Energy Economics 32 (2010) 524–531

Contents lists available at ScienceDirect

Energy Economics
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / e n e c o

Electricity consumption and economic growth in Burkina Faso:

A cointegration analysis
Idrissa M. Ouédraogo ⁎
Economics Department, University of Ouagadougou, 01 BP 1412 Ouagadougou, Burkina Faso

a r t i c l e i n f o a b s t r a c t

Article history: This study empirically establishes the direction of causality between electricity consumption and economic
Received 22 November 2008 growth in Burkina Faso for the period 1968–2003. The bounds test yields evidence of cointegration between
Received in revised form 20 April 2009 electricity consumption, GDP, and capital formation when electricity consumption and GDP are used as
Accepted 25 August 2009
dependent variable. Causality results indicate that there is no significant causal relationship between
Available online 1 September 2009
electricity consumption and investment. Estimates, however, detect in the long-run a bidirectional causal
JEL classification:
relationship between electricity use and real GDP. There is also evidence of a positive feedback causal
Q43 relationship between GDP and capital formation. Burkina Faso is therefore an energy dependent country. It is
C32 also a country in which electricity consumption is growing with the level of income. All of this shows that
electricity is a significant factor in socio-economic development in Burkina Faso; as such, energy policy must
Keywords: be implemented to ensure that electricity generates fewer potential negative impacts.
Electricity consumption © 2009 Elsevier B.V. All rights reserved.
Cointegration
Causality
Burkina Faso

1. Introduction
The analysis of the causal relationship between energy consump-
tion and economic activity is predominant in the literature related to
energy economics and has been subject to many empirical studies.
Interest in establishing the direction of causality between energy
consumption and economic growth was raised in the literature some
30 years ago (Kraft and Kraft, 1978; Akarca and Long, 1979, 1980).
This was inspired by the various oil crises of the seventies that
somehow slowed down economic activity all over the world. Beyond
its impact on world inflation, the increase in oil prices can have
recession effects on economies strongly dependent on energy, both in
the long-term and the short-term. These last years, because of the
events in the Persian Gulf and due to recent developments in
econometric techniques, the literature related to energy consumption
and economic growth has shown a renewal of interest (Asafu-Adjaye,
2000; Shiu and Lam, 2004; Jumbe, 2004; Paul and Bhattachargya,
2004; Oh and Lee, 2004; Lee, 2005; Narayan and Smyth, 2005;
Narayan and Singh, 2007; Akinlo, 2008; Odhiambo, 2009; Wolde-
Rufael, 2009).
Like most African countries, Burkina Faso is basically an agricul-
tural country with a relatively small industrial sector. The per capita
⁎ Tel.: +226 70211622; fax: +226 50312686.
E-mail addresses: midrissa@univ-ouaga.bf, idriss_mo@yahoo.fr.
Gross National Income (GNI) was estimated at 430 U.S. dollars in
2007. For the same year, the Gross Domestic Product (GDP) was
approximately 6.8 billion US dollars. The GDP grew annually by an
average of 5.6% between 1997 and 2007. Fig. 1 displays trends in
electricity consumption, real GDP, and Gross Capital Formation. The
three variables move in the same direction and describe a positive
trend during the period of the study. Over the last decade of the study
period, electricity consumption has outpaced real GDP, denoting the
growing demand for electricity. The Burkinabe economy is a young
and growing economy. Electric energy production is thus an essential
element to support its growth needs. The modernization of traditional
economic sectors and the continuous expansion of secondary and
tertiary sectors create new energy needs that increase the national
consumption of electricity. Electricity needs come from three main
sources: households, public administration, and the commercial and
industrial sectors. Over the period 1988–2003, the commercial and
industrial sectors accounted for 54.52% of the total electricity
consumption; that of households was 27.61%, and public administra-
tion consumed 17.37%. The consumption for the primary sector was
very low, about 0.50%.
To date, no study has been carried out on the causal relationship
between energy consumption and economic growth in Burkina Faso.
Yet, this type of analysis is necessary in the context of liberalization in
which the energy sector is involved these last years. This paper aims to
fill this gap. Using the bounds testing approach to cointegration and
the Granger causality test, this study empirically establishes the causal
relationship between electricity consumption and economic growth

