The nexus of foreign direct investment, economic growth and environment in Pakistan.
Raza, Syed Sundus ; Hussain, Anwar
This paper estimate the impact of sectoral FDI on economic growth
and carbon dioxide emissions in Pakistan. To this end, it uses time
series secondary data from 1972 to 2011 and applies Auto Regressive
Distributed Lag (ARDL) models. The results showed that FDI inflows in
manufacturing, transport, storage and communication sectors and energy
consumption have positive effect on the GDP growth of Pakistan, Besides,
FDI inflow in manufacturing, transport, storage and communication sector
and population density are responsible for the C[O.sub.2]emissions in
Pakistan. The results also validate Environmental Kuznet Curves in both
long and short run.
JEL Classification: E2, 04, Q5
Keywords: Sectoral FDI, CO: emissions, Environmental Kuznet Curves,
Gross Domestic Product Growth
1. INTRODUCTION
The growing concern for sustainable development diverted the world
concentration from conventional growth to environmentally affable growth
[Nasir and Rehman (2011)]. Environmental degradation has affected the
economic activities in serious manner. This increase in environmental
degradation is fueled by multi factors including the increasing trend of
foreign direct investment [Mabey and McNally (1999)].
Foreign firms target developing countries that have low
environmental standards, which attract investment in polluting sectors
leading to "pollution heaven hypothesis" (1) [Chakraborty and
Mukerjee (2010)]. Foreign firms choose to operate in developing
countries in order to gain benefit from low cost of production which in
turn effect environment negatively leading to "industrial flight
hypothesis." (2) But not all the FDI inflow is bad for environment
in developing world sometime it can be beneficial in form of
"pollution holes hypothesis." (3) This means that even if we
refuse these hypotheses there is considerable amount of environmental
damage associated with FDI [Shahbaz, et al. (2011)].
When foreign investment and trade amplify, it leads to extend the
net of economic activities. The increased level of economic activities
result in environmental degradation, which leads to scale effect. (4)
The emissions of C[O.sub.2] can be decreased by the use of environmental
friendly technology imported by foreign investors, so the international
investment and trade can lead to the environment friendly production, as
the competition increase the domestic producer also try to focus on
production and decrease per unit cost. This leads to technique effect.
(5) The FDI can also alter the industrial configuration of the economy
leading to composition effect (6) [Grossman and Krueger (1991)].
The FDI stimulates the domestic investment, human capital
formation, facilitate the technology transfer. Hence, the foreign direct
investment is considered as growth enhancing factor in developing
countries [Acharyya (2009); Falki (2009) and Asghar, el al. (2011)]. FDI
inflows have helped in boosting the economic growth through structural
makeover of the economy of Pakistan. It also helped in initiating the
industrial sector as well as lying foundations for agricultural sector,
supplied modern technology and technical support [Din (2007)].
There is inverted "U" shape relationship between
environmental degradation and economic growth, when economic growth
increases, income also increases which affect the environment
negatively. As a result of increased growth, the economy expands and
income rises. At high level of income people are more conscious about
environment so they demand to maintain clean environment. This
relationship is called as "Environmental Kuznets Curves" (EKC)
[Grossman and Krueger (1991)]. The same idea is also supported by Seldon
and Song (1994).
Different sectors have different effects on the economic growth
[Alfaro (2003)]. The type of FDI and sector in which it is going is very
important from both environmental degradation and economic growth point
of view. In this paper Carbon dioxide emissions are used as variable
representing environmental degradation and GDP to represent economic
growth. The FDI affect both environment and the economic growth. To test
the Environmental Kuznets Curve (EKC), many researchers have used Carbon
dioxide emissions as indicator for environmental quality.
There is very little work done on the sector specific FDI, economic
growth and environmental degradation. Therefore, this paper contributes
to empirically check the effect of sector specific FDI on economic
growth and environment followed by checking the existence of EKC in
Pakistan. For the analysis three sectors have been selected namely,
manufacturing sector, mining and quarrying sector and transport, storage
and communication sectors. Only those sectors are selected that have
high actual emissions (emission per unit of output).
In the past researchers tried to relate FDI with other economic
variables. Besides, they highlighted various influencing factors of
economic growth. Falki (2009) examines the effect of total FDI on
economic growth of Pakistan. The sector of economy in which FDI is
coming is very important with relationship to economic growth. The
outcomes in terms of economic growth can vary from sector to sector and
can be misleading if total FDI is used [Wang (2009)]. Studies by Alfaro
(2003); Ganges, et al. (2006); Chakraborty and Nunnenekamp (2007) and
Wang (2009)] found out that manufacturing sector contribute positively
towards economic growth, whereas there are insignificant contribution of
primary sector and ambiguous contribution of services sector towards
economic growth. Labour force and FDI have an important interaction and
labour force play an important role in the absorption of FDI
[Borenztein, el al. (1998)]. Energy consumption is a vital determinant
of economic growth as it is considered as an engine of economic progress
[Lee and Chang (2008)]. Economic growth and environmental degradation is
the area of concern from early 1990s, FDI tends to increase the level of
economic activity which a leads to environmental degradation [Pao and
Tsai (2010); Zhang, et al. (2011); Merican, et al. (2007) and Mulali
(2012)]. Environmental degradation is also related to population density
and the increase in population density trends to increase environmental
degradation [Shi (2003)].
