Interaction Between Environmental Kuznet Curve and Urban Environment Transition Hypotheses in Malaysia

The aim of this paper is to examine the interconnection bewteen the environmental Kuznet curve (EKC) and urban environment transition hypotheses in the Malaysian economy. Economic growth, CO2 emissions, energy consumption, financial development (FD), urbanization variables for the 1971-2013 period, F-bound tests, and multivariate Granger causality methods are used. The long-run relationships among the above series of variables are examined. Also, the findings confirm the existence of an inveretd EKC hypothesis in the Malaysian economy. Furthermore, the study implies that long-run urban sprawl can create environmental and health burdens in Malaysia. Moreover, the causality analysis finds bidirectional causality between CO2 emissions and energy consumption, between economic growth and urbanization, and from economic growth to FD. The overall results suggest that a rapid urbanization process can help to reduce the level of pollution and energy consumption by employing technical innovation and ecological modernization. In addition, the increased energy efficiency, implementation of energy savings projects, energy conservation, and energy infrastructure outsourcing reduce the level of pollution produced by urban areas. Also, alternative biofuels can reduce the emission intensity and fulfill future energy needs.


INTRODUCTION
Over the past two decades, many researchers have analyzed the relationships between energy consumption and macroeconomics variables. Numerous studies have examined the causal relationship between energy consumption and independent variables such as economic growth, financial development (FD), employment, and population Othman, 2014 andOztruk, 2015, 2016;Saidi and Hammami, 2015;Tang and Tan, 2015;Mahalik et al., 2017;Yue et al., 2019). A rapid pace of development is accompanied by a continuous rise in CO 2 emissions. Recently, the International Energy Agency (IEA) indicated that global CO 2 emissions from the energy sector had dropped in 2014 for first time during the last 40 years. The IEA suggested that this reduction in emissions was not tied to any economic downturn. Global emissions stood at 32.3 billion tonnes in 2014, unchanged from the preceding year. The study also attributed this decline in emissions to changing patterns of energy consumption in China and Organization for Economic Cooperation and Development (OECD) countries. In China in 2014, the greater generation of electricity was from renewable sources, such as hydropower, solar, and wind and less burning of coal. Furthermore, in OECD economies, recent efforts to promote more sustainable growth, including greater energy efficiency and more renewable energy, are producing the desired effect of decoupling economic growth from greenhouse gas emissions (International Energy Agency, 2015a).
In other words, the term urbanization does not fully reflect how many people settle in urban areas. The International Energy Agency (2014) also highlighted the high concentration of people, economic activities, and services in relatively small settlements. Urban economies of scale can promote better access to services, greater prosperity, and positive changes in lifestyle (Bekhet and Othman, 2017). However, it can also lead to increased slums and squatter settlements, social alienation, and environmental pollution. The positive and negative impact of urbanization is not distributed equally among the urban population; the rich and powerful draw more benefits from the positive effects and are better protected against adverse effects than the poor and marginalized. In the Association of Southeast Asian Nations (ASEAN) region, the countries are more ambitious in using economic integration as a tool for achieving broader regional prosperity and greater global competitiveness. However, many challenges remain despite the unique opportunities offered by three global megatrends: the ongoing expansion of cross-border trade, unprecedented urbanization, and the advent of multiple disruptive technologies (Woetzel et al., 2014). To support the urban planning, the Asian Development Bank is promoting city clusters by financing urban infrastructure to link cities and towns within a region to improve their economic development potential (Choe and Laquian, 2008).
In response to various economic and environmental challenges, the Malaysian economy has shifted its drivers of economic growth from energy-intensive industries toward services and higher value manufacturing, in addition to reforming fossil-fuel subsidies. Improvements in energy efficiency are also the result of implementation of dedicated measures that stem from Malaysia's commitment as a member of the Asia-Pacific Economic Cooperation countries to reduce its energy intensity. However, even with these ongoing efforts, challenges still lie ahead so careful future planning is needed to achieve sustainable growth. Figure 2 shows the upward trends of energy and CO 2 emissions in the Malaysian economy. In 2014, energy consumption and CO 2 emissions reached 145.31 billion ktoe and 0.4 billion kt, respectively. This means that a high gross domestic product (GDP) can be achieved through rapid development, but such development may threaten the environment due to the associated high CO 2 emissions. Fossil fuels in Malaysia's energy mix continue to be dominant, with the energy-related CO 2 emissions expected to double in the New Policies Scenario, increasing from 211 million tonnes (Mt) in 2013 to almost 430 Mt by 2040 (International Energy Agency, 2015b).
Urban growth is another driver of environmental challenges stemming from the rapid economic development in Malaysia.
In the past few years, Malaysia has achieved an average of 4%-5% economic growth; in addition, rapid urbanization has put considerable pressure on energy demand. Figure 3 shows the percentage share (%) of the rural and urban population in Malaysia in 1960-2018. The figure shows a huge gap between the rural and urban population growth rate; in the beginning, the rural population was larger than the urban population, but after reaching equilibrium in 1991, the pace of urbanization outstripped rural population growth. This urban sprawl has introduced various health and environmental issues in Kuala Lumpur and Klang Valley, where the annual air pollution cost is around US $1.0-1.6 billion (Brandon, 1994;Marcotullio and Lee, 2003). Urban expansion also brings economic prosperity as it enables Malaysian citizens to pursue a higher quality of life. To enhance the quality of life, during the Tenth Source: United Nations, Department of Economic and Social Affairs, Population Division (2018). World urbanization prospects Malaysia Plan, initiatives to enhance inclusivity included elevating the livelihood of the bottom 40% household income group (B40), building a progressive and more inclusive society, ensuring access to basic infrastructure and services, and promoting corridors as an engine of growth (Economic Planning Unit, 2010). Efforts were also undertaken to elevate the quality of life of rural households and enhance the economic participation of urban households through income-generating activities and micro-enterprise support programs.
