Energy demand in Pakistan: a disaggregate analysis.
Khan, Muhammad Arshad
|Publication:||Name: Pakistan Development Review Publisher: Pakistan Institute of Development Economics Audience: Academic Format: Magazine/Journal Subject: Business, international; Social sciences Copyright: COPYRIGHT 2008 Reproduced with permission of the Publications Division, Pakistan Institute of Development Economies, Islamabad, Pakistan. ISSN: 0030-9729|
|Issue:||Date: Winter, 2008 Source Volume: 47 Source Issue: 4|
|Product:||Product Code: 1312000 Natural Gas NAICS Code: 211111 Crude Petroleum and Natural Gas Extraction SIC Code: 1311 Crude petroleum and natural gas|
This study examines the demand for energy at disaggregate level
(gas, electricity and coal) for Pakistan over the period 1972-2007. Over
main results suggest that electricity and coal consumption responds
positively to changes in real income per capita and negatively to
changes in domestic price level. The gas consumption responds negatively
to real income and price changes in the short-run, however, in the
long-run real income exerts positive effect on gas consumption, while
domestic price remains insignificant. Furthermore, in the short-run the
average elasticities of price and real income for gas consumption (in
absolute terms) are greater than that of electricity and coal
consumption. The differences in elasticities of each component of energy
have significant policy implications for income and revenue generation.
JEL classification: Q41, Q42, C50
Keywords: Energy Demand, Disaggregate Analysis, Cointegration
Energy is considered to be the life line of an economy, the most vital instrument of socioeconomic development and has been recognised as one of the most important strategic commodities [Sahir and Qureshi (2007)]. Energy is not only essential for the economy but its supply is uncertain [Zaleski (2001)]. Energy is a strategic source that influenced the outcomes of wars, fueled and strangled economic development and polluted as well as clean up the environment.
In the era of globalisation, a rapidly increasing demand for energy and dependency of countries on energy indicate that energy will be one of the biggest problems in the world in the next century. This requires for alternative and renewable sources of energy. Traditional growth theories focus much on the labour and capital as major factor of production and ignore the importance of energy in the growth process [Stern and Cleveland (2004)]. The neo-classical production theories stresses that economic growth increases with the increases in labour, capital and technology. Today energy is indispensable factor and plays an important role in the consumption as well as production process,l Research suggests that energy plays an important role as compared to other variables included in the production and consumption function for countries which are at intermediate stages of economic development [IEA (2005)]. When we examine disaggregating components of energy demand, it is seen that electricity is the highest quality energy component and its share in energy consumption increases rapidly. Natural gas, petroleum and coal follow electricity respectively. This idea is supported by the results obtained when energy prices per unit are taken into consideration [Stern and Cleveland (2004) and Erbaykal (2008)].
The decisions of households and businesses regarding the use of energy have very important implications for long-run as well as short-run changes in economic activities. The nature of the demand for energy and the knowledge of its determinants are of crucial importance for accurate forecasting of its current and future needs. For this reason it is necessary to examine the nature of the relationship between energy consumption, output and the prices. The analysis is also important for the assessment of expenditures on energy consumption, energy demand management and development of strategies for future energy requirements.
Given the paramount importance of energy in the consumption patterns and productive activities, we examine the energy demand function at disaggregate level in the Pakistan over the period 1972-2007 using multivariate cointegration approach developed by Johansen (1988) and Johansen and Juselius (1990). In Pakistan, the economic structure, consumption patterns, available technologies, transport and rural-urban structure and life style that are generally different from those of well-developed countries. This situation demands the estimation of income and price elasticities of demand of each type of energy consumption, which indicate the possibilities and limitations of alternative energy control policies.
The relationship between energy consumption and economic growth has important implications at the theoretical, empirical and policy level. A large number of studies have focused on the relationship between energy consumption and real output. However, to date the results are mixed and conflicting. The variation in empirical findings could be due to different economic structure of particular countries being studied [Sari, et al. (2008)]. Another reason may be due to the fact that different economies have different consumption pattern and various sources of energy. Therefore, different sources of energy consumption might have varying impacts on the output of an economy [Ozun and Cifter (2007)]. Kraft and Kraft (1978) has found unidirectional causality running from GNP to energy consumption for United States for the period between 1947 and 1974. Their results indicates that the low level of energy dependence of US economy on energy enable US to pursue energy conservation policies which have no adverse effects on income [Jumbe (2004)]. Akarca and Long (1980) tested this relationship using the same data set for the USA and could not find relationship between the variables. Similar results were also found by Yu and Hwang (1984), Yu and Choi (1985), Erol and Yu (1987), Yu and Jin (1992), Cheng (1995), Asafu-Adjaye (2000), Soytas and Sari (2003), Altinay and Karagol (2004), Wolde-Rufael (2005), Lee (2006) and Soytas and Sari (2006). Erol and Yu (1987) examined the relationship between energy consumption and GDP for England, France, Italy, Germany, Canada and Japan for the period 1952-1982. They found bidirectional causality for Japan, unidirectional causality from energy consumption to GDP for Canada and unidirectional causality from GDP to energy consumption for Germany and Italy and no causality for France and England. In the context of developing countries Masih and Masih (1996) found evidence of Granger causality running from income to energy for Indonesia.
