The Impact of Economic Growth, Foreign Direct Investment, Urbanization and Trade Openness on CO2 Emissions in Sri Lanka

Climate change in Sri Lanka is a big issue in recent years and CO2 emission is continuously rising in Sri Lanka. The main objective of the study is to examine the impact of economic growth, foreign direct investment, urbanization and trade openness on CO2 emissions in Sri Lanka during the period 1978-2019. To accomplish the objective, applied Autoregressive Distributed Lag (ARDL) model to determine the long-run and short-run effects. CO2 emission is used as the dependent variable. Foreign direct investment, per capita GDP, trade openness and urbanization variables, which act as the independent variables in the model. The results indicate that GDP per capita, urbanization and trade openness could lead to increased environmental emissions in the long-run. However, GDP per capita could lead to decreased environmental emissions in the short-run. In order to prevent the increase of CO2 emissions caused by economic growth, urbanization and trade openness, the level of CO2 emissions should be considered when making policies to improve economic growth, and urbanization. Moreover, the government should design the trade reforms and policies to be accompanied by strong environmental policies in the long-run.


INTRODUCTION
CO 2 emission is one of the main factors of human induced climate change and increase global warming. The level of CO 2 in the earth has been on the increase since the industrial revolution (Ayoade, 2003). We should attempt to cut down on our personal CO 2 emission because climate change in Sri Lanka is a big issue in recent years. According to the Global Climate Risk Index 2019, Sri Lanka has been ranked the second most affected country by extreme weather events in the past 20 years (Eckstein et al., 2019, p. 6).
According to the definition of the World Development Indicator (2020), "Carbon dioxide emissions are those stemming from the burning of fossil fuels and the manufacture of cement and it includes carbon dioxide produced during consumption of solid, liquid, and gas fuels and gas flaring". Figure 1 represents the per capita CO 2 emissions in Sri Lanka has been showing an increasing trend since 1990s. The value for CO 2 emissions (metric tons per capita) in Sri Lanka was 1.31 as of 2019. According to the World Development Indicator Database (2020), over the past 41 years, this indicator reached a maximum value of 1.31 in 2019 and a minimum value of 0.2036 in 1989. Figure 2 represents the annual share of global CO 2 emission, which is measured as each country's emissions divided by the sum of all countries' emissions in a given year plus international aviation and shipping (known as 'bunkers ') and 'statistical differences 'in carbon accounts. Sri Lanka produces less than 0.1 percent of the world's carbon dioxide emissions. However, the level of CO 2 emission in the country is continuously rising in Sri Lanka in recent years.  The majority of studies have analyzed economic growth and energy consumption as the two most important variables related to environmental degradation (Osobajo et al., 2020;Mizra and Kanwal, 2017;Kasman andDunman, 2015 Dritsaki andDritsaki, 2014 etc.). However, economic growth and energy consumption alone may not enough to explain CO 2 emission (Ozturk et al., 2010). Therefore, it is very important to consider other macroeconomic variables such as foreign direct investment, urbanization and trade openness on CO 2 emission.
"Foreign direct investment (FDI) inflows play a strong impetus for economic development and growth and it is considered as one of the major driver of globalization" (Rathnayaka Mudiyanselage et al., 2021). The foreign direct investment is very important for both developed and developing country and believe that FDI can positively affect their economies. According to the World Development Indicator in Sri Lanka (2020), the net inflows of FDI in Sri Lanka increased from 1.47 million US dollars in 1978 to 758 million on US dollars in 2019. Data on FDI inflows have increased steadily from 2015 to 2018. However, it decreased in 2019 in line with the changing global economic situation and the political environment. Theoretically, the effect of FDI on the CO 2 could have two possible effects. The effect could be negative, in the sense that increased FDI inflows could lead to increased environmental emissions and increased FDI inflows could lead to decreased environmental emissions (Demena and Afesorgbor, 2020;Seker et al., 2015;Kim, 2020).
