Energy and economic growth in Pakistan.
Siddiqui, Rehana
Recent rise in energy prices, shrinking existing resources, and the
search for alternative sources of energy and energy conservation
technologies have brought into focus the issue of causality between
energy use and economic growth. The results of this study show that
energy, expansion is expected to lead to higher growth and its shortage
may retard the growth process. the impact of all sources of energy on
economic growth is not the same. The impact of electricity and petroleum
products as well as that of electricity only is high and statistically
significant. However, the reverse causality is critical for the
petroleum products
1. INTRODUCTION
Recent rise in energy prices, shrinking existing resources, and the
search for alternative sources of energy and energy conservation
technologies have brought into focus the issue of causality between
energy use and economic growth. (1) Energy expansion is expected to lead
to higher growth and its shortage may retard the growth process. (2,3)
Similarly economic growth may affect the demand for energy
significantly. In empirical literature, however, there is no consensus
about the direction of causality. For example, Asafu-Adjaye (2000)
examines the causal relationship between energy consumption, energy
prices, and economic growth for selected developing countries. The study
finds evidence of uni-directional Granger causality running from energy
to income for India and Indonesia, in the short run, and bi-directional
Granger causality between energy and income for The Philippines and
Thailand. The evidence for Pakistan also reveals that electricity
consumption affects economic growth significantly, and there is
bi-directional causality between economic growth and consumption of
petroleum products and no causal relationship between natural gas
consumption and economic growth [see Aqeel and Butt (2001)]. At the
sectoral level, the evidence shows that energy use affects the growth of
manufacturing sector of Pakistan, however, the substitution possibilities are limited among energy and non-energy inputs and between
electricity and gas for the period 1972-93 [see Mahmud (2000)].
Energy demand, particularly for households, responds positively and
significantly to economic growth [see Siddiqui (1999)]. The demand is
responsive to changes in energy prices also. Own price effect is
negative and the cross price elasticity estimates indicate substitution
between electricity and petroleum products and between natural gas and
petroleum products, especially for domestic users. The results for
commodity producing sectors like industry and agriculture, reported in
Siddiqui (1999), are supported by the findings of Mahmud (2000) that
there is limited substitutability between different sources of energy.
Thus, the rise in prices of energy has important implications for energy
use in Pakistan. (4)
In this paper, our objective is to examine the issue of causality
between economic growth and energy use for Pakistan, for the period
1971-2003. The study differs from earlier studies in three dimensions.
First, earlier studies, like Aqeel and Butt (2001), examine the issue of
causality for Pakistan but ignore the impact of changes in other sources
of economic growth. We intend to analyse the role of energy in economic
growth while controlling for changes in primary factors of production
and other sources of growth, viz., labour, capital, human capital
formation and exports. Second, earlier studies examine the impact of
total energy use on economic growth. The households are important users
of energy, however, this use may not contribute to economic growth.
Therefore, in this study, we exclude the household consumption of energy
and examine the impact of commercial use of energy on economic growth.
Third, unlike earlier studies, the present study constructs capital
stock series to examine the impact of capital formation on economic
growth.
The order of the study is as follows: Current issues and trends in
energy use are discussed in Section 2. Section 3 outlines the model,
methodology and data issues. The results are presented in Section 4.
Conclusions and policy implications of the results and the future
directions for research are discussed, briefly, in Section 5.
2. ENERGY TRENDS IN PAKISTAN (5)
Table A1 shows that in 1990s, the energy requirement in Pakistan
was lower relative to many developing and developed countries except for
Sri Lanka and Nepal. This may be a result of lower availability and/or
lower energy intensity in Pakistan. Per capita availability of energy,
used as indicator of prosperity, is low and it has remained constant
from 1993 to 1996 revealing that the country is relatively poor in terms
of availability of energy, but the growth rate of commercial use of
energy is similar to other developing countries in the region like China
and India. Table A1 also indicates that the gross domestic product (GDP)
per unit of energy is lower in Pakistan as compared to other countries.
The growth rate of energy use per unit of GDP was lower in Pakistan as
compared to Bangladesh, India and Nepal. In Pakistan, it declined from
3.3 percent in 1990 to 2.8 percent in 1997. (6)
In Pakistan, ratio of growth rate of energy use to growth rate of
output produced, viz., indicator of energy intensity, was around 0.96
during 1970-2003. However, energy intensity, of different sources of
energy viz., electricity, natural gas and petroleum products varied
significantly over time (7) (see Table 1). The energy coefficient for
electricity is 1.51 for total electricity use and 1.19 for the
commercial use of electricity. This shows that inclusion of domestic
consumption of electricity overestimates the energy coefficient. The
coefficients for gas and petroleum products also change when we exclude
household consumption from total, but the change is largest for
electricity. (8) Thus, in order to forecast energy needs for economic
growth, it is important to exclude domestic use from total energy use.
The coefficients may decrease as the efficiency of fuel use
increases either due to improvements in technology or due to reduction
in wastage. This indicates that Pakistan should concentrate not only on
the expansion of energy sources but also on efforts to improve the
efficiency of energy use. How efficiency can be improved--is an
important issue and the role of pricing mechanism may be important. (9)
Main features and critical issues for these components of energy are
discussed below:
2(a) Petroleum Products
Given initial gap in net supply and net consumption of energy, a
higher growth in consumption relative to growth in supply indicates a
widening gap between energy demand and supply. Table A2 shows that, on
average, the gap between demand and supply of petroleum products reduced
by 0.17 percent during 1991-2003. However, the growth rate of supply and
consumption, both, was negative in 1990-91, in 1996-97 and onwards from
2000-01. From Figure 1, we can see fluctuations in the supply and demand
for petroleum products. The demand and supply of petroleum products
moved together with supply lagging behind slightly.
[FIGURE 1 OMITTED]
In addition to economic growth, the rise and fall in supply and
demand for petroleum products has important implications for balance of
payment. In 1980s, approximately 90 percent of oil needs were fulfilled from imports of petroleum products. Despite the rise in prices the
import of petroleum products increased resulting in rising share of
petroleum products in total imports until 1999-2000 but declined
afterwards (see Table 2). The growth rate of total imports was negative
(-0.98 percent per annum) during 1996-2000, but the decline in import
growth rate, excluding petroleum products, is larger i.e., equaling
-3.09 percent, in 1996-2000. (see Table 2). This reflects the
significance of imports of petroleum products in total imports and its
implications for balance of payments.
In order to reduce import dependence and encourage exploration and
conservation efforts, Government of Pakistan announced petroleum
policies in 1990-91, 1994 and in 1997. In the 1990s, the emphasis of the
government policies was to exploit existing energy resources and to
build strong base for the domestic production and exploration. The main
objectives of the policies were to ensure adequate and cost-effective
provision of the energy with minimum environmental cost to the various
sectors of the economy. The efforts were directed towards cost
effectiveness, reduction in import dependence, promotion of
self-reliance through accelerated exploitation, minimum environmental
degradation, encouraging private foreign investment, creation of
qualitatively improved infrastructure in oil and gas industry,
development of an efficient and transparent management, deregulation of
downstream petroleum marketing sector, and rationalisation of prices and
LPG allocation.
In petroleum policy of 1994, the emphasis was on providing the
fiscal incentives to the petroleum industries and to suggest measures
for quality control, to minimise discretion and to increase transparency in the measures undertaken by the government. In 1996, a number of
fiscal incentives, given to petroleum exploring industries, were
withdrawn which affected the activities adversely. In the policy of
1997, the emphasis was on revival of those incentives and on encouraging
off-shore exploration activity. However, we can see that rapid growth in
import of petroleum products, despite rise in prices has resulted in
higher dependence on imports. The changing pricing policies with
deregulation efforts have also added uncertainty in the market for oil
products. (10)
The sectoral share for petroleum products also varies. The share of
transport sector varied between 48.6 percent (in 1990-91) and 49.1
percent in 2002-03. The share of power sector varied between 24.4
percent in 1990-91 and 48.8 percent in 2002-03. This shows that the
transport and power sectors are the main users of petroleum products as
the total use of these two sectors varied between 70-90 percent of total
consumption. The share of household sector was 9.5 percent in 1990-91
and it declined sharply to 1.72 percent in 2002-03. (11) The declining
share of domestic sector in consumption of petroleum products could be a
result of substitution of natural gas for domestic use [see Siddiqui
(1999)]. Thus, we expect that growth in supply and demand for petroleum
products is significantly correlated with growth in output, particularly
the growth of services sector.
