“Impact of FDI inflow, crude oil prices, and economic growth on CO2 emission in

Tunisia: Symmetric and asymmetric analysis through ARDL and NARDL

approach”

AUTHORS Tarek Ghazouani https://orcid.org/0000-0001-5879-100X

Tarek Ghazouani (2021). Impact of FDI inflow, crude oil prices, and economic
growth on CO2 emission in Tunisia: Symmetric and asymmetric analysis through
ARTICLE INFO
ARDL and NARDL approach. Environmental Economics, 12(1), 1-13.
doi:10.21511/ee.12(1).2021.01

DOI http://dx.doi.org/10.21511/ee.12(1).2021.01

RELEASED ON Thursday, 14 January 2021

RECEIVED ON Wednesday, 02 December 2020

ACCEPTED ON Tuesday, 12 January 2021

LICENSE This work is licensed under a Creative Commons Attribution 4.0 International
License

JOURNAL "Environmental Economics"

ISSN PRINT 1998-6041

ISSN ONLINE 1998-605X

PUBLISHER LLC “Consulting Publishing Company “Business Perspectives”

FOUNDER LLC “Consulting Publishing Company “Business Perspectives”

NUMBER OF REFERENCES NUMBER OF FIGURES NUMBER OF TABLES

49 3 10

© The author(s) 2021. This publication is an open access article.

businessperspectives.org
 Environmental Economics, Volume 12, Issue 1, 2021

Tarek Ghazouani (Tunisia)

Impact of FDI inflow, crude

BUSINESS PERSPECTIVES oil prices, and economic
LLC “СPС “Business Perspectives”
Hryhorii Skovoroda lane, 10,
Sumy, 40022, Ukraine
growth on CO2 emission
www.businessperspectives.org
in Tunisia: Symmetric
and asymmetric analysis
through ARDL and NARDL
approach
Abstract
This study explores the symmetric and asymmetric impact of real GDP per capita,
FDI inflow, and crude oil price on CO2 emission in Tunisia for the 1972–2016 period.
Using the cointegration tests, namely ARDL and NARDL bound test, the results show
that the variables are associated in a long run relationship. Long run estimates from
both approach confirms the validity of ECK hypothesis for Tunisia. Symmetric analy-
sis reveals that economic growth and the price of crude oil adversely affect the environ-
ment, in contrast to FDI inflows that reduce CO2 emissions in the long run. Whereas
the asymmetric analysis show that increase in crude oil price harm the environment
and decrease in crude oil price have positive repercussions on the environment. The
causality analysis suggests that a bilateral link exists between economic growth and
Received on: 2nd of December, 2020 carbon emissions and a one-way causality ranges from FDI inflows and crude oil prices
Accepted on: 12th of January, 2021 to carbon emissions. Thus, some policy recommendations have been formulated to
Published on: 14th of January, 2021 help Tunisia reduce carbon emissions and support economic development.
© Tarek Ghazouani, 2021 Keywords FDI inflow, crude oil prices, economic growth, CO2
emission, ARDL, NARDL, Tunisia

JEL Classification F21, Q43, Q56, C15

Tarek Ghazouani, Ph.D, Faculty of
Law, Economics and Management
of Jendouba University of Jendouba,
Tunisia.
INTRODUCTION
Climate change is among the greatest challenges facing humanity. It af-
fects all countries and can have devastating effects on communities and
individuals. Therefore, the achievement of sustainable development and
the abatement of global warming have challenged academics and deci-
sion-makers to focus on the drivers of CO2 emissions, in order to devel-
op the measures and policies needed to improve environmental quality.
The situation becomes even worse when it comes to promoting econom-
ic growth, since energy is an important determinant in achieving such
a goal (Stern, 2011). The environmental impact of economic growth is
based on a controversial assumption, provided by Grossman and Krueger
(1991) which so-called EKC (Environmental Kuznets Curve). It stipulates
This is an Open Access article, an inverted U-shaped link between level of development and emissions
distributed under the terms of the whereby a national pollution concentration increases as the economy de-
Creative Commons Attribution 4.0
International license, which permits velops to a certain level where it decreases again as the country uses its
unrestricted re-use, distribution, and increased affluence to further reduce pollution concentrations.
reproduction in any medium, provided
the original work is properly cited.
Conflict of interest statement: Similarly, given its role as a driver of economic growth, FDI inflows
Author(s) reported no conflict of interest could be detrimental to the environment of host countries, general-

http://dx.doi.org/10.21511/ee.12(1).2021.01 1
Environmental Economics, Volume 12, Issue 1, 2021

ly developing countries, where foreign firms producing polluting goods would seek to locate in those
countries with comparatively weak environmental regulations compared to their countries of origin
with environmental requirements. However, the environmental consequences of FDI flows can be
asymmetric, as the massive of FDI inflows does not necessarily have the same sign and magnitude of
effect as the sudden stop of FDI flows.

As Keynes (1936) states, the trend up-wards of any macroeconomic variable transforms unexpectedly
towards a negative trend, while a falling trend does not have the same abrupt turn towards an up-trend.
Assuming that increased FDI flows generate CO2 emissions through increased energy consumption,
however, it may have the same effect with their decrease. Shahbaz et al. (2014) attribute this impact to
the “ratchet effect” whereby energy consumption increases despite the fall in income caused by sudden
stop in capital flows. In addition, it may be that as a result of sudden reversals in FDI, domestic invest-
ment booms in response to domestic demand, leading in turn to increased energy consumption, which
increases CO2 emissions. In this respect, it is crucial to examine the possible asymmetry of the environ-
mental consequences of FDI flows in order to avoid inappropriate policies recommendations.

Building on the environmental and economic concerns, this study explores the symmetrical and asym-
metrical connection between economic growth and CO2 emissions with the incorporation of crude
oil prices and FDI inflows within the emissions function of the Tunisian economy. From an ecological
view, the crude oil price could indirectly affect energy use and environmental quality via its impact on
these factors. Indeed, two scenarios can occur and this depends on whether the country is a net oil‐ex-
porting or importing economies. According to Sturm et al. (2009), oil-exporting countries will increase
their economic activity as a result of the surge in their income following the rise in the price of crude
oil. Based on this idea, Nwani (2017) indicates that increase economic activity requires more energy
which means additional carbon emissions. However, for oil-importing countries, any drop in crude oil
prices could reduce their incentive to invest in environment-friendly energies and increase their use of
fossil fuel energy, thus contributing to CO2 emissions. Therefore, considered as a semi-exporter coun-
try, Tunisia is characterized by a strong vulnerability of the national economy to the volatility of inter-
national energy prices. Thus, this can change its behavior in terms of energy consumption, which will
ultimately reflect on the environment.

