Can green credit policy promote green total factor productivity? Evidence from China

Green credit, a market-driven environmental policy instrument and an essential component of the green financial system, has piqued academic and policymakers’ interest in whether it has successfully improved Chinese green total factor productivity (GTFP). Utilizing Chinese province panel data from 2006 to 2019, this study assesses GTFP using the slack-based model with the Global-Malmquist-Luenberger technique and investigates the influence of green credit on GTFP as well as its mechanism. The findings suggest green credit has a favorable influence on China’s GTFP. Green credit can boost GTFP through three mechanisms: upgrading industrial structure, stimulating green innovation, and optimizing energy consumption structures. Furthermore, green credit improves GTFP in eastern regions but has little impact elsewhere; the promotion impact is more effective in financial developed regions and legal developed regions. As a result, the Chinese government should encourage regionally differentiated green credit policy implementation.


Introduction
The Chinese economy has experienced rapid development since 1978, when the reform and opening up took place. Due to industrialization and urbanization, a growing number of ecological spaces are being occupied, and the ecological environment continues to deteriorate. In particular, the concentrated outbreak of environmental problems under the extensive development model puts the issue of sustainable economic development in front of the government and the public (Ma et al. 2021). How to achieve economic growth while considering environmental protection and resource conservation has become a difficult problem that the government must face. China's financial system is a typical bank-based financial system, which determines that green credit is a pillar of China's green financial system and is becoming a key driver of a sustainable eco-civilization and green growth transformation. According to our estimation, the balance of green credit accounts for more than 90% of the total balance of green financing. Therefore, the research conclusion based on green credit can represent the role of China's green financial system to a great extent.
Green credit is a financial policy in which commercial banks and other financial institutions establish environmental access thresholds in their credit activities and implement differentiated pricing strategies for different enterprises (Cilliers et al. 2010). Specifically, commercial banks offer preferential low-interest loans for energy-efficient and environmental-friendly enterprises, apply punitively high interest rates on loans to high-pollution and high-energy-consumption enterprises ("two highs"), and withhold financing to enterprises undertaking new projects that fall into restricted categories or categories that have been eliminated. Green credit utilizes the allocation function of the financial market to direct capital to resource-saving and environmentally friendly enterprises at the micro level (Zhang et al. 2021a, b, c) and profoundly influences their business decisions and innovation behavior. At the macro level, green credit has the potential to encourage clean production, transform the Responsible Editor: Eyup Dogan * Jiawang Zhang zhjw@snnu.edu.cn 1 economic development paradigm (Yin et al. 2019), and create a virtuous loop between finance and the environment. In China, green credit started much later than it did in Western countries. In 2007, the Chinese government formally introduced "green credit" in a statement titled "Policies and regulations on environmental protection to prevent credit risks.: According to the "Green Credit Guidelines" issued by the China Banking Regulatory Commission in 2012, financial institutions should promote and expand credit support for green economies. It also emphasizes the importance of financial institutions focusing on environmental issues such as pollution and excessive energy usage, which may be brought on by their customers and their important related parties in the process of production and operation. The promulgation of the Green Credit Guidelines implies that the government's green credit strategy has been further strengthened and now serves as a guide for the banking industry and enterprises (Wang et al. 2020). In 2020, the central economic work conference proposed to innovate and develop green finance to help China achieve carbon neutrality by 2060. China's green finance has grown significantly under the guidance and assistance of the policies mentioned (Ren et al. 2020). China's green credit scale surpassed 11 trillion RMB by 2020, according to the Central Bank of China, making it the largest globally.
Green credit is essentially a significant expansion and innovation of standard administrative-mandated environmental legislation. While GTFP is a novel accounting methodology for total factor productivity that combines energy and environmental restrictions. Namely, GTFP is aligned with the original intention of finding a balance between man and nature. In academic studies, GTFP is an effective tool for assessing the sustainability of an economy's development over the long run (Zhang et al. 2021a, b, c). Thus, it is of great theoretical significance and practical value to explore how to improve GTFP. According to Porter's hypothesis, the environment and the economy can develop synergistically, ultimately achieving a win-win situation. So after more than a decade of practice, does China's green credit policy help to boost GTFP? If the answer is affirmative, what could be the underlying mechanism? These questions need to be explored in depth both theoretically and empirically.
In comparison to previous research, this paper's key contribution is to examine whether the green credit is promoting GTFP, which adds to the body of knowledge regarding the determinants of GTFP. This paper also offers a comprehensive exploration of the intrinsic mechanisms underlying the impact of green credit on GTFP from both a theoretical and an empirical perspective, which deepens the literature on green credit's economic effect. Furthermore, considering regional heterogeneity, this study specifically compares the effects of green credit on GTFP based on region between mid-western and eastern areas, financial underdeveloped and developed areas, and legal underdeveloped and developed areas. In this regard, the findings of this study can serve as a useful guide for relevant policies.
The remainder of this paper is structured in the following way. "Literature review and theoretical analysis" reviews the current literature and analyzes green credit's impact on GTFP. "Research design" details the econometric model, variable selection, and data sources for this paper. In "Empirical analysis," the empirical analysis is presented, along with benchmark tests and robustness checks. Discussions of mechanisms and heterogeneities are included in "Additional analysis." The conclusion and implications are summarized in the last part of this study.