0140-9883/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.eneco.2009.08.011
 I.M. Ouédraogo / Energy Economics 32 (2010) 524–531
Fig. 1. Trend in electricity consumption, GDP and Gross Capital Formation, 1968–2003.
in Burkina Faso, and further elaborates on the implications and
possible forecasts for the new direction of the national energy policy.
This paper is organized as follows. Section 2 describes the energy
sector in Burkina Faso and Section 3 provides a brief review of
literature related to energy consumption and economic growth.
Section 4 outlines the methodology used in the study. Following,
Section 5 presents the empirical results of the research and the last
section concludes the study.
2. The energy sector in Burkina Faso
Burkina Faso is a landlocked country with energy resources limited
to some hydro-electrical installations. Without any domestic oil
resources, Burkina Faso must import all of the hydrocarbons needed
to fuel the country. For its oil products, the nation strongly relies on
the refineries and oil terminals of bordering countries that have access
to the sea (Côte d'Ivoire, Ghana, Togo, and Benin). Presently, Burkina
Faso's energy needs are about 8000 barrels/day.
Hydrocarbons and electricity constitute the main sources of energy
in Burkina Faso. The other sources represent only a small part of
energy production and contribute only incidentally to the production
of goods and services. Solar energy is used in two forms: the thermal
form and the electrical form. In the direct thermal form solar energy is
used for heating and drying fruits and vegetables, among other things.
This type of use has never been quantified in Burkina Faso. On the
other hand, the amount of solar energy in its electrical form (via
conversion of solar radiation) is estimated in an indirect way based on
the amount of solar energy equipment installed in the country for
various uses (refrigeration, television, video, communication, street
lighting, etc.). In this quantification, it is assumed that the demand for
electricity is a derived demand related to the stock of electrical
components available in the country. This method has been used to
estimate households' electricity demand in Burkina Faso (Ouédraogo
and Abdou Rafa, 2005). In 2002, the total estimated power installed
for solar energy collection was 1368 kWc. In 2005, less than 20
districts of the country were supplied with such systems.
Wood products are likewise used as a source of energy in the form
of firewood and charcoal. According to statistics available, wood
production is primarily intended for domestic use as firewood (more
than 80%). The demand is particularly important in rural areas where
need exceeds 85% of the total firewood demand.
Electricity as a factor of production is very important to the
production process of goods and services in Burkina Faso. Almost all
companies resort to electrical power. An important proportion of
these industrial companies (61.4%) exclusively use the electric power
provided by The National Company of Electricity (SONABEL) for the
production of their goods and services. Only 38.6% jointly use other
sources of energy, such as power generating units or steam turbines
(Ouédraogo et al., 2005). Electricity costs account, on average, for 15%
of the total expenses of the industrial and commercial companies
(Ouédraogo et al., 2005).
525
The National Company of Electricity is a State owned company
with a capital of 46 billion CFA francs, and is in charge of electricity
supply and distribution in Burkina Faso. It has a monopoly over the
production, import, transport, distribution, and marketing of electric-
ity in the country. Its total production in 2007 was 612,711,927 KWH.
For the production of electricity, SONABEL has 28 thermal power
stations and 4 hydroelectric power stations. However, it should be
noted that this production is insufficient to meet the country's needs
for electricity. SONABEL is thus obliged to import electricity from Côte
d'Ivoire and Ghana, and in 2007 SONABEL imported an estimated
123,910,359 KWH. In the same year, SONABEL provided 68 localities
of Burkina with electricity; its distribution network, about 7521 km
long, made it possible to provide 288,475 users with electricity.
As a result of the reforms undertaken within the framework of
economic liberalization, the government of Burkina Faso initiated a
profound reorganization of the electricity sector in December 1998 by
passing a law on the general regulation of the electric power supply in
Burkina Faso. In a specific way, this law suggests the repealing of the
monopoly on electricity production by SONABEL: it authorizes the
distribution of electricity in areas where no distribution company is
settled and it creates a fund for the development of electrification that
will be supplied by additional taxes on each kWh sold. A change in the
institutional status of SONABEL is planned for the years to come by the
creation of two companies: a state owned company that will be the
owner of assets and in charge of rural electrification, and a private
company that will be in charge of production and distribution of
electricity in urban areas.
These various measures imply for SONABEL a certain number of
changes at the level of production, management, and result in terms
of financial equilibrium. In this context, SONABEL carried out a series
of studies that led to the establishment of an optimum plan of
investment totaling 240 billions CFA francs for the period 2003–
2020. To carry out these investments, SONABEL will have to
significantly mobilize its own resources. With this purpose, it will
have to develop a self-financing capacity. To this end, SONABEL led
two studies, one on electricity price and the other on a prospective
financial analysis of the company. These two studies showed the need
for SONABEL to revise its tariffs in order to restore its financial balance.
This step will help the company undertake necessary investments for
its development.
If financial viability proves to be an imperative for SONABEL's
durability, its social and economic mission should not be ignored.
Indeed, it should be noted that electric power is today, not only a
necessary commodity for the Burkinabe population but also, a
primary production factor for many manufacturing units of the
country. Although electricity constitutes one of the main sources of
energy in Burkina Faso, only 13% of the households have access to it
and in rural areas only 1.1% of households use electricity as a source of
energy.
In September 2006, SONABEL decided to readjust its rates, a
change that became effective in October 2006. This readjustment was
the second in 2 years, as rate adjustments were first introduced in
October 2004. The increases of 2006 vary between 2.7% and 15.4% for
non-industrial consumption. For industrialists, the increase rate was
between 5.9% and 7.7%. In addition, there was an increase of 20% in the
service charge. According to SONABEL, the low rates applied to
industries were justified by the fact that the company did not want to
generate dysfunctions in the productive sectors. This increase was
further justified by the fact that 75% of managers of company supplied
by SONABEL say that the previous rise in the price of electricity
negatively influenced their sales. And since the majority of these
companies operate in a competitive environment, a high increase in
electricity prices can affect their competitiveness (Ouédraogo et al.,
2005). Electricity prices in Burkina Faso are now among the highest in
West Africa. From a policy point of view, the results of this study will
be of major importance for policymakers, in that it will provide them
526 I.M. Ouédraogo / Energy Economics 32 (2010) 524–531
with rational, empirical evidence that will help in the electricity sector
restructuring process. Carrying out reform in this sector without
having knowledge of the type of causal relationship that exists
between electricity consumption and economic growth could result in
ineffective and wasteful changes.
3. A brief review of related literature
