Professional Documents
Culture Documents
1 s2.0 S0264999314000911 Main
1 s2.0 S0264999314000911 Main
1 s2.0 S0264999314000911 Main
Economic Modelling
journal homepage: www.elsevier.com/locate/ecmod
a r t i c l e i n f o a b s t r a c t
Article history: This paper examines the long-run equilibrium and the existence and direction of a causal relationship between
Accepted 7 March 2014 carbon emissions, financial development, economic growth, energy consumption and trade openness for India.
Available online 12 April 2014 Our main contribution to the literature on Indian studies lies in the investigation of the causes of carbon emis-
sions by taking into account the role of financial development and using single country data. The results suggest
JEL classification:
that there is evidence on the long-run and causal relationships between carbon emissions, financial development,
C32
O53
income, energy use and trade openness. Financial development has a long-run positive impact on carbon emis-
Q43 sions, implying that financial development improves environmental degradation. Moreover, Granger causality
Q53 test indicates a long-run unidirectional causality running from financial development to carbon emissions and
Q56 energy use. The evidence suggests that financial system should take into account the environment aspect in
their current operations. The results of this study may be of great importance for policy and decision-makers
Keywords: in order to develop energy policies for India that contribute to the curbing of carbon emissions while preserving
Carbon emissions economic growth.
Financial development
© 2014 Elsevier B.V. All rights reserved.
Growth
Energy consumption
Trade
1. Introduction and forestry for livelihood. The Indian economy is also dependent on
natural resources and any adverse impact on these and related sectors
Climate change and global warming are the greatest and most con- will negate government's efforts to eradicate poverty and ensure sus-
troversial environmental issues of our times. There is broad consensus tainable livelihood for the population.
among scientists that accumulated carbon dioxide emitted from the India accords high priority to its development. The economy has
burning of fossil fuels, along with contributions from other human- been growing, on average, at 7.7% per year between 2000 and
induced greenhouse gas emissions, are warming the atmosphere and 2007, and fossil-fuel carbon emissions have increased by 125% be-
oceans of the earth (IPCC, 2007). The global effects of climate change tween 1950 and 2008, becoming the world's third largest fossil-
are already apparent in increasing the frequency of extreme weather fuel CO 2-emitting country. As outlined in India's 12th Five Year
events, altering precipitation patterns, heightening storm intensity, Plan (2012–2017), the government of India has provisionally set a
reversing ocean currents and a rising sea level. These changes, in turn, 9% GDP growth target, which will require energy supply to grow
can have significant impacts on the functioning of ecosystems, the at 6.5% per year. Being aware of achieving its growth trajectory in
viability of wildlife, and the well-being of humans. an environmentally sustainable manner, India has announced in
With the world's second largest population and over 1.1 billion December 2009 that it would aim to reduce the emissions intensity
people, India is one of the lowest greenhouse gas emitters in the of its GDP by 20–25% from 2005 levels by 2020. Therefore, India is
world on a per-capita basis. Its emission of 1.18 tonnes of carbon equiv- faced with the challenge of identifying the common ground between
alent per capita in 2008 was nearly one-fourth of the corresponding climate change policy and economic growth and pursuing measures
global average of 4.38 tonnes. However, India is highly vulnerable to that achieve both.
climate change, as a large population are dependent on agriculture However, to control the greenhouse gas emissions and to ensure
the sustainability of the economic development, it is important
to better understand the inter-temporal links in the environment–
⁎ Tel.: +33 1 69 47 79 22. energy–income nexus. In the literature, there have been few researches
E-mail address: mohamedamine.boutabba@univ-evry.fr. to explore the relationship between these variables in the case of India.
http://dx.doi.org/10.1016/j.econmod.2014.03.005
0264-9993/© 2014 Elsevier B.V. All rights reserved.
