Mediating Effects of Foreign Direct Investment Inflows on Carbon Dioxide Emissions

Abstract

In this research, the direct and indirect effects of foreign direct investment (FDI) inflows on carbon dioxide (CO₂) emissions in India are examined, covering the period from 1980 to 2014. To quantify the indirect outcome of the existence of FDI on CO₂ emissions, in this study, the three mediating channels of FDI are considered. The three broad mediating channels of FDI inflows are energy structure, industrial structure, and high-carbon technology, by which foreign direct investments affect India’s carbon dioxide emissions. In this study, the unit root test, the Johansen cointegration, the Granger causality technique, and the seemingly unrelated regression (SUR) are used for the empirical analysis. The findings discover a process of cointegration in the long-run and reveal unidirectional causation between FDI inflows and CO₂ emissions. The outcomes of the SUR estimation indicate that all the mediating factors substantially contribute to the level of CO₂ emissions. In this paper, the findings reveal that FDI inflows affect the level of India’s CO₂ emissions mainly via mediating factors compared to their direct effect. Finally, in this research, it is recommended that the concerned authorities should prioritize the redistribution of foreign direct investment from high carbon-intensive technologies to less carbon-intensive and cleaner technologies for India’s carbonless and sustainable future.

1. Introduction

Climate change, greenhouse gases (GHGs), future energy, and carbon dioxide emissions are all attracting interest from industrialized, developing, and impoverished countries (Narayanan & Sahu, 2016). As per the Future Earth and the Earth League report 2019, discussed at the Conference of Parties (COP) 25, the world has already been diverted from its track of meeting the Paris Agreements (to restrict global warming to 1.5 °C),and climate change is faster and more robust than has been expected1. Further, the report highlighted that this diversion will cause severe consequences for hundreds of millions’ food security, health, and threatens biodiversity. Thus, climate change has become a significant hurdle for accomplishing the United Nations’ Sustainable Development Goals (SDGs) for global development. Besides, with due time, natural scientists have investigated the fact that carbon dioxide (CO₂) emissions are the leading factor in climate change due to an intensified greenhouse effect and the degradation of the atmospheric system (IPCC, 2014). Measuring this severity, the World Bank (2018) reported an 84.38% increase in CO₂ emissions from 1980 to 2014, which has awakened the world to this unexpected change in CO₂ emissions. Thus, to control the extreme climatic events, the melting of the Arctic sea ice, rising sea levels, and other long-term and irreversible changes, it is indispensable to maintain the CO₂ emissions level and increase environmental sustainability (Mardani et al., 2019).

For decades, studies have pointed out that heavy industrialization and high energy consumption are the major causes of the increased CO₂ emissions (Muhammad & Khan, 2019). However, a trade-off occurs between environmental quality and higher economic prosperity. Growing economies have recently been significantly involved in internationalization, industrialization, and energy consumption to achieve their economic growth target (Sabir et al., 2020). Furthermore, the global and local environmental challenges produced by the ever-increasing usage of fossil fuels have reignited concerns, making it an enormous task to fuel economic growth in an environmentally friendly manner (Sahu & Narayanan, 2010). As CO₂ emissions are a global pollutant, recent researchers have emphasized the globalization–environmental quality nexus. At the same time, internationalization is the prime factor in speeding up industrialization, so the higher the level of globalization, the higher the production, consumption, and deterioration inthe quality of the environment (Haider et al., 2019). More specifically, according to the published studies, any initiatives by policymakers and government entities in developed and developing economies to optimize environmental quality will be ineffective unless the environmental impacts of globalization are considered, and CO₂ emissions will persist (Shahbaz et al., 2018). Several decades ago, researchers argued that an economy’s environmental degradation would reduce with a subsequent acceptable amount of economic growth (Grossman & Krueger, 1995). This theory, popularly acknowledged as the environmental Kuznets curve (EKC), examines the different dimensions of the CO₂ emissions on the macroeconomic variables, which can be sub-divided into CO₂ emissions and economic growth (Y. Wang & He, 2019; Hargrove et al., 2019), CO₂ emissions and modernization (Yao et al., 2018), and CO₂ emissions and FDI inflows (Muhammad et al., 2021).

In recent years, there have been two distinct groups of studies in the findings on the association between CO₂ emissions and FDI inflows in recent years. On one side, a group of studies has asserted a favorable correlation between these two factors (Guzel & Okumus, 2020). It has principally emphasized the costs of the growth of any economy in terms of deterioration in environmental quality (Mohsen et al., 2024). This relationship is known as the ‘pollution haven hypothesis’ (henceforth the PHH). This implies that polluting units come from developed and more stringently environmentally regulated countries to developing and less environmentally regulated countries with high environmental pollution. On the other side, a group of studies by J. Huang et al. (2017), Seetanah et al. (2019), and Ehigiamusoe (2020) have highlighted a negative affiliation between these two and claimed that FDI could bring more advanced and energy-saving innovations from developed to underdeveloped countries; thus, helping the economies in building sustainable development. This relationship represents the opposite of the PHH hypothesis and is known as ‘the pollution halo effect’ (henceforth the PHHH).Examining affluent nations like Turkey, France, and Australia, alongside several emerging and developing countries, in the current research has emphasized a positive relationship and confirmed that FDI inflows weaken environmental quality (Jun et al., 2018; Rehman et al., 2021). Thus, this has supported the PHH hypothesis. Conversely, some studies have revealed a negative association between these two and validated the PHHH hypothesis (Nguyen et al., 2021).

Furthermore, Marques and Caetano (2020) pointed out that PHH occurs in developing countries. However, in recent times, the association between CO₂ emissions and FDI remains unclear. Thus, further empirical investigation is needed to cross-check the direction and its nature over time. This vast body in the existing literature on this nexus mainly focuses on the direct linkage of CO₂ emissions (increase/decrease) and FDI inflows. Simultaneously, the existing studies have mostly ignored the indirect association between these two, which primarily accounts for an economy’s different structures and components, such as industrial structure, energy consumption, and carbon technology (Luigi et al., 2021). These three components of any economy’s structure are the highly significant mediating channels to materialize any investment from abroad (Shen, 2018). To examine the effect of such mediating channels of FDI on CO₂ emissions, we follow Shen’s (2018) study in the development of this paper. However, the research is currently limited to the Chinese economy’s carbon productivity and cannot be extended or generalized to a growing democratic economy like India.

With significant historical data from 1980 to 2014, in this study, India is selected for the exploration of the direct and indirect impact of FDI in flows on CO₂ emissions. In this study, India was chosen for the following reasons. India’s economy is among the world’s fastest expanding economies (Chaturvedi, 2017) and is the most significant destination for FDI inflows in the globe, which touched USD 60.1 billion during 2016–17 (UNCTAD, 2018). At the same time, India is significantly vulnerable to the melting of the Himalayan glaciers’ ice caps (Tewari, 2021); the region is susceptible to climate change and changes in the monsoon wind system. Simultaneously, the Carbon brief profile report, India (2019) indicated that India is the third-highest emitter globally, providing 3571 m tons of CO₂ emissions. The World Bank (2018) highlighted that “India could see $1.2 tn of lost GDP, plus lower living standards for nearly half of its population by 2050, compared to a scenario with no climate change.” In addition, carbon dioxide emissions supplement environmental change that is likely to harm human health in growing economies like India (Prasad & Mishra, 2017). Moreover, Watts et al. (2018) indicated that the heat significantly damages the Indian agricultural output and work force by increasing CO₂ emissions and, ultimately, climate change. In relation to these conditions, the government of India has taken several initiatives to control CO₂ emissions and achieve net zero carbon dioxide emissions by 2070. Starting from the Paris Agreement, the economy is following the Nationally Determined Contribution (NDC) with a special focus on reducing carbon intensity by 45% by 2030. The economy is creating numerous carbon sinks by increasing the forest and tree coverage by 2030 and also targeted an increase in the share of renewable energy sources to 40% of installed electric power capacity by 2030; all these initiatives will help India to control greenhouse gases and maintain the temperature increase below 1.5 °C (MoEFCC, 2022; Invest India, 2024). However, the Climate Action Tracker (CAT) has assessed that India’s measures for managing the climatic issues are insufficient. Therefore, the Indian government has to take further active policy measures to meet its net zero emission targets and control the economy of five trillion by 2025 (PIB, 2018; Climate Action Tracker, 2024). However, the government of India has taken initiatives to achieve its dual goal of economic growth and environment sustainability. The Indian economy is the fastest-growing global economy and provides opportunities for global investors to invest in pooled investment funds, sustainability-linked bonds, and green financing investments (Invest India, 2024; Kaushik et al., 2024). India has attempted to minimize the trade-off between FDI inflows (in billion USD) and CO₂ emissions (in million kilotons) but has not succeeded. The general pattern of FDI inflows and the seriousness of CO₂ emissions in India are visible in Figure 1. In the figure, it is clear that the economy receives a positive amount of FDI inflows over time, ultimately contributing to higher industrialization, urbanization, and economic growth. At the same time, it can be seen that the level of CO₂ emissions in India is also increasing over time, which is a direct cost of economic growth and might be an indirect effect of foreign investment. Thus, minimizing the uncertain climatic events and maintaining a balance between foreign investment and India’s sustainable environment is necessary.