For understanding the costs and benefits of FDI in terms of
economic growth and environmental degradation, it is critical to study
their nexus. The studies of Baek and Ron (2008); Acharyya (2009) and
Honglei, et al. (2011) are worth mentioning who explored the presence of
"Pollution Heaven Hypothesis" and EKC's. The relationship
between the economic growth and environmental degradation was first
floated by Grossman and Krueger (1991) followed by Selden and Song
(1994). Different studies on EKC's have been included in this
regard are Lindmark (2002); Fodhaa and Zaghdoud (2011); Nasir and Rehman
(2011); Shahbaz, et al. (2011) and Hitam and Borhan (2012). The studies
regarding EKC's use different indicators for environmental
degradation like C[O.sub.2] and S02. Nasir and Rehman (2011) and
Shahbaz, et al. (2011), explored the validity of EKC for Pakistan but
with the nexus of energy consumption, economic growth and total FDI,
2. THEORETICAL BACKGROUND
2.1. FDI Inflow and Carbon Dioxide Emissions Model
According to Dasgupta, et al. (2002) "The environmental
Kuznets curve posits an inverted-U relationship between environmental
degradation and economic development. Kuznets' name was apparently
attached to the curve by Grossman and Krueger (1991), who noted its
resemblance to Kuznets's inverted-U relationship between income
inequality and development." The relationship between environmental
degradation and economic growth can be expressed as:
[Z.sub.t] = [[alpha].sub.0] + [[alpha].sub.1][Y.sub.t] +
[[alpha].sub.2][Y.sup.2.sub.t] + [e.sub.t] ... (1)
Where [Z.sub.t] can be any variable which represent environmental
degradation and [Y.sub.t] can be any variable which represent economic
growth. The linear and nonlinear terms of economic growth are added in
order to check the validity of EKC. Theoretically if the coefficients of
Y are positive and that of [Y.sup.2] is negative; it validates the
existence of the EKC hypothesis [Shahbaz, et al. (2011)].
To check the impact of sectoral FDI oil Carbon dioxide emission,
the additional variables namely FDI in manufacturing, FDI in mining and
quarrying, FDI in transport, storage and communication sector are added
to the model. The population density is added to the existing model
because population density is related to Carbon dioxide emissions [Shi
(2003)]. The final model is as follows:
[mt.sub.t] = f ([g.sub.t], [g.sup.2.sub.t], [man.sub.t],
[mn.sub.t], [tr.sub.t], [pd.sub.t], D, [e.sub.t]) ... (2)
Where
[mt.sub.t] is Carbon dioxide emissions in year t measured in metric
ton per capita.
[g.sub.t] is Real GDP per capita in year t and measured in million
rupees.
[g.sup.2.sub.t] is square term of real GDP per capita in year t and
measured in million rupees.
[man.sub.t] is FDI inflow in manufacturing sector in year /
measured in million rupees.
[mn.sub.t] is FDI inflow in mining and quarrying sector in year t
measured in million rupees.
[tr.sub.t] is FDI inflow in transport, storage and communication
sector in year t measured in million rupees.
[pd.sub.t] is population density in year t measured in per square
km of land area.
Whereas D is dummy variable which represent structural breaks
namely in year (1994, 2000, 2005, 2007, 2008, and 2009) in sectoral FDI
data. e, is error term.
2.2. FDI Inflow and GDP Growth Model
The neoclassical economist gave the theory of output (production)
function as follows;
Y = Af(K, L) ... (3)
Equation 3 represents Cobb Douglas production function where K
represents capital and L represents Labour. Energy variable was first
added to the economic theory by Roegen (1975). Then Kraft and Kraft
(1978) was first to use energy consumption variables in production
function. Further FDI is used in sectoral form in this study because
different sectors have different effects on economic growth.
Borensztein, et al. (1998) stressed on the importance of human capital
because it plays very important role in the absorption of foreign direct
investment, this is the reason for the inclusion of labour force in the
model.
The model is as follows
[GDP.sub.t] = f ([man.sub.t], [mn.sub.t], [tr.sub.t], [lab.sub.t],
[ene.sub.t], D,[e.sub.t]) ... (4)
Where;
[GDP.sub.t] is gross domestic product in year l measured in million
rupees.
[man.sub.t] is FDI inflow in manufacturing sector in year t and
measured in million rupees.
[mn.sub.t] is FDI inflow in mining and quarrying sector in year t
measured in million rupees.
[tr.sub.t] is FDI inflow in transport, storage and communication in
year t measured in million rupees.
[ene.sub.t] is energy consumption in year t measured in million
metric tons of oil equivalent.
[lab.sub.t] is labour force in year t measured in millions.
Whereas D is dummy variable which represent structural breaks
namely in year (1994, 2000, 2005, 2007, 2008, and 2009) in sectoral FDI
data, [e.sub.t] is error term.
3. DATA AND METHODOLOGY
3.1. Data and Sources
The data used in this study is time series from 1972 to 2011. Data
on per capita C[O.sub.2] emissions, population density is obtained from
World Development Indicators (WDI). The data for the energy consumption
is taken from Statistical, Economic and Social Research and Training
Centre for Islamic Countries (SESRIC). While the data for the labour
force, real per capita GDP, real GDP and sectoral FDI is taken from
State Bank of Pakistan (SBP).
3.2. Methodology
3.2.1. Test of Stationarity
Augmented Dickey Fuller (ADF) test is widely used to identify the
order of integration 1(d) of variables. The general form of Augmented
Dickey Fuller test is
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)
Where, [X.sub.t] denotes the time series variable to be tested,
used in model, t is time period, [DELTA] is first difference and [phi]
is root of equation, [beta]t is deterministic time trend of the series
and [alpha] denotes intercept. The numbers of augmented lags (p)
determined by the dropping the last lag until we get significant lag.
The Augmented Dickey Fuller unit root concept is illustrated through
equation [DELTA][X.sub.t] = ([rho] - 1) [X.sub.t-1] + [[epsilon].sub.t],
Where, ([rho] - 1) can be equal to [phi], if [rho] = 1 so series has the
unit root, so root of equation is [phi] = 0.