34 M.A. Boutabba / Economic Modelling 40 (2014) 33–41
Ghosh (2010) investigated the causal relationship between carbon Results indicated that, first, China's financial development constitutes
emissions and economic growth using ARDL bounds testing approach an important driver for carbon emissions increase, which should be
complemented by Johansen–Juselius maximum likelihood proce- taken into account when carbon emissions demand is projected. Sec-
dure in a multivariate framework by incorporating energy supply, ond, the influence of financial intermediation scale on carbon emissions
investment and employment. The result revealed the absence of a outweighs that of other financial development indicators but its
long-run causality between carbon emissions and economic growth; efficiency's influence appears by far weaker although it may cause the
however a bi-directional short-run causality between the two is change of carbon emissions statistically. Third, China's stock market
found. Alam et al. (2011) applied the Toda and Yamamoto causality scale has a relatively larger influence on carbon emissions but the influ-
test to examine the dynamic relationship between carbon emissions, ence of its efficiency is very limited. Finally, among financial devel-
economic growth, energy consumption, labour forces and gross fixed opment indicators, China's FDI exerts the least influence on the
capital formation. They found a bi-directional Granger causality change of carbon emissions, due to its relatively smaller volume
between energy consumption and carbon emissions in the long compared with income. Ozturk and Acaravci (2013) examined the
run but neither carbon emissions nor energy consumption causes causal relationship between financial development, openness, eco-
movements in economic growth. Jayanthakumaran et al. (2012) nomic growth, energy consumption and carbon emissions in Turkey
analyzed the long-run relationship between carbon emissions and for the period 1960–2007. Empirical results yielded evidence of a
other variables such as growth, energy, trade and endogenously long-run relationship between carbon emissions, energy consumption,
determined structural breaks. They found evidence for the exis- income, openness ratio and financial development. The results also
tence of an EKC hypothesis for India. However, they failed to derive supported the validity of EKC hypothesis in Turkish economy. However,
a clear picture regarding the association of structural change and financial development has no significant effect on carbon emissions in
carbon emissions. Kanjilal and Ghosh (2013) examine the EKC hy- the long- run.
pothesis using threshold cointegration with endogenously deter- For cross-country case studies, Talukdar and Meisner (2001) ex-
mined structural breaks. The study advocates that the existence amined the impact of private sector involvement on carbon emis-
of structural breaks in the period of the study can produce mislead- sions using data from 44 developing countries over nine years
ing results if they are not incorporated in the cointegration testing (1987–95). They found that both foreign direct investments and do-
models. mestic financial capital markets in an economy are likely to have
This paper extends the above-mentioned multivariate frame- positive impacts on the environment. Claessens and Feijen (2007)
work further by including the impacts of financial development analyzed the role of governance in reducing CO2 emissions and re-
into the nexus. To the best of our knowledge, there has never been ported that with the help of more advanced governance firms can
an attempt to investigate the causes of carbon emissions for India lower the growth of carbon emissions. They suggested that financial
by taking into account the role of financial development and using development might stimulate the performance of firms due to the
single country data. This study tries to fulfil this gap. In this respect, adoption of energy efficient technologies, which reduce carbon
we argue that the analysis of the relationship between carbon emis- emissions. Tamazian et al. (2009) investigated the linkage between
sions and financial development may reduce the problems of omit- financial development, economic development and environmental
ted variable bias in econometric estimation. This attempt may also quality for BRIC countries using panel data over the period 1992–
be of great importance for policy and decision-makers to better ap- 2004. Their results revealed that higher degree of economic and
prehend the determinants of carbon emissions in order to develop financial development decreases the environmental degradation.
effective energy policies that will palliate the impacts of human ac- Tamazian and Bhaskara Rao (2010) tested the role of economic,
tivities, and thereby contribute to the curbing of carbon emissions financial and institutional developments on environmental degrada-
while preserving economic growth. tion with a sample of 24 transition countries for the period from
The remainder of the paper is organized as follows. Section 2 1993 to 2004. Their findings showed that financial liberalization
presents a brief literature review related to financial development and may be harmful for environmental quality if it is not accomplished
carbon emissions. Section 3 describes the data and methodology. in a strong institutional framework. In addition, the findings confirm
Empirical results are given in Section 4 while the summary and the the existence of an EKC.
concluding remarks are outlined in Section 5. India is included in some of the above-mentioned panel data studies.
However, it is widely recognized that any potential inference drawn
2. A brief literature review from these cross-country studies provides only a general understanding
of the linkage between the variables, and thus are unable to offer much
The impact of financial development on environmental conditions guidance on policy implications for each country (Ang, 2008; Lindmark,
has gained increasing attention in the recent literature. Yuxiang and 2002; Stern et al., 1996). Hence, the aim of this research is to investigate
Chen (2011) used provincial data of Chinese economy to examine the impact of financial development on carbon emissions in the case of
the impact of financial development on industrial pollutants and India.
found improvements in environment due to financial development.
