Renewable and Sustainable Energy Reviews 44 (2015) 211–220
Contents lists available at ScienceDirect
Renewable and Sustainable Energy Reviews journal homepage: www.elsevier.com/locate/rser
Linking financial development, economic growth and energy consumption in Pakistan Rabia Komal, Faisal Abbas n Department of Management Sciences, COMSATS Institute of Information Technology (CIIT), Park Road, Islamabad 44000, Pakistan
art ic l e i nf o
a b s t r a c t
Article history: Received 15 August 2014 Received in revised form 23 November 2014 Accepted 12 December 2014 Available online 5 January 2015
This paper aims at exploring the finance–growth–energy nexus for Pakistan over the 1972–2012 period. By employing the system GMM estimation technique, the study tries to capture the impact of financial development over energy consumption through economic growth channel and includes energy prices and urbanization in the structural model. The study finds positive and significant impact of economic growth and urbanization on energy consumption, while the effect of energy prices over energy consumption is significant but negative. Financial development positively and significantly affects energy consumption through the economic growth channel. Our analysis is important for policy makers for effective energy demand planning and conservation policies that would ensure sustainable economic development as well as serve as motivation to search alternative energy sources to meet the bourgeoning energy demand in Pakistan. & 2014 Elsevier Ltd. All rights reserved.
Keywords: Energy consumption Financial development Economic growth GMM Pakistan Energy prices Urbanization
Contents 1. 2. 3.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Energy crisis of Pakistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Literature review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Finance–energy nexus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Finance–growth nexus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Energy–growth nexus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Theoretical framework and model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Data and methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Results and discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1. Energy consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2. Economic growth. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3. Summary of channel effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. Conclusion and policy implication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
211 212 212 212 214 216 216 216 216 217 217 217 218 218 219 220
1. Introduction n
Corresponding author. Tel.: +923233844383. E-mail addresses:
[email protected] (R. Komal),
[email protected] (F. Abbas). http://dx.doi.org/10.1016/j.rser.2014.12.015 1364-0321/& 2014 Elsevier Ltd. All rights reserved.
Literature reveals that energy is crucial for enhancing economic productivity [27,35,49]. Furthermore, growth in production activities stimulates energy demand due to increase in consumption
212
R. Komal, F. Abbas / Renewable and Sustainable Energy Reviews 44 (2015) 211–220
[52]. EIA [18] estimated increase in the global energy consumption by about 56 percent during the period 2010–2040. Most of this increase will take place in non-OECD countries, where energy consumption is stimulated by strong growth in the economy [23,29]. Industrial expansion and population growth has led to increased energy consumption in Asian economies in general, and in Pakistan in particular [61]. Pakistan, falls in low middle income group country, has been facing the worst energy crisis for the past few years and its energy demand has continuously increased with economic and population growth. Furthermore, as a result of financial reforms, Pakistan's financial sector has shown exceptional and notable growth, particularly with respect to the banking sector [30].1 This would also have implication for both economic growth and energy consumption [26]. Economy's long run growth potential is about 6.5 percent per annum but power outages have reduced it to 2 percent [21]. This indicates that growth of the economy is largely suppressed due to energy shortages. Kessides [28] has pointed out that electricity shortage has a negative impact on international competitiveness and exports, employment, and poverty alleviation in Pakistan. Recent energy literature includes financial variables when modeling energy consumption for an economy, asserting that financial variables can impact energy demand. It is, therefore, essential to consider financial variables in such a study to exclude the possibility of underestimation of energy demand.2 Moreover, studies on finance–energy nexus highlight the direct impact of financial development on energy consumption [23,52]. Financial development affects energy consumption indirectly via economic growth. This effect may be either positive or negative depending whether economic growth occurs in an efficient manner or not. For instance, growth in financial sector improves funds availability for investment projects that results in industrial growth leading to expansion in production activities. This in turn enhances economic growth, and increases the demand for new infrastructure and more energy, thereby positively influencing energy consumption [23,52]. However, the ability to adopt technological innovations in industrial sector development varies across countries that affect the intensity of energy consumption [52]. Thus, this study intends to explore the finance–growth– energy nexus in Pakistan i.e. how financial development affects energy consumption (i.e., positively or negatively) using economic growth channel? Empirical literature on finance–energy nexus follows one of the two approaches. The first approach estimates the model in terms of elasticity in the variables by including energy consumption and financial development jointly in a single equation without much theoretical base. The second approach estimates the model using conventional unit root, cointegration and causality tests. The present study is different from these approaches in that it uses system GMM technique to separately capture the impact of financial development over energy consumption through economic growth. It, therefore, prepares a strong theoretical ground for empirical analysis. It explores the channel variable (economic growth) through which financial development may likely affect energy consumption. This channel variable is used to capture the effect of change in financial development on energy consumption, and to infer if increased financial development is linked to more energy consumption in Pakistan or vice versa. To the best of our knowledge, there is no published study that captured the indirect relation of financial development and energy consumption. 1 Financial reforms are introduced in several areas relevant to financial markets and institutions with the core purpose to encourage competition, improve supervision and governance and adopt monetary, credit and exchange mechanism that ensure efficient allocation of financial resources [32]. 2 Present study uses energy consumption and energy demand interchangeably.
This paper is organized as follows: Section 2 discusses energy scenario of Pakistan. Section 3 provides an overview of empirical literature. Section 4 elaborates theoretical framework and model. While data, and econometric methodology for analysis is described in Section 5. Section 6 deals with empirical results and their discussion. Section 7 concludes the study along with policy implications and future research directions.
