The Impact of Green Finance on Energy Transition Under Climate Change

Abstract

In recent years, growing concerns over environmental degradation and deepening awareness of the necessity of sustainable development have propelled green and low-carbon energy transition into a focal issue for both academia and policymakers. By decomposing energy transition into the transformation of energy structure and the upgrading of energy efficiency, this study investigates the impact and mechanisms of green finance on energy transition across 30 provinces (municipalities and autonomous regions) in China, with the exception of Tibet. In addition, the impact of climate change is incorporated into the analytical framework. Empirical results demonstrate that green finance development significantly accelerates energy transition, a conclusion robust to rigorous validation. Analysis of the mechanism shows that green finance promotes energy transition through the facilitation of technological innovation and the upgrade of industrial structures. Moreover, empirical evidence reveals that climate change undermines the promotional influence of sustainable finance on energy system transformation. The magnitude of this suppression varies nonlinearly across provincial jurisdictions with differing energy transition progress. Regional heterogeneity analyses further uncover marked discrepancies in climate–finance interactions, demonstrating amplified effects in coastal economic hubs, underdeveloped western provinces, and regions with mature eco-financial markets. According to these findings, actionable policy suggestions are put forward to strengthen green finance and accelerate energy transition.

1. Introduction

In recent decades, China’s accelerated economic development and industrial transformation have triggered a substantial escalation in energy demand, coinciding with persistent structural complexities within the national energy framework. Confronted with escalating climatic anomalies and planetary temperature increases, the nation has formally instituted carbon peaking and neutrality targets, prompting coordinated governmental initiatives across administrative tiers to implement strategic decarbonization roadmaps and operational protocols. Urgent actions to mitigate overreliance on fossil fuels have become a global priority, intensifying public attention on energy transition. Irfan et al., (2023) define energy transition as the process of shifting energy consumption from traditional fossil fuels to renewable, low-carbon alternatives while improving energy efficiency [1]. Coal has long dominated China’s energy mix, creating structural imbalances that threaten economic development, environmental protection, and energy security. Overdependence on fossil fuels exacerbates climate challenges and exposes the nation to volatility in global energy markets. Additionally, insufficient innovation capacity and technological gaps in the energy sector hinder the development of high-quality clean energy and impede efficiency improvements. As energy constitutes a fundamental pillar of socioeconomic operations, the strategic imperative to advance energy system modernization through structural optimization and efficiency augmentation has emerged as a cornerstone of China’s sustainable transition toward ecological civilization.

Huang et al., (2021) emphasize that China’s energy transition remains nascent, constrained by significant funding gaps. Green finance, a key driver of long-term sustainable growth, plays a crucial part in providing funds for renewable energy projects [2]. As stated in the 2023 Report on Financial Institution Lending released by the People’s Bank of China, by 2023, the amount of green loans had reached CNY 30.08 trillion, a 36.5% year-on-year increase—26.4 percentage points higher than overall loan growth. Green finance eases funding barriers for clean energy projects through targeted capital allocation, market-driven investment incentives, and specialized risk instruments like green insurance, accelerating sustainable energy deployment. This study analyzes the part that green finance plays in energy transition dynamics and assesses regional implementation variances across China’s economically diverse provinces.

Alfonso (2023) highlights that extreme climate events increasingly destabilize energy systems and societal operations [3]. While decarbonization necessitates systemic integration of renewable sources into energy matrices, climate volatility introduces geophysical risks compromising both geological resource potential and supply chain resilience in sustainable power infrastructure. Climate volatility disrupts financial markets, imposing financing constraints on capital-intensive renewable projects with long payback periods. As Corner et al., (2011) note, climate risks amplify uncertainties about zero-carbon technologies’ reliability, deterring investment [4]. Extreme weather also damages energy infrastructure, reducing generation efficiency and disrupting distribution networks. Thus, understanding climate change’s role in the green finance–energy transition nexus offers critical insights for policymakers.

The existing literature often examines energy transition through singular lenses—either energy structure or efficiency—while neglecting holistic analysis. Studies integrating climate change with finance or energy sectors remain largely qualitative, lacking a unified framework. Methodologically, most research employs linear models like OLS, overlooking nonlinear climate impacts across regions at varying transition stages. This study bridges these gaps through comprehensive theoretical and empirical analyses of green finance, climate change, and energy transition.

The article may offer the following contributions: Firstly, it categorizes energy transformation into structural transformation and efficiency upgrading, thereby delving deeper into the intrinsic linkage between green finance and energy transformation. Secondly, this article innovatively integrates the three pivotal domains of green finance, climate change, and energy transformation into a cohesive research framework, broadening the research horizon and elucidating the role of climate change in facilitating energy transformation through green finance with greater clarity. By establishing such a comprehensive framework, it facilitates an in-depth exploration of the intricate interactions among these three entities, offering fresh theoretical perspectives and analytical approaches for subsequent research endeavors. Thirdly, leveraging panel quantile regression models, this article analyzes the nonlinear impacts of climate change in regions at varying stages of energy transformation. The findings provide a reference for accelerating energy transformation and fostering high-quality energy development, while also serving as a valuable source of experience and guidance for proposing practical solutions and financial support policies pertinent to energy transformation.

2. Literature Review

2.1. Energy Transition

Grübler (2004) identifies three hallmarks of energy transition: quantity, structure, and quality [5]. Binder et al., (2017) define it as “a combination of disruptive changes and gradual adaptations along specific pathways [6]”. Scholars widely recognize energy transition as essential for environmental improvement and economic transformation. Khan et al., (2021) argue that phasing out non-renewables while scaling modern renewables fosters long-term sustainability [7]. Zheng (2022) stresses that climate-resilient, low-carbon societies demand accelerated energy transition to achieve China’s 2030/2060 goals [8]. Liu et al., (2022) spatially analyze China’s transition, revealing a “dependency-intensifying, innovation-fluctuating” trajectory [9]. Lee et al., (2024) contextualize China’s energy landscape, defining transition through structural optimization, efficiency gains, and renewable technology advancement [10]. Wang et al., (2024) have noted that sustaining China’s past growth rate of clean energy may fall short of meeting the future clean energy demands necessary for achieving the “dual-carbon” goals. Consequently, to attain these goals, it is imperative to overcome the inertia within the energy and economic systems, expedite China’s energy transition, and elevate the share of clean energy in the overall energy mix [11]. Wang et al., (2024) highlighted that China’s energy transformation is marked by its long-term, complex, and arduous characteristics. In anticipation of the crucial period leading up to carbon peaking in the future, they put forward recommendations including ensuring energy supply security, advancing total quantity control and structural adjustment simultaneously, implementing multiple measures for energy conservation, emission reduction, and carbon sequestration, fostering dual-wheel-driven collaboration between the government and the market, and facilitating opening-up and cooperation to ensure smooth internal and external circulation. These suggestions serve as a reference for scientifically pursuing energy transformation efforts [12].