Literature review
Academics are increasingly interested in studying green total factor productivity as it pertains to long-term sustainability. Since the GTFP is an effective tool for assessing the sustainability of an economy's development over the long run, the influencing factors of the GTFP have been widely studied by scholars (Feng and Serletis 2014). According to the existing literature, the influencing factors of GTFP can be divided into three main research categories: economic streams, technical streams, and government streams (Zhang et al. 2021a, b, c). Economic streams consist of the economic development level, market factors, production factors, and industrial structure (Feng et al. 2019;Liu and Xin 2019). Technical streams mainly include technological progress and efficiency (Wang 2017). Government streams consist of the infrastructure level, environmental regulation, and fiscal decentralization (Farhadi 2015;Ghosal et al. 2019).
The past literature on green credit has mostly focused on two points. To begin with, at the macro and meso levels, green credit policy was studied for its economic and environmental consequences (Dafermos and Nikolaidi 2021;Ngo et al. 2021). As far as economic effects are concerned, a study by Wang et al. (2021a, b), for example, demonstrated that green credit policies lead to considerable improvements in regional economic development. According to the results of a DSGE model developed by Liu and He (2021), green credit has noticeable effects on overall economic production and wellbeing. Hu et al. (2020) noted that the growth of green credit makes it easier to transform industrial structures via capital and financial channels. Researchers found that green credit can help boost local green innovation to a certain extent, but this influence has no spatial spillover effect (Guo et al. 2019). However, according to Zhang et al. (2011), the macro-level and meso-level implementation of green credit policy is ineffective. Among China's main challenges in implementing its green credit policy are the wide-ranging influence of "two high" enterprises, imprecise implementation standards, and vague policy details. In terms of environmental effects, the research conclusions are quite unified. Studies by Ren et al. (2020) and Song et al. (2021) revealed that green credit policy has resulted in considerable improvements in air and water quality, carbon dioxide emission reduction, and high-efficiency energy utilization.
In the second research direction, companies and financial institutions are examined from the micro perspective of green credit policy implementation. By requiring banks to decrease loans to two high enterprises, the green credit policy directly impacts firms' ability to raise debt . Specifically, as a quasi-natural experiment, researchers have found an increase in debt financing costs as well as a decrease in loan maturity for two high enterprises (Xu and Li 2020). Furthermore, reduced long-term debt in polluting companies due to the Green Credit Guidelines has resulted in a considerable reduction in R&D input and innovation output (Ling et al. 2020). Green credit policy was also found to have a detrimental influence on total factor productivity and enterprise performance of two high enterprises (Wen et al. 2021). With regard to banks, scholars have found a causal link between the financial performance and banking sector sustainability in China (Weber 2017;Chatzitheodorou et al. 2021). Scholars have also discovered that green loans are less risky than non-green loans (Cui et al. 2018). In conclusion, implementing green credit can increase bank profits while lowering credit risk (Guan et al. 2017;Yin et al. 2021;Miroshnichenko and Brand 2021).
In conclusion, most of existing studies examined the green credit on micro-agent decision-making. Some literature explores the economic effects or environmental effects of green credit. However, few researches have investigated the connection between green credit and GTFP as well as their potential impact mechanisms. Given the gaps above, this study tries to deconstruct the internal logic of green credit and GTFP and conduct a systematic analysis in terms of what kind of relationship exists and the possible impact mechanisms and heterogeneity across provinces. A path to sustainable economic development is to improve GTFP. Therefore, this study not only expands the economic effect of green credit but also enriches the related research on the influencing factors of GTFP. The findings will also provide a scientific basis for policy makers to formulate measures for promoting GTFP and help to realize the sustainable development of economy and environment.

