According to Alshagri et al. [16], energy consumption is the primary way that financial development influences CO2 emissions. This relationship operates via two distinct pillars. The first pillar involves a positive association between financial development and energy use. Financial development stimulates economic growth, thereby driving higher energy demand. The second pillar, known as the technology effect, highlights the role of financial development in encouraging countries to develop and adopt energy-efficient technologies, thereby enhancing environmental quality. Hence, the level of energy consumption fluctuates in response to expansions or contractions in financial development, which then impact emissions [31]. However, financial development does not have a uniform impact on emissions, as the effect varies based on the “shocks” within the financial system.

In the first pillar, positive shocks enhance access to external capital flows and affordable credit. These shocks include fiscal easing, investment promotion policies, and commitments to economic openness. Positive shocks bring opportunities for households and firms to buy more products, boosting energy consumption to support economic growth [32]. In addition, positive shocks to financial development can encourage countries to prioritise highly profitable but environmentally harmful industries [33]. As a result, financial development in developing countries can worsen environmental deterioration [34]. Without strong institutions, such development may fund harmful industries instead of sustainable projects [35]. Moreover, credit expansion reduces financing constraints, boosting economic output and energy consumption and intensifying ecological stress and footprints [36]. Conversely, negative shocks, such as financial crises or strict investment policies, can decrease energy consumption. Under these conditions, businesses tend to reduce their activities, and consumers limit their spending due to concerns over restricted credit access and reduced future income [37], [38].

In the second pillar, financial development plays a critical role in promoting environmentally friendly and green technologies. It can provide a positive shock that efficiently mobilises capital and strengthens the financial capacity of enterprises, which may promote sustainability by facilitating investments in green technology innovation and efficiency [39]. Advanced financial mechanisms, such as green credits, green bonds, and environmental investment funds, have incentivised enterprises to adopt cleaner production methods [40], [41]. However, negative shocks to financial development, such as financial crises, often cause financial institutions to become risk-averse. Under these circumstances, they may focus on low-risk projects at the expense of clean technology investments, which may in turn affect CO₂ emissions [32], [42]. Such projects are critical for reducing emissions, but they are typically associated with higher risks and longer payback periods, which may limit their financial support [43], [44]. As a result, both positive and negative shocks within both channels can significantly affect environmental quality. Figure 2 illustrates the asymmetric effects of financial development on CO₂ emissions through these two distinct pillars.

Figure 2.

The asymmetric impacts of financial development (FINDE) on CO₂ emissions in the literature

In the empirical literature, research consistently demonstrates that financial development has complex and often contradictory impacts on environmental quality, particularly CO₂ emissions. Two primary viewpoints emerge: one highlights the environmental benefits of financial development, and the other emphasises its potential to exacerbate environmental degradation. These mixed outcomes reflect underlying gaps in scholarly understanding of the complex relationship between financial development and emissions.

On the one hand, several studies argue that financial development can improve the environment, primarily by facilitating investments in green technologies and sustainable projects. For instance, Lee et al. [45] find a significant negative association between financial development and CO₂ emissions. They attribute this beneficial outcome to effective financial management that supports cleaner production practices. Similarly, Sheraz et al. [46] provide evidence that financial development significantly reduced emissions among G20 nations from 1986 to 2018. This positive environmental effect often arises from advanced financial instruments, such as green bonds and credits, which are crucial for funding low-carbon initiatives [40], [47]. Collectively, these studies emphasise the “green technology” channel, suggesting that positive shocks to financial development encourage investments in sustainable technologies, thereby curbing emissions.

In contrast, a substantial body of research reveals that financial development can increase environmental degradation, as it generates increased economic activity and promotes reliance on fossil fuels. For example, Habiba et al. [23] observe a significant positive link between financial development and CO2 emissions, which they attribute to economic growth patterns driven by carbon-intensive activities. Similarly, Mukhtarov et al. [34] identify financial development as a driver of increased emissions, emphasising that financial expansion frequently stimulates fossil fuel-based production and consumption. These studies emphasise an “economic growth” channel, where financial development facilitates growth that leads to higher emissions. This outcome is particularly notable for economies that rely on traditional energy resources.