2.2. Green Finance and Energy Transition

Hosier and Dowd’s (1987) energy ladder theory posits that households transition to cleaner fuels as economies develop [13]. Lee Gi-Hyoung (2010) first conceptualized green finance as financial mechanisms harmonizing economic and environmental goals [14]. Shahbaz et al., (2021) link financial development to renewable growth, identifying green finance as pivotal for sustainable transition [15]. Madaleno et al., (2022) underscore the necessity of green investment for clean energy systems [16]. Empirical studies demonstrate green finance’s role in energy efficiency: Peng et al., (2021) find that it spurs green innovation, particularly in resource-rich, developed, and market-oriented regions [17]. Wang et al., (2022) employ DEA and Tobit models to confirm stronger efficiency gains in eastern China [18]. Li et al., (2023) validate green finance’s efficacy in pilot reform zones, emphasizing digital technologies’ mediating role [19]. For energy structure, Jiang et al., (2022) analyze listed energy firms, showing that green finance compensates for traditional finance’s shortcomings, fosters innovation, and optimizes industrial structures [20]. Sun et al., (2022) link green finance to energy consumption via system GMM, noting geographic heterogeneity [21]. Li et al., (2024) use spatial Durbin models to reveal green finance’s structural and spatial spillover effects [22].

2.3. Financial Markets and Climate Change

Climate change is operationally defined by international organizations (United Nations, IPCC) and China’s meteorological authority through measurable parameters including severe weather occurrences, thermal fluctuations, rainfall distribution, and atmospheric motion intensity. Cashin et al., (2017) find that El Niño elevates oil prices and inflation [23]. Guo (2020) warns that climate risks threaten China’s financial stability and monetary policy [24]. Jiang (2022) delineates macro–micro pathways: climate risks erode institutional balance sheets (macro) and amplify credit/market risks (micro) [25]. Conversely, Dhayal et al., (2024) use ARDL models to show green finance’s long-term climate mitigation potential under supportive policies [26]. Shahbaz et al., (2021) highlight China’s green financing in corporate decarbonization [15].

2.4. Climate Change and Energy Transition

Frauke Urban (2010) highlighted in her research that current climate change is inextricably linked to energy-related activities. China and India, as coal-dependent economies, must urgently transition to clean energy to prevent carbon lock-in and combat climate crises [27]. Lee et al., (2022) developed a composite climatic vulnerability metric leveraging provincial-level data across 30 Chinese administrative units, demonstrating how suboptimal climatic patterns in moderately developed regions create geoeconomic thresholds that constrain the efficacy of renewable technology deployment in facilitating low-carbon system transformation [28]. Zhao Huaitian et al., (2023) innovatively employed visual analytics to comprehensively assess the impact of energy transition on climate change. Their findings demonstrate that the growth of renewable energy (particularly in power generation) significantly benefits climate mitigation. Renewable energy deployment demonstrates significant CO₂ and PM2.5 reductions, directly advancing climate stabilization goals [29].

Current research lacks comprehensive studies examining how climate change acts as a mediator in the relationship between green finance and energy transition, particularly in systematically quantifying their interconnected dynamics. To address this gap, this paper incorporates climate change as a pivotal factor within the green finance framework, employs quantitative methods to investigate its moderating effect on the relationship between green finance and energy transition, and proposes policy recommendations to advance the green and low-carbon energy transition in China.

3. Theoretical Analysis and Research Hypotheses

Theoretically, the early stages of transitioning to low-carbon energy systems necessitate significant financial investments and infrastructural development because of the restricted market size. Shi Qianqian (2023) noted that green finance, unlike traditional finance prone to credit discrimination, offers lower-cost financing and optimized loan conditions for environmentally friendly renewable energy projects, thereby alleviating funding bottlenecks and stabilizing energy transition [30]. Furthermore, under the dual-carbon goals set by China, green finance incentivizes governments to adopt supportive measures such as tax reductions and subsidies, directing capital toward low-pollution industries and optimizing resource allocation for sustainable development (Li et al., 2018) [31]. On this basis, we put forward the following hypothesis:

Hypothesis 1:

Green finance exerts a positive effect on the process of energy transition.

Alharbi et al., (2023) argued that green finance, as an eco-friendly financial instrument supporting sustainable development, indirectly influences energy transition [32]. From a theoretical perspective, the resource-based view (RBV) posits that technological innovation and industrial upgrading constitute core strategic resources that enable firms to achieve competitive advantages in sustainable transitions (Widya et al., 2022) [33]. Within the technological innovation system framework, these two pathways serve as fundamental drivers that integrate financial resources, technological capabilities, and market demands to facilitate systemic energy transitions (Keller, 2022) [34].

First, green finance channels financial resources towards the renewable energy and energy-efficient sectors, easing financing pressures for technological innovation. Hou Shaobo (2021) demonstrated that technological advancements reduce energy intensity and enhance efficiency, accelerating structural shifts in energy consumption [35]. Cisneros (2023) further confirmed that green innovation lowers renewable energy costs and displaces fossil fuels through substitution effects [36]. This aligns with IRENA’s (2017) finding that technological breakthroughs are prerequisites for achieving cost-competitiveness in renewable energy [24].

Second, green finance facilitates industrial restructuring via regulatory mechanisms [10]. Yang Ling et al., (2021) emphasized that advanced financial systems prioritize capital allocation to firms adopting green strategies while restricting loans to fossil-fuel-dependent enterprises [37]. An et al., (2023) empirically verified that industrial structure upgrading serves as a critical mediating channel through which green finance reduces energy intensity [7]. From a consumption perspective, Yao Wang et al., (2016) examined provincial panel data spanning 2006–2021 and discovered that green finance enhances market competitiveness of low-carbon products through preferential policies, coupled with rising consumer demand for eco-friendly goods [38].

While alternative mechanisms exist, they exhibit distinct characteristics from the primary pathways. Policy spillover effects have been documented, where green finance development in one region positively influences energy transitions in neighboring areas through spatial diffusion (Wang et al., 2023) [39]. Consumer behavior also plays a role, as retailer access to green finance positively impacts consumer purchase intentions for green products, particularly among environmentally conscious individuals (Peng et al., 2023) [40]. However, these mechanisms primarily function as contextual moderators rather than direct mediators. The current study focuses on technological innovation and industrial upgrading as primary pathways due to their theoretical centrality in the RBV and innovation system frameworks, consistent measurement availability across contexts, and empirical evidence of their dominant explanatory power in the green finance–energy transition nexus. Thus, we posit the following:

Hypothesis 2:

Technological innovation and industrial upgrading play a mediating role in the relationship between green finance and energy transition.

Climate change intensifies financial and energy system risks, restricting renewable funding and raising green tech costs. This sustains fossil fuels’ cost edge, slowing energy transition. However, Dong Kangyin et al., (2022) observed regional heterogeneity: in areas heavily reliant on traditional energy with underdeveloped infrastructure and low public awareness, climate shocks severely delay transition by restricting fiscal support [41]. Conversely, in regions prioritizing sustainability—where clean technologies, industrial chains, and public acceptance are advanced—climate change exhibits negligible inhibitory effects.

Beyond regional variations, emerging evidence suggests climate change may exert nonlinear impacts on the green finance–energy transition nexus. This nonlinearity arises from threshold effects in climate risk accumulation and adaptive capacity development. Feng and Zhao (2022) demonstrated that environmental policy instruments often exhibit threshold effects, where their effectiveness changes discontinuously after crossing critical levels of implementation intensity [42]. Similarly, Gan and Voda (2022) identified dynamic nonlinear characteristics in green finance’s impact on carbon intensity, with effects strengthening significantly after surpassing certain development thresholds [43].

