1

The causal relationship between CO2 emissions, energy usage, and

output in Gulf Cooperation Council (GCC) countries

Helmi HAMDI*, CAE/CERGAM Aix-Marseille University and Financial Stability

Directorate Central Bank of Bahrain| P.O Box 27, Manama, Kingdom of Bahrain

Rashid Sbia. Ministry of Finance- UAE.

Abstract

This study examines the causal relationship between carbon dioxide emissions, energy

consumption, and real output within a panel vector error correction model for six Gulf

Cooperation Council (GCC) countries namely Bahrain, Kuwait, Saudi Arabia, Qatar and

United Arab Emirates over the period 1980–2009. In the long-run, energy consumption has a

positive and statistically significant impact on carbon dioxide emissions. Moreover, there is a

dynamic relationship between carbon emissions and income, which confirms the presence

of the Environmental Kuznets Curve (EKC) for GCC countries. The short-run dynamics

results reveal a bi-directional causality between carbon and energy usage but reject the

existence of EKC.

Keywords: Carbon dioxide emissions Energy consumption, Growth, GCC, EKC

2

*Corresponding author: Tel: (973) 17547947| Mob (973)36003466 Fax: (973) 17532274. Tel: (973) 17547947|

Fax: (973) 17532274 email: helmi.hamdi@cbb.gov.bh

1. Introduction

The Gulf Cooperation Council (GCC) was founded on 26th May 1981 in order to promote

cooperation and coordination between the 6 countries (Bahrain, Kuwait, Oman, Qatar, Saudi

Arabia, and United Arab Emirates) in all socioeconomic fields and to build a union. One of

the common characteristics of the 6 countries is oil and gas production (with different level).

In 2010, the region produce 21.7% of the total worldwide oil production and it holds 35.8% of

the world’s proven oil (The proved reserve becomes 54.4% if we include Iraq and Iran

according to BP (2010)) and 22.5% of the world’s proven gas reserves. Production of energy

is the main contributor to the wealth of GCC countries. In 2010, hydrocarbon sector

represents 25% of the Gross Domestic Product in Bahrain, 52% in KSA, 57% in Qatar, 52%

in Kuwait, 54% in Oman and 34% in UAE (IMF 2011).

During the past few years, consumption of energy has been increasing drastically in the gulf

countries passing from 157 Million tonnes in 1999 to 263 Million tonnes in 2010, thus an

increase of 67.5%. Regarding natural gas, the consumption was 11.9 Billion cubic feet per day

in 2003; it becomes 21.7 Billion cubic feet per day in 2010, thus a change of 82.35% (BP,

2010). Nowadays, the region as a whole is becoming among the most pollutants in the world.

According to statistics above is it clear that a relationship between carbon emission, energy

consumption and economic growth in GCC countries exists. Hence, this issue is the

motivation of our study.

Generally speaking, the relationship between energy (emission and consumption) and

economic growth has been widely analyzed in the modern literature. Firstly it was studied by

3

Kraft and Kraft (1978) and then researchers have applied the same framework for many

countries around the world. Some studied have used single countries (Jumbe, 2004; Ghali and

El-Sakka, 2004; Morimoto and Hope, 2004; Wolde- Rufael, 2004) by the use of time series

analysis (with VAR, VECM, Cointegration and causality tests), and some other have used a

group of countries (Lee, 2005; Al-Iriani, 2006; Sari and Soytas, 2007; Lee and Chiang, 2008;

Wolde-Rufael, 2008; Narayan and Smyth, 2008; Chontanawat et al., 2008; Pao and Tsai,

2010) by the use of panel data analysis (GMM, Panel cointegration and panel causality tests).

Results find evidence of unidirectional (Morimoto and Hope, 2004; Wolde- Rufael, 2004;

Lee, 2005; Al-Iriani 2006), bidirectional (Jumbe, 2004; Ghali and El-Sakka, 2004; Wolde-

Rufael, 2005; Akinlo, 2008 Mahadeven and Asafu-Adjaye, 2007), or no causality (Huang et

al., 2008) according to the sample studied.

