Carbon dioxide emissions and the prospects for Lao PDR’s economic self-development

Abstract

This article explores carbon dioxide (CO₂) emissions and economic development in Lao PDR, in the context of the country’s urgent need and prospects for economic self-development. By employing an auto-regressive distributed lag (ARDL) bounds model, the empirical results show that gross domestic product (GDP) has a statistically significant and positive effect on CO₂ emissions. The article also affirms the existence of co-integration among pertinent variables and substantiates the positive long-term shock between CO₂ and GDP. Within reality, these results confirm the presence of causality running from foreign direct investment (FDI) and GDP to CO₂. Despite its limitations, this study is a pioneering exploration of empirical evidence regarding Lao PDR’s yearning and prospects for economic self-development, toward which gravitate the article’s future research directions and policymaking recommendations.

Keywords

FDI ARDL Lao PDR

Sorphasith Xaisongkham, is a master’s degree Candidate at Jiangxi University of Finance and Economics. He graduated a bachelor of commerce from National University of Laos. He used to be an assistant of many researchers. He is of the enormous endowments in the field of international trade and economics and his zealous interest of environmental economics. No. 169 Shuanggang East Road, Xialuo, Changbei, Nanchang District, Jiangxi Province, P.R. China. Tel.: +86 131 5580 2676; Email: Sorphasithxsk@yahoo.com

Syvanh Phonasa, is a PhD Candidate at Dongbei University of Finance and Economics (DUFE), Dalian P.R. China. Before beginning his doctoral research, he worked as a lecturer of international economics at National University of Laos (NUOL). His area of interest be composing international trade and finance, public economics and development economics. No.217, Jiianshan, Street Shahekou District, Dalian, China, Tel.: +86 187 4207 7483; Email: phonasa@nuol.eud.la

Khaysy Srithilat, is a PhD Candidate at Dongbei University of Finance and Economics (DUFE), Dalian China. Before commencing his doctoral research, he worked as a lecturer of economics at National University of Laos (NUOL). He is the author of several publications and international conference proceeding. His area of interest includes monetary economics, International Economics, and Finance. No.217, Jiianshan, Street Shahekou District, Dalian, China, Tel.: +86 183 4226 6339; Email: k.srithilat@nuol.edu.la

Honglin Yuan, is Professor and Dean of Graduate School of Jiangxi University of Finance and Economics, Doctoral Advisor in School of International Trade and Economics. He is the executive director of China Society of World Economics, director of China Society of Commercial Economy, and executive director of China Society of Trade Economics Teaching Research. Prof. Yuan has published over 20 papers in domestic and international periodical journals, such as Guangming Ribao, International trade, Economic Management Journal, Contemporary Finance & Economics. He does the internationalization of SME research, serves as director, consultant, trainer for several companies. No. 169 Shuanggang East Road, Xialuo, Changbei, Nanchang District, Jiangxi Province, P.R. China. Tel.: +86 791 8381 6677; Email: yh16910@163.com

1 Introduction

Currently, climate change, carbon dioxide emissions (CO₂) and greenhouse gas phenomenon are pondered as the crucial trouble and outstanding matter that many nations are immensely concerned and accentuated on such a problem. It is of the critical matter, wherefore various countries globally have paid enormous attention and endeavoured to tackle on it that is, the government in numerous nations have been seeking for the appropriate policy and negative impacts as minuscule as possible of CO₂ emissions on FDI and economic development. Therefore, CO₂ emissions has been cogitated to be the global issue in the status quo and upcoming future that many countries have attempted to pursue the most potential tacks and proper measures to address on it, in order to establish the fitting balance between economic development and CO₂ emissions consistent with and in pursuant to the green-sustainable development intentions of the government internationally.

Similarly, many scholars and researchers have paid special attention in researching with respect to the stubborn problem and controversial issue of climate change as well as CO₂ emissions and greenhouse gas phenomenon. In which, many of their seminal articles were extensively published as the academic masterpieces in the miscellaneous journals, in term of the impact of economic development on CO₂ emissions and other implicated researches. Notably, recently [1] proposed a fresh perspective for creating energy policies that will boost economic development in south Africa, considering energy policy in the long run [2], through the orientation to saving energy could have negative impact on economic development, Naser [3] proposed that government and policymakers shed the light on energy efficiency strategies and carbon dioxide emissions reduction policy in the long run without impeding economic development in order to move towards environmental sustainability, Appiah [4] suggested that there is the need for establishment of regulations, proper institutional structures to set up policy framework, enforcement agencies to monitor to ensure compliance and creation of public awareness of the dangers posed by environmental degradation, Balibey [5] mentioned policy makers may take account exogenous impacts such as foreign investments to plan energy policies, and to maintain economic development against global climate warming, and the developing countries should formulate policies on the environment in order to accomplish economic sustainability [6].