Some scholars take another approach, arguing that the environmental impact of financial development is nonlinear and context-dependent. For example, Zioło et al. [48] suggest that whether financial development promotes the adoption of green technology depends significantly on a country’s emission levels. While financial development is beneficial in high-emission countries, it may have limited or even negative effects in lower-emission contexts. Omer et al. [49] and Prempeh [50] offer similar findings, indicating that financial development yields environmental benefits in certain contexts, such as economies heavily reliant on fossil fuels, but shows limited effectiveness in others. Tran [51], in an analysis of 148 countries, emphasises the need to align financial development strategies with environmentally friendly policies. The study reinforces that this relationship can vary depending on context. Similarly, Tao et al. [52], Xiong et al. [53], Patel et al. [54], and Nguyen et al. [55] indicate that nonlinear and asymmetric relationships are present. Thus, both positive and negative financial shocks can significantly influence environmental outcomes across diverse economic settings.

While existing studies have addressed financial development, they primarily focus on its symmetric and linear effects. As a result, they overlook the possibility that positive and negative financial shocks may cause different environmental outcomes. Additionally, previous studies have not attempted to identify these asymmetric effects in both the short and long term. This study addresses these gaps by applying the NARDL model to disaggregate financial development into positive and negative changes, allowing for a more precise assessment of its impacts. Furthermore, although many cross-country analyses have explored the link between financial development and emissions, limited research has addressed the specific context of emerging economies such as Vietnam. This study contributes by providing empirical evidence and policy insights relevant to countries undergoing rapid financial and structural changes, as these places are often overlooked in broader comparative research.

Energy is central to producing, delivering, and consuming essential goods [1], [56], but it predominantly comes from fossil fuels [4], [5]. Although energy consumption is vital for economic development, scholars debate its role in environmental degradation. Indeed, previous studies consistently show that energy consumption is a major driver of ecological degradation, as observed in various contexts, such as groups of nations [24], China [57], and the United States [58]. When energy from fossil fuels is consumed for industrial production, transportation, electricity generation, and household use, it releases CO₂ and other emissions into the atmosphere. These emissions contribute to global warming and climate change, particularly in low-developed and developing economies [24], [59]. However, attempts to curtail energy supply can decrease economic performance in various sectors, as noted by Yιldιrιm et al. [60].

Many researchers view transitioning from fossil-based energy sources to renewables as a promising strategy for mitigating environmental damage [9], [10]. Even so, empirical evidence on the effectiveness of renewable energy remains mixed. Some researchers have identified unexpected negative environmental impacts associated with renewable energy use, such as hydropower-induced emissions [61], negative impacts of wind energy on ecosystems and land use [62], [63], and unforeseen ecological effects of geothermal energy [64]. These findings underscore a critical problem: renewable energy consumption may not yield uniformly positive environmental outcomes and can sometimes create unintended ecological consequences.

According to recent evidence, asymmetric and nonlinear effects can increase the complexity of energy’s environmental impact. For instance, Azam et al. [65] highlight the nonlinear influence of diverse energy sources – including nuclear, natural gas, and renewables. Their study demonstrates that renewable energy consumption can effectively reduce CO2 emissions without compromising economic growth. Similarly, Majeed et al. [26] assert that negative shocks (i.e., reductions) in fossil fuel consumption significantly reduce the ecological footprints of countries. Their study emphasises the importance of examining asymmetries in energy-environment relationships. Further reinforcing this perspective, Alshehry et al. [66] incorporate asymmetric effects within an ARDL framework. They demonstrate bidirectional causality between energy consumption and emissions, further illustrating that traditional symmetric analyses may overlook crucial dynamics.

Despite valuable contributions from previous research, several important limitations remain. Many studies rely on linear and symmetric models. Such models assume that energy consumption has uniform environmental effects, regardless of whether energy use increases or decreases, and may fail to capture the complexities of the energy-environment relationship. Furthermore, studies often overlook the asymmetric relationships between fossil and renewable energy consumption in developing countries, leading to a lack of specific policies based on a country’s characteristics. This study uses NARDL and focuses on Vietnam, a developing country undergoing a rapid energy transition, to provide empirical evidence on how shifts in the energy mix influence CO₂ emissions under prevailing economic and policy conditions. It is one of the main contributions of this research.