To the best of our knowledge, the symmetric and asymmetric impact of oil crude price along with FDI
inflows on CO2 emissions in Tunisia has not been earlier explored. Aiming to fill this void, this study
conducts an in-depth examination of the impact of economic growth, FDI inflows, and crude oil pric-
es on Tunisia’s CO2 emissions using advanced econometric methods. Firstly, the linear ARDL (Auto-
regressive Distributed Lag) was implemented to investigate the symmetric short- and long-run associa-
tion among candidate variables. While, the non-linear ARDL (NARDL) by Shin et al. (2014) has used
to explore this association asymmetrically, as to assess effects of both positive and negative changes in
FDI inflows and crude oil price on Tunisian’s CO2 emissions, respectively. Finally, the block exogeneity
Wald test has employed to analyze the directional causality among candidate variable.

1. LITERATURE REVIEW Bölük & Mert, 2015; Baek, 2016; Pata, 2018; Usman
et al., 2019; Malik et al., 2020), while other papers
There are a substantial number of empirical inves- do not (Apergis, 2016; Kang et al., 2016; Balsalobre
tigations dealing with the environmental conse- et al., 2015; Adu & Denkyirah, 2019). The contra-
quences of economic growth by analyzing them dictory results given are often explained either
through the EKC assumption (see Shahbaz and by differences in a country’s level of development
Sinha (2019) for a literature survey). However, (Apergis, 2016) or by differences in the additional
among these studies, there are those that support variables included in the emission function model
this EKC assumption (Apergis & Ozturk, 2015; (Amri, 2018). Likewise, In the case of Tunisia, the

2 http://dx.doi.org/10.21511/ee.12(1).2021.01
 Environmental Economics, Volume 12, Issue 1, 2021

validity of the EKC assumption is evenly disputed, ergy use and carbon dioxide emissions in both
as Ben Jebli and Ben Youssef (2015), Farhani and the short and long run. The same findings were
Ozturk (2015), and Amri (2017, 2018) reject the obtained by Agbanike et al. (2019) when explor-
EKC hypothesis. In contrast, Shahbaz et al. (2014), ing this relationship for the Venezuela. Contrarily,
Kwakwa et al. (2018), and Mahmood et al. (2019) other studies have found that rising crude oil pric-
proved the existence of EKC in Tunisian economy. es have a positive influence on the environment
Kwakwa et al. (2018) concluded that the occur- (Mensah et al., 2019; Malik et al., 2020). Here, the
rence of an inverted U-curve depends on sources results found are explained by the fact whether an
of CO2 emissions in Tunisia. oil-exporting or oil-importing economy. For net
oil-importing economies, a drop in crude oil price
Regarding the relationship between FDI inflows means an over-use of energy and this is reflected
and the environment degradation, evidence re- in higher CO2 emissions and the other way round
mains also unclear. Some evidence found that (Vielle & Viguier, 2007). While for oil-exporting
FDI inflows affect positively CO2 emissions (for economies, higher crude oil prices would lead to
instance, Al-Mulali, 2012; Omri et al., 2014; Haug higher incomes, which would increase the energy
& Ucal, 2019; Malik et al., 2020; Essandoh et al., consumption generated by more economic activi-
2020). Closely, this category of studies supports ties, thus causing more CO2 emissions, which in
the famous “pollution haven hypothesis” accord- turn worsens the quality of the environment and
ing to which developing countries suffer more vice versa (Sturm et al., 2009; Nwani, 2017).
from environmental pollution caused by multi-
national companies that transfer high-pollution
industries (Copeland & Taylor, 1994). Conversely, 2. DATA AND METHODOLOGY
other evidence have found that FDI inflows af-
fect negatively CO2 emissions (such as Al-Mulali 2.1. Data and empirical model
& Tang, 2013; Sung et al., 2018; Jiang et al., 2018).
Consequently, this group of studies supports the The data adopted in the present paper is based on
known “pollution halo hypothesis”, which postu- 46 observations for Tunisia spanning the period
lates that multinational companies disseminate 1972–2016. Data includes CO2 emissions as an in-
their clean technologies to host countries through dicator of environmental deterioration expressed
the transfer of innovative technologies. However, in metric tons per-capita, GDP per-capita (in con-
there are also other studies arguing that FDI in- stant 2010 USD), the net inflow of FDI per-capita
flows does not detrimental to the environment (in constant 2010 USD), which were collected from
(Lee, 2013; Chandran & Tang, 2013). Similarly for Word Bank Indicators. In addition, they include
Tunisia, there are mixed results regarding the re- crude oil price (in USD per-barrel), which was ex-
lationship between these two variables, as Hakimi tracted from BP Statistical Review.
and Hamdi (2016) prove a positive impact of FDI
flows on pollution while Abdouli and Hammami To assess the environmental effect of FDI in-
(2016) show an insignificant negative impact. flows, crude oil prices and economic growth in
Tunisia, the following specific model has been
As for the crude oil prices-CO2 emissions nexus, implemented:
although the existence of a few empirical stud-
ies, the evidence also seems varied. Some stud- ln Ct = ( )
β 0 + β1 ( ln GDPt ) + β 2 ln GDPt 2 + (1)
ies reveal that a rise in crude oil prices leads to
+ β3 ( ln FDI t ) + β3 ( ln OIPt ) + µt ,
increase in energy consumption which in turn
causes higher CO2 emission. For instance, Nwani where
2
(2017) find that the crude oil price, energy con- ln Ct , ln GDPt , ln GDPt , ln FDI t , and ln OIPt
sumption, and other economic performance indi- , are the natural logarithms of CO2 emissions
cators are cointegrated in a long-run relationship per-capita, GDP per-capita, the square term of
with CO2 emissions in the Ecuadorian for the pe- GDP per-capita, FDI per-capita and crude oil pric-
riod 1971–2013. Further, he reveals that the price es, respectively. The square of real GDP per-capita
2
of crude oil positively and significantly affects en- ( ln GDPt ) is included in equation 1 to implement

http://dx.doi.org/10.21511/ee.12(1).2021.01 3
 Environmental Economics, Volume 12, Issue 1, 2021

