Causality linkages among energy poverty, income inequality, income poverty and growth: a system dynamic modelling approach.
Murtaza, Ghulam ; Faridi, Muhammad Zahir
The study aims to bring together the new dimensions of poverty
dynamics in a system that will probe the casual relationship among the
energy poverty, income inequality, income poverty and growth. The study
is based on the hypotheses; (I) Does energy poverty cause income
inequality? (II) Does energy poverty a factual impediment to growth?
(Ill) Does energy poverty spur income poverty? And, whether the
alterative also exist against these hypotheses? For this purpose, study
accomplishes an energy poverty measure at national or aggregated macro
level defined by International Energy Agency (IEA). For causality
analysis, study utilises annually time series data ranges from 1973 to
2012; the properties of time series data justify the appropriate method
for the causality analysis should be used. The present study models the
poverty dynamics in Granger non causality context evidenced from
augmented VAR system and follows Toda and Yamamoto (1995), multivariate
TY-procedure for estimation of VARs system at level though seemingly
unrelated regression (SUR) using modified Wald-stat (M-Wald) as the
traditional coefficient restriction. Whereas F-stat does not follow the
normal distribution when simple OLS is used for VARs system estimation
at level. Empirical findings of the study go over the main points; a
significant bi-variant causality linkage between growth and energy
poverty; uni-variant causality that runs from income poverty to energy
poverty and from income inequality to energy poverty is observed.
JEL Classification: O43, L25
Keywords: Energy Poverty, Energy Development Index (EDI), Growth,
Granger
1. INTRODUCTION
The energy services stipulation of a country discloses its
importance as a decency of course of action necessary for economic
prosperity, lessening the poverty and depolarising the social asymmetry
[Barnes, et al. (2011)]. The accomplishment of basic needs of energy
services that include excess to electricity, commercial use of energy
for production process as well as usage of electricity in the
residential areas and modern use of energy sources for cooking purposes
portraits an image of high-quality living standard of individuals and
offers a way forward to economic development. (1) The notion of pro-poor
growth is well documented in the recent literature for assurance of
thinning the poverty that is congregated through translation of growth
into the lives of poor by reshaping the income distribution (2) for
marginalised group of people. Ekouevi and Tuntivate (2012) and studies
of international agencies [AGECC (2010); WHO (2006); UNDP and WHO
(2009)] have preliminary acknowledged the need of improving the access
to reliable and affordable modern energy services in the developing
economies for economic prosperity and social welfare of individuals.
As for as social inequality is concern, energy poverty is of
enormous worth to address it as deficiency in supplying commercial
energy especially electricity, tends to emphasise the social asymmetry
in the society [Pereira (2010)]. While the energy development mitigates
the poverty as it provides the sustainability and enhances the
opportunities for growth that leads to better quality of life [Pereira
(2010)]. The significance of energy services in the mechanism of
structural transformation for development and trading off the old modes
of living for new ones has made the concept of energy poverty a leading
concern now a days. In the developing countries like Pakistan, energy
services supplies are not met perfectly that create social injustice by
depriving people form clear cooking facility that badly effect their
health conditions; as well as from education as new modes of training
and guidance demand electricity essentially. Comfort and ease of life
purely rely on the use of modern home appliance and on vehicles which
run from electricity and fuel accordingly. Thus unswervingly
availability of energy components (i.e., oil, gas, electricity and coal)
at affordable prices diminishes social asymmetry; eliminates poverty;
boosts up economic performance and ultimately up lifts the living
standard of people.
The above deliberation urges to find out the causality linkages
among energy poverty, income inequality, income poverty and growth for
Pakistan. Moreover, secondly, study intends to examine the energy
services conditions through construction of an Energy development index
(EDI) that measures the energy poverty in Pakistan at macro level.
Thirdly, study creates distinction on methodological grounds from rest
of the studies. Study follows multivariate TY- procedure for the
estimation of VAR system through seemingly unrelated regression (SUR)
using modified Wald test for the causality analysis.
After a brief introduction in the first section, trends and size of
energy services in Pakistan and its comparison with the rest of the
economies and regions is drafted under Section 2. Section 3 is about the
energy development index (EDI) and its construction. Review of
Literature is presented in Section 4. Data and methodology is provided
in Section 5 while the empirical results and discussion are presented in
Section 6. At the end, Section 7 is consisting on conclusions and policy
recommendations. 2
2. ENERGY POVERTY SCENARIO IN PAKISTAN
Per capita commercial energy consumption is thought-out well gauge
for energy development which gears up economic growth and eliminates
poverty. The present per capita energy use for Pakistan is near to the
ground. The per capita energy use is 481.61 Kg tons of oil equivalent
(Kg of Toe) for Pakistan while the average per capita energy use of
South Asia is 555 kg Toe; OECD members countries has a average of 4176
kg Toe; Sub-Saharan Africa region has 681 Kg Toe; and, World average per
capita energy consumption is 1890 kg Toe, the estimates of [WDI (2011)]
reveal. This picture depicts the situation of energy poverty in Pakistan
regarding use of energy as within the region, Pakistan energy
consumption is about 15 percent below than average energy consumption of
South Asia; 21 percent less than that of India; and, even less than Sri
Lanka equals to 5 percent nearly. With respect to world energy
consumption, Pakistan uses 75 percent less energy and in comparison to
OECD countries its value is 88 percent. The Figure 2.1 demonstrates the
situation of energy use for Pakistan as compared to different countries
of the world.
People access to electricity is considered first-rated indicator
for excess to modern energy services. The world development indicators
show 1.2 percent increase, from 67.4 percent to 68.6 percent, in
population accessed with the electricity in Pakistan for the year 2010
to 2011. Figure 2.2 displays an inclusive comparison of Pakistan with
different regions and countries to make energy poverty incidence clear
for Pakistan. Within the region of South Asia, Pakistan is providing
electricity less than India, Sri Lanka and Nepal. In contrast to
Malaysia and Unites Arab Emiratis who are providing electricity to whole
population almost, Pakistan has succeeded just 68.6 percent in providing
electricity to its population. Similarly, Pakistan is also 18 percent
below than middle income countries and almost 10 percent below than the
world average in percentage of providing excess to electricity.