They claimed that financial development improves environmental 3. Methodology and data
quality by increasing income and capitalization, exploiting new technol-
ogy and implementing regulations regarding environment. Jalil and Following the empirical literature in energy economics, it is plausible
Feridun (2011) investigated the impact of financial development, eco- to form the long-run relationship between carbon emissions, financial
nomic growth and energy consumption on CO2 emissions in the case development, economic growth, energy consumption, and foreign
of China from 1953 to 2006. The results of the analysis revealed a nega- trade in linear logarithmic quadratic form, with a view of testing the
tive sign for the coefficient of financial development, suggesting that long-run and causal relationships between these variables in India, as
financial development in China has not taken place at the expense of follows:
environmental pollution. On the contrary, it is found that financial
development saves the environment from degradation. Moreover, the 2
results confirm the existence of a long-run relationship between carbon CO2t ¼ β0 þ α F F t þ α Y Y t þ α Y 2 Y t þ α E Et þ α T T t þ εt ð1Þ
emissions, income, energy consumption and trade openness while
supporting the presence of EKC hypothesis. Similarly, Zhang (2011) where t and ε denote time and error, respectively. CO2 is carbon
explored the effect of financial development on carbon emissions. emissions (measured in metric tonnes per capita), F stands for
M.A. Boutabba / Economic Modelling 40 (2014) 33–41 35
financial development that is the total value of domestic credit to the bounds testing approach generally provides unbiased long-run esti-
private sector1,2 as a share of GDP, Y indicates per capita real GDP mates and valid t-statistics (Narayan, 2005). Fourth, the model takes a
(measured in local constant currency), Y2 is the square of per capita sufficient number of lags to capture the data generating process in gen-
real GDP, E means the energy consumption (measured as kg of oil eral to specific modeling frameworks (Laurenceson and Chai, 2003).
equivalent per capita), which is used as a proxy for economic growth, Fifth, the error correction model (ECM) can be derived from ARDL
and T represents trade openness, which is the total value of exports through a simple linear transformation, which integrates short run ad-
and imports as a share of GDP. justments with long run equilibrium without losing long run informa-
The parameters αF, αY, αY2, αE, and αT are the long-term elasticity tion (Pesaran and Shin, 1999).
estimators of CO2 emissions with respect to financial development, Basically, the ARDL approach to cointegration involves two steps for
per capita real GDP, the square of per capita real GDP, the per capita estimating long-run relationship. The first step is to investigate the
energy consumption and the trade openness, respectively. existence of long-run relationship among all variables in the equation
Financial development may be harmful for environmental quality under estimation. If there is an evidence of cointegration between vari-
αF N 0 (Talukdar and Meisner, 2001; Zhang, 2011), otherwise αF b 0 if ables, the second step is to estimate the long-run and short-run models.
the focus of financial sector is to improve environmental quality by
enabling firms in adopting advanced cleaner and environment friendly 3.2. Stationarity
techniques (Claessens and Feijen, 2007; Jalil and Feridun, 2011;
Tamazian et al., 2009). As discussed earlier, the ARDL bounds testing procedure can be
The EKC hypothesis suggests that αY N 0 and αY2 b 0. αY being posi- applied irrespective of whether the variables are I(0), I(1) (Pesaran
tive reveals the phenomenon wherein as income increases, the CO2 and Pesaran, 1997). However, according to Ouattara (2004), in the pres-
emissions increase as well; αY2 being negative reflects the inverted-U ence of I(2) variables the computed F-statistics provided by Pesaran
curve-shaped pattern of the EKC, where once income passes the thresh- et al. (2001) become invalid. This is because the bounds test is based
old, the CO2 emissions will decrease. on the assumption that the variables should be I(0) or I(1). Therefore,
The expected sign of αE is positive because a higher level of energy the implementation of unit root tests in the ARDL procedure is neces-
consumption should result in greater economic activity and stimulate sary to ensure that none of the variables is integrated at an order of
CO2 emissions. I(2) or beyond.
αY is expected to be negative or positive, depending on the level of It is well known that the presence of structural breaks in the series
economic development stage of a country. In general, developing coun- may bias the results toward non rejection of the null hypothesis of a
tries, which are abundant in labour and natural resources, attempt to unit root when there is none. This consideration is of particular impor-
promote heavy industries, which usually are pollution-intensive, by tance since the economic system in India has been subject to some dras-
accepting foreign direct investment of developed countries. In contrast, tic changes in policy and regulations. An alternative to the unit root test
developed countries change from energy-intensive industries to ser- against a single-break stationarity was proposed by Zivot and Andrews
vices and knowledge-based technology-intensive industries, which (1992). It was extended to a two-break stationarity alternative by
are environmentally cleaner (Grossman and Krueger, 1995). Lumsdaine and Papell (1997) and up to five-break stationarity alterna-
The sample period runs from 1971 to 2008 based on the annual time tive, with a priori unknown number of breaks, by Kapetanios (2005).
