Literature Review

Several paradigms, theories, and instruments exist in terms of socio-economic and environmental indicators to help promote sustainable development and sustainability [16-19]. Major projects and infrastructures have significant environmental and human health impacts [20]. In this sense, green hydrogen energy can play a key role in environmental protection and addressing climate change by reducing carbon emissions. Studies have also highlighted the importance of green hydrogen in transitioning to a carbon-neutral economy [21, 22]. Green hydrogen technology can also act as a catalyst for green innovation and reduce the cost of production to make the technology more affordable and easily accessible on a broader scale. Green hydrogen not only mitigates the effects of climate change but also strengthens energy resilience and supports local industries in several nations, presenting favorable prospects for a low-carbon economy [23, 24].

At a global level, the transition to green hydrogen aligns with sustainable development goals and the achievement of carbon neutrality by shifting away from fossil fuels [24, 25]. In terms of net-zero carbon emissions, green hydrogen derived solely from renewable sources is the only type of fuel that can be classified as zero-emission fuel [26]. Green hydrogen complements technologies such as solar, wind, and electrification. Other energy sources may have higher costs, efficiency, scalability, and suitability for different industries in terms of transportation and heavy manufacturing than green hydrogen in the future, allowing green hydrogen to be more effectively positioned in the broader context of decarbonization. This also reinforces the need to integrate national plans with China’s decarbonization objectives, as it provides a viable solution and sustainable practice toward the green economy goal and carbon neutrality by 2030.

Some pieces of literature have argued that to achieve decarbonization at industrial levels, integrating green investments and finances into different industries can substantially transform the industries towards sustainable production and practices [27]. Das (2023) and Butturi and Gamberini (2022) reported similar outcomes of reduced emissions in the transport sector and decarbonizing industries [28, 29]. Scholars have also highlighted the importance of innovation in green technologies and transitioning to green technologies, including carbon storage and hydrogen-based technologies, in addressing carbon emissions and achieving sustainability [30-32]. Green hydrogen may replace fossil fuels as a feedstock or energy source, significantly reducing emissions where other renewable energy sources might not be feasible.

The role of government policies and infrastructure development cannot be ignored. The current shift and urgency towards decarbonizing energy systems and technologies has prompted world leaders to accelerate investments in green technologies and deploying hydrogen-related infrastructures [33]. As evident from previous research, developing green infrastructure, urbanization policies, and strict environmental regulations and taxation may reduce carbon emissions [34]. Similar policies and strategies, including the promotion of a green financial system and green credit availability targeted at innovation in green technology, have also significantly reduced carbon emissions [35]. In addition, trade policies and regulations related to carbon emissions and technological advancements in trading nations are also recognized as effective mechanisms to control and check global carbon emissions [36]. Investments and incentives for developing green hydrogen production technologies are crucial to support the shift towards a low-carbon economy and decrease the environmental impact of industrial operations [37].

On a broader level, previous studies have also emphasized the correlation between green infrastructure and social services, such as human health and environmental services, offering a more comprehensive view of the advantages of sustainable infrastructure [38]. Although previous scholarly works, such as the study conducted by Anderson and Gough (2021), highlight the significance of green infrastructure in mitigating the effects of climate change, there is a necessity for more focused investigations into the precise role that green hydrogen infrastructures play in broader decarbonization initiatives [39]. This more comprehensive perspective may be included in research that specifically examines green hydrogen infrastructure. Although a significant amount of literature is available on green infrastructure and the adoption of green hydrogen, there is a lack of thorough and empirical research that specifically examines the impact of green hydrogen infrastructures on carbon emissions reduction.

There are several gaps in the literature despite the rising interest in both green hydrogen infrastructures and their impacts on CO₂ emissions. Empirical studies that explicitly investigate the impact of green hydrogen infrastructures on emissions are scarce. Most current research is focused on conventional energy infrastructure projects; however, there is a lack of focus on the potential of green hydrogen to improve sustainability and encourage the transition to clean energy [40]. Similarly, investments and green infrastructure projects, including green hydrogen infrastructures, which China has initiated as part of its decarbonization strategy, need thorough research, as studies in this regard are scarce. This study will contribute research using the DID approach to fill these gaps and determine the impact of green hydrogen infrastructure on CO₂, SO₂, and NOx emissions in different provinces of China.

There is a general need to address the urgent energy and environmental issues that China and other nations face in the twenty-first century, as well as to ignite fundamental shifts, encourage sustainable and equitable growth, and increase integration and collaboration between different regions in China and other countries. In this regard, adopting green hydrogen infrastructures, along with green technology innovation and investment policies, is crucial in global efforts to reduce carbon emissions and transition to a sustainable, low-carbon economy.