Even though the relation between electricity consumption and
economic growth has been the subject of intense studies, economists
are still unable to agree on the type of causal relationship existing
between energy consumption and economic growth (see Table 1).
Though the existence of causal relations between the various
variables has not yet been subject to debate, the direction of causality
between energy consumption and economic variables as well as
statistical methods used to establish this relationship remain
controversial. For example, in the case of the United States, while
Kraft and Kraft (1978) argue that the direction of causality runs from
GNP to energy consumption, Akarca and Long (1980) found a one-
way causality running from energy consumption to employment, and
Yu and Hwang (1984) and Soytas and Sari (2003) found no significant
causal relationship between energy consumption and income.
Further, using a multivariate framework Stern (2000) and Lee
(2006) found a bidirectional relationship between energy and GDP.
Similarly, for Canada, while Soytas and Sari (2003) establish that GDP
and energy consumption are independent, Lee (2006) found evidence
for unidirectional causality running from energy to income.
Similar conflicting results where found for European countries. For
instance, using a multivariate framework and an error correction
model to test the causality relationship between energy consumption,
GDP, and price in Greece, Hondroyiannis et al. (2002) found a
bidirectional relationship among these variables. Soytas and Sari
(2003) found a unidirectional causality running from energy to GDP in
France, Germany, and Turkey, but from GDP to energy in Italy. They
found no relationship between the two in Poland and the U.K. Lee
(2006) found a unidirectional causality running from energy
consumption to GDP in Belgium, the Netherlands, and Switzerland,
and a reversed unidirectional causality in France and Italy. He found a
bidirectional relationship in Sweden and no relationship at all in the
U.K. and Germany.
In Latin America, causality has been evidenced to run from GDP to
energy for Argentina (Soytas and Saris, 2003) and from energy to GDP
for Brazil (Cheng, 1997), whereas the neutrality hypothesis is
validated for Mexico (Soytas and Saris, 2003; Cheng, 1997), Venezuela
(Cheng, 1997), and Brazil (Soytas and Saris, 2003). In the case of
Venezuela, Squalli (2007) reports evidence of positive causality
running from electricity to GDP. In other studies, Apergis and Payne
(2009) used a panel cointegration and error correction model to
investigate, within a multivariate framework, the relationship
between energy consumption and economic growth for Costa Rica,
El Salvador, Guatemala, Honduras, Nicaragua, and Panama. They
discovered the existence of cointegration between real GDP, energy
consumption, the labor force, and real gross fixed capital formation,
and also the presence of both short-run and long-run Granger
causality from energy consumption to economic growth.
In Asia, Masih and Masih (1996) found cointegration between
energy and GDP in India, Pakistan, and Indonesia, but no cointegration
in Malaysia, Singapore, and the Philippines. For the latter three
countries, the neutrality hypothesis was supported. Granger causality
was shown to exist from energy to GDP in India but in the opposite
direction in Indonesia. Pakistan shows a bidirectional relationship.
While Cheng and Lai (1997), using the Granger Causality method in an
analysis of Taiwan found a unidirectional causality running from GDP
to energy consumption, Yang (2000) found, for the same country and
using the same method, a bidirectional causal relationship between
energy consumption and GNP. Using cointegration and error
correction techniques, a further project by Asafu-Adjaye (2000)
examines the relationship between energy consumption, energy
prices, and income in four Asian countries. He found a unidirectional
Granger causality running from energy to income for India and
Indonesia, and bidirectional Granger causality for Thailand and the
Philippines. This result is consistent with Squalli (2007) for Indonesia.
Additionally, Soytas and Sari (2003) found a unidirectional causality
running from energy to GDP in Japan and from GDP to energy in
Korea. They found no relationship in India and Indonesia.
Despite the growing literature on the study of causality between
energy consumption and economic growth, there are very few
individual causality studies concerning countries in the Middle East,
apart from Squalli (2007) and Zamani (2007). Using the bounds test
procedure and the non-causality test developed by Toda and Yamamoto
to investigate the long-run relationship between electricity consump-
tion and economic growth for the OPEC members, Squalli (2007) found
conflicting results for countries in the Middle East. Positive causality
running from GDP to electricity was evidenced for Kuwait and negative
causality from GDP to electricity was detected for Iraq. In Iran, Qatar, and
Saudi Arabia he found bidirectional causality with negative causality
running from electricity to GDP for Saudi Arabia. He also evidenced
negative causality from electricity to GDP for UAE. Using a vector error
correction model to investigate the relationship between overall GDP,
industrial and agricultural value added, and consumption of different
kinds of energy in Iran, Zamani (2007) found a long-run unidirectional
relationship from GDP to total energy and a bidirectional relationship
between both GDP and gas consumption and GDP and petroleum
products consumption for the whole economy. At the desegregated
level, he discovered that causality is running from value added to total
energy, electricity, gas, and petroleum products consumption, and from
gas consumption to value added in the industrial sector. He also found
that there are long-run bidirectional relations between value added and
total energy, electricity, and petroleum products consumption in the
agricultural sector.
In the case of African countries, Ebohon (1996) used the Granger
causality test to examine the causal directions between energy
consumption and economic growth for Nigeria and Tanzania. He
found a simultaneous causal relationship between energy and
economic growth for both countries. Jumbe (2004), applying coin-
tegration and error correction vector to detect causality between
electricity consumption and economic activity in Malawi, found that
there is bidirectional causality between electricity consumption and
economic activity. Wolde-Rufael (2005) used the Bounds test
approach to cointegration and the Toda–Yamamoto Statistical
inference in vector autoregressions to examine the causal relationship
between energy consumption and economic growth for nineteen
African countries. He found evidence of a unidirectional causality
running from economic growth to energy use for Algeria, the
Democratic Republic of Congo, Egypt, Ghana, and the Ivory Coast.
He reported a positive causality running from energy use to economic
growth for Cameroon, and a negative causality running from energy
consumption to economic growth for Morocco and Nigeria. Bidirec-
tional causality was detected for Gabon and Zambia and no causal
relationship was found for Benin, the Congo Republic, Kenya, Senegal,
South Africa, Sudan, Togo, Tunisia, and Zimbabwe.
Additionally, in a recent study Wolde-Rufael (2009) re-examined
the causal relationship between energy consumption and economic
growth for the same African countries in a multivariate framework
with labor and capital as additional variables. The introduction of
these variables changed the results. For example, in seven of the
countries where he did not find evidence of causality in any direction
in the previous study, he discovered evidence of Granger causality. In
Benin and South Africa, he found unidirectional causality running
from energy consumption to economic growth; in Senegal, Sudan, and
Tunisia causality was found to run from economic growth to energy
consumption; and in Togo and Zimbabwe bidirectional causality was
 I.M. Ouédraogo / Energy Economics 32 (2010) 524–531 527