Despite the abundance of literature and the importance of energy in GCC countries, there is

no article in our knowledge which analyzes the relationship between energy emission, energy

consumption and economic growth in the region. Al-Iriani (2006) investigated the causality

relationship between gross domestic product (GDP) and energy consumption in the six GCC

countries; he used two variables which are GDP per capita and energy consumption. Our

paper try to fill the gap and investigate whether there is a causal relationship between energy

consumption, dioxide carbon emission and output. The reminder of the paper is as follows:

section 2 we present the econometric methodology and data, section 3 analyzes the empirical

results and section 4 concludes.

4

2. The Econometric methodology

According to literature of energy economics, the long-run relationship between carbon

dioxide emissions, energy consumption, and real GDP can be expressed as a linear logarithm

quadratic form and specified as follows:

ititiitiitiitit LYLYLECLCo 2

3212 (1)

Where i=1,…,N for each country in the panel and t=1,…,T refers to the time period.

2Co is the carbon dioxide emissions (measured in metric tons per capita); LEC is the energy

use (measured in kt of oil equivalent per capita); Y is the per capita real GDP (measured in

constant 2000 US dollars); and Y2 is the square of per capita real GDP. Variables of the

equation (1) are in natural logarithms; the parameters 1 , 2 and 3 represent the long-run

elasticity estimates of emissions with respect to energy usage, real GDP, and squared real

GDP, respectively. Broadly, the expected sign on energy consumption is positive because a

higher level of energy consumption resulted from buoyant economic activity lead to an

increase of CO2 emissions. Under the EKC hypothesis, the sign 2 is expected to be positive

while the sign for 3 is expected to be negative. If the coefficient of 2LY is statistically

insignificant, this means a monotonic increase in the relationship between per capita CO2

emissions and per capita income (Halicioglu, 2009). Finally, the error term is assumed to

be independent and identically distributed with a zero mean and constant variance

The empirical study is based on a panel for six GCC countries, annual data from

1980 to 2008 from the World Bank Development Indicators (WDI). Table 1 summarizes the

main statistics associated with carbon dioxide emissions per capita, energy use per capita and

5

real GDP per capita for each country and graphs illustrate the trajectory of these indicators

during the period of our study.

Table 1: Descriptive statistics

CO2emission Energy CONS. Real GDP

Mean S.D Mean S.D Mean S.D

Bahrain 25.16606 2.729218 9057.926 745.7835 17758.1 6876.53

Qatar 49.39338 13.10246 16903.81 2603.630 48551.5 12914.42

Oman 8.321521 3.859203 2835.352 1508.194 13402.9 5080.275

UAE 29.78562 4.977357 10708.18 1468.997 35458.2 7243.309

KSA 14.47233 1.847681 4677.442 869.9490 16738.4 2798.996

Kuwait 22.34945 8.653611 8348.626 2872.455 29529.0 7174.207

The mean of carbon dioxide emissions per capita ranges from 8.32 in Oman to 49.39

in Qatar. In addition, Qatar reveals the most variation in carbon dioxide emissions (defined by

the standard deviation) and Saudi Arabia the least variation 13.10 and 1.85 respectively.

As for the mean energy usage per capita Qatar has the highest mean usage (16903.81)

while Oman the least (2835.35) with Kuwait displaying the greatest variation (2872.46) and

Bahrain the least amount of variation (745.78). In terms of real GDP per capita, the highest

mean per capita income levels is in Qatar (48551.5) followed by UAE (35458.2) and the

lowest income is in Oman (13402.9). The variations in real GDP per capita are quite large

with the standard deviation of per capita income in Qatar at 12914.42 and for KSA at 2798.99.

6

0

10

20

30

40

50

60

70

80 82 84 86 88 90 92 94 96 98 00 02 04 06 08

CO2BH CO2KSA CO2KUW

CO2O CO2Q CO2UAE

Fig. 1. Per capitaCO2 emissions (before taking logarithm)

(metric tons of carbon dioxide per capita).