In this paper, the following two broad research questions are addressed: first, does FDI inflow directly impact CO₂ emissions? Second, does FDI inflow have any substantial indirect effect on CO₂ emissions? In this research, we seek to determine the relationship between FDI and CO₂ emissions and verify both the “pollution halo” and “pollution haven” hypotheses. To analyze the indirect effects of FDI inflows on CO₂ emissions, in this study, the three mediating channels of FDI are considered of the energy structure, industry structure, and high-carbon technology of the Indian economy. In this study, energy used is considered in terms of the kg of oil equivalent to measure the energy structure. The industrial structure accounts for the whole industrial value-added as a crucial measure of value. The proportion of fossil fuel usage relative to overall energy consumption significantly contributes to carbon dioxide emissions from high-carbon technologies. The novelty of our research is two-fold. First, the mediating influence of FDI on CO₂ emissions is investigated, which assesses the indirect nexus between these two variables. However, while measuring the mediating effects of FDI, in this study, the mediating factors are assumed to be independent. Second, along with cointegration and Granger causality technique, the SUR model is employed to measure all equations together in a simultaneous framework. The outcomes of this research will contribute to the existing literature in several respects. In this study, the unexplored dimensions of FDI and CO₂ emissions will be illustrated through the mediating effects. This research will be helpful for policymakers, researchers, and academia to contextualize the relationship between FDI and CO₂ emissions in many dimensions. First, this study will help policymakers understand the direct relationship between FDI and CO₂ emissions; this will help the host country decide whether foreign investment should be approved depending on its contribution to the environmental quality. Second, the findings of mediating effects can also guide the host countries in deciding which route of foreign investment should be encouraged to increase the GDP and enhance the environmental quality. Third, this study’s outcomes can also provide a comprehensive framework to analyze the impact of FDI on the Indian economy’s GDP, urbanization, and environmental quality. Thus, this research could assess the contribution of foreign direct investments in achieving the sustainable development goals with a special focus on low-carbon development in India. Finally, this study will help academia and researchers to understand the mediating effects using the simultaneous regression models.

The remainder of this work is structured as follows. Section 2 provides a comprehensive assessment of the current theoretical and empirical literature. Section 3 delineates the data, variables, and methods. Section 4 outlines the empirical findings and analysis. The final section presents the conclusion and policy implications.

2. Review of the Literature

In this section, a detailed insight is provided into the theoretical and empirical literature on the FDI–environmental quality association and the effects of the mediating factors of FDI on CO₂ emissions.

2.1. Theoretical Literature

The preliminary premise highlights that the EKC theory associates economic advancement with environmental degradation. This hypothesis, proposed by Grossman and Krueger (1991, 1995), empirically demonstrated the association between environmental degradation and the economy’s expansion. The researchers explained the nexus to illustrate that in an initial phase of economic growth, the pollution rate escalates, resulting in an increased amount of environmental degradation up to a specific point. Then, the sources of economic development press for a reduction in pollution levels and represent the ‘inverted U-shaped’ EKC hypothesis. In contrast, many existing studies on EKC in the literature have also revealed no significant relationship between these two (Stern, 2004). The argument for this might be due to the usage of inappropriate and weak econometric approaches.

The theoretical EKC approach of Grossman and Krueger (1995) has proposed three channels to describe the influence of FDI regarding CO₂ emissions, including scale, composition, and technique effects. The scale effect underlines that FDI promotes industrial production and increases energy demand, increasing CO₂ emissions in the destination nations (Pao & Tsai, 2011). It becomes a threat to the environment as the pollution level is positively associated with the scale of its economy. It is due to the rise in the production of that economy that high production will lead to a higher level of environmental deterioration (Pazienza, 2019). However, some of the new studies in the literature have suggested that environmental destruction will reach a limit due to the economy’s advancement (Luigi et al., 2021). Similarly, Shahbaz et al. (2018) explained the structural effect on the economy’s structure with energy intensity. Additionally, studies have demonstrated the effect of composition and explained the changes in industrial design and how they affect free-market policy (Pazienza, 2019). In contrast, the existing findings have argued that the composition effect in a free trade economy can be either undesirable or encouraging, contingent upon the nation’s competitive edge and productive specialization (Cole & Elliott, 2003). The technique effect is centered on the dissemination of newer machinery via FDI. This effect indicates that the present production technique’s emission level determines the cost per unit of items produced. A free market can liberally invest in neighboring countries, and with locative efficiency, the countries will update their technology and adopt the development of more energy-efficient technology. The technique effect is considered environmentally beneficial for all (Shahbaz et al., 2018; Pazienza, 2019). However, in our analysis, our prime objective is to disseminate the impacts of these three mediating factors of FDI inflows on CO₂ emissions for India but not to concentrate on the EKC hypothesis.

From these three channels, the theoretical literature has been extended. If the scale and the composition effect outweigh the technique effect, this creates the pollution haven hypothesis. Suppose the technique and composition effect overcomes the scale effect, then, the pollution halo effect is created (Pazienza, 2019; Shahbaz et al., 2018). The PHH and the PHHH effect are the two main hypotheses used to describe the relationship amid FDI inflows and pollution. The PHH hypothesis, pioneered by Pethig (1976), argued that multinational corporations (MNCs) relocate pollution-intensive industries to less stringently environmentally regulated countries from more stringent environmentally regulated countries. Supporting the above argument, C. Zhang and Zhou (2016) revealed that MNCs are driven to shift their heavy and polluting activities to nations with more permissive environmental legislation to minimize regulatory compliance expenses in their home nation. However, the PHHH hypothesis demonstrates a positive connection between FDI and the quality of the environment as it provides higher economic growth and environmentally friendly and cutting-edge technologies, enhancing the environmental quality (Dinda, 2004; Zarsky, 1999). This strengthens the idea that Heil and Selden (2001) and Grimes and Kentor (2003) have proposed, that FDI originating from advanced countries comes with advanced and greener technologies, enhancing both economic prosperity and environmental quality.

2.2. Empirical Literature

This section on the empirical literature covers four elements in the existing literature. The first element in Section 2.2.1 covers the relationship between CO₂ emissions and FDI inflows; the link between CO₂ emissions and mediating factors is presented in Section 2.2.2. In the third sub-section under this heading, Section 2.2.3, the affiliation between FDI inflows and the mediating factors is explained. Section 2.2.4 narrates the association between CO₂ emissions and other control variables, such as economic growth and urbanization, followed in Section 2.2.5 with the research gap and hypothesis.

2.2.1. CO₂ Emissions and FDI

A notable segment of the empirical studies in this dimension has examined the correlation between FDI and environmental quality. Within the existing literature, a number of studies have supported the PHH, while others have supported the PHHH. In between, only a few studies have supported either, while others have mentioned no significant effect between these two. Employing the ARDL bounds testing and VECM Granger causality technique, Salahuddin et al. (2018) revealed that the economy’s expansion, increased electricity use, and FDI in Kuwait caused CO₂ emissions. Similarly, Lee (2009) showed that FDI induced CO₂ emissions in Malaysia. Gokmenoglu and Taspinar (2016) noted a two-way affiliation between these two in Turkey. Further, the existing findings in the case of Ghana evidence and support the PHH hypothesis (Solarin et al., 2017). The study used the ARDL method from 1980 to 2012, revealing a beneficial and substantial association between FDI and CO₂ emissions. A. Q. Khan et al. (2018) discovered that the usage of energy, degree of openness, FDI, and socio-economic–political internationalization substantially and positively impacted CO₂ emissions in Pakistan. However, urbanization and economic expansion considerably and negatively impact CO₂ emissions.

