View metadata, citation and similar papers at core.ac.

uk brought to you by CORE

provided by University of Southern Queensland ePrints

Is the long-run relationship between economic growth, electricity consumption, carbon

dioxide emissions and financial development in Gulf Cooperation Council Countries
robust?

Mohammad Salahuddin
Australian Digital Future Institute and School of Commerce, University of Southern
Queensland, Toowoomba, Australia.
Tel: +61470592674. Email: salahuddin.mohammad@usq.edu.au

Jeff Gow
School of Commerce
University of Southern Queensland, Toowoomba, QLD 4350, Australia and Department of
Agricultural Economics, Stellenbosch University, Stellenbosch, South Africa
Email: Jeffrey.Gow@usq.edu.au

Ilhan OZTURK
Faculty of Economics and Administrative Sciences,
Cag University, 33800, Mersin, Turkey.
Email: ilhanozturk@cag.edu.tr

Abstract
The relationship between carbon dioxide emissions, economic growth, electricity
consumption and financial development in the Gulf Cooperation Council (GCC) countries is
investigated in this study using panel data for the period of 1980-2012. A number of
econometric techniques: dynamic ordinary least squares (DOLS), fully modified ordinary
least squares (FMOLS) and the dynamic fixed effect model (DFE) are applied in order to
estimate the long-run relationship between the variables. The long-run relationship is found to
be robust across these different econometric specifications. No significant short-run
significant relationship was observed. Electricity consumption and economic growth have a
positive long run relationship with carbon dioxide (CO 2 ) emissions whilst a negative and
significant relationship was found between CO 2 emissions and financial development. The
findings imply that electricity consumption and economic growth stimulate CO 2 emissions in
GCC countries while financial development reduces it. Granger causality results reveal that
there is a bidirectional causal link between economic growth and CO 2 emissions and a
unidirectional causal link running from electricity consumption to CO 2 emissions. However,
there is no causal link between financial development and CO 2 emissions. Also, impulse
response and variance decomposition analysis outline forecasted impacts of economic growth
and electricity consumption on future CO 2 emissions.

Keywords: CO 2 emissions, economic growth, electricity consumption, GCC countries,

financial development
JEL codes: Q32, Q43

1
1. Introduction
Economic development through economic growth is characterized by the usually close
relationship with increasing levels of electricity usage and energy more generally, as well as
an associated increase in carbon dioxide emissions (CO 2 ). Additionally, an increase in the
level and breadth of a country's financial markets to fund economic development is
commonly witnessed. The extent of the influence and the interrelationship of these variables
does though vary from country to country.
In this study these relationships are investigated for the six Gulf Cooperation Council
countries (GCC): Saudi Arabia, United Arab Emirates (UAE), Qatar, Bahrain, Kuwait and
Oman. All of these countries have experienced rapid economic growth over the past 40 years
due mainly to their vast oil and gas reserves. These include approximately 40% of the world's
proven oil reserves and approximately 25% of the world’s natural gas reserves. They
contribute approximately 8% of world CO 2 emissions (Al-Saleh et al. 2012). Given the
pivotal role of oil and gas in driving the world economy and the rapid increase in economic
activity worldwide over the past 40 years, it is perhaps unsurprising that the GCC countries
are amongst the highest per capita carbon dioxide (CO 2 ) country emitters (Hertog and
Luciani, 2009).
The rapid level of economic development in the GCC countries has been associated
with high rates of economic growth, electricity consumption and CO 2 emissions. Rates of
economic growth and per capita electricity consumption have surpassed the levels of the
major developed economies of the Organization of Economic Cooperation and Development
countries (OECD) (World Bank, 2013).
The impact of the 1970s oil shocks upon the energy-mix of the major developed
economies was such that it negated the otherwise strong previous association between CO 2
emissions and economic growth as they sought to insulate their economies from high oil and
gas prices. This influence was absent in the GCC countries as their governments used their
cheap and abundant energy to rapidly develop their economies. This rapidly growing
domestic demand for energy in the GCC countries was particularly evidenced for electricity.
This situation has now led to three of the six GCC countries being the world’s highest CO 2
emitters. Given this scenario, the regions' commitment to sustainable energy policies appears
to be a priority.
This study investigates the short and long-run relationships amongst economic growth,
electricity consumption, CO 2 emissions and financial development in the region and also
determines the causal direction between the variables. Relatively little attention has been paid
to the environmental sustainability of this region despite their being significant sources of
global energy supply and the potential impacts of this consumption on the environment. This
study is an attempt to fill this gap and more importantly, offers a discussion on policy options
to achieve sustainability in regional energy systems.
This study contributes primarily by focusing exclusively on the GCC countries and a
discussion of solutions to ensure environmental sustainability in the region. A secondary
contribution is methodological as it applies a number of sophisticated econometric
techniques: the dynamic ordinary least squares (DOLS), the fully modified ordinary least
squares (FMOLS) and the dynamic fixed effect model (DFE) to estimate the long-run
relationship between the variables. Also panel Granger causality is employed to determine the
causal direction between the variables. The robustness of the causal link is checked by the
Innovation Accounting Approach (IAA) that consists of impulse response functions and
variance decompositions. Therefore, the objective of this study is to assess the
interrelationship of the variables of interest in the GCC countries over the period 1980-2012.
The rest of the paper is structured as follows: section 2 contains the literature review
while section 3 is dedicated to the discussion of the model and the estimation methods used.

2
Section 4 presents the results whilst section 5 offers a discussion of the results. The paper
ends with conclusions, policy implications and recommendations in section 6.