EKC hypothesis augmented by FDI inflows and of cointegration or not. Afterwards, the long- and
crude oil price. μt and t refer to the residual term short-run estimation will be performed if the
and the year, respectively. cointegration relationship has been confirmed.
The short-run model is presented by equation (3)
2.2. The symmetric analysis below in which the parameter δi reflects the long-
term equilibrium adjustment speed after a short-
To examine the symmetric association among FDI term occurred event.
inflows, crude oil price, economic growth, and
CO2 emissions, the paper apply the Pesaran et ∆ ln Ct =+ α δ i ECM t −1 +
al.’s (2001) cointegration test, namely ARDL test. n m

Characterized by its advantages, when compared

=i 1 =i 0
+ ϕ1∑∆ ln Ct −i + ∑
ϕ2 ∆ ln GDPt −i +
to other conventional cointegration methods, the m m (3)

scientists in recently years (for instance,

= Nwani,
∑
ARDL approach has become very popular amidst + ϕ ∆ ln GDP 2 + ϕ ∆ ln FDI +
3
i 0=i 0
t −1 4 ∑ t −i

2017; Agbanike et al., 2019; Malik et al., 2020). m

∑
Before implementing the ARDL model, wheth- + ϕ5 ∆ ln OIPt −i + ε t .
er linear or non-linear, the order of integration i =0

of variables must be identified. For this purpose,

some conventional unit root tests are used, such 2.3. The asymmetric analysis
as the ADF-test by Dickey-Fuller and the PP-test
by Phillips-Perron, along with the LM-test by Lee Since it is ignored by the linear ARDL model, this
and Strazicich (2003) to take into account for pos- paper uses the procedure of Shin et al. (2014) to
sible structural breaks in the series. Next step con- determine the possible asymmetric association
sist to assess the evidence of a long run association by conducting the NARDL model. Indeed, the
by conducting the F-test. The empirical expression NARDL model exhibits the same requirements for
of the ARDL bound test for cointegration is pre- the integration order of the variables as the ARDL
sented as follows: model. Thus, following Shin et al. (2014), equation
2 can be restated in the following form:
α β1 ln Ct −1 + β 2 ln GDPt −1 +
∆ ln Ct =+
α β1 ln Ct −1 + β 2 ln GDPt −1 +
∆ ln Ct =+
+ β3 ln GDPt −21 + β 4 ln FDI t −1 + β5 ln OIPt −1 +
n m + β3 ln GDPt −21 + β 4+ ln FDI t+−1 + β5− ln FDI t−−1 +
+ ∑ϕ ∆ ln C + ∑ϕ2 ∆ ln GDPt −i +
t −i (2) n
+ β 6+ ln OIPt −+1 + β 7− ln OIPt −−1 + ∑θi ∆ ln Ct −i +
1
=i 1 =i 0
m m i =1
+ ∑ϕ3 ∆ ln GDPt −21 + ∑ϕ 4 ∆ ln FDI t −i + m m (4)
=i 0=i 0 + ∑θi ∆ ln GDPt −i + ∑θi ∆ ln GDPt −21 +
m =i 0=i 0

+ ∑ϕ5 ∆ ln OIPt −i + ε t , m

i =0 + ∑(θ i+ ∆ ln FDI t+−i + θ i− ∆ ln FDI t−−i ) +

i =0
m
where n and m are the lag length obtained us- + ∑(θ i+ ∆ ln OIPt −+i + θ i− ∆ ln OIPt −−i ) + ε t .
ing SCI (Schwarz Information Criteria), and ε t i =0
indicates the white nose error term. While β1 to  m
  m

β5 and φ2 to φ5 represent the long- and short-run ∑
From equation (4), β i  θi  and β i  θi 
+

 i positive

+ −
∑ −

dynamic, respectively. The existence or not of a captures the long (short)-run and negative
i 
cointegrating relationship is based on the F-test. impact of FDIt and crude OIPt on Ct. As for ARDL
The null hypothesis (β1 = β2 = β3 = β4 = β5 = 0) is model, the bound test is resorted to determine
tested through Wald test based on F-statistics to whether the variables are asymmetrically co-inte-
verify whether there is cointegration or not. The grated or not. In addition, the Wald-test is called
compared F-statistics to the lower and the upper to assess the long (short) run symmetric linkage β
bound critical value would conclude the existence = β + = β – (θ = θ + = θ –) for FDI and OIP. Given the

4 http://dx.doi.org/10.21511/ee.12(1).2021.01
 Environmental Economics, Volume 12, Issue 1, 2021