2(b) Natural Gas
Natural gas is another critical source of energy and its net supply
and net demand increased during the 1990s (see Table A2). In early
1990s, as compared to late 1990s, the gap between supply and demand,
given the initial positive gap of 417060 (TOE), was large due to sharp
fluctuations in net supply (see Figure 2). In recent years the demand
for natural gas is rising whereas the supply is showing a downward trend
indicating a need for efforts to increase net supply through exploration
and conservation.
[FIGURE 2 OMITTED]
A number of steps were taken by the Government of Pakistan (GOP) to
improve gas supply. For example, the privatisation of distribution
system and establishment of Oil and Gas Regulatory Authority. However,
the privatisation process of the SSGC and SNGPL is progressing at a slow
pace. In order to speed up the process, Oil and Gas Regulatory Authority
(OGRA) is established which will specify the revenue requirements for
companies engaged in transmission and distribution and regulate the
functioning of the gas supply and distribution companies. (12)
The gas distribution agencies SSGC and SNGPL purchase gas from
producers at prices fixed under different pricing regimes. These
agencies sell gas to various categories of consumers according to tariff schedule determined by the GOP. Regardless of the difference in cost due
to location, all consumers within the same category pay the same price.
GOP also controls expansion of gas transmission system and connections
to new consumers. (13)
In 1990s, the use of natural gas in various sectors of the economy
shows fluctuations, except for the power sector. The increased use of
natural gas was also a result of inter-fuel substitution in transport
sector and in power sector (see Table 4). In 2002-03, the demand for
natural gas from different sectors was: 17.7 percent for households, 2.9
percent for commercial, 1.7 percent for cement, 24.4 percent for
fertiliser, 34. I percent for power, and 19 percent for industrial
sector.
2(c) Electricity
Growth rate of net supply and net consumption of electricity slowed
down after 1995-96, however divergence between two indicators increased
resulting in cyclical pattern. In fact the gap between supply and demand
was negative showing rapid growth in demand, relative to supply, in the
early two years of the current decade (see Table A2 in Appendix). This
may be a reflection of improvements in economic activity. The growth
rate in net supply and demand for electricity also exhibits sharp
fluctuations (see Figure 3).
In 1980s, despite increase in supply, Pakistan was unable to meet
rapidly growing demand for electricity due to financial and political
constraints. In 1985, under an agreement with the World Bank, a policy
for encouraging private power producers and attracting foreign direct
investment in the power sector was announced. However, the details with
the applicants from the international consortium and groups, were
finalised after the mid-1990s. In the Energy Policy announced in 1994,
incentives were given to independent power producers (IPPs) to set up
thermal power units in Pakistan. Initially after intensive negotiations,
13 projects with a capacity of 2700 MW were finalised. Among these 13
projects, 11 have started production, however, HUBCO with largest power
generating capacity of 1200 MW started production in 1997. Currently, 16
IPPs with generation capacity of 5794 MW are in operation. (14) In
2002-03, total power generating capacity of WAPDA and KESC was 6491 MW
for thermal power and 5045 MW for hydel power.
[FIGURE 3 OMITTED]
In the power sector reduction in transmission and distribution
losses and theft are also critical issues for improving the efficiency
of consumption and distribution. The losses were more than 20 percent in
1990s with a rising trend from 21.1 percent in 1990-91 to 23.4 percent
in 2002-03 (see Table A3 in Appendix). In order to reduce these losses,
WAPDA has taken a number of steps. For example, the induction of Army
Monitoring Teams to check the bills and the electricity theft and
transmission losses, according to Chairman of WAPDA, has resulted in 10
percent overall reduction in system losses resulting in monitory benefit
of about Rs 15 billion. Managed billing and collection has also
increased the revenue of WAPDA by Rs 133 billion in 1998-99 from Rs 117
billion last year. Similarly, renegotiated tariff of 4.75 cents, with
IPPs, also resulted in saving of $1.5 billion over the life of the
projects [see Bashar (2000)]. (15)
Like other energy products, the price of electricity also increased
sharply in the 1990s. The increase in price of electricity resulted in
substitution of natural gas for electricity, particularly in the
agriculture and industrial sector. The share of these sectors declined
from 35.61 and 17.82 percent in 1990-91 to 30.73 percent and 11.43
percent in 2002-03, respectively. Despite the rise in electricity tariff
the share of household sector in total increased from 33.01 percent in
1990-91 to 44.87 in 2003-03 (see Table 5). (16)
On the basis of this discussion a number of critical issues emerge.
For example, what is the impact of changes in energy use on economic
growth, what are the determinants of supply and demand of different
sources of energy, how pricing policies affect inter-fuel substitution,
and what is the role of foreign direct investment on energy sector. In
this study, we analyse the first issue in detail. (17)
3. METHODOLOGY
The model, methodology used for testing the causality between
economic growth and energy use and data issues are discussed in this
section.
3(a) Model
The impact of energy use on economic growth has become critical
after the energy shocks in 1970s and recent emphasis on shrinking energy
resources and search for energy efficient production technologies and
equipments. The role of energy in economic growth is highlighted in a
number of studies [see for example, Aqeel and Butt (2001); Moroney
(1992); Riaz (1986) and Stern and Cleveland (2003)]. The studies used
production function approach to examine the impact of energy on economic
growth. Following Moroney (1992) and Stern and Cleveland (2003), the
model assumes production function of the following form:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] ... (1)
where:
[Q.sub.t] is total output produced at 't';
[A.sub.t] is state of technology at time 't';
[K.sub.t] is capital stock at time 't';
[L.sub.t] is labour force at time 't';
[H.sub.t] is indicator of human capital;
[X.sub.t] is exports at time 't';
[E.sub.jt] is consumption of 'jth' energy at time
't' (where energy sources are electricity, natural gas and
petroleum products).
Taking logarithm on both side the equation, we can write Equation
(1) as:
ln([Q.sub.t]) = ln ([A.sub.t]) + [alpha]ln([K.sub.t]) + [beta] ln
([L.sub.t]) + [delta] ln ([H.sub.t]) + [eta] ln ([Xsub.t]) + [summation]
[[gamma].sup.j] ln([E.sub.jt]) ... (2)
'j' varies from 1,2,3 (for electricity (I), gas (2) and
petroleum products(3))
t = 1, ... ..., 33 (for the period from 1969-70 to 2002-03)
Taking first difference, on both sides, Equation (2) can be written
as following growth equation:
[G.sub.q] = [G.sub.a] + [alpha] [G.sub.k] + [beta] [G.sub.l] +
[delta] [G.sub.h] + [eta][G.sub.x] + [[summation].sub.l] [[gamma].sub.j]
[G.sub.ej] ... (3)
where:
[G.sub.q] = ln([Q.sub.t]) - ln([Q.sub.t-1]) (growth rate of output)
[G.sub.a] = ln([A.sub.t]) - ln([A.sub.t-1]) (growth rate of total
factor productivity)
[G.sub.k] = ln([K.sub.t]) - ln([K.sub.t-1]) (growth rate of capital
stock)
[G.sub.l] = ln([L.sub.t]) - ln([L.sub.t-1]) (growth rate of labour
force)
[G.sub.h] = ln([H.sub.t]) - ln([H.sub.t-1]) (growth rate of human
capital)
[G.sub.x] = ln([X.sub.t]) - ln([X.sub.t-1]) (growth rate of
exports)
[G.sub.ej] = ln([E.sub.t]) - ln([E.sub.t-1])), (growth rate of jth
energy source).