Using dynamic panel regressions on selected developed and developing nations, Singhania and Saini (2021) mentioned that FDI inflows significantly impacted environmental quality and evidenced a pollution haven hypothesis. Similarly, using the STVR and the SUR models for China’s macro-economy from 1980 to 2017, Pu et al. (2020) revealed natural catastrophes occurring with financial and large-industrial growth. Following the earlier study applying the SUR technique, Zangoei et al. (2021) established a substantial and progressive relationship between FDI and emissions in selected emerging countries. Y. Huang et al. (2022) highlighted a substantial and positive association between FDI inflows and CO₂ emissions in selected G20 countries. Supporting this, Sreenu (2024) highlighted that FDI inflows are associated with increased CO₂ emissions in BRICS countries. However, the recent research of V. Boamah et al. (2023) revealed that FDI inflows have a two-way effect on CO₂ emissions depending on the nature of the investment in selected South African countries. On the contrary, the findings of Ren et al. (2024) established a positive relationship between FDI inflows and environmental quality. Their study revealed that FDI inflows negatively affected CO₂ emissions in G7 countries. This implied that FDI into developed countries is more environmentally friendly and positively contributes to a sustainable environment. However, the study of M. Khan et al. (2023) argued that the relationship between FDI inflows and CO₂ emissions depends on the level of education. If the economies have a higher level of education after a certain threshold, FDI inflows can reduce CO₂ emissions. In the context of India, Acharyya (2009), applying a cointegration technique, revealed a substantial association between FDI and CO₂ emissions. In addition, Rana and Sharma (2019) identified the existence of the pollution haven hypothesis in India. Their study further highlighted that FDI inflows cause economic prosperity through CO₂ emissions. In contrast to the preceding findings, C. Zhang and Zhou (2016) demonstrated that FDI could reduce CO₂ emissions in China. Similarly, using panel data from 280 selected Chinese cities during 2003–2012, Ning and Wang (2018) revealed that FDI brought positive environmental knowledge externalities and improved China’s environmental quality. Additionally, using the simultaneous equation method, Ayamba et al. (2019) revealed that FDI and environmental quality had a good connection from 1995 to 2016 in China. In the paper, it was argued that FDI brings the advantages of advanced technology and green production.

Further, the concluding findings of a meta-analysis of FDI on environmental pollution highlighted that FDI inflows significantly reduce environmental quality degradation (Demena & Afesorgbor, 2020). However, Kathuria (2018) highlighted that FDI has no role in influencing the emission levels in India; instead, it helps to progress the environmental quality. However, a recent study of Rej et al. (2024) highlighted that the impact of FDI inflows on CO₂ emissions depends on the positive and negative shocks; while the positive shocks have increased the emission levels, the adverse shocks have had no significant contribution to the level of emissions in India. Similarly, using firm-level Indian manufacturing data, Bagchi and Sahu (2024) have suggested that foreign investment can improve the environmental quality in India. Substantially, using the information for selected OECD countries and BRICS countries, Apergis et al. (2023) highlighted that FDI inflows from developed countries to developing countries show a positive contribution to environmental quality. Contrary to this, the study of Kayani et al. (2024) highlighted that FDI adversely impacts carbon dioxide emissions and enhances renewable energy in BRICS economies along with technological advancement. Conversely, Ansari et al. (2019) did not identify any significant linkage between FDI and emissions for selected countries. Instead, their study confirmed that energy usage is the major cause of the high level of CO₂ emissions in those countries.

2.2.2. CO₂ Emissions, Energy Consumption, Industry, and Fossil Fuel Consumption

The existing empirical studies concerning CO₂ emissions and energy consumption are well-explored. Using various econometric techniques such as ARDL cointegration, VECM causal techniques, and Westerlund and Pedroni cointegration estimations, studies include those of Saboori et al. (2014) for OECD countries, Saboori and Sulaiman (2013) for ASEAN countries, Yousefi-Sahzabi et al. (2011) for Iran, and Mirza and Kanwal (2017) for Pakistan. These studies have mentioned that the variables are highly inter-related, and that energy usage is a substantial cause of pollution. Supporting these studies, Mohsin et al. (2022) highlighted that energy consumption significantly contributes to CO₂ emissions in selected European and Central Asian countries. However, Bunnag (2023) did not found any significant relationship between these two in the case of Thailand. Limited studies, such as Tiwari (2011), have empirically verified this nexus in the Indian context and highlighted that energy usage has a favorable and substantial influence on emissions. In contrast, Ouyang and Lin (2015), using the LMDI, highlighted that the industrial sector in China made a substantial contribution to CO₂ emissions while energy intensity helped to reduce the level of CO₂ emissions.

Similarly, recent studies have explored the empirical literature on industrial development and its influence on CO₂ emissions in different countries. Employing other econometrics such as the threshold regression model, ARDL, VECM models, and non-linear ARDL cointegration techniques, Liu et al. (2016) for China demonstrated a favorable and considerable association between industrialization and CO₂ emissions. These studies have argued that with the colossal destruction of resources, high energy consumption, and heavy transportation in production activities, industrialization causes higher CO₂ emissions. At the same time, adopting the STRIPAT framework in a selection of 73 countries revealed that industrialization causes CO₂ emissions (Li & Lin, 2015). Supporting these, Parikh et al. (2009) showed a constructive and noteworthy connection between industrialization and CO₂ emissions. Further, the electricity sector is responsible for India’s CO₂ emissions level. However, Dong et al. (2020) have argued that industrialization reduces CO₂ emissions through upgrading industrial units with advanced technology and new mechanisms. Thus, industrial structures can contribute to carbon mitigation and sustainable development.

Studies have also discovered the empirical connection between the usage of fossil fuels and CO₂ emissions. Using the ARDL test for fifteen developing Asian countries, the empirical result of Hanif et al.’s (2019) paper demonstrated that such consumption causes higher emissions. Supporting these with the help of the ARDL-bound testing technique in Pakistan, they evidenced a long-term link between economic prosperity, usage of fossil fuels, FDI inflows, and CO₂ emissions. They revealed that fossil fuels, FDI, and the development of an economy cause CO₂ emissions. In contrast, employing the Toda–Yamamoto estimation technique for Iran, Lotfalipour et al. (2010) demonstrated that although GDP and energy consumption cause carbon emissions, fossil fuel consumption had no leading role in influencing the emission level.

2.2.3. Industry, Energy Consumption, Fossil Fuel Consumption, and FDI

According to Dinda (2004) and Ghani (2012), FDI has two effects on energy usage. First, FDI induces investment, increasing production in the host country and raising energy consumption. Second, it might expand the production capacity with efficient energy use and enable efficient production technologies. It helps the production units’ transit from dirty or relatively polluted companies to green energy structures. The earlier impact of FDI on energy consumption is well explained as the scale effect. Still, the latter signifies its reverse and implies the composition and technical impact during the production (Zhu et al., 2016). Similarly, many empirical studies with different econometric techniques like the Fourier ADL and ARDL test and the VECM Granger technique, Leitao (2015) in Portugal, Uzar and Eyuboglu (2019) in Turkey, Mavikela and Khobai (2018) in Argentina discovered that FDI provides an encouraging and substantial basis for the higher level of usage of energy. Energy-intensive investors’ investments have shifted from more developed and environmentally stringent countries to less-regulated developing countries. In addition, Saint and Ajmi (2020) supported the existing findings and highlighted that FDI has a favorable and considerable impact on increasing energy consumption for 23 sub-Saharan Africa (SSA) countries. In contrast, Nepal et al. (2021), in the case of India, revealed a harmful link and mentioned that FDI had the significant potential to improve environmental quality by inducing energy efficiency. However, Sadorsky (2010) and Hubler and Keller (2010) highlighted that FDI does not significantly affect energy consumption in developing economies. A recent study of Irfan and Ojha (2023) analyzed how FDI inflows help diversify the energy structure in several countries.