validation of the non-linear relationship, the short us to apply the ARDL cointegration approach. The
run asymmetric association can be provided via result of the cointegration test presented in Table 2
the dynamic multiplier effect in the following way: shows a F-stat value (10.327) exceeding the critical
value of the upper bound of Narayan (2005) table
ω ln Ct − j at 1% level, which confirm the existence of cointe-
Ds+ = ∑ j =0
s
, gration among studied variables.
ω ln FDI t+−i
ω ln Ct − j The implementation of co-integration across vari-
Ds− = ∑ j =0
s
, ables leads us to assess the short and long run im-
ω ln FDI t−−i
(5) pact of GDP, FDI and OIP on carbon emissions
=s 0,1, 2,3, …. (C). Tables 3 and 4 outline that economic growth
....Noting that s → ∞, Ds+ → β 4+ , Ds− → β5− , affects significantly and positively CO2 emissions
in the long run. While, the quadratic term of real
ω ln Ct − j
Ds+ = ∑ j =0
s
, GDP per capita improve significantly the environ-
ω ln OIPt −+i ment, which leads us to conclude that the invert-
ω ln Ct − j ed U-shaped EKC for Tunisia stands only on long
Ds− = ∑ j =0
s
, run with a turn point equal to 4,590.126 USD per
ω ln OIPt −−i
capita. The validation of the EKC assumption cor-
=s 0,1, 2,3, …. roborates the findings by Shahbaz et al. (2014) and
...Noting that s → ∞, Ds+ → β 6+ , Ds− → β 7− . Mahmood et al. (2019). The coefficient of ECT(–1)
is statistically significant and negative indicating
3. EMPIRICAL RESULTS that around 56% of the short-term unbalance rec-
tified in the long term.
AND DISCUSSION
Table 2. F-Bound test for linear cointegration
ARDL (1,0,0,3,4)
3.1. The symmetric analysis results
Null hypothesis: No levels
F-Bounds Test
Checking the stationarity of the series is a prerequi- relationship
Significance
site for deciding which cointegration test to choose. I(0) I(1)
10% 2.2 3.09
Using the ADF and PP tests, the empirical out- F-statistic 10.327
5% 2.52 3.49
come in Table 1 shows that some variables are in-
2.5% 2.82 3.87
tegrated in level, that is I(0) and other after the first k 4
1% 3.29 4.37
difference, that is I(1), while, when using the LS
test which take into account for possible structural
breaks, all variables are integrated in same order, The crude oil price affects CO2 emissions positive-
that is I(1). Therefore, altogether, these results allow ly and significantly at 1% level, as the results show

Table 1. Unit root test

Tests ADF PP LS
Variables T–statistic T–statistic T–statistic TB1, TB2
C –2.498 –2.737 –2.262 1982, 1991
∆C –8.547* –8.396* –6.890* 1981, 1988
GDP –1.464 –1.516 –3.306 1978, 1995
∆GDP –9.215* –8.819* –6.272** 1989, 2009
GDP2 –1.206 –1.234 –3.176 1978, 1995
∆GDP2 –8.851* –8.524* –6.990** 1989, 2009
FDI –2.878*** –2.779*** –5.466 1982, 2006
∆FDI –8.938* –18.863* –7.280* 2004, 2009
OIP –3.031** –3.037** –4.738 1983, 2003
∆OIP –5.905* –5.906* –7.599* 1985, 2003

Note: *, **, *** indicates that value is significant at 1, 5, and 10% threshold, respectively. TB indicates time break.

http://dx.doi.org/10.21511/ee.12(1).2021.01 5
Environmental Economics, Volume 12, Issue 1, 2021

that a 1% increase in the OIP is liable to lead to which pushes countries to subsidy energy con-
an increment of around 0.027% in CO2 emissions sumption. A study conducted by the International
in the Tunisian economy. This positive impact ex- Monetary Fund has shown that the percentage of
erted by crude oil price on CO2 emissions might subsidies for fuel amounted to 13.7% of the state
be explained through two mechanisms. Firstly, budget in 2013. This study also revealed that the
the income generated by crude oil export rev- fuel subsidy policy in Tunisia leads to excessive en-
enues and the tax revenues of the oil companies ergy consumption and a reduction in investment
provide economic conditions that produce more incentives in the field of rationalization of con-
energy consumption. Indeed, considered as semi sumption and renewable energies.
oil-dependent economy, in 2018, Tunisia exported
USD 614 million of crude oil, making it the 50th FDI significantly and negatively affects carbon
largest crude oil exporter in the world (Energy emissions in the long run at the level of 5%, which
Information Administration). In the same year, leads to accept the presence of the halo hypothe-
crude oil was the 5th most exported product in sis. This might explain the contribution of FDI to
Tunisia. The oil revenue average value was 5.38 per the reduction of emissions by the fact that inves-
cent of GDP during the period from 1970 to 2018 tors, when they enter Tunisia, tend to respect en-
(World Bank Indicator, 2019). vironmental regulations and international stand-
ards. In addition, this can suggest that FDI inflow
In addition, and according to the Central Bank of contributes to the transfer of clean technologies
Tunisia report (2017), the tax revenue of oil com- which lead to improve energy efficiency and the
panies constitutes about 40% of total direct taxes, reduction of greenhouse gas emissions. This find-
thus improving the country total income. In this ing is in accordance with the evidence of Zhang
respect, the rise in crude oil prices contributes to and Zhou (2016) for the Chinese economy.
economic growth in Tunisia, which in turn gener-
ates more economic activity, particularly in ener- Table 3 documents also the diagnostic analy-
gy-intensive sectors. Brini et al. (2017) reveal that sis (lower part), were the Jarque & Bera-test and
a 1% increment in the oil price increases Tunisia’s Breusch & Godfrey L-M-test suggest normali-
economic growth by 1.175%. Secondly, the subsi- ty and absence of serial correlation in the model,
dization of energy consumption encourages more respectively, while the ARCH-test and Breusch-
energy use. The rise in the price of a barrel of oil Pagan-Godfrey heteroscedasticity test accept the
leads to additional costs for economic institutions, alternate hypothesis of heteroscedasticity. In addi-
Table 3. Linear ARDL model long-run results
Dependent variable: Ct
Variable Coefficient St. error T-stat
GDPt 14.806 1.509 9.807*
GDPt2 –0.878 0.095 –9.178*
FDIt –0.029 0.014 –2.023**
OIPt 0.027 0.014 1.870***
C –61.451 5.953 –10.322*
R2 0.9678
Adj-R2 0.963
F-stat 293.300*
Diagnostic analysis
Test F-statistic P-Value
Normal 0.219 0.896
Seial 0.761 0.476
ARCH 0.325 0.571
Hetero 1.396 0.236
CUSUM Stable
CUSUM-SQ Stable

Note: *, **, *** indicates that value is significant at 1, 5, and 10% threshold, respectively.