series data availability. The data originate from the world development However, these tests maintain the linearity assumption under the unit
indicator data base, the World Bank. All variables are employed with root null hypothesis. If a break exists under the null of a unit root, it
their natural logarithms form to reduce heteroskedasticity and to obtain will exhibit size distortions that not only “over-reject” the null hypoth-
the growth rate of the relevant variables by their differenced logarithms. esis of a unit root, but also will tend to estimate the break point incor-
rectly. To overcome this problem, Lee and Strazicich (2003, 2004)
have developed an alternative (at most two) endogenous break unit
3.1. Estimation strategy root test that uses the Lagrange Multiplier (LM) test statistic, and allows
for breaks both under null and alternative hypotheses. Thus, rejection of
This study employs the Autoregressive Distributed Lag (ARDL) the unit root null based on LM test provides a quite strong evidence of
bounds testing procedure recently developed by Pesaran et al. (2001). stationarity.
The ARDL has several advantages over other techniques of cointegration Lee and Strazicich (2003, 2004)'s unit-root test considers the data
such as Engle and Granger (1987) and Johansen and Juselius (1990). generating process as follows:
First, it can be applied irrespective of whether the underlying variables
are I(0), I(1) or a combination of both (Pesaran and Pesaran, 1997). Sec-
Δyt ¼ δ ΔZ t þ ϕe
0
St−1 þ ut
ond, the ARDL procedure is statistically a more significant approach to
determine the cointegration relation in small samples than that of the
where e St ¼ yt −ψ e −Z e
x t δðt ¼ 2; …T Þ and Zt is a vector of exogenous
Johensen and Juselius cointegration technique (Pesaran and Shin,
variables defined by the DGP; e δ is the vector of coefficients in the regres-
1999). Third, even where some of the model regressors are endogenous,
sion of Δyt on ΔZt respectively with Δ as the difference operator; and ψx ¼
1
y1 −Z 1 e
δ, with y1 and Z1 the first observations of yt and Zt respectively.
Domestic credit to private sector refers to financial resources provided to the private
The unit-root null hypothesis is described by ϕ = 0. The augmented
sector, such as through loans, purchases of non-equity securities, and trade credits and
other accounts receivable, that establish a claim for repayment. terms Δe St− j ; j ¼ 1; …k; were included to correct for serial correlation.
2
In the literature, there are many proxies used for representing financial development. The value of k is determined by the general-to-specific search proce-
For example, the monetary aggregate M2 as a ratio of nominal GDP is used in measuring dure. To endogenously determine the location of the break (TB), the
financial deepening. However, the availability of foreign funds in the financial system LM unit-root searches for all possible break points for the minimum
makes the monetary aggregate an inappropriate measure of financial development. An-
other commonly used variable is the ratio of deposit liabilities to nominal GDP, which cap-
(the most negative) unit-root t-test statistic, as follows:
tures the broad money stock excluding currency in circulation. But, this measure doesn't
TB
take into account the allocation of capital. Several studies have also employed the ratio
Inf e e ¼ Inf e
τ λ
of commercial bank assets divided by commercial bank plus central bank assets which λ τ ðλÞ; λ ¼ :
T
measures the importance of the commercial banks in the financial system. In this study,
we use the domestic credit to private sector as a percentage of GDP, which constitutes
the most common variable used in the literature to represent financial development. In
The critical values of the endogenous two-break LM unit-root test
fact, this measure represents more accurately the role of financial intermediaries in are reported in Lee and Strazicich (2003) and the critical values of the
channelling funds to private markets participant. one-break LM unit-root test are tabulated in Lee and Strazicich (2004).
36 M.A. Boutabba / Economic Modelling 40 (2014) 33–41
In the present study, when the two-break LM test results showed To gauge the adequacy of the specification of the model, diagnostic
that only one structural break is significant to at least the 10 per cent and stability tests are conducted. Diagnostic tests examine the model
level for some series, we perform the one-break LM test of Lee and for serial correlation, functional form, non-normality and hetero-
Strazicich (2004). This was done not only because the one-break LM scedasticity. As suggested by Pesaran and Pesaran (1997), the stability
test appears more appropriate in this case, but also because we wanted of the short-run and long run coefficients are checked through the
to determine if including two breaks instead of one can adversely affect cumulative sum (CUSUM) and cumulative sum of squares (CUSUMSQ)
the power to reject the unit root hypothesis for these countries. For the tests proposed by Brown et al. (1975). The CUSUM and CUSUMSQ statis-
same reason, when the one-break or two-break LM test results showed tics are updated recursively and plotted against the breaks points. If the
that no break is significant, we employ the Augmented Dickey–Fuller plots of the CUSUM and CUSUMSQ statistics stay within the critical
(ADF), Phillips–Perron, Augmented Dickey–Fuller GLS (ADF-GLS), and bounds of a 5% level of significance, the null hypothesis of all coefficients
Kwiatkowski–Phillips–Schmidt–Shin (KPSS) unit-root techniques. The in the given regression is stable and cannot be rejected.
first three techniques test the null hypothesis of a unit root against the
alternative of stationarity. The KPSS method tests the hypothesis that 3.4. Granger causality
the series is stationary against the alternative of non-stationarity.