2. Literature Review
2.1 Literature review on CO 2 emissions, energy consumption and economic growth
The recent literature has focused on the relationship between energy consumption,
environmental pollution and economic growth. Many panel data and time series studies have
been done on the relationship between these variables. Saboori et al. (2014) investigated and
estimated the bi-directional long-run relationship between energy consumption, carbon
dioxide (CO 2 ) emissions and economic growth in the road transport sector of all OECD
countries. Using the Fully Modified Ordinary Least Squares (FMOLS) method, this study
confirmed that there is a positive significant bi-directional relationship between CO 2
emissions and economic growth, road sector energy consumption and economic growth, and
between CO 2 emissions and road sector energy consumption. Also the authors found that
most of the CO 2 emissions occurred as a result of energy consumption. In addition, the study
stressed the need to shift to other options for energy, such as biofuel, renewable and nuclear
energy, and the importance of long-run policies that aim to enhance energy efficiency. Hamdi
et al. (2014) examined the relationship between electricity consumption, foreign direct
investment, capital and economic growth in the case of Bahrain. Their causality analysis
supported the feedback effect between electricity consumption and economic growth. Cowan
et al. (2014), in a study of the BRICS countries, found support for the neutrality hypothesis
for Brazil, India and China, indicating that there is no association between electricity
consumption and economic growth. However, regarding the GDP–CO 2 emissions
relationship, a feedback hypothesis for Russia, that is a one-way Granger causality running
from GDP to CO 2 emissions, and a reverse relationship were found for Brazil both resulting
in inconclusive policy implications.
Sbia et al. (2014) investigated the empirical relationship between foreign direct
investment, clean energy, trade openness, carbon emissions and economic growth using time
series data for the United Arab Emirates (UAE). They found that energy demand had negative
relationship with foreign direct investment, trade openness and carbon emissions while
economic growth and clean energy stimulated energy consumption. In the case of Saudi
Arabia, Alkhatlan and Javid’s (2013) study revealed a positive relationship between CO 2
emissions and economic growth. They also concluded that electricity produces less pollution
than other sources of energy. Hamdi and Sbia (2013) examined the causal relationship
between carbon dioxide emissions, energy consumption and real output for a panel of Gulf
Cooperation Council (GCC) countries namely Bahrain, Kuwait, Saudi Arabia, Qatar and
UnitedArab Emirates over the period 1980–2009. Their empirical exercise supported the
presence of Environmental Kuznets Curve (EKC) hypothesis for these countries only in the
long-run. They also found bidirectional causality between carbon emissions and energy usage
in the short-run. Ozcan (2013) tested the EKC hypothesis for 12 Middle East countries using
panel data for the period 1990-2008. The study provided evidence for a U-shaped EKC for
five countries and an inverted U-shaped curve for three countries. No causal link between
income and CO 2 emissions was observed for the other four countries. The direction of
causality was mixed for different countries. Al-Mulali and Tang (2013) tested the pollution
haven hypothesis for the GCC countries. In other words, they investigated the effect of FDI
on CO 2 emissions. Their results suggested that increased FDI reduces CO 2 emissions. Also
the study reported that energy consumption and GDP growth stimulate CO 2 emissions. Liao
and Cao (2013), in a large panel of 132 countries, revealed that factors like urbanisation,
population density, trade, energy mix, and economic environment affect the level of CO 2
emissions.

3
 Al-Mulali (2012) undertook a large time series study involving seven different regions
– East Asia and Pacific; Eastern Europe and Central Asia; Latin America and the Caribbean;
Middle East and North Africa; South Asia; Sub-Saharan Africa; and Western Europe. Using
the Fully Modified Ordinary Least Squares (FMOLS) method, the authors investigated the
relationship between urbanization, energy consumption and CO 2 emissions. Their results
indicated a positive long-run relationship between the variables in six of the regions, while
findings varied for the remaining region. Some of the regions demonstrated a negative
relationship between the variables, while others – especially the low income countries – did
not show any relationship at all. Ozturk and Acaravci (2010) studied the relationship between
income, energy consumption, CO 2 emissions and employment in Turkey. They found that
neither CO 2 emissions per capita nor energy consumption per capita Granger-cause real GDP
growth per capita.
The literature studying the relationship between CO 2 emissions, energy consumption
and economic growth involving only the GCC countries is relatively scarce, although there
have been several studies on the APEC (Association of Petroleum Exporting Countries), as
well as the MENA (Middle East and North African) countries, and high income countries
which include all or some of the GCC countries. The study of Omri (2013) on 14 MENA
countries found that there is a bi-directional causal link between energy consumption and
economic growth in the region. Ozcan (2013) tested 12 Middle East countries and found the
EKC hypothesis supported in only three countries and in a further six countries it found no
support and no causal link was found in the other three countries. Arouri et al. (2012), in a
study of 12 MENA countries, showed that energy consumption has a positive and significant
impact on CO 2 emissions and that real GDP demonstrates a quadratic relationship with CO 2
emissions. Narayan and Popp (2012) found that the EKC hypothesis was not supported for a
panel of 43 countries including Middle Eastern countries and observed that for the Middle
Eastern panel, the income elasticity in the long run is smaller than the short run estimate
implying that an increase in income causes a decline in CO 2 emissions. Jaunky (2011) tested
the EKC hypothesis for 36 high income countries including three MENA countries – Bahrain,
Oman and UAE, with the results indicating that CO 2 emissions decline with a rise in income
in the long run. Soytas and Sari (2009) investigated the association between carbon emissions,
income, energy and total employment in five selected OPEC countries, including Saudi
Arabia, and found a cointegrating relationship between the variables.

2.2 Literature review on CO 2 emissions and financial development

Chang (2015) examined the non-linear effects of financial development and income on
energy consumption. The study used five indicators of financial development for a panel of 53
countries for the period 1999-2008. The sample was split into two regimes: high income and
non-high income countries. The findings indicate that energy consumption increases with
higher levels of financial development when financial development is measured as the share
of GDP of private and domestic credit. Ziaei (2015) investigated the effects of two indicators
of financial development (credit and stock markets) on energy consumption and CO 2
emissions. The results reveal that financial development reduces CO 2 emissions when the
stock market is considered as an indicator of financial development. Boutabba (2014)
examined the long-run equilibrium relationship and causal link among CO 2 emissions,
financial development, energy consumption and trade openness for India. The findings
indicate a positive significant long-run impact of financial development on CO 2 emissions,
that is, financial development increases CO 2 emissions in an unidirectional causal link
running from financial development to CO 2 emissions.Omri et al. (2014) investigated the
causal link between economic growth, financial development and CO 2 emissions in a global
panel of 54 countries. Their findings indicate a bi-directional causal link between economic

