consensual.
Table 1

Studies on the effects between TS energy consumption, economic growth and CO2 emissions

Author(s)

Time and country(ies)

Methodology

Variables

Main findings

Chandran and Tang (2013)

1971–2008

5 ASEAN countries

Granger causality (VECM)

CO2 emissions

Road energy consumption (ROAD)

Foreign direct investment

GDP

CO2 ↔ ROAD long-run (Malaysia and Thailand)

ROAD → CO2 long-run (Indonesia)

CO2 ↔ Road short-run (Philippines and Thailand)

ROAD → GDP long-run (Indonesia and Thailand) and short-run (Singapore, and Indonesia)

GDP↔ROAD both short- and long-run (Malaysia)

GDP → ROAD short-run (Philippines)

Liddle and Lung (2013)

1971–2009

107 countries

Heterogeneous panel causality

CMG

TS energy consumption (TS_EC)

GDP

GDP → TS_EC

GDP has a positive effect on TS_EC

Ben Abdallah et al. (2013)

1980–2010 Tunisia

Johansen cointegration

Granger causality (VECM)

Transport value added (TVA)

Road energy consumption (ROAD)

CO2 emissions from TS

Road infrastructure

Fuel price

TVA ↔ CO2

TVA ↔ ROAD

CO2 ↔ ROAD

Saboori et al. (2014)

1960–2008

27 OECD countries

Fully Modified Ordinary Least Squares cointegration

CO2 from TS

GDP

Road energy consumption (ROAD)

GDP ↔ ROAD

GDP ↔ CO2

CO2 ↔ ROAD

Ibrahiem (2017)

1980–2011

Johansen cointegration test

Granger causality (VECM)

Road energy consumption (ROAD)

GDP

Urbanisation

Population growth

Short-run

GDP ↔ ROAD

Long-run

ROAD → GDP

Alshehry and Belloumi (2017)

1971–2011 Saudi Arabia

ARDL

Granger causality (VECM)

CO2 emissions from TS

Road energy consumption (ROAD)

GDP

CO2 ↔ ROAD

ROAD ≠ GDP

GDP ≠ CO2

ARDL autoregressive distributed lag, CO 2 carbon dioxide emissions, CMG correlated effects mean group, OECD organisation for economic co-operation and development, VAR autoregressive, VECM vector error correction mechanism, GDP gross domestic product

The need to understand the drivers of TS energy consumption has motivated the literature to go further. In fact, there is a set of country-specific studies that analyse the drivers of TS energy consumption, by considering a sector as a whole or subdividing it according to different infrastructures. For instance, Achour and Belloumi (2016) analysed the TS in Tunisia and came to the conclusion that the gross domestic product (GDP), the population, transports intensity, and transports structure expand the TS energy consumption, while the energy intensity effect decreases it, since the energy efficiency measures taken on transports are appropriate to reduce the use of fossil fuels. Additionally, despite only considering the road TS energy use, when speaking of the same country, this is positively affected by the vehicle fuel intensity, vehicle intensity, economic growth, urbanised kilometres, and national network (Mraihi et al. 2013). Another example of this proves that in the Chinese TS, energy consumption is boosted by the transport activity, whereas energy intensity decreases it (Zhang et al. 2011). Furthermore, Wu and Xu (2014) focused on the cargo transportation in China and found that both the intensity of goods carried and the cargo transportation infrastructure have a negative impact on cargo transport-related energy consumption, whereas the economic growth actually boosts it.

Although the policies aimed at promoting the reduction of energy use (efficiency or conservation) are robust on the TS decarbonisation (Xu and Lin 2015a, b; Shafiei et al. 2017; Talbi 2017), the effects of new energy sources, such as renewable fuels and electricity, are not consensual among the scientific community. For example, a simulation-based comparison between scenarios of the transition to hydrogen and electricity shows that the transition to electric mobility is preferable for the reduction of the total fuel use and the goals of economic benefits; however, the mixed transition to electric mobility and hydrogen proves to be desirable to achieve the goal of reducing emissions (Shafiei et al. 2017). Similarly, using a LEAP (long-range energy alternative planning) model, Azam et al. (2016) showed that the reduction of both the energy consumption in road TS and CO2 can be achieved by the natural gas scenario, followed by the biofuels scenario and hybrid electric vehicle scenario. However, Månsson (2016) supports that the strategies for new energies sources (biofuels and electricity) are affected by external factors. Biofuels can be inefficient on the decarbonisation if many countries increase the use of these sources, bearing in mind that in this case, the growth in demand actually increases the competition by a set of the fixed resources. As to TS, electrification is quite dependent on the technological upgrade in the other countries, which means that it is difficult to implement this technology. In addition, the penetration of alternative sources also aims to reduce the external energy dependence, mainly for non-oil-producing countries. For these countries, electric vehicles could be the most efficient technology to reduce the external energy dependence (Marques et al. 2016). The same authors also argue that for Norway, Saudi Arabia and Russia (oil-producing countries), the energy dependence is affected in the same way when different vehicles technology was assessed.

The literature has shown that the use of non-fossil combustibles can reduce the emissions of pollutant gases (Nocera and Cavallaro 2016). For instance, the use of biofuels in China and USA TS is contributing to the reduction of CO2 emissions (Zhang et al. 2016), which seems to corroborate the results obtained both by Zhang and Chen (2015) for Chinese TS and by Neves et al. (2017) for OECD countries. Regarding the European Union, there are several policies to promote the use of biofuels (Cansino et al. 2012), although the associated biofuels costs do remain higher than fossil fuel costs (Ajanovic and Haas 2011; Sanz et al. 2014). However, the reduction of global CO2 emissions will only be reached if all the countries reduce oil consumption. Otherwise, if only some countries reduce oil use, the objective of global oil use reduction will be achieved, despite the fact that a drop in global CO2 emissions will not be reached (Eliasson and Proost 2015).

