Journal of Applied Economics and Business Studies, Volume. 6, Issue 2 (2022) 77-94 https://doi.org/10.34260/jaebs.625 77 Journal of Applied Economics and Business Studies (JAEBS) Journal homepage: https://pepri.edu.pk/jaebs ISSN (Print): 2523-2614 ISSN (Online) 2663-693X Renewable and Non-renewable Energy Consumption and Economic Growth in Pakistan: A Disaggregated Analysis Shazia Farhat Durrani1, Inayatullah Jan2 & Sidra Pervez3 1 Institute of Development Studies (IDS), The University of Agriculture Peshawar, Pakistan 2 Institute of Development Studies (IDS), The University of Agriculture Peshawar, Pakistan 3 Department of Business Administration, Iqra University, Islamabad, Pakistan ABSTRACT This study addresses the research question of how renewable and non-renewable energy consumption (EC) affects economic growth (GDP) in Pakistan over a period of 1972-2015. The study extends the basic production function having labor and capital as the mainstream variables by adding major energy sources of Pakistan. The results of Toda-Yamamoto Granger causality test confirm that no causality exists between GDP and labor force and bidirectional causality exists between GDP and gross capital formation. Moreover, the results confirm a unidirectional relationship (growth hypothesis) between hydroelectricity consumption (HEC) and nuclear energy consumption (NEC) and a bidirectional relationship (feedback hypothesis) between fossil fuel consumption (FFC) and GDP, respectively. The findings suggest for an efficient utilization of existing energy resources along with diversification and expansion of the renewable energy resources. The study recommends for the government policy to avoid energy conservation as it can hamper GDP growth in Pakistan. Keywords Energy Consumption, Economic Growth, Casual Relationship, Disaggregated Analysis, Government Policy JEL Classification Q13, Q42, Q48 1. Introduction The mainstream economic theory of production considers labor and capital as the main factors of production. Nevertheless, the neo-classical aggregate production function complements energy as an additional and necessary input factor in the production model (Shahbaz et al., 2014; Chiou-Wei et al., 2016). In a broader sense, energy is a vital input for Shazia Farhat Durrani, Inayatullah Jan & Sidra Pervez 78 all production processes (Azam et al., 2015). It is required for domestic, agricultural, industrial, and transportation purposes (Kahouli, 2017). Thus, a secure, adequate, and accessible energy supply is important for socioeconomic development of a country (Jan et al., 2017; Rafindadi & Ozturk, 2017; Durrani et al., 2021; Li et al., 2022; Wang et al., 2022). Pakistan is an energy deficient country (Jan, 2012; Javed et al., 2016; Jan & Lohano, 2021). Over the last two decades, Pakistan is trapped in the worst crisis of energy (Javid et al., 2013; Jan et al., 2017) which has severely affected economic growth (GDP) in the country (Jan & Akram, 2018). Fossil fuels, renewable energy (hydroelectricity), and nuclear energy are the major sources constituting Pakistan’s total energy mix. In Pakistan, energy consumption (EC) and GDP growth are highly correlated (Jan & Akram, 2018). Figure 1 illustrates the historical trend in total EC and GDP in Pakistan. The figure shows that increase in GDP is accompanied by increased EC. During 1972-2015, Pakistan’s fossil fuels consumption has increased from 7.039 to 68.870 million tons of oil equivalent (MTOE) whereas hydroelectricity consumption (HEC) has increased from 4 to 34.6 Terawatt-hour (TWh). Until 1999, there was no considerable change in nuclear energy consumption (NEC). However, after 1999 the consumption of nuclear energy started to grow and reached to 4.7 TWh by 2015 (BPS, 2017). Figure 1: Disaggregated EC and GDP growth in Pakistan, 1972-2015 (BPS, 2016; World Bank, 2017) 0 50 100 150 200 250 0 10 20 30 40 50 60 70 80 1 9 7 2 1 9 7 3 1 9 7 4 1 9 7 5 1 9 7 6 1 9 7 7 1 9 7 8 1 9 7 9 1 9 8 0 1 9 8 1 1 9 8 2 1 9 8 3 1 9 8 4 1 9 8 5 1 9 8 6 1 9 8 7 1 9 8 8 1 9 8 9 1 9 9 0 1 9 9 1 1 9 9 2 1 9 9 3 1 9 9 4 1 9 9 5 1 9 9 6 1 9 9 7 1 9 9 8 1 9 9 9 2 0 0 0 2 0 0 1 2 0 0 2 2 0 0 3 2 0 0 4 2 0 0 5 2 0 0 6 2 0 0 7 2 0 0 8 2 0 0 9 2 0 1 0 2 0 1 1 2 0 1 2 2 0 1 3 2 0 1 4 2 0 1 5 B il li o n $ / M T O E T W h Fossil Fuel Consumption (MTOE) Hydroelectricity Consumption (Terawatt-hours) Nuclear Consumption(Terawatt-hours) GDP at market prices (in billions,constant 2010 US$) Journal of Applied Economics and Business Studies, Volume. 