2. Energy crisis of Pakistan The energy sector of Pakistan is in crisis and has been facing many challenges for the past few years. Circular debt, fragile financial situation of energy supply firms, intense reliance on gas/oil (above 80%), declining gas production, less utilization of cheap hydel and coal resources and unexploited power production capacity are a few major limitations contributing to energy scarcity [4,11,20,21,44]. Dependence on expensive furnace oil within thermal electricity production has increased that is coupled with volatile international oil prices which has adverse implications for cost structure of electricity production and could further undermine energy shortage in Pakistan [28]. According to GoP [21], rate of growth in net primary energy supply remained 1.8 percent while rate of growth in final energy consumption remained 2.9 percent for the past six years, that is clear evidence of energy shortage. Considerable increase in the usage of electric appliances has contributed to increase in domestic demand that led to increase in the share of native users in total electricity consumption from 23 percent in 1980–81 to about 43 percent in 2012–13. Kessides [28] has highlighted that industrialization, urbanization, agricultural and service sector growth, rising per capita income and rural electrification are considered among key factors of growth in energy demand. There are estimates that electricity shortage will rose to 8000 MW in 2017 and over 13,000 MW in 2020. Although technical and financial support has been provided by international donor agencies to enhance production capacities and performance of generation companies (GENCOs), but this enhancement is insufficient in the light of rapidly growing energy demand.
3. Literature review The existing literature on energy economics is mainly based on three nexus; finance–energy nexus, finance–growth nexus, and energy–growth nexus. We discuss these one by one below. 3.1. Finance–energy nexus Literature on finance–energy nexus highlights the ways by which financial development can potentially affect energy consumption. At the household level, it is easier for consumers to gain an easy and cheap access to borrowed funds to purchase energy consumable products that directly affect energy demand. At the industrial level, it is easier for entrepreneurs to gain access to financial capital in order to expand existing businesses or start a new one, thereby creating a business effect. Increased stock market activity is regarded as an indicator of economic growth because it increases risk diversification for consumers and businesses that result in increased fund availability for investment projects, and thereby creates a wealth effect. This builds up consumer and business confidence that leads to expansion in the economy and creates demand for energy intensive products [17,49]. Shahbaz and Lean [52] mention that growth in industrial sector raises energy demand in two ways; firstly due to cross-sectoral growth; and secondly, as with increase in labor
Table 1 Summary of empirical studies on finance–energy nexus. Econometric technique
Main findings
Single country studies Kakar et al. [27] Pakistan
1980–2009
Johnson Co-integration test, VECM Granger causality test
Significant long run relationship exists. Unidirectional causality from M2 to EC and EC to DC.
Islam et al. [23]
Malaysia
1971–2009
ARDL, VECM Granger causality test
Cointegration among FD, EG, and ED. Short run causality from FD to EC.
Tang and Tan [57] Shahbaz and Lean [52]
Malaysia Tunisia
1972–2009 1971–2008
Johansen–Juselius cointegration test ARDL, UECM Granger causality test
Energy and financial sector development in long run relation. Long run relationship exists among EG, FD, URB, IND and EC. Bidirectional causality between FD and EC.
Shahbaz et al. [54]
China
1971–2011
ARDL, Granger causality test
Long run relationship exists among EC, EG, FD, capital, and trade. Bidirectional causality between FD and EC.
Mehrara and Musai [40]
Iran
1970–2009
ARDL
Long run relationship exists among EC, EG, Capital stock, FD, Oil revenues in long run.
Multi-country studies Mudakkaret al. [41] SAARC countries Sadorsky[49] 22 emerging economies
1975–2011 1990–2006
Toda–Yamamoto–Dolado–Lutkephol (TYDL) Granger causality test Generalized method of moments (GMM)
Bidirectional causality between FD and EC in short run in Pakistan. Stronger impact of stock market variables on energy consumption. Greater impact of income on energy consumption.
Sadorsky [50]
9 Central and East European economies
1996–2006
Generalized method of moments (GMM)
Stronger impact of banking variables on energy consumption.
Coban and Topcu [17]
European Union (EU)
1990–2011
Generalized method of moments (GMM)
No significant association between FD and EC in EU. U-shaped pattern for banking variables with energy.
Omri and Kahouli [43] Al-mulali and Sab [5]
65 countries
1990–2011
Generalized method of moments (GMM)
Unidirectional causality from FD to EC for global panel.
30 SSA countries
1980–2008
Pedroni (Engle–Granger based) cointegration test, VECM Granger causality test
Increased EC increases EG and FD but with high pollution Long run relation and causality exist between EC, FD, EG and CO2 emissions.
Country
R. Komal, F. Abbas / Renewable and Sustainable Energy Reviews 44 (2015) 211–220
Sample period
Author
Note: EC¼ Energy consumption, FD ¼Financial development, EG ¼ Economic growth, M2 ¼ money supply, DC¼ Domestic credit, URB ¼Urbanization, IND ¼Industrialization, UECM ¼Unrestricted Error Correction Model.
213
214
R. Komal, F. Abbas / Renewable and Sustainable Energy Reviews 44 (2015) 211–220
Table 2 Summary of empirical studies on finance-growth nexus. Author
Country
Main findings
Sample period
Econometric technique
1971–2004
Positive and significant impact of financial depth on EG in long run and of investment share on EG in short run. ARDL, UECM, and Dynamic FD leads to higher EG through improved efficiency of investment. OLS FD2EG
Single country studies Khan et al. Pakistan [32] Ang [6] Malaysia
1960–2003
Rufael [48]
Kenya
1966–2005
Bojanic [16]
Bolivia
1940–2010
Anwar and Sun [7]
Malaysia
1970–2007
Multi-country Ahmed and Ansari [3] Gregorio and Guidotti [22]
studies Pakistan, India, 1973–1991 Bangladesh 98 countries 1960–1985, 1950–1985
King and Levine [33]
80 countries
1960–1989
Correlation, OLS regression FD stimulates EG via boosting rate of capital accumulation and promoting efficient capital use
Beck and Levine [14]
40 countries
1976–1998
OLS regression, GMM technique
Banks and stock market independently spur economic growth.
Wu et al. [60] 13 EU countries
1976–2005
Unit root test and PMG method
Long run relationship exists among banking sector, stock market and economic development. Effect of FD on output may be negative in long run but stable economic development may be recurred via improving risk diversification and information services of banks.
Ahmed [2]
1981–2009
System Generalized method of moments (GMM)
Negative relationship between FD and EG due to financial liberalization. Countries experience positive effects of liberalization that are having strong legal institutions, stable inflationary environment and higher human capital.
21 SSA countries
ARDL and ECM
VAR and TYDL causality test Cointegration, Granger causality and ECM
Bidirectional causality between intermediary financial development and economic growth.
