Transcript Sources:- Researchers` Estimation using Stata 10.

ENERGY CONSUMPTION AND
ECONOMIC GROWTH: THE NIGERIAN
CASE
Ekene Stephen Aguegboh
and
Stella Ifeoma Madueme, Ph.D
6th ANNUAL NAEE/IAEE
INTERNATIONAL CONFERENCE
1
1.
INTRODUCTION
The existence of energy and its related issues in
economic literature and the challenges of
economic growth despite the importance of
energy as an input in the Nigerian growth and
development process.
 The monocultural nature of the Nigerian economy
and the fall-outs therein which include global
warming, oil spillage, gas flaring and so on.
The perspective of the resource curse
phenomenon (paradox of plenty) being associated
with the scenario of the Nigerian economy.
2
The concern about the impact of energy
consumption on the economy leading economists
like Omotor (2004), Adenikinju (2006), Omisakin
(2008) and Adeniran (2009) in recent times to
investigate the energy-growth nexus.
The fact that energy-growth nexus has not been
evaluated in line accordance with the unified
models of the mainstream economic growth theory
and the ecological growth theory.
The broad objective of the study is to ascertain the
causal relationship between energy consumption
and economic growth in Nigeria.
3
2. THEORETICAL EVIDENCE
Physical Theory of Economic Growth
The Biophysical Theory of Economic
Growth
The Classical Theory of Economic Growth
The Neutrality Hypothesis
The Ecological Economics Approach
The Mainstream Economic Theory
The Unified Model of Energy and Growth
4
3. EMPIRICAL EVIDENCE
Foreign Literatures include the works of Kraft and
Kraft (1978); Stern (2000); Yang (2000); AsafuAdjaye (2000); Aqeel and Butt (2001); Anjum and
Butt (2001); Soytas et al (2001); Glasure (2002);
Hondroyiannis et al (2002); Ghosh (2002); Soytas
and Sari (2003, 2004); and others.
Within the radar of domestic literature, the
empirical works reviwed were the works of
Adenikinju (1999); Akinlo (2008); Omisakin (2008);
Omotor (2008); Adeniran (2009); Wolde-Rufael
(2009) and Esso (2010).
5
4.
METHODOLOGY AND DATA
The methodology adopted in this research work is the
Vector-Autoregressive Model (VAR) and the Multivariate
Cointegration Analysis of Time Series.
(A) The Model:
The augmented vector autoregression (VAR) process of
order k is given as;
Yt
=
Where Yt is an L x 1 vector of innovations, and {i = 1, 2, ...,
k}. In this case, L = 4 and Yt = {RGDP, EC} where each
variable denotes real gross domestic product (RGDP) and
6
energy consumption (EC) respectively.
(B)
ESTIMATION PROCEDURE
Unit Root Test:
The procedure adopted is the Augumented-Dickey Fuller
(ADF) test due to Dickey and Fuller (1979, 1981).
(ii) Cointegration Test:
The Johansen technique was adopted because it is VARbased and because it performs better than single-equation
and alternative multivariate methods (Lutkepohl, 2001)
(iii) Causality Tests
In this research work, the researchers shall be looking at a
case of Granger-causality that entails five (5) endogenous
variables namely real gross domestic product (RGDP),
petroleum consumption (PC), gas consumption (GP), capital
formation (CF) and labour force (LF).
7
(i)
(iv) Stability Test:
The necessary and sufficient condition for stability is that all
characteristic roots lie outside the unit circle. Then is of full
rank and all variables are stationary.
(v) Impulse Response Function (IRF):
The IRF traces the response of the endogenous variables to
one-standard deviation shock to one of the disturbance term
in the system. This shock is transmitted to all of the
endogenous variables through the dynamic structure of the
VEC models (Lutkepohl, 2001).
(vi) Data Source:
Data for petroleum and gas consumption were extracted from
International Energy Agency (IEA) while the data for real gross
domestic product (RGDP), gross fixed capital formation and
labour force were extracted from Cental Bank of Nigeria (CBN)
8
Statistical bulletin.
5. EMPIRICAL RESULTS
(A) Unit Root Tests at first difference
VARIABLES
ADF (Intercept and
Order of Integration
Trend)
Ln (RGDP)
-4.739 (-3.716)*
I (0)
Ln (PC)
-6.034 (-2.625)***
I (1)
Ln (GC)
-13.000 (-2.625)***
I (1)
Ln (CF)
3.122 (-2.986)**
I (0)
Ln (LF)
-4.243(-2.625)***
I (1)
Note:- * ** and *** denotes significance at 1%, 5% and 10% level respectively. Figures
within parenthesis indicate critical values. ∆ is the first difference operator. Mackinnon
(1991) critical value for rejection of hypothesis of unit root applied.