6, Issue 2 (2022) 77-94 https://doi.org/10.34260/jaebs.625 79 This study is motivated by the fact that EC is an important determinant of GDP growth. In countries like Pakistan, high rates of EC are highly correlated with high GDP growth (Jan & Akram, 2018). Considering this scenario, this paper attempts to address the research question that what is the causal link between energy consumption and economic growth in Pakistan? A number of studies have examined the EC-GDP nexus (Jan et al., 2020; Durrani et al., 2021; Husaini & Lean, 2022; Oryani et al., 2022; Wang et al., 2022; Zhang et al., 2022). However, this study is different from other studies in several aspects. Firstly, we determine the EC-GDP relationship by taking labor and capital as additional variables, and thereby, evade the problem of specification error that could possibly arise by the omission of relevant variables from the model. Secondly, we carry out a disaggregated analysis which gives us source specific results regarding different types of energy sources in Pakistan. To our knowledge, none or very few studies investigate the energy-growth relationship by employing fossil fuel consumption (FFC), HEC, and NEC altogether in a single study. Thirdly, we used Toda-Yamamoto (T-Y) causality test for detecting the direction of causal relationship. The T-Y causality approach allows us to test for cointegration even if the variables are integrated of order I(0) or I(1) or the combination of both orders i.e. I(0) and I(1). The approach can also be used disregarding either the variables are cointegrated or not. Fourthly and most importantly, the previous studies have explored the EC and GDP relationship without considering structural break in the analysis. We use Break Point test to determine the break point while estimating the energy-growth nexus. The combination of these different methodological approaches, different data period, and country specific outcomes and inferences make the paper novel and an original contribution to literature. This paper is organized in various sections in the following manner. After introduction, we present a literature review. In section three, we provide details of the methodology used during this research. In this section, we provide relevant information on the data, variables, and econometric technique used in this study. Section four is related to the empirical results of unit roots test and T-Y Granger causality test. In the last section, we provide conclusion and policy implications. 2. Literature Review Literature on energy-growth nexus provides evidence of mixed and conflicting results (Yuan et al., 2008; Jan et al., 2020; Oryani et al., 2022). The conflicting nature of results is because of the heterogeneity of data sets and temporal and methodological variations in various studies. Based on the causality between EC and GDP, four types of hypotheses have Shazia Farhat Durrani, Inayatullah Jan & Sidra Pervez 80 been identified in literature (Apergis et al., 2010; Ikhide & Adjasi, 2015; Marques et al., 2016; Thao & Chon, 2016; Zaidi & Ferhi, 2019; Durrani et al., 2021; Filippidis et al., 2021; Husaini & Lean, 2022; Zhang et al., 2022). In the first case, the results exhibit a causality that runs from GDP to EC (Kraft & Kraft, 1978; Al-Iriani, 2006; Li et al., 2011; Ouedraogo, 2013; Dudzevičiūtė & Šimelytė, 2017; Furouka, 2017). This kind of hypothesis is referred to as conservation hypothesis. The conservation hypothesis confirms that GDP is the major driver of EC (Oryani et al., 2022). This hypothesis recommends for energy conservation policies having little or no effects on GDP. In the second case, EC leads to GDP, and is referred to as growth hypothesis (Siddiqui, 2004; Kakar & Khilji, 2011; Arouri et al., 2014; Mutascu, 2016; Gozgor et al., 2018; Husaini & Lean, 2022). In this case, energy conservation policies are not recommended as they may negatively affect GDP. The growth hypothesis suggests for increased energy production and consumption which flourishes GDP. In the third case, a bidirectional relationship is asserted between EC and GDP. This type of relationship is summed up in the feedback hypothesis (Omri & Chaibi, 2014; Alper & Oguz, 2016; Kahia et al., 2016; Rodríguez-Caballero & Ventosa-Santaulària, 2016; Tiba & Omri, 2017; Durrani et al., 2021;). According to this hypothesis, EC and GDP are complementary to each other and a change in one causes a change in the other. This is the reason why feedback hypothesis emphasizes on energy exploration and efficiency policies. The fourth case, the neutrality hypothesis, confirms no causality between EC and GDP (Zhang & Cheng, 2009; Jalil & Feridun, 2014; Yildirim et al., 2014).The neutrality hypothesis calls for energy efficiency policies. A recap of literature on causality between EC and GDP is provided in Table 1. Table 1: Literature on EC-GDP Nexus Study Period Country Methods Relationship Omri & Chaibi (2014) 1990- 2011 Developed and developing countries DSEMs and GMM EC↔GDP for Pakistan Pin (2014) 1982-2011 OECD countries ARDL bounds test, VECM Granger causality Mixed results Ahmed & Azam (2015) 30 years, varying 119 countries Granger causality test Mixed results Alper & Oguz (2016) 1990–2009 New EU member countries Asymmetric causality test and ARDL bounds test ECGDP (Czech Republic) Destek (2016) 1971-2011 Newly Asymmetric EC− GDP https://www.sciencedirect.com/science/article/pii/S1364032116001787#! https://www.sciencedirect.com/science/article/pii/S1364032116001787#! https://www.sciencedirect.com/science/article/pii/S0960148116303512#! Journal of Applied Economics and Business Studies, Volume. 