Financing activities have played a vital role in the new planning projects regarding renewable energy and sustainable economic growth. The overall total gross financing raised by the private sector through the banking system and the capital market amounted to RM281 billion in the second quarter of 2014 (Bank Negara Malaysia, 2015). The development of financial markets indicates that a country is strong in an economic sense. Financial deepening plays a significant role in transferring created funds to real sectors. In this case, the Malaysian financial system has exhibited continued improvement in the global and domestic financial markets. Overall, the financial intermediation has remained well supported by sound financial institutions, orderly financial market conditions, and sustained confidence in the financial system. In Figure 4, two important financial indicators, domestic credit to the private sector (DP) and domestic credit provided by financial sectors (DF), are shown as a percentage of GDP. During 1960-2018, DP and DF show an annual growth rate of 4.52% and 4.47%, respectively. A slump in the financial market Source: World Bank Database (2019)   occurred during the Asian and global financial crises, but in 2011 the financial sector started to stabilize. Volatility in the domestic financial market also arose from uncertainties in the external environment.
The main objective of this paper is to examine the interconnection between environmental Kuznet curve (EKC) and urban environmental transition (UET) theories in Malaysia. This paper makes several contributions: Examining the energy and emission nexus by adding the urbanization perspective can shed light on problems in Malaysian cities caused by high concentrations of industry. Moreover, air, water, noise, and solid waste pollution problems have increased rapidly and can have dramatic impacts on the life and health of city inhabitants. This single-country study uses both a static and a dynamic framework to model the impact of urbanization on CO 2 emissions to compare whether Malaysia is improving through new energy policies with increasing economic growth or whether urban sprawl has impeded sustainable development. However, the understandable determination of the Malaysian government to deliver modern development to its citizens means that urban places will become critical for innovation and development in the low-carbon transition. Importantly, the results of this study can drive policy responses at the national level to this global dilemma. The current study adds a powerful policy thrust that emphasizes the role of technological innovation, such as improved financial policies in raising funds for greening cities, development of compact cities, reduction of auto-dependent systems, and green building (Brown and McGranahan, 2016). Last, the current study results can enhance ongoing efforts with respect to alternative sources of energy, and the development of modernism means that Malaysia must prioritize these policies for adaption in urban places so that the majority of people will be able to live a sustainable life style that can withstand the upcoming challenges.
The remainder of the paper is structured as follows: Section 2 describes both the theoretical and empirical literature. Section 3 explains the data and model construction used in the analysis. Section 4 analyzes the empirical findings and offers relevant discussion. Finally, the concluding remarks and policy implications of the findings are outlined in Section 5.

THEORETICAL BACKGROUND AND PREVIOUS STUDIES
A large body of research can be found on the relationships among GDP, CO 2 emissions, FD, urbanization, and energy use covering a range of countries and using different methodologies. This section classifies this research into two sub-sections: EKC and UET theories and previous findings.

EKC and UET Theories
The environmental implications of economic growth and the planning implications for reducing environmental burdens are under hot debate. The main challenge in this context is how various changes affect the energy and environment perspective. The EKC hypothesizes that income inequality first rises and then falls as economic development proceeds (Kuznets, 1955). This term was coined in the 1980s in response to the finding that environmental problems display a similar pattern, initially increasing with economic growth and then declining. In the early stages of economic growth, degradation and pollution increase, but beyond some threshold of income per capita, which will vary for different indicators, the trend reverses, so that at high income levels economic growth leads to environmental improvement. This implies that the environmental impact indicator is an inverted U-shaped function of income per capita. In other words, this application leads to modernization and the affluence concept, where as societies continue to evolve to higher stages of development, environmental damage becomes more important and societies seek more environmentally sustainable practices (Dietz and Rosa, 1997;. Second, based on UET theory, urban growth is also usually blamed for environmmental problems. Basically, there is a relationship between rapid sprawl of the urban population and environmental degradation. As cities continue to become wealthier, industrial pollution problems may be lessened via environmental regulations, technological innovation, or changes in the energy mix and production methods. Last, the urbanization concept includes compact city theory, mostly focused on the benefits of increased urbanization leading to changes in demographics (Zhou et al., 2015). Higher urban density facilitates economies of scale for public infrastructure but at the same time lead to lower environmental damage.