Capital and Labour are the primary factors of production. The
economic theory suggests that rise in capital and labour affects
economic growth positively and significantly. Vast literature on
endogenous growth models highlights the significance of human capital
also for economic growth suggesting that increase in human capital
increases output [see for example, Barro and Sala-i-Martin (1995)]. The
exports are included to capture the effect of external demand or the
changes in external environment or openness on domestic economy. The
theoretical and empirical literature suggests positive impact of export
expansion on domestic output [see Dornbusch (1992) and Todaro (2000)].
Although mainstream growth theory focuses on primary factors of
production only, after the energy crisis of 1970s, the role of energy in
economic growth became critical issue. For example, Moroney (1992)
attributes the decline in productivity in United Stated to changes in
energy market. The evidence for China shows that energy played a
critical role in her economic growth particularly due to changes in
production structure. The energy-growth relationship is also affected
due to application of better and efficient fuel and changes in
production technology [see Chandler and Gwin (2003) and Stern and
Cleveland (2003)]. Therefore, energy growth is categorised as critical
factor for economic growth. In order to examine the causality between
economic growth and energy use, we have divided the analysis in two
parts:
(1) Energy-Output relationship.
(2) Productivity Analysis where we focus on relationship between
energy use per unit of labour and output per unit of labour. For this
purpose, the model in Equation (1) is expressed in terms of labour. (18)
3(b) Estimation
Estimation based on time series data requires special attention. It
is well documented that application of standard estimation techniques on
non-stationary time series data can cause spurious correlation. This can
lead to wrong policy implications and incorrect forecasting. However, if
the data are not stationary then after appropriate adjustments Vector
Auto Regression (VAR), Error Correction Model (ECM) or Autoregressive Distributed Lag Model or other models may be applied [see Gujarati
(1995) and Maddala and Kim (1998)]. Therefore, for time series model
estimation involves two steps:
(i) Test for Data Stationarity and Causality
In order to test stationarity of each series, we first apply the
unit root test. We have selected Augmented Dicky Fuller (ADF) and
Phillips-Perron (PP) tests for this study. (19) The tests are applied to
the data at levels and first difference.
After testing for stationarity of each series, we examine the
direction of causality by applying Hsiao's Granger Causality Test.
For the test model is specified as:
[Y.sub.t] = [[summation].sub.i=1] [a.sub.i] [Y.sub.t-i] +
[[summation].sub.i=1] [b.sub.i] [X.sub.t-i]
[Xj.sub.t] = [[summation].sub.i=1] [c.sub.i] [Y.sub.t-i] +
[[summation].sub.i=1] [dj.sub.i] [Xj.sub.t-i] ... (4)
where Y is output, Xj is the set of explanatory variables
representing capital, labour, human capital, exports and energy sources,
i.e., electricity, natural gas and petroleum products. If
'[summation].[b.sub.i]' and '[summation].[c.sub.i]'
are not statistically significant then the explanatory variable do not
affect output and changes in output do not affect the explanatory
variables. If only one set of coefficients is significant then it is
concluded that there is unidirectional causality. We apply F-test and
the null hypotheses are: [H.sub.0]: [summation][b.sub.i] = 0 and
[H.sub.0]: [summation][cj.sub.i], = 0 against the alternate hypotheses
in each case. If the estimated F-value is greater than critical F-value
in both cases we conclude that Y and X cause each other and there is
bi-directional causality between economic growth and energy. If one of
the null hypothesis is not rejected then we have uni-directional
causality and if both are not rejected then there is no causal
relationship between changes in output and energy. (20)
(ii) Estimation of the Model
After applying unit root test, and determining causality we
estimate the casual relationship between economic growth and energy
controlling for changes in capital stock, labour, human capital and
exports. After the test for stationarity and causality, we apply
Autoregressive Distributed Lag (ADL) model for estimating the regression
model. (21)
The model is specified as:
[Y.sub.t] = [a.sub.0] + [summation] [a.sub.i] [Y.sub.t i] +
[summation] [summation] [b.sub.ji] [X.sub.t i] + [[epsilon].sub.t] ...
(5)
It is dynamic linear regression model. The basic issue is to decide
about the size of the lag for dependent and explanatory variables. The
standard assumptions for the model include that residual is random and
explanatory variables are uncorrelated with the error term
([[epsilon].sub.t]) and the forecasts are consistent with the theory.
3(c) Data
For estimation of the model, we use time series data, for the
period 1970-2003. The main sources of data are various issues of Energy
Year Book (various issues) and Pakistan Economic Survey (various
issues). (22) The variables are defined as follows:
(1) Gross domestic product (GDP), at factor cost, is used as
measure of output. The data series is at constant prices of 1980-81.
(2) The data on primary factors of production and other sources of
economic growth, including the various measures of energy include:
(a) Capital: Capital stock data are not available from secondary
sources. The series is constructed on the basis of available series on
domestic capital formation. The series for domestic capital formation is
available from 1959-60 to 2002-03. Assuming initial capital-output was
equal to 1.16 in 1959-60, we estimate capital stock for 1959-60 [see
Kemal and Ahmad (1992)]. (23) For the subsequent years the capital stock
series is computed using the following formula:
[K.sub.t] =(1-[delta]) [K.sub.t-1] + [I.sub.t]
where:
[K.sub.t] = capital stock in period t;
[delta] = rate of depreciation, assumed to be constant;
[K.sub.t] 1 = capital stock in year t-l; and
[I.sub.t] = Gross fixed capital formation in year t.
Since the study covers the period from 1969-70 to 2002-03, the
impact of assuming constant capital-output ratio in 1959-60 is expected
to be significantly lower in 1960-70 and afterwards.
(a) The labour force data are taken from the Pakistan Economic
Survey. It is in million of persons in the work force. (24)
(b) Human capital also affects economic growth significantly. We
have included five years lagged enrolment at the secondary level, as a
proxy for human capital.
(c) Exports of goods and services, at constant prices of 1980-81,
is included as a proxy for openness of the economy. Export expansion is
expected to have positive effect on economic growth.
(d) Energy sector is divided in three components: electricity,
natural gas and petroleum products. The electricity consumption is
measured in Gwh, gas consumption is measured in MCFT and consumption of
petroleum is measured in tonnes. (25) Since the commercial use of energy
is expected to affect economic growth, we have excluded the household
demand from total demand for each energy source.
As mentioned earlier, the sectoral decomposition of energy
consumption shows varying share of household over time and it differs
across energy sources.
Furthermore, since the household consumption of energy is not
expected to affect over all economic performance, we exclude domestic
consumption from each energy source and include only commercial use of
energy in the model.
For estimation, the variables are at constant prices of 1980-81 and
expressed either in natural-log form and in first difference of
log-form, i.e., in growth rates terms.
4. RESULTS
The discussion of the results is divided into two parts.
4(a) Output Growth Model
The results for unit root test are reported in Table 6. The tests
are applied to the level and fist difference of the data series. Both
ADF and PP statistics show that all the variables are stationary at
first difference implying the data series for growth rates are
stationary. Thus, we estimate the first difference model.
The results of causality test are reported in Table 7. (26) The
results show that growth in capital stock, in electricity consumption
and in petroleum products affects economic growth significantly. (27)
For the natural gas the effect of rise in consumption does not affect
economic growth.
The estimated results of the ADL model are reported in Table 8. The
results of five different equations are reported. The adjusted R-squared
and F-test show that the model is a good fit. The reported h-statistics
lies in the range of -1.96 and 1.96, therefore we can not reject the
null hypothesis of no autocorrelation. The results show that growth of
capital stock affects output growth positively and significantly. The
effect of labour growth is negative in Equation 1 and positive in
Equation 5 but, in both equations, it is statistically insignificant.