4
growth and CO 2 emissions for the sub panels of Middle East, North Africa and Sub Saharan
countries and a unidirectional causality running from CO 2 emissions to economic growth for
other regions.
Ozturk and Acaravci (2013) found that an increase in foreign trade to GDP ratio
results an increase in per capita CO2 emissions and the financial development variable has no
significant effect on per capita CO2 emissions in the long run for Turkey-. These results also
support the validity of the EKC hypothesis in the Turkish economy. Shahbaz et al. (2013a) in
a time series study applied the ARDL bounds testing approach to cointegration to examine the
influence of financial development on CO2 emissions in Malaysia. Their findings indicate a
positive and significant relationship between the variables. Their findings imply that an
economy with more developed financial markets tend to attract more investment and thus
facilitate more industrialization which contributes towards higher level of energy
consumption eventually leading to higher level of CO2 emissions. Shahbaz et al. (2013b) in
another study investigated the relationship amongst economic growth, energy consumption,
financial development, trade openness and CO2 emissions in Indonesia. Their results confirms
a long-run cointegrating relationship among the variables. The study further reports that
financial development reduces CO2 emissions, in other words, financial development
improves environmental degradation levelsIn a recent panel study (Al-Mulali et al., 2014)
found that financial development was one of the factors that increased energy consumption in
GCC countries. The results further observed a cointegrating relationship between GDP,
urbanization, total trade and financial development. Financial development was found to
stimulate energy consumption and CO2 emissions in sub Saharan African countries (Al
Mulali, 2012) through an increase of investment in energy intensive industries whilst Shahbaz
and Lean (2012) obtained similar results for Tunisia.
Zhang (2011) investigated the impact of financial development on CO 2 emissions for
China and found financial development was a significant factor. Sadorsky (2011) examined
the effect of financial development on energy consumption for a panel of nine Central and
Eastern European economies. The findings supported the positive influence of financial
development on energy consumption. Sadorsky observed similar findings in an earlier study
(Sadorsky, 2010) that investigated the effect of financial development on energy consumption
in emerging economies. Tamazian and Rao (2010) recognized financial development as an
important driver of environmental performance. They argued that a more financially
developed market would provide more resources for environmental projects at a cheaper
price. Tamazian et al. (2009) found that a high degree of financial development is associated
with better environmental conditions. Jalil and Feridun (2011) found that financial
development reduces CO 2 emissions whilst Zhang and Cheng (2009) found the opposite in
these two China studies. Yuxiang and Chen (2010) argued that a country with a more
developed and sound financial system would enable industries to adopt and use advanced
state-of-the-art technologies that are less carbon intensive. They further suggested that
financial development helps economies enforce environmentally friendly regulations.
Overall the review on the relationship between financial development and CO 2
emissions suggests that the results are mixed although most of the investigations support the
view that higher levels of financial development is positively associated with declining levels
of CO 2 emissions. As seen from the literature, there is a limited number of studies for the
GCC countries on this issue. Therefore, this study aims to fill this gap in the literature.

3. Empirical model and econometric methods

An econometric model of the following form is estimated:
C it = β i +β 1 lnE it +β 2 lnY it + β 3 FD it + ε it (1)

5
The coefficients , β 1 and β 2 represent the long run elasticity estimates of CO 2 emissions with
respect to energy consumption and per capita GDP, as an increase in electricity consumption
and income are expected to cause an increase in CO 2 emissions. The effect of financial
development on CO 2 emissions cannot be anticipated at this stage as the literature offers
inconclusive evidence about this relationship.
To estimate the model, the following actions were taken in a step wise process: (i) a
cross-sectional dependence (CD) test was performed to verify whether there is cross-sectional
dependence across the panel; (ii) once cross-sectional dependence is observed, an appropriate
panel unit root test (i.e. CIPS) was conducted to examine the stationarity of the series; (iii) the
Pedroni cointegration test which verifies the long-run relationship among the variables was
then conducted; (iv) panel DOLS and panel FMOLS were employed to estimate the long-run
relationship while the DFE estimation technique was applied to estimate the short-run and
long-run relationships among the variables v) a VECM Granger causality test was conducted
to assess causality between the variables and finally vi) the robustness of the causal direction
of the relationship was checked by using an Innovation Accounting Approach (IAA) through
impulse response functions and variance decomposition analysis.
To investigate the relationships, data for the following variables were sourced:
- Per capita CO 2 emissions (C)
- Per capita electricity consumption (E)
- Per capita real GDP (Y)
- Financial development (FD) - domestic credit available to the private sector as share of GDP.
The World Development Indicators database 2013 was the source of the data for all
six countries (World Bank, 2013). Real GDP per capita (Y) which is measured at constant
2000 US$ was used, per capita electricity use (kWh) and per capita CO 2 emissions were
estimated by dividing total electricity and CO 2 emissions by the mid-year population. The
variables were then transformed into natural logs. This transformation was intended to
overcome the problem of heteroscedasticity between the variables.

3.1. Testing for Unit Roots

It is argued that (Asteriou, 2009) long-run parameters are likely to demonstrate
cointegrating relationships among a set of I(1) variables. In other words, it is expected that the
macroeconomic variables in the model will be characterized by a unit root process (Nelson
and Plosser, 1982). Therefore, determining the order of integration of the variables is the next
priority in estimation and the conducting of unit root tests for all variables achieves that aim.
Cross-sectional dependence is to be expected amongst this group of six homogenous six
countries. An examination of the presence of contemporaneous correlation across the
countries was achieved by implementing a cross-sectional dependence (CD) test developed by
Pesaran (2004) who defines the CD statistic as:

(2)
where

in which is the pair-wise cross-sectional correlation coefficients of residuals from the

conventional ADF regression; T and N are sample and panel sizes, respectively.
Because the CD test indicates the presence of cross-sectional dependence in the panel, the
following cross-sectionally augmented Dickey-Fuller (CADF) regression was used:
, t= 1..............T and i=1.....N (3)
where,

6
is the cross-sectional mean of y it . The purpose of including the cross-sectional mean in the
above equation is to control for contemporaneous correlation among y it . This is a modified
version of the IPS test (Im et al., 2003) and is referred to as the cross-sectionally augmented
IPS (CIPS) test (Pesaran, 2007). The null hypothesis of the test can be expressed as H 0 : β 1 =0
for all i against the alternative hypothesis H 0 : β <0 for some i. The test statistic provided by
Pesaran (2007) is given as:

where t i (N, T) is the t statistic of β i in equation (2). The critical values of CIPS (N, T) are
available in Table II(c) of Pesaran (2007).

3.2. Panel Cointegration

If the results from the CIPS unit root test indicate a cointegrating relationships in the
dataset then several panel cointegration tests suggested by Pedroni (1997, 1999) need to be
conducted. The Pedroni cointegration test controls for country size and heterogeneity which
allows for multiple regressors of the cointegration vector to vary across various panel sections
(Pedroni, 1999). Seven panel cointegration statistics for seven tests are obtained. Four are
within-dimension tests, whilst three are between-dimension or group statistics approach. The
methodology used for the panel cointegration is reproduced in Salahuddin and Gow (2014,
49-50).

3.3. Estimation of panel cointegration regression

If a cointegrating relationship between the variables is found, the next step is to
estimate the long-run parameters. Since in the presence of cointegration, OLS leads to
spurious coefficients, a number of alternative econometric methods are proposed. One such
method is the panel dynamic OLS (DOLS) which is believed to provide better results for
cointegrated panels. However, one major weakness of DOLS (Kao and Chiang, 2000) is that
it does not consider the cross sectional heterogeneity issue. Pedroni (2000; 2001) proposed the
fully modified OLS (FMOLS) estimator for cointegrated panels which takes into account the
cross sectional heterogeneity, endogeneity and serial correlation problems. The FMOLS
technique is also believed to provide consistent estimates in small samples (Pedroni, 2001).