[FIGURE 2.1 OMITTED]
Figure 2.3 presents substantial dependence of developing countries
on biomass for cooking purposes. The statistics of World Energy Outlook,
2012 (IEA) and WHO database (2010) indicate that 2588 million people (38
percent of world population); over 1.8 billion people (equals to half of
developing Asia population); and, about 700 million people (80 percent
of the sub-Saharan Africa), who are using traditional biomass sources
for cooking purposes and deprived from clean cooking facilities. 64
percent of population (111 million people) of Pakistan is using
traditional biomass for cooking purposes. While in China, India,
Indonesia, Philippines, Vietnam and rest of developing Asia, 29 percent,
66 percent, 55 percent, 50 percent, 56 percent and 54 percent of
population is not availing clean cooking facilities respectively.
[FIGURE 2.2 OMITTED]
Comparative exploration of modern fuel sources available for
cooking purposes show the incidence of energy poverty in Pakistan.
Biomass dependence, in Pakistan, is almost double than that of china and
world average, almost equal to India and Africa region, 10 percent more
than Vietnam and developing Asia average, 60 percent more than Middle
East, 15 percent more than Philippines and developing countries. So,
large dependence on biomass consumption for cooking purposes designates
Pakistan a poor country who is failing in providing health and safe
cooking facilities. Yet, Pakistan has shown an improvement in that
indicator of energy poverty as in the list of developing Asian countries
Pakistan is keeping pace with China, Thailand and Vietnam where a
notable improvement in lessening biomass dependence is observed.
[FIGURE 2.3 OMITTED]
3. LITERATURE REVIEW
The leading intention of the paper is to present a comprehensive
review of prior work to confer a deep insight about the issue of energy
poverty and its integrating factors. The empirical studies on the issue
of energy poverty for the developing countries are not in surfeit.
However, study makes a healthy endeavour to present literature on prior
work done until now in the following.
The significance of the role of energy especially electricity as a
mean of economic development is dated back at least to 1950s. Supply of
electricity causes to stimulate human productivity and welfare that
ultimately improve economic status of population. It is considered that
poverty elimination, efficiency of productivity, pollution reduction,
and health improvement is the fruit comes from provision of modern
energy [United Nations (1954)].
After gaining the importance from a number of overseas development
agencies [World Bank (1985); WIN (2005); UNDP (2007, 2012); ADB (2010a,
b)], the energy related issues have, now, become the central focus for
economic development and social wellbeing of individuals. The UN General
Assembly has announced the years 2014-2024, to be "the decade of
sustainable energy for all" [United Nations (2014)].
Recent literature and UNDP reports have re-conceptualised the
poverty across-the-board that withdraw it from traditional perception in
which poor were jammed with the notion of earning less than 2 dollar a
day [Sovacool (2012)]. A number of factors have, now, encompassed in the
definition of poverty that include life expectancy, literacy, caloric
intake, housing quality and excess to energy [UNDP (2010)]. This
inaugurated the intuition of non-income dimensions of poverty such as
lack of excess to electricity and reliance on the traditional biomass
fuel for cooking [Joneset, et al. (2010); International Energy Agency
(2010)].
The health impacts of biomass combustion form cooking are observed
in a number of studies. The pragmatic studies of [Pokhreletal (2005,
2013); Shrestha and Shrestha (2005); WIN (2005); Joshi, et al. (2009);
Dhimal, et al. (2010); Mallaetal (2011)] come to a conclusion that
emissions from burning of biomass are harmful for individuals health
significantly, especially, for women and children health which reduce
life expectancy, productivity and efficiency. Besides this, searching
for biomass fuel is a time taking activity that restricts women and
children from any other productive activity [Saghir (2005); Barnes and
Toman (2006)].
Causality linkages of income inequality and energy poverty are well
examined in the studies of [Hussain (2011); Sovacool (2012); Larson and
Kartha (2000); Masud, et al. (2007)]. Studies narrated that income poor
pay eight times more than the other group of income for the same unit of
energy they use. It is estimated that on average 20-30 percent income is
spent on the energy services by the poor households directly while
additional 20-40 percent income is paid out indirectly in term of time
and health injury related with collection and use of raw energy material
respectively. On the other hand, in contrast, making use of modern
energy services in running heavy machinery, illumination of shops and
factories, refrigeration of products for preservation and development of
the mechanisation process has lifted up employment opportunity and
provided incentive to poor by decreasing inequality and increasing their
income level.
Savacool (2012) pointed out a significant relationship between
energy poverty and economic wellbeing of people in the developing
countries. Income poverty and energy deprivation move together, where a
significant proportion of income is allocated for availing energy
services. For an instance, in case of Nepal, the introduction of
renewable energy technologies is the centre focus of government policies
that has activated the balanced growth and helping out to eradicate
poverty [Malla (2013)]. The studies of [Roddis (2000); Cabraal and
Barnes (2006); World Bank (2002)] also drawn the same conclusion of
bi-directional causality between energy development and poverty.
Above narratives make us available a termination that energy
services must be the essential meeting point of any economic agenda and
planning for social development. This leads us to put up an augmented
system that will connect poverty, growth and inequality with the new
no-income dimensions of poverty that is--energy poverty. A plausible
causality linkage among these variables may leave new foresights for
economic planners.