Surprisingly, the effect of 'education', the indicator of
human capital, is though positive but statistically insignificant
indicating no significant effect of secondary education on economic
growth. The growth rate of exports contributes to output growth
positively and significantly, as expected. The growth in total energy
use does not affect economic growth but the results change when we
decompose energy by its components. The impact of growth rate of
electricity and petroleum products on output growth is positive and
statistically significant. However, gas consumption does not affect
economic growth significantly. Interestingly, inclusion of growth of
different sources of energy, separately, in the equation reduces the
coefficient of capital stock significantly. This may be an indicator of
substitutability between energy inputs and capital or changing energy
intensity with changes in capital stock. (28)
4(b) Productivity Growth Model
In the second part, we estimate the productivity growth model. The
results for unit root test, reported in Table 9, show that first
difference series are stationary at 5 percent. The results for level,
not reported here, show that the data series are not stationary at
level. Therefore, the first difference model is estimated where the
first difference in dependent variable becomes growth in labour
productivity. The adjusted R-squared and F-test show that the model is a
fit. The reported h-statistics indicates that null hypothesis of no
autocorrelation can not be rejected.
The results for causality, reported in Table 10, show that there is
a long run relationship between growth in capital-per-worker, and
productivity growth. Except for growth in human capital and consumption
of natural gas, all other variables are important and statistically
significant determinants of productivity growth.
The estimated productivity model shows that growth in
capital-labour ratio is the major determinant of productivity growth.
For the remaining variables also, the results are similar to the
output-growth model. Export growth contributes to productivity growth
positively. (29) Growth in electricity use per worker and petroleum
products use per worker also contributes to productivity growth
significantly. Interestingly, like the earlier model, the inclusion of
energy in the growth model reduces the magnitude of the coefficient of
growth of capital-labour ratio whereas the magnitude of the coefficient
of export growth is robust. If the constant represents the impact of
changes in technology, the coefficient represent insignificant impact on
productivity growth, but the size of the coefficient is sensitive to the
inclusion/exclusion of indicators of energy.
These results indicate that capital stock is the most important
determinant of economic growth. However, the coefficient is sensitive to
inclusion/exclusion of the growth rate of energy sources, like
electricity and petroleum products. This result may be an indicator of
interrelationship between energy use and use of capital stock.
Surprisingly, the effect of growth of human capital is not statistically
significant. The impact of export growth is positive and the coefficient
is robust, indicating that external economic environment or openness
plays a critical role in domestic economic expansion. The results show
that energy is an important contributor to productivity growth.
Keeping in view these conclusions, can we determine the required
growth rates of capital and energy input for achieving target growth
rate of output. For example, what will be the required growth rate for
these inputs for achieving target growth rate of GDP of 5 percent (low),
7 percent (medium) and 10 percent (high)? Utilising the estimates of
Equation 1 and Equation 5, reported in Table 8, and assuming that
changes in growth rates of all the other variables are in line with
target growth rate of GDP, we compute the required growth rates for
capital stock, for electricity use and for use of petroleum products
(see Table 12). (30) The results show that substantially higher growth
rate in capital stock, electricity and petroleum products would be
required to achieve even modest economic growth rate of 5 percent. For
achieving GDP growth rate greater then 5 percent targeted efforts will
be required to increase capital stock and energy, i.e., electricity and
petroleum products. Furthermore, in order to achieve target GDP growth
rate of 7 percent the employment growth rate should exceed four percent
and exports growth rate should be more than 7 percent. (31) If the
growth rates of all other explanatory variables falls behind then
achievement of target GDP growth will become even more difficult. Thus,
for the achievement of target GDP growth rates policy efforts are
required in various dimensions.
5. CONCLUSIONS, POLICY IMPLICATIONS, AND FUTURE DIRECTIONS
The issue of energy supply and demand is important not only for the
economic prosperity of the current generations but also for the future
generations. From the above analysis, we can see that energy is a
critical determinant of economic growth. Therefore, its shortage can
retard economic growth. However, in order to achieve high economic
growth rates, multidimensional policies are required and these policies
should not ignore the energy sector. In order to improve availability of
energy and balance of payment position, alternative sources of energy
should also be developed. Based on the discussion above, we can outline
following policy implications and areas for future research:
(1) The rise in supply of energy at affordable prices is important
for economic growth. Deregulation will have important implication for
pricing behaviour of the various sources of energy. The rise in prices
affects the demand and consequently the economic growth. Thus, the
pricing policies should take into account the impact on economic growth
also.
(2) The issue of renewable and non renewable sources of energy,
demand and supply of each component of energy, intensity and efficiency
of energy use, availability of substitutes, pricing mechanism and
balance of payment implications of energy use are important and the
issue should be examined in detail.
(3) The poverty reduction strategy should have clear strategy for
energy sector. Lamech and O'Sullivan (2002) suggest that energy
plays an important role in reducing poverty. However, the poverty
reduction strategies should emphasise on expanding the access to energy,
improve reliability and achieve fiscal sustainability by reducing the
claims of the sector on budget, reduce fiscal risk due to the energy
sector, eliminate subsidies to reduce government liabilities, improve
governance and ensure environmental sustainability. These efforts should
be accompanied by a set of monitoring indicators like availability and
affordability of energy related equipment, fiscal discipline for energy
utilities and regulatory framework. The poverty reduction strategy of
Pakistan concentrates on energy sector also. [see Pakistan (2001)].
However, the emphasis of the policy is on deregulation and
privatisation. As we have seen earlier that it has resulted in higher
prices. The privatisation process in Pakistan has resulted in loss of
jobs, at least in the short run [see Kemal (1999)]. This will impact the
desired outcomes of the poverty reduction strategy. Thus, there is a
need, at least in the short run, to develop a mechanism to mitigate the
adverse impact of deregulation of energy prices on the poor and
unemployed.
(4) In the recent decade the issues of energy conservation, its
pricing and the impact on environment have raised concerns in the
developed and developing countries. In Pakistan, crisis of energy can be
termed not only the supply issue but also the "Crisis of Energy
Pricing". The crises is not a result of only the mismatch in demand
and supply of energy, but an outcome of imbalance in government
policies. Since a significant proportion of energy is imported
particularly the petroleum and petroleum products, a sharp increase in
price of energy also indicate deterioration in the purchasing power of
our export earnings in Pakistan. For example, in 1971-72, I metric ton
of raw cotton export could buy 64 metric tons of crude oil in 1984-85,
but in 1999, it could buy only 14.79 MT of crude oil. This increase in
prices affects not only the economic growth but also the balance of
payments.
(5) The discussion in Section 2 indicates the possibility of
inter-fuel substitution which may be result of changes in price
structure resulting in changes in production technologies and/or changes
in production structure. It will be interesting to examine these issues,
in detail.
Author's Note: An earlier version of the paper was presented
at the Annual General Meeting of Pakistan Society of Development
Economists, in 2002, in Islamabad, Pakistan. The author is grateful to
Dr A. R. Kemal, Director, PIDE, for his useful comments and suggestions.
The comments of the participants of the PSDE meeting in Islamabad and
invaluable suggestions of an anonymous referee have helped to improve
the paper.
REFERENCES
Aqeel, A., and M. S. Butt (2001) The Relationship between Energy
Consumption and Economic Growth in Pakistan. Asia-Pacific Development
Journal 8:2 (December), 101-110.
Asafu-Adjaye (2000) The Relationship between Energy Consumption,
Energy Prices and Economic Growth: Time Series Evidence from Asain
Developing Countries. Energy Economics 22:6 (December), 615-626.
Banerjee, A., J. Dolando, J. Galbraith, and D. F. Hendry (1993)
Co-integration, Error Correction, and the Econometric Analysis of
Non-stationary Data. Oxford University Press.
Barro, R. J., and X. Sala-i-Martin (1995) Economic Growth. New
York: McGraw Hill, International Edition.
