Professional Documents
Culture Documents
F 2015 061 Original Paper 1641 1648
F 2015 061 Original Paper 1641 1648
F 2015 061 Original Paper 1641 1648
net/publication/292392031
CITATION READS
1 249
3 authors:
Bekir Ayyıldız
Bozok University
29 PUBLICATIONS 23 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
İklim Değişikliği ve Kuraklığın Bitkisel Üretime Etkileri Yeşilırmak Havzası Örneği View project
All content following this page was uploaded by Bekir Ayyıldız on 20 June 2016.
1641
© by PSP Volume 24 – No 5. 2015 Fresenius Environmental Bulletin
1990-1994, 2008. In these years, there was an increase of on total economic level. Whereas the studies found some
oil prices and economic crises. On the other hand, transport evidences for the causality in sectoral level.
and services sectors show increasing energy consumption
from 1990 to 2011. This paper, was performed on the total economy and
on main sectors, (industry, service, transport and agricul-
In several studies [6-8] were investigated the relation- ture). The analysis were tested for individually main sec-
ship between economic growth and energy consumption tors in the EU. The results may be important for shared pol-
both on total economy and the sectoral level. In these studies icy makers and appliers in EU.
were not found the relationship between energy and growth
FIGURE 2 - The plots energy consumption for the EU countries (000 tonnes of oil equivalent)
1642
© by PSP Volume 24 – No 5. 2015 Fresenius Environmental Bulletin
This study use three different panel unit root tests in-
cluding Im, Pesaran, and Shin [15] (hereinreferred to as
IPS); Levin, Lin and Chu [16] (herein referred to as LLC); , 3
Breitung [17] (herein referred to as BRT) to assess the sta-
tionary properties of the used variables.
Eq (1) shows the IPS analysis model: ,
, 4
∆ , ∆ , ,
1, … , , 1, … , 1
Where and represents country-specific effects
of i th each country in the panel and and are the
In this formula yit represents the series for Ith country. distortion terms. The correlation between variables are cal-
pi shows the number of lags selected for the Augmented culated by Applying Ordinary Least Square (OLS) to Eqs.
Dickey–Fuller (ADF) unit root regression and it are inde- (3) and (4).
pendently and normally distributed random variables for Cointegrated variable sets have a representation of er-
all i and t with zero means and finite heterogeneous vari- ror correction. Therefore Vector Corrected Model (VECM)
ances . IPS tests the null hypothesis of the unit root for allows to distinguish between a long-run and short-run re-
1643
© by PSP Volume 24 – No 5. 2015 Fresenius Environmental Bulletin
∆ ∆ , , ,
∆ 6
, , Where, ECT is the error correction term obtained from
the residuals of the estimated cointegrating [3]. ETC coef-
∆ 5 ficients provides information to the long-run equilibrium
levels.
TABLE 2 – Panel unit root tests for EU country energy and economic variables
1644
© by PSP Volume 24 – No 5. 2015 Fresenius Environmental Bulletin
3. RESULTS AND DISCUSSION series in levels would thus be null. But, when using the first
differences, the null of unit roots was violently rejected at
In this study, firstly, the statistical properties were the 1% significance level for whole series. Thus, it was de-
tested of the time series. We have applied the following duced that whole the series were non-stationary and inte-
unit root tests for panel data: IPS, BRT and LLC for the grated of order one.
stationarity of variables. The tests have been calculated un-
der two varied property that individual effects or individual Taking these results into account, we concluded that
effects and trends. The test results reported in Table 2.The the series were integrated of order one and proceeded to
unit root hypothesis couldnot be rejected when the varia- test for cointegration. The results of panel cointegration
bles were taken in levels and any causal deduction from the tests based on Eq. (2) are shown in Table 3. The analysis
Total Economy
Panel v 0.090 Group rho 0.252
Panel rho -1.867b Group PP -4.576c
Panel PP -6.831c Group ADF -4.816c
c
Panel ADF -8.631
Industry
Panel v -0.261 Group rho 0.463
Panel rho -1.431a Group PP -4.294c
Panel PP -6.429c Group ADF -4.784c
c
Panel ADF -7.319
Transport
Panel v 1.196 Group rho 1.291
Panel rho -1.509a Group PP -2.018b
Panel PP -3.815c Group ADF -4.447c
Panel ADF -5.464c
Services
Panel v 0.554 Group rho -0.534
Panel rho -2.538c Group PP -3.969c
Panel PP -4.814c Group ADF -4.864c
Panel ADF -5.271c
Agriculture
Panel v -0.215 Group rho 1.548
Panel rho -0.183 Group PP -2.850c
Panel PP -2.032b Group ADF -2.970c
Panel ADF -0.970
Notes: Trend assumptions, no deterministic trend.