Generalized method of moments (GMM)
Domestic capital stock is affected by level of FD that leads to EG.
Granger causality test
FD-EG
OLS regressions
FD leads to EG but effect changes with regions, time period and income level. Effect is stronger in middle and low income countries. FD hinders EG in absence of proper regulation.
Cointegration exists between FD and EG. Causality runs FD- EG.
Note: 2indicates bidirectional causality, -indicates unidirectional causality.
demand due to economic growth, income improves that boosts demand for energy consumable products, and thereby enhance energy consumption. However, evidence also implies that financial development lessens energy consumption by achieving efficiency in its use for which amendments in infrastructure is required. This comes from investment in research and development of advance technologies that is linked to the development of financial sector. Besides, if consumers use energy efficient products like home appliances, it lowers energy use [23]. Coban and Topcu [17] also assert that financial development makes accessibility to advance technology easier that leads to energy efficiency, hence reduces energy consumption. Kakar et al. [27] have asserted that financial development can significantly contribute to efficient economic growth by reducing energy consumption in Pakistan. Several studies have explored the impact of financial development over energy consumption incorporating other variables in the model (see Table 1). 3.2. Finance–growth nexus Literature on financial development and economic growth nexus discusses two channels with which financial development promotes economic growth i.e. capital accumulation and total factor productivity (TFP).3 Literature reveals that financial 3 Financial development increases domestic and foreign capital investment by enhancing confidence of investors over financial system. As borrowing and lending mechanisms increase in economy, investment is promoted so output is increased. This channel is referred as capital accumulation or quantitative channel. Efficient financial system does better screening of available investment opportunities.
development comes from stock market development, banking sector development, increased domestic and foreign direct investment, and financial deregulation through liberalization measures, that promotes economic growth. Levine [36] has stated that banks perform a significant role in enhancing growth at initial level of economic development and in weak institutional settings. Services offered by financial intermediaries' i.e. evaluating projects, mobilizing savings, monitoring managers, diversifying risk, and aiding transactions, are necessary for technological progress and economic development. Better performing financial intermediaries are efficient in directing credit from households (savers) to business enterprises (borrowers) that promotes economic growth [37]. Financial liberalization has positive impact on growth through improving monetary transmission mechanism, increasing savings and investment, and improving risk sharing that lowers the cost of capital [6,49,52]. However, countries with weak regulatory framework may experience the negative effect of liberalization which leads to domestic capital flight and increasing the risk of financial fragility [2]. Several studies have confirmed the existence of a positive (negative) significant association between financial development and economic growth. Besides, causality between these two variables has been investigated with mixed results (for example, see Table 2).
(footnote continued) Thereby, it facilitates adoption of advance or innovative technology by channeling funds to knowledge and technology intensive industries, so economy grows. This channel is referred as total factor productivity (TFP) [also called technological innovation] or qualitative channel [6,7].
Table 3 Summary of studies on energy-growth nexus. Country
Sample period
Econometric technique
Main findings
Single country studies Jamil and Ahmad [24]
Pakistan
1960–2008
Johansen cointegration and VECM causality test
Long run relationship exists among EG, EC, EP. EG-EC
Tang and Tan [58]
Malaysia
1970–2009
ARDL and VECM Granger causality test
Direct relationship between EC and EG. Bidirectional causality between EC and each variable.
Belaid and Abderrahmani [15]
Algeria
1971–2010
Johansen and Gregory–Hansen cointegration test, VECM causality test
EC2EG. Oil price X EC
Shiu and Lam [55]
China
1971–2000
Johensen cointegration test, ECM causality test
EC-EG
Multi-country studies Abbas and Choudhury [1]
Pakistan and India
1972–2008
Johansen cointegation test, ECM causality test
EC2Agriculture GDP (For Pakistan) Agriculture GDP-EC (For India)
Odhiambo [42]
South Africa, Kenya, Congo
1972–2006
ARDL-bounds testing approach and Granger causality test
EC-EG (SA & Kenya) EG-EC (Congo)
Fatai et al. [19]
New Zealand, Australia
1960–1999
ARDL and Toda and Yamamoto (TY) Granger causality test
EC X EG (For New Zealand) EG-EC (For Australia)
Masih and Masih [39]
Korea and Taiwan
1961–1990
Johansen cointegration test, VECM Granger causality test
Long run relationship and mutual causality exists among EC, EG and price level.
Masih and Masih [38]
6 Asian countries
1955–1991
Johansen cointegration test, VECM Granger causality test
Long run relationship between EG and EC exists only in Pakistan, India and Indonesia EC2EG (For Pakistan)
Kahsai et al. [26]
40 SSA countries
1980–2007
Pedroni cointegration test, Granger causality test
Direct relationship between energy demand and income level Low income countries: EC2EG Middle income countries: EC X EG
R. Komal, F. Abbas / Renewable and Sustainable Energy Reviews 44 (2015) 211–220
Author
Note: EC¼ Electricity consumption, EP ¼ Electricity price, X¼ No existence of causality.
215
216
R. Komal, F. Abbas / Renewable and Sustainable Energy Reviews 44 (2015) 211–220
3.3. Energy–growth nexus Discussion on growth–energy nexus was initiated by Kraft and Kraft [34] who concluded that economic growth was followed by rising energy demand in the United States (US) over the 1947– 1974 period. Afterwards, a number of researchers have examined the relationship through two way Granger causality, leading to four testable hypotheses, (a) growth hypothesis, (b) conservation hypothesis, (c) feedback hypothesis and (d) neutrality hypothesis [8,24,26,31,42,53]. Causality relationship is different for different countries due to a number of factors including model estimation techniques, problems caused by non-stationary data, model specification issues, choice of variables, study period, sample selection, development level of country being studied [19,26,38,54]. Omri and Kahouli [43] assert that causality between the two variables may be mutually determined, as higher economic growth needs more energy consumption, as well as efficient energy consumption needs higher level of economic growth, therefore, causality direction may not be judged earlier. Studies on this nexus have explored relationship in terms of cointegration and causality (see Table 3).