3.4. Dynamic Fixed Effect Model

One common shortcoming of both the DOLS and FMOLS methods is that they do not
estimate short-run relationships (Murthy, 2007). Alternative methods such as the pooled mean
group (PMG) regression, mean group regression (MG) and the dynamic fixed effect (DFE)
model are available to consider different levels of heterogeneity across countries while
estimating both the short-run and the long-run effects simultaneously. The DFE imposes
homogeneity restrictions on the long-run and short-run coefficients while allowing the
intercept to vary. Since GCC countries are characterized by similar macroeconomic structures
(oil-based economies), the application of the DFE model is justified in this case. There may
be different types of temporary shocks in different GCC countries due to local laws,
regulations and political regimes and this heterogeneity is captured by country-specific
intercepts.
In practice, contemporaneous correlation across residuals arises from omitted common
factors. To eliminate the influence of these common factors allowance for time-specific
effects in the estimated regressions are made. In order to comply with the requirements for

7
standard estimation and inference, equation 1 is embedded into an ARDL (p, q) model. In
error correction form, this can be written as follows:
p-1 q-1

∆(y i ) t = ∑ γ j ∆(y i ) t-1 + ∑ δ j ∆(x i ) t-1 + [(y i ) t-1 - β 1 (X i ) t-1 }]+ β 0 +

i i i i
μt +Ɛ it (4)
j=1 j=0
where y i is the dependent variable (CO 2 emissions), Xs are independent variables (electricity
consumption, economic growth and financial development), γ j and δ j are short run
i i

coefficients, β 1 are the long- run coefficients, β 0,

i
μ
t and Ɛ it are country-specific fixed
i

effects, time-specific effects and stochastic error term respectively.

3.5 Panel Granger Causality test

If the variables are found to be first difference stationary [I(1)], then to assess the
causal direction of the relationship between them further tests are required (Granger, 1969).
Information about the exact direction of the causal link enables a more nuanced discussion of
the policy implications of the findings (Shahbaz et al., 2012).

3.6 Impulse response and variance decomposition

One major weakness of the VECM Granger causality test is that it is unable to provide
reliable estimates of the causal strength of relationship between variables beyond the selected
sample period. Another limitation is that it provides only the direction of the relationship, not
the corresponding sign. To overcome these limitations, this study applies the Innovation
Accounting Approach (IAA) which consists of variance decomposition and generalized
impulse response functions. The generalized impulse response function is preferred over the
simple Choleski fractionalization impulse response analysis as the generalized impulse
response function is insensitive to the order of the VECM. It also indicates whether the
impacts of innovations are positive or negative or whether they have a short-run or long-run
effect. The general representation of this procedure is available in the seminal works of Sims
(1980, 1986) and Bernanke (1986). Although impulse response function traces the effect of a
one standard deviation shock on the current and future values of all the endogenous variables
through the dynamic structure of VECM, it doesn't provide the magnitude of such effect.
Consequently, the variance decomposition method is employed to examine this magnitude.
Variance decomposition (Pesaran and Shin, 1999) measures the percentage
contribution of each innovation to h-step ahead of the forecast error variance of the dependent
variable and provides a means to determine the relative importance of shocks in explaining
the variation in the dependent variable. Engle and Granger (1987) argued that the variance
decomposition approach produces more reliable results as compared to those from other
traditional approaches.

4. Results
Table 1 reports the summary statistics which shows that the data are well behaved. The
standard deviations show that the data are homogeneous.
<Please insert table 1 here>
Table 2 presents the results of the CD test and CIPS unit root test. The CD test results
confirms cross sectional dependence in two of the three series (GDPC and energy
consumption). The CIPS unit root test proves all variables to be stationary at first difference,
i.e. I(1).
<Please insert table 2 here>

8
The Pedroni panel cointegration test results are presented in table 3. Six of seven tests reject
the null hypothesis of no cointegration. The group rho statistic has the best statistical power of
all the tests (Gutierrez, 2003), and it also rejects the null of no cointegration. Therefore, there
is evidence that there is a long run cointegrating relationship among the variables.
<Please insert table 3 here>
Table 4 presents the results from the DOLS estimates. The estimates suggest positive and
significant long-run relationships of electricity consumption and economic growth with CO 2
emissions. Financial development demonstrates a negative and significant association with
CO 2 emissions.
<Please insert table 4 here>
FMOLS estimates which produced similar results to DOLS but with slightly different
coefficient values are reported in table 5. The FMOLS results indicate a positive and highly
significant relationship between electricity consumption and economic growth with CO 2
emissions. Financial development has a highly significant negative effect on CO 2 emissions
which means financial development reduces CO 2 emissions.
<Please insert table 5 here>
Table 6 provides the results from the DFE estimation. Overall results suggest that the long run
coefficient of CO 2 emissions to electricity consumption is 0.61 and this is significant at the
5% level. In other words, a 1% increase in electricity consumption enhances CO 2 emissions
by 0.61% in the long run. There is also a significant and positive long run relationship
between economic growth and CO 2 emissions. The long run coefficient of CO 2 emissions to
economic growth is 0.40 which means a 1% increase in real GDP per capita causes a 0.40%
increase in CO 2 emissions. There is no significant short-run relationship among these
variables.
<Please insert table 6 here>
In table 7, the panel vector error correction model (VECM) Granger causality findings are
reported. There is a bidirectional causal link between energy consumption and CO 2 emissions
and a unidirectional causal link from economic growth to energy consumption. The
relationship between economic growth and CO 2 emissions has no causal link.
<Please insert table 7 here>
From Figure 1, it can be seen that the standard deviation of per capita CO 2 emissions leads to
a positive increase in future per capita CO 2 emissions. The response of per capita CO 2
emissions to the increases in electricity consumption and per capita GDP demonstrate the
expected signs but with different magnitudes. The accumulated response of per capita CO 2
emissions to electricity consumption is positive and significant and to GDP per capita is also
positive and significant. The response of per capita CO 2 emissions to future shocks of
financial development is negative and significant. Thus, these findings are supportive of all
earlier econometric estimations.
<Please insert figure 1 here>
Results from the variance decomposition analysis are reported in Table 8. The study allows
for a 32 year forecasting horizon. Interestingly, at the 5-year forecasting horizon, about 94%
of the one-step forecast variance in per capita CO 2 emissions is accounted for by its own
innovations and altogether 6% is accounted for by economic growth, electricity consumption
and financial development. In the long-run, the response to own innovative shocks declines to
around 65% while the response of per capita electricity consumption to the shocks in per
capita CO 2 emissions, economic growth and financial development are expected to rise to
35% from the first 5-year forecast horizon of 6%. Amongst the 35% of the variance,
approximately 23% of variance is due to the shocks in per capita electricity consumption and
around 11% variations are attributed to GDP per capita while the rest, 0.76%, is due to the
shock in financial development. The findings reinforce that while per capita electricity