Data Collection

The model uses secondary data for the analysis of the model. Data are used from a panel consisting of 31 provinces in China from 2004-2022. We target official websites and open-access online resources for data collection. The data related to hydrogen infrastructure projects is accessed through the databank on the International Energy Agency (IEA) website [11]. Fig. 3 shows the specific provinces in China with operational green hydrogen infrastructure projects. It can be observed that the projects are concentrated in the northern regions but spread across east to west of China. The provinces with operational green hydrogen infrastructure projects include Hebei, Henan, Gansu, Inner Mongolia, Shanghai, Shanxi, and Xinjiang.

Data related to socio-economic indicators at the provincial level for all 31 provinces of China is collected from the official website of the National Bureau of Statistics of China [41]. For national-level data, both NBS China and the World Bank Development Indicators (WDI) database are accessed [42]. As controls for regional heterogeneity, dummy variables are also created. All variables are chosen based on their relevance and literature studies. The details of the variables, including the dependent and independent variables, are provided in Table 1.

Model Specification

The study adopts a difference-in-difference (DID) approach to estimate the model [43]. One of the most often used techniques in the social sciences for determining causal effects in non-experimental contexts is differences-in-differences (DID). In a regression model, DID is typically used as an interaction factor between the time and treatment group dummy variables. For estimations, we utilize China’s carbon action plan and provinces having operational green hydrogen infrastructure as the treatment group, while the rest are the control group. Since the commencement of the 12^th Five-Year Plan period (2011-2015), China has included binding objectives for lowering carbon intensity in the framework of plans for national economic and social growth [44, 45]. According to the State Council Information Office of China, in 2015, the government of China set its nationally determined action objectives of reducing carbon emissions by 2030 [46]. Since then, the government of China has implemented and adopted policies that have focused on achieving carbon neutrality through reduced emissions and addressing climate change. Different provincial governments adopted policies based on local requirements to implement and achieve the set objectives under the government action plan. Among the projects, green hydrogen infrastructure and units were introduced and implemented in different areas. Therefore, it is arguable that although different provinces implemented the policy, the tools and measures used by the provinces were different. Therefore, it is assumed that in this regard, provinces having operational green hydrogen infrastructure projects showed higher commitment (technological advantage) compared with the provinces that did not adopt green hydrogen infrastructures or technologies. By 2015, some provinces had already initiated and completed operational green hydrogen infrastructure projects. Therefore, for the estimates, 2015 is the base year, given some provinces’ operational green hydrogen infrastructure projects. Based on the provinces of China, we divide the provinces into two distinct groups: one exposed to the treatment (Operational Hydrogen Infrastructure Project in 2015) is the treated group, and the second without the treatment (No operational Hydrogen Infrastructure project in 2015) is the control group. For the DID estimations, we estimate the following model:

Fig. 3.

Green Hydrogen Infrastructure Projects in China-2015 (Source: Authors using IEA data with built-in function on Microsoft Office).

Where co2 is the dependent variable and stands for carbon emissions by province, which is the proxy for climate change, we generate two dummies—time and treated. The time dummy is 0 for all years before 2015 and 1 otherwise. Also, the dummy is 0 for the control group and 1 for the treated group (provinces with operational hydrogen infrastructure in 2015). X_k,i is a vector of our control variables, which included heterogeneity and robustness controls in the estimations. ε_i denotes the error term in the model.

For estimations, the study assumes that provinces with operational green hydrogen infrastructure projects emit fewer emissions, including CO₂, SO₂, and NOx. For this purpose, the study uses panel data from 31 provinces of China from 2004-2022. Based on the DID approach, a treatment group and a control group are selected [47]. As a treatment group for the green infrastructure, the study takes 2015 as the year of treatment, where the green hydrogen infrastructure in the treatment groups was operational. At the same time, the rest of the provinces were utilized as the control group, which did not have operational green hydrogen infrastructures by 2015.

Fig. 4.

DID Trends for CO₂, SO₂ and NOx Emissions.

Difference-in-Difference Regression

We run the DID regression based on three emissions: CO₂, SO₂, and NOx. The regression’s outputs are reported in Table 3. Columns (1), (3), and (5) show the estimates without control variables, while Columns (2), (4), and (6) represent the results with the control variables added.

The estimates for all three emissions suggest that the impact of green hydrogen infrastructure is negative and significant after 2015. The negative sign shows that provinces with operational green hydrogen infrastructures after 2015 had negative and significant impacts on CO₂, SO₂, and NOx emissions. The negative impact is highest for CO₂ emissions (-0.56), followed by SO₂ emissions (-0.405), and then NOx emissions (-0.19). These results suggest that China’s policies after 2015, especially in the provinces with operational green hydrogen infrastructure projects, are estimated to significantly reduce CO₂, SO₂, and NOx emissions compared to other provinces. In addition to other renewable energy sources, the transition to green hydrogen infrastructure projects may provide an alternative path to achieving carbon neutrality and sustainable development goals.