4. Theoretical framework and model Financial development affects growth of an economy [32,49,60], it means that it may influence energy consumption through economic growth channel. Furthermore, financial development increases the efficiency of a country's financial system [49]. Well-established financial system boosts the effectiveness as well as the efficiency of financial institutions through enhancing innovation in financial services delivery, reduction of information cost, efficient management of complex and risky transactions, enhancement of transparency between borrowers and lenders; thereby guaranteeing profitability of investment [2,23,52]. All these factors improve business investment and economic activities, thereby boosting domestic production and economic growth. This in turn leads to increase energy consumption [49,50,52,54]. However, energy consumption may be relatively reduced if use of efficient technology is encouraged. Shahbaz et al. [54] has asserted that entrepreneurs are the key agents behind innovation and technological progress in a free market system. This means that the impact of financial development on energy consumption depends on the efficiency of overall system that includes quality of labor, capital, technology, investment environment and government sector policies and institutions. Impact can either be positive or negative depending on whether the economic growth occurs in an efficient or not i.e. if less energy is consumed to produce more or the same level of output or not. Less energy will be consumed when in an economy industrialist or business can upgrade their technology through easily credit availability. Hence, emphasizing the role of financial sector in bringing the efficiency in energy utilization while improving growth in an economy. Available literature is directly linking financial development with energy consumption by inserting both variables in a single equation. This research study employs system GMM technique (simultaneous equation model) to estimate separate equations to capture the effect of financial development over energy consumption through economic growth. For Pakistan, a positive relationship between energy and financial development is postulated because energy saving mechanisms is not up to the mark. Conceptually, energy consumption in an economy is directly affected by economic growth, energy price and urbanization, while it is indirectly affected by financial development through economic growth channel. Thus, a multivariate framework, based on the literature reviewed, is employed to empirically study the
(indirect) impact of financial development over energy consumption. The following structural model is designed for the purpose of estimation: EC t ¼ β1 þ β 2 Y t þ β3 EP t þ β 4 URBt þ μt
ð1aÞ
Y t ¼ α1 þ α2 FDt þ α3 I t þ α4 Sit þ α5 T t þ α6 Sct þ υt
ð1bÞ
where ECt is energy consumption, Yt is real GDP, EPt represents energy price, URBt is urbanization, FDt is financial development, It denotes investment, Sit is government size, Tt is trade, Sct is schooling (human capital), μt and υt are stochastic error terms, t denotes time period in years. In the above model, two equations have been constructed where Eq. (1a) is the main equation that includes variables directly affecting energy consumption, while Eq. (1b) serves as a channel equation in which economic growth (Yt) serves as a variable through which financial development may influence energy consumption (positively or negatively). The motive for Eq. (1b) is to analyze finance-growth relationship by inserting control variables following literature [22,33,37,13]. Empirical literature has highlighted that increase in income is a contributing factor towards increase in energy consumption in developing economies [52]. As production activities increase, energy requirement for input in production process is created that leads to increase in energy demand, except when economic growth occurs in energy efficient manner that conserves energy [23]. Mehrara and Musai [40] have mentioned that economic activities shift from manufacturing to services sector with economic growth that would reduce energy use reinforcing a hypothesized positive relation between income and energy consumption. Masih and Masih [39] have asserted that change in price level alters energy consumption. So, it can be postulated that energy prices would have significant negative effect over energy consumption in Pakistan [42]. Pakistan's urban population has increased to 69.87 million in 2012–13 from 67.5 million in 2011–12 [21]. Urbanization involves swelling population that participates in economic activities thereby increases energy consumption [25,56]. Shahbaz and Lean [52] assert that rising population in the urban areas boosts energy use due to more household appliances. Hence, it can be postulated that urbanization is directly and significantly linked with increased energy consumption in Pakistan.
5. Data and methodology 5.1. Data Data is annual time series for Pakistan ranging from 1972–2012 because continuous data is available from this period onwards. Secondly, it includes reasonable time length of observations to employ model estimation techniques. Data is transformed into log form for analysis that makes the interpretation of results comparatively easier. Variables used for estimation purpose with their measurement, definition, source, and expected sign are mentioned in Table 4. Financial development on domestic level has two aspects; intermediary development and stock market development. The present study is limited to intermediary development only because data on stock market development based on different indicators in the literature is not available for initial years; rather it is available from late 1980s. Thus, present study uses financial development through intermediary development measured through domestic credit to private sector. This not only measures efficiency of financial intermediaries in credit provision [22], but also indicates private sector activities within the economy. According to Rufael [48], it measures opportunities for new firms, as it has the ability to
R. Komal, F. Abbas / Renewable and Sustainable Energy Reviews 44 (2015) 211–220
217
Table 4 Names, definitions, sources and expected signs of variables. Variables (Measurement)
Definition (Sign)
Sources
Energy consumption (kt of oil equivalent)
Utilizing primary energy that is equal to native production including imports and stock changes, and excluding exports and fuel supplied to aircraft and ships transport. Changes in cost to average consumer of obtaining a basket of goods and services that may or may not be fixed at particular periods. (Negative) It refers to collective gross value added by all local manufacturers in the country including product taxes and excluding subsidies not included in value of the products. (Positive) People residing in urban areas as percentage of total population of the country. (Positive)
WDI [59] GoP [21] WDI [59] WDI [59] WDI [59] WDI [59]
Energy Price (CPI) Gross Domestic Product (GDP in Billion Pakistani Rupees) Urbanization (% of urban population in total population) Financial Development (Domestic credit to private sector as % of GDP) Investment (Gross fixed capital formation as % of GDP) Government Size (Government final consumption expenditure as % of GDP) Trade openness (Trade as % of GDP) Human Capital (Gross Secondary School enrollment)
It means financial capital granted to private sector, i.e. through loans, trade credits, purchases of non-equity securities, and other accounts receivable that set up a claim for reimbursement. (Positive) Involves land developments, machinery, plant and equipment procurement; and construction of roads, railways, and other, including hospitals, schools, offices, private residential dwelling, and commercial and industrial buildings. (Positive) Government current expenses on purchase of goods and services (including compensation of employees). It WDI also includes expenses on national defense and security, but excludes government military expenses that are [59] component of government capital formation. (Positive/Negative) It refers to collective imports and exports of goods and services calculated as a share of GDP. (Positive) WDI [59] It refers to enrollment in secondary education in all programs, regardless of age. (Positive) WDI [59]
scrutinize unviable projects for financing. Furthermore, this measure isolates credit being channeled to government sector. King and Levine [33] point out that financial system that channels funds to the private sector firms will possibly offer more services as compared to financial systems issuing credit to government sector or state owned enterprises. However, range of variables is used in literature that measure different aspects of intermediary development, but each measure has its pros and cons. Take the example of size that is measured by liquid liabilities as a share of GDP. King and Levine [33] mention that pure size of the financial system may not be robustly linked with financial services like risk management and information processing; it just indicates the volume of financial sector; so it may not be a realistic measure. Energy price data is not available for Pakistan. For this reason, researchers have used consumer price index as proxy for energy price [see inter alia; 26,29, 39,42,49,58]. However, Sadorsky [49] has mentioned that CPI does not seem to be a good proxy for energy price; hence, in the present study, it is proxied by consumer price index (CPI) with respect to group i.e. Fuel and Lighting. This indicator is a reasonable proxy of energy prices, since it considers price of fuel and lighting only, while CPI considers the prices of other commodity groups and services as well. Splicing method is used to make year 2005 as common base year (i.e. 2005¼100).