From the first time when EKC was popularized in the World Development Report of 1992 (World Bank, 1992), it has been evident that not all the environmental burdens exhibit the same type of relationship with per capita income, either cross nationally or over time. A great deal of variation in environmental burdens is not related casually or statistically to economic status. Thus, in this context, more specifically as the basis for claiming that economic growth is inherently good for the environment or that economic growth affects the environment generally, the EKC has various implications (Ekins, 1997 and2000;Stern, 2004). Based on such issues, McGranahan and Singsore (1994) and McGranahan et al. (2001) proposed contrasting curves for sanitation, urban air pollution, and carbon emissions. The existing literature points to three theories (ecological modernization, UET, and compact city theories) that are useful for explaining how urbanization can affect the natural environment (Poumanyvong and Kaneko, 2010). The current study employs UET theory, highlighting the interaction with EKC and covering the three perspectives involving demographic, economic, and technological factors. Figure 5 identifies layers of changes in the relationships among affluence, technical innovation, and urban environmental burdens.The three curves indicate that the severity of different environmental burdens changes with affluence. The first curve shows a decline in local burdens and a move toward global burdens, the second shows the rise and fall of the city-regional burden known as EKC, and the third shows the rise of global burdens in the form of CO 2 emissions. The third curve reflects the tendency for urban environmental burdens to be more dispersed and delayed in higher income cities than in lower income cities. In summary, this suggests that in poor cities, environmental challenges are localized, immediate, and health threatening.
In middle-income, rapidly developing cities, environmental burdens are citywide or regional, somewhat more delayed in their impacts, and a threat to both health and ecological sustainability.
In affluent or high-income cities, environmental burdens are global, intergenerational, and primarily a threat to sustainability (McGranahan et al., 2001;Zi et al., 2016).
The EKC and UET theories can applied for different contexts, including economies in rapidly developing Asia (Bai and Imura, 2000;Marcotullio, 2003). These applications either simply describe differences between cities of different income levels or include changes in the speed of transitions. Moreover, some researchers have analyzed the energy consumption and production technology pattern changes under rapid urbanization (Elias and Victor, 2005). However, Figure 5 helps to explain why ecologists are associated with the claim that environmental burdens increase with affluence and economists are associated with EKC. The theories of ecological modernization and UET recognize that urbanization can have both positive and negative impacts on the natural environment, with the net effect being hard to determine a priori (Sadorsky, 2014a and 2014b).

Existing Literature
A close relationship exists between energy cosumption and economic development. With rapid development, the demand for energy also rises. Begum et al. (2015) highlighted this relationship in Malaysia. They demonstrated that both per capita energy consumption and per capita GDP have long-term positive impacts on per capita carbon emissions in Malaysia. Furthermore, Sebri and Ben-Salha (2014) confirmed long-run equilibrium relationships among economic growth, energy consumption, and CO 2 emissions in Brazil, Russia, India, and China (the BRIC countries). In contrast, Kivyiro and Arminen (2014) identified causal links among CO 2 emissions, energy consumption, economic development, and foreign direct investment (FDI) in six sub-Saharan African countries. They found varied results in these countries. Omri and Kahouli (2014) elaborated the FDI and economic growth in 65 countries. They found a positive impact on energy consumption with FDI flows.
The research also expands on the environment and energy in a causality relation. Bastola and Sapkota (2015) elaborated that the existing energy crisis, alleviating poverty, and mitigating the potential impacts of climate change have captured significant attention among policy makers in Nepal. Granger causality tests suggested the presence of long-run bidirectional causality running from energy consumption to CO 2 emissions and vice versa and a unidirectional causality running from economic growth to both CO 2 emissions and energy consumption. These findings implied that policies that boost energy consumption may not spur economic growth but rather are more likely to have adverse effects on the environment. Salahuddin and Gow (2014) suggested that energy consumption and CO 2 emissions Granger cause each other while a unidirectional causal link running from economic growth to energy consumption is also found. Both absolute and relative decoupling occurred in all the Gulf Cooperation Council countries except Saudi Arabia during the study period. Lin and Moubarak's (2014) results showed bi-directional long-term causality between renewable energy consumption and economic growth in China. Caraiani et al. (2015) investigated the causality relationship between energy consumption and GDP in the context of emerging European countries covering 1980-2013. They found that in some specific cases the results implied that energy consumption and economic growth are jointly determined via bi-direction causality (Kalimeris et al., 2014;Farhani et al., 2014;Heidari et al., 2015).
Today, researchers are also considering the FD element in the energy and economic growth relationship. In this context, Komal and Abbas (2015) explored the finance-growth-energy nexus for Pakistan. Their study identified significant implications while considering the impact of FD on energy consumption through the economic growth channel. The result suggested that FD positively and significantly affects energy consumption through the economic growth channel. Furthermore,  highlighted the relationship among FD, CO 2 emissions, and economic growth in Middle East and North Africa (MENA) countries. They found unidirectional causality running from FD to economic growth and from trade openness to CO 2 emissions. Alam et al. (2015) examined the relationship among energy consumption, economic growth, relative prices of energy, FDI, and different FD indicators in the South Asian Association for Regional Cooperation member countries, Bangladesh, India, Nepal, Pakistan, and Sri Lanka, in 1975-2011.The results showed a significant relationship among energy consumption, economic growth, FDI, and FD proxies; Economic growth has posed another key question in economic development related to urbanization, which creates increasing pressure on the energy supply and the natural environment . Many researchers have discussed this issue to bring a better understanding of the relationship among urbanization, energy consumption, and CO 2 emissions (e.g., Poumanyvong and Kaneko, 2010;Al-Mulali et al., 2012;Liu and Xie, 2013;Shahbaz et al., 2014;Wang, 2014;Ren et al., 2015;Wang et al., 2015;Rafiq et al., 2016;Shen and Lin, 2016;Lang et al., 2016;Yang et al., 2016;Bekhet and Othman, 2017;. They highlighted the related research regarding energy consumption patterns and the urbanization process in various studies. Moreover, some of these studies argued that the impact of urbanization on energy consumption is greater than the impact of CO 2 emissions in each region. Poumanyvong et al. (2012) investigated empirically the effects of urbanization on transport energy use and CO 2 emissions with consideration of the different development stages. They suggested that the impact of urbanization on energy use and CO 2 emissions varies across the stages of development. Surprisingly, urbanization decreases energy use in the low-income group, while it increases energy use in the middle-and high-income groups. The impact of urbanization on CO 2 emissions is positive for all the income groups, but it is more pronounced in the middle-income group than in the other groups. Sharma (2011) found that urbanization has a negative impact on CO 2 emissions in high-income countries.