The ARDL method tests the existence or absence of cointegration
3.3. Cointegration analysis relationship between variables, but not the direction of causality. If
we do not find any evidence for cointegration among the variables
The ARDL procedure involves the estimation of Eq. (1) as follows: then the specification of the Granger causality test will be a vector
autoregression (VAR) in first difference form. However, if we find
X
p X
p X
p X
p
2 evidence for cointegration then we need to augment the Granger-type
ΔCO2t ¼ a0 þ a1i ΔCO2t−i þ a2i ΔF t−i þ a3i ΔY t−i þ a4i ΔY t−i
causality test model with a one period lagged error correction term
i¼1 i¼0 i¼0 i¼0
X
p X
p (ECTt − 1). This is an important step because Engle and Granger
þ a5i ΔEt−i þ a6i ΔT t−1 þ λ1 CO2t−1 þ λ2 F t−1 þ λ3 Y t−1 ð2Þ (1987) caution that if the series are integrated of order one, in the pres-
i¼0 i¼0 ence of cointegration VAR estimation in first differences will be mislead-
2
þλ4 Y t−1 þ λ5 Et−1 þ λ6 T t−1 þ μ t ing. The augmented form of Granger causality test with ECM is
formulated in multivariate qth order of VECM model as follows:
where Δ denotes the first difference operator, a0 is the drift component, et 2 3 2 3 2 32 CO 3
CO2t b1 c11;i c12;i c13;i c14;i c15;i c16;i 2t−i
μt is the usual white noise residuals, and the variables CO2, F, Y, Y2, E, and T 6 7 6 7 6 c c c c c c 76 F t−i 7 6
6 F t 7 6 b2 7 6 21;i 22;i 23;i 24;i 25;i 26;i 76 7
are as defined earlier. The terms with summation signs represent the 6 7 6 7 6 76 7
6Y 7 6b 7 X q
6 c c c c c c 7 Y 7
error correction dynamics, while the second part of the equation with λ 6 t 7 6 31;i 32;i 33;i 34;i 35;i 36;i 76 7
ð1−BÞ6 2 7 ¼ 6 37 t−i
þ ð 1−BÞ 6 7
6 Yt 7 6 b 7 6 c c c c c c 76
6 41;i 42;i 43;i 44;i 45;i 46;i 76 Y t−i 7
2
corresponds to the long run relationship. This equation incorporates the 6 7 6 47 i¼1 7
6 7 6 7 6 76
4 c51;i c52;i c53;i c54;i c55;i c56;i 54 Et−i 7
time trend variable to capture the autonomous time-related changes. 4 Et 5 4 b5 5 5
The ARDL method estimates (p + 1)k number of regressions in order Tt b6 c61;i c62;i c63;i c64;i c65;i c66;i T t−i
to obtain the optimal lag length for each variable, where p is the maxi- 2 3 2 3
δ1 γ 1t
mum number of lags to be used and k is the number of variables in 6δ 7 6γ 7
6 27 6 2t 7
the equation. This is an appropriate lag selection based on criteria 6 7 6 7
6 δ3 7 6γ 7
such as Akaike Information Criterion (AIC) and Schwarz Bayesian Crite- þ6
6δ 7
7½ECT t−1 þ 6 3t 7
6γ 7 ð4Þ
rion (SBC). The bounds testing procedure is based on the joint F-statistic 6 47 6 4t 7
6 7 6 7
4 δ5 5 4 γ 5t 5
or Wald statistic that is tested the null hypothesis of no cointegration.
Pesaran et al. (2001) and Narayan (2005) individually report two δ6 γ 6t
sets of critical values for a given significance level. One set of critical
values assumes that all variables included in the ARDL model are I(0), where (1 − B) is the lag operator, ECT is the lagged error-correction
while the other is calculated on the assumption that the variables are term and yts serially independent random errors with mean zero and
I(1). If the computed test statistic exceeds the upper critical bounds finite covariance matrix.
value, then the H0 hypothesis is rejected. If the F-statistic falls into the The VECM allows us to capture both the short-run and long-run
bounds then the cointegration test becomes inconclusive. In this case, Granger causality. The short-run causal effects can be obtained by the
following Kremers et al. (1992) and Banerjee et al. (1998), the error F-test of the lagged explanatory variables, while the t-statistics on the
correction term will be a useful way for establishing cointegration. If coefficient of the lagged error correction term indicates the significance
the F-statistic is lower than the lower bounds value, then the null hy- of the long-run causal effect.
pothesis of no cointegration cannot be rejected. Two sets of critical
values are reported in Narayan (2005) for sample sizes ranging from 4. Empirical results and discussion
30 observations to 80 observations. Given the relatively small sample
size in the present study (38 observations), we extract appropriate 4.1. Unit root tests
critical values from Narayan (2005).