("Kuznets curve", 2021), there exists a U-shaped relationship between the level of CO2 emissions and economic growth, according to the Environmental Kuznets Curve (EKC) hypothesis, which was introduced by Kuznets (1950s and 1960s). The EKC hypothesis indicates that growing per capita GDP leads to increasing CO2 emissions and when reaching a high enough per capita GDP at the certain turning point, CO2 emission starts to decrease. The effect of economic growth on the CO2 could have two possible effects. The effect could be negative, in the sense that increased GDP per capita could lead to increased environmental emissions and increased GDP per capita could lead to decreased environmental emissions. The higher level of income increases demands for more energy or promotes industries, which effect environment pollution. On the other hand, environmental quality may improve if higher income leads to adoption of environmentally friendly production techniques and consumption of clean and green energy (Aye and Edoja, 2017).
Urbanization is a one of major change taking place in worldwide. There is a significant difference in urban population change between the more developed counties and the less developed countries. Many studies have indicated that urbanization as the one of important variable related to environmental degradation (Gasimli et al., 2019: Sharma, 2011Cole andNeumayer, 2004 andShahbaz et al., 2014). According to the theories of ecological modernization, urbanization can have positive as well as negative impact on the environment (Sadorsky, 2014). An increase in urban population causes a high level of CO 2 emission. On the other hand, it does not increase CO 2 emission. The impact of urbanization on the environment differs with the level of development (Poumanyvong and Kaneko, 2010). Considering the trend of urban population growth in Sri Lanka, the data reveal that Sri Lanka has a slow tempo of urbanization. There was an increase in urban population growth, reaching the 18.585 percent in 2019 (World Development Indicator, 2020).
The relationship of trade openness and CO 2 emission has been examined in many studies. Several studies have found the mixed results. Some researchers argue that trade openness is harmful for environmental quality; while some researchers found trade openness is good for environmental quality. Some others have found even no relationship between trade openness and CO 2 emission. Trade openness towards greenhouse gas emission is considered as an important issue within the context of human induced climate change (Naranpanawa, 2011 There are some studies that are conducted in Sri Lanka to examine the impact of trade, and urbanization on environmental degradation (Gasimli et al., 2019), the impact of liberalization on CO 2 emission (Kariyawasam and Sumanaratne, 2020), the relationship between trade openness and carbon emissions (Narampanawa, 2011). There is not study that examines the impact of economic growth, foreign direct investment, urbanization and trade openness on CO 2 emission together in Sri Lanka uses recent data and sophisticated econometric technique. Thus, this study tries to bridge this gap. Considering the trend of CO 2 emission, foreign direct investment, economic growth, urban population growth, and trade openness in Sri Lanka, it is very important to examine the impact of these factors on CO2 emission in Sri Lanka. In the study, seek to answer the following question; namely, are there any impact on economic growth, foreign direct investment, urbanization and trade openness on CO 2 emission in Sri Lanka during the period 1978-2019?
The objective of this study is to examine the impact of economic growth, foreign direct investment, urbanization and trade openness on CO 2 emission in Sri Lanka. The main hypothesizes are developed as follows; H A = Economic Growth has an impact on Co 2 emission H B = Foreign direct investment has an impact on CO 2 emission H C = Urbanization has an impact on CO 2 emission H D = Trade openness has an impact on CO 2 emission The paper is organized as follows: In the first part, introduction, research questions and objectives of the study. In the second part, a review of the literature field gives empirical evidence of earlier studies. Next, the research methodology is used data and sources, model specification and estimation techniques are covered. Further describes the results of the study and some concluding remarks are made in final section.

Country Specific Analysis
Gasimli et al., (2019) examined the nexus between energy, trade, urbanization and environmental degradation in Sri Lanka from 1978 to 2014 using ARDL bound testing method. In this study, carbon emission (metric tons) per capita is used as a proxy for environmental degradation, energy consumption is proxied by energy consumption per capita (kg of oil equivalent per capita), income level is proxied by real GDP per capita and trade openness is measured as the ratio of the exports plus imports to GDP. The urban population is used as a proxy for urbanization. The longrun results indicated that there is a U-shaped relationship existing between CO 2 emission and income in the long-run. Trade openness, energy consumption positively contributed to carbon emissions. Urbanization in Sri Lanka is very interesting, as urbanization was significant with a negative sign implying that urbanization does not aggravate environmental degradation throughout the years.