Similarly, using data from 12 franc zone and 11 non-franc zone African countries, Ngouhouo and Ewane (2020) concluded that FDI has a major and beneficial impact on industrial progress within the franc zone. However, the results become undesirable in the non-franc zone. In addition, by applying various cointegration and causality techniques, Adejumo (2020) for Nigeria, and Dash and Parida (2013) for India demonstrated a significant link between FDI and industrialization. The argument represented that foreign direct investment inflows significantly upgrade industrialization and boost the economy’s investment and development. A recent work of Ali et al. (2024) demonstrated that FDI inflows positively contribute to industrialization, and industrialization and foreign investment have an adverse impact on CO₂ emissions. In addition, Zangoei et al. (2021) explained that although countries have tried to find an alternative fuel source to moderate the dependency on fossil energy consumption, fossil fuels are a significant component in attracting FDI with environmental deterioration. Nevertheless, shifting the energy structure away from polluted fossil fuels to green sources remains difficult.

2.2.4. CO₂ Emissions, Economic Growth, and Urbanization

By employing the ARDL and Toda–Yamamoto test, Nkengfack and Fotio (2019) revealed a progressive and substantial energy and economic expansion impact on CO₂ emissions for Algeria, South Africa, and Egypt. Similarly, Jalil and Mahmud (2009) described that economic prosperity causes carbon dioxide emissions in the case of China. In supporting these, a systematic survey by Mardani et al. (2019) also revealed that economic growth has a substantial effect on CO₂ emission levels. However, Mosikari and Eita (2020) have indicated that economic prosperity in selected African economies favorably influences emissions at the early stages until a specific duration. Then, pollution can be controlled with high economic growth (Brahmi & Zouari, 2014). Their paper concluded that the urban population significantly contributes to improving environmental quality. In the recent study of Q. Wang et al. (2023), it was highlighted that the GDP per capita has a significant impact on FDI and CO₂ emissions. Similarly, using Panel ARDL models, Sikder et al. (2022) highlighted that GDP growth is associated with high levels of CO₂ emissions in selected developing countries. Similarly, the findings of Suhrab et al. (2023) stated that GDP makes a positive contribution to emissions. In the Indian context, using econometric techniques like the dynamic multivariate Toda–Yamamoto (TY) technique, ARDL, and VECM causality techniques, Rana and Sharma (2019) supported such a relationship in India.

Many empirical studies have scrutinized the linkages between urbanization and CO₂ emissions (Mohsen et al., 2023). Using various empirical techniques like ARDL, the VECM Granger causality technique, and the FMOLS technique, Al-Mulali and Ozturk (2015) in Europe demonstrated that urbanization is a significant factor in such emissions. Additionally, Franco et al. (2017) showed that urbanization has a greater significance in enhancing the standard of life while fostering economic expansion and significantly affecting the emission levels in India. In addition, Mahmood et al. (2020) supported the empirical results of this nexus for Saudi Arabia. Their paper demonstrated that urbanization positively and significantly contributes to CO₂ emissions. Using the Toda–Yamamoto non-causality methods, Salahuddin et al. (2019) highlighted a bidirectional association amid urbanization and emissions in South Africa. Similarly, using spatial econometric models, G. Zhang et al. (2018) demonstrated that population urbanization has a positive and large spillover effect on pollution in China. In addition, Sikder et al. (2022) showed that high-level CO₂ emissions contribute more to urbanization in selected developing countries (Mohsen et al., 2014). In addition, Suhrab et al. (2023) highlighted that urbanization positively affects CO₂ emissions. On the other hand, Y. Tang et al. (2022) demonstrated a U-shaped relationship between the level of emissions and the degree of urbanization in China. On the contrary, Shahbaz et al. (2016) mentioned that urbanization significantly caused CO₂ emissions and addressed a U-shaped association between urbanization and pollution in Malaysia. However, Singh et al. (2024) highlighted that urbanization has significantly reduced the level of carbon dioxide emissions in India but remains positive in the case of South Africa.

2.2.5. Research Gap and Hypotheses

Though there exists an extensive corpus of writing relating to the PHH and the PHHH hypothesis for the rest of the globe, there are only a few studies on India where the results are mixed and unclear. India is a growing economy, the largest FDI receiver, and is one of the world’s highest emitting countries. To achieve a sustained equilibrium between FDI inflows and India’s CO₂ emissions, the country must prioritize globalization and increase sustainable energy consumption and industrialization. Therefore, the issue requires further investigation. Though the existing studies mostly measure the direct nexus between FDI inflows and CO₂ emissions, there are fewer studies focusing on the indirect effects of FDI inflows2. It is further identified that hardly any existing study has attempted to determine the indirect impact of FDI on CO₂ emissions, considering its mediating channels. Thus, it is essential to also measure their indirect nexus. Though several studies have examined the association between CO₂ emissions, industrialization, energy usage, and fossil fuel consumption, hardly any studies have attempted to find the impact of high-carbon technology based on fossil fuels as the mediating factors of FDI inflows on carbon dioxide emissions in India. In this study, the indirect effect of FDI inflows on CO₂ emissions is considered through three mediating channels to minimize this gap. In this study, the following four hypotheses are presented:

H₁: 

There exists a substantial association between FDI inflows and CO₂ emissions.

H₂: 

FDI inflows, energy structure, industrial structure, and carbon technologies are all linked in some way.

H₃: 

Energy structure, industrial structure, and carbon technologies significantly affect carbon dioxide emissions.

H₄: 

Inflows of FDI have a large mediating effect on CO₂ emissions.

3. Empirical Framework

The data and variables are described in this section. It presents the methodology to demonstrate its relevance to the econometric methods of unit root test, cointegration test, the Granger causality technique, and the unrelated regression estimate to verify the abovementioned hypotheses.

3.1. Data and Variables

In this study, India’s year-wise data are employed from 1980 to 2014. The sample duration is limited to 1980–2014 for two reasons. First, in 1980, India started its initial liberalization before 1990 (Panagariya, 2008). Second, the maximum limit is constrained up to 2014, depending upon the availability of CO₂ emission data in the World Bank data source. To maintain the consistency of all variables, data were collected from the World Development Indicators (WDIs) released by the World Bank. The primary dependent variable in our study is CO₂ emissions. CO₂ emissions are the commonly used indicator among other components to analyze the degree of environmental deterioration due to the long data availability period. Following Bakhsh et al. (2017), in this study, carbon dioxide emissions are considered as a significant dependent variable inour empirical analysis. The primary explanatory variable in our study is FDI inflows. Following Salahuddin et al. (2018),we measure FDI inflows as a portion of GDP in this estimation. The three important mediating variables are the energy structure (Wada et al., 2020), the industrial structure (Mahmood et al., 2020), and the high-carbon technology structure (Ali et al., 2024). The energy structure is represented in terms of energy usage. The industrial structure is represented as a proxy, and the industrial value is added to this study. Following Ali et al. (2024), our research uses fossil fuel consumption as a proxy to measure high-carbon technology.

To ensure our regression estimation results are free from a variable omission or are not affected by other unobservable factors, in our paper, two commonly influential factors of CO₂ emission are controlled, namely economic growth and urbanization (Rustemoglu & Andres, 2016; Mahmood et al., 2020). Similarly, the country’s growth is considered with the GDP per capita and signifies the consequence of economic expansion on CO₂ emissions. Simultaneously, the economy’s urban structure is a proxy of the growth in the urban population. During the estimation, all data are turned into the natural logarithm. In this paper, the CO₂ and FDI show the natural logarithm of carbon dioxide emissions and FDI inflows in their natural logarithmic form, respectively. ES indicates the natural logarithm of the energy structure of the economy, IS indicates the natural log of the industrial structure, and HCT shows the natural log of the high-carbon technology structure. The natural logarithm of gross domestic development (economic development) is captured through GDP, and URBAN shows the natural logarithm of the degree of urbanization in this paper.

Table 1. Description of variables, sources, and expected signs.