With reference to the electricity penetration within road transportation, it will be beneficial for the environment if the electricity used actually comes from renewable sources (Ajanovic and Haas 2016). The shift from fossil fuels consumption to electricity within TS could raise new challenges for the electricity systems, mainly due to the fact that they cannot be able to deal with an additional demand caused by electric mobility. Nevertheless, literature is stating that with a controlled plug-in vehicle loading in out-off peak hour, the impact on electricity costs will be less than 5% and there is no need to increase the installed capacity (Razeghi and Samuelsen 2016). Likewise, taking into account the environmental impacts, the promotion of vehicle charges is necessary, when there are high levels of renewables production and the adoption of Time-Of-Use tariffs (TOU), so as not to compromise the sustainability of the electricity system (Coffman et al. 2017).

To sum up, the reduction of the environmental impacts associated with energy consumption has caught the attention of the literature. The transition to a low-carbon economy has been hindered by the TS because it remains highly powered by fossil fuels. The analysis of the relationships between CO2 emissions, economic growth, and TS energy consumption has inspired the literature, by considering the TS as whole or subdividing it in the different infrastructures (see Table 1). The promotion of more efficient technologies, as well as the use of alternative sources, such as renewable fuels and electricity, has been pursued so as to counteract the harmful effect of this sector on the environment. In fact, energy efficiency measures can be effective at the level of TS decarbonisation. However, the literature is not harmonious as to the role played by the alternative sources (renewable fuels and electricity) on the transition towards low-carbon sector both in an economically and in an environmentally sustainable way. Additionally, the literature has not analysed the effects resulting from conventional and alternative (renewable fuels and electricity) TS energy sources on the economic growth and CO2 emissions, by using historical data.

3 Data and methodology

This paper used yearly panel data, comprising a time span from 1990 to 2014 for 21 high-income OECD countries. The period under analysis started in 1990, considering that it is a milestone, namely as far as environmental protection is concerned. For instance, the Kyoto Protocol highlights that GHG emissions should be reduced as compared with values of 1990. Therefore, following the data availability criterion for the entire period, the selected countries were as follows: Australia, Austria, Belgium, Canada, the Czech Republic, Denmark, Finland, France, Germany, Greece, Ireland, Italy, the Netherlands, Norway, Poland, Portugal, Spain, Sweden, Switzerland, UK, and the USA. All the variables were converted into their per capita value. Since all the variables were converted into their natural logarithm and considering there is a set of zeros on the database, a constant of 1 was added to each variable in order to solve the issue of loss of observations. The prefix “L” shall hereinafter mean a natural logarithm, whereas “D” shall mean a first difference of the variables. Table 2 shows the variables’ description, descriptive statistics, and the sources of the variables.
Table 2

Variables’ definition and descriptive statistics

Variable

Description

Obs

Mean

Std. Dev.

Min

Max

Source

LGDP

Ratio between GDP (constant LCU) and total population

525

10.78

1.03

9.44

13.22

WDI

LFF

Ratio between transports’ fossil fuels consumption and total population (kg of equivalent oil/person)

525

6.60

0.44

5.17

7.64

IEA

LELE

Ratio between transports’ electricity consumption and total population (kg of equivalent oil/person)

525

2.32

0.81

0.33

3.64

IEA

LRES

Ratio between transports’ renewable fuels consumption and total population (kg of equivalent oil/person)

525

1.31

1.45

0

4.70

IEA

LEN

Ratio between total energy consumption (except in TS) and total population (kg of equivalent oil/person)

525

7.65

0.40

6.78

8.43

OECD statistics

LCO2

Ratio between total CO2 emissions (from consumption of oil, gas and coal) and total population (kg of carbon dioxide equivalent/person)

525

4.10

0.96

1.82

5.298

BP statistics

WDI World Development Indicators, IEA International Energy Agency (IEA Headline Global Energy Data, 2016 edition), LCU local currency unit, and OECD Organization for Economic Co-operation and Development

The gross domestic product per capita (GDP), measured into constant local currency unit, was used as economic growth proxy, as usual (see, e.g. Saboori et al. 2014). The TS energy consumption was subdivided into: fossil fuels (FF), electricity (ELE) and renewable fuels (RES1), and these were expressed in kg of equivalent oil per capita, as frequently stated in the literature(see, e.g. Saboori et al. 2014; Achour and Belloumi 2016). CO2 emissions from consumption of oil, gas, and coal are expressed in kg of CO2 equivalent. Moreover, the total energy consumption in the economy except in TS was used as a control variable.

According to a panel data approach, the technical features of both variables and crosses (countries) must be checked in order to avoid biasing the results. Accordingly, the adopted procedure included checking of: (1) cross-sectional dependence test (CD test), (2) panel unit root tests (see Table 3), (3) correlation matrix values, and (4) variance inflation factor (VIF).
Table 3

Cross-sectional dependence test (CD test) and second-generation unit root test (CIPS)

 

CD test

CIPS

CD test

Corr

Abs (corr)

Lags

Without trend

With trend

LGDP

67.51***

0.932

0.932

0

0.636

− 2.863***

1

− 1.130

− 1.500*

LFF

32.85***

0.453

0.542

0

1.571

0.732

1

0.622

0.922

LELE

5.94***

0.082

0.503

0

− 0.042

1.228

1

0.842

1.611

LRES

64.20***

0.886

0.886

0

− 1.287*

− 3.325***

1

− 0.243

− 1.960**

LCO2

34.34***

0.474

0.544

0

− 4.131***

− 2.797***

1

− 1.588*

− 1.593*

LEN

29.04***

0.401

0.509

0

− 2.577***

− 4.664***

1

− 1.164

− 3.226***

DLGDP

42.66***

0.601

0.601

0

− 9.095***

− 6.065***

1

− 6.069***

− 3.861***

DLFF

19.99***

0.282

0.314

0

− 12.820***

− 10.779***

1

− 5.525***

− 3.466***

DLELE

1.71*

0.024

0.165

0

− 12.686***

− 11.222***

1

− 5.844***

− 4.047***

DLRES

13.44***

0.189

0.246

0

− 13.985***

− 12.319***

1

− 8.398***

− 6.038***

DLCO2

18.69***

0.263

0.316

0

− 15.398***

− 14.491***

1

− 9.796***

− 8.501***

DLEN

26.82***

0.378

0.394

0

− 15.905***

− 14.550***

1

− 11.505***

− 10.889***

CD test was performed according to the null hypothesis of the cross-sectional independence. The second-generation unit root test was performed under the null hypothesis wherein the variables are I(1)