Theoretical support for nonlinear climate moderation derives from the environmental Kuznets curve literature, which posits that environmental impacts change nonlinearly with economic development. Extending this framework, Xie et al., (2021) documented an inverted N-shaped relationship between energy transition depth and green productivity, indicating that both insufficient and excessive transition paces may yield suboptimal outcomes [44]. Applied to climate moderation, this suggests that climate change impacts might accelerate disproportionately after exceeding ecological resilience thresholds, creating non-symmetric response patterns.

Empirically, Zhang et al., (2022) confirmed threshold effects in green finance’s carbon intensity reduction, showing that capital stock accumulation enhances green finance effectiveness only after reaching critical levels [45]. Transposing this logic, climate change could similarly create threshold-dependent moderation: below certain climate risk levels, green finance mechanisms may adapt adequately through incremental adjustments; however, beyond critical thresholds of climate disruption, systemic risks overwhelm adaptive capacities, causing the green finance–energy transition relationship to deteriorate abruptly. This creates the hypothesized nonlinear negative moderating effect. Accordingly, we propose the following:

Hypothesis 3:

Climate change exerts a nonlinear negative moderating effect on the relationship between green finance and energy transition.

4. Design of Research

4.1. Model Specification

4.1.1. Benchmark Regression Model

Following the methodological framework of Lee et al., (2024), this study employs a two-way fixed effects model to investigate the influence of green finance on energy transition [10]. The regression equation is structured as follows:

$${\mathrm{ET}}_{\mathrm{it}} = {\mathrm{\beta }}_0 + {\mathrm{\beta }}_1 {\mathrm{GFI}}_{\mathrm{it}} + {\sum }_{k=2}^6{\beta }_k{Control}_{it}+ {\mu }_i+ {\phi }_t+ {\epsilon }_{it}$$

(1)

Among them, i stands for the province and t stands for time. The response variable ET_it represents energy transition, which includes the conversion of energy structure (ESC) and the improvement of energy efficiency (EEI). The explanatory variable GFI_it represents green finance. Control_it is used to represent a series of control variables that might affect energy transition, including the level of economic progress (P-GDP), overseas direct investment (FDI), urban development (Urban), financial infrastructure (FI), and government backing (Gov). In addition, the model incorporates provincial fixed effects ${\mu }_i$ and time fixed effects ${\phi }_t$, β₀ stands for the constant term, and ε stands for the error term.

4.1.2. Mediation Effect Model

From the theoretical analysis of Hypothesis 2, it can be seen that there are two potential ways for green finance to facilitate energy transition, namely promoting technological innovation and upgrading the industrial structure. Consequently, this study constructs the following mediation effect model to examine the intermediate transmission mechanism.

$${\mathrm{Med}}_{\mathrm{it}} = {\mathrm{\alpha }}_0 + {\mathrm{\alpha }}_1 {\mathrm{GFI}}_{\mathrm{it}} + {\sum }_{k=2}^6{\alpha }_k{Control}_{it}+ {\mu }_i+ {\phi }_t+ {\epsilon }_{it}$$

(2)

In Model (2), Med represents the mediating variable, which includes technological innovation (Tech) and the upgrading of the industrial structure (Indus). α₀ is the intercept, representing the expected value of the mediator (Med) when GFI and control variables are zero. Among the coefficients of the other variables, our focus is on α₁, which is the coefficient of GFI, indicating the effect of a one-unit increase in GFI on Med, holding control variables constant. The definitions of the other variables remain consistent with those in Model (1), and will not be repeated here.

$${\mathrm{ET}}_{\mathrm{it}} = {\mathrm{\gamma }}_0 + {\mathrm{\gamma }}_1 {\mathrm{GFI}}_{\mathrm{it}} +{\mathrm{\gamma }}_2 {\mathrm{Med}}_{\mathrm{it}} + {\sum }_{k=3}^7{\gamma }_k{Control}_{it}+ {\mu }_i+ {\phi }_t+ {\epsilon }_{it}$$

(3)

In Model (3), γ₀ represents the intercept, capturing the baseline level of ET when other variables are zero. γ₁ measures the effect of green finance (GFI) on ET and γ₂ captures the effect of the moderator variable (Med) on ET.

4.1.3. Moderation Effect Model

In order to accurately investigate the moderating impact that climate change exerts on the connection linking green finance and energy transition, this study refers to the research methods of Kazaz and Istil, introduces the variable of sunshine hours (Sun), and constructs an interaction variable combining green finance and sunshine hours (GFI × Sun) to examine the moderating effect of climate change [46]. The model is expressed as follows:

$${\mathrm{ET}}_{\mathrm{it}} = {\delta }_0 + {\delta }_1 {\mathrm{GFI}}_{\mathrm{it}}+ {\delta }_2 {\mathrm{Sun}}_{\mathrm{it}}+ {\delta }_3 {\mathrm{GFI}}_{\mathrm{it}} \times {\mathrm{Sun}}_{\mathrm{it}} + {\sum }_{k=4}^8{\delta }_k{Control}_{it}+ {\mu }_i+ {\phi }_t+ {\epsilon }_{it}$$

(4)

The above model focuses on the parameter δ₃ of the interaction variable. If δ₃ is significantly negative, it suggests that climate change exerts an inhibitory influence on the process of green finance promoting energy transition.

4.1.4. Quantile Regression Model for Panel Data

Recognizing the potentially nonlinear regulatory effects of climate change, we develop a panel quantile regression framework to systematically explore the heterogeneous manifestations of climate impacts across energy transition phases. The methodological design enables precise identification of developmental thresholds and dynamic interaction patterns, thereby informing differentiated governance frameworks that align with province-specific energy restructuring trajectories.

$$Q_{{ET}_{it}}^{\left({\tau }_q\right)} = {\theta }_1({\tau }_q) {ET}_{it}+ {\theta }_2({\tau }_q) {Sun}_{it}+ {\theta }_3({\tau }_q) {\mathrm{GFI}}_{\mathrm{it}}\times {\mathrm{Sun}}_{\mathrm{it}}+{\sum }_{k=4}^8{\theta }_4\left({\tau }_q\right) {Control}_{it}+ {\mu }_i+ {\phi }_t+ {\epsilon }_{it}$$

(5)

Model (5) stands for the panel quantile regression model considering fixed effects. Here, τ represents the quantile, where $\tau$ ∈ (0, 1), and θ₃ represents the moderating effect of climate change. Quantile regression extends beyond conventional OLS methodology by enabling the estimation of covariate effects across conditional quantiles, rather than restricting analysis to the conditional mean relationship. By comparing the regression results at different quantiles, it can provide a solid basis for formulating more-targeted policies.

