Third, current analyses of EETE driving factors typically rely on conventional econometric techniques (Wang et al. 2023; Yang et al. 2023; Lan et al. 2025), which are often limited in explaining high-dimensional nonlinear interactions. To overcome this limitation, this study introduces the SHAP method from the field of machine learning. This approach enables transparent quantification of the marginal contributions and interactive effects of multiple features on EETE, effectively addressing the interpretability challenges of complex nonlinear relationships. By visualizing attribution results, it also offers micro-level empirical evidence to support differentiated policy design.
EETE reflects the relationship between output generated to meet human production and living needs and the input of resources and environmental factors. It serves as a key indicator for achieving a balance between economic growth and environmental protection. This concept has been recognized by major international organizations such as the World Business Council for Sustainable Development and the Organization for Economic Co-operation and Development (OECD), and it has gradually become a central topic in the integrated study of regional economic and environmental systems. Existing research primarily focuses on two aspects: the measurement of EETE and the identification of its influencing factors.
Currently, the measurement of EETE primarily relies on two methodological approaches: parametric and non-parametric models. Among these, non-parametric methods are particularly suitable for ecological systems characterized by multiple variables and objectives, due to their flexibility and independence from specific functional forms. The data envelopment analysis (DEA) model, originally proposed by Charnes et al. (1978), is the most widely used non-parametric approach. Although DEA was initially developed for static efficiency evaluation, recent studies have increasingly combined it with the Malmquist index to assess dynamic changes in efficiency. To better reflect real-world environmental characteristics, researchers have incorporated undesirable outputs into the DEA framework. For example, Tian et al. (2025) included pollutants such as sulfur dioxide in the Super-SBM model, significantly enhancing the interpretability of the results.
The Super-SBM model has now been extensively applied to the assessment of regional EETE, with a growing body of empirical literature based on this approach. Yang and Yu (2025) suggested that green total factor productivity could serve as a proxy for EETE and employed the SBM-GML model for calculation. Yao et al. (2020), Yang et al. (2024a), Zhang et al. (2024), and Feng et al. (2022) all adopted the Super-SBM model to evaluate EETE levels and regional disparities, while identifying key influencing factors. In terms of model extension, Zhao et al. (2022) integrated the super EBM model with social network analysis (SNA), which not only revealed the efficiency levels of provinces in China but also highlighted the spatial interdependencies across regions. Lei et al. (2025) introduced the by-production technique to evaluate EETE performance in emerging economies from the perspective of carbon trading systems. Chen et al. (2023) focused on China's coastal areas and applied a three-stage DEA model to examine the significant impact of external environmental variables on marine EETE. Building on these efforts, researchers have also attempted to construct more comprehensive frameworks to capture the multidimensional structure of EETE. Sun et al. (2024a) combined the two-stage Super-NSBM model with the window DEA model and established an evaluation system encompassing environmental, economic, health, and social welfare dimensions, thereby improving the coverage and structural integrity of the indicator system. Sun et al. (2024b), from the perspective of ecosystem service supply and demand, developed a hybrid approach that integrates the entropy method with the InVEST model, and empirically validated in Shenzhen the positive correlation between EETE and the matching degree of ecosystem service supply and demand.
In addition to non-parametric methods, parametric approaches have also demonstrated practical value in cross-national efficiency comparisons. DiMaria (2019), Wang et al. 2023, and Xu et al. (2021) employed the stochastic frontier analysis (SFA) framework to conduct international assessments of EETE, thereby providing a theoretical and methodological foundation for global comparative studies of ecological efficiency.
In summary, although the methodological landscape for EETE measurement has become increasingly diverse in terms of model types and application contexts, mainstream approaches remain predominantly static, with SBM or Super-SBM models being most widely adopted. These models face structural limitations, such as insufficient responsiveness to temporal dynamics and limited compatibility with evolving policy objectives. To address these issues, this study develops a two-stage Dynamic DDF model that offers several critical improvements. First, the model incorporates intertemporal dynamic linkages. By introducing a carry-over mechanism for sustainability-related variables such as capital, the model captures technological inheritance and path dependency between decision-making periods. This approach overcomes the temporal discontinuity inherent in SBM models and better reflects the gradual and cumulative nature of regional ecological transitions. Second, the model supports direction-specific policy alignment. The DDF framework allows for the flexible specification of direction vectors in accordance with region-specific or stage-specific policy goals, offering a high degree of strategic adaptability. In contrast, the SBM model, due to its radial optimization structure, has limited capacity to reflect trade-offs in multi-objective decision processes. Third, the model provides greater flexibility in structural decomposition. By adopting a two-stage decomposition of EETE, it facilitates the identification of subsystem-level structures and sources of efficiency variation, thereby offering a more targeted diagnostic basis for heterogeneous regional contexts.