***, **, and * denote statistical significance level at 1, 5, and 10%, respectively

First-generation unit root tests are not trustworthy in the presence of cross-sectional dependence. Accordingly, when this phenomenon was found, the second-generation unit root test (CIPS) proposed by Pesaran (2007) should be performed. As stated in Table 2, this phenomenon was detected for all the variables with 1% level of statistical significance, except for DLELE, which is only statistically significant at 10%. For this variable, both first- and second-generation unit root tests were performed, and both of them suggest that the variable is I(1). Overall, the results presented in Table 3 show that all the variables were stationary in their first differences. The correlation matrix values and the variance inflation factor (VIF) were analysed so as to certify that both correlation and multicollinearity did not deserve concern for the estimation.

Faced with potentially endogenous variables, i.e. it is likely that the variables have a simultaneous causality, the use of panel data vector autoregressive (PVAR) is suitable. The estimator proposed by Love and Zicchino (2006) supports stationary endogenous variables as well as the unobserved individual heterogeneity. As can be seen in Sect. 4, the presence of the endogeneity was confirmed by the blocks of exogeneity analysis. This implies that the error term was correlated with the independent variables. Accordingly, the panel VAR was appropriated to deal with these data features, and the estimation can be explained as follows:
$${\rm Z}_{{it^{{}} }} = \varGamma_{0} + \varGamma_{1} {\rm Z}_{it - 1} + f_{i} + d_{c,t} + \varepsilon_{t} ,$$
(1)
where \({\rm Z}_{it}\) denotes the vector of the endogenous used variables (DLGDP, DLFF, DLELE, DLRES, DLCO2, and DLEN), \(\varGamma_{0}\) represents the vector of the constants, \(\varGamma_{1} {\rm Z}_{it - 1}\) denotes the matrix polynomial, \(f_{i}\) represents the fixed effects, \(d_{c,t}\) denotes the time effects, and \(\varepsilon_{t}\) represents the error term.

The presence of fixed effects was tested by using the Hausman test. The null hypothesis predicts that the random effects estimator is appropriated. All variables were tested both as dependent and as independent variables. The existence of the fixed effects was only detected for the model where DELE is dependent. Although the presence of fixed effects raises correlation problems between the regressors, this methodology allowed to remove them by using the “Hermelet procedure” as proposed by Arellano and Bover (1995). According to this technique, data loss is minimised, once the mean for future observations available was removed (Love and Zicchino 2006). Therefore, the system was estimated, based on a generalised method of moments (GMM) and with the regressors lagged as instrumental variables.

The Granger causality test, based on the Wald test (Abrigo and Love 2015), was performed, showing that the null hypothesis is the absence of causality. Furthermore, impulse response functions (IRFs) were estimated by using a Gaussian approximation based on the Monte Carlo simulations. The orthogonalized impulse response functions were based on the Cholesky decomposition, and the standard errors and the confidence intervals were estimated according to the 1000 Monte Carlo simulations. The function revealed reaction of one variable to the shock in another variable. After that, the forecast error variance decomposition (FEVD) was performed, based on a Cholesky decomposition of the residual covariance matrix, using 1000 Monte Carlo simulations, and for 15 periods. This function allowed us to understand the percentage that each endogenous variable explains of the forecast error variance of the other specific variable. After carrying out the analysis of the exogeneity blocks, the VAR—Cholesky ordering of variables was used, by placing the variables in the decreasing order of the exogeneity.

4 Results

Following the three lags selection criteria proposed by Andrews and Lu (2001), namely Bayesian information criterion (MBIC), Akaike information criterion (MAIC), and Hannan and Quinn information criterion (MQIC), the selected optimal lags in the PVAR estimation were 1 (see Table 4). Indeed, lag 1 minimises all criteria (MBIC, MAIC, and MQIC).
Table 4

Lag order selection criteria

Lag

CD

J

Jp value

MBIC

MAIC

MQIC

1

0.421

153.928

0.002

− 492.880

− 62.072

− 232.695

2

0.616

88.856

0.087

− 342.349

− 55.144

− 168.892

3

0.706

33.963

0.566

− 181.640

-38.037

− 94.911

The first-order PVAR was estimated with an impulse dummy for 2010 as an exogenous variable. The inclusion of this dummy aims to correct the residuals of the estimations since they suffered a breakdown in this year caused by economic recuperation after the economic crisis. The stability of the first-order PVAR was checked. The results are shown in Fig. 1. The stability condition is accomplished once the values are inside the circle. As mentioned by Abrigo and Love (2015), this implies that the impulse response functions and forecast error variance decomposition have a known interpretation (Table 5).
Fig. 1

Summary of the causalities according to the Granger causality. Notes: → denotes the causality with a statistical significance of the 1 and 5%. ⇢ denotes the causality with a statistical significance of the 10%

Table 5

Eigenvalue stability condition

All the eigenvalues lie inside the unit circle. pVAR satisfies stability condition