4.2. Selection of Variables

4.2.1. Response and Explanatory Variables

In this study, the response variable is energy transition (ET). Apergis et al., (2012) pointed out that the focus of China’s promotion of energy transition and the execution of the green and low-carbon circular economic development system mainly cover the following two aspects: first, paying great attention to energy saving and vigorously improving energy utilization efficiency; moreover, comprehensively developing sustainable energy and promoting supply-side energy structure reform [47]. Therefore, this study draws on the experience of Liang Peng (2023) and conducts a comprehensive analysis of energy transition from the two dimensions of structural transformation and efficiency upgrade [48]. First is the conversion of energy structure (ESC). The use of fossil fuels such as coal serves as the primary contributor to carbon emissions. Meeting decarbonization commitments requires systemic energy restructuring, prioritizing renewable-centric infrastructure development while implementing phased reductions in fossil fuel consumption shares. This dual-track approach ensures structural alignment with climate governance objectives. Given this, this paper selects the share of clean energy usage in the aggregate energy consumption as the proxy variable for the conversion of energy structure. The second aspect is the improvement of energy efficiency (EEI). At present, China has not made a breakthrough in the field of massive and cost-effective energy storage technology. At the same time, non-fossil energy shows intermittent and volatile characteristics in terms of power supply. While advancing clean energy transitions, coal remains critical for ensuring a stable electricity supply through upgraded combustion technologies and emission controls. This paper uses the proportion of GDP in relation to total energy consumption (i.e., the reciprocal of energy intensity) as a measure of energy efficiency.

In this paper, green finance (GFI) serves as the independent variable. As a pivotal institutional innovation in sustainable financial systems, green finance demonstrates transformative potential in China’s energy transition. Empirical studies confirm its dual capacity to enhance energy productivity by 18–22% while accelerating structural decarbonization at a 2.3% annual rate, thereby fortifying national energy resilience. Post−2016 regulatory frameworks have catalyzed extensive academic explorations employing policy evaluation models and system dynamics approaches. Initially, green finance was mainly measured by individual indicators like green credit, low-carbon capital flow, and so on. In order to address the drawbacks of conventional research, later scholars went on to build a multi-dimensional comprehensive assessment system for green finance. This paper draws lessons from Han Zhongxue et al., (2024). The entropy method is employed to develop a green finance index that encompasses four aspects—green credit, green insurance, green investment, and green governance—in order to more comprehensively measure the development level of green finance in each province [49].

Specifically, we processed these four indicators as follows:

• Data Normalization. Min-max normalization was employed to eliminate dimensional differences among indicators, transforming all values to the [0, 1] interval:

  $$x_{ij}^{\prime }={\displaystyle \frac{x_{ij}-min\left(x_j\right)}{\max \left(x_j\right)-\min \left(x_j\right)}}$$

  (6)

  where $x_{ij}$ represents the original value of indicator j in province i. This method was chosen for its ability to preserve data variability while ensuring comparability, consistent with recommendations by Pan et al., (2022) [50] and Wang et al., (2023) [51] for entropy-based index construction.
• Calculate Indicator Proportions.

  $$p_{ij}={\displaystyle \frac{x_{ij}^{\prime }}{{\sum }_{i=1}^n{x}_{ij}^{\prime }}}$$

  (7)

  where p_ij is the normalized proportion of the j-th indicator for the i-th object, reflecting its relative weight within that indicator across all samples.
• Compute Information Entropy.

  $$e_j=-k\sum _{i=1}^n{p}_{ij}$$

  (8)

  where e_j denotes the entropy value of the j-th indicator, measuring the degree of dispersion; a lower e_j indicates greater variability and more useful information. k = ${\displaystyle \frac{1}{ln\left(n\right)}}$ and n is the number of provinces.
• Determine Weight Coefficients.

  $$g_j=1-e_j$$

  (9)

  $$w_j={\displaystyle \frac{g_j}{{\sum }_{j=1}^m{g}_j}}$$

  (10)

  where $g_j$ represents the degree of difference for indicator j, and $w_j$ is the final weight. This approach minimizes subjective bias by leveraging the information entropy of indicator values, as demonstrated in Jiang et al., (2020) [52] and Pang et al., (2023) [53].
• Index Aggregation. The final green finance index was constructed as a weighted sum:

  $${GFI}_i=\sum _{j=1}^m{w}_j{x}_{ij}^{\prime }$$

  (11)

  The specific meanings of each dimension are presented in

  Table 1. Dimensions of green finance.

  Indicator	Definition	Unit
  Green Credit	The ratio of interest expenses in six high-energy-consuming industries to industrial industry interest expenses.	%
  Green Insurance	The ratio of agricultural insurance revenue to the overall agricultural production value.	%
  Green Investment	The ratio of environmental pollution control investment to GDP.	%
  Green Governance	The ratio of government environmental protection expenditure to total government expenditure.	%

  .

4.2.2. Control Variables

Drawing on existing relevant papers, the control variables chosen in this research include the level of regional economic progress (P-GDP), financial infrastructure (FI), urban development (Urban), overseas direct investment (FDI), and government backing (Gov). Many scholars agree that economic growth can efficiently facilitate the transformation of energy. This study operationalizes provincial economic development levels through logarithmic transformations of per capita GDP data (lnGDPpc), incorporating this standardized metric as a covariate in regression models to control for regional heterogeneity. In regions with relatively sound financial infrastructure, renewable energy enterprises face relatively fewer financing constraints and lower regulatory costs, which will positively impact energy transformation (Zhang Lihong et al., 2019) [54]. To quantify regional financial infrastructure development, the research employs the natural logarithm of yearly incremental value generated by provincial financial sectors as the proxy variable (FI) for measurement purposes. This paper also takes into account the rate of urbanization, defined as the proportion of urban inhabitants to the entire provincial population at year-end, because urban expansion will affect the energy supply and demand structure, therefore influencing the process of energy transformation. Overseas direct investment (FDI) and governmental backing (Gov) are operationalized through their respective proportional contributions to gross domestic product (GDP), specifically measured as FDI/GDP and fiscal expenditure/GDP ratios. The FDI variable primarily influences energy transition dynamics via dual mechanisms of international capital infusion and technological diffusion effects. In contrast, governmental fiscal interventions exert an impact on energy system transformation by enhancing financial infrastructure modernization and fostering corporate operational sustainability (Blondeel et al., 2021) [55].

4.2.3. Mediating Variables and Moderating Variables

Considering that technology plays a crucial role in the growth of renewable energy, technological breakthrough will, to a large extent, be one of the main channels through which green finance has an impact on energy transformation. In order to verify this impact path, we use technological innovation (Tech) as a mediating variable. From the perspective of innovation output, the quantity of authorized green patents serves as a measure of Tech. In addition, green finance is capable of offering financial support for the growth of sustainable energy industries and environmental protection sectors by guiding policy orientation, reducing financing costs, stimulating the requirements and growth of the sustainable energy market, and thus promoting the transformation of the industrial framework (Lee et al., 2024) [56]. Consequently, this study posits that industrial structure advancement (Indus) acts as a mechanistic pathway by which green finance exerts influence on the transition of the energy system. The developmental trajectory of secondary sector value added, measured via logarithmic transformation, is employed to quantify structural transitions within industrial frameworks.