The lag length has been chosen based on minimum AIC Lag selection: based on Modified Akaike
a,b,c
Significant at the 10%,5%,1% respectively.
Short-run Long-run
Total Economy ∆GDP ∆TOTEC ECT
∆GDP - 2.659c -4.967c
∆TOTEC 4.733c - -3.009c
Industry ∆IND ∆INDEC ECT
∆IND - 0.351 -5.368c
∆INDEC 28.236c - -1.002
Transport ∆GDP ∆TRSEC
∆GDP - 0.536 -6.764c
∆TRSEC 3.606c - -5.393c
Services ∆SER ∆SEREC ECT
∆SER - 8.275c -5.429c
∆SEREC 5.545c - 3.680
Agriculture ∆AGR ∆AGREC ECT
∆AGR - 1.943 -1.455
∆AGREC 2.443a - -3.392c
Note: The heteroschedasticity of the error terms is corrected by using white robust
standard errors both in periods. To correct for possible autocorrelation was used
the Newey-West estimator of the weighting matrix in the GMM criterion.
Note: a,b,cSignificant at the 10%,5%,1% repectively.
1645
© by PSP Volume 24 – No 5. 2015 Fresenius Environmental Bulletin
of cointegration on bivariate models for the EU supported Different from these studies, our study proved that
the being of a long-run relationship to emphasise the statis- there was a strong relationship between energy consump-
tical stability of the connection between the variables ex- tion and economic growth in the EU for both short and
amined here. long-runs. This is an important result which the EU should
We obtained Wald F-test on the significance of the co- take into consideration in the course of preparing a shared
efficients by estimated the VECM for all the sectors and all energy policy.
economy, We carried two different Granger causality rela-
tionships: a short-run causality, testing the significance of 3.2 Industry sector
the coefficients related to the lagged economic and energy We have found evidence for uni-directional long-run
variables (H0: 0 and 0 = for all i in eq. [5 and 6]),a as well as short-run causality in the industry sector. This
long-run causality related to the coefficient for the ECT term relationship is from IND to INDEC for short-run and from
(H0: = 0 for all i in eq. [5 and 6]), INDEC to IND for long-run. The change in the industry
In Table 4 and Table 5 the results of the VECM with sector of EU will effect energy consumption of the sector
two simultaneous equations of the causal relationships be- for short run. On the other hand, changes in the energy con-
tween energy consumption and economic growth for the all sumption of the sector will effect the growth of the sector
economy and the four energy sectors for the EU are re- in long run. In this case, it can be said that Union’s energy
ported. policies which will be applied in all members of the EU are
highly important for the dynamism of the industry sector.
Table 4 provides knowledge in terms of Wald F-test
on the coefficients. On the other hand in Table 5, the same
3.3 Transport sector
results are briefed to the short and long-run elasticities
when the coefficients are statistically significant. According to the empirical results, there was a uni-di-
It was found that there were different results in Table rectional causality from GDP to TRSEC for short run where
4 and 5 for each of the sectors. These results supported our as it was bi-directional between them for long run in the
research hypothesis. Each result was interpreted for each transport sector. Our results for the transport sector empha-
sector and economy level below. sizes that energy consumption is inevitable in EU. We have
noticed that Gross [8] has made a similar study for the US
3.1 Economy level and reported the same relationship.
There was a bi-directional causality relationship for 3.4 Services sector
both short and long runs between general economic growth
and total energy consumption in the EU. This result showed We have found a bi-directional causality in short term
that there was a fine relationship between economic growth and a uni-directional causality from SER to SEREC in long-
and total energy consumption in the EU. In a previous run for services sector. Results meant that the growth of ser-
study, the way and the presence of the causality for short vices sector in short term depends on energy consumption
and long runs on country basis was researched for 15 of the and the change in the energy consumption of the sector
EU member countries and different results for each country would change upon the growth speed of the sector. This
were recorded [4]. In another study [3] carried out for 12 Eu- result was also valid for the long-run. The first ranking sec-
ropean countries with panel methodology, a bidirectional but tor according to sectoral growth rates for especially both
weak causality was found between energy prices and GDP developing and developed economies is the services sector.
and electricity consumption and energy prices. Menagaki [5] In that case, the importance of energy consumption is clear
studied the causal relationship has been both studied be- for the sustainable growth of the services sector.