0

4,000

8,000

12,000

16,000

20,000

24,000

80 82 84 86 88 90 92 94 96 98 00 02 04 06 08

ENERGYBH ENERGYKSA ENERGYKUW

ENERGYO ENERGYQ ENERGYUAE

Fig. 2. Per capitaenergy use (before taking logarithm)

(kg of oi lequivalent per capita).

7

0

10,000

20,000

30,000

40,000

50,000

60,000

70,000

80,000

90,000

80 82 84 86 88 90 92 94 96 98 00 02 04 06 08

GDPBH GDPKSA GDPKUWGDPO GDPQ GDPUAE

Fig. 3. Per capita real GDP (before taking logarithm)

Our empirical investigation has two dimensions. The first is to examine the long-run

relationship between carbon dioxide emissions, energy usage and real GDP (Y), while the

second is to examine the short-run dynamic causal relationship between the variables. The

basic testing procedure requires three steps. The first step is to test whether the variables

contain a panel unit root to confirm the stationarity of each variable (Engle and Granger,

1987). This is done by using the Levin and Chu test, (LLC, 2002), the Im et al. test (2003)

(Im, Pesaran and Shin (IPS, 2003)) the Augmented Dickey–Fuller (F-ADF) (Maddala and

Wu, 1999; Choi, 2001) and finally Philips–Perron (PP) tests (1998). The second step is to test

whether there is a long-run cointegrating relationship between the variables. This is done by

the use of the Johansen-Fisher (Maddala and Wu, 1999; Kao, 1999; Pedroni, 1999, 2004)

methods. Finally, the last step, if all variables are I(1) (integrated of order one) and

cointegrated (Masih and Masih, 1996), short-run elasticities can be computed using the vector

error correction model (VECM) method suggested by Engle and Granger (1987). In this case,

8

an error correction mechanism exists by which changes in the dependent variables are

modeled as a function of the level of the disequilibrium in the cointegrating relationship,

captured by the error-correction term (ECT), as well as changes in the other explanatory

variables to capture all short-term relations among variables (Pao and Tsai, 2010).

3. Empirical results

3.1. Panel unit roots and cointegration tests for GCC

We use the panel unit root tests as proposed by Levin and Chu (LLC, 2002), Im, Pesaran and

Shin (IPS, 2003), the Augmented Dickey–Fuller (F-ADF) and finally Philips–Perron (PP,

1998). The results are displayed in Table 2. The test statistics for the log levels of CO2, EC,

GDP and GDP2 are statistically insignificant. When we apply the panel unit root tests to the

first difference of the four variables, all four tests reject the joint null hypothesis for each

variable at the 1 per cent level. Thus, from all of the tests, the panel unit roots tests indicate

that each variable is integrated of order one.

Table 2: Results of the unbalances Panel Unit Roots Tests for GCC Countries LLC IPS F-ADF PP Order of

Integration

level 1st diff level 1st diff, level 1st diff, level 1st diff,

lco2 -1.96 -4.68*** -1.03 -5.61*** 16.23 51.64*** 13.51 204.64*** I(1)

LEC -1.43 -3.56*** -0.56 -6.06*** 16.62 45.61*** 14.11 112.09*** I(1)

LY -1.77 -3.02*** -1.25 -4.07*** 20.16 37.70*** 13.23 74.79*** I(1)

lY2 -1.51 -3.13*** -0.97 -4.06*** 18.97 37.61*** 11.20 75.36*** I(1)

Note: The lag lengths are selected using SBC

*** Denotes the rejection of the null hypothesis at 1%level of significance

After checking the integration of our four variables at order one, I(1), the Pedroni, Kao and

Fisher tests for balanced (GCC) panel date are used.

9

Pedroni (1999, 2004) suggests two sets of tests for cointegration: the between and the

within dimensions. The within approach includes four statistics panel v- statistic, panel r-

statistic, panel PP-statistic, and panel ADF-statistic. These statistics pool the autoregressive

coefficients across different countries for the unit root tests on the estimated residuals taking

into account common time factors and heterogeneity across countries. The between approach

includes three statistics: group r-statistic, group PP-statistic, and group ADF-statistic. These

statistics are based on averages of the individual autoregressive coefficients associated with

the unit root tests of the residuals for each country. All seven tests are distributed

asymptotically as standard normal.