Variables	Description	Representation	Source	Signs
CO2	CO2 emission	Carbon dioxide (CO2) emissions (kg per 2010 USD of GDP)	WDI, WB	+/−
FDI	FDI inflows	Foreign Direct Investment, net inflows (% of GDP)	WDI, WB	+/−
ES	Energy structure	Energy use (kg of oil equivalent per capita)	WDI, WB	+/−
IS	Industrial structure	Industry (including construction), value added (constant 2010 USD)	WDI, WB	+/−
HCT	High carbon technology	Fossil fuel energy consumption (% of total energy)	WDI, WB	+/−
GDP	GDP per capita	GDP per capita (Constant 2010 USD)	WDI, WB	+/−
URBAN	Urban structure	The urban population (% of the total population) growth	WDI, WB	+/−

Note: ‘+’ sign and ‘−’ signs indicate positive and negative effects, respectively; the sign (+/−) means that impacts may be positive or negative. Source: authors’ contribution.

lists all variables’ measurements, definitions, sources, and predicted indications.

3.2. Methodology

In this section, the unit root test, cointegration test, causality test, and seemingly unrelated regression equations are included. The unit root test is the first measure to ensure it is stationary before any time-series estimation. The previous research concentrated on the interaction between FDI and CO₂ emissions at the aggregate level, such as at the single country or cross-country level, to verify the equilibrium and causality (Singhania & Saini, 2021; Nguyen et al., 2021) and it has been mentioned that detecting cointegration is the first step. After ascertaining the equilibrium or cointegration relationship, the second phase entails deciding on appropriate Granger causality (Pu et al., 2020). For the Granger causality technique, two strands of tests are available. The first aspect changes the test statistics but still does not distribute. It contains Toda and Yamamoto’s (1995) and Dolado and Lütkepohl’s (1996) modified Granger causality test, a simple method for calculating a Wald statistic by adding VAR delays to the conventional asymptotic distribution. Consequently, the following segment alters the distribution rather than the statistics. In the present study, the Granger causality technique is applied to survey the preliminary causal link age involving FDI and CO₂ emissions. Finally, in this study, the seemingly unrelated regressions (SURs) are applied considering the bootstrap causality and measure the bootstrap samples’ structure.

3.2.1. Unit Root Test

To determine if the time series included in this investigation is stationary, it is prudent to per forma stationary test. In this work, the augmented Dickey–Fuller (ADF) test (1981) and the Phillips–Perron (PP) test (1988) are used to ascertain the stationary features of this research. The ADF test screens for unit roots while allowing residual correlation. Simultaneously, the ADF test addresses more intricate models compared to the Dickey–Fuller test (1979). The investigation incorporates intercept (α) and trend (T) with a lag order of one to identify the unit root issue. The ADF test is predicated on the subsequent equation, as follows:

$$∆Y_t=\alpha +\gamma T+\delta Y_{t-1}+{\beta }_1{∆Y}_{t-1}+\dots +{\beta }_m∆Y_{t-m}+{\mu }_t$$

(1)

where $∆$ denotes a difference operator; ${\mu }_t$ denotes residual term at time t; $Y_t$ denotes time-series variable; and ${∆Y}_{t-1}+\dots +∆Y_{t-m}$ is used to correct the serial correlation. However, the complexity lies in the ADF test as it is prone to produce a high rate of type-1 error if not managed carefully (Naz et al., 2019). We employed the PP test and managed the type-1 error to maintain the protection from this ADF limitation. Utilizing the ADF methodology, Phillips and Perron (1988) introduced a unit root test that addresses the issues of heteroscedasticity and serial correlation. In addition, to remove the serial correlation, it provides rank to the residuals, and the equation of the PP test is as follows:

$$∆Y_t=\alpha +\gamma T+\delta Y_{t-1}+{\mu }_t$$

(2)

The null assumption is dismissed in this instance, indicating that the series is stationary. In this study, automatic lag length selection is utilized in the residuals through the Akaike Information Criterion (AIC).

3.2.2. Cointegration Test

Cointegration is a time-series econometric feature and a prerequisite for any long-term equilibrium relationship among various factors, and in this study, the Johansen cointegration technique of Engle and Granger (1987) is used to assess the long-run link amid FDI inflows and CO₂ emissions. The Johansen (1988) cointegration test will be used in our research to check the collaborative flow between these two variables. This method provides the two tests of trace (J_Trace) and max-Eigen value (J_Max). The trace and the max-Eigen statistical values are used (Johansen, 1988; Johansen & Juselius, 1990). The max-Eigen statistical value test is preferred (K. B. Boamah et al., 2018). The reason is that the max-Eigen test has greater power than the trace-Eigen test (Johansen & Juselius, 1990). Two statistical values of the Johansen cointegration test can be represented mathematically by Equations (3) and (4), as follows:

$$J_{Trace}\left(K\right)=-T\sum _{j=K+1}^m\ln \left(1-{\lambda }_j\right)$$

(3)

$$J_{Max}\left(K+1\right)=-Tln\left(1-{\lambda }_{K+1}\right)$$

(4)

Here, K is the cointegrating vector, while T represents the sample size. The jth most significant canonical correlation is presented in ${\lambda }_j$. The criterion for rejecting the null hypothesis is based on asymptotic critical values at a 5% significance level, mostly accessible through econometric software tools.

3.2.3. Granger Causality

Verifying the enduring relationship link between CO₂ emissions and FDI inflows does not end once the research has confirmed it. The long-term relationship between variables is inadequate for a definitive judgment regarding two-time series (Granger, 1969). Thus, our objective’s next logical step is to check for any causal connection between these two variables when the time series is cointegrated (Rahman & Kashem, 2017). In this study, the Granger (1969) framework is followed to examine this nexus. Following Wen and Dai (2020), we used the Granger causality technique between our variables in an enhanced VAR framework. Granger causality provides the opportunity intended for both unidirectional and bi-directional causality. However, our prime motive is to check whether FDI inflows cause CO₂ emissions in India.

3.2.4. SUR Regression

In this research, the impacts are examined of the mediator variables of FDI inflows on CO₂ emission. For simplicity, in this study, a single-step multiple mediator model is followed for the analysis of the internal mechanism of FDI inflows affecting India’s carbon emission levels. One approach to measuring such an effect is the apparently unrelated regression equations.

The SUR method is a multivariate regression technique. According to Zellner (1962), SUR is helpful when several equations seem irrelevant, but a secret and significant relationship exists between them. The reasons might be as follows: (1) some coefficients should be the same or equal to zero; (2) throughout the whole equations, the error terms or residuals are interconnected; (3) standard independent variables are used (Zellner, 1962). However, Bartels et al. (1996) applied the SUR technique to cross-section units such as household expenditure on gas and electricity. Further, Fernández and Harvey (1990) reported that the SUR model is helpful when dealing with models that are directly modeled in terms of components of interest, such as trends, seasonal, and cycles, known as structural time-series models. This model is widely used in time-series analysis (Engle, 1978; Gersch & Kitagawa, 1983; Harvey & Durbin, 1986). In these studies, the unobserved components in each time series are permitted to be contemporaneously associated with each other to quantify them in multivariate structural models. Extending further, Pu et al. (2020) have addressed that the SUR is favorable for estimating a model when the time-series data have3 fewer variables, and there is a correlation between the error terms of the dependent variables (M. A. Khan et al., 2014). The SUR model is a set of FGLS, considered either a simplified version of the general linear model or a generalization of the simultaneous equation regressions. Thus, this model helps measure the equations’ jointers and the associated covariance. The empirical analysis accounts for the following model specifications:

$${\mathrm{C}\mathrm{O}}_{2,t}=\alpha +c{FDI}_t+u_t$$

(5)

$$M_{i,t}={\beta }_i+a_i{X}_t+u_{i,t}\left(i=1,2,3\right)$$

(6)

$${\mathrm{C}\mathrm{O}}_{2,t}=\gamma +c^{\prime }{lnX}_t+\sum _{i=1}^3{b}_i{M}_{i,t}+{u\prime }_t$$

(7)

where $\alpha$, ${\beta }_i$, $\gamma$ are the intercepts; $u$, $u_i$ and $u^{\prime }$ are residual errors, and time is denoted as t; ${FDI}_t$ is the independent factorand implies FDI inflows; the dependent variable is presented in ${\mathrm{C}\mathrm{O}}_{2, t}$ and shows the CO₂ emissions; $M_i$ comprises the mediator variables where M₁ is the energy structure; M₂ is the industrial structure; and M₃ is the economy’s high-carbon technology. In Equation (5), we include economic growth and urbanization as two control variables. $c$ is the total effect of FDI inflows; $a_{i }$ is the effect of FDI inflows on M_i; b_i is the effect of M_i of CO₂ emissions, after controlling the influence of X; a_ib_i is the specific indirect effect of FDI inflows on CO₂ emissions via M_i. The direct effect of FDI inflows on emissions after controlling the mediators is presented with c′ and ${\sum }_{i=1}^3{b}_i{M}_i$ is the total mediation effect.