Furthermore, Tang [7] also indicated that FDI and income are the major determinants of CO₂ and the adoption of clean technologies by foreign investment is pivotal aspect in curtailing CO₂ emissions in the country and sustaining economic development at the same time, Yanchun [8] China must be strict with the inflow of foreign capital. Moreover, Ahmad [9] revealed that higher energy consumption and lower GDP may spawn to environmental matters such as air and water pollution and prevention action should be taken to minimize the environmental degradation and Phouphet [10] mentioned that climate change has serious impact of Lao economy in term of declining GDP in Laos.

Consequently, the major aspiration of this empirical research is to explore carbon dioxide emissions and the prospects for Lao PDR’s economic self-development by employing ARDL bounds approach, where it is of the indispensable intention to be found the empirical evidence in Lao PDR’s aspiration that can accommodate to be consistent with the de fact and imminent application. Nevertheless, the limitation of this study is merely pondered two major determinants affecting to carbon dioxide emissions, whereas there are also other pertinent indicators. Similarly, authors merely accentuate on the analysis of ARDL bounds approach to cointegration, but not considered the environmental kuznets curve (EKC) model into the additional analysis. equally important, we also have limitation regarding to data analysis. that is, this research has adopted small time series component due to lacking of data and existing information (inadequate data). wherefore authors utilized in the period of 1988–2016 and merely applied the quantitative analysis. Furthermore, this article has investigated only the single co-integration relation (ARDL single long run relation), while authors has not taken into account the multiple co-integration relation (ARDL multiple long run relation) and this research is a pioneering exploration of empirical evidence regarding Lao PDR’s yearning and prospects for economic self-development, toward which gravitate the article’s future research directions and policymaking recommendations etc.

2 Literature review

In spite of there were many researchers and scholars had investigated the relation between CO₂ emissions and economic development internationally, but such those previous studies discovered the disparate aspects in the empirical literatures markedly: Appiah [4] investigation of the relation between economic development and carbon dioxide emissions as economic structure changes: evidence from Ghana by employing ordinary least square found that CO₂ emissions level in Ghana is influenced by GDP. Subsequently, Miti [11] had studied a cointegration analysis of real gross domestic product and carbon dioxide missions in transitional countries, by using dynamics and fully modified ordinary least square indicated that an increase in GDP leads to accumulate carbon dioxide emissions in positive effect and the results clearly also suggest the existence of a statistically significant long run cointegrating relation between CO₂ emissions and real GDP, Lægreid [12] introduced theoretical insights from environmental political science research, which suggests that CO₂ emissions models would gain explanatory leverage if moderators gauging political institutions were considered and estimating the potentially moderating effects of democracy, corruption, veto points and players, and civil society activity, the results revealed a positive and linear per capita GDP-CO₂ relation, which is barely affected by any variations in political and institutional aspects. The only significant moderator in this analysis is bicameralism in democratic, low corrupt countries, which generates a stronger effect at low levels of GDP per capita, Khobai [1] had studied the relation between energy consumption, economic development and carbon dioxide emissions by adopting VECM and the outcomes manifested the presence of the long run relation between energy consumption, CO₂ emission and economic development, bidirectional causality flowing between energy consumption and economic development in the long run and unidirectional causality running from energy consumption, CO₂ emissions to economic development.

Similarly, Ahmad [9] had examined the impact of economic activities on CO₂ emissions by applying OLS, the results exhibited that FDI has no significant relation with CO₂ emissions while GDP, energy consumption and trade openness have significant relation and a joint effect of all variables towards CO₂ emissions is also found to be significant, Wolde [13] studied energy consumption, carbon dioxide emissions and economic development, which the outcome attested the existence of one co-integration among variables and the result of causality test was found energy consumption causes economic development in Ethiopia, Cederborg [14] is there a relation between economic development and carbon dioxide emissions? where the findings revealed the presence of relation between economic development and environmental degradation and growing per capita GDP leads to increasing carbon dioxide emissions, Mir [15] CO₂ emissions and economic development: production-based versus consumption-based evidence on decoupling finds evidence suggesting a decoupling of production-based CO₂ emissions and development, consumption based CO₂ emissions are monotonically increasing with per capita GDP, Uddin [16] causal relation between education, carbon dioxide emissions and economic development in Bangladesh, VECM technique was applied in this analysis and the outcomes demonstrated there exists cointegration among the variables and strong positive relation among environmental pollution, education expenditure and economic development and, Lin [17] had also investigated energy consumption trends and decoupling effects between carbon dioxide emissions and GDP in south Africa etc.

Furthermore, there were additional empirical literatures notably, Bakirtas [18] studied economic development and carbon emission: a dynamic panel data analysis which found that a country reduces CO₂ emissions leads to accrue in income, Hasson [19] had examined energy consumption, trade openness, economic development, carbon dioxide emissions and electricity consumption: evidence from south Africa discovered the presence of a positive associationship between energy consumption and economic development and the results display trade reduces overall pollutions caused by carbon dioxide emissions. Balibey [5] investigated relations among CO₂ emissions, economic development and foreign direct investment and the environmental kuznets curve hypothesis in Turkey, wherefore this investigation confirmed the causality relation display that FDI and economic development have a significant effect on CO₂ emissions, impulse-response functions and variance decompositions of VAR model support these relation among GDP, CO₂ and FDI, Borhan [20] displayed pollution indicator carbon dioxide emissions, the environmental kuznets curve was found, Evran [21] revealed the long run estimation results suggest that economic development and energy consumption have positive impacts on CO₂ emissions and this research did not find any significant relation between foreign direct investment and CO₂ emissions, Mahmood [22] the study found the long run relation in the model but short run relation did not exist. foreign direct investment, manufacturing value added and population density have positive impact on carbon dioxide emissions.