In contrast, the studies inter alia by Fatai, et al. (2004), Stern (1993, 2000), Yu and Choi (1985), Soytas, et al. (2001), Soytas and Sari (2003), Asafu-Adjaye (2000), Wolde-Rufael (2004) and Lee (2005) found supportive evidence of causality running from energy consumption to income. However, many researchers have reported that the relationship between energy-income may be characterised bi-directional causality. For example, Erol and Yu (1987) reported bi-directional causality for Italy and Japan and similar results are reported by Hwang and Gum (1992) for Taiwan, Masih and Masih (1996) for Pakistan, Soytas and Sari (2003) for Argentina, Ghali and El-Sakka (2004) for Canada, Wolde-Rufael (2005) for Gabon and Zambia, Lee for US and Asafu-Adjaye (2000) for Thailand and Philippines. Siddiqui (2004) concludes that the impact of all sources of energy were not same on economic growth. The impact of electricity and petroleum products were high and significant on economic growth with reverse causality between petroleum products and economic growth. Paul and Bhattacharya (2004) examined causality between energy consumption and economic growth for India over the period 1950-1996 applying both Engle and Granger (1987) and Johansen (1988) cointegration approach. The results supported the evidence of unidirectional causality from energy consumption to economic growth. Results based on Engle-Granger cointegration test exhibited unidirectional causality running from GDP to energy consumption in the long-run and no causality evidence was found in the short-run. They pointed out that when Engle-Granger approach combined with standard Granger causality test, the evidence of bi-directional causality between energy consumption and economic growth was found. The authors concluded that the long-run causal relation running from GDP to energy consumption and the short-run causal relation running from energy consumption to GDP.
The rest of this paper is organised as follows: Section 2 shed lights on the energy market in Pakistan. Model, methodology and data are discussed in Section 3. Empirical results and their interpretation are given in Section 4, while concluding remarks and policy implications are given in the final section.
2. ENERGY SECTOR IN PAKISTAN
Pakistan's energy infrastructure is under-developed, insufficient and poorly managed. (2) Presently Pakistan has been facing severe energy crisis. Despite strong economic growth and rising energy demand during the past decade, no serious efforts have been made to install new capacity of generation. Consequently, the demand exceeds supply and hence load-shedding is a common phenomenon through power shutdown [Haq and Hussain (2008)]. Pakistan needs around 14,000 to 15,000 MW electricity per day, and the demand is likely rise to approximately to 20,000 MW per day by 2010. Presently, it can produce about 11,500 MW per day and there is a shortfall of about 3000 to 4000 MW per day. This shortage is badly affecting industry, commerce, daily life and posing risks to the economic growth [Haq and Hussain (2008)]. The overall requirement of Pakistan is expected to be about 80 MTOE in 2010, up by 50 percent from the 54 MTOE of the current year. During the past 25 years energy supply in Pakistan has been increased by about 40 times but still the demand outstrips supply. With the increase in economic activities, per capita energy consumption had also been increased. Industrialisation, growth in agriculture and services sectors, urbanisation, rising per capita income and rural electrification has resulted in a phenomenal rise in energy demand [NBP (2008)]. Inefficient use of energy and its wastages has further widened the demand-supply gap and exerts strong pressure on the energy resources in the country. The annual growth of primary energy supply increased from 3.17 percent to 4.3 percent during 1997-98 to 2006-07. The share of natural gas reached to 48.5 percent, followed by oil 30.0 percent, hydro electricity 12.6 percent, coal 7.3 percent, nuclear electricity 0.9 percent, LPG 0.5 percent and imported electricity by 0.1 percent during the year 2006-07. Figure 1 presents the shares of primary energy supply in Pakistan.
It can be clear from Figure 1 that energy supply in Pakistan is highly dependent on Oil and Gas, which together contributes more than 77 percent of the total primary energy supplied. The average share of gas and oil are respectively 44.36 percent and 32.58 percent during the period 1997-98 to 2006-07. The remaining sources of energy supply consist of hydro-electricity and coal and their shares in total energy supply are around 12 percent and 6 percent respectively during the corresponding period. During 2006-07, total primary energy supply was 60387776 TOE. However, the energy supply for the final consumption is equal to 36005255 TOE.
It is now globally recognised that energy plays an important role in the production process. In Pakistan, agriculture, industry, trade and services sectors have been growing rapidly over the past few years. Given the pace of economic growth, energy demand is expected to increase. During the 1980s about 86 percent of the energy demand was met by domestic sources of energy and remaining 14 percent gap was filled by the imports. Since then, the demand-supply gap has been widening and reached around 47 percent by the end of 2000 [SBP (2006)].
At present Pakistan meets 75 percent of its energy needs by domestic resources including gas, oil and hydroelectricity production. Only 25 percent energy needs were managed through imports and oil taken major share alone; and imported oil may likely maintain important share in the future energy mix. Natural gas has emerged as the most important fuel in the recent past and the trends indicate its dominant share in the future energy mix [Sahir and Qureshi (2007)]. To sustain the pace of economic growth rate of over 7 percent over the next 25 year, Pakistan needs to expand its energy resource base. Figure 2 highlights the percentage share of the source-wise energy consumption in Pakistan during the period 1997-98 to 2006-07.
Figure 2 suggest that the average percentage share of oil in energy consumption was 40.9 percent during 1997-98 to 2006-07, followed by gas 34.6 percent, electricity 15.7 percent, coal 7.5 percent and LPG 1.3 percent during the same period. Significant changes took place among the inter-sectoral patterns of energy consumption. The change in pattern is evident from the data presented in Figure 3. It is evident from Figure 3 that on average industrial sector consumed 37.3 percent of energy, followed by transport sector with share 32.2 percent and domestic sector with share 22.2 percent. The agriculture sector, government and the commercial sector respectively consumed 2.6 percent, 2.5 percent and 3.3 percent. Though the annual growth rate of energy consumption has come down from 10.8 percent in 2004-05 to 6.1 percent at the end of 2006-07, still at present Pakistan faces deep energy crisis due to demand-supply gap. To steer the economy out of this crisis and to meet the future challenges there is urgent need to expand and upgrade the domestic resource base, accelerate exploitation and exploration of additional indigenous resources, increase the share of coal and hydroelectric in the energy mix, promote alternative renewable energy sources, improve energy efficiency and conversation, promote public private partnership in the energy sector and insure the necessary human resource development.