Although the relationship between foreign direct investment and CO₂ emissions has been extensively studied, the empirical evidence remains inconclusive, with findings varying across contexts and levels of analysis. The “pollution havens” hypothesis suggests that foreign direct investment inflows may increase emissions by transferring polluting firms to underdeveloped nations with less stringent rules as a way to maintain profitability [19], [20]. Paradoxically, developing countries may have to maintain weak environmental standards to attract foreign direct investment that promotes economic growth, thereby increasing environmental degradation [67]. A clear example is found in the study of Zhang et al. [25], who used the vector error correction model approach to examine the impact of foreign direct investment on China’s CO₂ emissions. They found that foreign direct investment exacerbated pollution, validating the “pollution havens” hypothesis for China. Aust et al. [68] also identified a negative relationship between foreign direct investment and the probability of achieving a sustainable climate in 44 African countries. Similarly, Faheem et al. [69] used data from 19702018 to confirm the “pollution havens” hypothesis in Malaysia. In Vietnam, Ngoc et al. [70] considered the role of foreign direct investment and urbanisation in shaping CO₂ emissions, and Bui et al. [71] specifically analysed the determinants of nitrous oxide emissions and highlighted the significant role of foreign direct investment. The latter study suggests that while foreign direct investment can drive economic growth and industrialisation, it also has complex implications for environmental quality. Thus, careful management is necessary to mitigate negative environmental impacts.

In contrast, other studies challenge the “pollution havens” narrative or even support the “pollution halo” hypothesis. In the latter hypothesis, scholars argue that foreign direct investment inflows lead host countries to adopt environmentally friendly technologies and management practices that reduce carbon emissions [21], [22], [72]. For instance, Apergis et al. [73] found that foreign direct investment originating from France, Germany, and Italy resulted in lower carbon emissions in BRICS countries between 1993 and 2012. Although foreign direct investment inflows tend to increase CO₂ emissions, Huang et al. [74] imply that the level of governance and environmental regulations can mitigate these negative impacts. The authors demonstrate that this occurred for G20 economies during the 1996–2018 period.

Furthermore, Wang et al. [75] assert that foreign direct investment exists in a nonlinear form when the quality and quantity of foreign direct investment inflows are considered. Their study focused on China during the 2006–2016 period. They found that when the quality and quantity of foreign direct investment were low, the level of pollution increased regardless of changes in science and technology. The results indicate that it is vital to improve a country’s science and technology capabilities so it can fully benefit from foreign direct investment inflows. Recently, Wang et al. [27] estimated the impact of foreign direct investment for 67 countries with different incomes and found that the impact on carbon emissions gradually changes from positive to negative at different GDP levels. This finding aligns with the inverted U-shape hypothesis of foreign direct investment with GDP per capita, which indicates that promoting economic growth can reduce the effect that foreign direct investment has on emissions. The above findings emphasise the importance of considering nonlinear models that effectively capture the complex impacts of foreign direct investment on environmental outcomes.

In summary, the literature review has highlighted significant gaps regarding how financial development, energy consumption, and foreign direct investment influence environmental quality, particularly CO₂ emissions. Existing literature primarily adopts linear or cross-country methodologies, and as a result, it often neglects the nonlinear impacts of positive and negative shocks on these factors. To address these gaps, this study employs a NARDL model to comprehensively investigate the asymmetric relationships between financial development, energy consumption, foreign direct investment, and CO₂ emissions within a unified analytical framework. Specifically, the research hypothesises that financial development, energy consumption, and foreign direct investment each exert asymmetric impacts - either enhancing environmental quality or causing it to deteriorate, depending on the nature and direction of shocks. By focusing on Vietnam, a rapidly developing economy with specific economic and environmental characteristics, this study provides context-specific insights into how financial mechanisms, energy use patterns, and investment dynamics jointly influence emissions through growth and technology channels. Through this integrated nonlinear approach, the research offers significant and robust evidence to support the development of targeted policies for sustainable economic and environmental development.

The data for Vietnam from 1990 to 2021 was collected using variables sourced from the World Development Indicators (WDI) of the World Bank. The dependent variable CO2 emissions per capita m CO 2 , (measured in kilotons), was used to represent environmental degradation. Independent variables included financial development (FINDE), as measured by domestic credit to the private sector (% of GDP); energy use (ENERGY), measured in kilograms of oil equivalent per capita; foreign direct investment inflows (FDIIN), measured in percent of GDP; and economic development, represented by GDP per capita (GDPPC) in United States dollars (USD). This comprehensive dataset made it possible to analyse the interrelationships among these factors over the study period. Table 1 presents a detailed description of the potential variables.