And Shaari [6] manifested the existence of relation between the variables (FDI, CO₂ and GDP), foreign direct investment did not have any effect on CO₂ emissions, granger causality based on VECM was applied and the results suggest there is no effect of FDI and GDP on carbon dioxide emissions in the short run, Tang [7] insisted the existence of long run equilibrium among the variables of interest, energy consumption and income positively impact CO₂ emissions, there are bidirectional causalities between CO₂ emissions and income, and between FDI and carbon dioxide emissions. Moreover, energy consumption is found to granger-cause carbon dioxide emissions in the short run and long run, Yanchun [8] found that foreign capital penetration has a significant negative effect on CO₂ emissions, exports aggravate the pressure on CO₂ emissions and the results also found the negative effect of domestic investment and agriculture shares on total carbon dioxide emissions over the same period, Yassine [23] substantiated there is a cointegration relation between CO₂ emissions and economic variables and the existence of a unidirectional causality from carbon dioxide emissions to the independent variables and Peng [24] explored economic development, foreign direct investment and CO₂ emissions in China: panel granger causality analysis etc.

As aforementioned contents above, despite many researchers had widely explored as to the impact of economic development on CO₂ missions. However, there has been no any empirical investigation in case of Lao PDR, in other words, it has not been confirmed as the accurate and authoritative evidence. Hence, this article will be studied by employing auto-regressive distributed lag (ARDL) bounds model approach in order to study and capture carbon dioxide emissions and the prospects for Lao PDR’s economic self-development, in both short-run dynamics and long-run relation. Equally important, the researchers are definitely intent to be expected that the results would be salutary to be properly utilized in practice as well as the essential component to be formulated the appropriate policy and consistent with taking care of CO₂ emissions for Lao PDR’s volition, in the status quo and imminent future.

3 Data and methodology

To attain the intentions of this research, it is of the vital ingredient to acquire the actual data and information adequately. Wherefore, such those time series data (all the variables) are obtained from World Bank (WB), in the spanning of 1988–2016 especially: carbon dioxide emissions (CO₂) measures as metric ton per capita, economic development (GDP) calculates as dollar (take logarithm of GDP), and foreign direct investment (FDI) computes to be dollars (take logarithm of FDI) in each year. Similarly, the special analytic techniques and advanced econometric model approach (i.e. ARDL bounds approach to cointegration) Pesaran [25], Shin [26] must be employed in order to investigate and capture the relation among variables in the estimated models.

3.1 Unit root

The property of unit root is of the essential part to be selected whether the specific techniques and models are fit application, by avoiding the various spurious estimation and wrongly applied in practice. Similarly, it is main problem in the time series analysis, which most of time series data have non-stationary characteristic. However, there are many methods of unit root testing that have been proposed by economists to handle whether a time series is stationary characteristic. Therein, the most popular technique to take care of unit root is well known as the new version of Dickey and Fuller (augmented Dickey Fuller: ADF), Dickey [27] and [28], which can be displayed as follows: $Δ y_{t} = α_{0} + β y_{t - 1} + \sum_{t = 1}^{k} γ_{i} Δ y_{t - 1} + ɛ_{t}$ (3.1)

Here, we are assumed that equation (3.1) is considered as the equation with intercept and the equation with trend and intercept can be rewritten as below: $Δ y_{t} = α_{0} + δ T + β y_{t - 1} + \sum_{t = 1}^{k} γ_{i} Δ y_{t - 1} + ɛ_{t}$ (3.2)

Where Δy_t denotes the first difference operator at a time, α₀ is the intercept (constant), δ, β and δ are coefficients, k represents as the lagged value of Δy and ɛ_timplies the white noise error term. Similarly, the first step of unit root testing process will be performed at the order of integration I (0) or at level in each variable. Nevertheless, if each of them is non-stationary characteristic at the order of I (0), then the testing process will be re-tested again at the first different level, in other words it must be conducted at the order of integration I (1) and will be continued at the second difference if the series is not stationary characteristic at the order of integration I (1). However, time series are predominantly stationary at the first difference.

The existence of unit root can be tested whether a variable has stationary characteristic, is considered from the result of F-statistics testing compared to the critical value at the 1%, 5% and 10% significant level respectively (this study is pondered 5% significant level). That is, if F-statistics value exceeds the critical value (p-value must be lower than 0.05) at the 5% significant level, implying that the null hypothesis of unit root (non-stationary) will be rejected. In other words, the alternative hypothesis will be accepted the stationary characteristic.