The test results of Pedroni displayed in table 3 reveal the rejections of the null of no

cointegration for all tests at 5 % level of significance except the group rho-tests and panel v-

test. However, according to Pedroni (2004), the two Pedroni test statistics which do not reject

the null hypothesis may have a very low power in the case of small time dimension. This fact

is also observed by Al-Iriani (2006) and Pao and Tsai (2010). Therefore, one may conclude

that our model is in fact panel cointegrated.

Table 3: Results of the balanced Panel Cointegration tests for GCC countries

Pedroni Residual Cointegration Test Statistics

Panel v-Statistic Weighted Statistic -1.360724

Panel rho-Statistic Weighted Statistic -1.821637

Panel PP-Statistic Weighted Statistic -2.54301 ***

Panel ADF-Statistic Weighted Statistic -2.05354**

Group rho-Statistic -0.37359

Group PP-Statistic -1.73254**

Group ADF-Statistic -1.392477**

Kao Test.

ADF -2.8082 (0.0025)***

Johansen Fisher Panel Cointegration Test

Null Hypo. Max-Eigen. Trace

10

None 44.90 (0.0000)*** 46.61 (0.0000)***

At most 1 10.73 (0.5526) 13.85. (0.3103)

At most 2 8.321 (0.7596) 10.07 (0.6095)

At most 3 16.45 (0.1716) 16.45 (0.1716)

Note: The optimal lag lengths are selected using SBC. Figures in parenthesis are

probability values.

Trace test and Max-eigenvalue test indicate 1 cointegrating vector at the 0.01 level

*** Denotes the rejection of the null hypothesis at 1%level of significance

The Kao test also suggests panel cointegration at 1% level of significance. In addition,

the Johansen Fisher test suggests the existence one cointegrating vectors at 1% of

significance. However, before the application of cointegration test, we selected the optimal lag

length of underlying VAR using the conventional model selection criteria. These criteria

established that the optimal lag length is one.

Overall, there is strong statistical evidence in favor of panel cointegration among CO2

emissions, energy consumption, GDP and GDP2 for GCC countries.

3.2. Panel Long run and short run

Generally speaking, the existence of cointegration signifies that there is at least one long-run

equilibrium relationship among the variables. In this case, Granger causality exists among

these variables in at least one way (Engle and Granger, 1987). The VECM is used to correct

the disequilibrium in the cointegration relationship, as well as to test for long and short-run

causality among cointegrated variables. The correction of the disequilibrium is done by the

mean of the Error correction term (ECT).

To test for panel causality, a panel-based VECM is specified as follows:

ttit

s

i

iit

r

i

i

p

i

it

q

i

iitit ectLYLYLECLCoLCo 111

2

1

1

1

1

1 1

111 22

(2)

11

ttit

s

i

iit

r

i

i

p

i

it

q

i

iitit ectLYLYLECLCoLY 212

2

1

2

1

2

1 1

222 2

(3)

ttit

s

i

iit

r

i

i

p

i

it

q

i

iitit ectLYLYLECLCoLY 313

2

1

3

1

3

1 1

333

22

(4)

ttit

s

i

iit

r

i

i

p

i

it

q

i

iitit ectLYLYLECLCoLEC 414

2

1

4

1

4

1 1

444 2

(5)

Where ECT is expressed as follows:

2

32102 LyLyLecLcoECT tttt , where t=1..T, denotes the time period (6)

The results of the long-run equilibrium relationship are presented in table 4 below. It shows

that the coefficient of lgdp is 61.02, which is positive and significant at the level of 1%. It

means that a 1% increase in per capita real GDP will increase per capita emissions by 61.02%

in the long- run. The coefficient of lgdp2 is negative (-2.92) and statistically significant at the

level of 1%. This shows that when the real GPD per capita reach a certain level a 1% increase

of its level will reduce the per capita emissions by 2.92%.