Having found that there exists a long-run relationship between the Table 1 reports the unit root results from the two-and one-break LM
variables, the next step is to estimate the error-correction model: tests. We tested each variable for a unit root using the two-break LM test
at the 1-, 5- and 10 percent levels of significance. As noted above, when
X
p X
p X
p X
p this test showed that only one structural break is significant we
2
ΔCO2t ¼ a0 þ a1i ΔCO22t−i þ a2i Δ F t−i þ a3i ΔY t−i þ a4i ΔY t−i employed the one-break LM test at the same levels of significance. In
i¼1 i¼0 i¼0 i¼0
ð3Þ order to determine the number of lags, we used a “general to specific”
X
p X
p
þ a5i ΔEt−i þ a6i ΔT t−i þ η ECT t−1 þ μ 1t procedure at each combination of break points for the two-break test,
i¼0 i¼0 and at each single break point for the one-break test.
As shown in Table 1, the unit root hypothesis with two structural
where η measures the speed of adjustment to obtain equilibrium in the breaks cannot be rejected for CO2 in level. Similar results were found
event of shock(s) to the system and ECTt − 1 is the residuals that are for Y and T, all of which have experienced one break in their term struc-
obtained from the estimated cointegration model of Eq. (1). tures. However, if we take the first differences, the unit root null for all
M.A. Boutabba / Economic Modelling 40 (2014) 33–41 37
Table 1 Table 3
Two/one-break minimum LM unit-root tests. The results of f-test for cointegration.
CO2 −5.182 1986 2000 −7.310⁎⁎⁎ 1975 2000 Note: The critical value ranges of F-statistics are 2.152–3.296, 2.523–3.829 and
Y −3.055 1989 – −7.266⁎⁎⁎ 1990 – 3.402–5.031 at 10%, 5% and 1% levels of significance, respectively, which are taken
T −4.665 1980 – −5.369⁎⁎⁎ 1981 – from the Appendix in Narayan (2005).
Note: Results are based on the model C, which allows for two changes in the level and
trend of the series. the ARDL model, (ii) the number of explanatory variables,(iii) whether
⁎⁎⁎ Rejection of the null hypothesis at the 1% significance level. the ARDL model includes an intercept and/or time trends, and (iv) the
sample size.
the series can be rejected at the 1% level, suggesting thereby that they The calculated F-statistic F(CO2/F Y Y2 E T DU1986 DU2000) = 2.731
are integrated of order 1, i.e. I(1). lies between the lower and upper bounds of critical values, indicating
Table 2 presents the ADF, PP, KPSS and ADF-GRS test results for F, Y2 that it is inconclusive whether or not the null hypothesis of no
and E and, for which the results from the two- and one-break LM tests cointegration relationship should be rejected. In this case, as mentioned
showed no significant breaks. The results reveal that all the four tests earlier, the error-correction term is a useful way of establishing
almost unanimously indicate that all variables are non-stationary in cointegration (Banerjee et al., 1998; Kremers et al., 1992).
their level data. However, the stationarity property is found in the first
difference of the variables in 5% or 1% critical level.
While the first structural break for CO2 in 1986 is inexplicable in 4.3. Long- and short-run elasticities
terms of energy consumption, the second in 2000 reflects the steady
increase of energy use since the late 1990s which contributes to the The ARDL cointegration procedure was implemented to estimate the
increase of carbon emissions. parameters of the Eq. (2). The AIC criterion has been utilized to find the
The break date of 1989 for Y shows an upward trend, which may be coefficients of the level variables. Because, AIC is known as parsimoni-
related to the economic reforms undertaken by Rajiv Ghandi soon after ous model, as selecting the smallest possible lag length and it minimizes
taking over as Prime Minister in 1985. Reforms include the abolition of the loss of degree of freedom as well.
licences for some industries, sale of shares in selected public enterprises, The long-run results are reported in Table 4. Except for the coeffi-
removal of price controls and establishment of the Stock Exchange cient of T in the model, all estimated coefficients are statistically signif-
Board of India. icant and have correct signs as expected.