3.2 Optimal lag selection criteria

Generally, prior to performing ARDL bounds testing approach to cointegration, the optimal lag length selection is major consideration to be determined in order to acquire the proper model. Similarly, there are various techniques to be stipulated the appropriate lag for model estimation namely: Akaike information criterion (AIC), Schwarz Bayesian criterion (SBC) and Hannan-Quinn criterion (HQC) respectively. Therein, AIC method is one of the most popular application to select the appropriate lag length in the econometric model analysis which can be expressed the general form as below: $Cr (n) = \log {\tilde{σ}}_{u}^{2} (n) + c_{T} φ (n)$ (3.3)

From equation 3.3, we can re-write each form of the optimal lag selection as equation: $AIC (n) = \log {\tilde{σ}}_{u}^{2} (n) + \frac{2}{T} n$ (3.4) $HQ (n) = \log {\tilde{σ}}_{u}^{2} (n) + \frac{2 log log T}{T} n$ (3.5) $SC (n) = \log {\tilde{σ}}_{u}^{2} (n) + \frac{logT}{T} n$ (3.6)

Where ${\tilde{σ}}_{u}^{2} (n) = T^{- 1} \sum_{t - 1}^{T} {\hat{u}}_{t} {(n)}^{2}$ denotes the error variance estimator, φ (n) is the order of the fitted process, C_T means 2/T, 2 log T/T and log T/T for the Akaike information criterion (AIC), Hannan-Quinn criterion (HQC), and the Schwarz Bayesian criterion (SBC) respectively. Likewise, in this article we are pondered (AIC) as the criterion or optimal lag length. That is, the appropriate lag selection is based on the lower value of Akaike information criterion (AIC) [29, 30]. Hence, the optimal lag in this empirical study will be selected as the smallest value of Akaike information criterion (AIC) for model estimations.

3.3 ARDL bounds testing approach

The classical estimation (regression model) is common use in econometric model under the property of variables are stationary at the order of integration I (0) (at level or no unit root). In contrast, if the underlying variables are not stationary at level, either stationary at the first difference or second difference and then the regression estimation (OLS) can not apply to be estimated. That is, it will lead to be misleading inference and result in the spurious regression estimation etc.

Similarly, when considering the long run relation. Therein, the traditional co-integration property or long run relation between variables was firstly proposed by [31, 32] and Johamen [33] co-integration technique to examine the long run relation between the underlying variables in the estimated models. This technique can be applied in the event of all variables have no stationary at level and stationary characteristic at the order of integration I (1). However, if the underlying variables have the mixture of I (0), I (1) Johamen [33] cannot be utilized and that may lead to the spurious and inefficient estimation. Subsequently, Pesaran [25], Shin [26] proposed ARDL bounds testing approach to co-integration. Where the advantageous feature of this technique can be adopted irrespective of whether all variables are I (0), I (1) or mixture of both, but none of them must be I (2). Likewise, this procedure can be authoritative and efficient estimation when likened to Johamen [33] co-integration technique. Therefore, this empirical study will be explored the short run dynamics and long run relation by employing ARDL bounds testing approach can be written in the shape of multiple regression as follows: $C 2_{t} = β_{0} + β_{1} {lnGDP}_{t} + β_{2} {lnFDI}_{t} + ɛ_{t}$ (3.7)

From equation (3.7) above. Hence, it can be transformed into the form of ARDL bounds testing approach models as following equations: $\begin{matrix} Δ CO 2_{t} = β_{1} + \sum_{i = 1}^{k} δ_{A} Δ CO 2_{t - i} \\ + \sum_{i = 1}^{k} δ_{B} Δ {lnGDP}_{t - i} + \sum_{i = 1}^{k} δ_{C} Δ {lnFDI}_{t - i} \\ + γ_{1} CO 2_{t - i} + γ_{2} \ln {GDP}_{t - 1} + γ_{3} \ln {FDI}_{t - 1} + μ_{t} \end{matrix}$ (3.8) $\begin{matrix} Δ \ln {FDI}_{t} = β_{2} + \sum_{i = 1}^{k} δ_{E} Δ \ln {FDI}_{t - i} \\ + \sum_{i = 1}^{k} δ_{F} Δ CO 2_{t - i} + \sum_{i = 1}^{k} δ_{G} Δ \ln {GDP}_{t - i} + \\ γ_{4} \ln {FDI}_{t - i} + γ_{5} CO 2_{t - i} + γ_{6} \ln {GDP}_{t - i} + μ_{2 t} \end{matrix}$ (3.9) $\begin{matrix} Δ {lnGDP}_{t} = β_{2} + \sum_{i = 1}^{k} δ_{H} Δ {lnGDP}_{t - i} \\ + \sum_{i = 1}^{k} δ_{I} Δ CO 2_{t - i} + \sum_{i = 1}^{k} δ_{J} Δ {lnFDI}_{t - i} + \\ γ^{7} {lnGDP}_{t - i} + γ^{8} CO 2_{t - i} + γ^{9} {lnFDI}_{t - i} + μ_{3 t} \end{matrix}$ (3.10) Where Δ represents the first difference operator, β_m denotes the constant of each model, μ_qt implies the white noise process, and δ_n and γ_p is the short run and long run coefficient respectively. Similarly, the existence of long run relation between variables of each estimated model can be tested under the following hypothesizes:

