A NEW LOOK AT THE REMITTANCES-FDI- ENERGY-ENVIRONMENT NEXUS IN THE CASE OF SELECTED ASIAN NATIONS

1. Introduction

The growing hazards associated with climate change have been among the foremost global concerns for several decades (Antonakakis et al., 2017). The relationship between financial development, GDP growth, trade, energy use, FDI, and pollution is a key subject of study in the ecological economics literature. Our work explores the impact of remittances, energy use, and FDI on the environment. A growing body of research has shown a robust association between climate change and CO₂ emissions that is produced through energy production and consumption. For instance, many studies employing panel data or time series, such as (Alam et al., 2016; Cansino et al., 2015; Dogan and Seker, 2016; Chen et al., 2016), deduced a positive relation among energy use and pollution emissions. Fei et al. (2011) found that a 1% surge in per capita raises the use of energy around 0.48–0.50% and leads to release CO₂ about 0.41–0.43% in China in the long-term. Energy is one of the essential inputs of production. In the last decade, there has been a considerable volume of literature on the connection among energy use and GDP growth, as well as among GDP growth and environment degradation. However, the empirical evidence leaves debate concerning many aspects, as these works vary considerably; for instance, some consider a single country and others groups of countries.

The linkage among energy use, FDI, growth, and CO₂ has received substantial attention from policymakers and researchers, and there is abundant related literature. Many studies, such as Zhu et al. (2016), Sapkota and Bastola (2017), Neequaye and Oladi (2015), have attempted to shed light on the energy–growth–emissions nexus by applying various techniques. Their results reveal a positive effect of energy use and GDP growth on CO₂ emissions. Baek (2016) estimated the cause of energy use, FDI, and income on CO₂ emission in 5 ASEAN nations using pooled mean group methods and found that FDI leads to a rise in CO₂. Many other studies, such as Abdouli and Hammami (2017), Sun et al. (2017), Abdouli and Hammami (2017) and Sapkota and Bastola (2017), have also revealed the consequences of FDI on environment and recognized as the FDI–income–environment relation; all of these studies find that the inflow of FDI harms the environment in developing countries. However, Zhang and Zhou (2016) examined the influence of FDI on the release of CO₂ in China by utilizing panel data at the provincial level and deduced that FDI reduces carbon emissions. Their results hold true the pollution halo hypothesis, that multinational companies (MNCs) from advanced nations export green technologies to less developed nations.

The association between remittances and the environment has not yet been addressed. South Asia recently received approximately 117 billion dollars in 2015, placing the developing countries at the top of the list of remittance recipients. Among the Asian countries, a significant amount of remittances flows to India, China, Bangladesh, Philippines, Pakistan, and Sri Lanka (World Bank). Though they remained elastic throughout a period of crisis, remittances continue to rank second after FDI as a key source of external financing for many developing nations (Goschin, 2014). For many developing nations, remittances contribute the main share of capital inflows, exceeding export revenues, FDI, and external aids (Meyer and Shera, 2016). In our study along with remittances, we will also check the pollution haven hypothesis, that state that FDI leads less developing nations to become pollution havens for developed nations, and halo hypothesis according to which FDI will bring clean technologies (Gill et al., 2018). As according Lim and Basnet (2017) stated that remittances lead to raise the income level of the people in the case of permanent or transitory incomes hypotheses, as the former one states that an increase in income permanently increases recent consumption, while the later one states that an augmentation in income is saved or smoothed over a life time. In the case of Asian economies, migrants are short-term workers who save money and invest it when they return to their home countries, which positively contributes to the economic growth of Asian economies (Makun, 2017; Meyer and Shera, 2016; Lim and Basnet, 2017; Ofeh and Muandzevara, 2017; Donou-Adonsou and Lim, 2016). Remittances increase GDP per capita, which leads to high demand for energy consumption and in turn affects the environment in developing economies.

For many developing nations, growth and CO₂ emissions go hand in hand. For instance, many other studies, such as Narayan et al. (2016), Ahmad et al. (2017), Ahmad et al. (2018), Balsalobre-Lorente et al. (2018), Appiah (2018), Ito (2017), Rahman and Ahmad (2019) and Azevedo et al. (2018), used different techniques and sample sizes of panel and time series data and deduced that economic growth, energy use, and per capita income positively affect CO₂. To date, no study has yet considered remittances to influence environmental quality. This work tries to fill the gap in the existing studies by taking remittances as a tool of economic development and examining their role in carbon emissions. We use the following conceptual framework to clarify the possible association between remittances and carbon emissions in six top remittance-receiving countries in Asia. Among these economies, the three countries of China, India, and Pakistan are among the top CO₂ emitters, while China and India are the top energy consumers and FDI destination nations worldwide.