Theoretical framework: green credit and GTFP
Based on the macro perspective, green credit mobilizes capital accumulation through differentiated financial policies, such as credit inclination and floating interest rates, to form green investment and provide capital elements for economic growth. Green credit policy encourages lenders to take into account all environmental risks associated with loan projects when making lending decisions. That is, loans are not granted to projects with significant pollution and energy consumption, but rather to projects that save energy and are environmentally benign. This can often achieve large-scale capital concentration within a short length of time and help spur investments in green industries and related emerging industries, as well as cultivate new economic growth points ). In addition, the decrease in financing costs of green projects encourages funds to flow to green projects with high efficiency, energy conservation, and low pollution, which enhances the quality of economic growth by optimizing the economic structure. In summary, using capital scale-oriented and capital cost-oriented mechanisms, green credit facilitates the "green allocation" of financial resources to realize the green transition of the economy.
Based on the micro perspective, through reviewing the existing literature, we believe that green credit can improve the enterprises' green efficiency by reducing transaction costs, lowering the risk of enterprise green innovation, and conducting pre-loan inspection and post-loan supervision on the energy-saving and emission-reduction effects of borrowing enterprises. First, based on the green requirements of regulators and their own profit-seeking needs, financial institutions professionally collect and process enterprise information and look for low-risk, high-yield projects that meet green standards for loans ). These measures can alleviate adverse selection and moral hazard and reduce the transaction costs of green enterprises (Yin et al. 2021). Second, enterprises need to invest significantly in green transformation and often face greater risks. In the early stage of green transformation, traditional finance cannot play its role and bring direct economic benefits to enterprises. Green credit can provide sufficient financial support for green R&D, transformation, and applications, thus reducing the risk of enterprise green transformation . Finally, financial institutions strictly screen the energy-saving and emission-reduction potential of green loan applicants and track and supervise the energy-saving and emission-reduction effects after the green credit is released. Enterprises with good energy-saving and emissionreduction effects can appropriately increase loans (Luo et al. 2021). For enterprises with poor rectification, loans can be stopped in time to increase the allocation of funds allocated. It should be pointed out that there is a one-to-one correspondence between the above three micro-level functional paths and the following three assumptions put forward in this paper at the meso-level, that is, the three influencing mechanisms of green credit promoting GTFP that this paper is going to verify. On the basis of the foregoing analysis, we hypothesize: Hypothesis 1. Green credit can promote green total factor productivity. Furthermore, through the above comprehensive analysis from macro and micro perspectives, we believe that green credit influences regional GTFP via industrial structure upgrading, green technology innovation, and energy consumption structure optimization. Among them, the mechanism of promoting industrial structure upgrading and optimizing energy consumption structure comes from the design of green credit policy itself. Green credit policy requires financial institutions to invest in green and low-energy industries while the green technology innovation mechanism is based on the pecking order theory and the debt heterogeneity theory. The pecking order theory points out that the large amount of capital investment, long duration, and high risk of innovation activities form a sharp contradiction with the risk aversion preference of creditors and thus deduces the view that debt financing is not conducive to stimulating innovation (Myers and Majluf 1984;Opler and Titman 1994). The debt heterogeneity theory indicates that relational creditors such as commercial banks have the characteristics of innovation and tolerance, pay attention to the long-term growth of enterprises and the relationship between banks and enterprises, and can achieve a win-win situation between banks and enterprises by promoting technological innovation of enterprises. Trading creditors, such as bondholders, pay more attention to the fluctuation of bond prices and thus lack motivation to pay attention to corporate innovation activities (David et al. 2008). Figure 1 displays the precise impact routes. Green credit facilitates the improvement of industrial structure by promoting capital formation, capital orientation, and information transmission. First, finance can gather idle funds in society to form industrial capital through savings and provide capital support for green industry growth (Fan et al. 2021). Second, green credit steers funds towards green companies by managing the volume and direction of loan supply to eliminate highenergy-consuming and pollution-producing projects, and finally complete green and advanced industrial structure transformation. Finally, financial institutions use their professional ability in information collection, analysis, and evaluation to identify potential investment projects.