Bashar, Amanullah (2000) Pakistan Energy Conference--Special
Report. Pakistan and Gulf Economist. April 24-30.
Chandler, W., and H. Gwin (2003) Helping China Manage Energy
Growth. [downloaded from internet].
Dornbusch, R. (1992) The Case for Trade Liberalisation in
Developing Countries. Journal of Economic Perspectives 6, (Winter),
69-85.
Ebinger, C. K. (1981) Pakistan: Energy Planning in a Strategic
Vortex. Bloomington, USA: Indiana University Press.
Gujarati, D. N. (1995) Basic Econometrics. (Chapters 17, 21 and
22), (3rd edition). New York: McGraw Hill, International Edition, USA.
Kazmi, S. (2000) PEPCO. Pakistan and Gulf Economist 19:18, May.
Kemal, A. R. and S. I. Ahmad (1992) The Sources of Economic Growth
in Pakistan. Study prepared for the Ninth Five-Year Plan. Pakistan
Institute of Development Economics, Islamabad. (May).
Kemal, A. R. (1999) Privatisation in Pakistan--Social Effects and
Restructuring. Pakistan Institute of Development Economics, Islamabad.
(September). (Unpublished paper.)
Lamech, R., and K. O'Sullivan (2002) Energy (Chapter 21 and
Technical Note 'Q') In Jeni Klugman (ed.) A Sourcebook for
Poverty Reduction Strategies: Vol 2: Macroeconomic and Sectoral
Approaches. Washington, D. C.: The World Bank.
Maddala, G. S., and In-Moo Kim (1998) Unit Roots. Cointegration,
and Structural Change. (Chapter 2). Cambridge: Cambridge University
Press, USA.
Mahmud, S. F. (2000) The Energy Demand in the Manufacturing Sector
of Pakistan: Some Further Results. Energy Economics 22:6 (December),
641-648.
Moroney, J. R. (1992) Energy, Capital and Technological Change in
the United States. Resource and Energy 14:4 (December), 363-380.
Pakistan, Government of (2001) Energy Year Book. Islamabad:
Government of Pakistan.
Pakistan, Government of (2001) Interim Poverty Reduction Strategy
of Pakistan. Islamabad: Government of Pakistan.
Pakistan, Government of (Various Issues) Pakistan Economic Survey.
Islamabad: Ministry of Finance, Government of Pakistan.
Riaz, T. (1984) Pakistan: The Energy Sector--A Study in Sector
Planning. Lahore: Ferozesons Limited, Pakistan.
Siddiqui, R. (1999) Demand for Energy and the Revenue Impact of
Changes in Energy Prices. Pakistan Institute of Development Economics,
Islamabad. (Research Report No. 174.)
Stern, D. I., and C. J. Cleveland (2003) Energy and Economic
Growth. [downloaded from internet]. December.
Todaro, M. P. (2000) Economic Development. (Seventh Edition). New
York: Addison-Wesley.
World Bank (2002) Pakistan: Country Assistance Strategy FY03-05.
Washington, D. C.: The World Bank.
(1) See, for example, Stern and Cleveland (2003): Asafu-Adjaye
(2000); Aqeel and Butt (2001): Mahmud (2000): Riaz (1984) and Siddiqui
(1999).
(2) See Ebinger (1981).
(3) Increased use of better energy sources saves time, helps people
in improving the quality of life and environment. It improves the social
services delivery, like effective utilisation of modern health related
equipment and better facilities in educational institutions and others.
The availability of modern and better fuels improves the lives of the
females and children who spend time on collection of traditional fuel
like wood. Therefore, it can be claimed that energy sector and the
energy services have important implications for poverty reduction. For
example, in the small scale industry, mostly concentrated in rural
areas, electricity provision can increase the length of working day and
increase the productivity of the resources. Furthermore, energy sector
itself is an important source of employment generation. Thus, use of
energy as an input is expected to have direct positive impact on output
and indirect positive impact on poverty and quality of life through the
employment generation.
(4) Energy is also an important source of government revenue that
can improve the fiscal deficit in the country, and can also result in
increased availability of tile resources for public investment.
(5) Due to data availability for net supply and net demand, the
discussion in this section covers the period from 1991 to 2003 only.
(6) A simple correlation coefficient between growth rate of various
sources of energy and output growth rate, in Pakistan, varies between
0.612 (for electricity), 0.437 (for petroleum products) and almost
negligible for natural gas.
(7) Non-commercial sources of energy, viz., wood, cotton sticks,
begasses, and crop roots, are less economical and inefficient because of
lower energy content and higher energy losses. Due to non availability
of data we can not include these sources in the analysis
(8) See Table 3, Table 4 and Table 5, discussed later.
(9) According to Riaz (1984), present energy pricing structure does
not provide incentives to improve efficiency.
(10) Initially, the prices of the major petroleum products were
fixed by the Ministry of Petroleum. Government of Pakistan. Later prices
were linked to the Singapore Mean FOB value of the petroleum products
and then linked to the average price of the petroleum products during
the last quarter in the Arabian-Gulf countries. The variation in prices
of petroleum products from different sources were reflected in variation
in development surcharge. In 1999, fixed development surcharge and sales
tax was imposed on petroleum products that has resulted in sharp rise in
prices of petroleum products. The price of furnace oil was deregulated
from I st July 2000 which resulted in almost doubling of the price.
(11) On average, during 1990s, the share of different sectors in
consumption of petroleum products was: 4.3 percent households, 12.7
percent industry, 1.9 percent agriculture, 47.5 percent for transport.
31 percent for power and 2.6 percent for other government.
(12) The producer (well-head) price and consumer price are still
determined by the Government of Pakistan (GOP). Gas prices which were
initially linked with the energy pricing policies at the time of
commercialisation of the well, are now linked with the international
fuel prices. This adds variability to gas price due to variability of
the real exchange rate. Currently, the government of Pakistan regulates
all gas prices, both for producers (at the well-head) and for consumers.
The gas price is fixed neither on economical nor on financial basis.
Increase in consumer gas prices has not kept pace with the increase in
average purchase prices implying that the gap between revenue for gas
companies and average gas prices at the well-head has declined over
time. The difference between the two prices is used to meet operating
expenses of the companies that are ensured a fixed rate of return, the
balance reverts to the GOP and is distributed among the provinces as Gas
Development Surcharge.
(13) For allocation of gas the sectoral priority is: residential
and commercial users; feedstock in fertiliser industry: replacement of
HSD in power generation; general industry- and replacement of furnace
oil in power generation. In order to reduce pressure on the natural gas
during the peak period of domestic demand for natural gas. The gas
distribution companies started regulating the supply of gas. For
example, to some industrial units, gas is supplied on a nine-months
supply basis SNGPL adopted the policy of nine-months supply to all new
industrial units in 1988. Later on same policy was adopted by SSGC. For
fertiliser plants the gas supply is stopped from December to February.
These units have to use alternative fuel during this period and it
results in higher cost of production. For Cement industry also. the
supply of gas is curtailed.
(14) The controversy over the tariff rates between IPPs and WAPDA
resulted in long conflict. The private power producer claim that high
oil prices, in Pakistan. are responsible for high electricity tariff
charged from WAPDA and KESC. The guaranteed 18 percent rate of return to
IPPs has created financial problems for WAPDA and pats a heavy burden on
future generations.
(15) According to Bashar (2000), Chairman WAPDA claimed that:
"The capacity has increased, beyond demand growth, but it is an
irony that electric power is not affordable for a majority of the
population and has become an increasingly non-economic input for
industry".
(16) The growth rate of electricity prices was, on average, 12
percent per annum during 1993-2003.
(17) The issues like pricing and inter-fuel substitution are
discussed partially in Siddiqui (1999) and Mahmud (2000).
(18) In the productivity model, all the variables and growth rates
are in terms of per unit of labour.
(19) For details of the tests, see Banerjee, et al. (1993) and
Maddala and Kim (1998).