Narampanawa
(2011) has examined the relationship between trade openness and carbon emissions in the case of Sri Lanka using Autoregressive Distributed Lag (ARDL) bounds testing approach and the Johansen-Juselius maximum likelihood approach over the period 1960 to 2006. The results suggested neither a long run equilibrium relationship, nor a long term causality exists between trade openness and carbon emissions in Sri Lanka. However, the results indicated that there is a short run relationship between trade openness and carbon emissions.
Kariyawasam and Sumanaratne (2020) have examined the impact of liberalization on CO 2 emission in Sri Lanka from 1977-2011 using ARDL model. The variables, which are Gross Domestic Product (GDP), Trade Openness (total trade as a % of GDP), Per capita Carbon Dioxide (CO 2 ) emission and per capita energy consumption used in the model. The long-run results indicated that all variables have positive and statistically significant impact on CO 2 emissions.
Kazilkaya (2017) has examined the relationship between carbon dioxide emissions, economic growth, foreign direct investment and energy consumption in Turkey over the period of 1970-2014 using ARDL bound testing method. Turkey is considered as a developed country. The results indicated that economic growth and energy consumption have positive and statistically significant impacts on CO 2 emissions in the long-run and short-run. However, there is not any significant relationship between foreign direct investment and CO 2 emissions in the long-run and short-run. He suggested some policy recommendations that produce and implement more effective and more efficient policies in the context of sustainable development in Turkey.
To examine the long and short-run linkages between economic growth, energy consumption and CO 2 emission in Tunisian (Chebbi and Boujelbene, 2008) conducted a study using VECM technique over the period 1971-2004. The results indicated that there is a positive linkage between output and energy use and CO 2 emission and energy consumption are positively related in the long-run. The results of the short run, economic growth positively affects the growth of energy consumption.
Kim (2020) has examined the effects of foreign direct investment, economic growth, industrial structure, renewable and nuclear energy, and urbanization on Korean greenhouse gas emission from 1981 to 2014 using ARDL bound test. The dependent variable is CO 2 emissions per capita, which is a proxy for greenhouse gas emissions. Independent variables are the net inflows of FDI, the share of renewable and nuclear energy in primary energy supply as a proxy for renewable and nuclear energy, the urban population in the total population as a proxy for urbanization and the share of manufacturing in GDP as a proxy for industrial structure in the model. The results indicated that economic growth, FDI and urbanization increase of greenhouse gas emissions, while manufacturing industry share, renewable energy and nuclear energy contributed to the reduction of GHG emissions in the long-run. In the short run results indicated that economic growth increases in GHG emissions, while renewable and nuclear energy have contributed to the reduction in GHG emissions. FDI and urbanization did not play a role in increasing HG emissions in the short term. Seker et al., (2015) have examined the impact of foreign direct investment, gross domestic product. The square of GDP and energy consumption on carbon dioxide emissions in Turkey over the period 1974-2010 using the autoregressive distributed lag model. The long-run results indicated that the effect of FDI on CO 2 emissions is positive, while the effects of the GDP and energy consumption on CO 2 emissions are positive and statistically significant. In the short-run, the results are found to be similar to those of the long-run model. Granger causality test causality test results indicated the existence of a causality running from all explanatory variables to CO 2 emissions in the long run. They suggested that Turkey should promote energy efficiency with sustainable growth, and encourage more FDI inflows particularly in technology-intensive and environment-friendly industries to improve environmental quality.

Cross-country Analysis
A large number of studies have been conducted on the relationship among foreign direct investment, economic growth, urbanization and trade openness on carbon emission in developed and developing countries as crosscountry analysis. Baek et al. Furthermore, it reveals that urbanization has a negative, but an insignificant impact on carbon emissions in the low income, middle income and high-income panels. GDP per capita has a positive impact on CO 2 emission in the low income and middleincome panels.