The results of Granger causality, following the first-order PVAR, are listed in Table 6. The null hypothesis predicts the absence of the causality. The TS fossil fuels consumption and total energy consumption (except in TS) show the bidirectional causality with the economic growth. This paper agrees with the findings of Camarero et al. (2015) that energy consumption is actually a critical variable to explain the economic growth in both aggregate and sectoral levels. Moreover, the use of electricity and renewable fuels on the TS is also causing the economic growth. However, these results also sustain that the electricity use on the TS is not significantly dependent on the economic performance from a statistical approach, since the economic growth is only causing the TS electricity use at 10% level of significance. Conversely, the use of renewable fuels on the TS is caused by the economic growth, and vice versa, supporting a bidirectional causality.
Table 6

Granger causality test

 

DLELE

DLCO2

DLEN

DLRES

DLFF

DLGDP

DLELE does not cause

1.406

11.388***

20.566***

5.650**

5.265**

DLCO2 does not cause

0.012

0.047

4.626**

13.396***

13.548***

DLEN does not cause

1.016

9.144***

4.600**

5.430**

19.377***

DLRES does not cause

1.880

0.578

0.460

2.789*

5.790**

DLFF does not cause

7.840***

5.124**

12.453***

12.206**

29.012***

DLGDP does not cause

3.159*

67.235***

26.748***

15.627***

10.530***

ALL

10.258*

112.266***

81.295***

43.411***

32.499***

54.461***

***, **, and * denote statistical significance at 1, 5, and 10%, respectively

Regarding CO2 emissions, a bidirectional causality is shown with fossil fuels use. Moreover, there is a unidirectional causality running from energy consumption except in TS to the CO2 emissions. In fact, this finding proves the harmful effect of energy use on the environment. The use of renewable fuels is caused by CO2 emissions, although the opposite is not true. Although the use of renewable fuels aims to reduce CO2 emissions, this paper indicates that this effect is not taking place now. This result indicates that the renewables penetration within the TS is being promoted by CO2 emissions. With reference to electricity consumption on the TS, there is no relationship with CO2 emissions. Figure 1 shows a summary of the causalities found according to the Granger causality.

Taking into account that the Granger causality is not able to reveal all the information about the relationships established between the variables, the impulse response functions (IRFs) were carried out (see Fig. 2). They provide both information about how one variable reacts (response), faced to a shock or innovation in another variable (impulse), while also revealing the time needed to return to equilibrium. Subsequently, the forecast error variance decomposition (FEVD) was also performed (see Table 7). The results allow us to understand the percentage of the forecast error variance that each of the variables explains, faced with a shock or innovation in one specific variable. Moreover, it also indicates both the time needed to and the percentage that each variable contributes to achieve the equilibrium.
Fig. 2

Impulse response functions (IRFs)

Table 7

Forecast error variance decomposition (FEVD)

Impulse variable

Forecast horizon

Response variable

DLELE

DLCO2

DLEN

DLRES

DLFF

DLGDP

DLELE

1

1

0

0

0

0

0

2

0.98928

0.00060

0.00038

0.00092

0.00263

0.00619

5

0.98835

0.00071

0.00039

0.00150

0.00277

0.00628

10

0.98831

0.00072

0.00039

0.00151

0.00278

0.00629

15

0.98831

0.00072

0.00039

0.00151

0.00278

0.00629

DLCO2

1

0.00116

0.99884

0

0

0

0

2

0.00925

0.78911

0.04447

0.00097

0.04623

0.10998

5

0.02030

0.71993

0.06048

0.00911

0.08289

0.10729

10

0.02070

0.71706

0.06030

0.00995

0.08396

0.10802

15

0.02070

0.71702

0.06030

0.00996

0.08397

0.10803

DLEN

1

0.01678

0.30063

0.68259

0

0

0

2

0.05948

0.24148

0.57019

0.00037

0.07148

0.05700

5

0.05914

0.24530

0.54514

0.00660

0.08312

0.06072

10

0.05930

0.24493

0.54347

0.00710

0.08382

0.06138

15

0.05931

0.24492

0.54345

0.00710

0.08383

0.06138

DLRES

1

0.00378

0.00160

0.02396

0.97066

0

0

2

0.02704

0.00489

0.02110

0.91903

0.00566

0.02228

5

0.03059

0.00633

0.02083

0.90302

0.00803

0.03121

10

0.03079

0.00656

0.02079

0.90161

0.00860

0.03165

15

0.03079

0.00656

0.02079

0.90160

0.00861

0.03166

DLFF

1

0.00500

0.07308

0.01342

0.00529

0.90321

0

2

0.02289

0.09796

0.01159

0.00892

0.82921

0.02943

5

0.02973

0.09962

0.01168

0.01784

0.79140

0.04973

10

0.03037

0.09981

0.01161

0.01879

0.78830

0.05112

15

0.03038

0.09981

0.01161

0.01880

0.78826

0.05114

DLGDP

1

0.01156

0.039566

0.05771

0.00022

0.13319

0.75776

2

0.04046

0.07082

0.04501

0.02185

0.28966

0.53219

5

0.04929

0.08270

0.04169

0.03888

0.29543

0.49202

10

0.05017

0.08319

0.04123

0.04032

0.29577

0.48932

15

0.05018

0.08319

0.04122

0.04034

0.29577

0.48929

As shown in Fig. 2, faced with a shock or innovation in one variable, all the variables return to the equilibrium. This result supports the stationarity of the variables under study. With reference to an impulse on the DLGDP, all the variables respond positively, except DLRES, thus meaning that they are achieving the equilibrium in 5 periods, except DLELE has managed to achieve it in 3 periods. Conversely, considering the response of the DLGDP, facing a shock in the other variables has a positive response in all the variables, except in DLRES and DLEN.

As to the fossil fuels used on the TS, all the variables react positively faced to a shock on the DLFF, except DLELE. Although the negative effect of electricity on the fossil fuels consumption has a lower magnitude, this result supports the perspective that the electrification of the transports sector could reduce the use of fossil fuels. Faced with a shock in DLCO2, the return to the equilibrium occurs quickly, approximately in 3 years, except for DLGDP and DLFF that achieve such equilibrium in about 7 years. Facing an impulse in DLCO2, special attention is needed for the negative response of the DLRES. In contrast, the economic growth and the use of TS fossil fuels reacted positively.