4. THE ENERGY DEVELOPMENT INDEX (EDI)
Ecological scientists and social welfare organiser always put forth
the need of understanding energy poverty to mitigate it [Pachauri and
Spreng (2011)]. It requires apparatus and structure in which it could be
measured, monitored, recorded and reported. A number of scientists, over
last 20 years, are involved in the energy and development issues to
understand the concept of energy poverty [Bravo, et al. (1979);
Bazilian, et al. (2010); Saghir (2004); Krugmann and Goldemberg (1983);
Pachauri and Spreng (2004); Goldemberg (1990); Pachauri and Spreng
(2011); Foster, et al. (2000)]. The present study construct Energy
Development Index (EDI) to measure (3) the energy poverty at national
level for Pakistan following the definition and computation methods of
[IEA (2004); Malla (2013)]. The EDI is a composite index consists of
four indicators or components that are equally weighted but this study
assigned the weight to each indictor on the basis of principal component
analysis (PCA). The Table 4.1 briefly describes the definitions, proxies
and measuring units of indicators of energy poverty for Pakistan. Each
indicator is normalised first by using the following formula;
Indicator = Actual value - Minimun value/Maximun value-Mimmum value
The principal component analysis (PCA) is utilised on all
normalised indicators of energy services to find weights for computing
the energy development index (EDI). The outcomes of PCA show that (PC 1)
explain 97 percent of the standardised variance, the Eigen values of (PC
1) reveal. While (PC 2), (PC 3) and (PC 4) explain standardised variance
equals to 0.018 percent, 0.006 percent, 0.0006 percent respectively. So
the first component (PCI) is best for assigning the weights to
normalised indicators. The individual share of each indicator to EDI is
given as under;
[FIGURE 4.1 OMITTED]
The results of ordinary correlates (provided in Appendix-I) call
for a composite index. The outcomes of the Energy Development Index
(EDI) are graphed for each year as shown in Figure 4.1. The trend of EDI
indicates the development of energy services over the time .Yet this
growth in not in line with the growth rates of other developing
countries. It is observed that from 2007 to onward a decrease in the
trend points out the incidence of energy crisis. The shortage of energy
supply, especially of electricity has increased the magnitude of energy
poverty in Pakistan.
5. DATA AND METHODOLOGY
The study intends to find out the causality linkages among energy
poverty, economic growth, income inequality and income poverty in case
of Pakistan. A number of studies have presented a system that provides
the scheme in which the poverty, growth and inequality are well studied.
The present study augments this system by incorporating the new
dimension of poverty that is--energy poverty. Thus, the study estimates
the dynamic Granger non-causality relationship between poverty, growth,
income inequality and energy poverty by employing multivariate Tota and
Yomamto (1995),TY-modeling.
5.1. Data
The study uses annually time series data for Pakistan ranges from
1973 to 2012. The data are sourced from Economic Survey of Pakistan
(various issues), the World Development Indicators database CR-ROM,
Jamal (2006) and Pakistan labour force survey (various issues),
depending upon the availability of data while some absent values of data
are interpolated by using software, Eviews 7.0 package.
The study uses four variables for the analysis. GDP Per Capita
(GDPPC) is the income per individual measured in Pak rupees, Income
Inequality (INEQ) indicates the distribution of income among different
income groups of people of country proxies by Gini-coefficient (in
percentage), Income Poverty (POV) is measured with head count ratio
(percentage) while the energy poverty (EDI) is expressed with the help
of energy development index (EDI) (4) measured in percentage. All the
variables are expressed in percentage after taking the natural log of
GDPPC.
5.2. Time Series Properties of Data
Before proceeding to multivariate TY-procedure, it requires the
time series properties of data to be scrutinised for obtaining the
maximum order of integration of series. The study uses augmented Dickey-
Fuller (1979), ADF test as well as Phillips Perron (1988), PP test for
robustness of unit root results.
The ADF test works in the following specification where optimal lag
length is selected on the basis of Schwars information criteria (SIC);
[DELTA] [S.sub.i,t] = C + [rho][V.sub.i,t-1] + [k-1.summation over
(j=1)] [GAMMA] [iSi.sub.i,t - j] + [beta]T + [epsilon].sub.i,t] (1)
Where [S.sub.i,t] indicates the respective time series variables
i.e., GDPPC, POV, INEQ, EDI. T specifies time trend, [DELTA] shows first
difference operator and [[epsilon].sub.i,t] is the white noise error.
The Equation (1) tests the Null hypothesis ([rho] = 0) for the existence
of a unit root process in the series against the alternative hypothesis
of ([rho] [not equal to] 0) mean-stationary.
For an exogenous shock to a time series that already has a
deterministic trend (T), the under-rejection of the hypothesis is
inevitable that may not supply robust results [Philip and Perron
(1988)]. So, permitting for dependence and heterogeneity in the error
term, following specification presents the non-parametric adjustment to
ADF test statistic;
[S.sub.i,t] = C + [beta] {t - c/2} + [rho] [S.sub.i,t-1] +
[[epsilon].sub.i,t] (2)
Where, [S.sub.i,t], is the corresponding time series (i.e., GDPPC,
POV, INEQ, EDI), {t - c/2) is the time trend, c stands for sample size
and [[??[.sub.i,t] is white noise error.
5.3. Econometrics Methodology
Existing Literature presents a variety of methodologies available
for causality inferences depending on the characteristics of time series
data. Granger non-causality, Johnson and Juselius (1990) ECM causality,
ARDL modeling causality suggested by Pesaran and Shin (1998), TY-
multivariate model causality and DP nonparametric causality proposed by
Diks and Panchenko (2006) are considered the standard causality tests
available.