estimation techniques that involves instrumental variables may lead to the attainment of consistent and unbiased parameter estimates. Instrumental variables provide a set of variables that are correlated with independent variables of the equation but are uncorrelated with disturbances. Instruments eliminate the correlation between independent variables and disturbances. Therefore, estimates obtained are reliable and consistent. Arellano and Bond [10] and Arellano [9] proposed Generalized Method of Moments (GMM) estimator that is both single equation and system estimator. It is preferred over other estimators of its class because of several reasons. Firstly, GMM offers a simple substitute to other estimators, particularly when it is problematic in writing maximum likelihood estimator. Secondly, GMM covers many standard estimators, thereby offers valuable framework for their evaluation and comparison. Thirdly, GMM is a robust estimator since it does not require information about accurate distribution of error terms. Fourthly, GMM is asymptotically unbiased and consistent estimator regardless of weighting matrix used. Separate instruments are employed for both equations of structural model that are the lagged values of the variables included in that particular equation.
5.2. Methodology
Instrumental variable estimation technique, that is Generalized Method of Moments (GMM), has been employed in this paper to jointly estimate the parameters of the structural model. Separate instruments have been used for both the equations of structural model that are the lagged values of the variables. Results of both the equations are reported in Table 5.
Earlier studies have employed cointegration and causality approaches to estimate structural parameters of a single equation model. Those techniques allow for estimation of relationship in the long run and short run. The present study focuses to capture the indirect effect of financial development on energy consumption through economic growth channel. Instrumental variable estimation technique such as Generalized Method of Moments (GMM) has been used for estimation of parameters. Our approach is to estimate structural parameters while in the estimation of structural model, economic growth variable is treated as endogenous while financial development variable is treated as exogenous. Endogenous variables and disturbances are mutually correlated in simultaneous equation models that create the problem of simultaneity or endogeneity bias. Consequently, inconsistent and biased parameter estimates are obtained using ordinary least square (OLS) regressions that leads to violation of one of the assumptions of classical linear regression model (CLRM). However, the use of
6. Results and discussion
6.1. Energy consumption Parameter estimates of the energy consumption equation (1) are reported in the second column of Table 5. The results are consistent with existing findings in the literature (Jamil and Ahmed [24], Kahsai et al. [26], Masih and Masih [38], Masih and Masih [39], Tang and Tan, [58]). Estimated coefficient of Eq. (1a) is significantly positive. It shows that one percent increase in economic growth has led to about 0.57 percent increase in energy consumption. Energy price has negative and statistically significant effect over energy consumption which shows that energy consumption decreases by about 0.03 percent with increase in
218
R. Komal, F. Abbas / Renewable and Sustainable Energy Reviews 44 (2015) 211–220
Table 5 GMM estimates for the model. Independent variables
Economic growth Energy price Urbanization Financial development Investment Government size Trade School (Human capital) AR(1) AR(2) R2 Adjusted R2 J-statistic (p-value)
Dependent variables Energy consumption (1a)
Economic growth (1b)
0.568nnn (21.116) 0.027nn (-2.193) 2.465nnn (13.097) – – – – – 0.942 [0.000] 0.280 [0.35] 0.9984 0.9981 0.090 (1.000)
– – – 0.043nnn (4.923) 0.048 (1.413) 0.0008 (0.053) 0.044nn (2.122) 0.265nnn (7.749) 0.982 (0.000) – 0.9989 0.9987 –
Note: Column (1) explains main equation of the model. Column (2) explains channel equation of the model. *** and **, indicate significance at 1 and 5 percent respectively. Robust t-statistics are reported in parenthesis. P-values for autoregressive tests are shown in square brackets.
energy price by one percent. This result is in line with economic theory that as price of commodity increases, its consumption decreases. However, the estimated coefficient is too small which indicates that the energy consumption in the country has dropped very little with increase in energy prices. This is justifiable in the current energy scenario where energy demand far exceeds the supply due to which energy prices are increasing. Also, energy is highly subsidized for low end consumers in Pakistan, and hence any change in energy prices is likely to have little impact on energy consumption as is evident from our empirical estimates. Urbanization is significantly and positively linked with energy consumption. Positive relation of urbanization with energy consumption is in accordance with past studies (for instance, Parikh and Shukla [45], Poumanyvong et al. [47], Sadorsky [51], Zhang and Lin [62]). Possible justification for somewhat high value of coefficient is that urban residents are relatively high consumers of energy goods due to TV, refrigerator, and mobile phones etc. With reference to the income level, the present study supports the findings of Poumanyvong et al. [47] and Sadorsky [51] who reported that urbanization is positively linked with energy consumption for developing countries like Pakistan and opposes the finding of Poumanyvong and Kaneko [46] that urbanization is negatively linked with energy consumption for high income countries. 6.2. Economic growth The results of economic growth equation are reported in column three of Table 5. Estimates show that financial development is positively and significantly linked with economic growth. One percent increase in financial development led to 0.04 percent increase in GDP in Pakistan. Possible justification for the small effect of financial development over economic growth is that only one indicator of financial development has been employed in the present study rather than a range of indicators. Underlying intuition regarding this effect is that when financial development happens, investor's confidence in the financial system grows which incentivizes their investment in the productive or economically viable projects via financial institutions, due to which real output increase and economic growth occurs. Consequently, income level improves that causes the aggregate demand to rise leading to higher energy demand.