4. Empirical Results and Discussion

4.1. Summary Statistics

The summary statistics are presented in the original form of the variables in

Table 2. Descriptive statistics.

Variable	Observation	Mean	Std. Dev.	Min	Max
CO2	35	1.136	0.089	1.01	1.292
FDI	35	0.800	0.89	0.003	3.620
ES	35	413.112	98.091	286.164	636.570
IS	35	257.340	163.739	77.641	611.834
HCT	35	59.252	9.949	39.383	73.577
GDP	35	833.321	356.818	422.904	1640.181
URBAN	35	2.839	0.369	2.329	3.889

Source: authors’ calculation.

. This empirical analysis is based on 35 observations. The average value of carbon dioxide emissions is about 1.136 units, while their maximum value is about 1.292 units, the lowest value is 1.010 units, and the standard deviation (SD) is 0.089 units. The average amount of FDI is 0.800, and the SD is 0.889. Its maximum value is about 3.620 units; the lowest score is 0.002. The average energy structure, industrial structure, and high-carbon technology score are 413.111, 257.341, and 59.252. At the same time, the standard deviation values of these three variables are 98.090, 163.739, and 9.948 units, respectively. The highest and lowest values of the energy structure lie between 636.570 and 286.164. The industrial structure has the highest and lowest values of 611.834 and 77.641, respectively. Similarly, the higher value of HCT is 73.576, and the lowest value stands at 39.383. The average GDP is 833.321, the SD is 356.818, the upper value is 1640.181, and the lowest is 422.904. URBAN’s average size and SD values are 2.839 and 0.369, respectively. At the same time, the highest and lowest values of URBAN are 3.889 and 2.329, respectively.

4.2. Correlation Results

Before the mediating effect analysis, in this study, the dependent variable CO₂ emissions, the prime independent variable FDI inflows with its mediating factors, and the control variables are correlated.

The correlation and covariance results for the FDI inflows and mediating factors are reported in

Table 3. Correlation results of FDI inflows with its mediating factors.

	CO2	FDI	ES	IS	HCT	GDP	URBAN
CO2	1.000
FDI	−0.248 (0.150)	1.000
ES	−0.391 ** (0.020)	0.861 *** (0.000)	1.000
IS	−0.415 ** (0.013)	0.898 *** (0.000)	0.989 *** (0.000)	1.000
HCT	−0.180 (0.301)	0.907 *** (0.000)	0.938 *** (0.000)	0.959 *** (0.000)	1.000
GDP	−0.456 *** (0.006)	0.879 *** (0.000)	0.993 *** (0.000)	0.997 *** (0.000)	0.941 *** (0.000)	1.000
URBAN	0.058 (0.742)	−0.829 *** (0.000)	−0.906 *** (0.000)	−0.900 *** (0.000)	−0.958 *** (0.000)	−0.886 *** (0.000)	1.000

Source: authors’ calculation. Note: ** and *** indicate the significant level at 5% and 1% significance levels, respectively.

. This indicates that FDI negatively correlates with CO₂ emissions; however, its p-values indicate that it is insignificant in affecting CO₂ emissions. The three mediating factors of FDI have also shown a negative association with the dependent variable, where all variables remain significant at a 5% significance level except high-carbon technology. GDP per capita has shown an adverse association with carbon dioxide emissions and remains substantial at a 1% significance level. The urban factor positively correlates with the dependent variable but is insignificant. Based on the findings, it is worth noting that there is a substantial beneficial link between FDI inflows and ES, FDI inflows and IS, and FDI inflows and HCT. Table 3 illustrates a favorable and considerable link between per capita GDP and FDI. However, URBAN presents a negative association with FDI but remains insignificant. The correlation analysis also shows that IS and ES are correlated positively and significantly. Finally, the results noted that IS and HCT have a beneficial and substantial correlation. Similarly, the correlation findings also highlighted that three mediating factors are positively and significantly associated with GDP but possess a negative but significant association with urbanization. The correlation between GDP and URBAN has shown a negative and significant association. It can be noted that all the correlation coefficients of mediating factors are significant at a 1% level with FDI inflows. All three factors show a positive association with FDI inflows, requiring further attention for empirical investigation regarding CO₂ emissions.

4.3. Unit Root Results

The ADF and PP unit root test results with intercept and trend and intercept are shown in

Table 4. ADF and PPunit root test results.

Variables	Level		First Difference
	Intercept	Intercept and Trend	Intercept	Intercept and Trend
ADF Test Statistics
CO2	−0.087 (0.670)	−0.137 (0.674)	−0.959 (0.000)	−1.016 (0.000)
FDI	−0.089 (0.643)	−0.479 (0.123)	−1.400 (0.000)	−1.401 (0.001)
ES	0.028 (0.999)	0.003 (0.995)	−0.602 (0.111)	−0.947 (0.001)
IS	0.004 (0.986)	−0.270 (0.173)	−0.719 (0.003)	−0.734 (0.011)
HCT	−0.054 (0.000)	−0.108 (0.134)	−0.521 (0.052)	−1.008 (0.000)
GDP	0.021 (1.000)	−0.074 (0.916)	−0.884 (0.000)	−1.097 (0.000)
URBAN	−0.076 (0.329)	−0.272 (0.131)	−0.677 (0.003)	−0.709 (0.013)
Phillips–Perron Test Statistics
CO2	−0.087 (0.623)	−0.137 (0.685)	−0.959 (0.000)	−1.016 (0.000)
FDI	−0.089 (0.746)	−0.479 (0.123)	−1.114 (0.000)	−1.114 (0.000)
ES	0.028 (0.999)	0.003 (0.989)	−0.812 (0.000)	−0.947 (0.000)
IS	0.004 (0.984)	−0.186 (0.495)	−0.719 (0.004)	−0.733 (0.016)
HCT	−0.053 (0.000)	−0.108 (0.096)	−0.598 (0.002)	−1.008 (0.003)
GDP	0.021 (1.000)	−0.073 (0.979)	−0.884 (0.000)	−1.097 (0.000)
URBAN	−0.096 (0.117)	−0.246 (0.138)	−0.677 (0.004)	−0.709 (0.018)

Source: authors’ calculation. Note: p-values are in parentheses.

. The ADF results show that all variables remain non-stationary at a level with intercept except HCT. Similarly, none of the variables are stationary at a level with intercept and trend. However, the PP test also reveals that some are non-stationary at levels with intercept except HCT. The PP test shows that the variables are non-stationary at the level with intercept and trend except the HCT. In Table 4, the ADF result exhibits that all the variables become stationary at the first difference with intercept except the ES. However, in both test cases, all variables become stationary in their first difference with intercept and trend.

The ADF and PP unit root test findings give a stable conclusion of the stationary level at the first difference, allowing the cointegration analysis to be conducted in this study.

4.4. Cointegration Test Result

The Johansen test is preferred to identify any long-run, steady, balanced link between CO₂ emissions and FDI inflows in India. The cointegration test results are shown in

Table 5. Cointegration test results.

Test Form (c, t, p)	Statistics Value	5% Critical Value	Null Hypothesis (H0)	Alternative Hypothesis (H1)	Eigenvalue	Hypothesized No. of CE (s)
Trace Statistics
(1, 0, 1)	15.791 **	15.494	h = 0	h $\ne$ 0	0.303	None
(1, 1, 1)	27.156 **	25.872	h = 0	h $\ne$ 0	0.502	None
(0, 0, 1)	21.538 ***	12.321	h = 0	h $\ne$ 0	0.391	None
Max-Eigen statistic
(1, 0, 1)	4.608 **	3.841	h = 0	h $\ne$ 0	0.138	At most 1
(1, 1, 1)	20.229 **	19.387	h = 0	h $\ne$ 0	0.213	None
(0, 0, 1)	6.126 ***	11.224	h = 0	h $\ne$ 0	0.392	None

Note: ‘h’ represents the cointegration rank; ‘c, t, p’, “c = 0” means “no constant”, “c = 1” means “a constant”; “t = 0” means no trend, and “t = 1” means having a trend; “p” represents the “lagged rank”; ***, ** indicates 1% and 5% significance level, respectively. Source: authors’ calculation.