China's economic development and urbanization created increased pressure on the energy supply and the natural environment. Tan et al. (2016) detailed the literature regarding the urbanization process. They found that an energy transition from high-pollution coal to clean electricity is also present in China, although the fundamental transition to renewable energy is still in its infancy. From a regional perspective, urbanization has asymmetric impacts on provincial energy use so that energy policies associated with urbanization should be province-specific. Xu and Lin (2015) confirmed an inverted U-shaped nonlinear relationship between industrialization and CO 2 emissions in the three regions in China.
Urbanization follows an inverted U-shaped pattern with CO 2 emissions in the eastern region and a positive U-shaped pattern in the central region. However, the nonlinear impact of urbanization on CO 2 emissions is insignificant in the western region. Chen et al. (2016) also highlighted the National New Urbanization Plan (2014-2020) unveiled by the Chinese central government, which revealed a new path for urbanization that accommodated unique Chinese characteristics. The most notable aspect was the transfer from land-centered urbanization to people-oriented urbanization.
As an emerging economy, the Malaysian economy is going through a transformation process which has brought various energy and environmental challenges. In this context, Rafiq et al. (2016) analyzed the impact of urbanization and trade openness on emissions and energy intensity in 22 increasingly urbanized emerging economies. They employed three second-generation heterogeneous linear panel models as well as recently developed nonlinear panel estimation techniques allowing for cross-sectional dependence. They found that population density and affluence increase emissions and energy intensity while renewable energy seems to be dormant in these emerging economies, but non-renewable energy increases both CO 2 emissions and energy intensity.  further suggested that ASEAN countries need rational planning of urban development and energy-efficiency improvement strategies. For Malaysia cases,  investigated the impact of urbanization on energy consumption by applying the stochastic impacts by regression on population, affluence, and technology (STIRPAT) method. The F-bound test approach was applied to examine integrating properties and long-run relationships in the presence of structural breaks.
The results validated the existence of cointegration and exposed urbanization as a major contributor to energy consumption. Also, affluence raises energy demand. The causality analysis found that urbanization Granger causes energy consumption. This was justified by , who confirmed that the relationship between urbanization and CO 2 emissions is U-shaped in Malaysia (i.e., urbanization initially reduces CO 2 emissions, but after a threshold level, it increases CO 2 emissions). The causality analysis also suggested that the urbanization Granger causes CO 2 emissions. Utilizing a different framework, Bekhet and Othman (2017) combined the augmented Cobb-Douglas production and ecological modernization theories in a single framework investigate the link between urbanization growth and CO 2 emissions. They revealed the inverted U-shaped relationship between CO 2 emissions and urbanization in the long run and it was reported similar to . Instead of that, various studies on urban planning in Malaysia have also been conducted (Hossain, 2011;Azam et al., 2015;Morris et al., 2016;Shuid, 2016).
The relationship among economic growth, energy, and the CO 2 emissions nexus has been studied extensively regarding EKC. Appendix 1 summarizes various studies in the context of the EKC hypothesis. The existing literature provides panel as well as single-country case studies. These studies provide effective energy policies and emission-reduction strategies. However, most existing studies have focused on the effects of urbanization regarding CO 2 emissions and energy consumption and have not fully considered other important factors such as demographic, growth, and technical advancements in Malaysia. This paper fills the gap in the existing literature by applying the EKC and UET theoretical perspectives in Malaysia for in-depth analysis. Thus, based on the above literature review and to achieve the objective of the current paper, we have formulated the following hypotheses: H 1 : Significant EKC influence exists in Malaysia. H 2 : Significant UET influence exists in Malaysia. H 3 : Significant long-and short-run relationships exist between CO 2 and its determinants in Malaysia. H 4 : Significant long-and short-run causalities exist between CO 2 and its determinants in Malaysia.

MODEL FRAMEWORK AND DATA SOURCES
In the literature, much attention has been paid to Ehrlich and Holdren's (1971) IPAT model, which suggests that the environmental impact (I) is the product of three factors, population (P), affluence (A), and technology (T), in the analysis of the effect of economic activity on energy consumption and the environment. This can be written as in equation (1).
Where I denotes the impact, P represents demographic, A is the economic, and T is the technological factors. The pivotal limitation of IPAT is that it does not permit hypothesis testing because the known values of some terms determine the value of the missing term (York et al., 2003). To overcome the limitations of the IPAT model, Dietz and Rosa (1997) proposed the STIRPAT model as in equation (2).
Non-linear equation (2) can be transformed to become linearized by taking logarithms (L), as shown in equation (3).
where a, b, c, and d represent the coefficients term, ε denotes the error term, and the subscripts i and t display as the country, province, and sectors and time, respectively. The IPAT model allows for assessing factors' impact on the environment (CO 2 ). In the current paper, this model is extended by incorporating various variables, as shown in Figure 6.