Model 1 H₀: δ_A = δ_B = δ_C = 0 (the non-existence of co-integration or long run relation)

H₁: δ_A ≠ δ_B ≠ δ_C ≠ 0 (the existence of co-integration or long run relation)

Model 2 H₀: δ_D = δ_E = δ_F = 0 (the non-existence of co-integration or long run relation)

H₁: δ_D ≠ δ_E ≠ δ_F ≠ 0 (the existence of co-integration or long run relation)

Model 3 H₀: δ_G = δ_H = δ_I = 0 (the non-existence of co-integration or long run relation)

H₁: δ_G ≠ δ_H ≠ δ_I ≠ 0 (the existence of co-integration or long run relation)

As hypothesis above, in order to draw a conclusion whether the null hypothesis is rejected, we are considered from F-statistics. That is, if F-statistics value exceeds upper bounds critical value, the null hypothesis (H₀: δ_A = δ_B = δ_C = 0, δ_D = δ_E = δ_F = 0, δ_G = δ_H = δ_I = 0) will have been rejected (the non-existence of co-integration or long run relation). In other words, the alternative hypothesis (H₁: δ_A ≠ δ_B ≠ δ_C ≠ 0, δ_D ≠ δ_E ≠ δ_F ≠ 0, δ_G ≠ δ_H ≠ δ_I ≠ 0) will have been accepted (the existence of co-integration or long run relation) and the result is substantiated the presence of co-integration or long run relation between variables (variables are co-integrated). By way of contrast, if F-statistics value is less than lower bounds critical value, denoting that we cannot reject the null hypothesis and then we can draw a conclusion that there is no cointegration or long run relation between the underlying variables (variables are not cointegrated). Moreover, in case of F-statistics value lies in between lower and upper bounds critical value, the result of co-integration testing is inclusive and the order of integration will be examined again whether the underlying variables are stationary characteristic at level or first difference.

In addition, the steps of estimated models have been divided into seven essential elements such as: (1) unit root test (2) optimal lag selection, (3) bounds testing, (4) long run and short run relation, (5) short run and long-run causality relation (6) variance decomposition and (7) impulse response function respectively. Equally important, this study must be checked the relevant testing problems as well as the diagnostic statistics in the estimated models specifically: heteroskedasticity, autocorrelation or serial correlation, non-normal distribution, multicollinearity, CUSUM test and CUSUM of square test and pertinent testing processes respectively.

Nevertheless, all of the estimated processes are based on the existing hypothesis of their particular properties. Above all mentioned have been expressed regarding the thoroughly analyzed steps. Hence, the next steps can be estimated the relation among variables made use of the models.

4 Results

4.1 Unit root testing

Before running the stipulated properties in the specific models, it is of the primary strand to check unit root whether the series has stationary characteristic, by applying the most popular and well known as augmented Dickey Fuller (ADF), Dickey [27] and Phillips [34]. The result of unit root testing can be embodied as follows:

Table 1.1 above, demonstrates the result of unit root testing and we found that all the variables have non-stationary characteristic at the level or I (0) (the presence of unit root). Nevertheless, they became to be stationary characteristic after altered to be the first difference or the order of integration I (1) at 1% and 5% significant level respectively.

Moreover, before conducting to ARDL bounds testing approach to cointegration, it is of essential constituent to stipulate the appropriate lag length. In this case, the optimal lag length is 4, by according to the lowest value of (AIC) [29]. Hence, this research will be utilized lag 4 to be estimated the ARDL bounds approach to cointegration.

4.2 ARDL bounds testing approach

The result of ARDL bounds testing approach is presented in Table 1.2 above embodies that, F-statistics value of all endogenous variables (all models) are greater than lower bounds and upper bounds at 1% of the significant level. Therefore, we can definitely reject the null hypothesis of no-cointegration (no-long run relation) and the result was strongly asserted the existence of cointegration. In other words, it has long run relation among variables at the 1% significant level.

4.3 Long run relation

Table 1.3 displays the long run relation among variables in the estimated model. The empirical results discovered that, FDI is negatively associated with CO₂ that is, a 1% increase in FDI conduces to decrease 0.12% of CO₂ [7, 8], due to the major foreign direct investment (FDI) are obtained from hydro-power and mining, which highly cover at 29.06 and 23.79% of total investements from 1989–2014. More significantly, Lao government has also implemented the various strategic plans and encouraged the numerious industries to participate and contribute on the climate change issues, as well as maintaining the environmental circumstances. whereas GDP has a positive impact and statistically significant, that is to say, if GDP accrues by 1% will result in CO₂ accumulates 0.33%, with 1% of the significant level, which this empirical literature has been in line with [1 , 17] and Uddin [16] respectively. However, if we are determined other aspects to be constant, CO₂ will decrease 4.63% with the 1% of significant level.