6 http://dx.doi.org/10.21511/ee.12(1).2021.01
 Environmental Economics, Volume 12, Issue 1, 2021

Table 4. Linear ARDL model short-run results

Dependent variable: Ct
Variable Coefficient St. error T-stat
∆GDP t 7.481 5.527 1.946
∆GDP t2 –0.461 0.355 –1.296
∆FDIt –0.016 0.011 –1.397
∆OIP t 0.004 0.024 0.179
ECT(–1) –0.559 0.134 –4.176*
C 0.018 0.009 1.946

Note: * refers to the significance at 1% level.

tion, the CUSUM and CUSUMSQ tests presented emissions (Tables 6 and 7). Similarly, the NARDL
in Figure 1, show that the global model is stable model supports the reversed U-curve over the
which reinforces the robustness of our results. long term, as does ARDL model.
Table 6. NARDL model long-run results
3.2. The symmetric analysis results
Dependent variable: Ct
The NARDL-related test findings reported in Variable Coefficient St. error T-stat
Table 5 indicated the existence of long term rela- GDPt 20.453 2.249 9.091*
tionships across the variables. The coefficient of GDPt2 –1.270 0.140 –9.056*
ECT(–1) is statistically significant and negative in- FDIt+ –0.048 0.025 –1.955**
dicating that around 64% of the short‐term short- FDIt- –0.024 0.034 –0.709
OIPt+ 0.251 0.072 3.459*
term unbalance rectified in the long term.
OIPt- –0.057 0.043 –1.757**
Table 5. F-Bound test for non-linear cointegration C –82.058 9.080 –9.037*
R2 0.985
Null hypothesis: No levels Adj-R2 0.978
F-Bound test relationship F-stat 147.854*
Significance I(0) I(1) Diagnostic analysis
F-statistic 6.447
10% 1.99 2.94 Test F-statistic P-Value
5% 2.27 3.28 Normal 0.114 0.944
2.5% 2.55 3.61 Serial 2.164 0.135
k 6
1% 2.88 3.99 ARCH 0.270 0.606
Hetero 0.967 0.503
CUSUM Stable
Regarding the long and short term estimation of CUSUM-SQ Stable
the NARDL model, the results show similarities
Note: *, ** indicates that value is significant at 1 and 5%
with those found in the ARDL methodology with threshold, respectively.
respect the impact of economic growth on carbon
20 1.4

15 1.2

1.0
10
0.8
5
0.6
0
0.4
–5
0.2
– 10
0.0
– 15
– 0.2
– 20 – 0.4
1980 1985 1990 1995 2000 2005 2010 2015 1980 1985 1990 1995 2000 2005 2010 2015

Cusum 5% Significance Cusum 5% Significance

Figure 1. Plots of CUSUM and CUSUMSQ tests

http://dx.doi.org/10.21511/ee.12(1).2021.01 7
Environmental Economics, Volume 12, Issue 1, 2021

Table 7. NARDL model short-run results

Dependent variable: Ct
Variable Coefficient St. error T-stat.
∆GDPt 4.579 5.368 0.852
∆GDPt2 –0.285 0.344 –0.828
∆FDIt+ –0.008 0.015 –0.557
∆FDIt- 0.004 0.028 0.153
∆OIPt+ 0.077 0.033 2.346**
∆OIPt- –0.043 0.030 –1.420
ECT(-1) –0.647 0.128 –5.047*
C 0.0112 0.012 0.867

Note: * and **, *** refer to the significance at 1 and 5% level respectively.

As for the asymmetric FDI-CO2 emissions link- CO2 emissions in the short term besides the long
age, only the positive shocks in FDI (FDI+) cast term, while the negative shock in the partial sum
significant effect as regards carbon emissions of crude oil price (OIP-) is significantly negative,
which suggests that the Halo hypothesis is valid meaning that any decrease in crude OIP decreases
for the Tunisian economy. However, the absence in CO2 emissions in the long-run (Tables 6 and
of an asymmetrical association is verified by the 7). Both the Wald test (Table 8) and the dynam-
Wald test that evaluates the equality between the ic multiplier graph of OIP (Figure 2) support the
coefficients of the two FDI shocks (Table 8), which asymmetrical association, since it is significant
is quite clear in Figure 2 who presents the dynamic long run.
multiplier graph of FDI.
Table 11 presents equally diagnostics analy-
As for the asymmetric crude oil price-CO2 emis- sis output (lower part) and shows no issue of
sions nexus, the positive shock in the partial sum normality, serial correlation, specification, ho-
of crude oil price (OIP+) is significantly positive, moscedasticity, and stability for the studied
meaning that any increase in crude OIP increases model (Figure 2).
Table 8. NARDL Wald test

Short-run Long-run
Exogenous variable
F-stat. Prob. F-stat. Prob.
FDI 2.268 0.143 0.427 0.518
OIP 2.123 0.156 3.997 0.055*

Note: * refers to the significance at 10% level.

.4
.08

.3
.04

.00 .2

–.04 .1

–.08 .0

–.12 –.1
1 3 5 7 9 11 13 15 1 3 5 7 9 11 13 15

Multiplier for FDI (+) Multiplier for FDI (+)

Multiplier for FDI (–) Multiplier for FDI (–)
Asymmetry plot (with C.I.) Asymmetry plot (with C.I.)

Figure 2. NARDL dynamic multiplier effect graphs

8 http://dx.doi.org/10.21511/ee.12(1).2021.01
 Environmental Economics, Volume 12, Issue 1, 2021

Table 9. Causality analysis

Dependent variable
Independent variables
Ct GDPt GDPt2 FDIt OIPt
Ct – 5.284*** 4.812*** 4.239 0.366
GDPt 14.833* – 7.838** 6.285*** 3.839
GDPt2 14.220* 6.456** – 6.056*** 3.444
FDIt 10.767*** 1.003 1.036 – 1.046
OIPt 9.247*** 1.302 1.392 21.244* –
Diagnostic analysis F-stat. Prob.
Normality 5.275 0.872
Heteros. 258.781 0.677
Serial Corr. 0.888 0.618

Note: *, **, *** indicates that value is significant at 1, 5, and 10% threshold, respectively.

Table 10. FMOLS and CCR estimates

FMOLS CCR
Variable Coefficient St. error T-Stat. Coefficient St. error T-Stat.
GDP 14.804 1.674 8.842* 14.830 1.448 10.238*
YSQ –0.878 0.106 –8.280* –0.880 0.092 –9.528*
FDI –0.030 0.016 –1.923*** –0.033 0.019 –1.735***
OIP 0.038 0.018 2.099** 0.0380 0.017 2.131**
C –61.45 6.612 –9.293* –61.566 5.685 –10.829*
R2 0.966 0.966
Adj-R2 0.962 0.962

Note: *, **, *** indicates that value is significant at 1, 5, and 10% threshold, respectively.