Table 1
An overview of selected studies.

Authors Countries Empirical method Results

North America and Europe

Kraft and Kraft (1978) U.S. Sims causality test GDP→EC
Akarca and Long (1980) U.S. Granger test EC→EM
Yu and Hwang (1984) U.S. Granger test Neutral
Stern (2000) U.S. VAR EC↔GDP
Hondroyiannis et al. (2002) Greece Error-correction model GDP↔EC
Soytas and Sari (2003) Italy Error-correction model GDP→EC
France, Germany, Turkey EC→GDP
Poland, U.S., U.K., Canada Neutral
Altinay and Karagol (2005) Turkey VAR/Granger causality test EC→GDP
Lee (2006) U.K., Germany Toda–Yamamoto Neutral
Sweden, U.S. GDP↔EC
Canada, Belgium, EC→GDP
Netherlands, Switzerland
France, Italy EC→GDP
GDP→EC.
Zachariadis and Pashourtidou (2007) Cyprus GDP↔EC

South and Central America

Cheng (1997) Brazil Granger causality test EC→GDP
Mexico, Venezuela Neutral
Soytas and Sari (2003) Argentina ECM GDP→EC
Brazil, Mexico Neutral
Squalli (2007) Venezuela Bounds test/Toda Yamamoto EC→GDP
Apergis and Payne (2009) Costa Rica, El Salvador, Panel Vector error correction model EC→GDP
Guatemala, Honduras,
Nicaragua, Panama

Asia and Oceania

Masih and Masih (1996) India Error-correction model EC→GDP
Indonesia GDP→EC
Pakistan GDP↔EC
Malaysia, Singapore, Neutral
Philippines
Cheng and Lai (1997) Taiwan Granger causality test GDP→EC
Yu and Hwang (1984) Philippines Granger test EC→GDP
Soytas and Sari (2003) Korea Error-correction model GDP→EC
Japan EC→GDP
India, Indonesia Neutral
Yang (2000) Taiwan Granger causality test GDP↔EC
Asafu-Adjaye (2000) India, Indonesia Error-correction model EC→GDP
Philippines, Thailand GDP↔EC
Shiu and Lam (2004) China EC→GDP
Narayan and Smyth (2005) Australia Multivariate Granger causality test GDP→EC
GDP→EM
Lee (2006) Japan GDP→EC
Narayan and Singh (2007) Fiji Island Multivariate Granger causality test EC→GDP
GDP↔LF
Squalli (2007) Indonesia Bounds test/Toda–Yamamoto EC→GDP

Middle East
Squalli (2007) Iran, Qatar, Saudi Arabia Bounds test/Toda–Yamamoto GDP↔EC
Iraq, Kuwait, GDP→EC
UAE EC→GDP
Zamani (2007) Iran Vector error correction model GDP→EC

Africa
Ebohon (1996) Nigeria, Tanzania Granger causality test GDP↔EC
Soytas and Sari (2003) South Africa ECM Neutral
Jumbe (2004) Malawi GDP↔EC
Wolde-Rufael (2005) Algeria, Congo DR, Egypt, Bounds test/Toda–Yamamoto GDP→EC
Ghana, Ivory Coast
Cameroon, Morocco, Nigeria GDP→EC
Gabon, Zambia EC→GDP
Benin, Congo Rep, Kenya, Senegal, GDP↔EC
South Africa, Sudan, Togo, Neutral
Tunisia, Zimbabwe Neutral
Neutral
Wolde-Rufael (2006) Cameroon, Ghana, Nigeria, Senegal, Bounds test GDP→EC
Zambia, Zimbabwe
Benin, Congo, DR, Tunisia GDP→EC
Egypt, Gabon, Morocco EC→GDP
Algeria, Congo Republic, Kenya, GDP↔EC
South Africa, Sudan Neutral
Neutral

(continued on next page)