3.2.2. Test of Cointegration
For finding the cointegration among the variables, Pesaran, el at.
(2001) has proposed bound test through ARDL approach to test the
co-integration. Through ARDL bound testing approach, the long run and
short run dynamic association between the variables can be estimated at
a same time by estimating the unrestricted error correction model
(UECM).
Following is the general form of ADRL model of co-integration or
UECM;
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)
Where; Y is dependent variable and X is vector of independent
variables, Pesaran, el al. (2001). Following two hypotheses will be
tested to check the co-integration between variables, u, is normally
distributed with zero mean and constant variance (0, [[sigma].sup.2]).
We have applied the restrictions on Equation 6 to check the following
hypotheses.
[H.sub.0.sup.1]: [[phi].sub.x][X.sub.t-1] =
0([[phi].sub.x][X.sub.t-1] is of lag of independent variables equal to
zero)
[H.sub.0.sup.2]: [[lambda].sub.y][Y.sub.t-1] = 0
([[lambda].sub.y][Y.sub.t-1] dependent variable lag equals to zero)
[H.sub.0.sup.2]: [H.sub.0.sup.1] [intersection] [H.sub.0.sup.2]
[H.sub.a]: [H.sub.0.sup.1] [union] [H.sub.0.sup.2]
We check [H.sub.0.sup.1] and [H.sub.0.sup.2] jointly. First to
check the co-integration joint hypothesis; [H.sub.0] is tested through
F-statistics, by comparing with critical values of F for bound test
[Pesaran, et al. (2001)]. There are two bound for each level of
significance, 1 (1) upper bound and I (0) lower bound. If F-statistics
lies outside the upper bound 1(1), the null of hypotheses is rejected.
If it lies below the lower bound 1 (0), the null hypothesis cannot be
rejected and if it lies between the two bound then results are
inconclusive.
In next step log run estimates can be calculated from UECM by
normalising the variables.
[Y.sub.t] = c + [beta] + [[phi].sub.x][X.sub.t-1] + [[mu].sub.t]
... (7)
Where; c is constant and [beta]t is trend. [[phi].sub.x][X.sub.t-1]
is vector of independent variables. Finally, short run dynamics are
estimated from the UECM as follows;
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (8)
So finally UECM for FDI inflows and C[O.sub.2] can be estimated as
follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (9)
[a.sub.0] and [beta] are the intercept and trend respectively.
Whereas [a.sub.1] to [a.sub.7] are the long run coefficients and
[[delta].sub.1] to [[delta].sub.7] are short run coefficients, [omega]
is the coefficient of dummy variable and [[mu].sub.t] error term.
UECM for FDI inflows and GDP growth can be estimated as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (10)
[a.sub.0] and [beta] are the intercept and trend respectively.
Whereas [a.sub.1] to [a.sub.6] are the long run coefficients and
[[delta].sub.1] to [[delta].sub.7] are short run coefficient. [omega]
Coefficient of dummy variable and [[mu].sub.t] error term.
3.3. Bound Test Procedure
The first step in the ARDL bounds testing approach is to estimate
Equation (9 and 10) by ordinary least squares (OLS) in order to test for
the existence of a long-run relationship among the variables by
conducting an F-test for the joint significance of the coefficients of
the lagged levels of the variables, i.e., [H.sub.0]'.
[[delta].sub.1] = [[delta].sub.2] = [[delta].sub.3] = [[delta].sub.4] =
[[delta].sub.5] = 0 against the alternative [H.sub.1]: [[delta].sub.1
[not equal to] [[delta].sub.2] [not equal to] [[delta].sub.3] [not equal
to] [[delta].sub.4] [not equal to] [[delta].sub.5] [not equal to] 0. We
denote the test which normalise on mt by [F.sub.mt] (mt/c,t, g,
[g.sup.2], man, mn, tr, pd) and normalised on GDP by [F.sub.GDP](GDP/c,
t, man, mn, lab, ene) for second mo del. A symptotic critical values
bounds provide a test for co-integration when the independent variables
are 1(d) (where 0 [less than or equal to] d [less than or equal to] 1):
a lower value assuming the regressors are 1(0) and an upper value
assuming purely 1(1) regressors. If the F-statistic is above the upper
critical value, the null hypothesis of no long-run relationship is
rejected. Conversely, if the test statistic falls below the lower
critical value, the null hypothesis cannot be rejected. Finally, if the
statistic falls between the lower and upper critical values, the result
b inconclusive.
In the second step, once co-integration is established the
conditional ARDL ([p.sub.1], [q.sub.1], [q.sub.2], [q.sub.3], [q.sub.4])
long model for the dependent variable is estimated. The long run model
of FDI inflow and C[O.sub.2] emission is as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (11)
Long run model of FDI inflow and GDP growth is as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (12)
Short run dynamics of FDI inflows and Carbon dioxide emissions is
as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (13)
Short run dynamics of FDI inflow and economic growth is as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (14)
4. RESULTS AND DISCUSSION
It can be seen from Table 1 that FDI in manufacturing sector (man),
FDI in mining and quarrying sector (mn) and population density (pd) are
stationary at level whereas Carbon dioxide emissions (mt), Gross
Domestic Product (GDP), real GDP per capita (g), Square of GDP per
capita([g.sup.2]), FDI in transport, storage and communication (tr),
Energy consumption (ene) and labour force (lab) are stationary at first
difference. The results explores that the order of differencing of these
variables are not the same, so ARDL model is appropriate to use.
4.1. ARDL Model for FDI Inflow and Carbon Dioxide
First the UECM is estimated that contains Carbon dioxide emissions
per capita as dependent variable as shown in Table 2. The estimated UECM
is given below which includes long run as well as short run
coefficients. This is parsimonious form of equation, from which
insignificant terms are deleted. The outcome of test depends on the lag
selection that is p=l, selected on the basis of AIC (Akaike criterion).