.

The cointegration test outcomes for CO₂ emissions and FDI inflows reveal a fairly long cointegration link between these variables in India’s case. This outcome is in line with the expectations for India of Salahuddin et al. (2018), C. F. Tang and Tan (2015), and Acharyya (2009). This implies that the long-term FDI inflows can affect the level of CO₂ emissions either positively or negatively in India. Thus, policymakers need to understand the causal direction of FDI inflows toward emissions while establishing a regulatory framework to assess FDI approval in India.

4.5. Granger Causality Test Result

Once the cointegration results are obtained, this investigation must examine the causal relationship between FDI inflows and CO₂ emissions. Thus, the next step of our paper is to examine whether FDI inflows cause carbon emissions in India. Our estimation applies the Granger (1969) causality test to explore this relationship for this study’s variables.

Table 6. Granger causality test results.

Lag Order	F Statistics	p-Values	No. of Observations
1	3.47 *	0.072	34
2	2.553 *	0.095	33
3	2.155	0.118	32
4	1.296	0.302	31
5	1.076	0.404	30

Source: authors’ estimation. Note: * presents significant results at a 10% significance level.

shows that in two cases, i.e., lag one and lag two, our results show that FDI inflows positively cause CO₂ emissions. This implies that one unit change in FDI demonstrated a 3.47 unit change in CO₂ emission at the first lag and a 2.553 unit change at the second lag. In other words, FDI has a favorable and consistent effect on CO₂ emission levels in the sample country. This result is consistent with Salahuddin et al. (2018), Acharyya (2009), and Lee (2009) and supports the PHH. These results mean that FDI also influences the increasing carbon dioxide emission levels. Thus, it is important to understand its importance on foreign investment policies for those coming on the direct route. As the findings support that the current FDI inflows are increasing without environmental safeguards, it is high time for the economy to screen the FDI mechanism and assess its contribution to environmental quality through different mediating channels.

4.6. SUR Regression Result

This calculation will examine how FDI inflows affect carbon dioxide emissions with different mediating factors. In this research, the SUR method is followed to verify the multiple effects of FDI inflows.

Table 7. SUR regression results.

Dependent Variables →	Model 1 CO2	Model 2 ES	Model 3 IS	Model 4 HCT	Model 5 CO2
Independent Variables↓
FDI	0.014 * (0.013)	0.107 *** (0.0106)	0.315 *** (0.026)	0.088 *** (0.007)	0.014 ** (0.006)
ES					0.856 *** (0.144)
IS					−0.716 *** (0.064)
HCT					1.246 *** (0.096)
GDP	−0.398 *** (0.040)
URBAN	−0.956 *** (0.113)
R-squared	0.790	0.741	0.807	0.823	0.883
Adj. R-squared	0.770
F-statistics or Chi2-statistic	38.93 ***	100.40 ***	146.49 ***	162.04 ***	264.50 ***
No. of Observations	35	35	35	35	35

Source: authors’ calculation. Notes: the parenthesis values are the robust standard errors. ***, **, and * present the significance level at 1%, 5%, and 10%, respectively.

displays the OLS findings and the results of the SUR model. Model 1 gives the OLS model estimation. Models 2–4 provide the impact of FDI on the mediating variables, and model 5 presents the mediating effect of FDI on CO₂ emissions.

In Table 7, model 1 shows that the total effect on the carbon emission levels is beneficial and consistent. The OLS measurement of FDI demonstrates a0.014 unit change due to one unit change in FDI. This shows that FDI significantly affects carbon emission levels in India due to direct investment from abroad. This result is consistent with Shahbaz et al. (2018) for France. This outcome aligns with Acharyya’s (2009) previous findings and imply the existence of the PHH hypothesis in India’s case. The empirical results demonstrate that one unit change in economic development offers a significant negative effect of −0.398 units on CO₂ emissions. This finding supports the recent literature (Z. Wang et al., 2018). Similarly, the coefficient of urbanization also highlights a significant negative effect of −0.956 units on CO₂ emissions. This outcome supports the existing literature (Yao et al., 2018). This implies that urbanization can improve the quality of the environment with eco-friendly techniques and environmental awareness in urban India. However, the OLS regression is limited to explaining the effect of the mediating channels on CO₂ emissions. Thus, in our study, the SUR estimators are employed.

The SUR technique simultaneously provides the results for four models (models 2–5) in Table 7. Model 2 demonstrates coefficient a₁, i.e., the direct impact of FDI on energy structure is 0.107 at the 1% significance level. Similarly, model 5 can identify coefficient b₁, i.e., the direct effect of energy structure on carbon dioxide is 0.856, which is still considerable at 1%. This is consistent with Saboori et al. (2014) and A. Q. Khan et al. (2018). This implies that energy consumption has significantly pushed the emissions line upward. Since a₁ and b₁ are substantial, the indirect effect through the energy structure is crucial, a₁b₁ = 0.092. This implies that FDI inflows use higher energy in terms of kg of oil use. It can be assumed that this might burden India’s transportation sector heavily4. This highly energy-intensive sector has become one of India’s significant contributors to CO₂ emissions. Therefore, India’s government must take immediate measures to reduce the energy intensity in oil consumption and substitute with renewable energy or less intensive resources. It needs high investment in energy-consuming sectors and demands serious action to build an unfavorable connection between energy resources and environmental quality through FDI. This is possible with the advancement of adaptive, less energy-intensive technologies.

Similarly, model 3 represents the coefficient a_2, which is 0.315, where a₂ presents the direct effect and shows that FDI makesan important contribution to increasing the industrial structure by increasing the production value. This result supports the earlier findings of Adejumo (2020), Dash and Parida (2013), and Gui-Diby and Renard (2015), which emphasized the direct and upward relationship between FDI inflows and industrial development. In other words, these papers claimed that FDI boosts industrialization and modernizes the economy with new investment and growth. Similarly, model 5 shows that the coefficient of b₂ is significant and has a negative value (−0.716), which indicates that CO₂ emissions have been adversely affected by the industrial structure. These findings are supported by Lin et al. (2017). This indicates that with environmentally friendly and energy-efficient technologies and skilled workers, industries could reduce environmental degradation and help countries sustain their economies. Thus, the mediating effect of industrial structure is substantial, with a value of a₂b₂ = −0.226. This illustrates that while FDI positively influences the industrial structure, the industrial structure harms CO₂ emissions. The implication is that FDI inflows increase the value-added of industries by introducing innovative technologies and environmentally friendly methods; thus, the industrial sector improves India’s environmental quality. The direct impact of FDI on the third mediating factor (high-carbon technology) coefficient values is shown in model 4. The coefficient value demonstrates that a₃ is 0.088, which becomes statistically relevant at the 1% level. This means that an uptick in the inflow of FDI would be beneficial. This implies that an increase in FDI inflows significantly affects the promotion of high-carbon technology. This result is consistent with the earlier empirical findings of Zangoei et al. (2021). This demonstrates that developing countries use more fossil fuel energy to generate economic growth from foreign investment.

At the same time, the mediating factor coefficient value b₃ is 1.250 and remains significant at a 1% level. The interpretation indicates that the rise in consumption of fossil resources (Brahmi & Luigi, 2022) as a percentage of total energy can significantly increase India’s carbon emission levels. This finding follows Mert and Bölük (2016) and Ahmad et al. (2016). The findings in these studies revealed that economic development, fossil fuel, and CO₂ emissions have a positive and strong correlation. Thus, the high-carbon technology indirect impact is substantial, where a₃b₃ = 0.110. From this result, it can be inferred that India’s FDI inflows favorably influence fossil fuel consumption and the adoption of high-carbon technology. Thus, investment through FDI in fossil fuel-intensive units generates rising carbon emissions in India. These results can be verified with the previous findings of Pao and Tsai (2011), who suggested that nations with fossil energy-oriented FDI inflows should shift their paradigm to more eco-friendly investments and control the level of emissions without compromising their economic growth. Along with the significant three mediating coefficients, our SUR model presents a substantial direct influence of FDI inflows on CO₂ emissions, c′ = 0.014 in model 5. This indicates that in addition to the three mediating variables, other factors may impact this mechanism, where FDI can influence India’s carbon dioxide emission levels.