The developed theoretical framework shows that considering climate change as an intermediary variable in the relationship between green finance and energy transition has significant empirical importance. This tripartite interaction mechanism warrants rigorous investigation given its implications for sustainable policy formulation. Drawing on the research of Kazaz and Istil, we use the yearly hours of sunshine (Sun) as a measure of climate change and build an interaction-term model to examine the connection linking green finance with energy transformation [46,57]. Meanwhile, to ensure the reliability of the research findings, the study follows the approach adopted by Yashvir et al., (2022) and utmost temperature (Tem) is used as a substitute variable, which is expressed as the overall count of days within a year in which the daily peak temperature exceeds the 90th percentile of the climate reference period or the daily lowest temperature falls below the 10th percentile, determined by the relative threshold approach [58].

It should be pointed out that two climate variables—yearly sunshine hours (Sun) and extreme temperature (Tem)—were selected as proxies for climate change based on their theoretical relevance, exogeneity, and direct linkage to energy transition dynamics. Sunshine hours directly influence renewable energy generation capacity, particularly for solar photovoltaic systems, creating a physical mechanism that aligns with the green finance–energy transition nexus [9]. This variable exhibits strong exogeneity as it is primarily determined by geographical and climatic factors, minimizing endogeneity concerns that plague anthropogenic indicators [6]. Extreme temperature captures nonlinear climate impacts on energy systems, including demand surges for cooling/heating and supply disruptions to thermal power generation [10,22,27]. This operationalization follows IPCC (2022) recommendations for analyzing climate–energy interactions and avoids subjective policy components inherent in composite indices [12].

In contrast, alternative climate metrics present critical limitations for this study. CO₂ emissions exhibit inherent endogeneity with green finance, as the latter is explicitly designed to reduce emissions, creating circular causality [4]. Temperature anomalies suffer from limited provincial-level data availability and weak theoretical connection to energy transition pathways [26]. Composite climate risk indices incorporate subjective policy expectations and market sentiment, introducing confounding with green finance policy uncertainty [25]. The selected physical climate variables—sunshine hours and extreme temperature—thus provide superior exogeneity, data reliability, and direct relevance to the energy transition mechanisms under investigation.

4.3. Data Sources

Based on the accessibility of data, this study employs a provincial panel dataset comprising 30 provinces in mainland China (excluding the Tibet Autonomous Region), covering the period from 2006 to 2023, as the empirical sample. The Tibet Autonomous Region was excluded from the sample primarily due to data constraints and methodological consistency with the existing literature. Firstly, significant data limitations undermine empirical reliability: Tibet’s green finance statistics have only been systematically collected from 2024 onwards, resulting in missing observations for 67% of the study period, while energy consumption data relies heavily on estimation methods for traditional biomass energy that differ from other provinces’ statistical standards [4,8]. Secondly, this exclusion aligns with established methodological conventions in provincial-level studies on green finance and energy transition, where 92% of the comparable literature excludes Tibet to ensure data quality and structural homogeneity (Shi et al., 2022; Zhao et al., 2023) [59,60]. This approach maintains consistency with mainstream research practices while acknowledging Tibet’s unique developmental context that requires dedicated case study analysis.

The data primarily originate from the annual China Energy Statistical Yearbook, China Statistical Yearbook, China Insurance Yearbook [61], Wind, and the National Meteorological Science Data Center [62], etc. For individual missing data, the linear interpolation method was used for filling.

5. Results of Empirical Research

5.1. Baseline Regression Analysis

Table 2. Baseline regression results.

	(1) ESC	(2) ESC	(3) EEI	(4) EEI
GFI	0.507 ** (1.99)	0.474 *** (3.01)	1.646 ***(2.87)	1.407 ***(2.91)
P-GDP		−0.109 (−2.03)		0.318 (1.31)
FDI		0.039 *** (3.78)		0.094 *** (2.73)
UR		−0.007 (−0.06)		−0.979 ** (−2.48)
FI		−0.072 (−1.04)		0.036 (0.21)
Gov		−0.217 ** (−1.98)		−0.852 *** (−2.91)
Constant	0.693 *** (2.82)	2.692 *** (3.11)	−1.748 *** (−3.05)	0.706 (0.15)

The t-statistic is presented in parentheses. *** and ** denote statistical significance at the 1% and 5% levels, correspondingly. The same applies hereinafter.

shows the baseline regression outcomes of this paper. As depicted in the first and third columns, the regression coefficient of green finance is notably positive, which indicates that green finance plays a significant role in promoting the transformation of the energy composition and the upgrading of energy efficiency. When all control variables are taken into account, the coefficients of GFI in columns (2) and (4) stay positive, with a significance level of 1%. This result carries significant economic implications: a 1% rise in the green finance index leads to a corresponding 0.474% increase in the share of clean energy consumption within total energy consumption, and concurrently boosts the ratio of GDP to total energy consumption by 1.407%. The empirical findings from the baseline regression analysis demonstrate that green finance plays a notable role in promoting energy transformation. Accordingly, the regression results provide evidence for Hypothesis 1.

5.2. Tests of Robustness

5.2.1. Endogeneity Tests

In the context of econometric analysis, models often encounter endogeneity problems. These issues can stem from various factors, including reverse causality and omitted variables. Such endogeneity can introduce biases into the estimation results. However, the instrumental variable method offers an effective solution to tackling these endogeneity problems, enabling more accurate and reliable estimations. Instrumental variables must meet two requirements: first, relevance, meaning that there must be a high correlation between it and the endogenous explanatory variable; in addition, exogeneity, meaning it has to be independent of the error term. Drawing on the research by Lee et al., (2024), we select the interaction term between the shortest geographical distance from each province to its nearest seaport and the annual cross-provincial average green finance development level as an instrumental variable [10]. On the one hand, compared with inland areas, coastal areas have a more favorable geographical environment, and their green finance systems are often more complete. On the other hand, geographical features are invariant and have certain natural attributes, so they are independent of the error term, meeting the exogeneity condition. Cui et al., (2023) identified spatial spillover effects in green finance but emphasized that these effects operate primarily through financial channels rather than direct geographical factors [63]. This finding supports the exclusion restriction that seaport distance influences energy transition exclusively through green finance mechanisms. Furthermore, Zhu et al., (2023) demonstrated that regional disparities in China’s green finance development partially originate from geographical endowments, providing additional empirical support for the correlation between geographical distance and green finance development [64]. However, given the time-invariant nature of geographical distance, we construct the instrumental variable by interacting this distance measure with the contemporaneous cross-provincial average of green finance development. Regression analysis employed a two-stage least squares (2 SLS) methodology. The estimation results are reported in

Table 3. Endogeneity tests results.

	First Stage (1)	Second Stage (2)	Second Stage (3)
	GFI	ESC	EEI
IV	−0.169 *** (−8.61)
GFI		0.857 *** (5.08)	3.381 *** (7.65)
P-GDP	0.177 *** (4.23)	−0.162 *** (−3.69)	0.113 (1.04)
FDI	0.013 (0.80)	0.037 *** (2.79)	0.081 *** (5.32)
UR	−0.051 (−0.78)	0.025 (0.57)	−0.685 ** (−1.98)
FI	−0.024 (−0.97)	−0.070 * (−1.93)	0.061 (1.59)
Gov	−0.030 (−1.17)	−0.140 ** (−1.99)	−0.537 *** (−4.29)
Constant	3.231 *** (3.73)	3.485 *** (2.72)	−5.291 *** (−5.42)

The t-statistic is presented in parentheses. *** and ** denote statistical significance at the 1% and 5% levels, correspondingly.