The results of the forecast error variance decomposition (FEVD) are listed in Table 7. In fact, the results allow us to understand the percentage that each endogenous variable explains of the forecast error variance of the other specific variable.

Renewable fuels, as well as electricity consumption on the TS, are self-explanatory as to the most important part of their forecast error variance. In the first period, the DLELE and DLRES are explaining 98.93 and 97.066% of their respective forecast error variance. The other endogenous variables are not significant on the explanation of the forecast error variance. In fact, as to the new equilibrium point, DLELE contributes in 98.831% for their respective forecast error variance, while DLRES explains 90.160% of their forecast error variance. This means that the penetration of alternative energy sources on TS energy consumption is not significantly dependent neither on the other TS energy sources nor on the economic performance.

Faced with a shock on DLCO2, in the first period, the forecast error variance is explained in 0.116% by the TS electricity use and in 99.884% by the DLCO2. After a tenth period, the forecast error variance is explained in 71.706% by DLCO2, 6.03% by DLEN, 8.396% by DLFF, and 10.802% by DLGDP. Indeed, economic growth, energy use except in TS and the use of fossil fuels are the most important contributors to CO2 emissions, responding with greater magnitude with a shock in the DLCO2. Although the IRF shows that renewable fuels and electricity use on the TS respond negatively faced with a shock or innovation in CO2 emissions, FEVD results indicate that this variable contributes to a low percentage in explaining the forecast error variance.

With reference to a shock in the DLFF variable, in terms of achievement of equilibrium achievement, after a 10-year period, the variables that are explained in the largest part of the forecast error variance are the TS use of fossil fuels (78.830%), CO2 emissions (9.981%), economic growth (5.112%), and TS electricity consumption (3.037%). As regards a shock in the economic growth, it is self-explanatory in 75.776%, in the first year. As to the equilibrium, the largest part of the forecast error variance is explained by DLGDP and DLFF, accounting for 49.202 and 29.543%, respectively, thus showing the importance of the use of TS fossil fuels for the economy.

5 Discussion

So far, the analysis of the effects of TS energy consumption on the economic growth and CO2 emissions has been deserved much of the attention of the literature. However, none of these studies have analysed the effects of both conventional and alternative energy sources, on an individual basis. In fact, this approach could provide crucial guidelines for policymakers so as to achieve a low-carbon TS.

It is a well-known fact that the TS is vastly powered by fossil fuels, namely oil, which is harmful to the environment. Although there are several efforts to improve the efficiency of the internal combustion engines to reduce the pollutant gases emissions (Abdul-Wahhab et al. 2017), this paper corroborates the conclusion that the use of these sources is increasing CO2 emissions. Moreover, the use of fossil fuels on TS is contributing to the economic growth, which is in line with Saboori et al. (2014). Indeed, this outcome shows the importance of TS for the dynamics of the entire economies. However, in order to decarbonise this sector and the economy, it is mandatory to reduce fossil fuels consumption.

Regarding the effects of renewable fuels consumption on the TS, this paper supports that they are actually reducing the use of fossil fuels, i.e. there is a substitution effect of fossil fuels by renewable fuels, although with a low level of significance. In fact, the penetration of the alternative sources is still faced with several challenges, namely social and technical (Bae and Kim 2017), something which can explain the low levels of significance found by this research. In other words, our findings sustain that the goal of reducing the use of fossil fuels on the TS could really be achieved by promoting the use of renewable fuels, contrary to what could happen in the electricity generation due to the need of backup from controllable fossils, as stated by Boccard (2009) and Flora et al. (2014). Nevertheless, this paper also indicates that renewable fuels are apparently hampering economic growth. Indeed, this outcome could result from excessive costs associated with supporting these sources, as highlighted by Ajanovic and Haas (2011). The findings of this paper also provide some guidelines to make the renewable fuels more attractive and competitive. The advancements to increase their market share are required. Currently, their use remains small, something which could explain the absence of (statistically significant) relationship with CO2 emissions. Although the use of these sources still does not contribute to directly reducing CO2 emissions, they are actually contributing to reducing the use of conventional sources, which consequently may reduce CO2 emissions. At the same time, they are apparently hampering economic growth. This could indicate that more research on the renewable fuels is required. First, it is mandatory to improve the renewable fuels efficiency so as to make their performance competitive with the conventional fuels with an aim towards enlarging their social acceptance. The improvement of the number of octanes accomplished with high heating value could reduce the need for additive conventional fuel. Second, their cost-effectiveness also needs to be enhanced so as to avoid the negative effects on the economic growth. Therefore, investments in research and development (R&D) of the renewable fuels could be an efficient way to counteract the undesirable effects found in this paper.

This paper indicates that electricity consumption on transports actually affects the economy on a positive basis. Nevertheless, this paper also supports that electricity penetration on the TS energy mix is not highly significant dependent on the economic performance. This means that, during the period under study, the electrification of the TS is mainly a case of policy decision-making. Additionally, the electricity use on TS does not have any statistically significant relationship with CO2 emissions, something which is not expected. In fact, according to the period under study, the electricity use on the TS has occurred mainly on the railways. This unanticipated finding could indicate that CO2 savings in the tailpipe achieved by using electricity on TS have actually resulted in an increase in the CO2 emissions caused in the electricity generation process. As stated by Ajanovic and Haas (2016), it is expected that the environmental benefits associated with the electricity use on the TS will only be reached if the electricity is generated from renewable sources.