Table 4: CO2 Emission long-run elasticities

Dependent Variable: LCO2

Regressors coefficients t-value

LENERGY

1.31

-2.340***

LGDP 61.02 -3.821***

LGDP2 -2.92 3.699***

Note; ** and *** denote significance of coefficients at 5% and 10% levels of significance

The respective positive and negative signs of the income and its square term together confirm

the existence of Environmental Kuznets Curve in GCC countries. Accordingly, carbon

12

emissions increase essentially with increase in income, reaches to its stabilization point, and

then starts to decline with further increase in income.

As the objective of the study is to examine the dynamic relationship between dioxide

emission, energy consumption and growth it is opportune to study the hypothesis of the

presence of Environmental Kuznets Curve. Table 5 illustrates the results in which Dlc is the

dependent variable. Given that the optimal lag length was one, the short-run results are also

presented for one lag only of each variable. Results show that only energy act positively and

significantly at the level of 1% to co2 emission.

The results in Table 5 advocate that the Environmental Kuznets Curve (EKC) hypothesis does

not hold in the short-run for GCC.

It is also evident from Table 5 that the error correction term, although having the right sign, is

statistically not significant. The coefficient of the error-correction term is -0.004521,

suggesting that when per capita emission is above or below its equilibrium level, it adjusts by

almost 0.45% within the first year. Thus, the speed of adjustment towards equilibrium is not

enough fast in case of any shock to emission equation.

Table 5: CO2 emissions short-run elasticities for GCC countries

Dependante Variable Δlco2

Regressors Coefficient t-stat

Δlgdp(-1) 0.418984 0.10271

Δlgdp2(-1) 0.006764 0.03368

Δlenergy(-1) 0.339961 3.28182***

Δ Intercept -0.011295 -0.64327

Ect (-1) -0.004521 -0.43969

13

The existence of a panel long-run cointegration relationship among emissions, energy

consumption, GDP and GDP2 suggests that there must be Granger causality in at least one

direction. Thus, the next concern is to inspect the direction of causality amongst these

variables. The results of causality tests based on the VEC model are reported in Table 6. The

table has three major blocks illustrating the short-run effects, long-run effects represented by

the error correction coefficients, and the joint short-run and long run effects, respectively.

Table 6: Results of causality tests based on VECM.

Variable Short run (F-stats) ECT

(t-stats)

Joint short and long run (F-stats)

Δlco2 Δlgdp Δlgdp2 Δlenergy Δlco2 Δlgdp Δlgdp2 Δlenergy

Δlco2 - 0.01 0.001 10.77*** -0.43 - 0.11 0.15 5.7***

Δlenergy 3.22* 0.06 0.132 - 1.20 2.18 0.85 0.78 -

Δlgdp 6.31** - 1.211 0.001 1.90* 4.64** - 2.64* 1.84

Δlgdp2 5.96** 0.001 - 0.008 1.62 4.02** 1.92 - 1.356

The F-statistics for the short-run dynamics reveals a bi-directional causality between carbon

and energy usage. This results support our findings reported in Table 5 in which energy usage

is the only explanatory variable significant at the level of 1%. The results further show

energy; GDP and GDP2 are influenced by CO2 emission. Based on these results, we may

conclude that, in the short-run, there is unidirectional causality between growth and CO2

emissions.

Regarding error correction results, it is observed that deviation from the long-run equilibrium

is only corrected by GDP per capita; the other variables appears to be weakly exogenous. This

reveals the fact that any changes in CO2 emission, energy consumption and GDP2 that disturb

long-run equilibrium are corrected by counter-balancing changes in the real GDP per capita.

14

In this context, it may be concluded that GDP is caused by carbon emissions, energy

consumption, and GDP2 but these three variables are not caused by the former.

Turning now to the right side of table 6, the joint Wald F-statistics results indicate in the

carbon emission equation, error correction term and energy consumption are jointly

significant at a level of 1%. On the other hand, each of GDP and GDP2 combined with error

correction term are statistically insignificant. However, in the GDP and GDP2 equations,

carbon emission equation and error correction term are jointly significant at a level of 5%.