528 I.M. Ouédraogo / Energy Economics 32 (2010) 524–531
Table 1 (continued)
Authors Countries
Africa
Squalli (2007) Algeria, Libya
Nigeria,
Akinlo (2008) Gambia, Ghana, Senegal
Sudan, Zimbabwe, Congo
Cameroon, Cote d'Ivoire,
Nigeria, Kenya, Togo
Odhiambo (2009) Tanzania
Wolde-Rufael (2009) Algeria, Benin, South Africa
Egypt, Ivory Coast,
Morocco, Nigeria, Senegal,
Sudan, Tunisia, Zambia,
Gabon, Ghana, Togo, Zimbabwe
Cameroon, Kenya
Notes: EC→GDP denotes causality runs from energy (or electricity) consumption to GDP; GDP→EC denotes causality runs from GDP to energy (or electricity) consumption;
GDP↔EC denotes bidirectional causality between GDP and energy (or electricity) consumption; EC→EM/EC→LF denotes causality runs from energy (or electricity) to Employment
or Labor force.
found. Moreover, the direction of causality was reversed in Algeria,
Morocco, and Nigeria. In Ghana, causality was reversed from
economic growth to energy consumption to bidirectional causality.
In Zambia, the bidirectional casualty was reversed to unidirectional
causality running from economic growth to energy consumption. In
Cameroon, causality was reversed from energy consumption to
economic growth to no causality. The findings from the original
study were confirmed only in four countries: Egypt, the Ivory Coast,
Gabon, and Kenya.
Analyzing electricity consumption and economic growth for OPEC
countries, Squalli (2007) found positive causality from electricity to
GDP for Nigeria and negative causality from GDP to electricity for
Algeria and Libya. Akinlo (2008) used the autoregressive distributed
lag (ARDL) bounds test to investigate the relationship between energy
consumption and economic growth for eleven countries in sub-
Saharan Africa. His findings showed a two-way causal relationship
between energy consumption and economic growth for Gambia,
Ghana, and Senegal. At the same time, he found unidirectional
causality running from economic growth to energy consumption in
Congo, Sudan, and Zimbabwe. For Cameroon, Cote d'Ivoire, Nigeria,
Kenya, and Togo no causal relationship was found. Similarly,
Odhiambo (2009) using the ADRL bounds testing approach for
Tanzania, found that there is a unidirectional causal flow from total
energy consumption to economic growth and a prima-facie causal
flow from electricity consumption to economic growth.
Besides analyses related to the total consumption of energy, the
causal relationship between electricity consumption and growth were
also addressed (Jumbe, 2004; Shiu and Lam, 2004; Narayan and
Smyth, 2005; Narayan and Singh, 2007; Zachariadis and Pashourtidou,
2007; Squalli, 2007). Here again, empirical results are mixed. As stated
above, Jumbe (2004) found evidence of bidirectional causality
between electricity consumption and economic activity in Malawi.
Shiu and Lam (2004) established, in the case of China, the existence of
a one-way causal relationship running from electricity consumption
to economic growth. Using multivariate Granger causality tests for
Australia, Narayan and Smyth (2005) found a unidirectional causality
running from GDP to electricity consumption and also from GDP to
employment. Narayan and Singh (2007), with the same method for
Fiji, concluded that there is a unidirectional causality running from
energy consumption to GDP in the long-run, and a bidirectional
causality running from GDP to labor force in the short-run. Altinay and
Karagol (2005) investigated the causal relationship between electric-
ity consumption and real GDP for Turkey and found evidence for
unidirectional causality running from electricity consumption to
income. Zachariadis and Pashourtidou (2007) analyzed electricity
consumption in Cyprus and found bidirectional causality between
residential electricity consumption and private income. Squalli (2007)
Empirical method Results
Bounds test/Toda Yamamoto GDP→EC
EC→GDP
Bounds test/VAR GDP↔EC
GDP→EC
Neutral
Neutral
Bounds test EC→GDP
Toda–Yamamoto EC→GDP
GDP→EC
GDP→EC
GDP→EC
GDP↔EC
Neutral
used the bounds test procedure to investigate the relationship
between electricity consumption and economic growth for OPEC
member countries and had mixed results.
Wolde-Rufael (2006) used the bounds testing approach to
cointegration to investigate the long-run causal relationship between
electricity consumption and economic growth for 17 African countries
for the period of 1971–2001. He found evidence of a long-run
relationship when real GDP per capita was used as the dependent
variable for five countries, and for four other countries when
electricity consumption per capita was used as the dependent
variable. The empirical evidence established that for Cameroon,
Ghana, Nigeria, Senegal, Zambia, and Zimbabwe there was a positive
unidirectional causality running from economic growth to electricity
consumption. For Benin, the Democratic Republic of the Congo, and
Tunisia the results indicate unidirectional causality running from
electricity consumption to economic growth. Finally, there was
bidirectional causality between the two variables in the case of
Egypt, Gabon, and Morocco. For the remaining countries (Algeria,
Congo Republic, Kenya, South Africa, and Sudan), no causal relation-
ship was found.
It is obvious that the results of the above-cited studies do not make
it possible to conclude definitely the direction of causality between
energy consumption and the variables of the economic activity. Yet, it
is known that this causality is of major importance for effective energy
policy design and implementation (Asafu-Adjaye, 2000; Jumbe, 2004;
Paul and Bhattachargya, 2004; Narayan and Smyth, 2005). A country
that is energy dependent, i.e. a country in which the direction of
causality runs from energy consumption to GDP, will have a cautious
energy policy because any negative shock on energy supply will have
certain negative effects on economic activity. On the other hand, in an
economy where energy consumption is determined by economic
activity, an energy conservation policy will have very little affect on
economic growth (Asafu-Adjaye, 2000; Jumbe, 2004; Paul and
Bhattachargya, 2004; Narayan and Singh, 2007). There are also
cases in which it is difficult to establish a causal relationship between
energy consumption and economic activity. It is the neutrality
hypothesis according to which energy-saving-policy has no effect
on the economic activity (Yu and Choi, 1985; Asafu-Adjaye, 2000).
4. Methodology
4.1. Variables specification and data sources
In recent studies, a multivariate approach, rather than a bivariate
approach, is used to investigate the causal relationship between
energy consumption and economic growth. It is argued that in a
bivariate framework, omitted variables could lead to erroneous
 I.M. Ouédraogo / Energy Economics 32 (2010) 524–531 529

conclusions. Energy, GDP, labor, and/or capital are the variables that Table 2
are generally included in a multivariate model. In fact, energy is Summary of unit root tests.