1.1. Conceptual framework

There is no literature available so far on the relationship between remittances and CO₂. Therefore, we have explained this relationship in (Figure 1) with the help of our hypothesis formulation below.

As the permanent income hypothesis affirms that increases in income permanently increase current consumption, while the transitory income hypothesis indicates that increases in income through remittances are saved or smoothed over a life time (Lim and Basnet, 2017). Remittances can also lead to high GDP per capita, which can lead to high demand for energy use and a return cause the environmental problems in developing economies. A researcher (Guo, 2017) who scrutinized the association between household income inequality and the release of CO₂ in China found that the increase in household income disparity positively affects CO₂ emissions. The causes for this positive impact are an upsurge in the disparity in household income with lower consumer demand and a higher scale of investment, which leads to excess capacity, increased waste in energy use and, hence, arise in CO₂ emissions. Remittances lead to economic growth, and for many developing nations, GDP growth and carbon emissions are positively related, and GDP per capita and energy use have a positive association with release of CO₂. In their studies, Ratha (2013) and Prabal and Dilip (2012) examined the relationship between remittances and household income. They found that on average, remittances increase a household’s income and consumption levels. Further, Iheke (2012) studied the effect of remittances on per capita in Nigeria, and his findings revealed that remittances improve per capita income and lead to increased household consumption. In contrast, Baldé (2013) concluded that remittances increase the level of individual savings and investment.

Increases in consumption and savings induce aggregate demand and bank deposits (Bhole, 2009; Sawyer and Sprinkle, 2015). Increases in aggregate demand and bank deposits increase production and improve the financial sector (Hahnel, 1999; Herr and Kazandziska, 2011). The boost in production leads to a rise in the use of energy, which is one of the key factors for environmental degradation. Many researchers have also considered the connection among CO₂ and financial development for different sample nations. Their results showed mixed outcomes, indicating that financial development can decrease or increase CO₂ emissions. Researchers such as Mulali et al. (2015), Saidi and Mbarek (2017) and Charfeddine and Khediri (2016) found that growth in financial development mitigates CO₂. Other scholars (Shahzad et al., 2017; Ahmad et al., 2019; Li et al., 2015; Haseeb et al., 2018) confirmed that financial development boosts the CO₂ emissions. Increases in financial development and industrial production also positively contribute to economic growth (Kedourie, 2013; Lakhera, 2016). Economic growth and industrialization can lead to rise or decline in CO₂ emissions. Other studies, including (De Bruyn et al., 1998; Heil and Selden, 2001; Stern and Common, 2001; Friedl and Getzner, 2003; Dinda, 2004; Tamazian and Rao, 2010), have been conducted on the association between financial development and the release of CO₂ and concluded that there can be positive and negative associations between these variables. So, from our earlier discussion we see that, thus, remittances are the main determinant of financial development and economic growth. Hence, it is essential to use an appropriate model to inspect the significance of remittances on the release of CO₂ in Asian economies. Furthermore, various studies have also identified that FDI inflow as an essential element of economic growth in some countries (Gunby et al., 2017). The associations between FDI and CO₂ are addressed by a few researchers according to the pollution haven or halo hypotheses. The former declares that the inflow of FDI causes environmental issues in the home country through the movement of dirty industries from advanced nations to developing nations; hence, developing nations become pollution havens for developed nations due to their loose environmental regulations. In contrast, the pollution halo hypothesis claims that FDI inflows to the host nation bring clean and advanced technologies that lead to a cleaner environment as MNCs import green technologies from advanced countries to developing nations and carry out business in an environmentally friendly way (Rahman, Chongbo, andAhmad, 2019).

The main contributions of our work are, first, our investigation of the impact of remittances, FDI, and energy consumption on CO₂ emission by employing the ARDL model techniques for the top remittance-receiving nations in Asia. ARDL-bound tests are generally employed to check the long-term cointegration association amongst variables. The ARDL technique is considered to have more significance and accuracy for small samples and avoids endogeneity problems related to the variables. Important events related to the economy or policy shift can leads breaks in the time series, by disregarding structural breaks in the series can lead to the omission of model settings and biases so, we find the breakpoint (structural break) and incorporate it into our model. We also choose the major samples of Asian countries, which consist of top CO₂ emitters, remittance and FDI recipients, and energy consumers.