This paper follows Toda and Yomamota (1995) to employ
TY-multivariate modeling because of a number of advantages over other
methodologies. Unlike Johnson ECM causality which necessitates same
order of integration of all time series, TY-Procedure is feasible even
when the order of integration of time series is mixed. Thus TY-Procedure
is free from pre-testing of co-integration of the series. Likewise, in
ECM Granger causality, use of standard Wald F-Stat for coefficient
restrictions on parameter after estimating VAR system from OLS, confers
non- standard asymptotic distribution of Wald F-stat that may involve
nuisance parameters if one or more series contain a unit root [Toda and
Phillips (1993); Sims (1990)]. So, TY- modeling is preeminent procedure
for causality inferences as it does not demand any co-integration test
and presents an augmented VAR system narrated as VAR (k + [d.sup.max])
through which restrictions are implemented with the help of modified
Wald Test (MWALD) on VAR(k) after estimating augmented VAR system from
Seemingly unrelated Regression (SUR) at level. Here, k is the number of
lags and [d.sup.max] represents the maximum order of integration among
all the time series. Kuzozumi and Yamamoto (2000) asserted that the
model will be valid until the condition; k > [d.sup.max] holds.
We examine the dynamic causality among energy poverty, growth,
inequality and income poverty by applying the TY- procedure, speified as
follows;
[S.sub.t] = [[phi].sub.s] + ([[phi].sub.1] [S.sub.t-1] +
[[phi].sub.2][S.sub.t-2] + [[phi].sub.3][S.sub.t-3] + ... +
[[phi].sub.n][S.sub.t-n] + [[onega].sub.[tau],t] (3)
Specifying this generalised version of TY-procedure for our
concerned variables (i.e.,EDI, INEQ, GDPPC and POV), we obtain the
following augmented VAR system of equations;
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (4)
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (7)
After the augmented VAR system is constructed, it is estimated from
seemingly unrelated regression(SUR).Standard MWALD is used for the
parameter restrictions on VAR(R) from VAR(k+[d.sup.max]) to get the
value of chi-square statistic that is asymptotically normally
distributed [Zapata and Rambaldi (1997)].
To demonstrate how MWALD works, we consider equation (4) where we
can test the hypothesis that income inequality (INEQ) does not Granger
cause energy poverty (EDI) if [[phi].sub.l1], = 0[for all]i; likewise,
income poverty (POV) does not Granger cause energy poverty (EPI) if
[[gamma].sub.i], = 0[for all]i; similarly, growth (GDPPC) does not
granger cause energy poverty (EDI) if [[psi].sub.i] = 0[for all]i. The
same mechanism is extended for the Equations (5), (6) and (7).
5.5. The Innovation Accounting System
This system demonstrates how a variable retorts from a shock that
comes across in other variables within the system and whether this shock
dies or continues over the time. Following Pesaran and Shin (1948) and
Koop, et al. (1996), we have employed generalised impulse response
function (GIRF) to gauge the comparative potency of causality in an
out-of-sample period as the TY-procedure tests only the long run
causality within the sample period. The generalised impulse response
function (GIRF) has advantages of other standard impulse response
functions [Ewing and Payne (2005)].
6. RESULTS AND DISCUSSIONS
The empirical evidences of Granger non-causality among poverty,
growth, inequality and energy poverty call for a dynamic system as
designed in TY-modeling. This representation persists an augmented VAR
(k+[d.sup.max]) system. For this sake, to find the values of k and
[d.sup.max] for estimating augmented VAR (k+[d.sup.max]), unit root
properties and lag length selection of variables are thin slices of this
segment.
6.1. Stationarity of Data and Lag Length Selection
For any time series analysis, the identification of the unit root
in the time series is important. Study used ADF and PP tests for
scrutinising the order of integration of series. Results are reported in
Table 6.1. Maximum order of integration of concerned variables is
([d.sup.max]=1) which fulfill the requirement of TY-Procedure for
Granger non-causality inference.
Next is to find out the maximum lag length (k) of the time series
variables for the estimation of augmented VAR (k+[d.sup.max]). Different
criterions are available for lag length selection consisting on Akaike
information criteria, Likelihood Ratio, Hannan-Quinn, Final prediction
error and Schwarz information criterion (SIC). Taking small sample size
into account, we supply [1 3] interval for unrestricted VAR output and
same for finding maximum lag length (k). Results are reported in Table
6.2 which shows that consistent maximum lag length is (k=2).
For dynamic Granger non-causality inferences, we have estimated the
augmented VAR (k+[d.sup.max]) that is--VAR(3) in level. The stability
condition of VAR(3) as well Diagnostic tests for each separate equation
of VAR system are performed.
Now, the diagnostic tests are carried out reported in Table 6.3 for
the estimated VAR of order 3. Results indicate that the VAR system is
free from any biasness of regression results. The test of stability of
VAR(3) shows that roots does not lie outside the unit root circle as
confirmed in Figure 6.1. In the same way, we have also applied the
diagnostic tests on each endogenous equation of VAR system before
proceeding to Granger non-Causality tests. Results are presented in
Table 6.4 which indicates that each equation passes the diagnostic
tests.
6.2. Granger Causality Results
The results of Granger non-causality are reported in Table 6.5.
Results provide interesting causality relationship between energy
poverty, growth and income poverty and income inequality for Pakistan
and exemplify worthy integration of variables within the dynamic system
to locate the net collision. We are noteworthy interested in the
direction of causality among economic growth, energy poverty and income
poverty besides a number of other results. The results show
bi-directional long run causality between economic growth and energy
poverty; running from energy poverty to economic growth and vis verse.
It explores the fact that excess to modern energy services are highly
significant for the economic prosperity of Pakistan as energy is
considered the main driver of any economic activity that wheel up the
production process many fold. Similar results are observed for
industrialised, less developed as well as for developing countries like
Nigeria, India, Pakistan and Bangladesh [(Paul and Bhattacharya (2004);
[Worrell, et al. (2001); Mozumder and Marathe (2007); Ojinnaka (1998);
Shahbaz and Feridun (2011); Javid, et al. (2013); Faridi and Murtaza
(2013)].
On the other hand, results reveal that economic well being may
ultimately leads to greater resources to be had to meet the energy
demand challenges and to endow the easiness of life regarding clean
cooking facilities and making more use of modern home appliances.