To examine the impact of interaction among the EKC and UET hypotheses, energy consumption (EC), economic growth (GDP), FD, and urbanization (UR) in relation to environmental pollution (CO 2 ) are used. Then, Equation (3) is extended as in equation (4).
where β is the intercept and t is the time period, while ψ stands for a residual or error term with constant variance and zero mean that is assumed to be normally distributed and α i s (i= 1,…., 6) are the coefficients of the variables. The EKC hypothesis between the economic growth level and CO 2 emissions (inverted U-shape) and the theory of UET recognize that urbanization can have positive and negative impacts on the natural environment. If urbanization has a statistically insignificant impact on CO 2 emissions, then urbanization will have no meaningful impact on CO 2 emissions. This is consistent with the positive and negative effects of urbanization on CO 2 emissions canceling each other, so the hypotheses regarding urbanization and CO 2 emissions are accepted in Malaysia if α 2 >0, α 3 <0 and α 5 >0, and α 6 <0, respectively. Annual data from the World Bank Development databases (2019) were used for all reserach variables by employing E-views and Microfit packages. This will involve more regression analysis and the sample size consists of only 42 observations, so the critical values are not available to compare the cointegrtaion result in a Pesaran table. However, this can be done by using the Microfit package.
To determine the level of stationarity [I(0), I(1), and I(d)], augmented Fuller [ADF] (1979,1981), Phillip and Perron [P-P] (1988), and Kwiatkowski, Phillips, Schmidt, and Shin [KPSS] (1999) statistical tests are used. To examine the equilibrium relationship among the variables above, the ARDL bounds testing model (Pesaran and Pesaran, 1997;Pesaran and Shin, 1999;Pesaran et al., 2001) is employed. This is because it can be used with a small sample size, it estimates long-run and short-run relationships simultaneously, and it allows for testing the existence of relationships among variables for stationary data at various level, I(0) and I(1), or both (Bekhet and Othman, 2018;Bekhet & Abdul, 2018;Bekhet & Harun, 2017;Hamdi et al., 2014;Pesaran and Shin, 1999;Pesaran and Pesaran, 1997). Equation (5) is constructed to examine the long-run and short-run relationships among the variables. LGDP LGDP LGDP LGDP Where ∆ is the first difference operator, β i s represent the intercepts, ϕ ij s denote the long-run coefficients of the variables, α ij s denote the short-run coefficients, ε it s denote the error terms which are normally distributed, k represents the optimal lag length, s is the lag order, and i,j=1,…, 7. The optimal lag structure of the first differenced regression is selected by the Akaike information criteria.
The • If the calculated F-statistics value is >the upper bounds value I(1), the H 0 is rejected, which means that the variables included in the model share long-run relationships. • If the calculated F-statistics value is <the lower bounds value, I(0), the H 0 is accepted, which means that the variables included in the model do not share long-run relationships.

• If the calculated F-statistics value falls in the range I(0)
≤F-statistics value ≤I(1), the decision is inconclusive with regard to accepting or rejecting the long-run relationship. In this case, Boutabba (2014) and Banerjee et al. (1998) argued that an efficient way to establish co-integration among the variables is to test the error correction term (ecM t-1 ) and confirm it with past studies.
Furthermore, Johansen and Juselius (1990) argued that the maximum likelihood method is more appropriate in a multivariate system to identify the number of cointegrated vectors in the model. However, the cointegration test, Johansen and Juselius (1990), is utilized to determine whether the variables in the vector are cointegrated in one or more long-run relationships using trace and maximum eigenvalue tests. The decision rule is to reject the H 0 of no co-integration when the estimated likelihood test statistic (trace and maximum eigenvalue tests) exceeds its critical value (Brooks, 2008;Johansen andJuselius, 1990, Juselius, 1991). In other words, the trace test aims to provide a single co-integration vector among the variables to test the H 0 of no co-integration (H 0 : r = 0) and the alternative hypothesis (H 1 : r=1). This test can be written as in equation (6).
However, the maximum eigenvalue test aims to determine the minimum co-integration (Brooks, 2008). Equation (7) shows that the H 0 of no co-integration is accepted if (H 0 : r = 0) and the alternative hypothesis is (H 1 : r≥2).
λ max (r, r+1)=−T Ln(1−λ r+1 ) Where r is the number of co-integrating vectors and T is the number of observations.
Since co-integration between the series exists, the long-run and short-run causality directions among the variables can be determined (Bekhet and Al-Smadi, 2014;Engle and Granger, 1987). Thus, the vector error correction model is applied as in equation (8). LGDP LGDP where δ i s denote the intercepts, β ij s present the short-run coefficients, and γ i s represent the coefficients of ecM t-1 and i, j=1,…, 7. In addition, ecM t-1 is the lagged value of the residuals derived from the F-bounds test among the variables. This is used to detect the long-run causality among the variables, while the joint χ 2 statistic for the first differenced lagged independent variables is used to test the direction of short-run causality between the variables (Bekhet and Harun, 2017;Bekhet and Al-Smadi, 2015;Hamdi et al., 2014;Boutabba, 2014). For exampel, ΔLEC t does not Granger cause ΔLCO 2t , if β 12 = … = β 21 , i =0.
In the final step, the cumulative sum of recursive residuals (CUSUM) and the cumulative sum of squares of recursive residuals is used to check the stability of the long-run parameters' movements for the ΔLCO 2t model. In addition, the impulse response functions test is used to trace the effect of a 1-time shock to one of current and future values of the endogenous variables. This provides the time path of the impact of a shock on the future values of all the variables in the multivariate dynamic system, and the IFR should provide better insights into short-term and long-term linkages among the variables employed in each of the models (Hamdi et al., 2014;Pesaran and Shin, 1999;Pesaran and Pesaran, 1997).