In the model dummy variables are also included to check the impact of
structural breaks in data. The significant dummies were in year 2007 and
2008. The year 2007 dummy represent structural break in FDI inflow in
transport, storage and communication sector. The year 2008 dummy show
structural break in FDI inflow manufacturing sector. Both dummies are
significant. As stated by the Board of Investment Pakistan, Foreign
direct investment inflow in the country was at 485 million dollars
during 2001-02, following which there was a rise in FDI inflow in the
country for the subsequent six years. The FDI inflow spiked in the year
2007-08, attaining a massive level of 5409 million dollars. After that,
there was a gradual fall till 2011-12 level. If the spike through
2007-08 is taken as a point of reference among 2001 and 2012, 10-15
percent increase was recorded till 2007-08, after that there was a
decline of 89 percent till 2011-12. One of the reasons was the
democratic government in Pakistan which gained foreign confidence and
engrossed foreign direct investment in Pakistan. Secondly the democratic
government failed to solve the problems of the energy sector. Energy
crisis has increased in the past three years. Continuous power cut downs
and riots took place in Pakistan, specifically at Punjab. This situation
influenced all economic sectors from manufacturing sector to transport
sector, where FD1 inflow was concentrated. After estimating the UECM,
long run relationship has been checked, through testing the hypothesis
that H0:[beta]=[a.sub.2]= [a.sub.3]= [a.sub.4]= [a.sub.5]= [a.sub.6]=
[a.sub.7]= [a.sub.8] = 0 by applying the F-test on lagged variables and
comparing its values with the critical bound values provided by Pesaran,
el al. (2001). The F calculated F = 12.8. As there are k = 6, the
[F.sub.III] (unrestricted intercept and no trend) has critical values of
upper 1 (1) and lower bound I (0) that are (2.45 3.61), so calculated F
is greater than the upper bound critical value.
4.1.1. Normalise! Long Run Estimates
In the next step, long run equation is estimated whose coefficients
are estimated by normalising it on dependent variable (mt). The
normalised long run estimates are given in Table 3 which shows that one
million rupees rise in GDP per capita income will increase per capita
Carbon dioxide emissions by 72.01 metric ton per capita. The coefficient
is also significant at 1 percent level of significance. The long run
results are also in line with the study conducted by Fodha and Zaghdoud
(2011).
The results shows that one million rupees increase in FDI inflow in
transport, storage and communication sector will increase Carbon dioxide
emissions by 0.0011 metric ton per capita. The coefficient of transport,
storage and communication (tr) is also significant at 1 percent level of
significance and results are also in line with the study done Gallagher
(2004). According to Gallagher (2006) "the increased emissions from
transport sector mainly depended on the non-provision of clean
technologies by the foreign firms" (pp. 28). Transport sector have
high emissions rate amongst all sectors and accountable for quarter of
C[O.sub.2] emissions in Pakistan. Emissions control in transport sector
is decisive for management of Climate Change [Draft National Climate
Change Policy (2011)].
Besides, one million rupees increase in mining and quarrying sector
will increase emissions by 0.0013 metric ton per capita but coefficient
of mining and quarrying sector (mn) is insignificant at 1 percent, 5
percent and 10 percent level of significance. Similarly if FDI inflow in
manufacturing sector increases by one million rupees the Carbon dioxide
emissions will increase by 0.0012 metric ton per capita. The coefficient
of manufacturing sector (man) is also statistically significant at 1
percent level of significance. The results are also in line with the
study done by Jorgenson (2007). In developing countries the foreign
firms use more pollution technology both in manufacturing sector and in
transport sector [Jorgenson (2007)]. The results may also get support
from that only Carbon dioxide emissions from the manufacturing sector in
Pakistan stands at 42.2 (million metric tons) in year 2011. Further it
is also suggested that the industrial sector contribute positively
towards Green House Gases (G1IG) [Draft National Climate Change Policy
(2011)].
The results show that if the population density is increased by one
unit then the emissions will increase by 0.014 metric ton per capita.
Similar results were also found by Shi (2003) who proposed that
population density is positively related with the Carbon dioxide
emissions in long run.
Pakistan is in the list of most vulnerable countries against
climate change. The recent United Nations Framework Convention on
Climate Change (UNFCCC) Paris conference 2015 has agreed to set-up a
special "Technology development and transfer mechanism" for
the development and transfer of new technologies from developed to
developing countries [Draft Paris outcome (2015)]. In case of Pakistan
there is lack of policy regarding the clean technology transfer through
FDI in manufacturing, mining and quarrying and transport, storage and
communication sectors. There is lack of mechanisms which can keep a
check and balance on the capital equipment coming in the form of FDI.
The aforementioned discussion confirms the effect of sectoral FDI
on environmental degradation in terms of Carbon dioxide emissions.
Furthermore, the results show that the sign of variable (g) is positive
and sign of variable ([g.sup.2]) is negative which validate the
existence of Environmental Kuznets Curves. The results for the
Environmental Kuznets Curves are in line with the study done by Nasir
and Rehman (2011).