In alignment with prior research on the factors influencing m CO 2 emissions by Arjun et al. [76] and Patel et al. [54], this study models CO₂ emission m CO 2 as a function of the variables under consideration as follows:

In eq. (1) α represents the intercept, and ε indicates the error term at time t.

The NARDL model, introduced by Shin et al. [77], is highly regarded for its capacity to identify nonlinear relationships and long-run equilibria. It has been widely used in macroeconomic research to examine variables such as inflation, economic growth, and labour markets, revealing nonlinear relationships and long-term equilibrium effects. Consequently, it provides valuable insights into economic phenomena. The NARDL methodology was selected for this study for multiple reasons: (i) it incorporates nonlinear asymmetry and cointegration within a unified equation; (ii) it evaluates the impact of positive and negative variations in decomposed variables on the dependent variable; (iii) it is appropriate for small sample sizes; (iv) it provides flexibility by not necessitating that variables be integrated in the same order; and (v) it operates as a dynamic error correction representation, yielding robust empirical results despite constrained sample sizes. Unlike linear approaches, asymmetric NARDL approaches account for both positive and negative impacts of the explanatory variables on dependent variables [29].

To employ NARDL, this study has to pass tests to examine unit root tests; the augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) tests were used for this purpose [78], [79]. To assess nonlinearity, the study utilised the BDS test, which was published in Broock et al. [80]. The logarithmic transformation of eq. (1) accounts for both increases and decreases in financial development, energy consumption, and foreign direct investment while controlling for economic growth. It is represented as follows:

In eq. (2), α represents the intercept, β denotes the coefficients of the variables, and ε indicates the error term at time t. The study separates financial development, energy consumption, and foreign direct investment into two distinct components: positive partial sums and negative partial sums. Based on this separation, four new time series were constructed as follows.

This study expresses the asymmetric effects of financial development as:

and

This study expresses the asymmetric effects of energy consumption as:

and

This study expresses the asymmetric effects of foreign direct investment as:

and

Then, the NARDL framework of equation can be written as:

In eq. (9), the conventional Wald test is used to evaluate the long-run symmetry (ω₂ = ω₃ for FINDE, ω₄ = ω₅ for ENERGY, and ω₄ = ω₅ for FDIIN) and asymmetry (ω₂ ≠ ω₃ for FINDE, ω₄ ≠ ω₅ for ENERGY, and ω₄ ≠ ω₅ for FDIIN). In the long run, positive coefficients (ω₂, ω₄, and ω₆ > 0) show that increases in FINDE, ENERGY, or FDIIN raise carbon emissions, while negative coefficients (ω₃, ω₅, and ω₇ < 0) indicate that their decreases also significantly increase emissions. Therefore, to compute the short-run NARDL coefficients with an error-correcting procedure, this study utilises the subsequent equation:

In eq. (10), &#x02205; represents the error correction term, which also shows the long-run equilibrium speed of adjustment after the shock in the short-run, σ, τ, ∩ represent asymmetry coefficients in short-run and ω represents asymmetry coefficient in long-run. Moreover, the conventional Wald test was used to evaluate the short-run symmetry (σ k + = σ k − for FINDE, τ k + = τ k − for ENERGY, and ∩ k + = ∅ k − for FDIIN) and asymmetry (σ k + ≠ σ k − for FINDE, τ k + ≠ τ k − for ENERGY, and ∩ k + = ∅ k − for FDIIN).

Upon confirming a long-term association, the dynamic multiplier effect was assessed. The effects of asymmetric cumulative dynamic multipliers on m CO 2, coming from a singular unit alteration in ENERG Y t + , ENERG Y t − , FDII N t + , and FDII N t − , can be ascertained by the subsequent methodologies:

In eqs. (11), (12), and (13), h = 0,1,2 ... → ∞, then m h + → β 1 + and m h − → β 2 − for FINDE, m h + → β 3 + and m h + → β 4 + for energy, and m h + → β 5 + and m h + → β 6 + for FDIIN . One should note that β 1 + = σ k + θ k ,β 2 − = σ k − θ k ,β 3 + = τ k + θ k ,β 4 − = τ k − θ k , and β 3 + = ∩ k + θ k,β 4 − = ∩ k − θ k. The asymmetric cumulative estimated dynamic multipliers m make it possible to analyse shocks affecting the system, illustrating the adjustment process both toward and away from a new equilibrium.