Likewise, uni-directional causality among energy poverty, income poverty
and income inequality; running from income poverty and income inequality
to energy poverty is observed. This indicates that low income
households, in Pakistan, are not able to afford fully the modern energy
services as essentially they have to devote a large share of their
income for energy services payments as their there exist high income
inequality. The causality linkages also explain that growth is not pro
poor in Pakistan as an increase in national income is not translated
into lives of the poor because growth is not reducing the size of income
distribution imbalances. Consequently, retaining people income poor
makes people energy poor depriving them from clean cooking fuel and
other modern energy services.
After the investigation of causality between energy poverty,
growth, income poverty and inequality, we also estimated the generalised
impulse response function to find the response of a shock of a variable
to other variable within the dynamic VAR system. In order to find the
standard errors, Monte Carlo Simulation is used with 5000 replications.
The results shown in Figure 6.1 verified that the long run causality
that the shock impacts are persistent for a longer period of time. The
impact of income poverty on energy poverty involves a two year lags
after that it gets persistent. Yet response of energy poverty to
inequality is for shorter period of time and dies out after 5 to 6
years.
[FIGURE 6.1 OMITTED]
7. CONCLUSIONS AND POLICY SUGGESTIONS
The present study probes the dynamic causality among energy
poverty, growth, income poverty and income inequality for Pakistan using
the data ranges from 1973 to 2012. The analysis adopts the advanced
TY-modelling in a multivariate framework that overcomes the problem of
variables omission biasness. The extract of the study goes over the main
points that a significant bi-variant causality linkages between growth
and energy poverty; uni-variant causality that runs from income poverty
to energy poverty and from income polarisation to energy poverty is
observed. This furnishes a clear message for the economic planner that
for any social and economic policy, state of energy services must be
considered indispensably. There is urgent need of pro poor growth
policies to depolarise the unfair income distribution and to mitigate
the income poverty so that the fruits of growth may be transferred to
poor and the excess to modern energy services may become possible to
them. That's why, high commercial energy consumption; modern
cooking fuel availability--that saves time and protects health of
households; excess to electricity especially in rural areas are the
limbs of new social and economic development policies that Pakistan
should follow for all these concerned intents and purposes.
Ghulam Murtaza <gm.qaui@gmail.com,
ghulammurtaza_14@pide.edu.pk> is PhD Scholar at Pakistan Institute of
Development Economics, Islamabad. Muhammad Zahir Faridi
<zahirfaridi@bzu.edu.pk> is Associate Professor of Economics,
Bahauddin Zakariya University, Multan.
APPENDIX-I
Fig. 6.1. Inverse Roots of AR Characteristics Poly Nominal
Principal Components Analysis
Sample Size : 1973-2012
Cumulative Cumulative
Number Value Difference Proportion Value Proportion
1 3.897334 3.822513 0.9743 3.897334 0.9743
2 0.074821 0.049537 0.0187 3.972155 0.9930
3 0.025284 0.022722 0.0063 3.997439 0.9994
4 0.002561 -- 0.0006 4.000000 1.0000
Variables PC 1 PC 2 PC 3 PC 4
Electrification Rate 0.258698 0.823500 0.255691 -0.115485
Fusel Fuel 0.243443 -0.014865 -0.684998 0.526499
Consumption
Per Capita Electricity 0.259862 -0.451212 0.649089 0.353877
Consumption in
Residential Areas
Per Capita Energy Use 0.253636 -0.343562 -0.209958 -0.764352
Ordinary Correlations
Electrification Fusel Fuel
Variables Rate Consumption
Electrification Rate 1.000000
Fusel Fuel 0.961786 1.000000
Consumption
Per Capita Electricity 0.936884 0.970316
Consumption in
Residential Areas
Per Capita Energy Use 0.945549 0.990971
Per Capita
Electricity
Consumption in
Residential Per Capita
Variables Areas Energy Use
Electrification Rate
Fusel Fuel
Consumption
Per Capita Electricity 1.000000
Consumption in
Residential Areas
Per Capita Energy Use 0.988608 1.000000
REFERENCES
ADB (2010a) Inclusive Growth: Impact Stories from Nepal. Manila:
ADB.
ADB (2010b) Clean Energy in Asia: Case Studies of ADB Investments
in Low-Carbon Growth. Manila: ADB.
AGECC (2010) Energy for a Sustainable Future. Advisory Group on
Energy and Climate Change (AGECC). New York: United Nations.
Awan, S. (2013) Determinants of Multidimensional Energy Poverty in
Pakistan. The Pakistan Development Review 52:4.
Barnes, D. F., S. R. Khandker, Sand H. A. Samad (2011) Energy
Poverty in Rural Bangladesh. Energy Policy 39, 894-904.
Barnes, Douglas, and Michael Toman (2006) Energy, Equity and
Economic Development. In Ramon Lopez and Michael Toman (Eds.) Economic
Development and Environmental Sustainability: New Policy Options.
Oxford: Oxford University Press.
Bazilian, M., P. Nussbaumer, A. Cabraal, R. Centurelli, R. Detchen,
G. Gielen, H. Rogner, M. Howells, H. McMahon, V. Modi, N. Nakicenovic,
B. O'Gallachoir, M. Radka, K. Rijal, M. Takada, and F. Ziegler
(2010) Measuring Energy Access: Supporting a Global Target. Earth
Institute, Columbia University, NY, USA.
Bravo, V., G. G. Mendoza, J. Legisa, C. E. Suarez, and I.
Zyngierman (1979) Estudiosobrerequerimientosfuturos no convencionales de
energiaen America Latina. Project RLA/74/030, Report to the UNDP,
Appendix 9. FundacionBariloche, Buenos Aires.
Dhimal, M., P. Dhakal, N. Shrestha, K. Baral, and M. Maskey (2010)
Environmental Burden of Acute Respiratory Infection and Pneumonia Due to
Indoor Smoke in Dhading. Journal of Nepal Health Research Council 8,
1-4.