3.3. Causality analysis results 3.4. Robustness check

Table 9 presents the combine causality outcomes To check the robustness of these results, the FMOLS
of the long and short-run which has examined (Fully Modified Ordinary Least Square) and CCR
by the Block Exogeneity Wald test. The analysis (Canonical Cointegration Regression) are applied
shows a two-way causal direction among GDP to re-estimate the specific model (1). Table 10 pre-
per-capita and CO2 emissions, thus supporting sents their results and shows that the results ob-
the feed-back effect, while a one-way causali- tained by both FMOLS and CCR are almost in
ty running from FDI flow and crude oil price line with the results of long run ARDL estimates
to CO2 emission. This result is consistent with in terms of sign and significance which proves the
the ARDL and non-linear ARDL evidences. robustness of the model. Moreover, these two es-
However, no issues occurred during the diag- timation techniques affirm that the EKC assump-
nostic analysis. tion is well established in the Tunisian economy.

CONCLUSION
Using annual data covering the 1971–2016 period, this study explores the symmetric and asymmetric
association of real GDP per capita, FDI inflow, and crude oil price on CO2 emission of Tunisia. Different
from previous studies on Tunisia, this is the first attempt to examine the individual effect of FDI inflow
and the crude oil price on CO2 emissions, while giving more solid evidence for the EKC hypothesis by
skipping energy use as an explanatory variable that induces a systematic volatility in the coefficients of
the model (Itkonen, 2012). Results from the linear ARDL have suggested a positive effect of crude oil
price as regards CO2 emission in the long term, while the negative and significant component of FDI
inflow validates the pollution Halo hypothesis for Tunisia. Whereas, Results from the NARDL indicate

http://dx.doi.org/10.21511/ee.12(1).2021.01 9
Environmental Economics, Volume 12, Issue 1, 2021

that in the long term, a rise in crude oil price intensifies CO2 emission, while the decrease in oil price
mitigates CO2 emission in Tunisia. Results from causality analysis divulge that there is feedback effect
between economic growth and CO2 emission, while a one-way causal direction ranges from crude oil
price to CO2 emission. Findings from the robustness check (FMOLS and CCR) reinforce the results of
long run ARDL estimates as well as the validity of the EKC hypothesis. Thus, from these findings, some
policy recommendations can be drawn.

POLICY RECOMMENDATIONS
The results of this study warrant attention from Tunisian policymakers. A notable policy implication
from the EKC findings is that Tunisia must never back down in its attempts to spur economic expansion,
as the results confirm that country can in fact develop from overall carbon emissions through steady
income growth. From the standpoint of FDI inflows, the negative effect on carbon emissions should en-
courage Tunisia to attract more foreign direct investment. For example, preferential taxes, in addition
to access to the domestic market, could further attract foreign direct investment.

The result of crude oil prices effect is not surprising because of subsidies on fuel prices on the one hand,
and my be also because of the growth in economic activity generated by income from oil export and
tax revenues of oil companies on the other hand. Consequently, Tunisian policy makers should review
its current fuel pricing policies to insure that national fuel prices reflect international market prices
and thereby keeping CO2 emissions under control. Furthermore, it’s interestingly that income from oil
exports and from the taxation of oil companies could be used to develop clean and environmentally
friendly energy sources. The promotion of renewable sources can expand opportunities for cleaner al-
ternatives power and reduce the build-up of greenhouse gases, thus sustains environment.

AUTHOR CONTRIBUTIONS
Conceptualization: Tarek Ghazouani.
Formal analysis: Tarek Ghazouani.
Investigation: Tarek Ghazouani.
Methodology: Tarek Ghazouani.
Visualization: Tarek Ghazouani.
Writing – original draft: Tarek Ghazouani.
Writing – review & editing: Tarek Ghazouani.

REFERENCES
1. Abdouli, M., & Hammami, S. 40(2), 281-288. https://doi. Middle East: a panel data analysis.
(2017). Economic growth, FDI org/10.1016/j.kjss.2017.12.008 Energy, 44(1), 564-569. https://doi.
inflows and their impact on the org/10.1016/j.energy.2012.05.045
3. Agbanike, T. F., Nwani, C., Uwazie,
environment: an empirical study
U. I., Anochiwa, L. I., Onoja T-G. 5. Al-mulali, U., & Tang, C. F.
for the MENA countries. Quality
C., & Ogbonnaya, I. O. (2019). Oil (2013). Investigating the validity
and Quantity, 51, 121-146. https://
price, energy consumption and of pollution haven hypothesis
doi.org/10.1007/s11135-015-
carbon dioxide (CO2) emissions: in the gulf cooperation council
0298-6
insight into sustainability (GCC) countries. Energy Policy, 60,
2. Adu, D. T., & Denkyirah, E. K. challenges in Venezuela. Latin 813-819. https://doi.org/10.1016/j.
(2019). Economic growth and American Economic Review, 28, 8. enpol.2013.05.055
environmental pollution in West https://doi.org/10.1186/s40503-
6. Amri, F. (2017). Intercourse across
Africa: testing the environmental 019-0070-8
economic growth, trade and
Kuznets curve hypothesis.
4. Al-mulali, U. (2012). Factors renewable energy consumption
Kasetsart Journal of Social Sciences,
affecting CO2 emission in the in developing and developed

10 http://dx.doi.org/10.21511/ee.12(1).2021.01
 Environmental Economics, Volume 12, Issue 1, 2021

countries. Renewable and environmental Kuznets curve and urbanization in Tunisia.