Data Quality, Stationarity and Co-integration Results
The first aim of any statistical procedure is to provide a succinct description of the data being analyzed. Also, in the time series data, the descriptive statistical tests, normality, and interrelationship matrix are very important to determine whether the data collected are adequate and sufficient for time series analysis (Bekhet and Al-Smadi, 2015). Table 1 shows that (LCO 2t , LEC t , LGDP t , LFD t , LUR t ) variables are positively and significantly interrelated with each other within the acceptance range of correlation coefficients. Furthermore, it reveals that these variables have no possibility of the negative effects of multicollinearity and that all variables are normally distributed, as confirmed by skewness and Jarque-Bera normality tests (Brooks, 2008). Table 2 shows the ADF, P-P, and KPSS unit root test results, which indicate that the above variables are stationary at I(1) with intercept and trend at different significance levels (1%, 5%, and 10%, respectively). Table 3 shows that the F-statistic values for the LCO 2t , LEC t , LGDP t , and LFD t models are co-integrated. This means that long-run relationships exist among the variables in these models.
In addition, the F-statistic values for the LUR t model show that the    (2009), (2) ***, **, *as defined in Table 2. Sources: (a) The Output of the Micro-fit 4.1 econometric software packages. (b) The output of the E-views 7.2 econometric software calculated F-statistic value falls in the range I(0) ≤ F-statistics value ≤ I(1), indicating that it is inconclusive with respect to accepting or rejecting the H 0 . Thus, the results of this model compared with those of past studies confirm the long-run relationship among the variables. Also, the researchers followed Boutabba (2014) to verify the co-integration among the variables by applying the stationarity test for the ecM t-1 . Furthermore, the results of the multivariate Johansene-Juselius test also confirmed that long-run relationships among the variables exist in all models except the LUR t model. In summary, the four models depict co-integration but we have focused on the first model (CO 2 model) due to the relevant objective of the study.
In summary, Table 3 shows that the variables in the CO 2 emissions, energy consumption, economic growth, and FD models have an equilibrium relationship. The table also shows that with the pace of development in Malaysia these variables move together. Also, it highlights that despite the rapid economic development challenges remain regarding how to manage energy consumption and control CO 2 emissions. Ozbugday and Erbas (2015) argued that the scale of economic activity measured by real income and industrialization has a significant positive effect on CO 2 emissions.
FD promotes investment, which raises energy demand due to economic growth. This means that it may influence energy consumption through the economic growth channel in Malaysia.
A well-established financial system boosts the effectiveness and efficiency of financial institutions through enhanced innovation in financial services delivery, reduction of information cost, efficient management of complex and risky transactions, and enhancement of transparency between borrowers and lenders, thereby guaranteeing the profitability of investment (Islam et al., 2013;Shahbaz, 2013). All these factors improve business investment and economic activities, thereby boosting domestic production and economic growth, while the urbanization variable predicts the need for careful planning regarding rapid urban sprawl in Malaysia.

Short and Long Run Results
The results of the co-integration relationship among the variables as in the first model (Table 3) are used to estimate the coefficients of the long-and short-run relationship between the above variables. Table 4 shows that energy consumption is positively associated with CO 2 emissions. This means that a 1% rise in energy consumption increases CO 2 emissions by 1.33%. This highlights the need for technical innovation and renewable energy production methods.
These findings are in line with Heidari et al. (2015). Moreover, the impact of income is positive on CO 2 emissions. Both linear and non-linear terms of GDP confirm the EKC hypotheses that the inverted-U relationship between economic growth and CO 2 emissions in Malaysia exists Yasmin, 2013a and2013b;Begum et al., 2015). The results indicate that a 1% rise in GDP will increase CO 2 emissions by 0.006% while the negative sign of the squared term seems to corroborate the delinking of CO 2 emissions and GDP at higher levels of income in the Malaysian economy. This shows the relevance of Vision 2020 in that the Malaysian economy is growing fast and people's cosumption patterns are changing. This evidence confirms that in Malaysia CO 2 emissions increase in the initial stage of economic growth and decline after a threshold is reached.