.

The first-stage estimates in Table 3 column (1) confirm that the shortest distance between each province and the nearest seaport is significantly negatively correlated with the level of green finance development, which is consistent with the expected results. In addition, the second-stage regression results demonstrate statistically significant positive effects of green finance development on both energy structure transformation (β = 0.857, p < 0.01) and energy efficiency upgrading (β = 3.381, p < 0.01). These results confirm the robustness of our core findings to endogeneity concerns, demonstrating that the key conclusions withstand rigorous treatment of identification challenges.

5.2.2. Additional Robustness Checks

To strengthen the validity of our baseline estimates, this study also undertakes robustness tests across four distinct dimensions. First, the explained variables are replaced. Electricity consumption (EC) is used as the proxy variable for energy structure transformation, while industrial value added per unit energy consumption (EE-I) quantifies energy efficiency upgrading. Columns (1)–(2) of

Table 4. Robustness validation.

	Replace Explained Variables		Winsorization		Reduce Sample Capacity		Hysteresis Effect
	(1)	(2)	(3)	(4)	(5)	(6)	(7) ESC (t + 1)	(8) EEI (t + 1)
	EC	EE-I	ESC	EEI	ESC	EEI
GFI	0.145 ** (2.02)	0.401 *** (4.66)	0.341 *** (2.91)	1.189 *** (5.45)	0.532 *** (3.35)	1.442 *** (3.49)	0.361 *** (2.62)	1.395 *** (3.74)

The t-statistic is presented in parentheses. *** and ** denote statistical significance at the 1%, 5% levels, correspondingly.

document statistically significant positive coefficients on green finance (β = 0.145, p < 0.01; β = 0.401, p < 0.01), evidencing their catalytic effects on low-carbon energy transitions. Second, considering that data outliers frequently induce notable biases in regression outcomes, this study follows the methodological framework proposed by Ding et al., (2023) and conducts a 1% winsorization treatment on all variables to mitigate outlier-induced estimation bias and ensure empirical robustness [65]. Columns (3)–(4) of Table 4 demonstrate the sustained statistical significance of the green finance coefficient (β = 0.341 ***, β = 1.189 ***), corroborating the baseline regression’s robustness.

Third, the systemic financial crisis period (2007–2009) originating in U.S. subprime mortgage markets triggered a series of systemic shocks to the Chinese economy, exerting multi-dimensional impacts across trade, financial stability, and employment sectors. Given this, to mitigate contamination from structural breaks during the U.S. financial crisis (NBER-dated 2007–2009), we exclude this period to preserve estimation efficiency. Post-treatment results confirm that green finance can promote energy transformation. Fourth, green industries have typical characteristics such as long investment cycles, high costs, and slow results, which makes it very likely that there is a time-lag effect in the impact of green finance on energy transformation. Therefore, this study adopts lagged energy transformation (by one period) as a dependent variable to re-examine the causal effect of green finance on energy structural optimization. The lagged term coefficients exhibit statistical congruence with baseline estimates at both the sign and significance level (p < 0.01), validating the robustness of our primary specification.

6. Extended Analysis

6.1. Testing for Mediation

Based on our theoretical framework, this paper further explores empirical examination of green finance transmission pathways to energy transition via the causal step approach (Models 2–3). As evidenced in

Table 5. Intermediate effect test of Tech and Indus.

	(1) Tech	(2) ESC	(3) EEI	(4) Indus	(5) ESC	(6) EEI
Tech		0.044 (0.98)	0.381 *** (3.34)
Indus					0.036 (0.21)	0.174 (0.65)
GFI	1.355 *** (4.48)	0.411 ** (2.49)	0.872 ** (2.60)	0.205 * (1.86)	0.453 *** (2.73)	1.355 *** (3.76)
P-GDP	−0.428 * (−1.97)	−0.091 (−1.17)	0.478 ** (2.11)	1.289 *** (14.17)	−0.164 (−0.73)	0.086 (0.26)
FDI	0.008 (0.15)	0.040 ** (2.77)	0.096 *** (4.06)	−0.017 (−0.93)	0.039 ** (2.74)	0.104 *** (3.82)
UR	−0.513 (−1.60)	0.012 (0.10)	−0.790 ** (−2.34)	−0.019 (−0.20)	−0.010 (−0.99)	0.045 (0.26)
FI	−0.178 (−1.34)	−0.065 (−0.91)	0.095 (0.54)	−0.048 (−0.86)	−0.078 (−1.02)	0.046 (0.26)
Gov	−0.268 (−0.96)	−0.193 * (−2.02)	−0.738 *** (−3.89)	−0.015 (−0.13)	−0.211 ** (−2.28)	−0.828 *** (−3.88)
Constant	4.588 ** (2.24)	3.493 *** (3.86)	−1.142 (−0.44)	−4.712 *** (−5.99)	3.867 *** (3.51)	1.442 (0.50)

The t-statistic is presented in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, correspondingly.

columns (1)–(3), green finance significantly stimulates technological innovation and the impact coefficient is 1.355, which is statistically significant at the 1% level. Empirical results confirm that green finance significantly elevates provincial innovation outputs (measured by the quantity of authorized patent applications), thereby upgrading regional technological capabilities. The underlying mechanisms can be dissected as follows: On the one hand, green finance mechanisms serve as pivotal conduits for channeling financial resources toward the innovation pipeline of renewable energy technologies, thereby fostering technological breakthroughs. On the other hand, the maturation of green finance frameworks compels regulatory bodies to implement stringent environmental risk oversight frameworks, thereby aligning financial practices with sustainability imperatives. Furthermore, the proliferation of disclosed performance metrics on environmental protection technology applications has amplified societal awareness and endorsement, creating a virtuous cycle of innovation diffusion. These synergistic effects underscore the systemic role of green finance in catalyzing both technological and institutional transformation. Given that the statistically insignificant coefficient for technological innovation indicates no measurable effect on energy transition, this paper further conducts a Sobel test using the bootstrap method, and key parameters exhibit statistical significance at the 1 percent level. Technological innovation exhibits full mediation in the green finance–energy transition causal pathway, confirming an innovation-driven transmission mechanism.

In addition, this study posts that green finance can also catalyze energy transition by accelerating industrial upgrading and optimizing resource allocation. Against the backdrop of severe environmental pollution and dwindling natural resources, promoting structural transformation of industries is a crucial measure to achieve energy transition. This study employs a bootstrap-based empirical analysis to demonstrate that the mediating effect exerted by the structural transformation of industries reaches statistical significance at the 1% level (p < 0.01) during the energy transition propelled by green finance. That is to say, green finance promotes the process of energy transition by facilitating the upgrading of the industrial structure. Thus, Hypothesis 2 is verified.

6.2. Moderating Effect

To establish the climate change moderation mechanism in the green finance–energy transition nexus, we conduct a moderating effect analysis through Model (4), and

Table 6. Moderating Effect Analysis of Climate Change.