Sustainable Energy Review, 69, in Turkey: an ARDL approach. Environmental Science and
527-534. https://doi.org/10.1016/j. Renewable and Sustainable Energy Pollution Research, 22, 15663-
rser.2016.11.230 Review, 52, 587-595. https://doi. 15676. https://doi.org/10.1007/
org/10.1016/j.rser.2015.07.138 s11356-015-4767-1
7. Amri, F. (2018). Carbon dioxide
emissions, total factor productivity, 15. Boufateh, T. (2019). The 22. Grossman, G. M., & Krueger, A.
ICT, trade, financial development, environmental Kuznets curve by B. (1991). Environmental impacts
and energy consumption: considering asymmetric oil price of a North American free trade
testing environmental Kuznets shocks: evidence from the top agreement (NBER Working Papers
curve hypothesis for Tunisia. two. Environmental Science and No. 3914). Retrieved from https://
Environmental Science and Pollution Research, 26(1), 706-720. www.nber.org/papers/w3914
Pollution Research, 25, 33691- https://doi.org/10.1007/s11356-
33701. https://doi.org/10.1007/ 23. Hakimi, A., & Hamdi, H.
018-3641-3
s11356-018-3331-1 (2016). Trade liberalization,
16. Brini, R., Amara, M., & Jemma- FDI inflows, environmental
8. Apergis, N. (2016). Environmental li, H. (2017). Renewable energy quality and economic growth: A
Kuznets curves: new evidence on consumption, International trade, comparative analysis between
both panel and country-level CO2 oil price and economic growth Tunisia and Morocco. Renewable
emissions. Energy Economics, 54, inter-linkages: The case of Tunisia. and Sustainable Energy Reviews,
263-271. https://doi.org/10.1016/j. Renewable and Sustainable Energy 58, 1445-1456. https://doi.
eneco.2015.12.007 Reviews, 76, 620-627. https://doi. org/10.1016/j.rser.2015.12.280
9. Apergis, N., & Ozturk, I. (2015). org/10.1016/j.rser.2017.03.067
24. Haug, A. A., & Ucal, M. (2019).
Testing environmental Kuznets 17. Chandran, V. G. R., & Tang, C. F. The role of trade and FDI for CO2
curve hypothesis in Asian (2013). The impacts of transport emissions in Turkey: nonlinear
countries. Ecological Indicators, 52, energy consumption, foreign relationships. Energy Economics,
16-22. https://doi.org/10.1016/j. direct investment and income 81, 297-307. https://doi.
ecolind.2014.11.026 on CO2 emissions in ASEAN-5 org/10.1016/j.eneco.2019.04.006
10. Baek, J. (2016). Do nuclear and economies. Renewable and
25. Itkonen, J. V. A. (2012). Problems
renewable energy improve the Sustainable Energy Reviews, 24,
estimating the carbon Kuznets
environment? Empirical evidence 445-453. https://doi.org/10.1016/j.
curve. Energy, 39(1), 274-280.
from the United States. Ecological rser.2013.03.054
https://doi.org/10.1016/j.en-
Indicators, 66, 352-356. https://doi. 18. Cherni, A., & Jouini, S. E. (2017). ergy.2012.01.018
org/10.1016/j.ecolind.2016.01.059 An ARDL approach to the CO2
26. Jiang, L., Zhou, H. F., Bai,
11. Balsalobre, D., Álvarez, A., & Can- emissions, renewable energy and
L., & Zhou, P. (2018). Does
tos, J. M. (2015). Public budgets economic growth nexus: Tunisian
foreign direct investment drive
for energy RD&D and the effects evidence. International Journal of
environmental degradation
on energy intensity and pollution Hydrogen Energy, 42(48), 29056-
in China? An empirical study
levels. Environmental Science and 29066. https://doi.org/10.1016/j.
based on air quality index from
Pollution Research, 22, 4881-4892. ijhydene.2017.08.072
a spatial perspective. Journal of
https://doi.org/10.1007/s11356- 19. Copeland, B. R., & Taylor, M. S. Cleaner Production, 176, 864-
014-3121-3 (1994). North-south trade and 872. https://doi.org/10.1016/j.
12. Ben Jebli, M., & Ben Youssef, the environment. The Quarterly jclepro.2017.12.048
S. (2015). The environmental Journal of Economics, 109(3), 755-
27. Kang, Y. K., Zhao, T., & Yang, Y.Y.
Kuznets curve, economic growth, 787. Retrieved from http://www.
(2016). Environmental Kuznets
renewable and non-renewable jstor.org/stable/2118421
curve for CO2 emissions in China:
energy, and trade in Tunisia. 20. Essandoh, O. K., Islam, M., & A spatial panel data approach.
Renewable and Sustainable Energy
Kakinaka, M. (2020). Linking Ecological Indicators, 63, 231-
Review, 47, 173-185. https://doi.
international trade and foreign 239. https://doi.org/10.1016/j.
org/10.1016/j.rser.2015.02.049
direct investment to CO2 ecolind.2015.12.011
13. Ben Jebli, M., Ben Youssef, S., emissions: Any differences
28. Keynes, J. M. (1936). The General
& Ozturk, I. (2016). Testing between developed and
Theory of Employment, Interest and
environmental Kuznets curve developing countries. Science
Money. Macmillan: London, UK.
hypothesis: the role of renewable of The Total Environment, 712,
Retrieved from https://www.files.
and nonrenewable energy 136437. https://doi.org/10.1016/j.
ethz.ch/isn/125515/1366_Keynes-
consumption and trade in OECD scitotenv.2019.136437
TheoryofEmployment.pdf
countries. Ecological Indicators, 60,
21. Farhani, S., Ozturk, I. (2015).
824-831. https://doi.org/10.1016/j. 29. Kwakwa, P. A, Alhassan, H., &
Causal relationship between
ecolind.2015.08.031 Aboagye, S. (2018). Environmental
CO2 emissions, real GDP,
Kuznets curve hypothesis in
14. Bölük, G., & Mert, M. (2015). The energy consumption, financial
a financial development and
renewable energy, growth and development, trade openness,

http://dx.doi.org/10.21511/ee.12(1).2021.01 11
Environmental Economics, Volume 12, Issue 1, 2021

