1. Introduction

1.1. Research Background

As the global community confronts the accelerating threat of climate change, the urgency to transition toward a low-carbon economy has become increasingly apparent. Central to this transition are new energy industries, including solar photovoltaics, wind power, hydrogen energy, and other renewable technologies. These sectors are expected to play a pivotal role in decarbonizing energy systems, reducing greenhouse gas emissions, and achieving long-term climate targets such as those outlined in the Paris Agreement and various national "net-zero" pledges.

However, the capital-intensive nature of new energy projects—characterized by long payback periods, high initial costs, and technological uncertainty—poses significant barriers to their development, especially in emerging economies. In response, green finance policies have emerged as critical mechanisms to address these funding gaps and to channel financial resources into environmentally sustainable projects. These policies encompass a wide range of instruments, including green bonds, carbon trading schemes, renewable energy subsidies, and sustainable banking regulations.

Despite the proliferation of such instruments, evaluating the investment efficiency of new energy sectors under these green finance regimes remains challenging. Conventional financial metrics often fall short in capturing the full spectrum of environmental, social, and governance (ESG) considerations. Moreover, efficiency is not only a function of financial input and energy output but also reflects the capacity of firms and regions to effectively convert policy incentives into tangible performance gains.

In this context, it becomes crucial to adopt dynamic, multi-dimensional evaluation models that can capture the temporal evolution of investment efficiency and disentangle the contributions of technological progress, scale, and managerial practices. Among the methodological approaches available, the Data Envelopment Analysis (DEA) combined with the Malmquist Index stands out for its ability to assess productivity changes over time and to identify sources of efficiency variation.

1.2. Literature Review

The rise of green finance as a research domain has prompted extensive theoretical and empirical work. Taghizadeh-Hesary and Yoshino (2019) provided a foundational framework, emphasizing the role of financial systems in facilitating low-carbon investments and mitigating risks for green projects. Their studies highlight the synergistic effects of public and private financing mechanisms, as well as the importance of institutional support in achieving sustainability targets.

In parallel, a growing body of literature has applied efficiency analysis methods to assess performance in the energy sector. The DEA-Malmquist index approach, in particular, has gained prominence due to its ability to decompose total factor productivity (TFP) change into components such as technical efficiency, technological progress, and scale efficiency. For example, Wang et al. (2021) applied this method to assess productivity in China's renewable energy firms, demonstrating the utility of the Malmquist index in capturing intertemporal performance shifts.

However, several important gaps remain in the existing research:

Most studies lack dynamic temporal analysis, focusing instead on static efficiency levels at a single point in time.

Few analyses explicitly link green finance policies to changes in investment efficiency, making it difficult to assess policy effectiveness.

Regional disparities and the heterogeneity of policy impact across different subnational units (e.g., provinces, municipalities) are often underexplored.

These gaps underscore the need for a comprehensive, policy-sensitive, and regionally nuanced approach to efficiency measurement in new energy sectors.

1.3. Research Objectives

This study seeks to fill the aforementioned gaps by conducting a dynamic and empirical assessment of investment efficiency in new energy industries, driven by the implementation of green finance policies. Specifically, it pursues the following objectives:

To evaluate the temporal evolution of investment efficiency in new energy sectors across different regions, using a DEA-Malmquist index framework;

To assess the differential impact of various green finance policy instruments, such as subsidies, carbon markets, and green bond issuance, on investment performance;

To identify regional disparities in efficiency gains and losses, highlighting the role of institutional, technological, and managerial factors;

To decompose efficiency change into its constituent drivers, distinguishing between improvements due to technological innovation versus managerial or scale-related factors.

Through this approach, the study aims to provide evidence-based insights for policymakers, investors, and energy firms seeking to optimize capital allocation, improve productivity, and align investment strategies with long-term sustainability goals.

2. Methodology: DEA-Malmquist Index Framework

2.1. Theoretical Basis

The analytical framework adopted in this study combines Data Envelopment Analysis (DEA) with the Malmquist Productivity Index (MPI) to assess both the static and dynamic dimensions of investment efficiency in new energy industries. This approach allows us not only to identify which decision-making units (e.g., provinces or firms) are efficient at a given point in time but also to evaluate how their productivity evolves over time in response to green finance policies.

Data Envelopment Analysis (DEA) is a non-parametric linear programming technique widely used to assess the relative efficiency of decision-making units (DMUs) based on multiple input and output indicators. DEA constructs an empirical production frontier and evaluates each unit's distance from this frontier, thereby identifying technical efficiency levels. In this study, the DEA-BCC (Banker, Charnes, Cooper) model is used under variable returns to scale (VRS), which is more appropriate for heterogeneous regional contexts and differing industrial scales.