4.4 Short run relation

The results of estimated model in the short run relation among variables are demonstrated in Table 1.4, it is found that GDP in each lagged period has a negative influence on CO₂, demonstrating that one unit of increasing in GDP at each lagged period of time will engender to decrease one unit of CO₂ by 1% significant level. On the contrary, at each lagged period of time for FDI are statistically positive impact on CO₂ at the significant level of 1% etc.

To view the error correction term ECT_t-1 as the speed of adjustment from short run towards long run equilibrium. As presented in Table 1.4 manifests that ECT_t-1 has negative effect and statistically significant. That is, the starting point of the variation from short run to long run equilibrium has been corrected by 57.44% in each year, this indicates the high speed of adjustment to long run equilibrium.

4.5 Long run and short run causality relation

Moreover, since we found the presence of co-integration relation among variables that was embodied in Table 1.2, this can be allowed to apply VECM granger causality approach in order to identify the short run and long run causality relations among variables made use of the stipulated models. Likewise, the long run causality relation is determined by the significant level of error correction term in one lagged period (ECTt_-1) that is, one lagged period of error correction term (ECTt_-1) must be negative impact and statistically significant. Whereas the short run causality can be examined the joint significance by each lagged period of the exogenous variables (wald coefficient test). Following this further is the result of VECM causality relation:

The result of VECM causality relation is presented in Table 1.5 above, which exhibits that ECT_t-1 is negative and statistical significance and this can be substantiated the presence of a unidirectional causality relation between GDP and FDI on CO₂ in the long run. Similarly, this article also confirmed the existence of unidirectional causality relation running from FDI and GDP to CO₂ in the short run. In other words, CO₂ is caused by GDP and FDI in both short run and long run [35]. Conversely, there is no short run and long run causality relation running from CO₂, GDP to FDI and CO₂, FDI to GDP (In the event of FDI and GDP are the proxy of endogenous variables).

More accurately, residual and stability diagnostics are significant part to be checked in order to avoid misspecification, spurious and inefficient estimation. As presented in Table 1.6, the result corroborated the estimated models have no serial correlation, heteroskedasticity, non-normality distribution and multicollinearity. More precisely, the results of CUSUM test and CUSUM of square test are embodied in Fig. 1.1 can be averred that both testing procedures have the stable characteristic in residual variance, due to (the middle line) they shift to be fluctuated within the lower and upper bounds line at 5% significant level.

Fig.1.1

the result of CUSUM test and CUSUM of square test.

4.6 Variance decomposition

We also applied the variance decomposition approach in order to attest that how much of the forecast error variance in the target variables can be elucidated by the exogenous impulses or shocks. As reported in Table 1.7, we found that the own shock of CO₂ can be explained 62.79% fluctuation and the impulse in CO₂ is described by 11.50 and 25.71% of FDI and GDP (In the fourth period). Likewise, the innovation to CO₂ is explained by its own innovation accounts for 56.19% and can be caused by 23.87 and 19.94% of both FDI and GDP respectively (in the twentieth period).

Pertain to GDP, our result discovered that GDP impulse can be explained by its own shock equals 84.65% volatility, GDP can also be contributed by CO₂ and FDI cover at 14.96 and 0.39% fluctuation (in the fourth period). Whereas, the shock to GDP is described by 90.98% of its own and can be contributed 8.46 and 0.55% variation by CO₂ and FDI respectively (In the twentieth period). Moreover, this article is also found FDI can be well-explained by CO₂ and GDP more than its own shock in the long run.

4.7 Impulse response function analysis

Impulse response function approach was also adopted to insist how the pertinent variables react to a shock in another variable at each period of time (the effect of one-time shock to the innovation on current and future value of the endogenous variables). That is to say, how the pertinent variables response to each other under one standard impulse at over the time period etc.

The results of empirical study are reported in Fig. 1.2 found that, a positive shock of one standard deviation error term to CO₂ reacts to its own in the positive direction, in the long run (starting from the next ten years). likewise, one standard deviation shock of GDP indicates the negative response towards CO₂ in the first four years and becomes the positive effect over the time period (in the long run).

Fig.1.2

Impulse response function.

Furthermore, this result also exhibits the negative response of FDI to CO₂ due to one standard deviation impulse at the first four years, transforms to be the positive impact after three years and finally turns into the negative shock over the time period, in the long run. This empirical result also discovered that the reaction of FDI to CO₂ and CO₂ to its own have identical fluctuation and negative impulse to one another in the long run.