9
consumption is likely to have a very strong forecasted impact on per capita CO 2 emissions,
the impact of economic growth is also likely to be evident in the future. However, the
forecasted impact of financial development seems to be weak.
<Please insert table 8 here>

5. Discussion
This study investigated the effects of economic growth, electricity consumption and
financial development on CO 2 emissions in GCC countries using panel data for the period of
1980-2012. CIPS panel unit root tests were conducted that account for cross sectional
dependence and find that all variables first difference stationary. The Pedroni cointegration
test confirms a cointegrating relationship among the variables. Group DOLS andFMOLS
were employed to estimate the long-run relationship among the variables. The panel
econometric technique, the DFE model, was estimated to examine both the short-run and the
long-run relationship between CO 2 emissions and economic growth, electricity consumption
and financial development. Group DOLS and FMOLS were also employed to test the
robustness of the long-run relationship among the variables. Economic growth and electricity
consumption were found to have a positive significant impact on CO 2 emissions in the long-
run while no significant short-run relationship between these variables was observed. The
findings of the long-run association between electricity consumption and economic growth
are in contrast with the results of a recent study on GCC countries (Hamdi and Sbia, 2014).
Financial development was found to reduce CO 2 emissions in the long-run. A bi-directional
causal link was found between economic growth and CO 2 emissions. This implies that
although the GDPs of the GCC countries are largely oil based, their oil based revenues
generate high incomes for their citizens and a massive influx of foreign workers both of which
leads to a sharp rise in energy demand. To meet the growing energy demand, enormous
amounts of electricity are generated, mostly from fossil fuel sources.
A unidirectional causal link running from electricity consumption to CO 2 emissions is
found to exist. No causal link was found between financial development and CO 2 emissions.
Impulse response functions and variance decomposition analysis reveal that per capita
electricity consumption and economic growth will continue to impact CO 2 emissions
significantly into the future while the impact of financial development is expected to be of
little magnitude. Therefore, the GCC countries will have to look for alternative sources of
power generation as well as undertaking measures to reduce CO 2 emissions. The overall
results imply that economic growth and electricity consumption contribute towards CO 2
emissions in the GCC countries. No such relationship was found for financial development.

6. Conclusions and policy implications

This study aimed to examine the effects of economic growth, electricity consumption
and financial development on CO 2 emissions in GCC countries using panel data for the period
of 1980-2012. Unit root test that account for cross sectional dependence was conducted.
Pedroni cointegration test confirmed a cointegrating relationship between the variables. A
panel econometric technique, the DFE model, was estimated to examine both the short-run
and the long-run relationship between CO 2 emissions and economic growth, electricity
consumption and financial development. Economic growth and electricity consumption were
found to have a positive significant impact on CO 2 emissions in the long-run while no
significant short-run relationship between these variables was observed. Financial
development was found to reduce CO 2 emissions in the long-run. Group DOLS and FMOLS
provided evidence in support of the DFE results.
The Granger causality results suggested a bidirectional causal link between economic
growth and CO 2 emissions. A unidirectional causal link running from electricity consumption

10
to CO 2 emissions was found to exist. No causal link was found between financial
development and CO 2 emissions. Impulse response functions and variance decomposition
analysis revealed that per capita electricity consumption and economic growth would continue
to impact CO 2 emissions significantly into the future while the impact of financial
development is expected to be of little magnitude. Overall results demonstrated that economic
growth and electricity consumption contributed towards CO 2 emissions in the GCC countries.
No such relationship was found for financial development.
The findings of this study have very important policy implications for GCC countries
for not only to be able to efficiently deal with current climate challenges but also for their
post-oil future. Emissions are already causing sea levels to rise and affecting coastlines and
marine lives resulting in increasing levels of salinity. This situation will eventually cause a
scarcity of the availability of fresh water. The GCC countries are already running a large
number of desalinization plants which are very expensive to operate and are also harmful to
the environment as they need huge amounts of electricity to run.
Also, as these countries' energy supply is predicted to reduce with the passage of time,
the opportunity cost of huge government subsidies on current energy consumption is likely to
be more and more financially unsustainable. Although a trade-off between these opportunity
costs and the political reality of these countries, which are mostly ruled by monarchies, may
be difficult to envisage. Since, these countries are under potential threat as a result of their
alarming levels of emissions and their responses to combating emissions appear to be
inadequate so far, they can't afford to waste time. There is a need to act promptly to promote
energy efficiency and the use of renewable resources, in other words, they must do everything
possible to reduce their economies dependence on fossil fuels and to introduce newer more
environmentally friendly technologies to meet their energy needs.
Based on the findings of the study, it is suggested that the GCC countries should
reduce CO 2 emissions by a variety of measures. There are alternative potential measures for
electricity generation that will enable the region to achieve higher levels of energy efficiency.
It is already evident that GCC countries can reduce CO 2 emissions and gain energy efficiency
in three ways: a) promoting CCUS (carbon capture, utilization and storage) plants, b)
promoting the use of renewable resources and c) building nuclear energy plants.
The CCUS method has already proved its potential to reduce CO 2 emissions in the
region (Al Saleh et al. 2012). Another recent study of Saudi Arabia also has recognized the
potential of CCUS to significantly reduce CO 2 emissions (Mansouri et al. 2013). The GCC
countries also have clear advantage over the rest of the world in renewable resources,
especially solar and wind energy. The region is characterized by an enormous amount of
sunlight and wind and for more than 80% of days in a year its sky is cloud free or clear. The
average solar radiation of the region is 2200kWh (th)/m2 (Hertog, 2010). Therefore, solar and
wind are the two most significant potential renewable sources for energy in the region.
Although, Saudi Arabia and the UAE have already been pursuing research on this potential,
other GCC countries also need to recognize and tap this opportunity. It is already evident that
use of solar photovoltaic (PV) can significantly save CO 2 emissions in the electricity sector of
Saudi Arabia and the UAE (Mansouri et al., 2013; Mondal et al., 2014). Solar PV is a very
good technology option for long term investment in the power sector. It will potentially
enable GCC countries to achieve their renewable generation targets (Mondal et al. 2014).
Building nuclear energy plants is another viable option for the GCC countries to
combat emissions. Since all these countries' economies are characterized by large foreign
capital reserves thanks to their oil revenues, investment in such projects should not be
considered too ambitious for them. The UAE has already decided to integrate nuclear energy
into its electricity generation portfolio and a recent study by (Alfarra and Abu-Hijley, 2012)

11
showed through a number of scenario analysis that the use of nuclear energy would not only
reduce CO 2 emissions but also reduce per unit electricity generation costs.
Finally, this study recommends that the GCC countries need to significantly boost
investment for research in clean energy technologies and build energy expertise. This is not
only to address the prevailing climate challenges and meet their current renewable energy
targets only but also to deal with further challenges in the post-oil age. Long-term investment
in building a university under the potential name of 'GCC University of Energy Research and
Technology' could be a vital and sustainable contribution towards the achievement of such
goal.