6, Issue 2 (2022) 77-94 https://doi.org/10.34260/jaebs.625 81 industrialized countries causality approach (Brazil and Malaysia) Luke (2016) 1990-2011 Sub-Sahara African countries Panel cointegration, Granger causality test ECGDP Marques et al., (2016) 1965-2013 Global ARDL bound test and T-Y causality test EC↔GDP Nadeem & Munir (2016) 1972-2014 Pakistan ARDL bound test, Granger causality test Mixed results Narayan (2016) 1984-2010 135 countries Panel data predictive regression model EC− GDP Tang et al., (2016) 1971-2011 Vietnam Cointegration and Granger causality EC→GDP Thao & Chon (2016) 1990-2012 OECD countries Stochastic distance function EC→GDP Carmona et al., (2017) 1980-2013 Oil exporting countries Cointegration and Granger causality EC↔GDP Koçak & Şarkgüneşi (2017) 1990–2012 Black Sea and Balkan countries Panel cointegration, Heterogeneous panel causality approach EC−GDP (Turkey) Gozgor et al., 2018 1990-2013 OECD countries Panel ARDL and PQR EC→GDP Jan et al., 2020 1972–2015 Pakistan ARDL bound test EC→GDP Durrani et al., 2021 1972-2015 Pakistan T-Y causality test EC↔GDP Husaini & Lean, 2022 1980-2017 Asia Threshold estimation EC→GDP Oryani et al., 2022 1976-2017 Iran ARDL, cointegration test, causality ECGDP Note: EC→GDP shows the direction of causality running from EC to GDP. EGGDP signifies the direction of causality running from GDP to EC. EC↔GDP symbolizes bidirectional causality between GDP and EC. EC−GDP means no causality between EC and GDP. ARDL means autoregressive distributed lag, VECM means vector error correction model, T-Y means Toda-Yamamoto causality test, DSEMs means dynamic simultaneous- equation models and GMM means generalized method of moments. https://www.sciencedirect.com/science/article/pii/S030142151630550X#! https://www.sciencedirect.com/science/article/pii/S030142151630550X#! Shazia Farhat Durrani, Inayatullah Jan & Sidra Pervez 82 3. Methodology 3.1.Data Sources and Variables In this study, we employed the basic production function containing only two variables, i.e. capital and labor and extended this basic function by adding FFC, HEC, and NEC in it. Annual time series data on GDP, gross capital formation (K), labor (L), FFC, HEC and NEC for Pakistan for the period 1972-2015 has been used in this study. Data on gross domestic product (GDP) and gross capital formation was obtained from the World Development Indicators (WDI) database (WB, 2017). Data on labor force was taken from various issues of the Economic Survey of Pakistan (published by the Ministry of Finance, Government of Pakistan). Data on fossil fuels consumption, HEC, and NEC was retrieved from British Petroleum’s (BPs) Statistical Review of World Energy 2016 (BPS, 2017). We use GDP in constant US$ 2010 as a dependent variable (Li et al., 2011). Explanatory variables include labor, gross capital formation, FFC, HEC, and NEC. Labor is measured in million whereas gross capital formation is measured in constant US$ 2010. Fossil fuels consumption is measured in Million Tons of Oil Equivalent (MTOE) whereas HEC and NEC are measured in Terawatt-Hour (TWh) (Jan et al., 2020; Durrani et al., 2021). All the variables were measured in natural logarithms. E-Views v.10 was used for data analysis. 3.2.Model Specification We examined the direction of causality between Pakistan’s major sources of EC (at disaggregated level) and GDP using the following basic model: ( , , , , ) t t t t t t Y f L K FFC HEC NEC= (1) Where Y denotes GDP, L denotes labor, K denotes gross capital formation, FFC means fossil fuel consumption, HEC means hydroelectricity consumption, and NEC means nuclear energy consumption. All of the study variables are converted into log form. The econometric model to be estimated is: 1 2 3 4 5t o t t t t t t LNY LNL LNK LNFFC LNHEC LNNEC      += + + + + + (2) Where 𝛽0 = intercept, 𝛽1 to 𝛽5= coefficients that are interpreted as elasticity in logarithmic models, 𝜀𝑡 = error term in time t. 3.3.Unit Root Tests It is essential to check time series data for the unit root. In the presence of a unit root, the model will generate spurious, biased and meaningless results (Gujarati & Porter, 2009). We Journal of Applied Economics and Business Studies, Volume. 6, Issue 2 (2022) 77-94 https://doi.org/10.34260/jaebs.625 83 used the Augmented Dickey-Fuller (ADF) (Chiou-Wei, 2008) and Break Point (BP) (Lee, 2006) unit root tests to avoid unit root problem. We conducted unit root tests both with (Eq. 3) and without allowing for a time trend (Eq. 4) (Oh & Lee, 2004). Where, ∆Yt means 1st differenced value of variable to be tested in time t; α means intercept; βt means time trend; Yt−1means the first lag of variable; δ means parameter to be estimated; p means number of lags; and εt means error term in time t. Our null and alternative hypotheses are: H0: δ = 0 (depicting nonstationarity) and HA: δ < 0 (depicting stationarity). The null hypothesis of unit root tests is compared with the significance levels of 1%, 5%, and 10%. In order to reject the null hypothesis, the probability value of ADF or BP statistics should be less than the specified significance level. Besides, if the probability of trend is found significant at 1%, 5%, and 10% level, then the results of model with intercept and trend are accepted (i.e. Eq. 3). Nonetheless, if the trend is found insignificant, then the decision about stationarity of a variable is made on the basis of model with intercept only (i.e. Eq. 4). 