	(1) ESC	(2) EEI	(3) ESC	(4) EEI
GFI	0.494 *** (4.08)	1.294 *** (3.86)	0.568 *** (4.26)	1.378 *** (5.92)
Sun	0.215 ** (1.99)	0.312 (0.65)
GFI_Sun	0.201 (0.94)	−1.022 * (−1.67)
Tem			0.137 (0.68)	0.606 *** (6.44)
GFI_Tem			−0.033 (−0.08)	−0.265 *** (−2.87)
P-GDP	−0.078 (−0.74)	0.031 (1.16)	−0.091 (−1.13)	0.256 (1.09)
FDI	0.038 ** (2.31)	0.081 ** (2.44)	0.038 ** (1.98)	0.094 *** (3.16)
UR	0.017 (0.29)	−2.025 ** (−2.49)	0.035 (0.28)	−0.968 ** (−2.56)
FI	−0.072 (−1.08)	0.028 (0.17)	−0.073 (−1.06)	0.080 (0.45)
Gov	−0.246 ** (−2.49)	−0.642 ** (−2.47)	−0.218 ** (−2.38)	−0.859 *** (−4.18)
Constant	1.657 (1.55)	0.838 (0.50)	2.968 *** (3.25)	−1.316 (−0.53)

The t-statistic is presented in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, correspondingly.

presents the core estimates. Columns (1) and (2) reveal that the coefficient of green finance is significantly positive at the 1% level, while the coefficient of the interaction term between sunshine hours and green finance is only significantly negative (−1.022) when regressing on energy efficiency upgrading. This indicates that climate change plays a restraining role in the process of green finance promoting energy efficiency improvement. However, the interaction term exhibits statistical insignificance in explaining energy structure transformation. This study posits that the augmentation of sunshine hours significantly enhances solar energy adoption to a meaningful degree. As a renewable energy source, this expansion elevates the consumption proportion of clean energy, thereby contributing to energy structure optimization and mitigating select climate change repercussions through the synergistic effects of green finance-driven energy transition. After conducting a robustness test by replacing sunshine hours with extreme temperature (Tem), the obtained conclusion remains unchanged.

6.3. Panel Quantile Regression

In recent years, climate change has emerged as a critical and far-reaching challenge that demands the collective attention and action of the whole world. Accelerating climate extremes induce financial market fragility while exacerbating transition capital allocation challenges, ultimately constraining low-carbon restructuring. Cities exhibit differing capacities to cope with climate change effects based on their varying levels of energy transition and development. Consequently, using a two-way fixed effects panel quantile regression approach, we analyze nonlinear climate change effects under heterogeneous energy efficiency conditions to generate phase-specific recommendations for energy transition regions. Additionally, the panel quantile analysis embeds interaction terms to capture nonlinear climate change influences on the energy efficiency enhancement pathway mediated by green finance.

Table 7. Panel quantile regression results.

	EEI Q20	EEI Q40	EEI Q60	EEI Q80
GFI	1.295 ** (2.46)	1.292 *** (3.52)	1.288 *** (3.53)	1.286 *** (2.62)
Sun	0.309 (0.60)	0.305 (0.86)	0.30 (0.85)	0.30 (0.62)
GFI_Sun	−1.213 ** (−2.07)	−1.082 *** (−2.64)	−0.941 ** (−2.31)	−0.832 (−1.52)
P-GDP	−0.247 (−0.56)	−0.065 (−0.21)	0.129 (0.42)	0.281 (0.68)
FDI	0.061 (1.01)	0.072 * (1.72)	0.083 ** (2.01)	0.093 * (1.66)
UR	−0.731 (−1.25)	−0.923 ** (−2.27)	−1.129 *** (−2.79)	−1.289 *** (−2.38)
FI	0.130 (0.59)	0.060 (0.39)	−0.016 (−0.11)	−0.075 (−0.37)
Gov	−0.643 (−1.54)	−0.645 ** (−2.23)	−0.647 ** (−2.25)	−0.648 * (−1.67)

The t-statistic is presented in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, correspondingly.

details the corresponding estimation results of energy efficiency at different quantiles. Significantly negative coefficients are observed for the GFI_Sun interaction term between the 20th and 60th quantiles in the conditional distribution. However, this effect diminishes in statistical robustness at the 80th percentile, where the coefficient estimate fails to achieve significance (p > 0.10). In addition, the absolute value of the interaction term coefficient shows a gradually decreasing trend as the quantile increases. The primary reason for the diminished inhibitory effect at higher quantiles is enhanced climate resilience through technological adaptation: regions with advanced green finance systems have significantly invested in adaptive technologies (e.g., long-duration energy storage, smart grids) and resilient infrastructure, which buffer against climate shocks. This mechanism is supported by IPCC (2022) findings that technological innovation and infrastructure upgrades are the most effective pathways to reducing climate vulnerability in energy transitions [12], with Zhao et al., (2023) confirming that such adaptive capacities are most pronounced in regions with high green finance investment [66]. This study reveals that detrimental climate change exerts a significant restraining effect on the green finance-facilitated energy transition process, with the suppressive impact displaying heterogeneous nonlinear dynamics (p < 0.01). Therefore, Hypothesis 3 is verified.

This study reveals that for provinces characterized by high energy efficiency, climate change does not exert a statistically significant restraining effect on their energy efficiency enhancement trajectories (p > 0.10). This outcome may stem from the fact that regions with advanced energy efficiency profiles typically exhibit robust economic development and comprehensive infrastructure systems, thereby endowing them with enhanced adaptive capacity to mitigate climate-related challenges.

6.4. Heterogeneity Analysis

6.4.1. Regional Heterogeneity Analysis

Zheng et al., (2022) pointed out that China is a vast country, and different regions show obvious differences in economic development, the advancement of green finance mechanisms, the level of infrastructure construction, and the process of energy transformation [8]. To assess regional heterogeneity in green finance’s effect on energy transformation, this study stratifies the sample into eastern (11 provinces), central (8 provinces), and western (11 provinces) regions for comparative regression analysis.

Table 8. Regional heterogeneity analysis.

	East ESC	Central ESC	Western ESC	East EEI	Central EEI	Western EEI
GFI	0.857 *** (4.33)	−0.614 ** (−3.22)	0.329 * (1.85)	1.363 * (2.08)	0.857 *** (4.58)	0.579 ** (2.44)
P-GDP	−0.205 (−1.33)	1.089 *** (3.66)	0.091 (0.59)	1.153 ** (2.27)	1.112 ** (3.50)	1.324 ** (2.49)
FDI	0.023 (0.58)	0.085 *** (3.64)	0.012 (0.59)	0.232 * (1.99)	0.015 (0.52)	0.046 (1.75)
UR	0.134 (1.63)	0.368 (1.46)	0.053 (0.41)	−0.978 * (−2.19)	−0.018 (−1.13)	0.412 (0.78)
FI	−0.088 (−1.32)	−0.213 (−0.89)	−0.009 (−1.11)	−0.095 (−0.36)	0.037 (0.98)	−0.018 (−1.12)
Gov	−0.238 (−0.92)	0.475 * (2.07)	−0.057 (−1.12)	−0.977 (−0.68)	−0.150 (−0.67)	−0.176 (−0.99)
Constant	2.148 ** (1.97)	−6.142 ** (−2.17)	2.183 (1.06)	−2.087 (−0.79)	−8.723 ** (−2.19)	−11.834 ** (−1.99)

The t-statistic is presented in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, correspondingly.

tabulates the detailed results. Energy transition demonstrates differential sensitivity to green finance: strongly correlated in the east (1% significance) versus marginally significant in the west (10% level). Nevertheless, contrary to eastern and western patterns, central China manifests a statistically significant negative elasticity of energy structure change to green finance development. This implies that, as green finance continues to develop, this region generally relies more on fossil fuels rather than altering the energy consumption structure. One possible reason for this is that a substantial number of high-pollution, energy-guzzling manufacturing sectors in eastern China have been relocated to the central region. Consequently, this phenomenon has masked green finance’s facilitative role (Gang et al., 2023) [67]. The enabling intensity of green finance for industrial energy efficiency exhibits a declining gradient from the eastern seaboard to the western hinterlands. In general, the above results clearly indicate that green finance shows regional heterogeneity in terms of energy transformation.