The per capita consumption of energy by different sources of energy is reported in Table 1. It is clear from the Table 4 that per capita consumption of oil during 1997-98 to 2003-04 fell from 4.0 kg to 1.6 kg, whereas per capita consumption of natural gas stood constant at 1.0 (MMBtu). The per capita consumption of LPG and electricity shows an increasing trend. Pakistan's economy has been growing at an average of 7.6 percent per year over the last three years. To sustain future growth of over 7 percent, the demand for energy is expected to grow at 1.2 times the economic growth rate, amounting to over 8 percent growth per year [ISSI (2007b)]. (3) However, the excess demand for energy has been increasing year-by-year and creating alarming situation for the country [Looney (2007)]. It is clear from the Figure 4 that of the excess demand for energy has increased overtime. The average excess demand for energy is equal to 0.48 QBtu for the period 1980-2005. According to Pakistan's Energy Security Plan (2005-2030), the total primary energy consumption in Pakistan is expected to increase seven-fold from 55 MTOE to 360 MTOE and over eight-fold increase in the requirement of power by 2030 [ISSI (2007b)].
[FIGURE 4 OMITTED]
Thus the country would be facing the shortage of more than 31 percent of energy in the future. In Pakistan the current energy crisis stems from the decline in hydro sources of energy and over-reliance on the expansive source of electricity. Presently, oil-based thermal plants accounts for 68 percent of generating capacity, hydroelectric plants for 30 percent and nuclear plants for only 2 percent [Looney (2007)]. This has led to a huge generation costs, which in turn adversely affect the economy over the past eight years. Rise in the oil prices pushing electricity tariff very high. As a result, manufacturing costs and inflation are at the rising trend, export competitiveness is eroded and the pressure on the balance of payments is increasing. These factors adversely affect the present growth trajectory of the economy [Loonely (2007) and NBP (2008)].
3. MODEL, METHODOLOGY, AND DATA
The Energy demand is function of various factors such as real income, relative prices and structure of the economy, the available technology and life style [Howard, et al. (1993) and Jorgenson and Wilcoxen (1993)]. However, energy demand studies frequently employs GDP and energy price as an argument to calculate income and price elasticities. These elesticities have been used to understand demand behaviour, demand management, energy forecast and policy analysis [Varian (1988)]. The estimated elasticities have relevant for designing appropriate pricing policies. Following the conventional neo-classical microeconomic theory [Bentzen and Engsted (1993); Mohammad and Eltony (1996); Beenstock, et al. (1999); Clements and Madlener (1999); Silk and Joutz (1997); Al-Faris (2002); Narayan and Smyth (2005); De Vita, et al. (2006); Dergiades and Tsoulfidis (2008) and Ziramba (2008)] the demand for energy is modeled as the outcome of a utility maximisation process undertaken by consumers. The solution of utility maximisation problem yields the following general demand function.
[q.sup.j.sub.t] = [[beta].sub.0] + [[beta].sub.1][ry.sub.t] + [[beta].sub.2][p.sub.t] + [[epsilon].sub.t] (1)
Where [q.sub.t], [ry.sub.t] and [p.sub.t] are respectively per capita energy consumption, per capita real income and domestic price level at time t. [[epsilon].sub.t] is the random term assumed to be normal and identically distributed. (4) j = E, G, P and C denote the electricity, gas, petroleum and coal consumption respectively. The lower case letters represents the logarithmic values of the variables included in Equation (1). The coefficients [[beta].sub.1] and [[beta].sub.2] represents the elasticities of real output per capita and price level.
We employ Johansen (1988) and Johansen and Juselius (1990) multivariate cointegration method to examine the cointegration between various components of energy, real output per capita and price level. We will not offering a detailed explanation of Johansen's methodology because it has well documented in the existing literature. If the null of no cointegation is rejected, then we estimate the dynamic energy demand model by using the following error-correction model:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)
such that [lambda] [not equal to] 0
If the null of no cointegration is not rejected, then we employ short-run vactor autoregressive (VAR) Granger causality/block exogeneity Wald test by estimating the following equation
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)
such that [[beta].sub.1i], [[beta].sub.2i] [not equal to] 0
The study is based on annual data covering the period 1972-2007. Data on per capita electricity consumption (Gwh), per capita petroleum consumption (tones), per capita consumption of natural gas (mm cft excluding LPG) and per capita coal consumption (thousand of metric tone) are calculated as each source of energy divided by population. Real income is calculated as nominal GDP divided by consumer price index (2000=100). Real income per capita is calculated as real income divided by population. Since the data on prices of each source of energy is not available, we proxied it by the consumer price index [see Asafu-Adjaye (2000); Hondroyiannis, et al. (2002); Akinlo (2008) and Galindo (2005)]. Data on energy sources are taken from Pakistan Economic Survey (various issues) and data on GDP, CPI and population are taken from International Financial Statistics (i.e., IFS CD-ROM- 2008).
4. EMPIRICAL ANALYSIS
We first examine the order of integration using Augmented Dickey-Fuller (ADF) unit root test. Table 2 report the results.