Dickey, D. and W. A. Fuller (1979) Distribution of the Estimate for
Autoregressive Time Series with a Unit Root. Journal of American
Statistical Association 74, 427-31.
Diks, C. G. H. and V. Panchenko (2006) A New Statistic and
Practical Guidelines for Nonparametric Granger Causality Testing.
Journal of Economic Dynamics and Control 30, 1647-1669.
Ekouevi, K. and V. Tuntivate (2012) Household Energy Access for
Cooking and Heating: Lessons Learned and the Way Forward. Washington,
DC: World Bank.
Ewing, B. T. and J. E. Payne (2005) The Response of Real Estate
Investment Trust Returns to Macroeconomic Shocks. Journal of Business
Research 58, 293-300.
Faridi Z. and G. Murtaza (2013) Disaggregate Energy Consumption,
Agricultural Output and Economic Growth in Pakistan. The Pakistan
Development Review 52:4.
Foster, V., J. P. Tre, and Q. Wodon (2000) Energy Prices, Energy
Efficiency, and Fuel Poverty. Unpublished paper Latin America and
Caribbean Regional Studies Program. The World Bank, Washington, DC.
Goldemberg, J. (1990) One Kilowatt Per Capita. Bulletin of the
Atomic Scientists 46:1.
Goldemberg, J., E. L. La Rovere, and S. T. Coelho (2004) Expanding
Access to Electricity in Brazil.
Holdren, J. P. and K. R. Smith (2000) Energy, the Environment, and
Health. In Kjellstrom Tord, Streets David, Wang Xiadong (eds.) World
Energy Assessment Energy and the Challenge of Sustainability. New York:
United Nations Development Programme, 61-110.
Hussain, F. (2011) Challenges and Opportunities for Investments in
Rural Energy. Presentation to the United Nations Economic and Social
Commission for Asia and the Pacific (UNESCAP) and International Fund for
Agricultural Development (IFAD) Inception Workshop on "Leveraging
Pro-Poor Public-Private-Partnerships (5Ps) for Rural Development,"
United Nations Convention Centre, Bangkok, Thailand.
IEA, International Energy Agency (2004) World Energy Outlook.
Paris.
Jamal, H. (2006) Does Inequality Matter for Poverty Reduction?
Evidence from Pakistan's Poverty Trends. The Pakistan Development
Review 45:3, 439-459.
Johansen, S. and K. Juselius (1990) Maximum Likelihood Estimation
and Inference on Cointegration: With Application to the Demand for
Money. Oxford Bull Econ Stat 52, 169-210.
Jones, R. (2010) Energy Poverty: How to Make Modern Energy Access
Universal? Special Early Excerpt of the World Energy Outlook 2010 for
the UN General Assembly on the Millennium Development Goals. Paris:
International Energy Agency/OECD.
Joshi, H. D., R. Pandey, and B. Dhakal (2009) Health Impact of
Indoor Air Pollution. Journal of Nepal Health Research Council 7, 69-75.
Khandker, S. R., D. F. Barnes, and H. A. Samad (2012) Are the
Energy Poor Also Income Poor? Evidence from India. Energy Policy 47,
1-12.
Koop, G., M. H. Pesaron, and S. M. Potter (1996) Impulse Response
Analysis in Nonlinear Multivariate Models. Journal of Econometrics 74,
119-47.
Krugmann, Hartmut and Jose Goldemberg (1983) The Energy Cost of
Satisfying Basic Human Needs. Technological Forecasting and Social
Change 24, 45-60.
Kuzozumi, E. and T. Yamamoto (2000) Modified Lag Augmented
Autoregressions. Econometric Review 19, 207-31.
Larson, E. D. and S. Kartha (2000) Expanding Roles for Modernised
Biomass Energy. Energy Sustain Dev. 4:3, 15-25.
Lutkepohl, H. (1993) Introduction to Multiple Time Series. 2nd ed.
Berlin: Springer.
Malla, M. B., N. Bruce, E. Bates, and E. Rehfuess (2011) Applying
Global Cost-benefit Analysis Methods to Indoor Air Pollution Mitigation
Interventions in Nepal, Kenya and Sudan: Insights and Challenges. Energy
Policy 39, 7518-7529.
Malla, S. (2013) Household Energy Consumption Patterns and Its
Environmental Implications: Assessment of Energy Access and Poverty in
Nepal. Energy Policy 61, 990-1002.
Masud, J., D. Sharan, and B. N. Lohani (2007) Energy for All:
Addressing the Energy, Environment, and Poverty Nexus in Asia. Manila:
Asian Development Bank.
Mirza, B. and A. Szirmai (2010) Towards a New Measurement of Energy
Poverty: A Cross-Community Analysis of Rural Pakistan. United Nations
University, Maastricht Economic and social Research and Training Centre
on Innovation and Technology. (UNU-MERIT Working Paper Series 024).
Mozumder, P. and A. Marathe (2007) Causality Relationship Between
Electricity Consumption and GDP in Bangladesh. Energy Policy 35,
395-402.
Nussbaumer, P., M. Bazilian, and V. Modi (2012) Measuring Energy
Poverty: Focusing on What Matters. Renewable and Sustainable Energy
Reviews 16, 231-243.
Pachauri, S. and D. Spreng (2004) Energy Use and Energy Access in
Relation to Poverty. Economic and Political Weekly 39:3, 17-23.
Pachauri, S. and D. Spreng (2011) Measuring and Monitoring Energy
Poverty. Energy Policy 39, 7497-7504.
Pachauri, S., A. Muller, A. Kemmler, and D. Spreng (2004) On
Measuring Energy Poverty in Indian Households. World Development 32:12,
2083-2104.