4. Concluding remarks

This study aims at analyzing the dynamic relationship between carbon dioxide emissions,

energy consumption, and real GDP for a panel of 6 GCC countries over the period 1980–

2008 and to obtain policy implications of the results. First set of tests show the existence of a

cointegration relationship and results of the long-run elasticities demonstrate that GDP per

capita is positive and significant at the level of 1%. This means that a 1% increase in per

capita real GDP will increase per capita emissions by 61.02% in the long- run. Indeed, these

results confirm the existence of Environmental Kuznets Curve in GCC countries in the long-

run. Turning now to the short-run dynamics; results reveal a bi-directional causality between

carbon and energy usage. In the short run, energy usage is the only explanatory variable

significant at the level of 1%. Our findings advocate that the Environmental Kuznets Curve

(EKC) hypothesis does not hold in the short-run for GCC. Thus, the absence of causality from

emissions to growth suggests that GCC countries can control their carbon emissions without

troubling their economic growth.

15

Disclaimer

The opinions expressed in this article are those of the authors and they do not necessarily

represent the official position of their institutions.

References

Akaike, H., 1969. Fitting autoregressive models for prediction Ann. Inst. Stat. Math. 21, 243-

247.

Akinlo, A.E., 2008. Energy consumption and economic growth: evidence from 11 African

countries. Energy Economics 30, 2391–2400.

Al-Iriani, M., 2006. Energy–GDP relationship revisited: an example from GCC countries

using panel causality. Energy Policy 34, 3342–3350.

Auty, R., 2001. The political state and the management of mineral rents in capital-surplus

economies: Botswana and Saudi Arabia. Resources Policy 27, 77–86.

Beaudreau, B., 2005. Engineering and economic growth. Energy Economics 16, 211–220.

BP (2011) BP Statistical Review of World Energy, November 2011, London.

Chiou-Wei, S.Z., Chin, C., Zhu, Z., 2008. Economic growth and energy consumption

revisited: Evidence from linear and nonlinear Granger causality. Energy Economics

30, 3063–3076.

Chontanawat, J., Hunt, L.C., Pierce, R., 2008. Does energy consumption cause economic

growth? Evidence from systematic study of over 100 countries. Journal of Policy

Modelling 30, 209–220.

Dickey, D.A., Fuller, W.A., 1979. Distribution of the Estimators for Autoregressive Time

Series with a Unit Root. Journal of the American Statistical Association, 74, 427- 431

16

ECA (Economic Commission for Africa), 2004. Economic Report: Unlocking Africa's Trade

Potential.

ECA (Economic Commission for Africa), 2008. Energy for Sustainable Development: Policy

Options for Africa UN-ENERGY/Africa: A UN collaboration mechanism and UN

sub-cluster on energy in support of NEPAD.

Ghali, K.H., El-Sakka, M.I.T., 2004. Energy use and output growth in Canada: a multivariate

cointegration analysis. Energy Econ. Vol. 26 (2), pp.225–38.

Granger, C.W.J., 1969. Investigating Causal Relationship by Econometric Models and Cross-

Spectral Methods. Econometrica 37, 424-438.

Granger, C.W.J., 1986. Developments in the study of cointegrated economic variables.

Oxford Bulletin of Economics and Statistics 48, 213-228.

Granger, C.W.J., 1988. Some recent developments in a concept of causality. Journal of

Econometrics 39, 199-211.

Guttormsen, A.G., 2004. Causality between Energy Consumption and Economic Growth.

Agricultural University of Norway, Department of Economics and Resource

Management, Discussion Paper #D-24.

Howells, M., Shrattenholzer, L., Joseph, E.A., 2008. Introduction, special section energy and

development in Africa. Energy Policy 36, 2771–2772.

Huang, B., Hwang, M.J., Yang, C.W., 2008. Causal relationship between energy consumption

and GDP growth revisited: a dynamic panel data approach. Ecological Economics 67,

41–54.

IEA (International Energy Agency), 2005. World energy outlook: energy and poverty, Paris,

2002 IEA (International Energy Agency). Energy Statistics.

17

IMF, 2010. Gulf Cooperation Countries: Enhancing Economic Outcomes in an uncertain

Global Economy. Middle East and Asian Department. IMF Washington.