(20) For details of the test, see Gujarati (1995).
(21) According to Maddala and Kim (1998). the ADL model, Vector
Autoregression (VAR) model and Error Correction Model (ECM) give similar
results [see Chapter 2, for detail, in Maddala and Kim (1998)].
(22) Both are published by the Government of Pakistan.
(23) The value of capital-output ratio is reported in Kemal and
Ahmad (1992) They report capital-output equal to 1.16 in 1959-60.
(24) Since capital stock and labour force are the primary factors
of production. The growth in services of these factors are assumed to be
proportional to the growth in the physical units of these services.
(25) The data on energy, measured in terms of tonnees of oil
equivalent (TOE) are also available. However, due to change ill
methodology around 1988 the data for different sources of energy,
measured in TOE. shows sharp changes in 1988. Therefore, in estimation
of the model we do not include the different components of energy,
measured in TOE.
(26) Here we report only the results for causality from explanatory
variables to growth, The results for the reverse causality are not
reported. However, the results show that for exports, petroleum products
and human capital reverse causality holds. Similarly for the
relationship within the energy group we see that gas consumption affects
consumption of electricity and petroleum products significantly only in
the short run. This is a surprising result, given the substitutability
between gas (CNG) and gasoline. We intend to explore the issue of
bi-directional in a later study.
(27) We estimated the models for one to eight years lags, however,
we report only the results for two years lag model for two reasons.
First, the main conclusions do not change and secondly, the R-sqaured
and F-ratio for 2-years lag model is higher.
(28) Three interactive dummy variables with petroleum products were
included to capture the effect of petroleum policies of 1990, 1994 and
1997. However, the estimated coefficient was statistically insignificant
in each case.
(29) The Granger Causality test shows that reverse causality also
holds.
(30) For different growth scenarios, the growth rates of other
explanatory variables are expected to change in the same proportion as
the ratio of target GDP growth rate and average GDP growth rate. The
growth rates, for the period examined in the study were: 2.64 percent
per annum for labour three, 5.00 percent per annum for exports. 4.89
percent per annum for one year lagged GDP-growth and 4.98 percent per
annum for 2 years lagged GDP-growth. For different growth scenarios, the
growth rates of other explanatory variables are expected to change in
the same proportion as the ratio of target growth rate and average
growth rate.
(31) In order to achieve target growth rate of 10 percent,
employment should increase by more then 5 percent and export should
increase by 10 percent per annum.
Rehana Siddiqui is Chief of Research at the Pakistan Institute of
Development Economics, Islamabad.
Table 1
Coefficient of Energy Intensity (1971-2003)
Electricity Natural Gas
Growth Rate of (Gwh) (mcft)
Gross Domestic
Product E1 E2 G1 G2
1970-1980 4.63 1.63 1.29 1.84 1.74
1980-1990 6.12 1.76 1.52 1.12 1.00
1990-2000 4.40 1.07 0.49 1.11 0.93
2000-2003 3.34 1.47 1.90 2.05 2.31
1970-2003 4.89 1.51 1.19 1.38 1.27
Petroleum Products Total
(Tonnes) Energy
P1 P2 T1
1970-1980 1.05 1.04 1.29
1980-1990 1.50 1.58 0.74
1990-2000 1.35 1.57 1.04
2000-2003 -0.76 -0.66 0.65
1970-2003 1.18 1.27 0.93
Notes: Energy intensity is defined as the ratio of growth rate of
energy to growth rate of output.
E1=Coefficient ofenergy intensity for total electricity use.
E2=Coefficient ofenergy intensity for electricity use (excluding
household use).
G1=Coefficient ofenergy intensity for total natural gas.
G2=Coefficient ofenergy intensity for natural gas (excluding
household use).
P1=Coefficient ofenergy intensity for total petroleum products.
P2=Coefficient ofenergy intensity for petroleum products
(excluding household use).
T1=Coefficient oftotal energy intensity. (Total Energy is in TOE
units).
Table 2
Import of Petroleum Products (1990-91-2002-03
Import of Petroleum Total Imports (at
Products (at Constant Constant Prices
Prices 1990-91=100) 1990-91=100)
Rs Million Growth Rs Million Growth
Rate (%) Rate (%)
1990-1991 37823 -- 171114 --
1995-1996 60829 9.50 214349 4.51
1999-2000 70401 3.65 206073 -0.98
2002-2003 60336 -5.14 230800 3.78
Total Imports Excluding
Import of Petroleum
Products (at Constant Share of
Prices 1990-91=100) Petroleum
Products in
Rs Million Growth Total
Rate (%) Imports (%)
1990-1991 133291 -- 22.10
1995-1996 153520 5.83 28.38
1999-2000 135672 -3.09 34.16
2002-2003 170465 7.61 26.14
Source: Pakistan (Various Issues) Pakistan Economic Survey.
Table 3
Sectoral Share in Consumption of Petroleum Products (1991-2003)
Households Industry Agriculture Transport Power
1990-91 9.5 11.5 2.7 48.6 24.4
1998-99 2.9 12.8 1.5 47.2 33.2
2002-03 1.72 9.75 1.2 49.1 36.6
Sources: Pakistan (Various Issues) Pakistan Energy, Yearbook.
Table 4
Sectoral Share in Consumption of Natural Gas
Household Commercial Cement
1990-91 14.3 2.6 2.8
1998-99 20.7 3.4 1.2
2002-03 17.6 2.6 0.4
Fertiliser Power Industry
1990-91 23.2 27.9 19.1
1998-99 26.3 28.9 19.1
2002-03 16.2 38.5 19.9
Sources: Pakistan (Various Issues) Pakistan Energy Yearbook.
Table 5
Sectoral Share in Consumption of Electricity (1991-2003)
Household Commercial Industry Agriculture
1990-91 33.0 6.8 35.6 17.8
1998-99 44.8 5.6 27.9 13.0
2002-03 44.9 6.1 30.7 11.4
Sources: Pakistan (Various Issues) Pakistan Energy Yearbook.
Table 6
Results for Stationarity--Unit Root Test
Augmented Dicky Fuller Test
Level 1st Difference
Gross Domestic Product -0.171 -4.397
Capital Stock -1.341 -3.662
Labour -1.882 -4.615
Exports -2.319 -4.163
Education -1.445 -3.942
Electricity -0.437 -4.026
Natural Gas -1.756 -4.020
Petroleum Products -0.440 -4.116
Critical Value (at 5%) (-3.561) (-3.561)
Phillip-Perron Test
Level 1st Difference
Gross Domestic Product -0.615 -5.033
Capital Stock -2.653 -3.662
Labour -2.250 -6.675
Exports -1.520 -6.371
Education -2.487 -4.621
Electricity -0.621 -4.835
Natural Gas -2.110 -9.020
Petroleum Products -0.997 -4.423
Critical Value (at 5%) (-3.551) (-3.551)
Note: Since each variable is measured in natural log. 1st difference
series represent annul average growth rate for each variable.
Table 7
Granger Test for Causality
Null Hypotheses F-Value Probability Decision
Growth in capital stock does
not cause growth 10.85 0.0004 Rejected
Growth in labour force does not
cause growth 0.21 0.813 Not rejected
Growth in exports does not
cause growth 2.53 0.051 Rejected
Growth in human capital does
not cause growth 1.36 0.280 Not rejected
Growth in electricity use does
not cause growth 2.99 0.070 Rejected
Growth in use of natural gas
does not cause growth 0.09 0.940 Not rejected
Growth in use of petroleum
products does not cause
growth 3.36 0.070 Rejected
Note: Causality in reverse direction was also tested and null
hypotheses were rejected in most cases except for electricity and
petroleum products.
