View Conference Presentation - United States Association for

Energy and Human Development: Panel Co-integration
and Causality Testing of the Energy, Electricity and
Human Development Index (HDI) Relationship
Nadia Ouedraogo
PhD student
Centre of Geopolitics of Energy and Raw Materials (GEMP)
University of Paris-Dauphine
30th USAEE/IAEE North America Conference,
OCT. 9-12, 2011
• 1. Overview
• 2. Methodology
• 3. Results
• 4. Conclusion
Overview (1)
• The increasing attention given to global energy issues and the
international policies needed to reduce greenhouse gas emissions
have given a renewed stimulus to research interest in the linkages
between the energy sector and economic performance at country
• The existence or non-existence of a long run causal relationship
between energy consumption and economic growth in these
countries should lead to the choice of an optimal energy policy for
energy poverty reduction, economic growth and climate mitigation.
• They may exist 3 types of causality:
– Unidirectional causality runs from energy consumption to growth
– Unidirectional causality runs from economic growth to energy
– Finally a bi-directional causality: energy causes economic growth and
growth leads to increase of energy consumption
1. Overview (2)
• The causal relationship between energy consumption
and income is a well-studied topic in the literature of
energy economics. The causality is in the sense of
Granger causality (Granger, 1969).
– Granger-causality implies causality in the prediction
(forecast) sense rather than in a structural sense. It
starts with the premise that ‘the future cannot cause
the past’; if event A occurs after event B, then A
cannot cause B (Granger 1969).
• The large number of studies in this area,
unfortunately, found different results for different
countries as well as for different time periods within
the same country.
1. Overview (3)
• However, very little attention in the literature
has been paid to the development indicators
other than GDP, particularly the HDI. This can
be partly explained by the difficulties in terms
of data availability. For instance, although the
HDI index was developed in 1990, the UN
undertook several major revisions of the
index, so that the data from different years are
not comparable over time and cannot be used
as a single series.
Overview (4)
• The purpose of this paper is to investigate the
relationship between economic growth, energy
use, poverty alleviation and development.
• To perform that, we are using:
– a recently developed panel unit root
– panel cointegration and panel causality techniques to
the relationship between human development index
and the total energy consumption
as well as the electricity and oil price
for the fifteen (15) Economic Community of West African
States (ECOWAs) from 1988 to 2008.
2. Methodology and Data
Panel Data
Panel unit root
Yes: Unit
No: Unit Root
1. Levin, Lin and Chu (2002)
2. Breitung (2000)
3. Im, Pesaran and Shin (2003)
4. Maddala and Wu (1999) and Choi (2001)
5. Hadri(1999).
No panel cointégration
First difference or
secondary différence
Panel cointegration Test
Pedroni (1999, 2004)
Kao (1999)
Panel cointegration modèle estimation
Pedroni (2000,2001) [FMOLS & DOLS]
Chiang and Kao (2000,2002)[DOLS]
Pesaran & alii (1999) [PMG]
Engel et Granger (1987), [VECM]
• HDIit=ait+βit+ d1iLENER+ d2iLELEC+ d3iLPX+ eit
The observable variables are in natural logarithm form, t =1,.....T is time periods;
i=1,.....N members of the panel; αi is the country-specific effects, di is the
deterministic time trends and eit is the estimated residual.
Data used in this analysis are annual time series on Human Development Index
(hereafter referred to as HDI); per capita energy consumption (referred to as ENER
hereafter) and per capita electricity consumption (referred to as ELEC) for 15
ECOWAS countries for the years 1988 to 2008. HDI data is obtained from the
United Nation Development Program (UNDP), and the energy data is obtained
from ENERDATA. International energy price in us $ /brent is from Statistical review
of World Energy 2010
All variables used, except the HDI, are in natural logarithm.
Unit root Results
Cointegration Results (1)
Cointegration Results (2)
FMOLS & DOLS Results
FMOLS and DOLS models estimation give different results. The t-statistics of
FMOLS model are systematically lower than that of the DOLS, especially when the
model is estimated without trends.
We must notice that the DOLS method has the drawback of reducing the number
of degrees of freedom by including leads and lags in the variables studied. This
reduces the robustness of estimations. As, the size of our sample is already low in
both dimensions of time and the number of countries, the DOLS results would
therefore not robust.
The DOLS estimation method, however, allows us to confirm the general trend and
direction of causality obtained by the FMOLS method.
It is interesting to note that the within-dimension results do not differ from
between- dimension results.
Modeling intra-dimension (Within) allows taking into account the heterogeneity of
individuals in their temporal dimension and / or individual. Within estimator
eliminates the individual effects (persistent differences between the countries over
the period). He favors the temporal information.
Panel Results (1)
The panel long-run income elasticity is 0.8
in energy model, which is statistically
significant at the 5% level, and the effect is
This implies that a 1 % increase in per
capita energy consumption decrease the
HDI by 0,5%, Moreover, the panel longrun energy price elasticity is -0,11 in this
model, which is statistically significant at
the 5% level, and the effect is negative.
This implies that a 1% increase in energy
price reduces the HDI by around 0.11%,
when the dependent variable is energy
In the electricity per capita model, the
panel long-run income elasticity is 0.22 %
which is statistically significant at the 5%
level, and the effect is positive. Hence,
panel long-run electricity consumption
increases the HDI by 0.22%.
Panel Results (2)
This negative impact of energy on the HDI supports our assertion. Indeed,
several hypotheses can be formulated to explain the negative impact of energy
on the HDI:
-excessive consumption of energy in unproductive sectors of the economy;
-a capacity constraint;
-inefficient supply of energy.
Regarding our sample of countries, one of the most obvious first explanations is
the inefficiency of energy supply. In fact, energy consumption in the region is
composed of 80% biomass. Outside, the use of biomass has a negative and hard
on many aspects of the HDI:
-expectancy at birth (through its negative impact on health and nutrition, for
-level of education (through such non-schooling of girls whose time is devoted to
tasks such as searching for wood but also through the reduction of study time
due to lack of lighting, the lack of access to the New Technologies etc.).
The coefficients for electricity consumption and ECT in the electricity equation
are significant at the 5% level, respectively, and the two variables are jointly
statistically significant at the 1% level. This clearly shows that there is a
unidirectional Granger Causality running from electricity consumption to HDI in
the long-run.
• The long-term results allow us to draw some conclusions:
-A Long-term increase in the quantities consumed energy is necessary for
economic growth but an improvement in the quality of this consumption is vital,
especially if we are targeting the human development of people involved (the
current structure of consumption has a negative impact on HDI and we have
demonstrated the positive impact of electricity on economic development in the
-Thus, an improvement in income, followed by non-availability of supply of
electricity, cooking gas and other forms of energy that can reduce the pressure on
biomass is not sufficient for sustainable development. Measures target the decline
in the share of biomass in energy consumption should be encouraged because the
use of this form of energy is also a real threat to the environment.
By making electricity accessible to all, this could help reducing poverty but also to
improve the quality of living.
The relatively low magnitudes of the estimated own price elasticity suggest that
the potential implementation of pricing policies to curtail energy demand may not
be useful. The small price sensitivities also indicate that little substitution between
alternative energy options is possible.

similar documents