considered as a variable of a production function. Energy is expected Variables ADF test P–P test
to play a significant role in achieving economic progress as a (trend and constant) (trend and constant)
complement or a substitute to other factors of production. Knowledge Level First difference Level First difference
of the exact relationship between energy and the other factors of
lnELEC − 3137 − 2546 − 2341 − 4132⁎⁎
production is of major importance for energy policy formulation lnGDP − 2611 − 7896⁎⁎⁎ − 2394 − 8678⁎⁎⁎
(Ebohon, 1996). There is evidence that energy and labor are lnINVES − 2546 − 5487⁎⁎⁎ − 2546 − 4567⁎⁎⁎
substitutable, as are capital and labor. However, the relationship
between energy and capital is subject to debate. But, in the case of Critical values
1% critical value⁎⁎⁎ − 4.2242
developing countries, because of their special economic features, one 5% critical value⁎⁎ − 3.5348
could argue that energy complements capital in these countries 10% critical value⁎ − 3.1988
(Ebohon, 1996). Wolde-Rufael (2009) found that in eleven of the
⁎⁎⁎, ⁎⁎ and ⁎denote significance level at 1%, 5%, and 10% respectively.
seventeen countries he had investigated, energy is no more than a
contributing factor to output growth, and not an important factor
when compared to capital and labor. The Granger causality test is then implemented based on the results
Recognizing the importance of these issues, and following Narayan of the above tests. There will be evidence of long-run relationships
and Smyth (2008), Lee et al. (2008), this study examines the among the variables if, in the equations given above, H0 is rejected. In
relationship between electrity consumption and economic growth that case, the Granger causality test should be augmented with a one
within a production function approach by incorporating a capital stock period lagged error correction term (ECTt-1). In the case of no evidence
variable. Because of data availability and consistency, labor is not of a long-run relationship among the variables, the traditional Granger
included in the model. For the key variables, i.e., those measuring the causality test will be conducted, i.e., a VAR in first difference form.
economic activity and energy consumption, we take as proxies the real
GDP and total consumption of electricity (ELEC), respectively. To these, 5. Empirical results
we add Gross Capital Formation (INVEST) as a proxy of capital stock.
The use of gross capital formation to proxy the stock of physical 5.1. Unit root test
capital follows the work of Soytas and Sari (2006), Narayan and Smyth
(2008), Apergis and Payne (2009), and Wolde-Rufael (2009). It is Before performing the bound test, it is essential to check for the
argued that since in the perpetual inventory method the rate of stationarity of the data series used. This is important to obtaining an
depreciation is assumed to be constant, changes in investment are unbiased estimation from the Granger causality tests, and also because
closely related to changes in capital stock. Thus, capital formation can the bound test is used only when variables are I(0) or I(1). The
be used as a reliable proxy for changes in capital stock.1 Augmented Dickey–Fuller (ADF) test and the Philipps–Perron test were
The data used in this study are annual data; they cover the period applied to test for the existence of unit roots. In addition, p, the optimal
from 1968 to 2003. Real GDP and Gross Capital Formation are in Local lag length, is determined by Aikaike or Schwarz information criterion
Currency Unit and are from the Africa Development Indicators (2007) (see Perasan and Perasan, 1997). The optimal lag length is the one that
of the World Bank. The yearly consumptions of electricity measured in minimizes the two criteria AIC and SBC. The estimations give p = 2 for
GWh are drawn from the Statistics of the National Company of electricity consumption and lead to a simple DF test for the other two
electricity production. For the estimates, the natural logarithm of series. After the optimal lag is known for each series, estimates are
these variables is used. performed. The results indicate that the three series are I(0).
However, the ADF test is often considered a non robust test. To
4.2. The ARDL bound test procedure solve this problem, the Phillips–Perron test is applied. The results
prove the presence of a unit root in each of the three series. These
The Autoregressive Distributed Lag (ARDL) bounds testing results suggest that the stationarity of first differences of the three
approach to cointegration proposed by Pesaran et al. (2001) is used series should be tested. All of the results are presented in Table 2.
to analyse long-run relationships among the variables of interest. This Considering the results of the P–P test, each of the three series is I(1).
procedure is adopted because it has better small sample properties
than alternative methods. Moreover, it can be used irrespective of the 5.2. Cointegration test
order of integration of the regressors. The bounds testing procedure
consists of estimating an unrestricted error correction model with the As shown in the methodology above, the bound test for cointegra-
following generic form in which each variable comes in turn as a tion consists of testing hypothesis H0 for Eq. (1) with each variable
dependent variable: considered in turn as the dependent variable. Estimates must be made
after selecting the adequate model and the optimal lag length. For the
choice of model, five possibilities are given (Pesaran et al. 2001): the
X
P X
p X
p
Δ ln GDP = α 0G + biG Δ ln GDPt − i + ciG Δ ln ELECTt − i + diG Δ ln INVESt − i model with no intercept and no trend (case I); the restricted intercept
i=1 i=1 i=1
and no trend model (case II), the unrestricted intercept and no trend
+ γ1 ln GDPt − 1 + γ2 ln ELECt − 1 + γ3 ln INVESt − 1 + e1t ð1Þ

Table 3
To test for the existence of a long-run relationship among the Results of the bound test.
variables, an F-test for the joint significance of the lagged level variables
F-statistics (k = 2 and T = 36) CVs at 1% CVs at 5%
coefficients will be conducted. The null hypothesis of no cointegration
among the variables, H0, against the alternative, H1, for each equation is I(0) I(1) I(0) I(1)