We started with 4 lags and tested down to zero lag and selected the model using the optimum lags and no serial correlation in the residuals. The t-ADF column gives the values of the test and if these are higher than critical values in absolute terms, the unit root hypothesis is rejected. The results suggest that except per capita consumption of petroleum ([pol.sub.t]) all other variables are stationary at their first difference, implies that all the series are integrated of order one (i.e., I (1)). Per capita consumption of petroleum ([pol.sub.t]) remains non-stationary at its first difference, implies that this variable is integrated of order two (i.e., I (2). Based on the results of unit root test we estimates natural gas, electricity and coal demand functions for Pakistan using Johansen (1988) and Johansen and Juselius (1990) multivariate cointegration method to determine the long-run relationship among I (1) variables.
(i) Natural Gas Demand Function
Natural gas has become an important and largest source of energy in Pakistan with demand and imports growing rapidly. Pakistan is likely facing major energy crisis of natural gas, electricity and oil in the next three to four year that could choke the economic growth. The major shortfall is expected in the natural gas supplies. During the period 1997-98 to 2006-07, average share of natural gas in total energy consumption was 35 percent and currently its demand is increased to 44 percent. The demand function of this important source of energy depends on real income per capita and domestic price level. To estimate the natural gas demand equation we begin with a lag structure of order 4 of all three variables included in the gas demand function (i.e. [q.sup.gas.sub.t], [ry.sub.t], [p.sub.t]) and the model was made parsimonious by reducing the number of lags on the basis of Akaike Information Criteria (AIC) and sequential F-tests for model reduction. Based on AIC and sequential F-tests we select optimal lag length of order 3. To determine the number of cointegration relationships we employ trace test adjusted for the degrees of freedom. (5) The results are reported in Table 3.
The trace test supports the evidence of one significant cointegrating vector, which implies the existence of a long-run and stable relationship between per capita gas consumption, per capita real income and domestic price level. Normalising the first cointegrating vector on [q.sup.gas.sub.t], gives the long-run gas demand function, indicates the presence of positive link with real income per capita and a negative but inelastic elasticity with respect to domestic price level.
[q.sup.gas.sub.t] = 9.62 + 1.05 [ry.sub.t] - 0.003 [p.sub.t] s.e [(1.46).sup.*] [(0.47).sup.*] (-0.18) (4)
The demand elasticities of natural gas consumption with respect to real income per capita and domestic price level possess expected signs. The coefficient of real income per capita is equal to 1.05 and statistically significant; confirming the role of income in influencing demand for natural gas in the long-run. However, the relative large size of the coefficient indicates that demand for natural gas is elastic with respect to income. The coefficient of price level is negative implies that there is negative relation relationship between gas demand and domestic price level. However, the size of this coefficient is very small and statistically insignificant. This suggests that changes in domestic price level exert almost no impact on gas consumption. These finding indicates that the demand for gas increases as the level of real income increases significantly, while changes in domestic price level produces no impact on natural gas demand in the long-run. This finding implies that gas demand is price inelastic and natural gas is necessity good. These findings are consistent with the earlier findings of Iqbal (1983) and Siddiqui and Haq (1999). (6)
Since all the variables included in the gas demand function are stationary at their first differences. Therefore, we estimate an error-correction model and the results are given by Equation (5) and t-statistics are reported in parentheses.
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)
The results suggest lagged natural gas consumption, real income per capita and domestic price level are the important determinants of natural gas demand in the short-run. Changes in the past three period's gas consumption exerts positive and negative effect on current gas consumption respectively. The overall impact of past period's gas consumption is positive in the short-run. (7) The large size of the coefficients of lagged dependent variable suggests the presence of inertia in the adjustment process in the demand for natural gas. The overall impact of real income growth exerts negative impact on gas demand in the short-run. This result suggests that in the short-run consumption of natural gas is luxury rather than necessity good. This result could be possible because natural gas connections are not provided in majority of the rural villages and remote areas; only big cities are connected with gas pipe lines. Thus for rural population, gas is luxury good and for urban population gas may be necessity good. Furthermore, as the income increases population living outside the cities substitutes firewood, kerosene oil and bio-fuel for natural gas. As a consequence, natural gas consumption reduces as the per capita real income increases.
The overall impact of price changes is negative on gas consumption in the short-run. The coefficient of lagged error-correction term has expected negative sign, implying that the deviations of [q.sup.gas.sub.t] from its long-run equilibrium values have the negative feedback effect of restoring equilibrium in the subsequent periods.
(ii) Electricity Demand Function
Electricity is another important source of energy in Pakistan. The average share of electricity in total energy consumption is about 18 percent during 1997-98 to 2006-07. Electricity consumption grew in all economic sectors during the last five years. Currently Pakistan has facing severe energy crisis, particularly electricity crisis and the electricity shortfall has gone up to 3000 to 4000 MW. This could be due to the mismanagement of electricity demand and supply. For the efficient management of electricity demand and its future needs, the knowledge of demand elasticities is necessary. The accurate estimates of the demand elasticities can be obtained by estimating the electricity demand function.
To estimate electricity demand function we begin with a lag structure of order 4 of per capita electricity consumption ([q.sup.elec.sub.t]), per capita real income ([ry.sub.t]) and domestic price level ([p.sub.t]). The model was made parsimonious by reducing the number of lags on the basis of AIC and sequential F-tests for model reduction. Based on AIC and sequential F-tests we select optimal lag length of order 2. To determine the number of cointegration relationship among [q.sup.elec.sub.t], [ry.sub.t] and [p.sub.t] we employ trace test adjusted for degrees of freedom. The results are reported in Table 4.