Paul., S. and R. N. Bhattacharya (2004) Causality Between Energy
Consumption and Economic Growth in India: A Note on Conflicting Results.
Energy Economics 26, 977-983.
Pereira, M. G., M. A. V. Freitas, and N. F. D. Silva (2010) Rural
Electrification and Energy Poverty: Empirical Evidence from Brazil.
Renewable and Sustainable Energy Review 14, 1229-1240.
Pesaran, M. H. and Y. Shin (1998) Generalised Impulse Response
Analysis in Linear Multivariate Models. Economics Letters 58, 17-29.
Pesaran, M. H. and Y. Shin (1998) Impulse Response Analysis in
Linear Multivariate Models. Energy for Sustainable Development 8:4,
86-94.
Phillips, P. C. B. and P. Perron (1988) Testing for a Unit Root in
Time Series Regression. Biometrika 75:2, 336-346.
Pokhrel, A. K., K. R. Smith, A. Khalakdina, A. Deuja, and M. N.
Bates (2005) Case-control Study of Indoor Cooking Smoke Exposure and
Cataract in Nepal and India. International Journal of Epidemiology 34,
702-708.
Pokhrel, A. K., M. N. Bates, S. P. Shrestha, I. L. Bailey, R. B.
Dimartino, and K. R. Smith (2013) Biomass Stoves and Lens Opacity and
Cataract in Nepalese Women. Optometry and Vision Science 90, 257-268.
Practical, Action (2010) Poor People's Energy Outlook.
Sagar, A. D. (2005) Alleviating Energy Poverty for the World's
Poor. Energy Policy 33, 1367-1372.
Saghir, J. (2004) Energy and Poverty: Myths, Links, and Policy
Issues. Energy Sector Notes. Energy and Mining Sector Board, 4, World
Bank, Washington, DC.
Scheikman, J. A. (2002) An Agenda Perdida-diagno' sticos e
propostas para a retomada do crescimento com maiorjustic.a social. Rio
de Janeiro, Instituto de Estudos do Trabalho e Sociedade (IETS).
Shahbaz, M. and M. Feridun (2011) Electricity Consumption and
Economic Growth Empirical Evidence from Pakistan. Qual Quant.
Shrestha, I. L., and S. L. Shrestha (2005) Indoor Air Pollution
from Biomass Fuels and Respiratory Health of the Exposed Population in
Nepalese Households. International Journal of Occupational and
Environmental Health 11, 150-160.
Sims, C. and J. Stock (1990) Watson M. Inference in Linear Time
Series Models with Unit Roots. Econometrica 58, 113-44.
Sovacool, B. K., C. Cooper, M. Bazilian, K. Johnson, D. Zoppo, S.
Clarke, J. Eidsness, M. Crafton, T. Velumail, and H. A. Raza (2012) What
Move Sand Works: Broadening the Consideration of Energy Poverty. Energy
Policy 42, 715-719.
Toda, H. Y. and P. C. B. Phillips (1993) Vector Autoregressions and
Causality. Econometrica 61, 1367-93.
Toda, H. Y. and T. Yamamoto (1995) Statistical Inference in Vector
Autoregression with Possibly Integrated Processes. Journal of
Econometrics 66, 225-250.
UNDP (2007) Energy and Poverty in Nepal: Challenges and the Way
Forward. Regional Energy Programme for Poverty Reduction. Kathmandu:
UNDP.
UNDP (2012) Energy to Move Rural Nepal Out of Poverty: The Rural
Energy Development Programme Model in Nepal. Kathmandu: UNDP.
UNDP, WHO (2009) The Energy Access Situation in Developing
Countries: A Review Focusing on the Least Developed Countries and
Sub-Saharan Africa. New York.
United Nations (1954) Rural Electrification. Prepared by the
Secretariat of the Economic Commission for Asia and the Far East.
Geneva: Department of Economic Affairs.
United Nations Development Program (2010). Human Development
Report. New York: UNDP.
WHO (2006) Fuel for Life: Household Energy and Health. Geneva:
World Health Organisation (WHO).
WIN (2005) Household Energy, Indoor Air Pollution and Health
Impacts: Status Report for Nepal. Kathmandu: Winrock International Nepal
(WIN).
World Bank (1985) Energy Assessment Status Report: Nepal. World
Bank, Washington, DC. (Report No. 028/84).
World Bank (2002) Rural Electrification and Development in the
Philippines: Measuring the Social and Economic Benefits. Energy Sector
Management Assistance Programme (ESMAP), 255/02. World Bank, Washington,
DC.
Worrell, E., L. Price, and LBNL (2001) Pew Centre/NCEP 10-50.
Improving Industrial Energy Efficiency in the U.S.: Technologies and
Policies for 2010 to 2050. Workshop Proceedings. "The 10-50
Solution: Technologies and Policies for a Low-Carbon Future."
Zapata, H. O. and A. N. Rambaldi (1997) Monte Carlo Evidence on
Cointegration and Causation. Oxford Bulletin of Economics and Statistics
59, 285-98.
Authors' Note: The first author is indebted to Prof. Dr.
Khurshed Hassan Khawar, Govt. Gorden College, Rawalpindi, for his
interminable efforts and worthy support in pulling-off my aspirations.
The earlier version of this paper is presented at the 30th AGM and
Conference 2014, PSDE, PIDE, Islamabad. The authors much appreciate the
helpful comments of conference participants and further benefitted from
the remarkable comments and suggestions of Ms. Rashida Haq, Discussant,
PSDE, PIDE.
(1) International Energy Agency report (2010) declares that 1.3
billion people living without excess to electricity and about 2.6
million people who are not provided clean cooking facilities globally.
This indicates a serious impediment to social and economic development
and must be addressed uncompromisingly for the achievement of UN
Millennium Development Goals [Dagoumas and Kitsios (2014)].
(2) According to Scheikman (2002), the prudent government policies
formulated with the aim of reducing poverty and income inequality
account education and health substances a lot and these issues cannot be
accomplished without required energy services.