Post-estimation for the DID estimations is run using parallel trend plots. Parallel trend plots provide a graphical representation of the trend in the treated and untreated groups before and after the intervention. We plot the DID regression estimates for the overall model based on CO₂, SO₂, and NOx emissions. Fig. 4 shows the trends in CO₂ emissions for all provinces of China from 2004-2022. It can be observed that the outcomes followed a parallel trend before 2015 for the groups. After introducing reforms and operations of green hydrogen projects, we see a change in the trends between the two groups. With 2015 being the year for environmental policy shifts and operational green hydrogen infrastructure projects, it can be observed that there is an overall steep decline in CO₂ emissions after 2015. In particular, in the case of provinces with green hydrogen infrastructure projects, the reduction in CO₂ emissions continued significantly more than the rest after 2015, and by 2017, the emissions were lower than the rest of the provinces. The gap between the two groups widened after 2020, suggesting the impacts of COVID-19 lockdowns on the overall emissions. Xu et al. (2024) also highlighted a similar increase in overall CO₂ emissions after 2020 in China, pointing out a rapid expansion of exports, driven by epidemic prevention materials and “stay-at-home economy” products being the key drivers of the CO₂ emissions [48]. By 2022, there was a slight increase in emissions for provinces without green hydrogen infrastructure, but the trend continued to decrease for provinces with green hydrogen infrastructure projects. Similar parallel trends before 2015 can be observed in the case of SO₂ and NOx emissions, which started to deviate from the parallel trend afterward, suggesting significant impacts of the green hydrogen infrastructures and policy reforms in 2015 and onwards. It can also be observed that the gap between the treated and untreated provinces in the case of SO₂ and NOx emissions increases further after 2020.

Here, it is important to note that the diverse impacts of COVID-19 were felt across the globe because of lockdown measures, which led to the shutdown of factories and disrupted the supply chains. Similarly, China was under lockdown for the first quarter of 2020 and opened afterward. The increase in CO₂, SO₂, and NOx emissions after the lockdowns can be attributed to various factors; one may be due to a sudden increase in demand in other countries that remained locked down while China opened its factories and production units as it was producing and supplying not only general tradable goods but also medical supplies to cope with COVID-19. Another may be due to supply chain disruptions, and recovery trends may have contributed to the sudden emissions.

We also used a parallel trend test to assess this assumption further. The result with a Prob > F = 0.725 indicates insufficient data to refute the parallel-trends’ null hypothesis, and thus the parallel-trends assumption is accepted. Furthermore, the Granger causality test is used to determine if the control or treatment groups alter their behavior in expectation of treatment. We lack sufficient evidence to reject the null hypothesis that there was no behavior change prior to treatment based on the test results (Prob > F = 0.6453). These results, along with our earlier diagnoses, imply that we may have faith in the accuracy of our DID estimations.

Fig. 5.

DID-Regional Trends for CO₂ Emissions.

Robustness and Heterogeneity

To further validate the study’s outcomes, we checked the robustness and heterogeneity of the estimates based on three criteria: 1) We estimate separate models for CO₂, SO₂, and NOx based on provinces in three distinct regions of China, including eastern, central, and western regions. 2) We replace the population with GDP per capita (cgdp) to test the impact and robustness of the estimates. 3) We use robust standard errors for the estimations of each model.

Table 4 presents the estimates for the DID regression for CO₂ based on Eastern, Central, and Western regions of China in Columns (1), (2), and (3), respectively. Estimates suggest a slight disparity in the three regions. All three regions experienced a decline in CO₂ emissions because of the green hydrogen infrastructure and policies after 2015, as shown by the significant and negative signs for all three regions. As a result of the intervention, the decline in CO₂ emissions in the eastern regions was higher, followed by the central and western regions. These outcomes may be attributed to the government’s key focus on the eastern regions, which included the main metropolitan areas with ports and higher population density, where the emissions were higher. Based on these estimates, it can be predicted that favorable policies coupled with green infrastructure projects like green hydrogen infrastructures can play a significant role in achieving decarbonization.

Fig. 5 highlights the trends of the regional DID estimates and provides further validation of the parallel trends before 2015 and deviation afterward due to the introduction of policy reforms and operational green hydrogen infrastructure projects. It is highlighted that post-2015, there is a decline in CO₂ emissions for all three regions, which is predicted to result from green hydrogen infrastructure and policy intervention. However, CO₂ emissions show a rise after 2020 but are still lower, especially in the eastern and central provinces with operational green hydrogen infrastructure, than in the provinces without green hydrogen infrastructure. Previous studies also predicted similar trends [49].