The trace test does not reject the null of no cointegration among the variables included in the electricity demand function. This means that there is no long-run relationship between per capita electricity consumption ([q.sup.elec.sub.t]), per capita real income ([ry.sub.t]) and domestic price level ([p.sub.t]).
In the absence of cointegration among the variables we now test the hypothesis of whether the real income per capita and domestic prices play any role in determining the per capita electricity consumption. For this purpose causality among the per capita electricity consumption, real income per capita and domestic price level and the most parsimonious results are represented by Equation (6).
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)
The results presented in Equation (6) suggest that the demand for electricity is significantly determined by the lagged electricity demand, lagged real income growth and lagged domestic price changes in the short-run. The electricity consumption lagged by two and three year exerts positive impact on the current electricity consumption. Similarly, the growth of real income per capita influences current electricity consumption growth positively. However, changes in real income per capita take one year to produce changes in current electricity consumption per capita. The effect of domestic price changes on current electricity demand lagged by two and four years remains negative and positive and significant respectively. However, the overall impact of price changes remains negative in the short-run. The short-run electricity demand function passes all the diagnostic tests.
We also employ VAR Granger causality/Block exogeneity Wald tests and the results suggest that both real income per capita and domestic price level causes electricity demand in the short-run. However, neither per capita electricity consumption and domestic price level causes real GDP per capita nor per capita electricity consumption and real GDP per capita causes domestic price level in the short-run. This result suggests that income and pricing policies play an important role in the determination of electricity consumption.
(iii) Coal Demand Function
Coal is mainly used in power, brick-kilns and cement industries. In 2006-07, the share of coal in overall energy mix is 7.5 per cent only. During 2007-08, about 53 percent of total coal production is being utilised by brick-kilns industries and 44.6 percent coal is consuming by cement industry, while power sector consuming only 2.2 percent. About 80 per cent of cement industries has switched over to coal from furnace oil due to high furnace prices. This has generated the demand for coal around 2.5 to 3.0 million tones per annum [Pakistan (2007-08)]. The consumption of coal is related to GDP and coal is used in industries that contribute to economic growth. Therefore, an econometric model is required to determine the impact of GDP and domestic price level on the consumption of coal.
To examine the coal demand, we start with 4 lags and tested down sequentially. The optimal lag length of order 2 is chosen on the basis of AIC and sequential F--statistic. To determine the cointegration between per capita coal consumption, per capita real GDP and domestic price level, we employ trace test adjusted for degrees of freedom following Cheung and Lai (1993) procedure. Table 5 reports the cointegration results for per capita coal consumption.
It can be seen from the Table 8 that the trace test does not reject the null of no cointegration between per capita coal consumption ([q.sup.coal.sub.t]), real GDP per capita ([ry.sub.t]) and domestic price level ([p.sub.t]). This result implies that there is no long-run relationship between the variables included in the coal demand function.
In the absence of cointegration among the variables we now test the hypothesis of whether the real income per capita and domestic prices play any role in determining the per capita coal consumption. To this end, causality between the per capita coal consumption, real income per capita and domestic price level is examined and the most parsimonious results are represented by Equation (7).
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (7)
The results reported in Equation (7) suggest that coal demand is significantly determined by real income and domestic price level significantly in the short-run. The sum of the short-run elasticities of coal demand with respect to real income is positive and greater than unity. However, the impact of real income per capita passes on coal consumption after one and two years. This finding suggest that coal demand is income elastic which means that when income increases the demand for coal is also increases but more than proportionately. Similarly, the sum of short-run elasticities with respect to domestic price level is negative and very small (i.e., 1.28-1.40 = -0.12). This implies that the demand for coal is price inelastic for industries consuming coal in the short-run. The estimated equation passes all the diagnostic tests and there is no econometric problem.
To examine the causality we employ VAR Granger causality/block exogeneity Wald test. The result suggests that both real income and domestic price level causes coal demand significantly in the short-run. However, no VAR causality has been observed from coal consumption and domestic price level to real GDP or from coal consumption and real GDP to domestic price level in the short-run. This result implies that income and pricing policies play very important role in the determination of coal demand.
In this study we analysed the energy demand at disaggregate level using annual data covering the period 1972 to 2007. We find long-run relationship only in the case of gas demand. The results of the gas demand equation suggest that in the long-run only real income per capital exerts positive impact on gas consumption, while domestic price play no role at all to influence the gas demand in the long-run. However, in the short-run average impact of real income per capita and domestic price remains positive and negative significantly. The error-correction term is negative and significant supporting the evidence of long-run causality between gas consumption, real income and domestic price level.
No evidence of cointegration observed for the case of electricity and coal demand functions. Therefore, we have estimated short-run dynamic demand functions for electricity and coal. In both cases the overall impact of real income and domestic price level remains positive and negative respectively. The average income elasticity of gas and coal is higher than that of electricity (in absolute terms). The average price elasticity of gas consumption is much higher than that of electricity and coal consumption (in absolute terms).The differences in the price elasticities for each component of energy have clear implications for taxation and income generation. In the short-run the average price and income elasticities of electricity and coal (in absolute terms) are small than that of gas with may indicate that in Pakistan electricity and coal is consider as necessity good. These findings are very important for income and pricing policies. To design appropriate energy pricing policy, up to date estimates of price and income elasticities of gas, electricity and coal demand that this study provides, will prove useful. The policymakers and private investors could be benefit from this study because it provides useful information regarding the market for energy demand.
Ahn, S. K. and G. C. Reinsel (1988) Nested Reduced Rank Autoregressive Models for Multiple Time Series. Journal of American Statistical Association 83, 849-856.
Akarca, A. T., and T. V. Long (1980) On the Relationship between Energy and GNP: A Reexamination. Journal of Energy Development 5,326-331.