3.4.Optimum Lag Selection In this study, the Akaike’s Information Criterion (AIC) has been used for selecting optimum number of lags for the model. The AIC is significant over other criteria if the number of observations is small (Liew & Khimm, 2004). For 60 observations or below, AIC is a more reliable and accurate criterion. 3.5.Toda-Yamamoto Causality Test The direction of causality was determined by using Toda-Yamamoto causality test (Toda & Yamamoto, 1995; Leiva & Liu, 2018). T-Y test is preferred for determining causality because this test can be applied without considering the integration order of the selected variables. It means that we can use this test if all variables are integrated at levels or at difference or both. This test can also be applied regardless of the presence or absence of cointegration (Soytas & Sari, 2009). The following general form of the equations has to be estimated: 1 1 11 p t t t t ti Y Y Y     − −=  = + + + + (3) 1 1 11 p t t t ti Y Y Y    − −=  = + + + (4) 1 1 t h d k d t i t i j t j i j Y Y X    + + − − = = = + + +  (5) Shazia Farhat Durrani, Inayatullah Jan & Sidra Pervez 84 Where, d is the maximum order of integration of the variables; h and k are optimum lags of Y and X, and εt is the error term. 4. Analysis and Discussion 4.1.Unit Root Tests The results of ADF and BP unit root tests for GDP (Y), L, K, FFC, HEC, and NEC along with their order of integration are illustrated in Table 2. The table confirms that the unit root tests produce mixed results about the variables being I(0) and I(1). Integration order of each variable is decided following ADF test that does not consider structural break in the data series and a Break Point unit root test that considers a single structural break when testing for unit root. Table 2: Unit Root Tests Results Variable ADF Decision BP Decision Level Ι LNY -1.858222 -0.2666 (1992) LNL -0.696935 -3.862556 (1996) LNK -3.382422** I(0) -4.532775 (2004) I(0) LNFFC -3.268340** I(0) -4.407659 (1978) LNHEC -2.267476 -3.713505(2003) LNNEC -1.403973 -4.306537**(1999) Ι & Γ LNY -1.020664 -1.013616 (2009) LNL -1.681005 -0.719640 (1996) LNK -2.316333 -5.130789** (1991) LNFFC 0.396322 -2.726607 (2005) LNHEC -1.966210 -4.169948(1986) LNNEC -2.900759 -5.606595**(2002) First difference Ι LNY -4.359346*** -5.162705** (1992) LNL -6.962707*** -7.16885*** (2010) LNK -5.752976 -4.84811*** (1993) LNFFC -4.214846*** -5.82469*** (2004) LNHEC 6.966780*** -7.29408***(1988) LNNEC -6.905983*** -8.06017***(1999) Ι & Γ LNY -4.781955*** I(1) -5.161639 (2003) I(1) LNL -6.909658*** I(1) -8.75798*** (1996) I(1) 1 1 t h d k d t i t i j t j i j X X Y    + + − − = = = + + +  (6) Journal of Applied Economics and Business Studies, Volume. 6, Issue 2 (2022) 77-94 https://doi.org/10.34260/jaebs.625 85 LNK -5.952431*** -6.34264*** (2005) LNFFC -4.988945*** -5.519503** (2003) I(1) LNHEC -7.432628*** I(1) -7.57556***(1979) I(1) LNNEC -6.881003*** I(1) -10.2694***(1999) I(1) Note: ***, **, * represents significance at 1%, 5%, and 10% respectively. The years in the parentheses indicates break year. Ι shows intercept and Ι & Γ show intercept and trend. Table 2 shows that the two unit root tests provide mixed and somewhat contradicting results. Both tests concur that GDP, labor, HEC and NEC are stationary at first difference and are integrated of order I(1). However, the results are contradicting for FFC. For FFC, the results of ADF test indicate stationarity at level, i.e. I(0), whereas the results of the BP test show startionarity and first difference, i.e. I(1). For labor, both ADF and BP unit root tests are showing the same order, i.e. I(0). Comparing the results of two unit root tests, the results of BP test are preferred due to incorporation of structural break in it, and hence, are used to decide integration order of variables. According to BP unit root test, it is concluded that except for labor, all other variables failed to reject the null hypothesis at level. Hence, the unit roots results indicate that the dependent variable and all explanatory variable except labor are stationary at first difference and are I(1). 