2. Data Source and Description

We utilized yearly data for the period 1982–2014 for the top six remittance-receiving economies in Asia, namely, China, India, The Philippines, Pakistan, Sri Lanka, and Bangladesh. Our sample size is balanced with respect to all countries and variables. We choose these countries because they include top CO₂ emitters, remittance and FDI recipients, and energy consumers. We collect the data from the World Development Indicators (World Bank). Table 1 reports some common statistics of the data utilized in the analyses.

3. Specifications of the Model and Methodology

We used the following simple regression model :

where $i=1,\dots ,6$ indicate the countries and $t=1982,\dots ,2014$ denotes the time period. ${\beta }_0$ is a constant, ${\beta }_1$, ${\beta }_2$, ${\beta }_3$ indicate coefficients and $\epsilon$ is the residuals. In this work, coefficient ${\beta }_1$ is of most interest to us to empirically check the effect of remittances on CO₂. We claim that remittances can lead to growth in developing economies, cause CO₂ emissions and, as they increase the level of individual income, result in higher energy use and production, which can also lead to increased CO₂. Likewise, if remittances are used for financial development, then CO₂ can increase or decrease. Hence, it is anticipated that ${\beta }_1>0$ or ${\beta }_1<0$. Moreover, FDI can augment CO₂ through a rise in the production of goods or the transfer of the dirty industry to the host country, or it can decrease CO₂ through the import of green and advanced technologies to the host country. Hence, it is also expected that ${\beta }_3>0$ or ${\beta }_3<0$ (Baek, 2016). Energy use is predicted to have a positive effect on CO₂; thus, ${\beta }_4>0$.

3.1. ARDL bound test

To analyze the short and long-runs effects, in this work, we employ the ARDL method recommended by Pesaran et al. (1999). This technique is considered effective for checking cointegration association and its robustness regardless of the integration order like, I(1), I(0), or mixed order cointegrated, but it is not valid for the I(2) series. Equation (1) can be written in ARDL form:

Next, the error correction models can be estimated for short-term dynamics as follows:

$$\begin{array}{l}{\mathrm{\Delta }\mathrm{\text{CO}}}_{2it}={\alpha }_0+\sum _{i=1}^p{\alpha }_1{\mathrm{\Delta }\mathrm{\text{CO}}}_{2t-i}+\sum _{j=1}^q{\alpha }_2{\mathrm{\Delta }\mathrm{\text{REM}}}_{t-j}\\ +\sum _{m=1}^q{\alpha }_3{\mathrm{\Delta }\mathrm{\text{ENU}}}_{t-m}+\sum _{r=1}^q{\alpha }_4{\mathrm{\Delta }\mathrm{\text{FDI}}}_{t-1}+\theta {\mathrm{\text{ECT}}}_{t-1}+{\epsilon }_{it},\end{array}$$(3)

where $\mathrm{\Delta }$ represents the first deference operators, while ${\theta }_i$ indicates the coefficients of error correction terms (ECT) for short-term dynamics.

3.2. Estimation procedure

In estimating Equation (2), we perform a Wald test to verify the presence of long-term associations among variables. $H_0$: ${\beta }_1={\beta }_2={\beta }_3=0$ suggests that no cointegration exists, while Ha: ${\beta }_1\ne {\beta }_2\ne {\beta }_3\ne 0$ is opposite to the null hypothesis. The $F$-test value is compared with the lower and upper bound critical values calculated by (Pesaran et al., 2001). The higher the $F$-test value than the upper bound will lead to rejection of $H_0$ of no cointegration relations, and vice versa.

4. Empirical Results and Analysis

4.1. Unit root test

To check the stationary issues of the series, we applied the Augmented Dickey–Fuller (ADF) test. The results, reported in Table 2, verify that all of the variables are non-stationary at the level, but after taking the first difference, the non-stationary variables become stationary.

4.2. Breakpoint unit root test

The conventional unit roots provide biased results due to missing information regarding structural breakpoints in series. In this work, we applied Narayan and Popp (2013) proposed test with two structural breaks in the slope and level, whose time of occurrence is supposed to be unknown and that are modeled as innovational outliers and hence take effect gradually. Employing Monte Carlo simulations, Narayan and Popp (2010) stated that the test has stable power and correct size and identifies structural breaks more precisely than tests used in previous studies, such as Zivot and Andrews (1992), Lumsdaine and Papell (1997) and Lee and Strazicich (2003). They employed two conditions for the deterministic component as; model (M1) allows for two breaks in the level, represented in Table 3, and model (M2) permits two breaks in the level as well as the slope of the deterministic trend component, represented Table 3.