5 Conclusion and discussion

The main intention of this research is to explore carbon dioxide emissions and the prospects for Lao PDR’s economic self-development by applying ARDL bounds testing approach to cointegration. This empirical investigation discovered that GDP has statistically significant and positive impact on CO₂ in the short run and long run, our result also found the presence of cointegration among variables, in other words there occurs the long run relation in the case of Lao PDR. Likewise, this article has also been substantiated the feedback effect between CO₂ and FDI, GDP has the positive impulse on CO₂ in the long run. In addition, this empirical research merely found the short run and long unidirectional causality relation flowing from GDP and FDI to CO₂. However, there is no short run and long run causality relation running from CO₂, GDP to FDI and CO₂, FDI to GDP.

6 Research limitations

In this study, authors are accentuated in few major influences affecting on carbon dioxide emissions by merely employing ARDL model analysis. however, we are not considered only the environmental kuznets curve (EKC) technique and other major determinants. equally important, we also have limitation regarding to data analysis that is, this research has adopted small time series component due to lacking of data and existing information (inadequate data). Which authors utilized in the period of 1988–2016 and merely applied the quantitative analysis. Moreover, this article has investigated only the single co-integration relation (ARDL single long run relation), while authors has not taken into account the multiple co-integration relation (ARDL multiple long run relation) and this study is a pioneering exploration of empirical evidence regarding Lao PDR’s yearning and prospects for economic self-development, toward which gravitate the article’s future research directions and policymaking recommendations etc.

Worldwide, most business enterprises and other societal organizations are forced to form themselves and to function under the inhumane, top-down imposition of the aye-unnatural and faulty, bureaucratically-hierarchized authority and power sham, a delusion of organization that subjugates and enslaves people, in lieu of a pragmatically ecological decision-making structure of collegial control and responsibility. In a worldwide response, pioneering business enterprises and other societal organizations in many industries are transforming themselves into truly ecological decision-making structures of collegial control and responsibility.

Moreover, the major feature of this article has been employed the special advanced econometric model as well as auto-regressive distributed lag (ARDL) technique that can be more superior than other techniques under considering the underlying variables. Whereas, most researchers had applied other procedures to be estimated. More accurately, this research applies perfect features for analytic methods in order to explicitly identify the impact of latent reactions among pertinent variables in the short run and long run relation. Therein, some of our techniques and econometric tools have not probably been employed in the previous studies. Therefore, authors are zealous to desire that this article would be the potential impact for publishing as a seminal article in the imminent future.

7 Policymaking recommendations

Even though, Lao PDR is a developing country. It is crucial to instill public awareness and various pertinent sectors to pay enormous attentions and to participate on social corporate responsibility for retaining the environmental degradation and its impact on human being. As a result, carbon dioxide emissions have played an indispensable issue role in attaining economic development in the current circumstance and forthcoming future. Therefore, Lao government should formulate the appropriate policy in both the short-term and long-term implementation in order to establish the proper balance between economic development and foreign direct investment on carbon dioxide emissions alongside with maintaining foreign direct investment promotion and the green-sustainable development aspirations in Lao PDR.

Footnotes

Appendixes

Table 1.7

Variance decomposition of CO2, GDP and FDI

Period(year)	CO₂			lnGDP			lnFDI
	CO₂	lnGDP	lnFDI	CO₂	lnGDP	lnFDI	CO₂	lnGDP	lnFDI
1	100	0.00	0.00	9.91	90.09	0.00	83.01	1.15	15.84
4	62.79	25.71	11.50	14.96	84.65	0.39	74.42	4.37	21.21
8	70.79	12.77	16.45	7.60	92.25	0.15	46.54	41.21	12.24
12	61.10	19.35	19.56	7.10	92.63	0.27	46.61	39.81	13.58
16	56.18	20.33	23.49	8.08	91.43	0.49	50.06	33.80	16.14
20	56.19	19.94	23.87	8.46	90.98	0.55	51.29	31.60	17.10

Source: acquiring from author’s calculation.

References

Khobai

, Le

R.P.

, The relationhip between energy consumption, economic growth and carbon dioxide emission: The CASE OF South Africa, 7 (2017), 102–109.

Obradović

and Lojanica

, Energy use, CO 2 emissions and economic growth – causality on a sample of SEE countries, Econ Res Istraživanja 30 (2017), 0. doi: 10.1080/1331677X.2017.1305785

Naser

, Analyzing long-run relationhip between energy consumption and economic growth in the Kingdom of Bahrain, (2017), 7001. doi: 10.1051/e3sconf/20172307001

Appiah

, Du

, Musah

A.I.

, Afriyie

, Investigation of the relationhip between economic growth and carbon dioxide (CO 2) emissions as economic structure changes: Evidence from Ghana, 7 (2017), 160–167. doi: 10.5923/j.re.20170706.02

Balibey

, Relationhips among CO 2 emissions, economic growth and foreign direct investment and the environmental kuznets curve hypothesis in Turkey, 5 (2015), 1042–1049.

Shaari

M.S.

, Hussain

N.E.

, Abdullah

, Kamil

, Relationhip among foreign direct investment, economic growth and CO2 emission: A panel data analysis, 4 (2014), 706–715.

Tang

C.F.

, Tan

B.W.