4.2.Optimum Lag Selection We use VAR lag order selection criteria to select the appropriate number of lags (Razzaqi et al., 2011). We used different lag order selection criteria to decide the lag length (Zhang & Cheng, 2009). The results of VAR lag order selection criteria for VAR model are provided in Table 3. The table confirms that the number of lags selected by AIC is two. In our study, three among five criteria are selecting one as optimal lag. However, we select two lag as optimum because the model run with two lags perform better and pass all the diagnostic tests. In contrast, model with one lag exhibits the problem of serial correlation and dynamic instability. The auto-regressive (AR) root graph and other relevant tests applications confirm that the model is dynamically stable at two lags (see Figure 2) and is free from non-normality, serial correlation, and heteroskedasticity issues. Shazia Farhat Durrani, Inayatullah Jan & Sidra Pervez 86 Figure 2: Inverse Roots of Auto-regressive Characteristic Polynomial Table 3: VAR Lag Order Selection Criteria Lag LogL LR a FPEb AICc SCd HQe 0 148.5196 NA 4.00e-06 -6.757055 -6.339111 -6.604863 1 230.7338 36.3552* 8.84e-08 -10.57238 -9.987257* -10.35931* 2 235.5195 7.470385 8.57e-08* -10.61071* -9.858407 -10.33676 3 238.3091 4.082382 9.21e-08 -10.55166 -9.632187 -10.21684 * indicates optimum lags selected by the criterion (at 5% level). aSequential modified LR test statistic (LR), bFinal prediction error (FPE), cAkaike information criterion (AIC), dSchwarz information criterion (SC), and eHannan-Quinn information criterion (HQ) 4.3.Results of T-Y Granger Causality Test The T-Y Granger causality test was carried out by using modified WALD (MWALD) test to investigate the direction of causality between EC and GDP (Alper & Oguz, 2016). In our model, the maximum integration order is I(1) and maximum lag length is 2 lags. Table 4 shows results of the T-Y Granger causality test. The results confirm that in panel A, all -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 Journal of Applied Economics and Business Studies, Volume. 6, Issue 2 (2022) 77-94 https://doi.org/10.34260/jaebs.625 87 variables except labor reject the null of non-Granger causality at 10%, 1%, and 5% levels respectively. These results suggest that labor does not cause GDP. However, the remaining variables such as gross capital formation, FFC, HEC, and NEC are significantly Granger causing GDP of Pakistan. The results are in consensus with those of Gozgor et al., (2018) who found that both renewable and non-renewable energy consumption are positively associated with economic growth in 29 OECD countries. In panel B, gross capital formation, FFC, and NEC are significantly Granger causing labor at 5% level and the rest of the variables fail to reject the non-causality null in case of GDP and labor, and HEC and labor. In case of panel C, however, the results are opposite to panel B. In panel C, the GDP and HEC are Granger causing the gross capital formation at 5% and 1% level respectively. Contrary to that, labor, FFC, and NEC fail to cause the gross capital formation, as none of them reject the null of non-causality at any prescribed significance level. In case of panel D, with FFC as dependent variable, only GDP is significantly causing the FFC at 10% level. All other variables including labor, gross capital formation, HEC, and NEC are failed to reject the null of non-causality. In panel E, no other variable other than NEC is significantly Granger causing the HEC. In penal F, where NEC is taken as a dependent variable in MWALD test, only FFC was causing NEC. The remaining four variables failed to reject the null of non-causality. Similar results were reported by Husaini & Lean (2022) and Oryani et al., (2022). Table 4: Results of T-Y Granger Causality Test Dependent variable Excluded variables Chi-square Probability Panel A LNY LNL 2.287107 0.3187 LNK 4.808387* 0.0903 LNFFC 5.721649* 0.0572 LNHEC 10.28256*** 0.0059 LNNEC 6.577388** 0.0373 Panel B LNL LNY 0.229032 0.8918 LNK 6.280045** 0.0433 LNFFC 8.934992** 0.0115 LNHEC 3.461423 0.1772 LNNEC 8.002591** 0.0183 Panel C LNK LNY 7.021431** 0.0299 LNL 1.738061 0.4194 LNFFC 3.949095 0.1388 LNHEC 11.32432*** 0.0035 LNNEC 3.507478 0.1731 Panel D LNFFC LNY 5.678164* 0.0585 Shazia Farhat Durrani, Inayatullah Jan & Sidra Pervez 88 LNL 0.604023 0.7393 LNK 2.020159 0.3642 LNHEC 0.754212 0.6858 LNNEC 1.918292 0.3832 Panel E LNHEC LNY 0.108051 0.9474 LNL 3.515819 0.1724 LNK 0.556998 0.7569 LNFFC 2.811650 0.2452 LNNEC 6.832299** 0.0328 Panel F LNNEC LNY 0.781551 0.6765 LNL 3.379927 0.1845 LNK 2.911404 0.2332 LNFFC 6.330418** 0.0422 LNHEC 4.471844 0.1069 Note: ***, **, * represents respective significance at 1%, 5%, and 10% levels. We provide an overview of the above results and the associated direction of causality in Table 5. The table shows that no causality exists between GDP and labor force. However, a bidirectional causal relationship between GDP and gross capital formation occurs. The results further validate a growth hypothesis for HEC and GDP and NEC and GDP. The findings are in consensus with those of Aqeel & Butt (2001) and Wolde-Rufael (2004) who confirmed growth hypothesis between electricity use and GDP growth for Pakistan and Shanghai, respectively. Likewise, the study by Omri & Chaibi (2014) supported growth hypothesis between NEC and GDP for Belgium and Spain. In case of FFC and GDP, a feedback hypothesis is confirmed for Pakistan. Similar findings were reported by Bildirici & Bakirtas (2014) who found bidirectional causality between coal consumption and GDP for China. Table 5: Direction of Causality between Variables No Granger Causality Unidirectional Granger Causality Bidirectional Granger Causality LNY – LNL LNHEC → LNY LNY ↔ LNK LNHEC − LNL LNNEC → LNY LNY ↔ LNFFC LNFFC − LNK LNK → LNL LNNEC − LNK LNFFC → LNL LNHEC − LNFFC LNNEC → LNL LNHEC → LNK LNFFC → LNNEC LNNEC → LNHEC Note: −, →, ↔ indicate no, unidirectional, and bidirectional Granger causality Journal of Applied Economics and Business Studies, Volume. 