To capture the intertemporal changes in productivity, we employ the Malmquist Productivity Index (MPI), which decomposes total factor productivity (TFP) change between two time periods into:

Technological Change (TC): Reflects shifts in the production frontier, often driven by innovation or policy-induced modernization.

Technical Efficiency Change (TEC): Measures whether a DMU is moving closer to or further away from the best-practice frontier.

Scale Efficiency Change (SEC): Assesses the extent to which changes in output are proportional to changes in inputs, reflecting the optimal scale of operation.

The Malmquist index is particularly useful in this study as it does not require price data (a common limitation in developing energy markets) and allows for a non-parametric decomposition of efficiency sources, aligning well with the policy-heterogeneous nature of green finance interventions.

2.2. Model Specification

To operationalize the DEA-Malmquist model, a comprehensive set of inputs, outputs, and policy-related control variables are defined, drawing from the unique attributes of new energy industries and the mechanisms of green finance.

Input Variables:

R&D Expenditure (in million RMB): Represents technological investment and innovation effort.

Green Credit Volume (in billion RMB): Measures the amount of bank loans directed toward green projects.

Government Subsidies (in million RMB): Captures direct fiscal support to renewable energy firms or projects.

Output Variables:

New Energy Capacity (MW): Reflects physical expansion and production capability in renewables.

Carbon Emission Reductions (tons): Measures environmental effectiveness as a non-monetary return.

Revenue Growth (year-on-year %): Indicates economic performance tied to investment decisions.

Policy Variable:

Green Finance Index (GFI): A composite regional index constructed to reflect the intensity and comprehensiveness of green finance policy frameworks. It includes indicators such as:

Number and volume of green bond issuances

Presence and generosity of green tax incentives

Regulatory stringency of local green finance guidelines (e.g., ESG disclosure mandates)

The inclusion of the GFI allows for heterogeneity across regions and enables investigation into how variations in policy environments mediate changes in investment efficiency. The GFI is introduced as an external environmental variable in the second-stage analysis, which could involve regression or clustering to assess correlations between policy strength and MPI results.

2.3. Data Sources

The study relies on a balanced panel dataset covering 30 provincial-level regions in China from 2015 to 2022. This timeframe captures key policy shifts, including the acceleration of carbon trading pilot programs, the rollout of green bonds, and increasing regulatory attention to ESG frameworks.

Primary data sources include:

China’s National Bureau of Statistics (NBS): For macro-level inputs and energy capacity data.

CSMAR (China Stock Market & Accounting Research): For firm-level R&D, subsidies, and financial performance data.

Wind Energy Database: For technical data on renewable projects, emission reductions, and project-level output indicators.

People’s Bank of China and China Green Finance Committee Reports: For data on green credit and bond issuance.

Local policy bulletins and government work reports: For qualitative and quantitative data feeding into the construction of the Green Finance Index (GFI).

Data cleaning involved deflating financial variables to constant 2015 prices using the GDP deflator and standardizing emission reduction measures to ensure cross-provincial comparability. Outlier detection was conducted via the interquartile range method, and missing values were imputed using regional means or firm-level interpolation where appropriate.

3. Empirical Results and Discussion

3.1. Overview of Efficiency Scores

Using the DEA-Malmquist index, we compute Total Factor Productivity (TFP) changes for 30 Chinese provinces over the period 2015–2022. The average Malmquist Index (MPI) value across all provinces is 1.042, indicating an overall 4.2% annual increase in productivity, driven primarily by technological improvements in renewable energy deployment.

Table 1. MPI and its decomposition into Technical Efficiency Change (TEC), Technological Change (TC), and Scale Efficiency Change (SEC).

The results suggest that Technological Change (TC) is the dominant factor driving productivity growth across all regions. In contrast, Technical Efficiency Change (TEC) remains relatively stagnant, and Scale Efficiency Change (SEC) hovers around unity, implying that most provinces are already operating near optimal scale.

3.2. Temporal Trends and Turning Points

To visualize temporal patterns, we compute the annual average MPI values and plot the trajectory of investment efficiency over the 8-year period.

The sharpest gains occurred between 2016 and 2018, coinciding with the introduction of China’s first green bond issuance guidelines and expansion of carbon market pilots.

A temporary dip is observed in 2020, likely due to pandemic-related disruptions and short-term declines in project investment.

Recovery resumes post-2021 with the strengthening of ESG reporting standards and the national dual-carbon (双碳) targets set for 2030/2060.

These results provide empirical validation that green finance policies not only mobilize capital but also enhance investment efficiency over time.

3.3. Regional Disparities in Green Investment Efficiency

Significant inter-provincial disparities are observed. Beijing, Jiangsu, and Guangdong lead the rankings, driven by high GFI scores and concentration of tech-oriented firms. Meanwhile, Gansu, Qinghai, and Ningxia show slower growth despite strong resource endowments, suggesting managerial and institutional constraints.