The break date of 1983 for T coincides with the second oil shock, The results indicate that financial development has a long-run posi-
which deteriorated India's terms of trade as well as its balance-of- tive impact on per capita CO2 emissions. A 1% increase in domestic credit
payment. to private sector will lead to about 0.182% increase in per capita CO2
As none of the variables is integrated of order two, the ARDL bounds emissions, which is significant at the 1% level. This suggests that finan-
procedure can be used to examine the existence of a long-run relation- cial development improves environmental degradation. This finding is
ship in the following step. different from Tamazian et al. (2009), whose study uses panel data
over period 1992–2004 and reveal that higher degree of financial devel-
opment decreases the environmental degradation in the BRIC countries.
4.2. Cointegration test results However, our result lends support to Zhang (2011), who note that the
bank loans provide solid support for Chinese companies to access exter-
Given the establishment of structural breaks in the CO2 series, Eq. (2) nal finance and enhance investment scale. This boosts economic growth
incorporates two dummy variables (DU1986 and DU2000). The and carbon emissions which depend on the bank asset scale expansion.
cointegration test under the bounds testing approach involves compar- Both linear and non-linear terms of real GDP provide evidence in
ing the F-statistics against critical values. Given that the value of the supporting the inverted-U relationship between economic growth and
F-statistic is sensitive to the number of lags imposed each time on the CO2 emissions. The result indicates that a 1% rise in real GDP will raise
differenced variables (Bahmani-Oskooee and Nasir, 2004), we select CO2 emissions by 11.984% at the 1% significance level while a negative
the optimal order of lags of the model based on the Akaike Informa- sign of squared term seems to corroborate the delinking of CO2 emissions
tion (AIC) and the Schwarz–Bayesian(SBC) information criteria as and real GDP at the higher level of income. These evidences support the
suggested by Pesaran et al.(2001). The results of the lag selection EKC hypothesis, revealing that CO2 emissions increase in the initial
criteria indicate that the optimal number of lags is one. stage of economic growth and decline after a threshold point, i.e.
The calculated F-statistics, together with the critical values, are 19,380 Indian rupees. This finding is consistent with Jayanthakumaran
reported in Table 3. The F-test has a non-standard distribution that et al. (2012) and Kanjilal and Ghosh (2013), whose studies do not in-
depends on four factors, namely (i) the order of variables included in clude financial development, and with those various studies in which
Table 2
Conventional unit root tests.
Level First difference Level First difference Level First difference Level First difference
Test statistics Test statistics Test statistics Test statistics Test statistics Test statistics Test statistics Test statistics
Note: ADF: Augmented Dickey–Fuller test. PP: Phillips–Perron test. KPSS: Kwiatkowski–Phillips–Schmidt–Shin. ADF-GLS: Elliot–Rothenberg–Stock Dickey–Fuller GLS detrended. ADF, PP
and DF-GLS critical values are taken from MacKinnon (1991). KPSS critical values are sourced from Kwiatkowski et al. (1992). All null hypotheses except KPSS are unit root; while, in KPSS
null is stationarity.
⁎⁎⁎ Rejection of the null hypothesis at the 1% significance level.
⁎⁎ Rejection of the null hypothesis at the 5% significance level.
38 M.A. Boutabba / Economic Modelling 40 (2014) 33–41
3 4
In order to check the model robustness, the energy variable is dropped from the equa- According to the World Bank's Doing Business 2011 Report, with an overall rank of
tion since it may explain most of the CO2 emissions. The exclusion of energy variable does 134, India was among the top 50 countries in terms of obtaining credit and protecting
not affect the findings. These results are available upon request. investors.
M.A. Boutabba / Economic Modelling 40 (2014) 33–41 39
Table 5 10
Short-run estimation results.
Regressor Coefficient
ΔF −0.023 5
(0.067)
ΔY 7.340***
(2.432)
ΔY2 −0.368*** 0
(0.124)
ΔE 1.003***
(0.332)
ΔT −0.061
-5
(0.058)
ΔDU1986 −0.002
(0.016)
ΔDU2000 0.013
(0.016) -10
Constant 0.026** 2001 2002 2003 2004 2005 2006 2007 2008
(0.010)
ECM(−1) −0.640*** CUSUM 5% Significance
(0.155)
2
R 0.614
SE of regression 0.021 Fig. 1. Plot of CUSUM.
5. Conclusion
0.4
Table 6
Granger causality results.
Notes: ***, ** and * indicate that the null hypothesis of no causation is rejected at the 1%, 5% and 10% significance levels, respectively. The numbers in parentheses are standard errors. Δ is
the first difference operator. The number of appropriate lag is one according to Akaike information criterion, Schwarz information criterion and Hannan–Quinn information criterion.