Table 2 presents the descriptive statistics that were calculated to assess the performance of the underlying variables from 1990 to 2021. These findings emphasise the variables’ central tendency and variability. GDPPC demonstrates the greatest average value, followed by ENERGY, FINDE, FDIIN, and m CO 2. All variables exhibit negative skewness, with the exception of ENERGY and FDIIN. The data in each series exhibits a normal distribution. In terms of volatility, as measured by standard deviation, m CO 2 exhibits the most volatility, followed by FINDE, GDPPC, ENERGY, and FDIIN. These findings indicate a considerable degree of diversity among the examined variables.

This study used NARDL to examine the nonlinear impacts of financial development, energy consumption, and foreign indirect flows on the environmental conditions in Vietnam. Before assessing the NARDL methodology, it was essential to verify the integration of the time series dataset. According to Ibrahim [81], the cointegration F-statistic becomes invalid when any of the variables is found to be stationary at I(2). Table 3 displays the outcomes of the ADF and PP tests. The statistical results from these tests indicate that all variables demonstrate stationarity at either I(0) or I(1), with no variables exhibiting stationarity at I(2).

Before employing the NARDL model, it is essential to examine the dataset for nonlinearity. To achieve this, the nonlinear approach referred to as BDS, introduced by Broock et al. [80], is employed. Table 4 displays the outcomes of the BDS test, demonstrating that all variables possess nonlinear attributes. Thus, the NARDL model is considered suitable for further analysis.

Table 5 presents the findings of the NARDL-bound cointegration test. The study indicates that the calculated F-statistic (3.661) surpasses the upper bound value (3.5), achieving significance at the 5% level. Thus, the null hypothesis (H0) can be dismissed in support of the alternative hypothesis (H1), which asserts the presence of long-term cointegration among the variables. This finding affirms a long-term cointegration relationship among the analysed variables.

After nonlinearity was confirmed through the BDS test (Table 4) and a long-term relationship was established via the bounds cointegration test (Table 5), the NARDL estimation model was utilised to examine both short-term and long-term dynamics between m CO 2 emissions and the chosen control variables. Table 6 depicts the uneven impacts of various factors on environmental quality.

The model of error correction estimates (ECM), shown in Table 7, examines the short-term nonlinear relationships among the variables. The ECM coefficient was determined to be statistically negative at the 1% level, indicating a stable long-term connection between the variables. This calculation suggests that any short-term imbalances will adjust towards long-term equilibrium at a rate of 176.4%.

The Wald test evaluates the asymmetric effect in the long term [82]. Table 8 illustrates the relevance of asymmetry in the long-run parameters of all determinants of m CO 2 emissions, as evidenced by the rejection of the null hypothesis of a symmetric long-run relationship for FINDE, ENERGY, and FDIIN.

The confirmed asymmetric relationship is further substantiated by the effect of the cumulative asymmetric dynamic multiplier of financial development, energy consumption, and foreign indirect investment on m CO 2 emissions, as illustrated in Figure 3. In this figure, the solid black line denotes that FDIIN, ENERGY, and FINDE are experiencing a positive shock, whereas the dashed black line indicates a negative shock to these variables. The red dotted lines indicate the combinations of dynamic multipliers resulting from negative and positive shocks of deconstructed FDIIN, ENERGY, and FINDE on m CO 2 emissions. The spectrum of positive and negative shocks is indicated on the vertical axis, while the horizontal axis denotes the time period. NARDL multipliers illustrate the dynamic adaptations of variables towards a new equilibrium following positive and negative shocks. The figure incorporates asymmetry adjustments of m CO 2 in response to negative and positive shocks (black lines) and illustrates asymmetric patterns that represent the disparity between negative and positive shocks.

Figure 3.