as follows: H0: γ1 =γ2 =γ3 = 0 and H1: γ1 ≠γ2 ≠γ3 ≠ 0. 4.448 6.028 3.478 4.335
FELEC (ELEC\GDP, INVES) = 4.9457
1 FGDP (GDP\ELEC, INVES) = 6.5568
It is worth noting that the use of gross fixed capital formation as a proxy for capital
FINVES (INVES\GDP, ELEC) = 2.6290
stock is questioned by Lee et al. (2008). It is stated that because gross fixed capital
formation is a flow variable it does not accurately account for capital accumulation. Notes: The small sample CVs for bounds test are derived from Narayan (2005).
530 I.M. Ouédraogo / Energy Economics 32 (2010) 524–531
Table 4
Short-run and long-run relationships.
Short-run
Dependent variables ECTt-1 Δln ELEC Δln GDP
lnELEC − 0.31⁎⁎⁎ (− 3.32) − 0.032⁎⁎ (0.146)
lnGDP − 0.421⁎⁎ (− 2.561) 0.129⁎ (1.764) − 0.276⁎⁎⁎ (3.295)
lnINVES − − 0151 (− 1.178 0.779⁎⁎⁎ (2.983)
⁎⁎⁎, ⁎⁎, ⁎denote significance level at 1%, 5%, and 10% respectively.
model (case III); the unrestricted intercept and no trend model (case
IV); and the unrestricted intercept and unrestricted trend model (V). For
this study, provided that the coefficients of the trends are not significant,
the model with restricted intercept and no trend is retained. For the
choice of lag length, the usual criteria (AIC, HQIC, and SBIC) are used.
Using the last two criteria, we find that p = 1 is the optimal lag.
Results of the bounds test are reported in Table 3. The data show that
γi are significantly non null at the level of 5% when electricity
consumption and real GDP are used as the dependent variable. This
implies that there is a cointegrating relationship in these cases. Indeed,
in these two cases the calculated value of the F statistic is located above
the upper limit at the level of 5%. When capital formation is used as the
dependent variable, the calculated F statistic is below its theoretical
value. The existence of a cointegrating relationship between electricity
consumption, GDP, and investment suggests the existence of a Granger
causality relationship. But the existence of such a relationship does not
indicate the direction of the causality between the variables. For that
reason, a causality test is carried out.
5.3. Granger causality test
Given the fact that the F-test highlighted the existence of a
relationship when lnELEC and lnGDP are considered as the dependent
variable in the ARDL, causality test is conducted by estimating a Vector
Error Correction Model (VECM) within the ARDL framework. The
results presented in Table 4 show that the coefficients of adjustment
to equilibrium of real GDP and electricity consumption are significant
at the 5% level and have the expected sign.
The results also show that in both the short-run and the long-run
there is a positive causality running from electricity consumption to
economic growth. In the same way, there is evidence to support the
growth-led hypothesis. In other words, in Burkina Faso there is a
bidirectional relationship between electricity consumption and real
GDP, which means that energy use and economic growth complement
each other. Electricity consumption acts as a stimulus to economic
growth and a high level of economic growth demands a high level of
electricity use. The results also show that in the short-run economic
growth and capital formation mutually influence each other, which
suggests that a high level of investment leads to a high level of
economic growth and vice versa. There is no significant causal
relationship between electricity consumption and capital formation.
These findings indicate that in Burkina Faso energy policies aimed at
improving the energy infrastructure and increasing the energy supply
are the appropriate options. Energy conservation policies could
hamper social and economic progress.
6. Conclusion and policy implication
This study aimed at determining the causal relationship between
electricity consumption and economic growth in Burkina Faso. The
bound testing approach to cointegration proposed by Pesaran et al.
(2001) detected a cointegrating relationship between the series studied
when electricity consumption and GDP were considered as the
dependent variable. The results of Granger test showed that there is
no causal relationship between electricity consumption and investment.
Long-run
Δln INVES lnELEC lnGDP lnINVES
− 0.031 (− 0.364) − 2095⁎⁎⁎ (5.096) − 0.099 (− 0.358)
0.306⁎⁎⁎ (2.788) − 0.240 (1.502)
− − −
However, these results suggest that in the short-run and the long-run
there is a bidirectional relationship between electricity consumption
and real GDP. There is also evidence of a positive feedback relationship
between GDP and capital formation. These results suggest that Burkina
Faso is a country dependent on electric energy, and that it is also a
country where electricity consumption grows with the income level.
From a policy point of view, the results highlight the importance of
energy policy on economic growth, and household welfare particu-
larly, in the context of poverty alleviation. Given the fact that the
majority of the people of Burkina Faso are denied access to the use of
electricity combined with the obsolescence of the energy infrastruc-
ture, investments and measures meant to increase electricity supply
should be implemented. Thus, the current energy policy and the
electricity sector restructuring process should be designed to meet
this goal. The Government and SONABEL must implement policies in
such a way that electricity should not be an obstacle to development.
Relying mainly on imported electricity and fuels to meet the needs of
the country generally leads to electricity rationing. This type of energy
conservation policy can have a negative impact on economic activity
considering the high proportion of companies relying exclusively on
this source of energy. Developing other sources of energy such as solar
energy and bio-energy, could help in attaining this goal.
Acknowledgements
I am very grateful to Diarra Mahamadou for his invaluable contribution
to this paper. I am also thankful to Emeline Sanou Ouattara for her help in