Other determinants of economic growth are government size and human capital (enrollment at secondary school). Economic growth increases with increase in both the variables, but this relation is insignificant for government size and significant for secondary school enrollment. Review of literature shows that government consumption expenditure (size) has both positive and negative link with economic growth. The insignificant positive impact of government size with economic growth is consistent with Barro [12] who mentioned that government consumption expenditure particularly non-productive one (such as on defense) would have negative impact over economic growth because it has no direct impact on private productivity. Rather it lowers saving and growth as well as increases the cost of inputs through distorting effects such as high tax rates and government expenditure programs. However, this impact may be positive when it comes to productive government consumption expenditure (such as on services that act as input for private production and on implementation of property rights). Positive and highly significant effect of schooling (human capital) on economic growth is in line with Barro [13] who asserted that higher human capital matters for economic growth because of the evolution of new products and ideas that trigger technological innovation in an economy. Though investment is insignificantly related with economic growth, and trade has a positive and significant impact on economic growth in Eq. (1b). Results are consistent with Anwar and Sun [7]; Beck and Levine [14]; Levine et al. [37]. Insignificant impact of investment over economic growth is also justifiable since the share of public investment in Pakistan is quite large and this sector is quite inefficient. Investment is negligible in the productive sectors. Most of the time, government is borrowing from the central bank to meet the current expenditures. Political instability in the country is responsible for low return on investment. Problem of autocorrelation is normally encountered in time series data. This has been resolved by using autoregressive (AR) process.4 High p-value of J-statistics shows validity of instruments used in the estimation of structural equations (see Table 5). 6.3. Summary of channel effect The present study has gone for a different strategy from the existing literature that it has captured the impact of financial development over energy consumption through economic growth by working on a system of equations. However, the methodological approach used in the study allows estimating elasticity in the short run only in contrast with single equation estimation that indicates long run relationship as well. Two coefficients of interest are simultaneously put equal to zero under the null hypothesis i.e., rejection of null hypothesis indicates significance of joint coefficient determined by the test (see Table 6). Wald test is used to test the joint significance of two terms estimated in Eqs. (1a) and (1b). It computes test statistic based on unrestricted regression by imposing restrictions on the estimated coefficients and follows chi-square distribution under the null hypothesis of no relationship [10]. This test can be employed to test multiple parameters simultaneously. Wald test explored product of elasticity of effect of financial development on economic growth and of economic growth on energy consumption that came out to be 0.024 (see column 3 Table 6). It shows that one percent increase in financial development resulted in 0.024 percent rise in energy consumption 4 AR (1) is first order autoregressive process which means that present value is depending on preceding one value. AR (2) is second order autoregressive process which means that present value is depending on preceding two values. According to Arellano and Bond [10], AR (1) and AR (2) are test for autocorrelation in first differenced errors.
R. Komal, F. Abbas / Renewable and Sustainable Energy Reviews 44 (2015) 211–220
219
Table 6 Effect of financial development on energy consumption. Channel variable
Effect of financial development on channel variable (1)
Effect of channel variable on energy consumption (2)
Effect of financial development on energy consumption (3)
Economic growth Wald test (p-value)
0.043
0.568
0.024nnn (4.525) 20.474 (0.000)
Note: nnn
Indicates significance at 1% level. t-statistics are reported in parenthesis.
in the short run which is highly significant. In short, financial development affects energy consumption positively and significantly through economic growth channel. This positive relationship between financial development and energy consumption is in line with previous studies (Coban and Topcu [17], Islam et al. [23], Sadorsky [49], Sadorsky [50], Shahbaz and Lean [52]). On the contrary, the present study negates the finding of Kakar et al. [27] who reported the absence of significant short-run relationship between financial development, energy consumption, and economic growth in Pakistan. Results show that Pakistan's economy has experienced increase in its energy consumption because of financial development that shows that economic growth has not occurred in an efficient manner with respect to energy use. Funds funneled to entrepreneurs for carrying out business activities has played a pivotal role in increasing both economic growth and energy consumption in the last four decades. This result has the implication for energy consumption of country.