Table 8
Estimated Regression Model--Autoregressive Distributed Lag Model
Equation 1 Equation 2 Equation 3
Constant 0.944 0.758 0.79
(0.57) (0.394) (0.40)
Growth Rate of Capital 0.839 0.849 0.84
Stock-Gk (2.73) (2.68) (2.58)
Growth Rate of Labour -0.07 -0.063 -0.068
Force-GI (0.38) (0.31) (0.316)
Growth Rate of 0.058 0.058 0.057
Exports-Gx (1.96) (1.884) (1.89)
Growth Rate of Human 0.014 0.012
Capital-Gedu -- (0.020) (0.16)
Growth Rate of
Electricity-Ge -- -- --
Growth Rate of Natural
Gas-Gg -- -- --
Growth Rate of Petro.
Products-Gp -- -- --
Growth Rate of Total 0.008
Energy-Gte -- -- (0.16)
Y(t-1) -0.045 -0.045 -0.047
(0.35) (0.246) (0.27)
Y(t-2) 0.026 -0.022 -0.023
(0.169) (0.133) (0.14)
Adjusted R-squared 0.21 0.18 0.17
F-value 2.62 2.08 2.20
h-statistics -1.571 -1.468 -1.524
N 31 31 31
Equation 4 Equation 5
Constant 3.325 2.251
(2.55) (1.586)
Growth Rate of Capital 0.301 0.42
Stock-Gk (2.117) (2.469)
Growth Rate of Labour 0.152
Force-GI -- (0.887)
Growth Rate of 0.03 0.029
Exports-Gx (1.94) (2.24)
Growth Rate of Human 0.05
Capital-Gedu (0.89) --
Growth Rate of 0.186 0.153
Electricity-Ge (3.049) (2.236)
Growth Rate of Natural -0.10
Gas-Gg (0.837) --
Growth Rate of Petro. 0.198 0.168
Products-Gp (2.896) (2.491)
Growth Rate of Total
Energy-Gte -- --
Y(t-1) -0.159 -0.112
(1.162) (0.722)
Y(t-2) -0.227 -0.262
(1.662) (1.688)
Adjusted R-squared 0.569 0.52
F-value 5.945 4.536
h-statistics -1.515 -1.50
N 31 31
Note: t-values are reported in parentheses.
Y(t-1) and Y(t-2) are lagged values of dependent variable.
Table 9
Test for Stationarity: Unit Root Test
Augmented Dicky Fuller Phillip-Perron
Test--1st Difference Test--1st Difference
Gross Domestic Product -3.887 -5.182
Capital Stock -3.810 -5.241
Exports -3.927 -6.183
Education -4.814 -5.225
Electricity -3.74 -5.383
Natural Gas -4.293 -7.976
Petroleum Products -3.817 -4.632
Critical Value (at 5 %) (-3.556) (-3.556)
Note: All the variables are divided by labour force. Since each
variable is measured in natural log. 1st difference series represent
annual average growth rate for each variable.
Table 10
Granger Test for Causality-Productivity Growth
Null Hypotheses F-Value Probability Decision
Growth in capital-labour ratio
does not cause productivity
growth 2.10 0.048 Rejected
Growth in exports per unit of
labour does not cause
productivity growth 2.42 0.109 Rejected
Growth in human capital does
not cause productivity growth 0.19 0.82 Not rejected
Growth in electricity per unit
of labour does not cause
productivity growth 3.84 0.035 Rejected
Growth in natural gas per unit
of labour does not cause
productivity growth 0.23 0.80 Not rejected
Growth in petroleum products
per unit of labour does not
cause productivity growth 2.64 0.82 Rejected
Note: Causality in reverse direction was also tested and null hypotheses
were rejected in most cases except for electricity and petroleum products.
Table 11
Estimated Result for Productivity Growth--Autoregressive
Distributed Lag Model
Equation 1 Equation 2 Equation 3
Constant -0.154 -0.456 -0.401
(0.223) (0.647) (0.558)
Growth Rate of Capital-GkL 0.772 0.687 0.664
(4.369) (5.784) (3.545)
Growth Rate of Exports-GxL 0.068 0.076 0.074
(2.12) (2.394) (2.303)
Growth Rate of Human
Capital-GeduL -- 0.111 0.100
(1.51) (1.312)
Growth Rate of Electricity-GeL -- -- --
Growth Rate of Natural Gas-GgL -- -- --
Growth Rate of Petro.
Products-GpL -- -- --
Growth Rate of Total
Energy-GteL -- -- 0.04
(0.624)
YL(t-1) 0.041 -0.007 -0.026
(0.306) (0.056) (0.189)
YL(t-2) -0.052 -0.095 -0.102
(0.341) (0.712) (0.751)
Adjusted R-squared 0.482 0.506 0.494
F-value 7.968 7.144 5.87
h-statistics -1.57 -2.00 -1.53
N 31 31 31
Equation 4 Equation 5
Constant 0.617 0.557
(1.285) (1.002)
Growth Rate of Capital-GkL 0.448 0.423
(1.839) (2.693)
Growth Rate of Exports-GxL 0.02 0.023
-1.76 (1.840)
Growth Rate of Human
Capital-GeduL 0.103 0.048
(1.863) (0.793)
Growth Rate of Electricity-GeL 0.291 0.18
(4.412) (3.165)
Growth Rate of Natural Gas-GgL -0.112 --
(0.973)
Growth Rate of Petro.
Products-GpL 0.199 0.171
(3.344) (2.515)
Growth Rate of Total
Energy-GteL -- --
YL(t-1) -0.033 -0.005
(0.385) (0.793)
YL(t-2) -0.236 -0.192
(0.389) (0.702)
Adjusted R-squared 0.810 0.7745
F-value 16.94 13.496
h-statistics -1.52 -2.00
N 31 31
Note: t-values are reported in parentheses.
YL(t-1) and YL(t-2) are lagged values of dependent variable.
Table 12
Required Growth of Critical Sources of Economic Growth
Required
Required Growth Required Growth
Rate of Growth Rate of
Target Growth Capital Stock Rate of Petroleum
Rate of Gross Electricity Products
Domestic Product Equation 1 Equation 5 Equation 5 Equation 5
5 Percent 5.12 4.56 4.75 5.20
7 Percent 7.50 8.60 9.65 8.65
10 Percent 11.64 12.65 20.04 20.73
Actual Growth Rates of the Sources of Growth (1970-2003)
Min. Growth Rate 2.76 2.76 -7.73 -4.06
Max. Growth Rate 7.84 7.84 13.24 13.97
Avg. Growth Rate 5.51 5.51 5.67 6.04
Note: The equations are from Table 7.