4.3. Cointegration bound test

Integrating the information from the structural breaks, we compare the critical $F$-statistics against the lower and upper bounds of (Pesaran et al., 2001) the results are shown in Table 4. That clearly indicates a long-term co-integrating association between remittances, energy use, FDI, and CO₂. The calculated $F$-statistic for China is 13.98030, that is larger than the upper bound value; hence, the alternative hypothesis is accepted. The calculated $F$-statistics for India, Pakistan, The Philippines, Bangladesh, and Sri Lanka are 5.591886, 7.629929, 14.84171, 11.10835, and 5.440190, respectively, which are all larger than the upper bound of 4.37. Thus, $H_0$ is clearly rejected at the 1% significance level.

4.4. Long-term and short-term estimations

Tables 5 and 6 provide long-term and short-term results. From the bound test, the cointegration relation among all the variables is clear.

Table 5 indicates the long-term impacts of the particular variables on carbon emissions, and the six countries have mixed results. In the long term, energy use has a significant positive influence on the release of CO₂ in each country, and these are significant at the 1% level for China, The Philippines, and Pakistan, while for Bangladesh, India, and Sri Lanka, the influence is significant at the 5% level. Hence, energy use is the key determinant of CO₂ emissions in all sample nations. These results are corroborated with the work of Zhu et al. (2016), Sapkota and Bastola (2017) and Neequaye and Oladi (2015). Conversely, remittances have a significant positive effect on CO₂ in the case of Pakistan, Sri Lanka, Bangladesh, and The Philippines, as according to different studies, a significant number of immigrants of these countries go abroad, especially to the Gulf countries, and send money back to their home countries. Most of this money is not sent through regular channels, such as banks, so it does not improve financial development. Furthermore, these migrants are short-term workers abroad, and the money sent back is saved and invested when they return home. This investment leads to economic growth and improves the individuals’ per capita income, which can create high demands for energy that in turn can harm the environment. From the long-term coefficient, we can see that a 1% raise in remittances will raise CO₂ emissions by 0.2694% in Pakistan and 0.1355% in the Philippines at the 1% significance level. Furthermore, for Bangladesh and Sri Lanka, the long-term coefficients are found to be 0.2169% and 0.2438%, respectively, at the 5% significance level. For China and India, the coefficients are statistically non-significant. The results also reveal that FDI has a significant positive effect on the release of CO₂ in China, Sri Lanka, and India; hence, the pollution haven hypothesis is found to be true in the long term in these countries, as found by Abdouli and Hammami (2017), Sun et al. (2017), Solarin et al. (2017) and Sapkota and Bastola (2017). The governments of these countries should closely watch the inflow of FDI that harms their environment. Furthermore, the coefficient of FDI is found to be negatively significant for Bangladesh ($-$0.094) at the 5% significance level, so it has a negative association with CO₂; hence, a pollution halo does exist in the case of Bangladesh, which implies that most FDI to Bangladesh goes to the agricultural and service sectors. Therefore, the Bangladeshi government should not be concerned about the inflow of FDI, as it can enhance the country’s growth. These outcomes are also corroborated with the results of Zhang and Zhou (2016) in the case of China, who observed that FDI leads to the reduction of carbon emissions. For Pakistan and The Philippines, FDI has a non-significant effect on the release of CO₂ in the long-run. The findings show that each country should try to adopt a proper strategy for remittances and FDI that is beneficial to the green technology sector, service sector, and financial development.

Table 7 shows that the short-term estimation results are almost the same as the long-term estimates. In all sample countries, energy use has statistically significant positive effect on the release of CO₂. However, in the short term, the coefficients of remittances show a significant robust positive association with carbon emissions only in the cases of The Philippines, Pakistan, and Sri Lanka. A 1% raise in remittances will raise carbon emissions by 0.02447% in Pakistan and by 0.1179% in The Philippines, at the 1% significance level. Furthermore, a 1% increment in remittances will cause CO₂ to increase by 0.40581% in Sri Lanka, at 5%. However, for the other three countries, China, India, and Bangladesh, the coefficient is found to be non-significant in the short term. Conversely, FDI inflow has a significant positive impact on the release of CO₂ in the cases of China, Sri Lanka, and India, while it has a negative significant sign for Bangladesh. For Pakistan and The Philippines, the coefficient of FDI is found to be statistically non-significant in the short term. This follows the assumption that loose environmental laws in a nation probably attract more foreign investment in dirty sectors industries, as MNCs seek to avoid strict regulatory agreements in their home nations; hence, foreign investment inflow to be inclined and damage the environment in a host nation, as in the cases of India, Sri Lanka, and China (Baek, 2016).