The Co-integration Analysis (ARDL)
To investigate the impact of economic growth, foreign direct investment, urbanization and trade openness on CO 2 emission in Sri Lanka, we adopted the Autoregressive Distributed Lag (ARDL) bounds testing approach developed by Pesaran et al., (2001). The data are of mixed type of I (0) and I (1). This requirement fulfills employing the ARDL model to identify the long-run and the short-run relationship between variables.

Data and Sources
The data used in the empirical analysis was mainly secondary data available in the World Development Indicators. Data on the GDP per capita, foreign direct investment inflows, urbanization and trade openness on CO 2 emission from the year 1978 to 2019 were collected for the study purpose. The study is used only 41 period of data. The justification for the important variables used in this study is based on reviewing the existing theoretical and empirical studies.

Model Specification
This study adopted the Anwar et al., (2020) theoretical framework to examine the impact of economic growth, urbanization and trade openness on CO 2 emission. Furthermore, to examine the foreign direct investment on CO 2 emission modified the model adding foreign direct investment inflow variable in the model. The model could be specified as follows: Where, (CO 2 ) CO 2 emissions, (FDI) Foreign direct investment inflows, (GDPP) GDP per capita, (URBAN) is a proxy for urbanization and Trade openness (TR). We converted the variables of per capita CO 2 emissions, foreign direct investment inflows and GDP per capita into their natural logarithmic forms, as the natural logarithms of all variables smooth out the entire data used for analysis.
In an econometric form, equation (1)  ) represent the short-run dynamic of the model. In the next step of the estimation procedure, we obtain the short run dynamics of parameters and long run adjustment of the model by estimating the error correction version of ARDL model pertaining to the variables in the equation (4) is as follows: where, is speed of adjustment coefficient, t is pure random error term.

Estimation Techniques
In order to make the model and variables free from problems associated with time series data employed Augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) unit root test approaches to test stationary of the variables. The second step of the estimation bound testing procedure is employed in order to investigate the existence of a long-run relationship the bound tests approach developed by Pesaran et al. (2001). If the calculated Fstatistics is greater than the appropriate upper bound critical values, the null hypothesis is rejected implying co-integration. After establishing the evidence of the existence of the co-integration between variables, the lag orders of the variables are chosen by using the appropriate Schwarz Info Criteria. In the next step of the estimation procedure, we obtain the short run dynamics of parameters and long run adjustment of the model by estimating the error correction version of ARDL model pertaining to the variables in the equation (4). Moreover, Diagnostic tests were conducted to check whether the results are robust.

Descriptive Statistic
Descriptive statistics shows the basic characteristics of the data set. The descriptive statistics LGDPP and URBAN variables are positively skewed. Source: Author's computation using E-views 11

Unit Root Test
The Augmented Dickey-Fuller (ADF) test and Phillips Perron (PP) test were used for testing the stationarity of the data (See below Table 3). The results indicated that the null hypothesis of series contain a unit root cannot be rejected at levels for all variables except LFDI in ADF and PP unit root approaches. The results of the Unit Root Test confirmed that LFDI variable is stationary at the first differences, This reveals that LFDI is integrated in order zero I(0) while all other series, namely LCO 2, LGDPP, URBAN and TR variables are integrated in order one I(1). It means the data are of mixed type of I (0) and I (1). This requirement fulfills employing the ARDL model to identify the longrun and the short-run relationship between variables. -1.9 -3.5 -7.0 -3.5 I LFDI -7.5 -3.5 -6.3 -3.5 0 LGDPP -1.7 -3.5 -4.5 -3.5 I URBAN -1.8 -3.5 -5.9 -3.5 I TR -1.5 -3.5 -5.7 -3.5 I Note: t-Stat values are given in the table (Trend & Intercept) Source: Author's computation using E-views 11

Bound Test Approach to Cointegration Test
The results of the ARDL bound test are presented in table 3 below. The F-statistics in the entire model (6.2882) is higher than upper critical values at 1%, 2.5%, 5% and 10% level of significance. Therefore, the null hypothesis of absence of co-integration is rejected, which implies that there is a long-run relationship among the dependent variables and the explanatory variables in the models.  Note: Probability values are given in the Table. *, **, and *** imply the rejection of the null hypothesis at 1%, 5% and 10%, level of significance respectively.