Jumbe, C.B.L., 2004. Cointegration and Causality between Electricity Consumption and

GDP: Empirical Evidence from Malawi. Journal of Energy Economics 26(1): 26-68.

Konya, L., 2004. Export-Led Growth, Growth-Driven Export, Both or None? Granger

Causality Analysis on OECD Countries. Applied Econometrics and International

Development, Euro-American Association of Economic Development, vol. 4(1).

Konya, L., 2004. Unit-Root, Cointegration and Granger Causality Test Results for Export and

Growth in OECD Countries" International Journal of Applied Econometrics and

Quantitative Studies, Euro-American Association of Economic Development, vol.

1(2).

Kraft, J., Kraft, A., 1978. On the relationship between energy and GNP. Journal of Energy

Development 3, 401-403.

Lee, C.C., 2005. Energy consumption and GDP in developing countries: A cointegrated panel

analysis. Energy Economics 27, 415–427.

Lee, C.C., Chiang, C., 2008. Energy consumption and economic growth in Asian countries: a

more comprehensive analysis using panel data. Resource and Energy Economics 30,

50–65.

Lee, C.C., Chang, C.P., Chen, P.F., 2008. Energy–income causality in OECD countries

revisited: the key role of capital stock. Energy Economics 30, 2359–2373.

Loizides, J., Vamvoukas, G., 2005. Government expenditure and economic growth: evidence

from trivariate causality testing. Journal of Applied Economics 8, 125–152.

18

Lütkepohl, H., 1982. Non-causality due to omitted variables. Journal of Econometrics 19,

367–378.

Mahadeven, R., Asafu-Adjaye, J., 2007. Energy consumption, economic growth and prices: a

reassessment using panel VECM for developed and developing countries. Energy

Policy 35, 2481–2490.

Masih, A.M.M., Masih, R., 1996. Energy consumption, real income and temporal causality:

results from a multi-country study based on cointegration and error-correction

modelling techniques. Energy Economics 18 (3), 165-183.

Morimoto, R., Hope, C., 2004. The impact of electricity supply on economic growth in Sri

Lanka. Energy Economics 26 (1): 77-85.

Narayan, P.K., Smyth, R., 2008. Energy consumption and real GDP in G7 countries, new

evidence from panel cointegration with structural breaks. Energy Economics 30,

2331–2341.

Odhiambo, N.M., 2008. Financial depth, savings and economic growth in Kenya: A dynamic

causal linkage. Economic Modelling 25, 704–713.

Oh, W., Lee, K., 2004. Causal relationship between energy consumption and GDP revisited:

the case of Korea 1970–1999. Energy Economics 26, 51-59.

Pesaran, M.H., Shin, Y., 1998. Generalised impulse response analysis in linear multivariate

models. Economics Letters 58, 17–29.

Phillips, P.C.B., Perron, P., 1988. Testing for a unit root in time series regression. Biometrika

75, 335–346.

Prasad, G., 2008. Energy sector reform, energy transitions and the poor in Africa. Energy

Policy 36, 2785–2790.

19

Sari, R., Soytas, U., Ozlem, U., 2001. Energy Consumption and GDP Relations in Turkey: A

Cointegration and Vector Error Correction Analysis. Economies and Business in

Transition: Facilitating Competitiveness and Change in the Global Environment

Proceedings, s. 838-844: Global Business and Technology Association.

Shan, J., 2005. Does financial development ‘lead’ economic growth? A vector autoregression

appraisal. Applied Economics 37, 1353–1367.

Soytas, U., Sari, R., 2003. Energy consumption and GDP: causality relationship in G-7

countries and emerging markets. Energy Economics 25, 33-37.

Soytas, U., Sari, R., 2006. Energy consumption and income in G-7 countries. Journal of

Policy Modelling 28, 739–750.

Wolde-Rufael, Y., 2008. Energy consumption and economic growth: The experience of

African countries revisited, Energy Economics Volume 31, Issue 2, March 2009,

Pages 217–224