6, Issue 2 (2022) 77-94 https://doi.org/10.34260/jaebs.625 89 at1%, 5%, and 10% levels, respectively. 5. Conclusion This study probes causality between major sources of EC and GDP of Pakistan over the period 1972 to 2015. ADF and BP unit root tests were used for testing stationarity of the data series. Only gross capital formation was found stationary at level. The rest of all variables became stationary after differencing and were integrated of order I(1). Given the mixed integration order, T-Y Granger causality test was employed for investigating the direction of causality among the variables of interest. Results of the Granger causality tests confirm the existence of feedback hypothesis for GDP and FFC. Besides, the findings also confirm the growth hypothesis for hydroelectricity and NEC and GDP. Based on the findings of this study, it is recommended that government should focus on expanding the supply of energy along with the diversification of energy mix. The focus of the government policy should be on increasing the renewable energy sources in the total energy mix. This will promote economic growth as well as global environmental sustainability which are hampered by carbon emission from non-renewable energy consumption. 6. Acknowledgements The authors thank the anonymous reviewers for their constructive remarks which helped in improving the manuscript. The authors confirm that the study was not financed by any organization. References Ahmed, M. and M. Azam (2015). ‘Causal nexus between energy consumption and economic growth for high, middle and low income countries using Frequency Domain Analysis’, Renewable and Sustainable Energy Reviews, 60, pp. 653-678. Al-Iriani, M.A. (2006). ‘Energy-GDP relationship revisited: An example from GCC countries using Panel Causality’, Energy Policy, 34, pp. 3342-3350. Alper, A. and O. Oguz (2016). ‘The role of renewable energy consumption in economic growth: Evidence from Asymmetric Causality’, Renewable and Sustainable Energy Reviews, 60, pp. 953-959. Apergis, N., J. Payne, K. Menyah and Y. Rufael (2010). ‘On the causal dynamics between emissions, nuclear energy, renewable energy, and economic growth’, Ecological Economics, 69(1), pp. 2255-2260. Aqeel, A. and M.S. Butt (2001). ‘The Relationship between energy consumption and economic growth in Pakistan’, Asia-Pacific Development Journal, 8(2), pp. 101- 110. Arouri, M., G.S. Uddin, P. Kyophilavong, F. Teulon and A.K. Tiwari (2014). ‘Energy utilization and economic growth in France: Evidence from Asymmetric Causality Test’. IPAG working paper series, 102. Shazia Farhat Durrani, Inayatullah Jan & Sidra Pervez 90 Azam, M., A.Q. Khan, B. Bakhtyar and C. Emirullah (2015). ‘The causal relationship between energy consumption and economic growth in the ASEAN-5 countries’, Renewable and Sustainable Energy Reviews, 47, pp. 732-745. Bildirici, M.E. and T. Bakirtas (2014). ‘The Relationship among oil, natural gas and coal consumption and economic growth in BRICTS (Brazil, Russian, India, China, Turkey and South Africa) countries’, Energy, 65, pp. 134-144 British Petroleum Statistics. (2016). ‘BP Statistical Review of World Energy’. https://www.bp.com/en/global/corporate/energy-economics/statistica-review-of- world-energy/downloads.html [Accessed on 28 June 2019]. Carmona, M., E. Congregado, J. Feria and J. Iglesias (2017). ‘The energy-growth nexus reconsidered: Persistence and causality’, Renewable and Sustainable Energy Reviews, 71, pp. 342-347. Chiou-Wei, S.Z., C. Ching-Fu and Z. Zhu (2008). ‘Economic growth and energy consumption revisited - Evidence from linear and nonlinear Granger causality’, Energy Economics, 30, pp. 3063-3076. Chiou-Wei, S-Z., Z. Zhu, S-H. Chen and S-P. Hsueh (2016). ‘Controlling for relevant variables: Energy consumption and economic growth nexus revisited in an EGARCH-M (Exponential GARCH-in-Mean) model’, Energy, 109, pp. 391-399. Destek, M.A. (2016). ‘Renewable energy consumption and economic growth in newly industrialized countries: Evidence from Asymmetric Causality Test’, Renewable Energy, 95, pp. 478-484. Dudzevičiūtė, G. and A. Šimelytė (2017). ‘Export, energy consumption and economic growth inter-linkages: The case of Lithuania’, Scientific Annals Economics and Business, 64(3), pp. 395-410. Durrani, S.F., I. Jan and M., Ahmad (2021). ‘Do primary energy consumption and economic growth drive each other in Pakistan? Implications for energy policy’, Biophysical Economics and Sustainability, 6(8). https://doi.org/10.1007/s41247- 021-00090-x. Filippidis, M., P. Tzouvanas and I. Chatziantoniou (2021). ‘Energy poverty through the lens of the energy-environmental Kuznets curve hypothesis’, Energy Economics, 100 (105328). https://doi.org/10.1016/j. eneco.2021.105328. Furouka, F. (2017). ‘Renewable electricity consumption and economic development: New findings from the Baltic Countries’, Renewable and Sustainable Energy Reviews, 71, pp. 450-463. Gozgor, G., C.K.M. Lau and Z. Lu (2018). ‘Energy consumption and economic growth: New evidence from the OECD countries’, Energy, 153, pp. 27-34. Gujarati, D. and D. Porter (2009). Basic Econometrics. 