Currently, the transition for electric mobility on the road systems remains slight. Indeed, the greatest challenges are upon the social acceptance, the improvement of the charging infrastructure, the new business models and the research of the range extenders (Mahmoudzadeh Andwari et al. 2017). Therefore, in the next few decades or even years, the outcome of this study is expected to change, namely through the development of the life cycle and capacity of the electric vehicles’ batteries. Furthermore, it is also expected that the penetration of the electric vehicles could decrease their cost (Mahmoudzadeh Andwari et al. 2017), thus making the electric vehicles more attractive. Although electricity use is not directly contributing to reducing CO2 emissions, this paper indicates that it actually shows a substitution effect with fossil fuels sources. In other words, our findings sustain that the goal of reducing the use of fossil fuels on the TS could be achieved by promoting the use of electricity.

This paper indicates that both transport and electricity policies must be followed together. The promotion of the electric vehicles must be pursued. More investment in R&D in battery technology could be an efficient mechanism to improve the battery capacity and life cycle. It is expected that this progress could result in an increase in the share of electric vehicles in the automotive market. At the same time, policymakers should promote electricity generation through renewables sources. Indeed, the TS must use renewable electricity. Conversely, if the electricity used by TS is generated from conventional sources, the reduction of the CO2 emissions could not be achieved. This means that the consumption of the electricity must be coordinated with the natural resources availability, namely wind and solar photovoltaic. Policymaking must promote the charging of electric vehicles in periods of the high potential to the renewable generation. Users that charge their car in these periods must be encouraged. For instance, promotion of the existence of a charging station in the workplace could incentive the electric vehicle charging in these periods. Also, the existence of a differentiated electricity prices could be an efficient mechanism to achieve it. Actually, the existence of cheaper electricity when there is a high renewable generation will encourage users to charge their electric vehicles in these periods. The coordination of both transport and electricity policies could be helpful for both, in fact. The penetration of electric vehicles is essential on transition for low-carbon TS. Meanwhile, with the controlled charging process, electric vehicles could contribute to renewables accommodation.

6 Conclusions

The transition towards low-carbon TS has led policymakers to promote the use of alternative sources such as electricity and renewable fuels. However, the technical specifications of this sector actually act as a barrier and, as such, do hamper this energy transition. Therefore, this paper aims to provide some policy suggestions about how the conventional and alternative TS energy sources are interacting as well as their effects on CO2 emissions and economic growth. Based on an empirical approach, this paper applies a panel VAR for 21 high-income from 1990 to 2014. The results of this paper can be very helpful for political decision-making.

This paper supports that the use of conventional energy sources in the transport sector is enlarging the economic growth. However, thanks to the broadly documented literature, it also corroborates the harmful effect of these sources on the environment. Moreover, this paper indicates that the promotion of TS alternative sources must be pursued, despite the need to have further attention on this topic. With reference to renewable fuels, apparently this is hindering economic growth. Moreover, there is no evidence as to how these sources are obstructing CO2 emissions. Nevertheless, it also supports that these sources could actually contribute to reducing the dependence on fossil fuels. With reference to the electricity penetration on the TS, the conclusion is that it actually enlarges economic growth; however, this does not have a direct effect on CO2 emissions.

Nevertheless, it is important to make sure that the results of this paper reflect what has occurred in the past. In fact, the TS are faced with several challenges to transit for a low-carbon sector, something which is currently in rapid and constant transition, namely on the diversification of their energy mix. The results obtained in this paper can be kept in future or, alternatively, they are likely to evolve.

Footnotes

  1. 1.

    This variable comprises the direct use of renewable fuels by the transport sector and does not take into account the renewable electricity.

Notes

Acknowledgements

The financial support of the NECE-UBI, the Research Unit in Business Science and Economics, sponsored by the Portuguese Foundation for the Development of Science and Technology, Ministry of Science, Technology and Higher Education, Project UID/GES/04630/2013, is gratefully acknowledged. We also would like to express our acknowledgments to the anonymous reviewers for the valuable and useful comments and suggestions.