Dynamic multiplier graphs of FDIIN, ENERGY, and FINDE

Table 9 presents the outcomes of model residual diagnostic tests, encompassing evaluations for autocorrelation, heteroscedasticity, Ramsey RESET, and normality. The tests indicate that the null hypotheses regarding autocorrelation, heteroscedasticity, model stability, and normality are not rejectable.

Additionally, Figure 4 displays the CUSUM and CUSUMSQ tests for the model. These graphs demonstrate the model’s stability, as evidenced by the significant lines falling within the critical boundaries.

Figure 4.

Plots showing cumulative sum (left) and cumulative sum of squares (right)

The analysis in Table 6 reveals that both positive and negative fluctuations in FINDE significantly contribute to environmental degradation by increasing m CO 2 emissions in the long term. To be more precise, a positive shock of 1% in FINDE induces a 0.234% escalation in m CO 2 emissions, whereas an equivalent negative shock yields a 0.408% decline. These responses are statistically significant at the 10% and 1% confidence levels, respectively.

These findings extend the work of Patel et al. [54], who identify a nonlinear negative correlation between financial shocks and m CO 2 emissions. Partly consistent with Qin et al. [32] and Boufateh et al. [83], this study suggests that enhanced financial intermediation increases household and business access to energy-intensive goods, driving emissions upward. Moreover, the positive relationship between financial development and emissions supports the work of Ali et al. [84], who find that expanding financial sectors spur industrial activity that relies on fossil fuels, thus elevating emissions. Given Vietnam’s heavy fossil fuel dependence, financial expansion translates directly into higher emissions, whereas financial contractions might encourage sustainable consumption by restricting access to energy-intensive goods. Similar results have been observed in other Southeast Asian economies, such as Indonesia [85] and Thailand [86]. These findings indicate that decoupling financial growth from environmental harm is a regional challenge that requires adopting sustainable finance, embracing renewable energy, and strengthening environmental regulations [87]. Therefore, policymakers should focus on balancing financial growth with investments in green technologies to promote sustainable economic development.

In the present study, unbalanced correlations were detected between ENERGY and m CO 2 emissions, especially regarding the magnitude and direction of the alterations. Estimates indicate that a 1% reduction in ENERGY leads to an 11.374% rise in m CO 2 emissions, while increases in ENERGY exhibit no statistically significant impact on emissions. This asymmetry can be linked to Vietnam’s heavy reliance on fossil fuels – such as coal, oil, and natural gas – which constituted approximately 73% of the country’s total energy consumption in 2021, according to World Bank statistics. Negative shocks to energy consumption may lead to greater reliance on existing coal-fired power plants or trigger investments in additional high-emission facilities. In contrast, the limited impact of positive energy shocks may reflect that Vietnam is actively diversifying its energy portfolio and enhancing renewable sources, including wind and solar power, to limit emissions growth. The findings are consistent with those of Tang et al. [88], who identified a positive correlation between energy consumption and emissions, although their research did not specifically investigate the impact of negative energy shocks. Similar to the results regarding financial development, these findings on energy consumption mirror challenges in Southeast Asia. For example, energy shortages in countries like Indonesia and the Philippines often result in increased dependence on carbon-intensive fuels. These insights underscore the urgent need for regional policies that promote energy diversification, enhance efficiency, and accelerate the integration of renewables to lower carbon intensity and strengthen energy security.

Positive shocks to FDIIN result in a long-term decrease in m CO 2 emissions in Vietnam, whereas negative shocks to FDIIN do not exhibit statistically significant impacts on emissions. Specifically, a positive FDIIN shock was found to reduce m CO 2 emissions by 0.148%, corroborating the “pollution halo” hypothesis and suggesting that more foreign direct investment enhances environmental quality. This reduction likely occurs because foreign investors introduce cleaner technologies and advanced production processes as a result of research and development efforts. However, these findings contrast with studies by Bui et al. [71], Faheem et al. [69], and Ngoc et al. [70], which highlight the negative environmental impact of foreign direct investment in other contexts. The lack of statistically significant effects associated with negative foreign direct investment shocks in Vietnam could be attributed to the relatively small scale of these shocks, given the steady inflows of foreign investment, of which Vietnam remains a net recipient. Overall, the results underline the role of foreign direct investment in promoting cleaner technologies and improving Vietnam’s environmental sustainability in the long run.