Notably, the ECM confirms a negative coefficient for all of the sample nations. The significant coefficient of $-$0.5996 at the 1% level for China show, that any deviation from the long-term equilibrium between the variables will be corrected by approximately 59% each year. The ECM for India is $-$0.2205, which indicates adjustment of 22% per year towards the long-term equilibrium. Furthermore, the ECM coefficients are $-$0.9085 for Pakistan, $-$0.8702 for The Philippines, $-$0.4019 for Bangladesh, and $-$0.42618 for Sri Lanka.

4.5. Diagnostic tests

Table 7 presents several diagnostic tests that were performed to check biases in the model. All of the $R$-square and adjusted $R$-square values are greater than 50 for each country, which indicates that our models are well fitted.

We examined (${\chi }^2$ Ramsey Reset) for correct functional form for our estimated models. Also, Table 7 shows the test of serial correlation (${\chi }^2$ SC), the Jarque–Bera test of (${\chi }^2$ Normality), and a heteroskedasticity test based on (${\chi }^2$ ARCH) and (${\chi }^2$ B–G). All diagnostic tests show that there are no problems of functional form, serial correlation, normality, or heteroskedasticity, thus, indicating the non-existence of biases in the models.

4.6. Stability check

To ascertain the dynamic stability of the parameters in the estimated models by using the cumulative sum of the recursive residuals (CUSUM). For all countries (Figures 2–7), the parameters are stable during all sample periods.

5. Conclusion

There are many literatures on the association between FDI, energy, trade openness, growth, and carbon emissions, but thus far, no study has yet considered remittances in the model for top remittance-receiving Asian countries. Remittances have been found to enhance growth and decrease poverty level as per capita income increases; therefore, we empirically checked using the ARDL-bound test whether remittances can lead to CO₂ emissions. The main contribution of this work is its assessment of the impact of remittances, energy use, and inward FDI on CO₂ emissions in six Asian economies through application of the ARDL-bound test with structural breaks, using data for the period 1982–2014. The findings from the ARDL-bound test verify the presence of long-run cointegrating relations between variables. Both the long-term and short-runs results confirm that energy use has a significant positive effect on CO₂ emissions in all sample nations, namely, China, Pakistan, India, the Philippines, Bangladesh, and Sri Lanka. Furthermore, CO₂ emissions significantly increase with increased in remittances in the long-term in the cases of Pakistan, the Philippines, Sri Lanka, and Bangladesh, while the effect is significant in the short term only in the cases of The Philippines, Pakistan, and Sri Lanka. In contrast, inflow of FDI has positive significant effects on the release of CO₂ in China, India, and Sri-Lanka in the long as well as in the short term, hence, confirming the pollution haven hypothesis in these nations, while the result is statistically non-significant for other economies, such as Pakistan and The Philippines. Our analysis also shows that FDI has a significant negative impact on the release of CO₂ for Bangladesh both in the long and short-terms, thus, confirming the pollution halo hypothesis.

Hence, we conclude that the results are mixed for all samples, as the nations are heterogeneous in the way they receive FDI and remittances and consume energy in various economic sectors. The most important contribution of our work is its demonstration that the environmental impacts of rising economic activities may be varied and should be observed on a case by case basis, requiring further research. The study shows that developing nations should pay attention to environmental degradation while seeking to achieve development goals through beneficial energy use, FDI, and remittances. The key recommendations of this work are that nations should pursue diverse growth paths, and policymakers should consider these variables because remittances can be used for financial development that can mitigate CO₂ emissions. Remittances should be sent through formal channels and used for financial development, as they are in India and China, in which cases remittances do not lead to environmental degradation. Energy is an important part of the economy, but countries need to replace oil and fossil fuel consumption with substitute energy sources, like renewable energy and liquefied natural gas, as energy is one of the key factors of pollution. Hence, these governments and policymakers should formulate a special policy for renewable energy by taxing fossil energy and subsidizing renewable energy. Inward FDI plays a key role in the growth of developing nations, but it also adversely affects the environment. Thus, countries should attract FDI in green technology or to the financial or service sectors.

Acknowledgment

We would like to thank the editor and anonymous reviewer(s) for their valuable comments. This work is supported by National Natural Science Foundation of China (71773007, 71403024), Beijing Social Science Foundation (17YJB020), National Social Science Foundation of China (16ZDA026), and Beijing Normal University Cross-disciplinary Project.