Akinlo, A. E. (2008) Energy Consumption and Economic Growth: Evidence from 11 Sub-Sahara African Countries. Energy Economics 30, 2391-2400.
Al-Faris, A. R. (2002) The Demand for Electricity in the GCC Countries. Energy Policy 30, 117-124.
Altinay, G. and E. Karagol (2004) Structural Break, Unit Root, and the Causality between Energy Consumption and GNP in Turkey. Energy Economics 26, 985-994.
Aqeel, A and M.S. Butt (2001) The Relationship between Energy Consumption and Economic Growth in Pakistan. Asia-Pacific Development Journal 8: 2, 101-109.
Asafu-Adjaye, J. (2000) The Relationship between Energy Consumption, Energy Prices and Economic Growth: Time Series Evidence from Asian Developing Countries. Energy Economics 22, 615-625.
Beenstock, M., E. Goldin, and D. Nabot (1999) The Demand for Electricity in Isreal. Energy Economics 21, 168-183.
Bentzen, J. and T. Engsted (1993) Short- and Long-run Electicities in Energy Demand: A Cointegration Approach. Energy Economics 15:15, 9-16.
Cheng, B. S. (1995) An Investigation of Cointegration and Causality between Energy Consumption and Economic Growth. Journal of Energy and Development 21, 73-84.
Cheung, Y. W. and K. S. Lai (1993) Finite Sample Size of the Johansen's Likelihood Ratio Tests for Cointegration. Oxford Bulletin of Economics and Statistics 55, 313-328.
Clements, M. P., and R. Madlener (1999) Seasonality, Cointegration, and Forecasting UK Residential Energy Demand. Scottish Journal of Political Economy 46:2, 185-206.
De Vita, G., K. Endresen and L. C. Hunt (2006) An Empirical Analysis of Energy Demand in Namibia. Energy Policy 34, 3447-3463.
Dergiades, T. and L. Tsoulfidis (2008) Estimating Residential Demand for Electricity in the United States, 1965-2006. Energy Policy 30, 2722-2730.
Erol, U. and E. H. S. Yu (1987) On the Causal Relationship between Energy and Income for Industrialised Countries. Journal of Energy and Development 13, 113-122.
Engle, R. F. and C. W. J. Granger (1987) Cointegration and Error-Correction: Representation, Estimation, and Testing. Econometrica 55, 251-276.
Erbaykal, E. (2008) Disaggregate Energy Consumption and Economic Growth: Evodence from Turkey. International Research Journal of Finance and Economics, Issue 20, 172-179.
Erol, U. and E. H. S. Yu (1987) On the Causal Relationship between Energy and Income for Industrialised Countries. Journal of Energy and Development 13, 113-122.
Fatai, K., Oxley and F. G. Scrimgeour (2004) Modelling the Causal Relationship between Energy Consumption and GDP in New Zealand, Australia, India, Indonesia, The Philippines and Thailand. Mathematics and Computer Simulation 64, 431-445.
Galindo, L. M. (2005) Short- and Long-run Demand for Energy in Mexico: A Cointegration Approach. Energy Policy 33, 1179-1185.
Ghali, K. H. and M. I. T. El-Sakka (2004) Energy Use and Output Growth in Canada: A Multivariate Cointegration Analysis. Energy Economics 26, 225-238.
Ghosh, S. (2002) Electricity Consumption and Economic Growth in India. Energy Policy 30, 125-129.
Haq, Noor ul and Khalid Hussain (2008) Energy Crisis in Pakistan. Islamabad: Policy Research Institute.
Hondroyiannis, G., S. Lolos and E. Papapetrou (2002) Energy Consumption and Economic Growth: Assessing the Evidence from Greece. Energy Economics 24, 319-336.
Howard, R. B., L. Schipper and B. Anderson (1993) The Structure and Trends and Intensity of Energy Use: Trends in Five OECD Nations. The Energy Journal 14, 27-44.
Hwang, D. B. K. and B. Gum (1992) The Causal Relationship between Energy Consumption and GNP: The Case of Taiwan. The Journal of Energy and Development 16, 219-226.
Iqbal, M. (1983) Residential Demand for Electricity and Natural Gas in Pakistan. The Pakistan Development Review 22:1, 23-36.
International Energy Agency (IEA) Statistics Oil Information 2005.
Johansen, S. (1988) Statistical Analysis of Cointegrating Vectors. Journal of Economic Dynamics and Control 12, 231-254.
Johansen, S. and K. Juselius (1990) Maximum Likelihood Estimation and Inference on Cointegration-With Application to Demand for Money. Oxford Bulletin of Economic and Statistics 52, 169-210.
Jorgenson, D. W. and P. J. Wilcoxen (1993) Energy, the Environment and Economic Growth. In A. Kneese and J. Sweeney (eds.) Handbook of Natural Resources and Energy Economics. Vol. III, Elsevier NL, Amsterdam, The Netherlands, pp. 1267-1349.
Jumbe, C. B. L. (2004) Cointegration and Causality between Electricity Consumption and GDP: Empirical Evidence from Malawi. Energy Economics 26, 61-68.
Kraft, J and A. Kraft (1978) Note and Comments: On the Relationship between Energy and GNP. The Journal of Energy and Development 3, 402-403.
Lee, C. C. (2005) Energy Consumption and GDP in Developing Countries: A Cointegrated Panel Analysis. Energy Economics 27, 415-427.
Lee, C. C. (2006) The Causality Relationship between Energy Consumption and GDP in G-11 Countries Revisited. Energy Policy 34, 1086-1093.