References

  1. Abdul-Wahhab, H. A., Al-Kayiem, H. H., Aziz, A. R. A., & Nasif, M. S. (2017). Survey of invest fuel magnetization in developing internal combustion engine characteristics. Renewable and Sustainable Energy Reviews, 79, 1392–1399.  https://doi.org/10.1016/j.rser.2017.05.121.CrossRefGoogle Scholar
  2. Abid, M., & Sebri, M. (2012). Energy consumption-economic growth nexus: Does the level of aggregation matter? International Journal of Energy Economics and Policy, 2(2), 55–62.Google Scholar
  3. Abrigo, M. R. M., & Love, I. (2015). Estimation of panel vector autoregression in stata: A package of programs, p. 28.  https://doi.org/10.1017/cbo9781107415324.004.
  4. Achour, H., & Belloumi, M. (2016). Investigating the causal relationship between transport infrastructure, transport energy consumption and economic growth in Tunisia. Renewable and Sustainable Energy Reviews.  https://doi.org/10.1016/j.rser.2015.12.023.Google Scholar
  5. Ajanovic, A., & Haas, R. (2011). On the future relevance of biofuels for transport in EU-15 countries. WIT Transactions on Ecology and the Environment, 143, 97–108.  https://doi.org/10.2495/ESUS110091.CrossRefGoogle Scholar
  6. Ajanovic, A., & Haas, R. (2016). Dissemination of electric vehicles in urban areas: Major factors for success. Energy.  https://doi.org/10.1016/j.energy.2016.05.040.Google Scholar
  7. Alshehry, A. S., & Belloumi, M. (2017). Study of the environmental Kuznets curve for transport carbon dioxide emissions in Saudi Arabia. Renewable and Sustainable Energy Reviews, 75, 1339–1347.  https://doi.org/10.1016/j.rser.2016.11.122.CrossRefGoogle Scholar
  8. Andrews, D. W. K., & Lu, B. (2001). Consistent model and moment selection procedures for GMM estimation with application to dynamic panel data models. Journal of Econometrics, 101(1), 123–164.  https://doi.org/10.1016/S0304-4076(00)00077-4.CrossRefGoogle Scholar
  9. Arellano, M., & Bover, O. (1995). Another look at the instrumental variable estimation of error-components models. Journal of Econometrics, 68(1), 29–51.  https://doi.org/10.1016/0304-4076(94)01642-D.CrossRefGoogle Scholar
  10. Azam, M., Othman, J., Begum, R. A., Abdullah, S. M. S., & Nor, N. G. M. (2016). Energy consumption and emission projection for the road transport sector in Malaysia: An application of the LEAP model. Environment, Development and Sustainability, 18(4), 1027–1047.  https://doi.org/10.1007/s10668-015-9684-4.CrossRefGoogle Scholar
  11. Bae, C., & Kim, J. (2017). Alternative fuels for internal combustion engines. Proceedings of the Combustion Institute, 36(3), 3389–3413.  https://doi.org/10.1016/j.proci.2016.09.009.CrossRefGoogle Scholar
  12. Ben Abdallah, K., Belloumi, M., & De Wolf, D. (2013). Indicators for sustainable energy development: A multivariate cointegration and causality analysis from Tunisian road transport sector. Renewable and Sustainable Energy Reviews, 25, 34–43.  https://doi.org/10.1016/j.rser.2013.03.066.CrossRefGoogle Scholar
  13. Boccard, N. (2009). Capacity factor of wind power realized values vs. estimates. Energy Policy, 37(7), 2679–2688.  https://doi.org/10.1016/j.enpol.2009.02.046.CrossRefGoogle Scholar
  14. Burke, P. J., & Csereklyei, Z. (2016). Understanding the energy-GDP elasticity: A sectoral approach. Energy Economics, 58, 199–210.  https://doi.org/10.1016/j.eneco.2016.07.004.CrossRefGoogle Scholar
  15. Camarero, M., Forte, A., Garcia-donato, G., Men-, Y., & Ordo, J. (2015). Variable selection in the analysis of energy consumption-growth nexus. Energy Economics.  https://doi.org/10.1016/j.eneco.2015.10.012.Google Scholar
  16. Cansino, J. M., de Pablo-Romero, M., Román, R., & Yñiguez, R. (2012). Promotion of biofuel consumption in the transport sector: An EU-27 perspective. Renewable and Sustainable Energy Reviews, 16(8), 6013–6021.  https://doi.org/10.1016/j.rser.2012.06.011.CrossRefGoogle Scholar
  17. Chandran, V. G. R., & Tang, C. F. (2013). The impacts of transport energy consumption, foreign direct investment and income on CO2 emissions in ASEAN-5 economies. Renewable and Sustainable Energy Reviews, 24, 445–453.  https://doi.org/10.1016/j.rser.2013.03.054.CrossRefGoogle Scholar
  18. Coffman, M., Bernstein, P., & Wee, S. (2017). Integrating electric vehicles and residential solar PV. Transport Policy, 53, 30–38.  https://doi.org/10.1016/j.tranpol.2016.08.008.CrossRefGoogle Scholar
  19. Costantini, V., & Martini, C. (2010). The causality between energy consumption and economic growth: A multi-sectoral analysis using non-stationary cointegrated panel data. Energy Economics, 32(3), 591–603.  https://doi.org/10.1016/j.eneco.2009.09.013.CrossRefGoogle Scholar
  20. Eliasson, J., & Proost, S. (2015). Is sustainable transport policy sustainable? Transport Policy, 37, 92–100.  https://doi.org/10.1016/j.tranpol.2014.09.010.CrossRefGoogle Scholar
  21. European Commission. (2016). Transport in Figures 2015.  https://doi.org/10.2832/861735.
  22. Flora, R., Marques, A. C., & Fuinhas, J. A. (2014). Wind power idle capacity in a panel of European countries. Energy, 66, 823–830.  https://doi.org/10.1016/j.energy.2013.12.061.CrossRefGoogle Scholar
  23. Ibrahiem, D. M. (2017). Road energy consumption, economic growth, population and urbanization in Egypt: Cointegration and causality analysis. Environment, Development and Sustainability.  https://doi.org/10.1007/s10668-017-9922-z.Google Scholar
  24. Liddle, B., & Lung, S. (2013). The long-run causal relationship between transport energy consumption and GDP: Evidence from heterogeneous panel methods robust to cross-sectional dependence. Economics Letters (Vol. 121).  https://doi.org/10.1016/j.econlet.2013.10.011.
  25. Love, I., & Zicchino, L. (2006). Financial development and dynamic investment behavior: Evidence from panel VAR. Quarterly Review of Economics and Finance, 46(2), 190–210.  https://doi.org/10.1016/j.qref.2005.11.007.CrossRefGoogle Scholar