Figure 4. Provincial MPI Map (Average 2015–2022)(Insert a shaded map of China with provinces colored by average MPI value.)

This spatial variation highlights the importance of not just natural resources but also policy design, financial access, and human capital availability in driving green investment efficiency.

3.4. Role of Green Finance Policy Index (GFI)

To test the explanatory power of green finance policy environments, we conduct a second-stage panel regression using average MPI as the dependent variable and GFI as the key independent variable, controlling for GDP per capita and industrial structure.

Table 2. Fixed-effects regression showing the influence of GFI on MPI. Significance levels

***p<0.01, *p<0.05.

The results confirm a strong and statistically significant relationship between regional green finance strength and investment efficiency. Notably, regions with higher GFI scores tend to attract better-performing projects and demonstrate greater responsiveness to technological innovations.

Summary of Findings

Overall improvement in investment efficiency is observed across China’s new energy sectors, mainly driven by technological progress rather than scale expansion or managerial improvement.

Eastern regions outperform western ones, indicating potential returns to policy harmonization and interregional knowledge transfer.

Green finance policy intensity, as captured by the GFI, is a significant predictor of investment efficiency, supporting the role of institutional quality in green transition pathways.

4. Conclusion and Policy Implications

4.1. Summary of Key Findings

This study applies a DEA-Malmquist index framework to assess the investment efficiency dynamics of China’s new energy industries under the influence of green finance policies from 2015 to 2022. The key findings are as follows:

Overall improvement in productivity was observed across the sector, with an average annual MPI growth of 4.2%, primarily driven by technological progress (TC) rather than improvements in managerial efficiency (TEC) or scale optimization (SEC).

Regional disparities are evident: eastern provinces such as Jiangsu, Guangdong, and Zhejiang outperform central and western regions, indicating a need for more targeted policy support in less developed areas.

The constructed Green Finance Index (GFI) is shown to be positively and significantly correlated with investment efficiency, confirming the effectiveness of green financial instruments (e.g., green bonds, subsidies, carbon markets) in promoting productive investments in the renewable sector.

These findings contribute to the existing literature by offering a dynamic, regionally disaggregated evaluation of how green finance impacts new energy investment efficiency, highlighting both technological progress and institutional variation.

4.2. Policy Implications

Based on the empirical results, several actionable policy recommendations are proposed:

(1) Strengthen Regional Green Finance Ecosystems

Local governments should tailor green finance policies to regional industrial structures and financial capabilities. In underperforming provinces, capacity-building initiatives such as green finance talent programs, regional ESG disclosure mandates, and pilot subsidy platforms can help reduce institutional frictions.

(2) Incentivize Technological Innovation

Since technological change is the dominant driver of efficiency, policies that enhance R&D tax credits, streamline patent commercialization, and promote university-industry linkages are crucial. Moreover, support for emerging technologies such as green hydrogen, energy storage, and digital twin systems can further accelerate frontier productivity.

(3) Improve Access to Green Finance Instruments

The government should expand access to green credit and capital markets, particularly for SMEs in clean tech sectors. Mechanisms like green loan guarantees, interest subsidies, and carbon trading liquidity injections can reduce financing costs and risk perceptions.

(4) Monitor and Evaluate Policy Effectiveness Dynamically

Establish a green investment monitoring dashboard that tracks MPI components (TEC, TC, SEC) over time at the regional level. This would allow real-time policy adjustment and transparent public reporting, enhancing both accountability and investor confidence.

4.3. Limitations and Future Research Directions

This study has several limitations that open pathways for further research:

Data granularity: The current panel is constrained by data availability at the provincial level; future work can employ firm-level or project-level datasets for more precise efficiency attribution.

Policy heterogeneity: The constructed GFI aggregates multiple instruments, which limits the ability to distinguish the marginal effects of individual policies (e.g., green bonds vs. tax credits). Advanced econometric techniques or natural experiments can improve causal identification.

International comparisons: Extending the DEA-Malmquist framework to include cross-country comparisons could reveal how China’s green finance effectiveness compares globally, providing useful benchmarks for other emerging economies.

4.4. Final Remarks

As China progresses toward its dual carbon goals of carbon peaking by 2030 and carbon neutrality by 2060, ensuring that green capital is deployed efficiently will be paramount. This study confirms that green finance not only mobilizes investment but also enhances productivity in the renewable energy sector, especially when technological innovation is prioritized.

By aligning financial systems with environmental goals and correcting regional asymmetries in investment outcomes, policymakers can unlock the full potential of the green transition—ensuring both sustainability and economic resilience.