7. Conclusion and policy implication Pakistan's financial sector has seen notable and unprecedented growth particularly with respect to the banking sector since 1990s [30]. Islam et al. [23] point out that emerging economies that are developing their financial sector should anticipate growth in energy demand in addition to that the one caused by increasing income. This means that financial development also has influence over energy consumption because of its impact on economic growth. For such economies, estimation of energy demand without considering financial development would provide inaccurate picture. There has been increasing demand for energy over the years in Pakistan because of economic and financial development that has not been fulfilled due to which economy suffered, since lack of sufficient energy supply has suppressed growth rate of Pakistan's economy to a great extent. Rapid population growth and industrial growth has led to increase in demand for energy sources but the supply of energy has not increased with similar pace to meet the growing demand. The objective of present study is to investigate empirically the effect of financial development on energy consumption through economic growth channel in Pakistan. For this purpose, we estimated a system of equations to investigate this indirect relationship. Energy price and urbanization are also included in model pertaining to their potential influence over energy consumption. The structural energy consumption model has been estimated by employing Generalized Method of Moments (GMM) estimation technique. Annual time series data covering period 1972–2012 is taken for study. Short run parameter estimate of individual equation of the model, confirmed the existence of the hypotheses constructed for the present study. Results for the main equation indicated that energy consumption was positively and significantly linked with economic growth and urbanization, while negatively and significantly linked with energy prices. The results for channel equation indicated the positive and significant effect of financial development on economic
growth along with control variables namely, investment, government size, trade, and human capital, where trade and school affect economic growth in a positive and significant manner, while the effect of size and investment is positive but insignificant. Positive and significant effect of financial sector development over energy consumption suggests that this increase in energy consumption emerging from increase in financial development must be considered at the time of planning energy consumption for Pakistan economy; failing which would lead to the underestimation of energy consumption that is alarming for sustainable economic growth. It is also possible that policies to save energy may not meet desired targets if policy makers fail to consider the impact of financial development over energy consumption. However, in order to reverse/avoid energy and overcome energy crisis, financial sector can play a vital role in terms of intermediary development. Banks can ensure that funds are channeled primarily to those entrepreneurs who come up with innovative ideas and advanced technologies. Human capital is central in bringing technological innovation due to their involvement in generation of new research ideas. Funds should be funneled to research and development (R&D) sector of well-reputed firms to innovate and promote energy saving technologies. Efficient energy use would reduce cost of production for firms. In this manner, economic growth will be accelerated along with decline in overall energy consumption. Consequently, this would make cheap and easy accessibility to energy efficient products at household level that will further reduce energy consumption. However, it is to be noted that financial sector development comes with financial reforms. Provision of necessary infrastructure is important for sound financial development policy. Credible and reliable support system is essential to guarantee stable performance of financial sector, since accomplishment of financial sector policies may rely on effectiveness of institutions implementing them. Pakistan is an energy dependent economy whose economic growth has become crippled in the absence of adequate energy supply. Continuous supply of energy is required to support economic activities in Pakistan so that the economy may be able to achieve potential level of economic growth. Furthermore, urbanization has a strong positive influence over energy consumption which means that the energy requirement is likely to increase in future to meet the growing demand of increasing urban population. Government of Pakistan should take serious steps to ensure efficient utilization of available energy resources along with up-gradation of existing energy production capacity at macro level to mitigate energy shortage problem. Investment in renewable energy resources would promote access to cheap energy for carrying out economic activities, for which financial institutions can assist along with government. Moreover, rising energy prices have adverse effect for economy since production cost is increased that becomes a reason for the loss of competitive advantage for manufacturers in foreign market. Promotion of cheap renewable energy resources can mitigate this problem. This is necessary for achieving optimum and sustainable economic growth in Pakistan. The present study uses only one indicator for financial sector development. Therefore, it would be interesting to develop
220
R. Komal, F. Abbas / Renewable and Sustainable Energy Reviews 44 (2015) 211–220
financial development index in which foreign direct investment along with domestic investment and other indicators of financial development may also be included. References [1] Abbas F, Choudhury N. Electricity consumption-economic growth Nexus: an aggregated and disaggregated causality analysis in India and Pakistan. J Policy Model 2013;35(4):538–53. [2] Ahmed AD. Effects of financial liberalization on financial market development and economic performance of the SSA region: an empirical assessment. Econ Model 2013;30:261–73. [3] Ahmed SM, Ansari MI. Financial sector development and economic growth: the South-Asian experience. J Asian Econ 1998;9(3):503–17. [4] Ali G, Nitivattananon V. Exercising multidisciplinary approach to assess interrelationship between energy use, carbon emission and land use change in a metropolitan city of Pakistan. Renew Sustain Energy Rev 2012;16:775–86. [5] Al-mulali U, Sab CNBC. The impact of energy consumption and CO2 emission on the economic growth and financial development in the sub Saharan African countries. Energy 2012;39:180–6. [6] Ang JB. What are the mechanisms linking financial development and economic growth in Malaysia? Econ Model 2008;25:38–53. [7] Anwar S, Sun S. Financial development, foreign investment and economic growth in Malaysia. J Asian Econ 2011;22:335–42. [8] Apergis N, Tang CF. Is the energy-led growth hypothesis valid? New evidence from a sample of 85 countries Energy Econ 2013;38:24–31. [9] Arellano M. On Testing of Correlation Effects with Panel Data. J Econom 1993;59:87–97. [10] Arellano M, Bond S. Some tests of specifications for panel data: Monte Carlo evidence and an application to employment equations. Rev Econ Stud 1991;58:277–97. [11] Asif M. Sustainable energy options for Pakistan. Renew Sustain Energy Rev 2009;13:903–9. [12] Barro RJ. Government spending in a simple model of endogenous growth. J Polit Econ 1990;98(S5):103–25. [13] Barro RJ. Economic growth in a cross section of countries. Q J Econ 1991;106 (2):407–43. [14] Beck T, Levine R. Stock markets, banks, and growth: panel evidence. J Bank Finance 2004;28:423–32. [15] Belaid F, Abderrahmani F. Electricity consumption and economic growth in Algeria: a multivariate causality analysis in the presence of structural change. Energy Policy 2013;55:286–95. [16] Bojanic AN. The impact of financial development and trade on the economic growth of Bolivia. J Appl Econ 2012;15(1):51–70. [17] Coban S, Topcu M. The nexus between financial development and energy consumption in the EU: a dynamic panel data analysis. Energy Econ 2013;39:81–8. [18] EIA. International energy outlook. Washington DC: Energy Information Administration; 2013 [accessed 18.09.13]. [19] Fatai K, Oxley L, Scrimeour FG. Modeling the causal relationship between energy consumption and GDP in New Zealand, Australia, India, Indonesia, The Philippines and Thailand. Math Comput Simul 2004;64:431–45. [20] GoP. Economic Survey 2009–2010, Finance Division. Economic Advisor's Wing, Islamabad: Government of Pakistan; 2010. [21] GoP. Economic Survey 2012–2013, Finance Division. Economic Advisor's Wing, Islamabad: Government of Pakistan; 2013. [22] Gregorio JD, Guidotti PE. Financial development and economic growth. World Dev 1995;23(3):433–48. [23] Islam F, Shahbaz M, Ahmed AU, Alam MM. Financial development and energy consumption nexus in Malaysia: a multivariate time series analysis. Econ Model 2013;30:435–41. [24] Jamil F, Ahmad E. The relationship between electricity consumption, electricity prices and GDP in Pakistan. Energy Policy 2010;38:6016–25. [25] Jones DW. How urbanization affects energy-use in developing countries. Energy Policy 1991;19(7):621–30. [26] Kahsai SM, Nondo C, Schaeffer PV, Gebremedhin TG. Income level and the energy consumption–GDP nexus: evidence from Sub-Saharan Africa. Energy Econ 2012;34:739–46. [27] Kakar ZK, Khilji BA, Khan MJ. Financial development and energy consumption: empirical evidence from pakistan. Int J Trade Econ Finance 2011;2(6):469–71. [28] Kessides IN. Chaos in power: Pakistan's electricity crisis. Energy Policy 2013;55:271–85. [29] Khan MA, Ahmad U. Energy demand in pakistan: a disaggregate analysis. Pak Dev Rev 2008;47(2):437–55. [30] Khan MA, Khan S. Financial sector restructuring in pakistan. Lahore J Econ 2007 (Special edition). [31] Khan MA, Qayyum A. Dyn Model Energy Growth South Asia 2007. [32] Khan MA, Qayyum A, Sheikh SA. Financial development and economic growth: the case of pakistan. Pak Dev Rev 2005;44(4–2):819–37. [33] King RG, Levine R. Finance and growth: Schumpeter might be right. Q J Econ 1993;108(3):717–37.