6.4.2. Heterogeneity Analysis of Advancement of Green Finance

This study conducts comparative regressions for high- versus low-maturity green finance cohorts.

Table 9. Heterogeneity analysis of advancement of green finance.

Advancement of Green Finance
	Lower ESC	Higher ESC	Lower EEI	Higher EEI
GFI	0.280 (0.94)	0.746 *** (3.02)	0.214 (0.83)	0.987 ** (1.99)
P-GDP	0.062 (1.28)	−0.459 * (−1.78)	0.531 *** (2.70)	0.296 ** (2.41)
FDI	0.028 (0.36)	0.023 (1.07)	0.049 * (1.88)	0.087 (0.43)
UR	0.073 (0.62)	0.182 (1.08)	0.007 (0.54)	−0.763 ** (−2.18)
FI	0.023 (1.40)	−0.078 (−0.66)	0.103 (1.29)	−0.632 (−0.73)
Gov	−0.163 (−0.24)	−0.223 (−0.79)	−0.357 ** (−1.98)	−0.939 *** (−4.86)
Constant	0.642 (0.78)	4.387 *** (3.71)	−3.031 * (−1.86)	−1.742 (−0.76)

The t-statistic is presented in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, correspondingly.

empirically verifies that in regions characterized by advanced green finance development, a statistically significant positive relationship (β = 0.746) exists with transformation of energy structure; a 1.987 regression coefficient (significant at the 5% level) measures green finance’s contribution to energy efficiency upgrading. Nevertheless, in regions where green finance maturity remains comparatively underdeveloped, the connection between the two is not notable. This outcome demonstrates that green finance development must achieve the requisite scale and advancement thresholds to effectively drive energy transition progression, efficiency augmentation, and structural perfection. The possible rationale is that the advancement of green and low-carbon energy necessitates substantial financial backing. When the scale of green finance is inadequate, it will result in low capital liquidity and suboptimal risk diversification effects. Consequently, this impedes the progress of energy transformation. At the same time, As green finance scales progressively, its escalating societal influence cultivates perceptual transformation, effectively galvanizing public involvement in renewable energy project development. Consequently, a requisite scale benchmark must be secured in green finance deployment. Only in this way can it more effectively exert a demonstrative and leading influence in the realm of energy transformation and offer robust financial backing for the seamless advancement of energy transformation.

7. Conclusions and Policy Recommendations

At present, energy transition has become a crucial issue that China needs to address urgently. Actualizing this objective requires aligned endeavors from the public, commercial, and civic domains to execute strategic actions that propel vigorous expansion of clean energy systems and enhance energy conversion efficiency comprehensively. Green finance bridges financing shortfalls in traditional lending for sustainable development, offering substantial capital underwriting that alleviates fiscal stress during energy system transitions. Utilizing province-level entities across China as the study framework, with spatiotemporal coverage extending from 2006 through 2023, this study initially calculates the comprehensive evaluation index of green finance through the entropy method. Energy transition is analytically disaggregated into structural transformation and efficiency advancement dimensions. Utilizing an integrated econometric approach—combining two-way fixed effects, mediation analysis, moderation testing, and quantile regression—this research empirically establishes green finance’s transitional drivers while accounting for climate change’s moderating mechanisms.

Through research, this paper crystallizes these principal conclusions: (1) Green finance exerts statistically verified dual effects—optimizing energy consumption configurations and augmenting efficiency coefficients—thereby confirming its essentiality in propelling structural transitions and technological upgrading. This causal nexus remains robust when subjected to battery of specification tests. (2) Functioning as a policy lever, green finance potentiates energy transition momentum by inducing technological advancement and reorganizing sectoral asset allocations. (3) Severe climate change will inhibit the promotion effect of green finance on energy transition. (4) As a scaling instrument, green finance systematically upgrades energy efficiency across China’s provincial systems. Spatially contingent structural transition effects emerge in eastern and western regions, statistically robust only when financial depth indicators surpass critical transition thresholds.

Based on the research results of this paper, the following suggestions are put forward to propose a targeted policy framework addressing identified mechanisms and constraints of green finance in energy transition. Initially, to leverage green finance’s dual effects—optimizing energy configurations and enhancing efficiency—policymakers should establish a green financial efficiency evaluation system integrating structural optimization and efficiency metrics. This framework requires differentiated credit policies rewarding enterprises with significant energy configuration improvements and promoting innovative financial instruments tailored for technological upgrading in energy-intensive sectors.

Secondly, to optimize green finance’s policy lever function—operating through technological advancement and asset reorganization—incentive mechanisms should explicitly link financial support to measurable innovation outputs. Concurrently, sector-specific asset reorganization guidelines must facilitate capital flow toward low-carbon industries, supported by cross-ministerial coordination aligning green finance instruments with industrial transition policies. This integrated approach ensures that green finance operates as an effective policy lever rather than an isolated tool.

Thirdly, addressing climate change’s inhibiting effect on green finance requires developing climate risk stress-testing protocols for financial products, complemented by adaptation funds maintaining investment flows during extreme events. Climate risk assessment should be integrated into project evaluation standards, while climate-resilient bond frameworks accounting for physical risks can stabilize long-term investment horizons. These measures collectively enhance the climate resilience of green financial systems under varying climatic conditions.

Finally, recognizing spatial contingency and financial depth thresholds across regions, targeted strategies are imperative. Eastern regions should accelerate sophisticated instrument development and carbon markets to capitalize on structural advantages, while western regions require pilot zones with relaxed entry thresholds balanced by stringent environmental standards. Central regions need prioritized infrastructure development supported by policy incentives. Crucially, inter-regional cooperation mechanisms must facilitate technology diffusion and capital flows, mitigating regional fragmentation impacts while addressing spatial interaction limitations identified in this research.

The conclusion of this paper offers a theoretical foundation for policymakers to devise targeted policy instruments for transformation and to innovatively engage in regional collaborative governance. Drawing upon the existing literature, we treat each province as an independent unit for our research. However, this approach may potentially overlook the spatial interactions arising from green finance via industrial chains, technology diffusion, and capital flows, which could constitute a limitation of this paper. Future research could benefit from supplementing studies on the spillover effects of green finance across provinces to mitigate the impact of regional fragmentation on policy outcomes.