Looney, R. (2007) Sustaining Economic Expansion in Pakistan in the Era of Energy Shortfalls: Growth Options to 2035. OPEC Review 31: 2, 125-157.
Masih, A. M. M., and R. Masih (1996) Energy Consumption, real Income, and Temporal Causality: Results from and Multi-Country Study Based on Cointegration and Error-Correction Modelling Techniques. Energy Economics 18, 165-183.
Mohammad, Y. H. and M. N. Eltony (1996) The Demand for Natural Gas in the Gulf Cooperation Council (GCC) States. The Middle East Business and Economic Review 8, 41-48.
National Bank of Pakistan (2008) Pakistan's Energy Sector. Economic Bulletin. January-February. pp. 4-15.
Narayan, P. K. and R. Smith (2005) The Residential Demand for Electricity in Australia: An Application of the Bounds Testing Approach to Cointegration. Energy Policy 33, 467-474.
Ozun, A. and A. Cifter (2007) Multi-scale Causality Between Energy Consumption and GNP in Emerging Markets: Evidence From Turkey. Investment Management and Financial Innovations 4:2, 61-70.
Pakistan, Government of (Various Issues) Pakistan Energy Year Book. Islamabad.
Pakistan, Government of (Various Issues) Pakistan Economic Survey. Islamabad.
Paul, S. and R. N. Bhattacharya (2004) Causality between Energy Consumption and Economic Growth in India: A Note on Conflict Results. Energy Economics 26, 977-983.
Reimers, H. E. (1991) Comparisons of Tests for Multivariate Cointegration. Institute of Statistics and Econometrics, University of Kiel. (Working Paper No. 58).
Rufael, Y. W. (2006) Electricity Consumption and Economic Growth: A Time Series Experience of 17 African Countries. Energy Policy 34, 1106-1114.
Sahir, M. H. and A. H. Qureshi (2007) Specific Concerns of Pakistan in the Context of Energy Security Issues and Geopolitics of the Region. Energy Policy 35, 2031-2037.
Sari, R., B. T. Ewing and U. Soytas (2008) The Relationship between Disaggregate Energy Consumption and Industrial Production in the United States: An ARDL Approach. Energy Economics 30, 2302-2313.
Siddiqui, R. (2004) Energy and Economic Growth in Pakistan. The Pakistan Development Review 43: 2, 175-200.
Siddiqui, R. and R. Haq (1999) Demand for Energy and the Revenue Impact of Changes in Energy Prices. Pakistan Institute of Development Economics, Islamabad. (Research Report No. 174).
Silk, JU. I. and F. L. Joutz (1997) Short and Long-run Elasticities in US Residential Electricity Demand: A Cointegration Approach. Energy Economics 19, 493-513.
Soytas, U. and R. Sari (2006) Energy Consumption and GDP: Causality Relationship in G-7 Countries and Emerging Markets. Energy Economics 25, 33-37.
Soytas, U. and R. Sari (2003) Energy Consumption and GDP: Causality Relationship in G-7 Countries and Emerging Markets. Energy Economics 25, 33-37.
Stern, D. I. (1993) Energy and Economic Growth in the USA, A Multivariate Approach. Energy Economics 15, 137-150.
Stern, D. I. (2000) A Multivariate Cointegration Analysis of the Role of Energy in the US Economy. Energy Economics 22, 267-283
Stern, D. I. and J. C. Cleveland (2004) Energy and Economic Growth. Rensselaer Working Papers in Economics. 1-42.
The Institute of Strategic Studies (2007a) Inaugural Address: Mr. Shaukat Aziz. Prime Minister Islamic Republic of Pakistan, Printed in "Energy: Sources of Regional Cooperation and Competition", Conference Proceedings Organised by the Institute of Strategic Studies, Islamabad.
The Institute of Strategic Studies (2007b) Energy: Sources of Regional Cooperation and Competition. Conference Proceedings Organised by the Institute of Strategic Studies, Islamabad.
Varian, H. (1988) Microeconomic Analysis. Norton, New York.
Wolde-Rufael, Y. (2004) Disaggregate Energy Consumption and GDP, the Experience of Shangai 1952-1999. Energy Economics 26, 69-75.
Wolde-Rufael, Y. (2005) Energy Demand and Economic Growth: The African Experience of 19 Countries. Journal of Policy Modelling 27, 891-903.
Yu, E. S. H. and B. K. Hwang (1984) The Relationship between Energy and GNP: Further Results. Energy Economics 6, 186-190.
Yu, E. S. H. and J. Y. Choi (1985) The Causal Relationship between Energy and GNP: An International Comparison. Journal of Energy Development 10, 249-272.
Yu, E. S. H. and J. C. Jin (1992) Cointegration Tests of Energy Consumption, Income and Employment. Resource Energy 14, 259-266.
Zaleski, P. (2001) Energy and Geopolitical Issues. In D. B. Rao and D. Harshyita (eds.) Energy Security. New Delhi: Discovery Publishing House.
Ziramba, E. (2008) The Demand for Residential Electricity in South Africa. Energy Policy 36, 3460-3466.
Zou, G. and K. W. Chau (2006) Short and Long Run Effects between Oil Consumption and Economic Growth in China. Energy Policy 34, 3644-3655.
(1) In this paper we have used energy demand and energy consumption interchangeably.
(2) The energy sector of Pakistan is poorly managed, service quality is low, theft of power and gas is rampant and most utilities are still receiving subsidies.
(3) ISSI represents "The Institute of Strategic Studies", Islamabad.
(4) Lower case letters denote that the variables are expressed in logarithms.