Green projects with both investment value and environmental benefits are presented to investors to promote the optimal allocation of social resources. Industrial structure upgrading can promote resource conservation, drive a continuous increase in product quality as well as added value, and bring about an improvement in economic benefits (He et al. 2019). Moreover, the transition from industrialization to a service-oriented economy means the industry structure shifts from a laborintensive to a capital and technology-intensive. This changes the previous extensive development model, made full use of resources, reduced environmental pollution, and promoted the improvement in GTFP. On the basis of the foregoing analysis, we hypothesize: Hypothesis 2. Green credit improves GTFP by promoting industrial structure upgrading. Green credit guides capital flow towards environmental friendly areas through incentive and restraint mechanisms and profoundly affects the innovative behavior of relevant enterprises. For two high enterprises, a higher loan rate increases the financing cost. Moreover, due to high emissions, such firms come to face greater pressure from public opinion and moral condemnation, and the risk of corporate debt default increases, resulting in banks refusing or reducing loans to the two high enterprises (Nabeeh et al. 2021). This results in a decline in cash available to such enterprises for green innovation, which is not favorable to improving regional green innovation levels. "Clean" enterprises benefit from green credit because it provides them with low financing costs and abundant funds, alleviating their funding shortages and assisting in the advancement of regional green innovation. Besides its direct influence, green credit also indirectly affects the development of green technologies. Whether a company can actively conduct green innovation is heavily dependent on its investment in R&D and human capital, which is a strategic choice that is made after fully weighing the sacrifice of current economic interests for future ecological value. As the green credit guidelines set the threshold for enterprise environmental protection, when expanding the production scale, clean enterprises spend more funds on updating production equipment and green technology process development (Hong et al. 2021). This further enhances enterprises' capacity for green tech- Fig. 1 Impact mechanisms of green credit on GTFP nology innovation. The two high enterprises may also increase R&D investment. This is because the essence of green credit is a crucial innovation of conventional environmental legislation. Therefore, green credit has an innovative compensation effect ). High interest rates will force companies to improving their green innovation ability, thus promoting GTFP. On the basis of the foregoing analysis, we hypothesize: Hypothesis 3. Green credit improves GTFP by promoting green innovation. Green credit policy supports industrial enterprises in carrying out traditional energy transformation, promoting the green and low-carbon energy consumption structure, and encouraging renewable energy utilization, thereby guiding energy consumption structure optimization. The shift in energy consumption structure is a game between traditional energy consumption and new energy consumption (He et al. 2019), from the standpoint of the substitution effect. Under green credit policy, energy-related industries' supply conditions and prices are expected to change, and at the same time, the preference of energy consumption departments for renewable energy and clean energy will increase (Taghizadeh-Hesary and Yoshino 2020). Over time, consumers will decrease their use of fossil fuels and increase the demand for new energy (Liu et al. 2017), therefore optimizing energy consumption structures for their own heath, environmental welfare, and sustainable development. The impact of the energy consumption structure on green economy growth is reflected in two aspects. As an input element, energy can contribute to the scale and speed of economic growth. However, the increase of coal consumption will eventually result in environmental degradation and lower social welfare. As a result, reducing coal consumption will aid in the promotion of green economic growth. On the basis of the foregoing analysis, we hypothesize: Hypothesis 4. Green credit improves GTFP by optimizing energy consumption structure.