, The impact of energy consumption, income and foreign direct investment on carbon dioxide emissions in Vietnam, Energy 79 (2015), 447–454. doi: 10.1016/j.energy.2014.11.033

Yanchun

, FDI and China ’ s carbon dioxide emissions: 1978–2008 2 global climate change and carbon emissions, (2008), 289–293.

Ahmad

, Azreen

, Zulkifli

, Abdul

, Nik

, Yaseer

W.M.

, The impact of economic activities on co 2 emission, 2 (2016), 81–88.

10.

Phouphet

, Faculty

, Impact of climate change on poverty in laos, (2011), 1–22.

11.

Miti

, A cointegration analysis of real GDP and CO 2 emissions in transitional countries, (2017). doi: 10.3390/su9040568

12.

Lægreid

O.M.

, Are carbon dioxide emissions decoupled from GDPgrowth in well-functioning democracies? (2017).

13.

Wolde

E.T.

, Mulugeta

, Hussen

M.M.

, Energy consumption, carbon dioxide emissions and economic growth in ethiopia, (2016), 16.

14.

Cederborg

A.J.

, Snöbohm

. Is there a relationhip between economic growth and carbon dioxide emissions? (2016).

15.

Mir

, Storm

, Carbon emissions and economic growth: Production-based versus consumption-based evidence on decoupling, (2016).

16.

Uddin

M.M.

, Causal relationhip between education, carbon dioxide (CO 2) emission and economic growth in Bangladesh, 19 (2014), 60–67.

17.

Lin

, Beidari

, Lewis

, Energy consumption trends and decoupling effects between carbon dioxide and gross domestic product in South Africa, 2012 (2015), 2676–2687. doi: 10.4209/aaqr.2015.04.0258

18.

Bakirtas

, Bayrak

, Cetin

, Economic growth and carbon emission: A dynamic panel data analysis, (2014), 91–102. doi: 10.14207/ejsd.2014.v3n4p91

19.

Hasson

, Masih

, Energy consumption, trade openness, economic growth, carbon dioxide emissions and electricity consumption: Evidence from South Africa based on ARDL. MPRA Pap No 79424, 2017.

20.

Borhan

, Musa

, Hitam

, The impact of co 2 on economic growth in asean 8, 35 (2012), 389–397. doi: 10.1016/j.sbspro.2012.02.103

21.

Evran

, S, (2017).

22.

Mahmood

, Chaudhary

A.R.

, FDI, Population density and carbon dioxide emissions: A case study of pakistan 1, 3 (2012), 355–361. doi: 10.5829/idosi.ijee.2012.03.04.1760

23.

Yassine

, Mostefa

, Mohammed

, Factors affecting co2 emissions in the BRICS countries: A panel data analysis, 26 (2015), 114–125. doi: 10.1016/S2212-5671(15)00890-4

24.

Peng

, Tan

, Li

, Hu

, Economic growth, foreign direct investment and CO 2 emissions in China: A panel granger causality analysis, (2016). doi: 10.3390/su8030233

25.

Pesaran

M.H.

, Shin

, An autoregressive distributed lag modelling approach to cointegration analysis, Econom Econ Theory 20th Century Ragnar Frisch Centen Symp (1999), 1–31. doi: 10.1017/CCOL521633230

26.

Shin

, Smith

R.J.

, Bounds testing approaches to the analysis, 326 (2001), 289–326. doi: 10.1002/jae.616

27.

Dickey

D.A.

, Fuller

W.A.

, Distribution of the estimators for autoregressive time series with a unit root, 74 (2014), 427–431.

28.

Dickey

D.A.

, Fuller

W.A.

, Likelihood ratio statistics for autoregressive time series with a unit root, Econometrica 49 (1981), 1057. doi: 10.2307/1912517

29.

Akaike

, This week’ s citation classic: Akaike 1974, Curr Contents Eng Technol Appl Sci 51 (1981), 42. doi: 10.1098/rspa.1963.0204

30.

Nkoro

, Uko

A.K.

, Autoregressive distributed lag (ARDL) cointegration technique: Application and interpretation, J Stat Econom Methods 5 (2016), 63–91. doi: 10.1002/jae.616

31.

Granger

C.W.J.

, Some properties of time series data and their use in econometric model specification, Journal of Econometrics 16 (1981), 121–130.

32.

F BYR. Co-Integration and Error Correction: Representation, Estimation, and Testing Author (s): Robert F. Engle and C.W.J. Granger Reviewed work (s): Published by: The Econometric Society Stable URL: http://www.jstor.org/stable/236. yet drift 2012;55:251–76.

33.

Johamen

, Jtiselius

, Maximum likelihood estimation and inference on cointegration - with appucations to the demand for money, (1990), 2.

34.

Phillips

BYPCB

. Testing for a unit root in time series regression, (1988), 335–346.

35.

Ren

, Yang

, An empirical research on the relationhip between foreign direct, 809 (2013), 951–957. doi: 10.4028/www.scientific.net/AMR.807-809.951