tween economic growth and renewable energy and also green-
3.5 Agriculture sector
house gas emissions and employment for the period between
the years 1997–2007 for 27 European countries. The empiri- The relationship between the growth in agricultural
cal results did not confirm causality between renewable en- sector and energy consumption shows a very different pat-
ergy consumption and GDP, although panel causality tests tern from other sectors in the EU, which is the fact that it
unfold short-run relationships between renewable energy is shaped by a common European policy and supported by
and greenhouse gas emissions and employment. highly important preservation programs. This relationship
1646
© by PSP Volume 24 – No 5. 2015 Fresenius Environmental Bulletin
is from AGR to AGREC for both short and long runs. In REFERENCES
other words, it means that energy consumption will grow
as long as agricultural sector grows in EU or vice versa. [1] Kraft, J. and Kraft, A. (1978) On the relationship between en-
ergy and GNP. J. Energy Development 3, 401–403.
Energy is considered as an important input for agricul-
tural production and intensively used for producing, trans- [2] Vladimer, P. and Tokmozishvili, M. (2010) Russian Energy
portation, processing and preserving agricultural products. Politics and the EU: How to Change the Paradigm. Caucasian
On the other hand, modern applications in agricultural sec- Review of International Affairs 4, 103-105.
tor increase the energy consumption. Use of agricultural [3] Ciarreta, A. and Zarraga, A. (2010) Economic growth-electric-
instruments and machinery increases the use of fossil fuels, ity consumption causality in 12 European countries: A dy-
which are the main and the most important energy sources. namic panel data approach. Energy Policy 38, 3790–3795.
One of them is, as the outcome of this paper, we eval- [9] Mehrara, M., Pahlavani, M. and Elyasi, Y. (2011) Government
uated the relationship between energy consumption and Revenue and Government Expenditure Nexus in Asian Coun-
growth in EU countries generally and sectorally and ob- tries: Panel Cointegration and Causality. International Journal
of Business and Social Science 2, 199-207.
tained different results. Secondarily, we analysed all the
data of 27 EU countries for each variable by the panel data [10] Constantini, V., Martini, C. (2010) The causality between en-
system. ergy consumption and economic growth: A multi-sectoral
analysis using non-stationary cointegrated panel data. Energy
We found a bi-directional causality relationship be- Economics 32, 591–603.
tween total energy consumption and general economic
growth in the EU. Additionally, in sectoral analysis we [11] Arellano, M. and Bond, S.R. (1991) Some tests of specifica-
found a bi-directional causality relationship between en- tion for panel data: Monte Carlo evidence and an application
ergy consumption and growth in only transport sector for to employment equations. Review of Economic Studies 58,
277–297.
long-run and services sector for short-run.
Our findings showed that the relationship between en- [12] Lee, C.C. and Chang, C.P. (2007) Energy consumption and
ergy consumption and growth was contrarious for short and GDP revisited: a panel analysis of developed and developing
countries. Energy Economics 29, 1206-1223.
long runs. The energy consumption in agricultural sector
originates from growth, according to our analyses. The re- [13] Huang, B.N., Hwang, M.J. and Yang, C.W. (2008) Causal re-
sults of this paper is important for shared policy makers lationship between energy consumption and GDP growth re-
and appliers in the EU because we identified more energy visited: a dynamic panel data approach. Ecological Economics
67, 41–54.
demanding and less energy demanding sectors through a
strong methodology for both short and long-runs. [14] Chen, S.T., Kuo, H.I. and Chen, C.C. (2007) The relationship
In conclusion, energy safety should be sustainable in between GDP and electricity consumption in 10 Asian coun-
tries. Energy Policy 35, 2611–2621.
order to achieve a sustainable growth in economy in the
EU. [15] Im, K.S., Pesaran, M.H. and Shin, Y. (2003) Testing for unit
roots in heterogeneous panels. J. Econometrics 115, 53–74.
The authors have declared no conflict of interest. [16] Levin, A., Lin, C.F. and Chu, C. (2002) Unit root tests in panel
data: asymptotic and finite-sample properties. J. Econometrics
108, 1–24.
[17] Breitung, J. (2000) The local power of some unit root tests for
panel data. Advances in Econometrics 15, 161–177.
1647
© by PSP Volume 24 – No 5. 2015 Fresenius Environmental Bulletin
CORRESPONDING AUTHOR
Gülistan Erdal
Gaziosmanpasa University
Faculty of Agriculture
Department of Agricultural Economics
Tokat
TURKEY
Phone: +90-356-2521480
Fax: +90-356-2521488
E-mail: glerdal2@yahoo.com
1648