Sources:- Researchers’ Estimation using Stata 10.
9
(B) Cointegration Test
Lag-Selection Criteria
LAG
AIC
HQIC
SBIC
0
14.74260
14.81530
14.98050
1
4.21572
4.65207
5.64308
2
3.35999*
4.15998*
5.97682
Sources:- Researchers’ Estimation using Stata 10.
10
(C) Cointegration Test
Johansen Test for Cointegration
RANK
EIGENVALUE
TEST(TRACE)
TRACE
5% CRITICAL
STATISTIC
VALUE
0
0
101.0479
68.52
1
0.75125
62.0917
47.21
2
0.65706
32.1262
29.68
3
0.51543
11.8405*
15.41
4
0.32633
0.7801
3.76
5
0.02748
0
0
* denotes acceptance of the null hypothesis at the 0.05 percent Probability level
Sources:- Researchers’ Estimation using Stata 10.
11
(D) Causality Tests
F Statistics
Lag
p-value
PC does not cause
RGDP
22.352
2
0.000
GC does not cause
RGDP
6.082
2
0.048
CF does not cause
RGDP
18.064
2
0.000
LF does not cause
RGDP
1.2077
2
0.547
RGDP does not cause
PC
0.15131
2
0.927
GC does not cause PC
0.13613
2
0.934
CF does not cause PC
0.17838
2
0.915
LF does not cause PC
1.4121
2
0.494
RGDP does not cause
GC
1.7457
2
0.418
Sources:- Researchers’ Estimation using Stata 10.
12
Causality Tests (contd.)
PC does not cause GC
9.415
2
0.009
CF does not cause GC
9.7397
2
0.008
LF does not cause GC
3.788
2
0.150
RGDP does not cause
CF
PC does not cause CF
4.4487
2
0.108
5.9244
2
0.052
GC does not cause CF
0.9908
2
0.609
LF does not cause CF
3.6423
2
0.162
RGDP does not cause
LF
PC does not cause LF
0.19113
2
0.909
0.15046
2
0.928
GC does not cause LF
2.759
2
0.252
CF does not cause LF
0.84179
2
0.656
Sources:- Researchers’ Estimation using Stata 10.
13
(E) Stability Test
-1
-.5
0
Imaginary
.5
1
Roots of the companion matrix
-1
-.5
0
Real
.5
1
Sources:- Researchers’ Estimation using Stata 10.
14
(F) Impulse Response Function
(i) Response of Real GDP to Petroleum Consumption Shocks
varbasic: dlnpc -> lnrgdp
.5
0
-.5
0
5
step
95% CI for irf
10
irf
Sources:- Researchers’ Estimation using Stata 10.
15
(ii) Response of Real GDP to Gas Consumption Shocks
varbasic: dlngc -> lnrgdp
.6
.4
.2
0
0
5
step
95% CI for irf
10
irf
Sources:- Researchers’ Estimation using Stata 10.
16
(iii) Response of Real GDP to Capital Formation Shocks
varbasic: lncf -> lnrgdp
10
0
-10
-20
0
5
step
95% CI for irf
10
irf
Sources:- Researchers’ Estimation using Stata 10.
17
(iv) Response of Real GDP to Labour Force Shocks
varbasic: dlnlf -> lnrgdp
.4
.2
0
-.2
0
5
step
95% CI for irf
10
irf
Sources:- Researchers’ Estimation using Stata 10.
18
5. CONCLUSIONS AND POLICY IMPLICATIONS
 Petroleum and gas consumption causes real GDP to
some extent in line with Soytas and Sari (2001).
 The government of Nigeria should make rigorous
effort to encourage investment in energy
generation. Thus the deregulation of the
downstream sector is a policy the right direction.
 There is need to improve on infrastructure with
particular emphasis on the current energy
infrastructure.
 Until the elementary limitations such as lack of
institutions, rules, financing mechanism etc. that are
restraining the development of energy sector, energy
supply will still persists to be a major obstacle for the
19
economic and social development in Nigeria.
THANK YOU
FOR YOUR
RAPT
ATTENTION
20