the translation. Finally, I would like to express my sincere gratitude to the
Editor of the Journal and the two anonymous referees whose constructive
comments and suggestions have improved the paper tremendously.
However, any errors and/or omissions are my own.
References
Altinay, G., Karagol, E., 2005. Electricity consumption and growth: evidence from
Turkey. Energy Economics 27, 849–856.
Akarca, A.T., Long, T.V., 1979. Energy and employment: a time series analysis of causal
relationship. Resource Energy 2, 151–162.
Akarca, A.T., Long, T.V., 1980. On the relationship between energy and GNP: a re-
examination. Journal of Energy Development 5, 326–331.
Akinlo, A.E., 2008. Energy consumption and economic growth: evidence from 11 Sub-
Sahara African countries. Energy Economics 30, 2391–2400.
Apergis, N., Payne, J.E., 2009. Energy consumption and economic growth in Central
America: evidence from a panel cointegration and error correction model. Energy
Economics 31, 211–216.
Asafu-Adjaye, J., 2000. The relationship between energy consumption, energy prices and
economic growth: time series evidence from Asian developing countries. Energy
Economics 22, 615–625.
Cheng, B.S., 1997. Energy consumption and economic growth in Brazil, Mexico and
Venezuela: a time series analysis. Applied Economics Letters 4, 671–674.
Cheng, B.S., Lai, T.W., 1997. An investigation of cointegration and causality between energy
consumption and economic activity in Taiwan. Energy Economics 19, 435–444.
Ebohon, O.J., 1996. Energy, economic growth and causality in developing countries: a
case study of Tanzania and Nigeria. Energy Policy 24, 447–453.
Hondroyiannis, G., Lolos, S., Papapetrou, G., 2002. Energy consumption and economic
growth: assessing the evidence from Greece. Energy Economics 24, 319–336.
Jumbe, C.B.L., 2004. Cointegration and causality between electricity consumption and
GDP: empirical evidence from Malawi. Energy Economics 26, 61–68.
Kraft, J., Kraft, A., 1978. On the relationship between energy and GNP. Journal of Energy
and Development 3, 401–403.
Lee, C., 2005. Energy consumption and GDP in developing countries: a cointegrated
panel analysis. Energy Economics 27, 415–427.
 I.M. Ouédraogo / Energy Economics 32 (2010) 524–531
Lee, C., 2006. The causality relationship between energy consumption and GDP in G-11
countries revisited. Energy Policy 34, 1086–1093.
Lee, C., Chang, C., Chen, P., 2008. Energy-income causality in OECD countries revisited:
The key role of capital stock. Energy Economics 30, 2359–2373.
Masih, A.M.M., Masih, R., 1996. Energy consumption, real income and temporal
causality: results from a multi-country study based on cointegration and error-
correction modelling techniques. Energy Economics 18, 165–183.
Narayan, P.K., 2005. The saving and investment nexus for China: evidence from co-
integration tests. Applied Economics 37, 1979–1990.
Narayan, P.K., Singh, B., 2007. The electricity consumption and GDP nexus for the Fiji
Islands. Energy Economics 29, 1141–1150.
Narayan, P.K., Smyth, R., 2005. Electricity consumption, employment and real income in
Australia evidence from multivariate Granger causality tests. Energy Policy 33,
1109–1116.
Narayan, P.K., Smyth, R., 2008. Energy consumption and real GDP in G7 countries: new
evidence from panel cointégration with structural breaks. Energy Economics 30,
2331–2341.
Odhiambo, N.M., 2009. Energy consumption and economic growth nexus in Tanzania:
an ARDL bounds testing approach. Energy Policy 37, 617–622.
Oh, W., Lee, K., 2004. Causal relationship between energy consumption and GDP
revisited: the case of Korea 1970–1999. Energy Economics 26, 51–59.
Ouédraogo, I.M., Abdou Rafa, M.L., 2005. La demande d'électricité des ménages au
Burkina Faso, Revue CEDRES-ETUDE, Nouvelle série 48, 1–21.
Ouédraogo, I.M., Ouédraogo, G.J, Ouédraogo, J.P., Abdou Rafa, M.L., 2005. Etude des impacts
socio-économiques d'une hausse des tarifs du kWh de la SONABEL. Rapport définitif,
SONABEL.
Paul, S., Bhattachargya, R.N., 2004. Causality between energy consumption and economic
growth in India: a note on conflicting results. Energy Economics 26, 977–983.
Pesaran, M.H., Pesaran, B., 1997. Working with Microfit 4.0: Interactive Econometric
analysis. Oxford University Press, Oxford.
531
Pesaran, M.H., Shin, Y.C., Smith, R., 2001. Bound testing approaches to the analysis of
level relationships. Journal of Applied Econometrics 16, 289–326.
Shiu, A., Lam, L.P., 2004. Electricity consumption and economic growth in China. Energy
Policy 32, 47–54.
Squalli, J., 2007. Electricity consumption and economic growth: bounds and causality
analyses of OPEC members. Energy Economics 29, 1192–1205.
Soytas, U., Sari, R., 2003. Energy consumption and GDP: causality relationship in G-7
countries and emerging markets. Energy Economics 25, 33–37.
Soytas, U., Sari, R., 2006. Energy consumption and income in G7 countries. Journal of
Policy Modeling 28, 739–750.
Stern, D.I., 2000. A multivariate cointegration analysis of the role of energy in the US
macroeconomy. Energy Economics 22, 267–283.
Wolde-Rufael, Y., 2005. Energy demand and economic growth: the African experience.
Journal of Policy Modeling 27, 891–903.
Wolde-Rufael, Y., 2006. Electricity consumption and economic growth: a time series
experience for 17 African countries. Energy Policy 34, 1106–1114.
Wolde-Rufael, Y., 2009. Energy consumption and economic growth: the experience of
African countries revisited. Energy Economics 31, 217–224.
Yang, H.Y., 2000. A note on the causal relationship between energy and GDP in Taiwan.
Energy Economics 22, 309–317.
Yu, E.S.H., Choi, J.Y., 1985. The casual relationship between energy and GDP: an
international comparison. Journal of Energy Development 10, 249–272.
Yu, E.S.H., Hwang, B.K., 1984. The relationship between energy and GNP: further results.
Energy Economics 6, 168–190.
Zachariadis, T., Pashourtidou, N., 2007. An empirical analysis of electricity consumption
in Cyprus. Energy Economics 29, 183–198.
Zamani, M., 2007. Energy consumption and economic activities in Iran. Energy Economics
29, 1135–1140.