(3) Still there is no consensus on the issue of measuring energy
poverty [Nussbaumer, et al. (2012)]. Different studies on measuring the
Energy poverty based on; different approaches; definitions; data
availability are being cited as under for reference and not discussed in
detail as this is beyond the scope of this paper. [Bazilian, et al.
(2010); Foster, et al. (2000); Mirza and Szirmai (2010); Barnes, et al.
(2010); Practical Action (2010); Awan, et al. (2013); Pachauri, et al.
(2004); IEA (2004); World Energy Outlook (2010); United Nations
Development Program (2010); Jones (2010); Holdren and Smith (2000);
Khandker, et al. (2012); Sovacool, et al. (2012)].
(4) ED1 is measured with the help of a composite index consists of
four variables. Definitions, measuring units and proxies of all four
variables (indicators) are provided in Table 4.1 under Section 4 in
detail.
Table 4.1
Indicators for Energy Development Index (EDI)
Indicator Definition
Per Capita It is the amount of energy per capita used in
Commercial Energy the production process indicates the overall
Consumption economic development of the country.
Excess to Electricity People from total population availing
the facility of electricity which is an
indicator for social asymmetry, reliance
and ease of life.
Per Capita It is per capita consumption of
Electricity in electricity in the residential sector that
Residential Sector express the ability of the consumer for
the payment of electricity services and
basic reliability.
Share of Modern The excess of modern energy services
Energy Fuel in for cooking purposes out of total energy
Total Residential services provided to household instead
Energy Use of traditional biomass burning for
cooking. It includes the use of oil, gas
and electricity.
Units of
Indicator Proxy Measurement
Per Capita Commercial Energy Tonnes of oil
Commercial Energy Consumption Per equivalent (Toe)
Consumption Capita
Excess to Electricity Rate of Percentage
Electrification
Per Capita Per Capita Tonnes of oil
Electricity in Electricity equivalent (Toe)
Residential Sector Consumption in
Residential Sector
Share of Modern Share of Fossil Percentage
Energy Fuel in Fuel Energy
Total Residential Consumption in
Energy Use Total Consumption
Table 6.1
Stationarity of Data
At Level
Trend and
Intercept Intercept
Variable ADF PP ADF PP
RGDPC 0.33 1.65 -4.21 -4.42 *
EDI -2.69 -2.69 0.31 0.02
POV -0.70 -1.28 -2.35 -1.57
INEQ -2.63 *** -2.92 ** -2.85 -3.3 **
With First Difference
Trend and
Intercept Intercept Max. *
Lag Order of
Variable ADF pp ADF pp Length Integration
RGDPC -10.5 * -10.92 * -- -- 9 I(0)
EDI -4.23 * -4.24 * -- -- 9 I(1)
POV -1.73 -4.12 * -0.40 -4.1 * 9 I(1)
INEQ -- -- -- -- 9 I(0)
Source: Authors' calculations, * max lag length for ADF test is 9
where optimal lag length is chosen on the basis Schwarz info
criterion. For PP test, Bandwidth is opted on the basis of Newey-
West using Bartlett kernel. Critical values for different level of
significance are cited from MacKinnon (1996). *, **, *** represents 1
percent, 5 percent and 10 percent level of significance respectively.
Table 6.2
VAR Lag Order Selection Criteria
Lag LogL LR FPE
0 -362.7281 NA 4777.565
1 -136.2288 391.7825 0.055020
2 -112.3280 36.17433 * 0.037106 *
3 -95.76586 21.48596 0.039199
Lag AIC SC HQ
0 19.82314 19.99730 19.88454
1 8.444803 9.315569 * 8.751788
2 8.017727 9.585107 8.570302 *
3 7.987344 * 10.25134 8.785507
Source: Authors calculations.
Table 6.3
Diagnostic Test Results of VAR(3)
Diagnastic Tests Test Statistics p-values
Autocorrelation LM 261.90 .158
Residual Noramlity (J-B test) 13.35 .101
White Heteroskedasticity Test 22.98 .114
VAR Stability -No root lies outside the
unit circle-
Source: Authors calculations.
Table 6.4
Diagnostic Tests of Estimated Endogenous Equations
Autocorrelation- Residual
Equations LM Normality (J-B)
EDI .301 13.97
(.824) (.497)
INEQ 1.089 13.54
(.375) (0331)
GDPPC 1.051 .382
(.390) (.825)
POV 3.026 9.431
(0.042) (.097)
White
Equations Heteroskedasticity(ARCH) CUSUM Test
EDI 0.244 Within limits
(0.62)
INEQ 2.733 Within limits
(.107)
GDPPC .853 Within limits
(361)
POV 7.131 Within limits
(.0329)
Source: Authors calculations.
Table 6.5
Results of Dynamic Granger non-Causality
MWALD Test
Dependant Economic Income Energy Income
Variables Growth Poverty Poverty Polarisation
Economic 1 5.841 ** 16.482 * 1.948
Growth (0.053) (0.0003) (0.377)
Income Poverty 0.521 1 3.972 2.853
(0.770) (0.121) (.248)
Energy Poverty 17.140 * 10.160 * 7.719 *
(0.0002) (0.006) 1 (0.021)
Income 3.741 13.850 * 1.666 1
Polarisation (0.154) (0.001) (0.4346)
Dependant
Variables Causality Inferences
Economic Economic Growth [left arrow]
Growth Income Poverty
Economic Growth [left arrow]
Energy Poverty
Income Poverty --
Energy Poverty Energy Poverty [left arrow]
Economic Growth
Energy Poverty [left arrow]
Income Poverty
Energy Poverty [left arrow]
Income Polarisation
Income Income Polarisation [left arrow]
Polarisation Income Poverty
Source: Authors calculations. *, ** represent significance level of
1 percent and 5 percent respectively.