4.1.2. Short Run Estimates
Short run estimates of ARDL are given in Table 4. The coefficient
of Error Correction Term (ECT) is significant and negative. The
estimated coefficient of ECT shows disequilibrium is corrected or
adjusted with the speed of 78 percent in-between one year. The
significance ECT also confirms the long run relationship of variables as
estimated earlier. According to short run results, the [DELTA]g is
positively associated with the Carbon dioxide emissions in Pakistan and
these results are also in line with the study by Fodha and Zaghdoud
(2011). The first lag of FDI inflow in manufacturing sector
([DELTA][man.sub.t-1]) affect the Carbon dioxide emissions in short run
also. This means that previous year FDI in this sector will affect
Carbon dioxide emissions in current year. This is valid because FDI from
the previous year will also produce C[O.sub.2] emissions, therefore
adding to current year amount of emissions produced, the coefficient of
([DELTA][man.sub.t-1]) is also statistically significant at 1 percent
level of significance. [DELTA]tr is also positively related to
C[O.sub.2] emissions in short run. While results also showed that
difference and first lag of population density are also positively
related to the Carbon dioxide emissions in short run both of the
coefficients are statistically significant at 1 percent level of
significance. The results for population density are also in line with
the study done by Shi (2003). While interestingly the EKC exist in short
run also.
4.2. ARDL Results for the Impact of FDI Inflows on GDP Growth
First the UECM is estimated, that contains GDP as dependent
variable. The estimated UECM is given in Table 5 which includes long run
as well as short run coefficients. This is parsimonious form of
equation, from which insignificant terms are deleted. The outcome of
test depends on the lag selection that is p = 2, selected on the bases
of Akaike Info Criteria (AIC). In the model dummy variables are also
included to check the impact of structural breaks in data. Different
dummies were added to capture the effect of structure break. The
significant dummies were in year 2007 and 2008.
After estimating the UECM, long run relationship has been checked,
through testing the hypothesis that H0:[beta]=[a.sub.2]= [a.sub.3]=
[a.sub.4]= [a.sub.5]= [a.sub.6]= [a.sub.7]= [a.sub.8]=0 by applying the
F-test on lagged variables and compared its values with the critical
values bound provided by Pesaran, et al. (2001). As the value of F is
15.21, k is 5, the [F.sub.III] (Unrestricted intercept and no trend) has
critical values of upper 1(1) and lower bound I (0) that are (2.96
4.81), so the null hypothesis Ho, that there is no co-integration is
rejected at, 0.05 level of significance. This further concludes
existence of co-integration.
4.2.1. Long Rim Estimate
The results show that if FD1 in manufacturing sector is increased
by one million rupees, the GDP will increase by 24.27 million rupees.
The coefficient of FDI inflow in manufacturing sector (man) is also
statistically significant at 1 percent level of significance and it is
positively related to GDP growth in long run (Table 6). The results are
also in line with the study conducted by Chakraborty and Nunnekamp
(2008).
The results further showed that one million increases in FDI inflow
in mining and quarrying sector will increase GDP by 61.33 million
rupees. The coefficient of FDI inflow in mining and quarrying sector
(mn) is statistically insignificant at 1 percent, 5 percent and 10
percent level of significance. There is no significant relationship
between the FDI inflow in mining and quarrying sector and economic
growth, that is because of the fact that when foreign investment is
involved in this sector foreign firms take lions share from the host
countries that's the reason that FDI in this sector does not
contribute towards economic growth. The results are also in line with
the study by Chakraborty and Nunnekamp (2008).
Further results showed that one million rupees increase in FDI
inflow in transport, storage and communication sector will increase GDP
by 193.43 million rupees. The coefficient of FDI inflow in transport,
storage and communication sector (tr) is also statistically significant
at 1 percent level of significance. Similar result was also found by
Gangnes, et al. (2006).
One million increases in labour force variable will decrease GDP by
0.13 million rupees. Labour can contribute negatively towards GDP
growth, this happens when labour is not efficient. This idea was also
supported by Khan and Qayyum (2007).
Results showed that increase of one million metric tons of oil
equivalent in energy consumption will increase GDP by 0.54 million
rupees. This shows a positive relationship among energy consumption
variable and GDP growth variable, further the coefficient of energy
consumption (ene) is statistically significant at 1 percent level of
significance. The results are also in accordance with the results of
Glasure (1998) and Lee and Chang (2008).
4.2.2. Short Run Estimates
Short run estimates of ARDL are given below in Table 7. The ECT is
significant and negative. The estimated coefficient of ECT shows
disequilibrium is corrected or adjusted with the speed of 22 percent
in-between one year. The significance ECT also confirms the long run
relationship of variables as estimated earlier. The results showed that
FD1 inflow in manufacturing sector and transport, storage and
communication sector contribute positively towards GDP in short run
also. The difference and second lag of variable transport, storage and
communication sector (tr) is statistically significant at 10 percent and
1 percent level of significance which shows that FDI inflow in this
sector is positively related to GDP increase in short run. This argument
is valid because services sector is the largest contributor towards
Pakistan's GDP [Economic Survey of Pakistan (2011)]. When the
investment in this sector take place the effect can be seen in GDP
growth after one or two year that is the reason that second lags of
variable (tr) is statistically significant. The investment made in this
sector affect the GDP in coming years also.
Further the energy consumption is also positively associated with
the increase in GDP in short run. The variable of energy consumption is
also statistically significant at 1 percent level of significance. The
short run results of energy consumption are also in line with the
results of Lee and Chang (2008). Interestingly the second lag of labour
force is negatively affecting the GDP; the coefficient of second lag of
labour force is also statistically significant at 1 percent level of
significance. This happen due to labour inefficiency in Pakistan. This
demands human capital growth in the country.
5. CONCLUSION AND POLICY IMPLICATION
This study examined the effect of FD1 inflow on Carbon dioxide
emissions and GDP growth and checked the validity of EKC hypothesis in
Pakistan for the time period of 1972-2011. Per capita Carbon dioxide
emissions was used as indicator of environmental degradation and real
GDP as economic growth. The ARDL model was employed for the estimation
purposes. The findings revealed that FD1 inflow in manufacturing sector,
transport, storage and communication sector and population density have
positive impact on the environmental degradation in the long run. The
EKC hypothesis is also valid in the long-run for Pakistan. Further,
population density, FDI inflow in transport, storage and communication
sector and manufacturing sector variables add significantly to the
deterioration of environment in the short-run also. The EKC hypothesis
is valid in the short-run also. Furthermore, the coefficient of FDI
inflow in manufacturing sector, transport, storage and communication
sector and energy consumption are statistically significant and these
are the major influencing factors of GDP growth.