  26. Mahmoudzadeh Andwari, A., Pesiridis, A., Rajoo, S., Martinez-Botas, R., & Esfahanian, V. (2017). A review of battery electric vehicle technology and readiness levels. Renewable and Sustainable Energy Reviews, 78(February), 414–430.  https://doi.org/10.1016/j.rser.2017.03.138.CrossRefGoogle Scholar
  27. Månsson, A. (2016). Energy security in a decarbonised transport sector: A scenario based analysis of Sweden’s transport strategies. Energy Strategy Reviews, 13–14, 236–247.  https://doi.org/10.1016/j.esr.2016.06.004.CrossRefGoogle Scholar
  28. Marques, S., Reis, L., Afonso, J. L., & Silva, C. (2016). Energy rating methodology for light-duty vehicles: Geographical impact. Environment, Development and Sustainability, 18(5), 1501–1519.  https://doi.org/10.1007/s10668-016-9776-9.CrossRefGoogle Scholar
  29. Mraihi, R., ben Abdallah, K., & Abid, M. (2013). Road transport-related energy consumption: Analysis of driving factors in Tunisia. Energy Policy, 62, 247–253.  https://doi.org/10.1016/j.enpol.2013.07.007.CrossRefGoogle Scholar
  30. Neves, S. A., Marques, A. C., & Fuinhas, J. A. (2017). Is energy consumption in the transport sector hampering both economic growth and the reduction of CO2 emissions? A disaggregated energy consumption analysis. Transport Policy, 59(July), 64–70.  https://doi.org/10.1016/j.tranpol.2017.07.004.CrossRefGoogle Scholar
  31. Nocera, S., & Cavallaro, F. (2016). The competitiveness of alternative transport fuels for CO2 emissions. Transport Policy, 50, 1–14.  https://doi.org/10.1016/j.tranpol.2016.05.013.CrossRefGoogle Scholar
  32. Omri, A. (2014). An international literature survey on energy-economic growth nexus: Evidence from country-specific studies. Renewable and Sustainable Energy Reviews, 38, 951–959.  https://doi.org/10.1016/j.rser.2014.07.084.CrossRefGoogle Scholar
  33. Payne, J. E. (2010). A survey of the electricity consumption-growth literature. Applied Energy, 87(3), 723–731.  https://doi.org/10.1016/j.apenergy.2009.06.034.CrossRefGoogle Scholar
  34. Pesaran, M. H. (2007). A simple panel unit root test in the presence of cross-section dependence. Journal of Applied Econometrics, 22(2), 265–312.  https://doi.org/10.1002/jae.951.CrossRefGoogle Scholar
  35. Razeghi, G., & Samuelsen, S. (2016). Impacts of plug-in electric vehicles in a balancing area. Applied Energy, 183, 1142–1156.  https://doi.org/10.1016/j.apenergy.2016.09.063.CrossRefGoogle Scholar
  36. Saboori, B., Sapri, M., & bin Baba, M. (2014). Economic growth, energy consumption and CO2 emissions in OECD (Organization for Economic Co-operation and Development)’s transport sector: A fully modified bi-directional relationship approach. Energy, 66, 150–161.  https://doi.org/10.1016/j.energy.2013.12.048.CrossRefGoogle Scholar
  37. Sanz, M. T., Cansino, J. M., González-Limón, J. M., Santamaría, M., & Yñiguez, R. (2014). Economic assessment of CO2 emissions savings in Spain associated with the use of biofuels for the transport sector in 2010. Utilities Policy, 29, 25–32.  https://doi.org/10.1016/j.jup.2014.04.002.CrossRefGoogle Scholar
  38. Shafiei, E., Davidsdottir, B., Leaver, J., Stefansson, H., & Asgeirsson, E. I. (2017). Energy, economic, and mitigation cost implications of transition toward a carbon-neutral transport sector: A simulation-based comparison between hydrogen and electricity. Journal of Cleaner Production, 141, 237–247.  https://doi.org/10.1016/j.jclepro.2016.09.064.CrossRefGoogle Scholar
  39. Talbi, B. (2017). CO2 emissions reduction in road transport sector in Tunisia. Renewable and Sustainable Energy Reviews, 69, 232–238.  https://doi.org/10.1016/j.rser.2016.11.208.CrossRefGoogle Scholar
  40. Tang, C. F., & Shahbaz, M. (2013). Sectoral analysis of the causal relationship between electricity consumption and real output in Pakistan. Energy Policy, 60, 885–891.  https://doi.org/10.1016/j.enpol.2013.05.077.CrossRefGoogle Scholar
  41. Tiba, S., & Omri, A. (2017). Literature survey on the relationships between energy, environment and economic growth. Renewable and Sustainable Energy Reviews, 69, 1129–1146.  https://doi.org/10.1016/j.rser.2016.09.113.CrossRefGoogle Scholar
  42. World Energy Council. (2011). World Energy Scenarios: Global Transport Scenarios 2050. https://www.worldenergy.org/wp-content/uploads/2012/09/wec_transport_scenarios_2050.pdf.
  43. Wu, H.-M., & Xu, W. (2014). Cargo transport energy consumption factors analysis: Based on LMDI decomposition technique. IERI Procedia, 9, 168–175.  https://doi.org/10.1016/j.ieri.2014.09.058.CrossRefGoogle Scholar
  44. Xu, B., & Lin, B. (2015a). Factors affecting carbon dioxide (CO2) emissions in China’s transport sector: a dynamic nonparametric additive regression model. Journal of Cleaner Production, 101, 311–322.  https://doi.org/10.1016/j.jclepro.2015.03.088.CrossRefGoogle Scholar
  45. Xu, B., & Lin, B. (2015b). Carbon dioxide emissions reduction in China’s transport sector: A dynamic VAR (vector autoregression) approach. Energy, 83, 486–495.  https://doi.org/10.1016/j.energy.2015.02.052.CrossRefGoogle Scholar
  46. Zhang, H., & Chen, W. (2015). The role of biofuels in China’s transport sector in carbon mitigation scenarios. Energy Procedia, 75, 2700–2705.  https://doi.org/10.1016/j.egypro.2015.07.682.CrossRefGoogle Scholar
  47. Zhang, H., Chen, W., & Huang, W. (2016). TIMES modelling of transport sector in China and USA: Comparisons from a decarbonization perspective. Applied Energy, 162, 1505–1514.  https://doi.org/10.1016/j.apenergy.2015.08.124.CrossRefGoogle Scholar
  48. Zhang, M., Li, H., Zhou, M., & Mu, H. (2011). Decomposition analysis of energy consumption in Chinese transportation sector. Applied Energy, 88(6), 2279–2285.  https://doi.org/10.1016/j.apenergy.2010.12.077.CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media B.V., part of Springer Nature 2018

Authors and Affiliations

  1. 1.Management and Economics DepartmentUniversity of Beira InteriorCovilhãPortugal
  2. 2.NECE-UBIUniversity of Beira InteriorCovilhãPortugal

Personalised recommendations