5th ed. USA: McGraw-Hill Irwin. Husaini, D.H. and H.H. Lean 2022. ‘Renewable and non-renewable electricity-growth nexus in Asia: The role of private power plants and oil price threshold effect’, Resources Policy, 78(102850). https://doi.org/10.1016/j.resourpol.2022.102850. Journal of Applied Economics and Business Studies, Volume. 6, Issue 2 (2022) 77-94 https://doi.org/10.34260/jaebs.625 91 Ikhide, E. and C. Adjasi (2015). ‘The causal relationship between renewable and non- renewable energy consumption and economic growth: The case study of Nigeria’. Paper presented at the Economic Society of South Africa at UCT: South Africa. Jalil, A. and M. Feridun (2014). ‘Energy-driven economic growth: Energy consumption- economic growth nexus revisited for China’, Emerging Markets Finance and Trade, 50(5), pp. 159-168. Jan, I. and H.D. Lohano (2021). ‘Uptake of energy efficient cookstoves in Pakistan’, Renewable and Sustainable Energy Reviews, 137(110466). https://doi.org/10.1016/j.rser.2020.110466. Jan, I., S.F. Durrani and H. Khan (2020). ‘Does renewable energy efficiently spur economic growth? Evidence from Pakistan’, Environment, Development and Sustainability, 23, pp. 373-387. Jan, I. and W. Akram (2018). ‘Willingness of rural communities to adopt biogas systems in Pakistan: Critical factors and policy implications’, Renewable and Sustainable Energy Reviews, 81, pp. 3178-3185. Jan, I. (2012). ‘What makes people adopt improved cookstoves? Empirical evidence from rural northwest Pakistan’, Renewable and Sustainable Energy Reviews, 16(5), pp. 3200-3205. Jan, I., S. Ullah, W. Akram, N.P. Khan, S.M. Asim, Z. Mahmood, M.N. Ahmad and S.S. Ahmad (2017). ‘Adoption of Improved Cookstoves in Pakistan: A logit analysis’, Biomass and Bioenergy, 103, pp. 55-62. Javed, M.S., R. Raza, I. Hassan, R. Saeed, N. Shaheen, J. Iqbal and S.F. Shaukat (2016). ‘The energy crisis in Pakistan: A possible solution via biomass-based waste’, Journal of Renewable and Sustainable Energy, 8 (043102). Javid, A.Y., M. Javid and Z.A. Awan (2013). ‘Electricity consumption and economic growth: Evidence from Pakistan’, Economics and Business Letters, 2(1), pp. 21- 32. Kahia, M., M.S. Ben-Aissa and L. Charfeddine (2016). ‘Impact of renewable and non- renewable energy consumption on economic growth: New evidence from the MENA net oil exporting countries (NOECs)’, Energy, 116(1), pp. 102-115. Kahouli, B. (2017). ‘The short and long run causality relationship among economic growth, energy consumption and financial development: Evidence from South Mediterranean Countries (SMCs)’, Energy Economics, 68, pp. 19-30. Kakar, Z.H. and B.A. Khilji (2011). ‘Energy consumption and economic growth in Pakistan’, Journal of International Academic Research, 11(1), pp. 33-36. Kraft, J. and A. Kraft (1978). ‘Relationships between energy and GNP’, The Journal of Energy and Development, 3, pp. 401-403. Koçak, E. and A. Şarkgüneşi (2017). ‘The renewable energy and economic growth nexus in Black Sea and Balkan Countries’, Energy Policy, 100, pp. 51-57. Lee, C.C. (2006). ‘The causality relationship between energy consumption and GDP in G- 11 countries revised’, Energy Policy, 34, pp. 1086-1093. Shazia Farhat Durrani, Inayatullah Jan & Sidra Pervez 92 Leiva, B. and Z. Liu (2019). Energy and economic growth in the USA two decades later: Replication and reanalysis’, Energy Economics, 82, pp. 89-99. Li, C., T. Lin, Y. Chen, Y. Yan, Z. Xu (2022). ‘Nonlinear impacts of renewable energy consumption on economic growth and environmental pollution across China’, Journal of Cleaner production, 368(133183). https://doi.org/10.1016/j.jclepro.2022.133183. Li, Y., Feng, N., and Fang, N. (2008). ‘Relationship between energy consumption and economic growth: Empirical study based on data on Hebei Province from 1980 to 2008’, System Engineering Procedia, 1, pp. 117-123. Liew, V., and Khim, S. (2004). ‘Which lag length selection criteria should we employ? Economics Bulletin, 3(33), pp. 1-9. Luke, N. (2016). ‘Examining the renewable energy consumption-economic growth nexus in Sub-Saharan African countries’, MSc Honors Program Theses. http://scholarworks.uni.edu/hpt/241[Accessed on 15 June 2019]. Marques, M.L., A.J. Fuinhas and C.A. Marques (2016). ‘On the global energy consumption and economic growth nexus: A long time span analysis’, International Energy Journal, 16, pp. 143-150. Mutascu, M. (2016). ‘A bootstrap panel granger causality analysis