Lee, J., Strazicich, M., 2003. Minimum Lagrange multiplier unit root test with two structural Saboori, B., Suleiman, J., Mohd, S., 2012. Economic growth and CO2 emissions in Malaysia: a
breaks. Rev. Econ. Stat. 85 (4), 1082–1089. cointegration analysis of the environmental Kuznets curve. Energy Policy 51, 184–191.
Lee, J., Strazicich, M., 2004. Minimum Lagrange multiplier unit root test with one structural Shahbaz, M., Lean, H.H., Shabbir, M.S., 2012. Environmental Kuznets curve hypothesis
break. Working paper. Department of Economics, Appalachian State University, USA. in Pakistan: cointegration and Granger causality. Renew. Sust. Energ. Rev. 16 (5),
Lindmark, M., 2002. An EKC-pattern in historical perspective: carbon dioxide emis- 2947–2953.
sions, technology, fuel prices and growth in Sweden, 1870–1997. Ecol. Econ. Song, T., Zheng, T., Tong, L., 2008. An empirical test of the environmental Kuznets curve in
42, 333–347. China: a panel cointegration approach. China Econ. Rev. 19, 381–392.
Lotfalipour, M.R., Falahi, M.A., Ashena, M., 2010. Economic growth, CO2 emissions, and Stern, D.I., Common, M.S., Barbier, E.B., 1996. Economic growth and environmental degra-
fossil fuels consumption in Iran. Energy 35, 5115–5120. dation: the environmental Kuznets curve and sustainable development. World Dev.
Lumsdaine, R., Papell, D., 1997. Multiple trends breaks and the unit root hypothesis. Rev. 24 (7), 1151–1160.
Econ. Stat. 79, 212–218. Talukdar, D., Meisner, C.M., 2001. Does the private sector help or hurt the environment?
MacKinnon, J.G., 1991. Critical Values for Co-Integration Tests. In: Engle, R.F., Granger, C.W.J. Evidence from carbon dioxide pollution in developing countries. World Dev. 29,
(Eds.), Long-Run Economic Relationships. Oxford University Press, pp. 267–276. 827–840.
Narayan, P.K., 2005. The saving and investment nexus for China: evidence from Tamazian, A., Bhaskara Rao, B., 2010. Do economic, financial and institutional develop-
cointegration tests. Appl. Econ. 37, 1979–1990. ments matter for environmental degradation? Evidence from transitional economies.
Nasir, M., Rehman, F.-U., 2011. Environmental Kuznets curve for carbon emissions in Energy Econ. 32, 137–145.
Pakistan: an empirical investigation. Energy Policy 39, 1857–1864. Tamazian, A., Chousa, J.P., Vadlamannati, C., 2009. Does higher economic and financial
Ouattara, B., 2004. Foreign Aid and Fiscal Policy in Senegal. Mimeo University of development lead to environmental degradation: evidence from the BRIC countries.
Manchester. Energy Policy 37, 246–253.
Ozturk, I., Acaravci, A., 2013. The long-run and causal analysis of energy, growth, open- World Bank, k. Doing Business 2011: Making a difference for Entrepreneurs. http://
ness and financial development on carbon emissions in Turkey. Energy Econ. 36, www.doingbusiness.org//media/fpdkm/doing%20business/documents/annual-reports/
262–267. english/db11-fullreport.pdf.
Pesaran, M., Pesaran, B., 1997. Working with Microfit 4.0: Interactive Economic Analysis. World Economic Forum, 2012. New Energy Architecture: India. http://www3.weforum.
Oxford University Press, Oxford. org/docs/WEF_NewEnergyArchitecture_ExecutiveSummary_India.pdf.
Pesaran, H.M., Shin, Y., 1999. Autoregressive distributed lag modelling approach to Yuxiang, K., Chen, Z., 2011. Financial development and environmental performance: evi-
cointegration analysis. In: Storm, S. (Ed.), Econometrics and Economic Theory in dence from China. Environ. Dev. Econ. 16, 1–19.
the 20th Century: The Ragnar Frisch Centennial Symposium. Cambridge University Zhang, Y.J., 2011. The impact of financial development on carbon emissions: an empirical
Press, Cambridge. analysis in China. Energy Policy 39, 2197–2203.
Pesaran, M.H., Shin, Y., Smith, R.J., 2001. Bounds testing approaches to the analysis of level Zivot, E., Andrews, D., 1992. Further evidence on the great crash, the oil price shock and
relationships. J. Appl. Econ. 16, 289–326. the unit root hypothesis. J. Bus. Econ. Stat. 10, 251–270.