One should note that countries like Malaysia and Thailand have also benefitted from foreign direct investment inflows that introduce cleaner technologies, sustainable practices, and energy-efficient production processes. These innovations have contributed to environmental improvements over time [89]. However, these benefits are contingent on attracting high-quality foreign direct investment targeted at green industries, renewable energy, and eco-friendly infrastructure. The insignificant impact of negative foreign direct investment shocks in Vietnam may reflect the relatively small scale of environmentally harmful investments in the country, a trend also noted in some regional counterparts. This finding underscores the importance of regional efforts to attract environmentally responsible foreign direct investment while implementing strict environmental standards to maximise the long-term sustainability benefits of such inflows. Moreover, these observations strengthen the study’s relevance to Southeast Asia regions.

The GDP per capita in Vietnam exhibits a substantial positive correlation with m CO 2 emissions. A 1% increase in GDP per capita corresponds to a 0.920% increase in m CO 2 emissions, highlighting the detrimental environmental consequences of economic growth. With economic growth, energy-intensive industries such as manufacturing, transportation, and construction typically develop. Such industries frequently depend on fossil fuels like coal, oil, and natural gas for energy. This dependency amplifies energy consumption and emissions over time. Moreover, urbanisation and rising living standards contribute to increased consumption of goods and services, further exacerbating emissions. Without a significant transition to renewable energy sources or the adoption of sustainable practices, prolonged economic growth is likely to continue driving up m CO 2 emissions. These findings emphasise the need for Vietnam to integrate clean energy solutions and environmentally sustainable policies into its economic development strategy to mitigate the environmental consequences of growth.

In the short term, the results in Table 7 indicate that prior levels of m CO 2 emissions exert a considerable influence. Furthermore, financial development significantly influences m CO 2emissions in the short term, as demonstrated by both positive and negative variations in FINDE (represented as ∆FINDE⁺ and ∆FINDE^-). A 1% decline in financial development (∆FINDE^-) and its lagged value correlate with increases in m CO 2 emissions of 0.986% and 1.070%, respectively, signifying a decline in environmental quality. Conversely, a positive change in financial development (∆FINDE) also significantly raises m CO 2 emissions. The positive short-term relationship between financial development and m CO 2 emissions may be attributed to increased lending and investments directed towards carbon-intensive sectors such as manufacturing and energy production, which are integral to Vietnam’s growing economy. However, it is also possible for negative financial shocks or constraints to temporarily exacerbate emissions as firms shift their focus to cost-cutting measures over environmentally sustainable practices.

The asymmetric impacts of ENERGY on m CO 2 emissions in Vietnam reveal a complex pattern in the short term. A positive shock of ENERGY decreases m CO 2 emissions in the preceding period but increases emissions in the current period. However, over the long term, positive shocks of ENERGY show no statistically significant impact on m CO 2 emissions. These fluctuations suggest that energy consumption can both reduce and increase emissions depending on the timing and magnitude of demand changes. This finding suggests a challenge for policymakers, who face uncertainty in predicting the short- and long-term consequences of energy consumption on environmental outcomes.

Similarly, the NARDL model indicates that a negative shock to ENERGY leads to a reduction in m CO 2 emissions in the previous period, likely because the country scales down energy use in response to supply shortages or energy costs. However, these negative shocks are followed by an increase in m CO 2 emissions, and this pattern extends into the long run. The initial reductions may be unsustainable, forcing industries and households to rely on easily accessible but highly polluting energy sources, including coal-fired power facilities, to satisfy increasing energy requirements. This dependence is likely to escalate as Vietnam experiences swift economic expansion and modernisation, which will exert more pressure on the nation’s energy infrastructure.

A positive shift in foreign direct investment inflows significantly reduces m CO 2 emissions in the short term, mirroring the observed long-term impact. Specifically, a 1% increase in FDIIN will result in a 0.138% reduction in m CO 2 emissions, highlighting the environmental benefits of foreign capital. This decrease can be attributed to technology transfer and the adoption of more environmentally friendly practices introduced by foreign firms operating in Vietnam. Additionally, these corporations are frequently subject to stricter environmental regulations and standards for corporate social responsibility in their home countries, which may motivate them to implement greener practices in their Vietnamese operations.