(5) Since our data sample is small. As the sample size is small finite sample adjustment to critical values is warranted [Ahn and Reinsel (1988); Reimers (1991) and Cheung and Lai (1993)].
(6) Iqbal (1983) and Siddiqui and Haq (1999) concluded that in the context of Pakistan the income elasticity of gas demand is higher and price elasticity is lower.
(7) Sum of the short-run elasticities are positive i.e. 8.41+2.14-1.80 = 8.75.
Muhammad Arshad Khan
Table 1 Per Capita Household Energy Consumption Parameter 1997-98 1998-99 1999-00 2000-01 Population (in Million) 113 133 136 140 Oil (kg) 4.0 3.8 3.6 3.3 Gas (MMBtu) 1.0 1.0 1.0 1.0 LPG (kg) 1.2 1.2 1.3 1.4 Electricity (kWh) 114 146 157 163 Parameter 2001-02 2002-03 2003-04 Population (in Million) 143 147 150 Oil (kg) 2.4 2.0 1.6 Gas (MMBtu) 1.0 1.0 1.0 LPG (kg) 1.8 1.8 1.9 Electricity (kWh) 162 161 172 Source: Household Use of Commercial Energy (Report No. 320/06, World Bank). Table 2 Augmented Dickey-Fuller (ADF) Tests Series Optimum Lag T-ADF [beta] [ry.sub.t.sup.T] 0 -1.167 0.863 [p.sub.t] 0 -0.8633 -0.006 [pol.sub.t] 2 -1.586 0.052 [gas.sub.t] 1 -0.731 0.981 [elec.sub.t] 0 -2.444 0.964 [coal.sub.t] 3 -2.438 0.487 [[DELTA]ry.sub.t] 0 -5.400 -0.015 [[DELTA]p.sub.t] 0 -3.106 ** -0.318 [[DELTA]gas.sub.t] 1 -2.598 -0.637 [[DELTA]elec.sub.t] 0 -3.645 * 0.325 [[DELTA]elec.sub.t] 1 -2.661 *** 0.416 [[DELTA]coal.sub.t] 0 -6.816 * -0.241 t--[DELTA]Y Series [sigma] lag AIC Decision [ry.sub.t.sup.T] 0.037 -- -6.506 1 (1) [p.sub.t] 0.018 -- -4.873 1 (1) [pol.sub.t] 0.063 2.009 -2.591 1 (1) [gas.sub.t] 0.042 1.781 -6.226 1 (1) [elec.sub.t] 0.035 -- -6.460 1 (1) [coal.sub.t] 0.101 2.044 -4.411 1 (1) [[DELTA]ry.sub.t] 0.038 -- -6.454 1 (0) [[DELTA]p.sub.t] 0.028 -- -4.250 1 (0) [[DELTA]gas.sub.t] 0.064 -2.598 -2.568 1 (1) [[DELTA]elec.sub.t] 0.043 -- -6.250 1 (0) [[DELTA]elec.sub.t] 0.035 -1.955 -6.588 1 (0) [[DELTA]coal.sub.t] 0.108 -- -4.386 1 (0) Optimum lag equation for ADF: [DELTA][x.sub.t] = [alpha] + [micro]t + [beta][x.sub.t-1] + [p.summation over (i=1)][[gamma].sub.i][DELTA][x.sub.t-i] + [v.sub.t] Note: (a) Optimum lag is based on minimised Akaike Information Criterion (AIC). T stands for time trend. (b) The results for the first difference variables are reported without trend. Table 3 Results of Cointegration Tests Series: ([q.sup.gas.sub.t], [ry.sub.t], [p.sub.t],) and lags = 3 Log Engenvalue Likelihood Rank (p) Trace Test p-values 191.55 0 39.66 0.014 ** 0.574 205.63 1 19.17 0.069 0.506 217.27 2 2.24 0.729 0.089 218.81 3 -- -- Note: The VAR model includes restricted constant and no trend. We reported trace test adjusted for critical values following Cheung and Lai (1993). Table 4 Results of Cointegration Tests Series: ([q.sup.elec.sub.t], [ry.sub.t], [p.sub.t],) and lags = 4 Log Engenvalue Likelihood Rank (p) Trace Test p-values 214.72 0 28.99 0.202 0.539 227.10 1 13.51 0.332 0.399 235.24 2 3.33 0.531 0.153 237.91 3 -- -- Note: See note below Table 3. Table 5 Results of Cointegration Tests Series: ([q.sup.coal.sub.t], [ry.sub.t], [p.sub.t]) and lags = 2 Engenvalue Log likelihood Rank ([rho]) Trace test p-values 168.86 0 33.72 0.070 0.517 181.24 1 13.32 0.346 0.261 186.37 2 4.84 0.309 0.160 189.33 3 -- -- Note: See note below Table 3. Fig. 1. Percentage Share of Primary Energy Supply from 1997-98 to 2006-07 (in TOE) Oil 32.58 Gas 44.36 LPG 0.37 Coal 5.8 Hydro Electricity 12.11 Nuclear Electricity 0.77 Imported Electricity 0.1 Note: Table made from pie chart. Fig. 2. Share of Source-wise Energy Consumption during 1997-98 to 2006-07 (in % of total TOE) Oil 40.9 Gas 34.6 LPG 1.3 Coal 7.5 Electricity 15.7 Note: Table made from pie chart. Fig. 3. Energy Consumption by Sector (% of Total Energy) Domestic 22.2 Commercial 3.3 Industrial 37.2 Agriculture 2.6 Transport 32.2 Other Govt. 2.5
|Gale Copyright:||Copyright 2008 Gale, Cengage Learning. All rights reserved.|