Urbanization is widely regarded as one of the core determinants influencing regional EETE. On the one hand, urbanization promotes the improvement of infrastructure, economic agglomeration, and the development of green governance systems, thereby enhancing regional resource allocation efficiency and ecosystem service capacity. From the perspective of the digital economy, Shi et al. (2025) found that its coupling with urbanization can improve regional ecological efficiency through technological innovation. At the micro-level, Yang et al. (2024a) reported that urbanization has driven the expansion of cultivated land, significantly enhancing farmland ecosystem services. Li et al. (2023) revealed a nonlinear turning point in the relationship between urbanization and ecological quality, indicating that improvements in urbanization quality can contribute to ecosystem restoration. On the other hand, the ecological impact of urbanization does not manifest uniformly across regions; instead, it exhibits marked spatial heterogeneity. In the middle reaches of the Yangtze River, Zhang et al. (2024) found insufficient synergy between urbanization and EETE, with most cities reporting coupling degrees below 0.5. From the perspective of ecological intensity, Hu et al. (2024) suggested that the effect of urbanization on EETE varies depending on regional ecological carrying capacity. In areas with lower ecological intensity, urbanization tends to play a positive role, whereas in regions with higher ecological intensity, it may impose additional ecological stress, resulting in a suppressive effect.
In general, the influence of urbanization on EETE is characterized by multidimensional complexity. While it offers positive pathways through resource optimization and ecological incentives, it also poses potential risks such as increased ecological burdens and policy mismatches.
Green governance plays a multifaceted role in enhancing EETE, primarily through institutional constraints, optimization of resource allocation, and the management of externalities. In the dimension of green finance, existing studies have emphasized its role in directing capital flows toward environmentally friendly industries, thereby reducing ecological footprints and improving overall ecological performance (Wang et al. 2024b). More specifically, loans supporting petroleum and mineral resources development have been shown to significantly improve ecological efficiency in terms of human development (Coulibaly, 2024). Green innovation, as a key transmission channel of green finance, has demonstrated a significant positive effect in most developed countries (Liao et al. 2023). However, its effectiveness appears less robust in emerging economies. For instance, Pata et al. (2024), in a study on BRICS countries, found that technological innovation had not led to substantive improvements in ecological footprints, reflecting substantial regional differences in institutional embedding and environmental capacity for green governance. Such divergence is also evident at the urban scale. Moreover, the effectiveness of green governance is jointly constrained by the type of policy instruments employed and the regional stage of development. Han et al. (2024) demonstrated that green trade policies yield stronger marginal effects in countries with lower levels of ecological efficiency, suggesting a higher elasticity of policy response in low-baseline areas. The positive impact of carbon trading policies on regional ecological efficiency is primarily realized through the inclusion of urban green spaces in emission reduction output systems (Lei et al. 2025).
In summary, green governance affects EETE through multiple mechanisms, including financial incentives, institutional regulation, and ecological management. However, the effectiveness of these policies is significantly influenced by the foundational conditions of regional development and the degree of spatial coordination.