To protect environment from increasing Carbon dioxide emissions,
the government should consider sector specific FDI inflow in the economy
in their policy. Special attention should be given to population control
to lessen the pressure on the increasing Carbon dioxide emissions in the
country. Furthermore, to stimulate economic growth in both short and
long run, the FDI inflow in manufacturing and transport, storage and
communication sector must be promoted. The government must also invest
in human capital. This will not only increase the labour productivity
but also the quality of the labour.
REFERENCES
Acharyya, J. (2009) FDI, Growth and the Environment: Evidence from
India on C[O.sub.2] emissions during the Last Two Decades. Journal of
Economic Development 34:1.
Alfaro, L. (2003) Foreign Direct Investment and Growth: Does the
Sector Matter? Boston, MA: Harvard Business School. (Mimeographed).
Asghar, N., S. Nasreen, and H. U. Rehman (2011) Relationship
between FDI and Economic Growth in Selected Asian Countries: A Panel
Data Analysis. Review of Economics and Finance 1:1, 84-96.
Baek, J. and W. W. Kon (2008) A Dynamic Approach to the
FDI-Environment Nexus: The Case of China and India. Selected Paper
Prepared for Presentation at the American Agricultural Economics
Association Annual Meeting, Orlando, FL, July 27-29. Series/Report No:
Selected Paper 467514.
Borensztein, E., J. D. Gregorio, and J. W. Lee (1998) How does
Foreign Direct Investment Affect Economic Growth? Journal of
International Economics 45:1, 115-135.
Chakraborty, C. and P. Nunnenkamp (2008) Economic Reforms, FDI, and
Economic Growth in India: A Sector Level Analysis. World Development
36:7, 1192-1212.
Chakraborty, D. and S. Mukherjee (2010) Relationship between Trade,
Investment and Environment: A Review of Issues. MPRA Munich Personal
RePEcArchive. (MPRA Paper No. 23333).
Dasgupta, S., B. Laplante, H. Wang, and D. Wheeler (2002)
Confronting the Environmental Kuznets Curve. The Journal of Economic
Perspectives 16:1, 147-168.
Dickey, D. A. and W. A. Fuller (1979) Distribution of the
Estimators for Autoregressive Time Series with a Unit Root. Journal of
the American Statistical Association 74, 427-431.
Dietzenbacher, E. and K. Mukhopadhyay (2007) An Empirical
Examination of the Pollution Haven Hypothesis for India: Towards a Green
Leontief Paradox? The Journal of Environmental and Resource Economics
36:1,427-449.
Din, G. M. U. (2007) Impact of Foreign Capital Inflows (FCI) On
Economic Growth in Pakistan: [1975-2004], Journal of Independent Studies
and Research 5:1.
Draft Paris outcome, version 1 of 9 December (2015) at 15:00,
UNFCCC. Available at http://unfccc.int/resource/docs/2015/cop21
/eng/da01.pdf
Falki, N. (2009) Impact of Foreign Direct Investment on Economic
Growth in Pakistan. International Review of Business Research Papers
5:1, 110-120.
Fodhaa, M. and O. Zaghdoud (2011) Economic Growth and Pollutant
Emissions in Tunisia: An Empirical Analysis of the Environmental Kuznets
Curve. Energy Policy 38:2, 1150-1156.
Gallagher, K. S. (2006) Limits to Leapfrogging in Energy
Technologies? Evidence from the Chinese Automobile Industry.
Energy' Policy 34:4, 383-394.
Gangnes, B., I. Noy, and B. T. Vu (2008) Is Foreign Direct
Investment Good for Growth? Evidence from Sectoral Analysis of China and
Vietnam. Journal of the Asia Pacific Economy 13:4, 542-562.
Glasure, Y. U. and A. R. Lee (1998) Cointegration,
Error-correction, and the Relationship between GDP and Energy: The Case
of South Korea and Singapore. Resource and Energy Economics 20:1, 17-25.
Grossman, G. M. and B. A. Krueger (1991) Environmental Impacts of a
North American Free Trade Agreement. NBER Working Papers Series.
(Working Paper No. 3914).
Hitam, M. B. and and H. B. Borhan (2012) FD1, Growth and the
Environment: Impact on Quality of Life in Malaysia. Procedia--Social and
Behavioral Sciences 50:1, 333-342.
Flonglei, C., Z. Xaiorong, and C. Qiufeng (2011) Export-oriented
Economy and Environmental Pollution in China: The Empirical Study by
Simultaneous Equation Model. Energy> Procedia 5:1, 884-889.
Jorgenson, A. K. (2007) Does Foreign Investment Harm the Air We
Breathe and the Water We Drink? A Cross-National Study of Carbon Dioxide
Emissions and Organic Water Pollution in Less-Developed Countries, 1975
to 2000. Organisation Environment 137:20.
Kennedy, P. (2003) Guide to Econometrics. (5th ed). MIT Press:
Massachusetts.
Khan, A. K. and A. Qayyum (2007) Dynamic Modeling of Energy and
Growth in South Asia. The Pakistan Development Review 44:4, 481-498.
Kraft, J. and A. Kraft (1978) Relationship between Energy and GNP.
Journal of Energy' Development 3:2.
Lee, A. R. and Y. U. Glasure (1998) Cointegration,
Error-correction, and the Relationship between GDP and Energy: The Case
of South Korea and Singapore. Resource and Energy Economics 20:1, 17-25.