Long-run Coefficient Estimates
The regression results indicate, R squared value is 0.9865. This means that 98.65 percent of total variations in CO 2 emission Sri Lanka is explained by changes in economic growth, foreign direct investment, urbanization and trade openness. The F-statistic of with a p value is 252.3103 of 0.000 at 1 percent significance level reveals that all the independent variables were jointly significant in predicting CO 2 emission in Sri Lanka.  & Afesorgbor, 2020;Seker et al., 2015 andKim, 2020). Some researchers found that insignificant relationship between foreign direct investment and CO 2 emissions (Kazilkaya, 2017).
LGDPP with a coefficient of 1.1508 has a positive and statistically significant impact on CO 2 emissions in the long-run. This result is in line with these empirical studies (Kazilkaya, 2017;Kim, 2020;Sekeret al., 2015;Chebbi & Boujelbene, 2008;Fan et al. 2006, Kariyawasam and Sumanaratne, 2020and Sharma 2011. URBAN with a coefficient of 0.2051 has a positive and statistically significant impact on CO 2 emissions in the long-run. This result is in line with these empirical studies (Kim, 2020;Martínez-Zarzoso and Maruotti, 2011). TR has positive and statistically significant impacts on CO 2 emissions in the long-run. In the sense that increased trade openness could lead to increased environmental emissions in the long-run (Kariyawasam and Sumanaratne, 2020).  Table. *, **, and *** imply the rejection of the null hypothesis at 1%, 5% and 10%, level of significance respectively. Table 6 reveals that short-run coefficients.

Short-run Coefficient Estimates and Error Correction Representation
LGDPP variable is statistically significant at the 10% significance level.
LGDPP variable has a negative impact on CO 2 emissions in short-run. It indicated that increased LGDPP could lead to decreased CO 2 emissions in the short-run.
The error correct term (ECT(-1)) coefficient is significant at 1% significance level with the negative sign which is between 0 and −1, implying that the model can converge back to long-term equilibrium quickly after a shortterm shock. The value of -0.2725 indicates that the disequilibria from this period's shock can be adjusted in the next period about 27.25%.

Diagnostic Tests
The model is free from serial correlation and heteroscedasticity. Moreover, the functional form is correct and stochastic residuals are normally distributed. The estimated model satisfies all indispensable diagnostic tests.

The Results of Stability Test
The graphs of the CUSUM and CUSUM of squares test confirms that the model is stable since the residual plot lies between the lower and upper critical bounds at the 5% level of significance. That is, the selected model has stable parameters, which can be used for long-term forecasts. Source: Author's Computation Using E-views 11

CONCLUSION
This study analyzed the impact of foreign direct investment, per capita GDP, urban population growth is a proxy for urbanization and trade openness on CO2 emission in Sri Lanka over the period 1978-2019. The econometric methodology used in this paper is the Autoregressive distributed lag model. According to ARDL bounds test, the long-run equilibrium among the variables is confirmed. Then, the long-run and short-run coefficients were estimated using ARDL error correction model.
In the long run, the foreign direct investment can not significantly explain the variation in carbon dioxide emission throughout the years. However, the results indicated that increased economic growth, urbanization and trade openness could lead to increased environmental emissions in the long-run. However, GDP per capita could lead to decreased environmental emissions in the short-run. In order to prevent the increase of CO 2 emissions caused by economic growth, urbanization and trade openness, the level of CO 2 emissions should be considered when making policies to improve economic growth, and urbanization. Sri Lanka should design more green and sustainable urbanization and promote renewable energy consumption in urban area. Moreover, the results suggest the importance of trade reforms and policies to be accompanied by strong environmental policies in the long-run.