12
References
AlFarra, H. J., Abu -Hijleh, B., 2012. The potential role of nuclear energy in mitigating CO 2
emissions in the United Arab Emirates, Energy Policy 42, 272-285.
Alkhathlan, K., Javid M., 2013. Energy consumption, carbon emissions and economic growth
in Saudi Arabia: An aggregate and disaggregate analysis, Energy Policy, 62, 1525-1532.
Al-Mulali, U., Ozturk, I., 2014. Are energy conservation policies effective without harming
economic growth in the Gulf Cooperation Council Countries? Renewable and
Sustainable Energy Reviews 38, 639-650.
Al-Mulali, U., Tang, C. F., 2013. Investigating the validity of pollution haven hypothesis in
the gulf cooperation council countries, Energy Policy 60, 813-819.
Al-Mulali, U., 2012. Exploring the bi-directional long run relationship between urbanization,
energy consumption and carbon dioxide emission, Energy, 46, 156-167.
Al-Saleh, Y. M., Vidican, G., Natarajan, L., Theeyattuparampil, V. V., 2012. Carbon capture,
utilization and storage scenarios for the Gulf Cooperation Council region: A Delphi-
based foresight study, Futures 44, 105-115.
Arouri, M. E. H., Youssef, A. B., Mhenni, H., Rault, C. 2012. Energy consumption, economic
growth and CO 2 emissions in Middle East and North African countries, Energy Policy
45, 342-349.
Asteriou, D. 2009. Foreign aid and economic growth: New evidence from a panel data
approach for five South Asian countries, Journal of Policy Modeling 31, 155-161.
Bernanke, B. 1986. Alternative Explanations of the Money-Income Correlation, Carnegie-
Rochester Conference Series on Public Policy, 25: 49-99.
Boutabba, M. A., 2014. The impact of financial development , income, energy and trade on
carbon emissions: Evidence from the Indian economy, Economic Modelling 40, 33-41.
Chang, S., 2015. Effects of financial developments and income on energy consumption,
International Review of Economics and Finance 35, 28-44.
Cowan, W.N., Chang, T., Inglesi-Lotz, R., Gupta, R. 2014. The nexus of electricity
consumption, economic growth and CO 2 emissions in the BRICS countries. Energy
Policy 66, 359-368.
Engle, R.F., Granger, C.J., 1987. Cointegration and error-correction - representation,
estimation and testing. Econometrica 55, 251-278.
Granger, C. W. J., 1969. Investigating Causal Relations by Econometric Models and Cross-
spectral Methods, Econometrica 37, 424-438.
Gutierrez, L., 2003. On the Power of Panel Cointegration Test: A Monte Carlo Comparison,
Economics Letters, 80:105-111.
Hamdi, H., Sbia, R., Shahbaz, M., 2014. The nexus between electricity consumption and
economic growth in Bahrain, Economic Modelling 38, 227-237.
Hamdi, H., Sbia, R., 2014. The dynamic relationship between CO 2 emissions, energy usage
and Growth in Gulf Cooperation Council (GCC) countries: An aggregated analysis,
Economie Appliquée 67, 161-182.
Hertog, S., 2010. Princes, Brokers, Bureaucrats: Oil and State in Saudi Arabia, Ithaca, NY:
Cornell University Press.
Hertog, S., Luciani, G., 2009. Energy and Sustainability policies in the GCC, The Center for
the Study of Global Governance, Kuwait Program on Development, Governance and
Globalization in the Gulf States, November, 2009.
Im, K. S., Pesaran, M. H., & Shin, Y. 2003. Testing for unit roots in heterogeneous panels,
Journal of Econometrics 115, 53-74.
Jalil, A., Feridun, M., 2011. The impact of growth, energy and financial development on the
environment in China: a cointegration analysis, Energy Economics 33, 284-91.

13
Jaunky, V.C., 2011. The CO 2 emissions-income nexus: evidence from rich countries. Energy
Policy 39, 1228-1240.
Kao, C., Chiang, M. H., 2000. On the estimation and inference of a cointegrated regression in
panel data, In: Baltagi B. H. (Ed.), Advances in Econometrics: Nonstationary Panels,
Panel Cointegration and Dynamic Panels, vol. 15, pp. 179-222.
Liao, H., Cao, H. 2013. How does carbon dioxide emission change with economic
development? Statistical experiences from 132 countries. Global Environmental Change
23, 1073-1082.
Mansouri, N. Y., Crookes, R. J., Korakianitis, T., 2013. A projection of energy consumption
and carbon dioxide emissions in the electricity sector for Saudi Arabia: The case for
carbon capture and storage and solar photovoltaics, Energy Policy 63, 681-695.
Mondal, M.A.H., Kennedy, S., Mezher, T., 2014. Long-term optimization of United Arab
Emirates energy future: Policy implications, Applied Energy, 114: 466-474.
Murthy, N.R.V., 2007. Panel Cointegration Analysis: An Empirical Example in Cointegration
for the Applied Economist, in Rao, B. (ed.) Cointegration for the Applied Economics
(Second Edition): 222-256.
Narayan, P.K., Popp, S., 2012. The energy-real GDP nexus revisited: Empirical evidence
from 93 countries, Economic Modelling 29, 303-8.
Nelson, C., & Plosser, C., 1982. Trends and Random Walks in Macroeconomic Time Series:
Some Evidence and Implications. Journal of Monetary Economics 10, 130-162.
Omri, A., Nguyen, D. K., Rault, C., 2014. Causal interactions between CO 2 emissions, FDI,
and economic growth: Evidence from dynamic simultaneous-equation models,
Economic Modelling 42, 382-389.
Omri, A., 2013. CO 2 emissions, energy consumption and economic growth nexus in MENA
countries: Evidence from simultaneous equations models. Energy Economics 40, 657-
664.
Ozcan, B., 2013. The nexus between carbon emissions, energy consumption and economic
growth in Middle east countries: A panel data analysis. Energy Policy 62, 1138-1147.
Ozturk, I., Acaravci, A., 2010. CO 2 emissions, energy consumption and economic growth, in
Turkey. Renewable and Sustainable Energy Reviews 14, 3220-3225.
Ozturk, I., Acaravci, A., (2013), The Long-Run and Causal Analysis of Energy, Growth,
Openness and Financial Development on Carbon Emissions in Turkey. Energy
Economics, 36, 262-267.
Pedroni, P., 1997. Panel Cointegration; Asymptotic and Finite Sample Properties of Pooled
Time Series Tests, with an Application to the PPP Hypothesis: New Results, Working
paper, Indiana University, April.
Pedroni, P., 1999. Critical Values for Cointegration Tests in Heterogeneous Panels with
Multiple Regressors, Oxford Bulletin of Economics and Statistics 61, 653-670.
Pedroni, P., 2000. Fully-modified OLS for heterogeneous cointegration panel, Advanced
Econometrics 15, 93-130.
Pedroni, P., 2001. Purchasing power parity tests in cointegrated panels, Review of Economics
and Statistics 83, 727-731.
Pesaran, M. H., 2004. General diagnostic tests for cross section dependence in panels.
Cambridge Working Papers in Economics No. 435. University of Cambridge and
CESifo Working Paper Series No. 1229.
Pesaran, M. H., 2007. A simple panel unit root test in the presence of cross-section
dependence. Journal of Applied Econometrics 22, 265-312.
Pesaran, M.H., Shin, Y., Smith, R., 1999. Pooled Mean Group Estimation of Dynamic
Heterogeneous Panels. Journal of the American Statistical Association 94, 621-634.