As an external institutional foundation for ecological-environmental transformation, the macroeconomic environment exerts significant guiding and constraining effects on regional EETE. While economic growth can enhance the supply capacity of ecosystem services, it may also induce excessive resource consumption and pollution dispersion, indicating that its influence is not unidirectionally positive. On one hand, economic development promotes the optimization of resource allocation, improvement of infrastructure, and technological advancement, all of which are conducive to enhancing EETE. Jin et al. (2020), using nighttime light data, found that economic agglomeration significantly improves regional ecological efficiency. Jiang et al. (2024) further suggested that the co-evolution of economic growth and technological progress contributes more effectively to dynamic improvements in EETE. At the national level, Mehboob et al. (2024) found that GDP growth in Turkey had a significant positive impact on reducing ecological footprints, indicating that in some developing countries, growth does not necessarily come at the expense of the environment. On the other hand, the ecological effects of economic growth may exhibit complex nonlinear patterns. Zhou et al. (2024) found a U-shaped relationship between the level of digitalization and low-carbon transition efficiency, suggesting that macro-structural transformation has a staged threshold effect on ecological performance. Quan et al. (2024), in a study of OECD countries, reported that economic growth significantly suppresses ecological footprints in higher quantiles and revealed a bidirectional causal relationship. Ma et al. (2024), considering the role of economic policy uncertainty, pointed out that in countries with high levels of pollution, policy stability serves as a key moderating factor influencing EETE. In addition, institutional and structural policy instruments also function as important exogenous variables shaping EETE. Magazzino et al. (2025) empirically demonstrated that environmental taxation can indirectly enhance EETE by controlling ecological footprints and reducing environmental costs. Yang and Yu (2025) highlighted the positive relationship between local fiscal capacity and the coordination of ecological governance, offering a policy basis for optimizing economic governance structures. Smart city development, as a new pathway of urbanization, has also shown considerable potential in improving ecological performance. Song et al. (2022) found that smart city pilot programs primarily enhance resource use efficiency through energy-saving mechanisms, although further improvements in environmental performance are still required.
In summary, the influence of the macroeconomic environment on EETE has evolved beyond a singular focus on growth, toward a multidimensional characterization involving institutional flexibility, policy stability, and emerging economic paradigms.
In the process of regional ecological-environmental transformation, infrastructure serves not only as a material foundation but also plays a pivotal role in shaping developmental pathways. First, infrastructure enhances the organization and allocation of regional resources, thereby providing a more efficient platform to support ecological transitions. Feng et al. (2025) and Sheng et al. (2024), examining high-speed rail and charging facilities respectively, revealed that the optimization of transportation and energy networks can strengthen urban functional coupling and reduce ecological costs. Jiang et al. (2024), using underdeveloped counties as a case study, confirmed the foundational role of infrastructure in improving EETE, which is particularly crucial in areas with weak ecological governance capacities. Second, the construction of digital infrastructure is reshaping both the technological and institutional foundations of the green economy. Wang et al. (2024c) demonstrated that digital infrastructure significantly enhances ecological total factor productivity through the diffusion of green innovation and the agglomeration of green industries. Li et al. (2025a) further found that its impact on ecological resilience exhibits a cumulative time effect, thereby strengthening the dynamic adaptability of ecological governance. In addition, land use efficiency has become a critical dimension in assessing the ecological value of infrastructure. Su et al. (2024) and Xu et al. (2025) noted that the relationship between land development intensity and ecological efficiency is nonlinear, involving significant marginal constraints and risks of resource misallocation. From the perspective of urban agglomerations, Li et al. (2025b) analyzed the spatial heterogeneity between land use efficiency and EETE, emphasizing that the efficiency gains generated by infrastructure vary considerably across different stages of development.
In summary, the influence of infrastructure on EETE has evolved from a unidirectional technological input to a structural intervention mechanism. Its effectiveness depends on the dynamic alignment among resource endowments, regional functions, and institutional contexts.
The stability of ecological foundations is a fundamental prerequisite for ensuring the long-term sustainability of regional EETE. Focusing on key ecological elements such as water, forests, soil, and waste, existing studies have conducted systematic investigations from the perspectives of resource use efficiency, ecosystem resilience, and ecological service functions. Regarding water resources, Ji et al. (2025) emphasized that optimizing water allocation is critical to ecological sustainability, noting a significant positive correlation between ecological water use and levels of green development. The agricultural ecological base has also received growing attention. Gou et al. (2025) found that large-scale ecological agricultural infrastructure can effectively control soil erosion through water conservation and farmland protection, thereby enhancing regional ecological sustainability. Forest system restoration has been identified as a central factor in improving ecological quality. Based on remote sensing data, Zhang et al. (2025c) concluded that forest integrity and landscape restoration are key drivers of regional ecological improvement. Guo et al. (2024) and Jiang et al. (2025), focusing on solid waste utilization, demonstrated that resource recycling and industrial waste reuse exert a positive influence on regional ecological-environmental transformation.
Overall, the ecological foundation determines both the carrying capacity and the regenerative potential of regional ecosystems and constitutes the basis for sustained improvements in EETE. While current research has identified direct linkages between ecological components and EETE, an integrated analysis of the underlying mechanisms and their systemic interactions remains lacking.