In addition, the results reveal that an increase in FD leads to an increase in the effect of CO 2 emissions in the Malaysian economy in the long run. The coefficient of LFD t suggests that a 1% contribution of FD leads to 0.014% increase in CO 2 emissions during the estimation period. These findings are similar to earlier results from Ziaei (2015), , and Boutabba (2014). From the urban sprawl perspective, the coefficient of LUR t shows an inverse impact on CO 2 emissions in Malaysia. However, according to UET theory, the coefficient of LUR t 2 is positive, which means that in the long run Malaysia will still face pollution challenges due to the urbanization process. This suggests upcoming urban sprawl with various implications for the Malaysian economy. Azam et al. (2015) suggested that the population growth rate has a positive and statistically significant relationship with energy consumption only in the case of Malaysia among the ASEAN countries.  also argued that from the Malaysian perspective, due to the rapid urbanization process with a huge number of industrial areas, "over-urbanization" exists. 1***, **, * as defined in Table 2 The short-run dynamics results showed empirical evidence that energy consumption leads to an increase in CO 2 emissions in Malaysia at one lagged period. Note that a 1% rise in energy consumption increases CO 2 emissions by 0.001% while negatively affecting it at two and three lagged periods. The sign of coefficients of GDP and GDP 2 confirmed the existence of the inverted-U Kuznet curve also in the short run in Malaysia. In addition, the results showed that the short-run elasticities for CO 2 emissions are less than the long-run elasticities. In the long run, the result shows that urbanization lowers CO 2 emissions by 0.58%. However, in the short run, emissions will increase due to Malaysia's urbanization process. The study shows that land use development and urban density are the real sprawling problems. Landownership structures and the inability or unwillingness of government to intervene in these structures are perhaps the main factors contributing to 'illegal' settlements and chaotic urban sprawl. This result is consistent with Hossain (2011), who studied the urbanization process in newly industralized countries. In contrast, for the long run, the signs of UR and UR 2 highlight the UET application that cities becoming wealthier by increasing their manufacturing base leads to industrial polltion problems that affect the land, water, and air Ren et al., 2015;Wang et al., 2015. Conversely, with modernization and more wealth, the environmental problems can be lessened through technical innovation, strict regulations, and changes in the composition of economic sectors (McGranahan et al., 2001).
In addition, the coefficient of ecM t-1 is also negative and statistically significant at the 1% confidence level. This indicates that the ΔLCO 2t model is corrected from the short-run equilibrium toward the long-run equilibrium at high speed by 131%. This shows that the long-run equilibrium between CO 2 emissions and other variables would correct back within 7 months (1 divided by the estimated coefficient of ecM t-1 ). This is also consistent with the Malaysian policy target to cut CO 2 emissions by 40%. Furthermore, the results of diagnostic tools suggest no possibility for the non-normal distribution of error terms, serial correlation, functional form, and heteroscedasticity phenomena ( Table 5).

Causality Results
The multivariate Granger causality test results are reported in Table 5, which shows evidence of the long-run bidirectional Granger causality running between the variables in the models at the 1% and 5% significance levels. Engle and Granger (1987) argued that if a co-integration relationship exists among the variables, then causal directions between these variables must exist.
Based on above discussion, Figure 7 summarizes that short-run Granger causality directions among the study variables exist.
Bidirectional causality running from CO 2 emissions to energy consumption, economic growth, and urbanization and from economic growth to FD is confirmed. Unidirectional causality running from energy consumption to urbanization, from economic growth to energy consumption, from FD to CO 2 emissions, and from urbanization to FD is also confirmed. These findings are consistent with the empirical evidence of  for MENA countries, Ziaei (2015) for 13 European and 12 East Asia countries, Boutabba (2014) for India, Shahbaz et al. (2013) for Turkey, Saboori and Sulaiman (2013) for ASEAN countries, and Chang (2015) for 53 countries.
In addition, Figure 8 presents the results for stability of the long-run relationship for the CO 2 emissions model. It indicates that the CO 2 emissions model is stable in the long-run parameters because the above test results lie within critical bounds at the 5% significance level. This implies that the estimated parameters are valid for the 1971-2013 period. In other words, the ecM t-1 value is relatively stable.
(2) denote the bidirectional results variables for the LCO 2 , LEC, LGDP, and LFD models at the 5% significance level. The findings unveiled that energy consumption, economic growth, and FD have a positive effect on CO 2 emissions whereas urbanization has a negative effect. In the short run, the variables display different levels of significance. However, energy consumption, GDP, and urbanization have a positive relationship with emissions; only the finnacial component shows a negative relation. The causality analysis exposed the bidirectional causality from CO 2 emissions to energy consumption, GDP, and urbanization. This means that two-way relationships exist among these variables and any change in one affects the others. There is unidirectional causality running from energy consumption to urbanization, GDP to energy consumption, FD to CO 2 emissions, and urbanization to FD.
Based on the overall results of the current study in Malaysia, challenges have increased due to rapid economic growth resulting in high urbanization, energy consumption, and CO 2 emissions. In this context, it is important for Malaysia to promote the 'government take the lead' measure to drive energy efficiency.
The government should also engage in continuous cooperation with non-governmental bodies to promote energy efficiency and energy conservation measures in the residential sector. The increased energy demand in the country is expected to lead to both energy efficieny and new energy sources (Omri and Nguyen, 2015). Rafindadi and Ozturk (2015) highlighted that natural gas consumption and the economic growth nexus will enable the speedy actualization of the Kyoto target of reducing CO 2 emissions and ensure environmentally friendly working conditions. Various energy agencies, such as KeTHA, the Sustainable Energy Development Authority (SEDA), and the Suruhanjaya Tenaga (ST) Energy Commission, have undertaken joint efforts to establish a comprehensive national energy database to support the dissemination and distribution of energy statistics in Malaysia to local and international stakeholders and the public. This can help to enhance research work which will support effective long-term energy and environmental policy decisions.
In this context, the Eleventh Malaysian Plan (2015-2020) emphasizes energy security and economic efficiency as well as environmental and social considerations. The policy focuses on specifically four areas: green growth, sustainable methods of production to cut emissions, protection of natural resources, and strong responses to climate change challenges. The plan has further accelerated the urban and rural transformation process with a healthy environment. This means that the concept of cities as islands of privilege and opportunity is supported by indicators on health, education, and income, which generally reflect better outcomes in urban areas than rural areas. However, with increased globalization, the urban transition has widened the gaps between social groups and is making inequality more visible. Large cities generate creativity and solidarity, but also make conflicts more acute. Moreover, the eradication of poverty is also vital because measurement of poverty in both rural and urban areas is generally based on income, which does not necessarily provide an accurate picture of the scale and multidimensional nature of poverty. One view is that urban poverty is a transient phenomenon of ruralto-urban migration and will disappear as cities develop, thus absorbing the poor into the mainstream of urban society. This view is reflected in Malaysian national poverty reduction strategies, which tend to be rural-focused, with the result that interventions have had limited effect in reducing relative poverty, exclusion, and inequality in cities (Economic Planning Unit, 2015).