Table 5 shows the DID estimations for the Eastern, Central, and Western regions of China in terms of SO₂ emissions. The estimates suggest the presence of regional heterogeneity in terms of SO₂ emissions. The impact on SO₂ emissions in the provincial operational green hydrogen infrastructure under the action plan of 2015 is highlighted as significant in the eastern region at a 1% level, significant at a 10% level in the case of central regions, while it is insignificant for the western regions of China. This heterogeneity suggests the differences in the geographical areas of China. The eastern region, which is more industrialized and economically developed, indicates a more substantial drop in SO₂ emissions due to improved infrastructure and more complex implementation of environmental regulations [50]. In contrast, the western region’s insignificance may be attributed to its less established infrastructure and weaker concentration of green hydrogen projects compared to the eastern and central areas. Furthermore, discrepancies in regional economic trends and local government initiatives likely contribute to the observed disparities in SO₂ emission reductions [51]. Other factors may include distinctions in socio-economic development, technical level, and energy systems in different regions, as presented in past studies [49, 52].

Following the trends of CO₂ emissions, the parallel trend assumptions for SO₂ emissions are also valid for the eastern and western regions before 2015, as highlighted in Fig. 6. Regional heterogeneity can also be observed in SO₂ emissions between provinces possessing operational green hydrogen projects and those without green hydrogen projects. SO₂ emissions in the treated group of provinces tend to decline at an increased level compared to the untreated group of provinces, evidenced by the increasing gap after 2015. Like CO₂ emissions, the difference is noted to be greater after 2020, with the treated group showing signs of decline, while there is an increase in provinces without green hydrogen infrastructures and an even sharper rise after 2022 for all three regions. It could be argued that the sharp increase may be the result of countries opening after COVID-19 lockdowns, thus increasing demand and production.

Fig. 6.

DID-Regional Trends for SO₂ Emissions.

The regional outcomes based on DID estimates in terms of NOx emissions are presented in Table 6. The regional heterogeneity is evident from the estimates. The results of the DID show that the western regions had a greater and significant decline in the case of NOx emissions, while the central provinces had lesser reductions with significance at a 10% level. For the eastern region, the decline was insignificant, suggesting that policy change and green hydrogen infrastructures did not contribute significantly to NOx emissions in the eastern regions. These findings indicate that western areas may have adopted more specific strategies to decrease NOx emissions, possibly due to increased expenditures on pollution control technologies [53]. However, the insignificant outcome in the eastern region might be due to the saturation of pollution control measures or the challenges in lowering NOx emissions in heavily industrialized and densely inhabited areas [54, 55]. The efficacy of environmental interventions can be greatly influenced by regional policy variances and local government activities, as seen by the disparities in effects observed between different areas.

The regional heterogeneity is also evident in Fig. 7. Parallel trend assumptions are also followed for the three regions. In the case of the Eastern provinces, the blue line representing the untreated group is below the red line representing the treated group, indicating higher emissions in the untreated group. In the case of provinces located in the central and western regions, the treated lines are above the untreated groups, indicating a higher level of emissions, highlighting lower emissions in the untreated group of provinces. It can be observed that after treatment through the action plan of 2015, the NOx emissions in the treated group have declined drastically post-2015, as shown by the widening of the gap in the eastern province while the gap narrows for central and western regions. These outcomes further validate the outcomes of the DID model and provide evidence for the effectiveness of policy intervention and operational green hydrogen infrastructures in reducing emissions and decarbonization.

The outcomes of this study suggest that green hydrogen infrastructure can significantly impact emission reductions in China’s provinces. Devising policies targeted at decarbonization and projects based on renewable energy sources, including green hydrogen infrastructures, may have practical outcomes. China’s government can further enhance investments and cooperation in terms of green technology innovations and cleaner production alternatives using renewable energy sources.

The study has theoretical and empirical implications. This study highlights the importance of green hydrogen infrastructure in the broader context of sustainable development and environmental economics. It validates the concept that innovative approaches to energy may result in significant advantages for the environment, thereby verifying ideas that highlight the interdependent nature of technological innovation and the preservation of the environment. In terms of practical implications for policymakers, the study indicates that specific and targeted investments in green hydrogen projects might result in a notable reduction in adverse emissions, contributing to global efforts to reduce carbon dioxide levels and other emissions. China’s green hydrogen development might serve as a model for other countries experiencing comparable issues in attaining decarbonization and establishing sustainable energy infrastructure. China’s approach to finance, regional cooperation, and policy incentives may apply to other situations, particularly developing states. Moreover, the observed geographical differences suggest that adaptable and unique approaches, including local economic and policy conditions, are crucial for optimizing the impact of green hydrogen projects. By carefully directing investments to locations where they may have a significant impact, these insights can help drive future policy and advance the global objectives for combating climate change and transitioning to cleaner and sustainable energy sources.

Fig. 7.

DID-Regional Trends for NOx Emissions.