Lee, C. C. and C. P. Chang (2008) Energy Consumption and Economic
Growth in Asian Economies: A More Comprehensive Analysis Using Panel
Data. Resource and Energy' Economics 30:1, 50-65.
Lindmark, M. (2002) An EKC-pattern in Historical Perspective:
Carbon Dioxide Emissions, Technology, Fuel Prices and Growth in Sweden
1870-1997. Ecological Economics 42: (1-2), 333-347.
Mabey, N. and R. McNally (1999) Foreign Direct Investment and the
Environment: From Pollution Havens to Sustainable Development. A WWF-UK
August 1999. Retrieved September 11, 2012, from assests.panda.org
Mackinnon, J. G. (1991) Critical Values for Cointegration Test.
Chapter 13. In R. F. Engle and J. Granger Long Run Economic
Relationships: Readings in Cointegration. Oxford University Press.
Merican, Y., Z. Yusop, Z. N. Mohd, and L. S. Hook (2007) Foreign
Direct Investment and the Pollution in Five ASEAN Nations. International
Journal of Economics and Management 1:2, 245-261.
Mulali, U. A. (2012) Factors Affecting C[O.sub.2] emission in the
Middle East: A Panel Data Analysis. Energy 44:1, 564-569.
Nasir, M., and F. U. Rehman (2011) Environmental Kuznets Curve for
Carbon Emissions in Pakistan: An Empirical Investigation. Energy Policy
39:3, 1857-1864.
Pakistan, Government of (2011) National Draft on Climate Change.
Ministry of Environment.
Pakistan, Government of (2011) Pakistan Economic Survey, 2011. The
Finance Division. Ministry of Finance, Pakistan.
Pao, H. T. and C. M. Tsai (2010) Multivariate Granger Causality
between C[O.sub.2] emissions, Energy Consumption, FDI (Foreign Direct
Investment) and GDP (Gross Domestic Product): Evidence from a Panel of
BRIC (Brazil, Russian Federation, India, and China) Countries. Energy
36:1, 685-693.
Pesaran, H. M., Y. Shin, and R. J. Smith (2001) Bounds Testing
Approaches to the Analysis of Level Relationships. Journal of Applied
Econometrics 16:1, 289-326.
Roegen, N. G. (1975) Energy and Economic Myths: Institutional and
Analytic Economics Essay. New York Programme on Press Inc.
Selden, T. M. and D. Song (1994) Environmental Quality and
Development: Is There a Kuznets Curve for Air Pollution Emissions?
Journal of Environmental Economics and Management 27:2, 147-162.
Shahbaz, M., S. Nasreen, and T. Afza (2011) Environmental
Consequences of Economic Growth and Foreign Direct Investment: Evidence
from Panel Data Analysis. (MPRA Paper No. 32547).
Shi, A. (2003) The Impact of Population Pressure on Global Carbon
Dioxide Emissions, 1975-1996: Evidence from Pooled Cross-country Data.
Ecological Economics 44:1, 29-42.
State Bank of Pakistan (Various Issues) International Investment
Position of Pakistan. Statistics and Data Warehouse. Pakistan.
UNCTAD (2009) UNCTAD Training Manual on Statistics for FDI and the
Operations of TNCs. FDI Flows and Stocks. I. Retrieved February 20,
2013, from United Nations Conference on Trade and Development Division
on Investment and Enterprise:
http://unctad.org/en/Docs/diaeia20091_en.pdf
UNCTAD (2012) World Investment Report 2012: Towards a New
Generation of Investment Policies. Switzerland: United Nations
Publications.
UNCTAD (2012) World Investment Report, 2015. Towards A New
Genration of Investment Policies. United Nations, Switzerland.
UNFCCC (2015) Ad Hoc Working Group on the Durban Platform for
Enhanced Action Second Session Part 12, Paris.
Wang, M. (2009) Manufacturing EDI and Economic Growth: Evidence
from Asian Economies. Applied Economics 41, 991-1002.
World Bank (2009) The World Development Indicators 2011 (WDI)
Database. World Bank.
Zhang, J., M. A. Cole, and R .J .R. Elliott (2011) Growth, Foreign
Direct Investment, and the Environment: Evidence from Chinese Cities.
Journal of Regional Science 51:1, 121-138.
Syed Sundus Raza <simdasmanzar@gmail.com> is Visiting
Faculty, Fatima Jinnah Women University, Rawalpindi. Anwar Hussain
<amvar@pide.org.pk> is Assistant Professor, Pakistan Institute of
Development Economics, Islamabad.
(1) Companies move their official operations to less developed
economies to take benefit of weak environmental law's or developing
countries may put a low price on their environment to make new foreign
investment much attractive, which leads to over use of natural resources
and environmental degradation [Mabey and McNally (1999)].
(2) Companies move their operations to developing economies to take
advantage of lower cost of production [Shahbaz, el al. (2011)].
(3) The foreign firms may use better management and advance
technology that can result in clean environment in host country
[Shahbaz, et al. (2011)].
(4) When trade cause the expansion of economic activity thus trend
to increase pollution [Dietzenbacher and Mukhopadhyay (2007)].
(5) Trade can induce technological spillovers that can lead to the
adoption of "cleaner" production techniques by host countries
[Dietzenbacher and Mukhopadhyay (2007)].
(6) It is the change in the share of dirty goods in GDP, because of
a price change favouring their production [Acharrya (2009)].