[34] Kraft J, Kraft A. On the relationship between energy and gNP. J Energy Dev 1978;3:401–3. [35] Lee CC, Chang CP. Energy consumption and economic growth in Asian economies: a more comprehensive analysis using panel data. Resour Energy Econ 2008;30:50–65. [36] Levine R. Bank-based or market-based financial systems: which is better? Working paper #9138. 1050 Massachusetts Avenue, Cambridge, MA 02138, USA: National Bureau of Economic Research; 2002. [37] Levine R, Loayza N, Beck T. Financial Intermediation and Growth: Causality and causes. J Monetary Econ 2000;46:31–77. [38] Masih AMM, Masih R. Energy consumption, real income and temporal causality: results from a multi-country study based on cointegration and error-correction modeling techniques. Energy Econ 1996;18:165–83. [39] Masih AMM, Masih R. On the temporal causal relationship between energy consumption, real income, and prices: some new evidence from Asian-Energy dependent NICs based on a multivariate cointegration/Vector error-correction approach. J Policy Model 1997;19(4):417–40. [40] Mehrara M, Musai M. Energy consumption, financial development and economic growth: an ARDL approach for the case of Iran. Int J Bus Behav Sci 2012;2(6):92–9. [41] Mudakkar SR, Zaman K, Shakir H, Arif M, Naseem I, Naz L. Determinants of energy consumption function in SAARC countries: balancing the odds. Renew Sustain Energy Rev 2013;28:566–74. [42] Odhiambo NM. Energy consumption, prices and economic growth in three SSA countries: a comparative study. Energy Policy 2010;38:2463–9. [43] Omri A, Kahouli B. Causal relationships between energy consumption, foreign direct investment and economic growth: fresh evidence from dynamic simultaneous-equations models. Energy Policy 2014;67:913–22. [44] Pakistan energy year book. Hydrocarbon Development Institute of Pakistan. Publication and information dissemination; 2009. Available from 〈http://hdip. com.pk/contents/Publication-Information-Dissemination/20.html〉 [accessed 18.09.13]. [45] Parikh J, Shukla V. Urbanization, energy use and greenhouse effects in economic development; results from a cross-national study of developing countries. Glob Environ Change 1995;5(2):87–103. [46] Poumanyvong P, Kaneko S. Does urbanization lead to less energy use and lower CO2 emissions? A cross-country analysis Ecol Econ 2010;70: 434–444. [47] Poumanyvong P, Kaneko S, Dhakal S. Impacts of urbanization on national transport and road energy use: evidence from low, middle and high income countries. Energy Policy 2012;46:268–77. [48] Rufael WY. Re-examining the financial development and economic growth nexus in Kenya. Econ Model 2009;26:1140–6. [49] Sadorsky P. The impact of financial development on energy consumption in emerging economies. Energy Policy 2010;38:2528–35. [50] Sadorsky P. Financial development and energy consumption in Central and Eastern European frontier economies. Energy Policy 2011;39: 999–1006. [51] Sadorsky P. Do urbanization and industrialization affect energy intensity in developing countries? Energy Econ 2013;37:52–9. [52] Shahbaz M, Lean HH. Does financial development increase energy consumption? the role of industrialization and urbanization in Tunisia Energy Policy 2012;40:473–9. [53] Shahbaz M, Lean HH, Farooq A. Natural gas consumption and economic growth in Pakistan. Renew Sustain Energy Rev 2013;18:87–94. [54] Shahbaz M, Khan S, Tahir MI. The dynamic links between energy consumption, economic growth, financial development and trade in China: fresh evidence from multivariate framework analysis. Energy Econ 2013;40: 8–21. [55] Shiu A, Lam PL. Electricity consumption and economic growth in China. Energy Policy 2004;32:47–54. [56] Solarin SA, Shahbaz M. Trivariate causality between economic growth, urbanization and electricity consumption in Angola: cointegration and causality analysis. Energy Policy 2013;60:876–84. [57] Tang CF, Tan BW. The linkages among energy consumption, economic growth, relative price, foreign direct investment, and financial development in Malaysia. Qual Quant 2012:1–17. http://dx.doi.org/10.1007/s11135-012-9802-4 (Available from). [58] Tang CF, Tan EC. Exploring the nexus of electricity consumption, economic growth, energy prices and technology innovation in Malaysia. Appl Energy 2013;104:297–305. [59] WDI. World Development Indicator, The World Bank Group. Online available at: 〈http://data.worldbank.org/data-catalog〉; 2013. [60] Wu JL, Hou H, Cheng SY. The dynamic impacts of financial institutions on economic growth: evidence from the European Union. J Macroecon 2010;32:879–91. [61] Zaman K, Khan MM, Ahmad M, Rustam R. Determinants of electricity consumption function in Pakistan: old wine in a new bottle. Energy Policy 2012;50:623–34. [62] Zhang C, Lin Y. Panel estimation for urbanization, energy consumption and CO2 emissions: a regional analysis in China. Energy Policy 2012;49:488–98.