Table A1
Growth Rates of Energy Use Indicators: 1990-97 (Percentages)
Commercial
Energy Use
Countries Total Per Capita
Bangladesh 2.53 0.60
(20936) (190)
China 4.26 2.92
(866666) (763)
Egypt 3.66 1.27
(31895) (608)
India 4.22 2.05
(359846) (424)
Indonesia 5.82 3.77
(98846) (555)
Korea 11.58 10.28
(91402) (2132)
Malaysia 12.45 9.23
(23874) (1317)
Nepal 3.47 0.53
(5834) (311)
Pakistan 4.66 1.68
(43238) (400)
Sri Lanka 4.57 3.07
(5476) (322)
Thailand 10.59 9.01
(43706) (786)
World 1.53 -0.13
(8508414) (1705)
Low Income Countries 1.04 -1.25
(1122683) (6.07)
Middle Income
Countries 1.11 -0.35
(3297830) (1397)
High Income Countries 1.99 1.21
(4167901) (4996)
GDP/ Per Net Energy
Unit of Imports
Energy Use (% of
(PPP $ kg Commercial
Countries of TOE) Use)
Bangladesh 5.26 0
(5.0) (10)
China 10.65 20.09
(1.8) (-3)
Egypt 3.00 -6.86
(3.9) (72)
India 4.10 9.40
(3.3) (7)
Indonesia 4.78 -2.3
(3.4) (-69)
Korea -0.42 2.08
(4.0) (76)
Malaysia 0 -10.63
(4.0) (104)
Nepal 4.76 17.76
(2.8) (3)
Pakistan 2.82 3.62
(3.3) (21)
Sri Lanka 3.45 9.20
(6.2) (23)
Thailand -0.69 -1.23
(4.9) (39)
World
-- --
Low Income Countries
-- --
Middle Income
Countries
-- --
High Income Countries
-- --
Carbon Dioxide Emissions
Countries Total Per Capta
Bangladesh 6.91 12.25
(15.4) (0.1)
China 5.77 4.91
(2401.7) (2.1)
Egypt 4.45 3.29
(75.4) (1.4)
India 2.20 3.40
(675.3) (0.9)
Indonesia 6.80 0
(165.2) (7.0)
Korea 9.16 8.23
(241.2) (5.6)
Malaysia 13.64 10.96
(55.3) (3.0)
Nepal 17.76 0
(0.6) (0.1)
Pakistan 5.63 4.91
(67.9) (0.6)
Sri Lanka 10.50 12.25
(3.9) (0.2)
Thailand 13.58 12.25
(95.7) (1.7)
World 5.97 3.26
(16183.1) (3.3)
Low Income Countries 9.96 7.82
(1376.8) (0.7)
Middle Income
Countries 8.70 5.39
(5772.8) (2.7)
High Income Countries 2.91 0.55
(9033.6) (11.9)
Electricity
Transmission
Electricity and Distribution
Consumption Per Loss (% of
Countries Capita (kwh) Total Output)
Bangladesh 9.86 -12.75
(43.0) (34)
China 7.18 2.25
(471.0) (7)
Egypt 2.39 0
(697.0) (12)
India 6.13 0
(254.0) (I8)
Indonesia 13.24 -3.65
(156.0) (I-5)
Korea 14.05 -3.65
(2202.0) (5)
Malaysia 13.59 -1.74
(1096.0) (10)
Nepal 5.68 -0.58
(28.0) (29)
Pakistan 3.75 2.25
(267.0) (21)
Sri Lanka 6.80 0
(153.0) (I7)
Thailand 11.97 -3.29
(690.0) (11)
World 1.05 0
(1928.0) (8)
Low Income Countries -0.73 4.57
(373.0) (I3)
Middle Income
Countries 1.26 3.40
(1243.0) (9)
High Income Countries 2.05 -2.54
(72941.0) (7)
Source: World Bank (2000) World Develoment Report 21AA1-111.
* Initial values of each variable are reported in parenthesis.
Table A2
Growth Rates of Net Supply and Net Consumption of Energy
Initial
Net Supply
and Demand
(1989-90) 1990- 1991- 1992-
(TOE) 91 92 93
(a) Oil Products
Net Supply 8065360 5.14 11.33 8.14
Net Consumption 8066499 -3.20 8.63 7.53
Difference -1139 -1.94 2.70 0.61
(b) Natural Gas
Net Supply 5137893 9.31 -7.89 12.00
Net Consumption 4720833 7.16 3.57 8.40
Difference 417060 2.14 -11.46 3.60
(c) LPG
Net Supply 137294 14.57 -11.90 3.04
Net Consumption 137294 14.57 11.90 3.04
Difference 0.00 0.00 0.00 0.11
(d) Coal
Net Supply 1388352 -2.41 16.90 -10.80
Net Consumption 1388352 -2.41 16.90 -10.80
Difference 0.00 0.00 0.00 0.00
(e) Electricity
Net Supply 2978866 8.16 10.23 7.37
Net Consumption 2342996 9.18 7.17 7.44
Difference 635870 -1.01 3.06 -0.07
(r) Total Energy
Net Supply 17707764 1.89 5.77 7.58
Net Consumption 16655974 1.85 7.45 6.25
Difference 1051790 0.05 -1.68 1.33
Giowth Rate of GDP 422484 5.60 7.70 2.10
1993- 1994- 1995- 1996-
94 95 96 97
(a) Oil Products
Net Supply 5.98 1.97 8.34 -1.07
Net Consumption 5.15 4.29 10.17 -2.87
Difference 0.83 -2.31 1.83 1.80
(b) Natural Gas
Net Supply -12.02 30.04 6.74 0.63
Net Consumption 4.64 5.53 8.73 -3.60
Difference -16.66 24.51 -1.99 4.23
(c) LPG
Net Supply -5.01 29.90 17.19 -9.15
Net Consumption -5.01 29.90 17.19 -9.15
Difference 0.00 0.00 0.01 0.00
(d) Coal
Net Supply 8.06 -15.07 7.59 -1.18
Net Consumption 8.06 -15.07 7.59 -1.17
Difference 0.00 0.00 0.00 -0.01
(e) Electricity
Net Supply 3.33 5.70 5.82 3.31
Net Consumption 2.40 5.38 5.64 2.31
Difference 0.93 0.35 0.18 1.00
(r) Total Energy
Net Supply 0.74 9.57 7.42 0.19
Net Consumption 4.73 3.67 8.96 -2.26
Difference -4.00 5.91 -1.55 -2.44
Giowth Rate of GDP 4.41 5.10 6.60 1.70
1997- 1998- 1999- 2000-
98 99 2000 01
(a) Oil Products
Net Supply 0.41 6.27 2.64 -1.06
Net Consumption 0.41 5.43 3.85 -3.17
Difference 0.00 0.84 -1.21 2.11
(b) Natural Gas
Net Supply 6.76 0.74 4.70 6.44
Net Consumption 9.11 2.49 7.87 0.61
Difference -2.35 -1.74 -3.16 5.83
(c) LPG
Net Supply 9.87 13.09 10.47 7.90
Net Consumption 9.87 13.12 10.47 7.90
Difference 0.00 0.00 0.00 0.00
(d) Coal
Net Supply -12.65 7.99 -7.72 2.42
Net Consumption -12.66 7.99 -7.72 2.42
Difference 0.0 0.11 0.00 0.00
(e) Electricity
Net Supply 5.32 5.41 -0.09 3.08
Net Consumption 3.79 -2.91 5.62 6.37
Difference 1.52 8.00 -5.71 3.29
(r) Total Energy
Net Supply 2.69 4.52 2.24 2.39
Net Consumption 2.88 3.48 4.74 -0.11
Difference -0.19 1.04 -2.49 2.50
Giowth Rate of GDP 0.33 4.20 3.90 1.80
2001- 2002-
02 03 Avg.
(a) Oil Products
Net Supply -6.89 -3.29 2.13
Net Consumption 4.39 -2.03 2.29
Difference 2.49 -1.26 -0.17
(b) Natural Gas
Net Supply 1.84 0.22 4.58
Net Consumption 5.10 6.19 5.06
Difference -3.26 -5.97 -0.48
(c) LPG
Net Supply 23.10 4.57 7.59
Net Consumption 19.58 3.77 7.26
Difference 3.53 0.80 0.33
(d) Coal
Net Supply 13.88 13.01 1.52
Net Consumption 13.88 13.01 1.52
Difference 0.00 0.00 0.00
(e) Electricity
Net Supply 6.52 4.50 5.28
Net Consumption 4.11 3.94 4.65
Difference 2.41 0.56 0.63
(r) Total Energy
Net Supply 0.03 0.53 3.50
Net Consumption 1.36 2.73 3.52
Difference -1.30 -2.20 0.00
Giowth Rate of GDP 3.10 5.10 4.215
Table A3
Transmission and Distribulion Losses--Eleciricity
Auxiliary Transmission and
Years Consumption Distribution Losses
1991-1992 2.43 21.68
1992-1993 2.23 21.06
1993-1994 2.62 21.59
1994-1995 2.60 21.45
1995-1996 2.93 21.50
1996-1997 2.41 21.70
1997-1998 2.07 23.97
1998-1999 1.74 25.80
1999-2000 2.09 24.18
2002-2003 2.10 23.80
Source: Pakistan (Various Issues).