of energy consumption and economic growth in the G7 Countries’, Renewable and Sustainable Energy Reviews, 63, pp. 166-171. Nadeem, S. and K. Munir (2016). ‘Energy consumption and economic growth in Pakistan: A sectoral analysis. https://mpra.ub.uni-muenchen.de/74569/ [Accessed on 15 January 2018]. Narayan, S. (2016). ‘Predictability within the energy consumption-economic growth nexus: Some evidence from income and regional groups’, Economic Modelling, 54, pp. 515-521. Oh, W. and K. Lee (2004). ‘Causal relationship between energy consumption and GDP Revisited: The case of Korea 1970–1999’, Energy Economics, 26, pp. 51-59. Omri, A. and A. Chaibi (2014). ‘Nuclear energy, renewable energy, and economic growth in developed and developing countries: A Modeling Analysis from Simultaneous- Equation Models’. http://www.ipag.fr/fr/accueil/la-recherche/publications- WP.html [Accessed on 02 August 2022]. Oryani, B., H. Kamyab, A. Mоridiаn, Z. Azizi, S. Rezania, S. Chelliapan (2022). ‘Does structural change boost the energy demand in a fossil fuel-driven economy? New evidence from Iran’, Energy, 254(C). 124391. https://doi.org/10.1016/j.energy.2022.124391. Ouedraogo, N.S. (2013). ‘Energy consumption and economic growth: Evidence from the economic community of West African States’, Energy Economics,36, pp. 637-647. Pin, H.L. (2014). ‘Renewable energy consumption and economic growth in nine OECD countries: Bounds test approach and causality analysis’, The Scientific World Journal of Applied Economics and Business Studies, Volume. 6, Issue 2 (2022) 77-94 https://doi.org/10.34260/jaebs.625 93 Journal, 2014, pp. 1-6. Rafindadi, A.A. and I. Ozturk (2017). ‘Impacts of renewable energy consumption on the German economic growth: Evidence from Combined Cointegration Test’, Renewable and Sustainable Energy Reviews, 75, pp. 1130–1141. Razzaqi, S., F. Bilquees and S. Sherbazz (2011). ‘Dynamic relationship between energy and economic growth: Evidence from D8 countries’, Pakistan Development Review, 50(4), pp. 437-58. Rodríguez-Caballero, C.V. and D. Ventosa-Santaulària (2016). ‘Energy-growth long-term relationship under structural breaks: Evidence from Canada, 17 Latin American Economies and the USA’, Energy Economics, 61, pp. 121-134. Shahbaz, M., M. Arouri and F. Teulon (2014). ‘Short- and long-run relationships between natural gas consumption and economic growth: Evidence from Pakistan’, Economic Modelling, 41, pp. 219-226. Siddiqui, R. (2004). ‘Energy and economic growth in Pakistan’, Pakistan Development Review, 43(2), pp. 175-200. Soytas, U. and R. Sari (2009). ‘Energy consumption, economic growth, and carbon emissions: challenges faced by an EU Candidate Member’, Ecological Economics, 68, pp. 1667-1675. Tang, C.F., B.W. Tan and I. Ozturk (2016). ‘Energy consumption and economic growth in Vietnam’, Renewable and Sustainable Energy Reviews, 54, pp. 1506-1514. Thao, N.T.N. and V.L. Chon (2016). Nonrenewable, renewable energy consumption and economic performance in OECD countries: A Stochastic Distance Function Approach’. http://veam.org/wp-content/uploads/2016/08/70.-Nguyen-Thi-Ngan- Thao.pdf [Accessed on 7 July 2022]. Tiba, S. and A. Omri (2017). ‘Literature Survey on the relationships between energy, environment and economic growth’, Renewable and Sustainable Energy Reviews, 69, pp. 1129-1146. Toda, H.Y. and T. Yamamoto (1995). ‘Statistical inference in vector autoregressions with possibly integrated processes’, Journal of Econometrics, 66 (1-2), pp. 225-250. Wang, Q., Z. Dong, R. Li and l. Wang (2022). ‘Renewable energy and economic growth: New insight from country risks’, Energy. 238(C). 122018. https://doi.org/10.1016/j.energy.2021.122018. Wolde-Rufael, Y. (2004). ‘Disaggregated industrial energy consumption and GDP: The case of Shanghai, 1952-1999’, Energy Economics, 26, pp. 69-75. World Bank. (2017). World Development Indicators Database. https://databank.worldbank.org/data/reports.aspx/source=2&country=PAK [Accessed on 28 December 2017] Yildirim, E., D. Sukruoglu and A. Aslan (2014). ‘Energy consumption and economic growth in the next 11 countries: The Bootstrapped Autoregressive Metric Causality Approach’, Energy Economics, 44, pp. 14-21. Yuan, J.H., J.G. Kang, C.H. Zhao and Z.G. Hu (2008). ‘Energy consumption and economic Shazia Farhat Durrani, Inayatullah Jan & Sidra Pervez 94 growth: Evidence from China at both aggregated and disaggregated levels’, Energy Economics, 30(6), pp. 3077-3094. Zaidi, S. and S. Ferhi (2019). ‘Causal relationships between energy consumption, economic growth and CO2 emission in Sub-Saharan: evidence from dynamic simultaneous-equations models’, Modern Economy, 10, 2157-2173. Zhang, Q., S.A.R. Shah and L. Yang (2022). ‘An appreciated response of disaggregated energies consumption towards the sustainable growth: A debate on G-10 economies’, Energy, 254(A). 124377. https://doi.org/10.1016/j.energy.2022.124377. Zhang, X.P. and X.M. Cheng (2009). ‘Energy consumption, carbon emissions, and economic growth in China’, Ecological Economics, 68(10), pp. 2706-2712.