Table 8. Bootstrap test results.

Mediation Effect	Observed Coefficient	95% Confidence Interval	95% Biased Confidence Interval	Conclusion
a1b1	0.092	(0.051, 0.133)	(0.054, 0.134)	Significant
a2b2	−0.226	(−0.302, −0.150)	(−0.308, −0.165)	Significant
a3b3	0.110	(0.083, 0.138)	(0.083, 0.137)	Significant
a1b1 + a2b2 + a3b3	−0.024	(−0049, 0.001)	(−0.053, −0.005)	Significant

Source: authors’ calculation.

presents the mediating factors that significantly affect carbon emission levels. The bootstrap test results show that the individual and the combined indirect impact of FDI inflows on CO₂ emissions are substantial. While the individual mediating effect of the energy structure and the high carbon technology positively impacts CO₂ emissions, the industrial structure negatively affects India’s emission levels. However, the combined mediating effect shows an undesirable consequence of FDI inflows on CO₂ emissions. This suggests that FDI inflows in India are not directly deteriorating the environmental quality offered as the OLS estimators in model 1. Instead, different mediating channels help to improve the environmental quality. This finding is in keeping with Kong’s (2021) past research. This implies a positive and sustainable economy with high FDI inflows and improved environmental quality.

Finally, the findings demonstrate that the total mediation effect of FDI inflows is significant and accounts for the overall proportion of carbon dioxide in the Indian economy, which accounts for around 82.8 percent in India; thus, it can visualize the necessity of the proper structural distribution and its utilization of FDI inflows in India. The above analysis can be shown in a diagram (see Figure 2).

This figure (Figure 2) provides the flow directions of FDI inflows influencing CO₂ emissions directly and indirectly. The OLS regression findings are shown in the first part, and the impact on carbon emissions is also presented. Throughout the second part, the figure provides the mediating effect of FDI inflows by energy structure, industrial structure, and high carbon technology on the CO₂ emission level of India.

In summary, in this study, an indirect effect of FDI inflows on CO₂ emissions is highlighted after verifying its cointegration and unidirectional causal relationship from foreign direct investment inflows to carbon dioxide emissions in India. The novelty of this study lies in its investigation of the mediating effects of FDI inflows on emissions. With the apparently unrelated regression application, our study enables us to find the relationship between FDI and CO₂ emissions, energy structure, industrial structure, fossil fuel consumption, and the effect of these three mediating channels of FDI on CO₂ emissions. The findings also demonstrate that the two control variables in the first model are negative and significant. This implies that economic growth and urbanization have reached a threshold, and environmental awareness and energy-efficient technology have helped to reduce CO₂ emissions. This result follows those of Z. Wang et al. (2018) and Yao et al. (2018). These findings have several theoretical and practical implications concerning future studies and policy suggestions and are further elucidated in the next section.

5. Conclusions and Policy Implication

In this paper, research is covered on FDI inflows and environmental quality in India. The analysis focused on the FDI inflow’s direct and indirect impact on CO₂ emissions in the presence of three mediating channels. In this research, extensive time-series data are accounted for over a 35-year period. In this study, the ADF and the PP unit root test, the Johansen cointegration test, the Granger causality test, and the SUR method are used for the empirical analysis. After giving a brief of variables with descriptive statistics, our research found that FDI inflows and CO₂ emissions are stationary at the same integer. Our research employed the Johansen cointegration test, which shows that these two have a long-term association. Then, this research applied the Granger causality and evidenced a one-way causation from FDI to CO₂ emissions. In this study, the effect is estimated using the SUR technique to examine the mediating effects of FDI inflows. As per the SUR estimates, the mediating effect of FDI inflows on India’s carbon dioxide emissions is greater than the direct impact. The findings also show that foreign investors who invest through FDI in India are highly energy-intensive and consume significant amounts of oil. At the same time, our results evidenced that foreign investment in India is significantly fossil fuel-intensive and produces more carbon emissions with high-carbon technologies. This result supports the pollution haven theory in the internationalization and environmental quality nexus. However, the industrial structure’s mediating influence demonstrates an inverse link between FDI inflows and CO₂ emissions. This result contrasts the energy structure and the carbon technology, supporting the recent environmental and internationalization nexus mentioned in the pollution halo effect. Therefore, the impact of FDI inflows on CO₂ emissions depends on redistributing investments across the structures. Thus, in this study, several theoretical and practical implications are identified that can be used as a road map for researchers, industrial agents, policymakers, and government.

In this paper, the debate is highlighted between the “pollution haven” and “pollution halo” effects by synthesizing the question of determining whether FDI inflows are favorable or unfavorable for the quality of the environment. From the cointegration findings, it is suggested that as FDI inflows and CO₂ emissions have long-term associations in India, policymakers must prepare a regulatory framework that simultaneously considers both the environmental impact and the FDI project approval decisions. Then, the policymakers should consider the unidirectional causation of FDI inflows to CO₂ emissions. In this situation, the government and the policymakers should encourage the “green FDI” and investors to switch to advanced environmentally friendly technologies and practices with a high potential for reducing CO₂ emissions in India. The findings of this estimation control the mediating effects and explain that part of FDI inflows is invested in the energy-intensive and fossil fuel sectors. This further contributes to emissions and might provide polluting industries with a less environmentally friendly approach for investing in India. This indicates a ‘capital flight’ of polluted units from a highly regulated and high-production cost economy to India. However, the results also revealed that FDI in industrial sectors significantly boosts the industrial outcome. At the same time, the SUR estimates revealed that investment in the industrial sector is energy-efficient (Esposito et al., 2023) and helps control emissions. It is further noted that the net SUR bootstrap result shows that FDI inflows create a “pollution halo” in India after comprising the adverse effect of FDI inflows on the environment. Therefore, our findings suggest that studies should control the heterogeneous consequences of FDI inflows and its adverse outcomes on the sustainable environment. In this paper, some practical implications are also suggested for the industrial units. As FDI inflows in the industrial sector help to create a sustainable environment, industrial entities can increase their competency and attract more FDI. At the same time, the industrial sector can boost the economy for more output and employment sustainably. Considering its priorities, the Indian government and reputed industrial CEOs should encourage programs like “Mission Innovation (MI)” to formulate a new and more “transformative innovation policy (TIP)”, as mentioned by Carayannis et al. (2022) in the case of the US in 2015. This program aims to make clean energy accessible and generate more sustainable jobs and investment opportunities for enterprises.

A concrete policy implication of this study is explained in the following section. Being the most prominent global growing country, India needs more foreign investment to grow and develop faster. Thus, similar to Singhania and Saini (2021), our research suggests that government intervention with clear environmental disclosures by establishing stringent regulations is essential for reducing the direct volume of carbon emissions and GHG levels. Therefore, to achieve COP26’s “net zero” emission target by 2070 and attain sustainable development, India needs to encourage the ability to cut emissions and complement the best use of resources. Thus, a promotion of all measures may be encouraged at the macro and micro levels, with tax exemptions for cleaner units. Like Zhu et al. (2016), our research suggests that the government may impose strict legal and non-legal barriers to control various production units’ energy consumption. For example, a carbon tax can be imposed on those production units emitting more carbon dioxide. At the same time, MNCs should follow a strict disclosure index on environmental quality within the host country (Esposito & Brahmi, 2023). This will pressure domestic units to improve their environmental quality and stay more competitive in the world market (RobecoSAM, 2019). Simultaneously, as India is a renewable energy source, it may encourage production units to invest more in using solar, wind, and bio-fuel energy instead of fossil fuels. Thus, India should improve its economic performance and enable further FDI inflows with efficient energy use and the latest eco-friendly technologies to achieve a cleaner and sustainable environment.

The implications of the results have a broader scope for future research. Firstly, research can be undertaken for other countries or at the country level, considering this set of macroeconomic variables. Secondly, studies should investigate other mediating channels of FDI inflows while considering their environmental effects. Next, this study can be extended to verify such nexus at a more disaggregate level such as regional or districts, industries, or firm levels. Finally, the future research can encourage case studies for specific industries to change their competitiveness and competency. Therefore, this research will help researchers and policymakers choose where to invest in FDI inflows and generate more benefits with a good environmental quality.