natural resource extraction 41(3), 201-225. https://doi. 43. Stern, D. I. (2011). The role of
context: evidence from Tunisia. org/10.1111/opec.12102 energy in economic growth.
Quantitative Finance and Annals of the New York Academy
36. Omri, A., Nguyen, D. K., & Rault,
Economics, 2(4), 981-1000. of Sciences, 1219(1), 26-51.
C. (2014) Causal interactions
Retrieved from https://www. https://doi.org/10.1111/j.1749-
between CO2 emissions, FDI, and
researchgate.net/publica- 6632.2010.05921.x
economic growth: evidence from
tion/329777135_Environmental_
dynamic simultaneous-equation 44. Sturm, M., Gurtner, F. J., &
Kuznets_curve_hypothesis_in_a_
models. Economic Modelling, 42, Gonzalez, J. (2009). Fiscal
financial_development_and_natu-
382-389. https://doi.org/10.1016/j. policy challenges in oil-exporting
ral_resource_extraction_context_
econmod.2014.07.026 countries- A review of key issues
evidence_from_Tunisia
(ECB Occasional Paper Series
37. Pata, U. K. (2018). The influence
30. Lee, J. W. (2013). The contribution No.104). Frankfurt, Germany.
of coal and noncarbohydrate
of foreign direct investment Retrieved from https://papers.ssrn.
energy consumption on CO2
to clean energy use, carbon com/sol3/papers.cfm?abstract_
emissions: revisiting the
emissions and economic growth. id=1325245
environmental Kuznets curve
Energy Policy, 55, 483-489.
hypothesis for Turkey. Energy, 45. Sung, B., Song, W. Y., & Park,
https://doi.org/10.1016/j.en-
160, 1115-1123. https://doi. S. D. (2018). How foreign
pol.2012.12.039
org/10.1016/j.energy.2018.07.095 direct investment affects CO2
31. Mahmood, H., Maalel, N., & emission levels in the Chinese
38. Pesaran, M. H., Shin, Y., & Smith,
Zarrad, O. (2019). Trade manufacturing industry: evidence
R. J. (2001). Bounds testing
Openness and CO2 Emissions: from panel data. Economic
approaches to the analysis of level
Evidence from Tunisia. System, 42(2), 320-331. https://doi.
relationships. Journal of Applied
Sustainability, 11(12), 3295. org/10.1016/j.ecosys.2017.06.002
Econometrics, 16(3), 289-326.
https://doi.org/10.3390/
https://doi.org/10.1002/jae.616 46. Usman, O., Iorember, P. T.,
su11123295
& Olanipekun, I. O. (2019).
39. Shahbaz, M., & Sinha, A. (2019).
32. Malik, M. Y., Latif, K., Khan, Revisiting the environmental
Environmental Kuznets curve for
Z., Butt, H. D, Hussain, M., & Kuznets curve (EKC) hypothesis
CO2 emissions: a literature survey.
Nadeem, M. A. (2020). Symmetric in India: the effects of energy
Journal of Economic Studies, 46(1),
and Asymmetric Impact of Oil consumption and democracy.
106-168. https://doi.org/10.1108/
Price, FDI and Economic Growth Environmental Science and
JES-09-2017-0249
on Carbon Emission in Pakistan: Pollution Research, 26(13), 13390-
Evidence from ARDL and Non- 40. Shahbaz, M., Balsalobre-Lorente, 13400. https://doi.org/10.1007/
Linear ARDL Approach. Science D., & Sinha, A. (2019). Foreign s11356-019-04696-z
of the Total Environment, 726, direct Investment-CO2 emissions
47. Vielle, M., & Viguier, L. (2007).
138421. https://doi.org/10.1016/j. nexus in Middle East and North
On the climate change effects
scitotenv.2020.138421 African countries: Importance
of high oil prices. Energy Policy,
of biomass energy consumption.
33. Mensah, I. A., Suna, M., Gaoa, 35(2), 844-849. https://doi.
Journal of Cleaner Production, 217,
C., Omari-Sasub, A. Y., Zhuc, org/10.1016/j.enpol.2006.03.022
603-614. https://doi.org/10.1016/j.
D., Ampimaha, B. C., & Quar-
jclepro.2019.01.282 48. World Bank Indicator. (2019).
coo, A. (2019). Analysis on
World Development Indicators.
the nexus of economic growth, 41. Shahbaz, M., Khraief, N., Uddin,
Retrieved from https://datacatalog.
fossil fuel energy consumption, G. S., & Ozturk, I. (2014).
worldbank.org/dataset/world-
CO2 emissions and oil price in Environmental Kuznets curve
developmentindicators
Africa based on a PMG panel in an open economy: A bounds
ARDL approach. Journal of testing and causality analysis 49. Zhang, C., & Zhou, X. (2016).
Cleaner Production, 228, 161- for Tunisia. Renewable and Does foreign direct investment
174. https://doi.org/10.1016/j. Sustainable Energy Reviews, 34, lead to lower CO emissions?
jclepro.2019.04.281 325-336. https://doi.org/10.1016/j. Evidence from a regional
rser.2014.03.022 analysis in China. Renewable and
34. Nicolli, F., & Vona, F. (2019).
Sustainable Energy Reviews, 58,
Energy market liberalization 42. Shin, Y., Yu, B., & Greenwood-
943-951. https://doi.org/10.1016/j.
and renewable energy policies Nimmo, M. (2014). Modelling
rser.2015.12.226
in OECD countries. Energy asymmetric cointegration
Policy, 128, 853-867. https://doi. and dynamic multipliers in a
org/10.1016/j.enpol.2019.01.018 nonlinear ARDL framework. In
R. C. Sickles, & W. C. Horrace
35. Nwani, C. (2017). Causal
(Eds.), Festschrift in Honor of Peter
relationship between crude oil
Schmidt: Econometric Methods
price, energy consumption and
and Applications (pp. 281-314).
carbon dioxide (CO2) emissions
Springer, New York. https://doi.
in Ecuador. OPEC Energy Review,
org/10.1007/978-1-4899-8008-3_9

12 http://dx.doi.org/10.21511/ee.12(1).2021.01
 Environmental Economics, Volume 12, Issue 1, 2021

APPENDIX A
16 2.0

12
1.5
8

4
1.0

–4 0.5

–8
0.0
– 12

– 16 – 0.5
1990 1995 2000 2005 2010 2015 1990 1993 1996 1999 2002 2005 2008 2011 2014

Cusum 5% Significance Cusum 5% Significance

Figure A1. Plots of CUSUM and CUSUMSQ tests (NARDL model)

http://dx.doi.org/10.21511/ee.12(1).2021.01 13