14
Saboori, B., Sapri, M., Baba, M., 2014. Economic growth, energy consumption and CO 2
emissions in OECD (Organization of Economic Co-operation and Development)'s
transport sector: A fully modified bi-directional relationship approach. Energy, 66(1),
150-161.
Sadorsky, P., 2011. Financial development and energy consumption in Central and European
frontier economies, Energy Policy 39, 999-1006.
Sadorsky, P., 2010. The impact of financial development on energy consumption in emerging
economies, Energy Policy 38, 2528-2535.
Salahuddin, M., Gow, J., 2014. Economic growth, energy consumption and CO 2 emissions in
Gulf Cooperation Council countries, Energy 73, 44-58.
Sbia, R., Shahbaz, M., Hamdi, H., 2014. A contribution of foreign direct investment, clean
energy, trade openness, carbon emissions and economic growth to energy demand in
UAE. Economic Modelling 36, 191-197.
Shahbaz, M., Lean, H. M., 2012. The dynamics of electricity consumption and economic
growth: A revisit study of their causality in Pakistan, Energy 39, 146-153.
Shahbaz, M., Zeshan, M., Afza, T. 2012. Is energy consumption effective to spur economic
growth in Pakistan? New evidence from bounds test to level relationships and Granger
causality, Economic Modelling 29, 2310-2319.
Shahbaz, M., Solarin, S. A., Mahmood, H., Arouri, M., 2013a. Does financial development
reduce CO 2 emissions in Malaysian economy? A time series analysis, Economic
Modelling 35, 145-152.
Shahbaz, M., Hye, Q. M. A., Tiwary, A. K., Leitao, N. C., 2013b. Economic growth, energy
consumption, financial development, international trade and CO 2 emissions in
Indonesia, Renewable and Sustainable Energy Reviews 25, 109-121.
Sims, C.A., 1980. Macroeconomics and Reality, Econometrica 48(1): 1-49.
Sims, C.A., 1986. Are Forecasting Models Usable for Policy Analysis? Federal Reserve Bank
of Minneapolis Quarterly Review, 3: 1-16.
Soytas, U., Sari, R. 2009. Energy consumption, economic growth and carbon emissions:
Challenges faced by an EU candidate member. Ecological Economics 68, 1667-1675.
Tamazian, A., Chousa, J. P., Vadlamannati, C., 2009. Does higher economic and financial
development lead to environmental degradation: Evidence from BRIC countries,
Energy Policy 37, 246-253.
Tamazian, A., Rao, B. B., 2010. Do economic, financial and institutional developments
matter for environmental degradation? Evidence from transitional economies, Energy
Economics 32, 137-145.
World Bank, 2013. World Development Indicators, The World Bank, Washington DC, USA,
available from: <http://data.worldbank.org/data-catalog/world-development-
indicators/World Bank-2012>; 2013 (accessed on 03.08.2013)
Yuxiang, K., Chen, Z., 2010. Financial development and environmental performance:
evidence from China, Environment and Development Economics 16, 1-19.
Zhang, X.P., Cheng, X-M., 2009. Energy consumption, carbon emissions and economic
growth in China. Ecological Economics 68, 2706-2712.
Zhang, Y. J., 2011. The impact of financial development on carbon emissions: An empirical
analysis in China, Energy Policy 39, 2197-2203.
Ziaei, S. M., 2015. Effects of financial development indicators on energy consumption and
CO 2 emissions of European, East Asian and Oceania countries, Renewable and
Sustainable Energy Reviews 42, 752-759.

15
Figure 1. Impulse responses of CO 2 emissions to independent variables

Accumulated Response to Generalized One S.D. Innovations ± 2 S.E.

Accumulated Response of LCO2 to LCO2 Accumulated Response of LCO2 to LEU
1.6 1.6

1.2 1.2

0.8 0.8

0.4 0.4

0.0 0.0

-0.4 -0.4

-0.8 -0.8
5 10 15 20 25 30 5 10 15 20 25 30

Accumulated Response of LCO2 to LGDPC Accumulated Response of LCO2 to DOMESTIC_CREDIT_TO_PRIVA

1.6 1.6

1.2 1.2

0.8 0.8

0.4 0.4

0.0 0.0

-0.4 -0.4

-0.8 -0.8
5 10 15 20 25 30 5 10 15 20 25 30

16
 Table 1: Summary statistics of the variables
LCO2 LGDPC LFD LEPC
Mean 3.044914 10.00954 1.534847 3.844567
Median 3.117565 9.864770 1.551185 3.918153
Maximum 4.227340 11.31380 1.971023 4.241647
Minimum 1.492100 8.711520 0.833438 2.794986
Std. Dev. 0.614045 0.640451 0.209025 0.291089
Skewness -0.461721 0.116449 -0.453239 -1.042773
Kurtosis 2.913923 1.591193 3.473983 4.047534