However, this finding does not align with Hoa et al. [90], whose research finds that foreign direct investment has a positive impact on carbon emissions. Differences in methodology and timeframe can explain the contrasting findings. Using a NARDL model, the current study highlights that positive shocks to foreign direct investment reduce m CO 2 emissions, likely reflecting Vietnam’s recent efforts to attract sustainable investments and environmentally friendly projects. In contrast, Hoa et al. [90], employing a basic ordinary least squares regression, find that foreign direct investment inflows increase m CO 2 emissions, potentially because earlier stages of foreign direct investment are often concentrated in polluting industries. In contrast, negative shocks to foreign direct investment inflows do not exert a statistically significant effect on m CO 2 emissions in the short term, possibly due to Vietnam’s steady position as a net recipient of foreign investment.

Similarly, GDPPC in Vietnam demonstrates a positive impact on environmental quality, as can be seen by a 1% increase in GDPPC leading to a significant 2.630% reduction in m CO 2 emissions. This trend may reflect that during periods of rapid economic expansion, the initial focus is often on factors such as developing infrastructure, attracting foreign investments, and modernising industries, all of which contribute to improved production efficiency and a transition to cleaner technologies. Additionally, robust economic growth enables the government to allocate more resources to environmental protection initiatives and enforce stricter emissions regulations. Rising income levels also drive changes in consumer preferences. These changes may result in increased demand for environmentally friendly products and services, further supporting emissions reductions in the short term.

Vietnam’s significant dependence on fossil fuels, including coal, oil, and gas, exacerbates carbon emissions during disruptions in energy consumption, highlighting the structural challenges the country may face in its transition to cleaner energy sources. This study offers valuable insights for Vietnam as it pursues economic restructuring towards sustainability. This process involves optimising the advantages of financial development, foreign investment influx, and energy utilisation while alleviating their detrimental impacts. Policymakers should implement a comprehensive strategy to mitigate the negative environmental effects of financial development, energy consumption, and economic expansion while capitalising on the environmental advantages of foreign direct investment.

First, it is crucial that policymakers integrate environmental considerations into financial sector regulations in order to reduce the detrimental impact of financial development on CO2 emissions. Policymakers could incentivise green lending practices, promote sustainable investment portfolios, and establish a robust framework for green finance initiatives. These measures are particularly important given the study’s finding that positive shocks to financial development can reduce emissions while negative shocks increase them. These results demonstrate that it is possible for sustainable financial growth to align with environmental goals.

Second, accelerating the transition to a diversified and sustainable energy mix is imperative, given the rise in emissions that followed negative energy shocks in this study’s model. This finding suggests that Vietnam should implement energy efficiency measures, expand the adoption of renewable energy sources, and gradually phase out carbon-intensive fossil fuels, particularly coal. Such efforts would enhance energy resilience and render Vietnam’s energy consumption less carbon-intensive.

Third, to fully realise the environmental benefits of foreign direct investment, Vietnam should prioritise attracting foreign investments in clean technologies, renewable energy, and sustainable industries. Because the study supports the “pollution halo” hypothesis, Vietnam could implement targeted policies – such as stricter environmental regulations and incentives for adopting green practices – to encourage foreign firms to contribute more effectively to Vietnam’s sustainability goals.

Finally, while economic growth may temporarily reduce emissions by causing businesses to adopt cleaner technologies, long-term strategies should aim to decouple economic expansion from environmental degradation. This strategy requires promoting eco-friendly industrialisation, incentivising energy-efficient production processes, and fostering a shift toward a more sustainable, service-oriented economy. These measures would ensure that Vietnam’s economic growth aligns with its environmental and sustainability objectives.

The study describes the correlation among financial development, energy consumption, foreign direct investment, and CO₂ emissions in Vietnam. However, it has several significant limitations. First, the focus on a single country makes it difficult to generalise to other contexts. Second, the use of aggregated country-level data may mask regional or sectoral variation. Third, the short-term analysis does not fully reflect long-term trends or structural changes. To address these limitations, future studies should expand the scope to include additional countries, use more detailed data at the regional or sectoral level, and consider the role of policy, technological progress, institutional quality, and other moderating factors. Comparisons with emerging or developing economies could provide valuable lessons on balancing economic development, an energy economy, and environmental protection.