Table 1
Stationarity Results of the Study Variables
ADF Test Statistics
Order of
Variables Level First Difference Integration I(d)
Mt -3.02 -7.94 *** I(1)
GDP -1.06 -3.93 ** I(1)
G -2.15 -4.27 *** I(1)
[g.sup.2] -1.55 -3.90 *** I(1)
Tr 2.80 -12.76 *** I(1)
Man -4.16 *** -3.21 I(0)
Mn -3.46 ** -2.91 I(0)
Pd -3.84 *** -2.91 I(0)
Ene -2.63 -5.16 *** I(1)
Lab 0.02 -6.57 *** I(1)
Note: *, **, *** I0 percent, 5 percent and I percent level of
significance respectively. Both trend and intercept are included in
checking stationarity except for "pd" where only intercept is taken.
Table 2
Results of UECM for the Impact of FDI Inflow on Carbon Dioxide
Emissions
Dependent Variable = Amt (Metric Ton Per Capita)
Coefficient t-slat P values
Constant -1.05 -4.35 0.00
[mt.sub.t-1] -0.78 -6.89 0.00
[g.sub.t-1] -56.17 -5.83 0.00
[g.sup.2.sub.t-1] 268.9 2.54 0.01
[mn.sub.t-1] -0.000919 -1.53 0.13
[man.sub.t-1] -0.000785 -2.08 0.04
[tr.sub.t-1] -0.000866 -2.44 0.00
[pd.sub.t-1] -0.011 -4.00 0.00
[DELTA][mt.sub.t-1] 0.43 2.59 0.01
[DELTA]g 19.6 3.27 0.00
[DELTA][g.sup.2.sub.t-1] -461.04 -2.83 0.00
[DELTA]tr 0.00517 2.08 0.04
[DELTA][man.sub.t-1] 0.0078 3.75 0.00
[DELTA]pd 0.10 3.77 0.00
[DELTA][pd.sub.t-1] 0.0887 3.87 0.00
[D.sub.2007] 0.150 2.01 0.05
[D.sub.2008] 0.059 3.89 0.00
Breuseh-Godfrey Serial 0.02 (0.86)
Correlation LM Test,
F-statistics
R-square 0.70
Table 3
Normalised Long Run Results for the Impact of FDI Inflows on Carbon
Dioxide Emissions
Dependent Variable=Carbon Dioxide Emissions in Metric Tons Per
Capita (mt)
Coefficient t-stats P values
Constant 1.35 4.35 0.00
[g.sub.t] 72.01 5.83 0.00
[g.sup.2.sub.t] -344.74 -2.54 0.01
[T.sub.t] 0.0011 2.44 0.00
[Mn.sub.t] 0.0013 1.53 0.13
[Man.sub.t] 0.0012 2.08 0.04
[Pd.sub.t] 0.014 4.00 0.00
[D.sub.2007] 0.150 2.01 0.05
[D.sub.2008] 0.059 3.89 0.00
Table 4
Short Run Results of Impact of FDI Inflows on Carbon Dioxide
Emissions
Dependent Variable = [DELTA]mt
Coefficient t stats P values
Constant -1.05 -4.35 0.00
[DELTA]g 19.6 3.27 0.00
[DELTA][g.sup.2.sub.t-1] -461.04 -2.83 0.00
[DELTA]tr 0.0052 2.08 0.04
[DELTA][man.sub.t-1] 0.0078 2.01 0.05
[DELTA]pd 0.10 3.75 0.00
[DELTA][pd.sub.t-1] 0.088 3.88 0.00
[D.sub.2007] 0.150 2.01 0.05
[D.sub.2008] 0.059 3.89 0.00
ECT -0.78 -6.89 0.00
Table 5
Results of UECM for the Impact of FDI Inflow on GDP Growth
Dependent Variable = [DELTA]mt
Coefficient t stats P values
Constant -92.25 -2.88 0.00
[GDP.sub.t-1] -0.22 -2.74 0.00
[mn.sub.t-1] -13.45 -1.55 0.14
[man.sub.t-1] -5.34 -2.40 0.02
[tr.sub.t-1] -42.6 -5.83 0.00
[ene.sub.t-1] -0.12 -2.56 0.01
[lab.sub.t-1] 0.03 0.71 0.48
[DELTA]man 7.44 4.24 0.00
[DELTA]tr 3.23 1.95 0.06
[DELTA][lab.sub.t-2] -0.28 -3.41 0.00
[DELTA][tr.sub.t-2] 51.26 5.32 0.00
[DELTA][ene.sub.t-2] 0.38 6.78 0.00
[D.sub.2007] 0.73 2.34 0.03
[D.sub.2008] 0.50 4.41 0.00
Breusch-Godfrey Serial Correlation LM Test, 0.07 (0.72)
F-statistics
R-square 0.93
Table 6
Normalised Long Run Results for the Impact of FDI on GDP Growth
Dependent Variable = GDP (Million Rupees)
Coefficient t stats P values
Constant 419.3 2.88 0.00
[Man.sub.t] 24.27 2.40 0.02
[Mn.sub.t] 61.33 1.55 0.14
[Tr.sub.t] 193.43 5.83 0.00
[Ene.sub.t] 0.54 2.56 0.01
[Lab.sub.t] -0.13 -0.71 0.48
[D.sub.2007] 0.73 2.34 0.03
[D.sub.2008] 0.50 4.41 0.00
Table 7
Short Run Results of Impact of FDI Inflows on GDP Growth
[DELTA]GDP = Gross Domestic Product (Million Rupees)
Coefficient t stats P values
Constant -92.25 -2.88 0.00
[DELTA]man 7.44 4.24 0.00
[DELTA]tr 3.23 1.95 0.06
[DELTA][lab.sub.t-2] -0.28 -3.41 0.00
[DELTA][tr.sub.t-2] 51.26 5.32 0.00
[DELTA][ene.sub.t-2] 0.38 6.78 0.00
[D.sub.2007] 0.73 2.34 0.03
[D.sub.2008] 0.50 4.41 0.00
ECT -0.22 -2.74 0.00