Jarque-Bera 6.737876 15.97199 8.196517 42.66683

Probability 0.034426 0.000340 0.016602 0.000000

Sum 572.4439 1881.794 288.5513 722.7785

Sum Sq. Dev. 70.50864 76.70310 8.170278 15.84502

Observations 188 188 188 188

Table 2: CIPS unit root test results

P CD CIPS CIPS (1st Diff.)
LGDPC 0.488 2.09** -1.501 -2.637***
LCO 2 0.298 -0.52 -1.747 -2.664***
LEPU 0.832 18.50*** -1.283 -2.860***
FD 0.506 11.03*** -1.811 -3.006***

Table 3: Pedroni Residual Cointegration Test

Alternative hypothesis: common AR coefs. (within-dimension)
Weighted
Statistic Prob. Statistic Prob.
Panel v-Statistic 2.679511 0.0037 1.777153 0.0378
Panel rho-Statistic -1.588999 0.0560 -0.742916 0.2288
Panel PP-Statistic -2.961571 0.0015 -1.791329 0.0366
Panel ADF-Statistic -1.357293 0.0873 -0.834728 0.2019
Alternative hypothesis: individual AR coefs. (between-dimension)
Statistic Prob.
Group rho-Statistic -0.907086 0.1822
Group PP-Statistic -2.939609 0.0016
Group ADF-Statistic -1.579396 0.0571

17
Table 4: Results from panel Dynamic Ordinary Least Squares (DOLS) estimation
Variable Coefficient Std. Error t-Statistic Prob.
LEC 0.648209 0.100480 6.451117 0.0000
LGDPC 0.390644 0.202823 1.926033 0.0562
LFD -0.006200 0.002251 -2.754503 0.0067
R-squared 0.949922 Mean dependent var 3.041422
Adjusted R-squared 0.926752 S.D. dependent var 0.603062
S.E. of regression 0.163215 Sum squared resid 3.569641
Long-run variance 0.039696

Table 5: Results from Fully Modified Ordinary Least Squares (FMOLS) estimation
Variable Coefficient Std. Error t-Statistic Prob.
LGDPC 0.405094 0.131383 3.083311 0.0024
LFD -0.655508 0.168888 -3.881318 0.0001
LEPC 1.338018 0.169979 7.871661 0.0000
R-squared 0.907085 Mean dependent var 3.050169
Adjusted R-squared 0.902738 S.D. dependent var 0.602068
S.E. of regression 0.187766 Sum squared resid 6.028805
Durbin-Watson stat 1.597326 Long-run variance 0.075695

Table 6: Results from Dynamic Fixed Effect (DFE) and Mean Group (MG) estimations
Dynamic Fixed Effect Mean Group
Variables Long Run Short Run Long Run Short Run
Error Correction -0.402*** -0.499***
(0.0562) (0.161)
Δ GDP Per Capita 0.217 0.525***
(0.254) (0.158)
Δ Electric Power Consumption 0.885*** 0.163
(0.120) (0.220)
Δ Financial Development 0.104 0.0378
(0.177) (0.0784)
GDP Per Capita 0.405* 3.372
(0.013) (2.570)
Electric Power Consumption 0.617*** -0.737
(0.107) (0.850)
Financial Development -0.131 -0.0354
(0.058) (0.0761)
Constant -2.448*** -4.168**
(0.785) (2.044)
Observations 192 192 192 192

18
Table 7: Panel VECM Granger Causality
Dependent variable: D(LCO2)
Excluded Chi-sq df Prob.
D(LEPU) 5.287331 1 0.2215
D(LGDPC) 0.329650 1 0.5659
D(FD) 6.121831 1 0.0134
All 11.23275 3 0.0105
Dependent variable: D(LEPU)
Excluded Chi-sq df Prob.
D(LCO2) 7.352636 1 0.0067
D(LGDPC) 5.152295 1 0.0232
D(FD) 0.142911 1 0.7054
All 11.69896 3 0.0085
Dependent variable: D(LGDPC)
Excluded Chi-sq df Prob.
D(LCO2) 0.882288 1 0.3476
D(LEPU) 1.985224 1 0.1588
D(FD) 0.867994 1 0.3515
All 2.547835 3 0.4667
Dependent variable: D(FD)
Excluded Chi-sq df Prob.
D(LCO2) 2.618363 1 0.1056
D(LEPU) 2.679576 1 0.1016
D(LGDPC) 0.196280 1 0.6577
All 3.735083 3 0.2915

19
Table 8. Variance Decomposition of CO 2 emission for GCC countries: 1980-2012
DOMESTIC_CREDIT_TO_
Period S.E. LCO2 LEU LGDPC PRIVA
1 0.141899 100.0000 0.000000 0.000000 0.000000
2 0.184542 99.69382 0.117580 0.146193 0.042407
3 0.207631 98.49373 1.021031 0.386978 0.098266
4 0.222774 96.56938 2.526875 0.763157 0.140583
5 0.234046 94.24441 4.339306 1.250446 0.165842
6 0.243119 91.78623 6.232153 1.805594 0.176026
7 0.250761 89.36774 8.066943 2.389722 0.175592
8 0.257369 87.08277 9.772029 2.975648 0.169548
9 0.263173 84.97231 11.31842 3.547021 0.162251
10 0.268323 83.04643 12.70142 4.095226 0.156919
11 0.272926 81.29900 13.92888 4.616515 0.155613
12 0.277064 79.71645 15.01417 5.109955 0.159422
13 0.280802 78.28272 15.97245 5.576120 0.168714
14 0.284191 76.98166 16.81866 6.016312 0.183360
15 0.287276 75.79829 17.56666 6.432122 0.202927
16 0.290094 74.71916 18.22883 6.825195 0.226808
17 0.292674 73.73249 18.81608 7.197110 0.254327
18 0.295044 72.82801 19.33787 7.549329 0.284796
19 0.297227 71.99686 19.80240 7.883180 0.317560
20 0.299240 71.23135 20.21677 8.199860 0.352015
21 0.301102 70.52484 20.58709 8.500444 0.387625
22 0.302827 69.87154 20.91864 8.785902 0.423923
23 0.304427 69.26639 21.21600 9.057107 0.460509
24 0.305915 68.70496 21.48314 9.314856 0.497045
25 0.307299 68.18335 21.72352 9.559875 0.533255
26 0.308589 67.69810 21.94016 9.792832 0.568911
27 0.309793 67.24614 22.13569 10.01434 0.603833
28 0.310917 66.82473 22.31241 10.22498 0.637879
29 0.311968 66.43142 22.47235 10.42529 0.670943
30 0.312951 66.06400 22.61730 10.61576 0.702943
31 0.313872 65.72048 22.74883 10.79687 0.733827
32 0.314734 65.39905 22.86833 10.96906 0.763557

20