Furthermore, from the environmental perspective on FD, the findings show a strong co-integration. This justifies that any planned financial system must encourage economic growth through effective investment policies. Also, a developed financial sector promotes economic growth through investment channels. It also plays a vital role in cutting CO 2 emissions by applying eco-friendly techniques during production processes. Frankel and Romer (1999) argued that FD attracts FDI from the developed world to emerging economies. Moreover, the financial sector motivates local firms to adopt environmentally friendly techniques for production processes. This implies that a sound financial sector can improve environmental quality through the new-technologyusing effect but this effect is nullified during periods of financial instability. Regarding the financial element in an economy, Shahbaz (2013) argued that to avoid financial instability and its impact on environmental degradation, financial sector reforms should be implemented step by step with great care.
In Malaysia, the financial sector promotes many investment schemes in green technology. In 2010, Green Technology Malaysia (2016) established the Green Technology Financing Scheme (GTFS) to accelerate expansion of the green technology industry in the country by providing easier access to financing from financial institutions. RM3.5 billion in loans was allocated for use through 2015. The Scheme, which targets both producers and users of green technology, offers a rebate of 2% per annum on the interest rate or Source: The Output of the Micro-fit 4.1 econometric software Figure 8: Cumulative sum of recursive residuals and cumulative sum of squares of recursive residuals tests for the (1971-2013) period profit rate charged by the financial institutions and a government guarantee of 60% on the financed amount. Moreover, Bank Negara Malaysia has worked closely with the Ministry of Finance and the Ministry of Energy, Green Technology, and Water to establish the GTFS. Efforts are also underway to strengthen the capacity of financial institutions to provide advisory support to businesses. To achieve comprehensive financial allocation of green technology funds, the various phases consist of research and development, manufacturing processes, and end-use technology.
Last, from the urbanization perspective, our analysis elaborated that with urban expansion there is an element of modernazation in Malayisa, but in the long run this needs careful planning. UNHABITAT (2016a and 2016b) highlighted the national urbanization policy which aimed to ensure that urban centers play an optimal role as engines of the nation's economic growth by providing a high quality of life through a systematic and planned urbanization process. Overall, city growth provides opportunities for modernization and cultural enrichment and accelerates social change. Rapidly growing cities, especially larger ones, include various generations of migrants, each with a diversity of social and cultural backgrounds. A diverse global network of vibrant cities needs continual adjustment and adaptation to the mix of traditional values and contemporary perspectives of communities and governments. The creation of cities with a vision which supports harmonious communities and living conditions through sustainable urban development is beneficial. Both urbanization and globalization have brought drastic changes to the image of major cities, as justified in UET theory, which suggests that urbanization policy should consider land use which reduces emissions and will lead to achieving future green technology targets in Malaysia. This can result in sustainable green energy growth. Various environmental challenges still exist in Malaysia that are related to urban sprawl; these can be handled by effective policy initiatives. The World  highlighted that Malaysia needs to establish lead transport agencies at the metropolitan level to spearhead integrated approaches to the planning and delivery of both public and private urban transport, identify and implement sustainable financing mechanisms, and align public transport with incentives to discourage the use of private transport in congested areas, such as London, Singapore, and other major world cities have done successfully. This includes projects such as mass rapid transit construction and establishment of the Land Public Transport Commission (SPAD), which has developed into a capable, multi-disciplinary planning and regulatory agency. Institutional challenges remain, but further policy steps can ensure that urban transport planning is integrated across modes and administrative boundaries at the metropolitan level.
In Malaysia, most urban centres are still weak at portraying a distinct identity and exceptional character. The influx of external influence has created more cities with a global character, such as Kuala Lumpur, Georgetown, and Johor Bahru. With the rapid pace of urbanization expected by 2020, urban governance will face various complex challenges ahead. These challenges require that the respective parties be more focused in undertaking each and every responsibility in urban development. However, the involvement of multiple agencies and departments in urban management has made it difficult to coordinate many actions and in turn compromises the effectiveness of those actions. To strengthen Malaysia's role in the new urban agenda, the best opportunity is to collaborate with United Nations -Habitat. To this end, Malaysia will host the World Urban Forum (WUF9) in 2018. The government of Malaysia hopes that by hosting WUF9, a major and concrete legacy will be created for sustainable urban communities (UNHABITAT, 2016b).
The present study opens up new dimensions for future research; for example, the model used for this paper can be improved by including other variables (e.g., governance, institutional quality, political scenarios, credit management) to examine their impact on CO 2 emissions. This idea can be enhanced to investigate the effect of the US financial crisis on environmental degradation in developed and developing economies (Bekhet and Yasmin, 2014).
The availability of data on domestic credit to the private sector at the provincial level will make it possible to conduct a study to examine the effect of financial instability on the environment at the provincial level. This might provide new ways for policy makers to develop sound financial markets to contribute to saving the environment.