Empirical model
In this paper, we examine the impact of green credit on GTFP. Considering the different conditions in each province and the large time span, the following two-way fixed effects model is constructed for empirical research.
For formula (1) and the following formulas, i denotes the province and t denotes the time. GTFP refers to green total factor productivity. Gcredit represents green credit. X reflects control variables affecting GTFP. According to Wang et al. (2021a, b) and Song et al. (2021), the control variables include human capital, urbanization, foreign direct investment, fiscal expenditure scale, environmental regulation, economic scale, and financial development. α 0 , α 1 , and β are estimated coefficients. λ i , γ t , and ε it are region fixed effect, year fixed effect, and random perturbation term.

Dependent variable
GTFP is the combined efficiency that comprehensively considers the production factor inputs, resource consumption, and environmental costs. This study assesses GTFP using the slack-based model with the Global-Malmquist-Luenberger technique in the benchmark regression. The GTFP evaluation index system includes three types of indicators: input, desired output, and non-desired output (Li and Chen 2021;Luo et al., 2022). Table 1 lists the detailed evaluation index system of GTFP.

Main independent variable
At present, the measurement of green credit in the academic field mainly includes four measures: the proportion of green credit from financial institutions, the proportion of loans for energy-saving and environmental protection projects, the green credits in industrial pollution control investments, and the proportion of interest expenses in the six high energy-intensive industries. The first two measures are derived from the social responsibility reports of five large commercial banks and some joint-stock banks in China. Scholars use this data to study from the national level. The green credits in industrial pollution control investments data will not be counted after 2010. Thus, the first three measures should not be adopted. Considering the continuity and availability of data and based on the existing literature (Hu et al. 2020), this paper selects the ratio of interest expense of six major energy-intensive industries in each province to the total interest expense of industrial industries as a reverse indicator to measure green credit (Gcredit). Gcredit is the reverse indicator of green credit. A higher implementation intensity of green credit policy means less credit is given to heavily polluting industries, resulting in a smaller Gcredit. Compared with the total amount of loans in heavily polluting industries, the loan interest can better reflect the borrowing cost of funds, and the realization of differentiated borrowing costs is the basic feature of green credit policy (Zhang et al. 2021a, b, c).

Control variables
Based on the previous research (Li and Xu 2018;Wang et al. 2021a, b), this paper has controlled the other variables that affect GTFP in the econometric model. Economic scale (GDP): some scholars believe that the current economic growth is achieved at the expense of environment, which has a negative impact on GTFP. However, some economists think that with economic growth, people's environmental awareness will be enhanced and therefore have a positive impact. Thus, this index is introduced in the econometric model and measured with the natural logarithm of real GDP. Human capital (HUM): the promotion of human capital is beneficial to improve the efficiency of the use of technology, thus enhancing the GTFP. The average years of education are used to calculate human capital. In China, primary school lasts for 6 years, middle school lasts for 9 years, high school lasts for 12 years, and college lasts for 16 years. Thus, Human = H1 × 6 + H2 × 9 + H3 × 12 + H4 × 16. In the formula, H1, H2, H3, and H4 represent the proportion of the total population in each province with primary, middle, high school, college, and above education levels. Urbanization rate (URB): it is generally believed that with the advancement of urbanization, the urban development mode gradually adopts low-carbon and green development mode, thus promoting the overall sustainable development process and realizing the promotion of GTFP. This variable is equal to the share of the population that lives in urban areas. Foreign direct investment (FDI): the pollution heaven hypothesis assumes that polluting or declining industries are transferred to low-income countries, which will reduce environmental welfare of host countries. Meanwhile, there is another hypothesis that "demonstration effect" and "learning effect" can bring advance techniques to host country through technology spillover effect. This variable is quantified by dividing actual foreign investments by GDP. Fiscal expenditure size (EXP): when the financial expenditure is used for education investment and infrastructure improvement, it will be beneficial to the promotion of regional GTFP. However, when fiscal expenditure is used for administrative management, it may lead to the distortion of resource allocation, thus resulting in the loss of efficiency. Fiscal expenditure is quantified by fiscal spending as a percentage of GDP. Environmental regulation (ENV): in order to establish an ecologically balanced economic system and reduce environmental pollution emissions, government has continuously strengthened environmental regulation. Environmental regulation is quantified by the percentage of GDP spent on environmental protection. Financial development (FIN): it is generally believed that the regions with higher financial development level have more developed service industry and higher resource allocation efficiency, which is helpful to promote GTFP. Financial development is calculated by dividing the proportion of a financial institution's loan balance by GDP.

Sample and data
Due to the fact that the green credit data began in 2005, and to ensure that the analysis is comprehensive, the research sample here is 30 provinces (excluding Tibet, Hong Kong, Macao, and Taiwan)

Benchmark regression
GTFP in China is obtained using the selected input-output indicators (Fig. 2). In 2005, China's GTFP was 1.02. Affected by the international financial crisis, in 2009, China's GTFP plunged to its lowest value during the sample period (below 1), reflecting the retrogression of GTFP. After that, the value of China's GTFP rose in fluctuation. In 2019, GTFP reached the highest value in the sample period, exceeding 1.05.
The Hausman test is required before regression analysis to confirm whether fixed effects model should be adopted. In this paper, the Hausman test value of the fixed effects and random effects model is 60.45, and it was significant at 1% level, indicating that the fixed effects model should be selected instead of the random effects model. Thus, this study uses the fixed effects model.
The impact of green credit on GTFP is shown in Table 3. The empirical results without any control variables are presented in the first two columns, while the results after controlling for other variables are reported in the last two columns. Individual fixed effect results are tabulated in columns (1) and (3), while columns (2) and (4) are regression results with two-way fixed effects models. As shown in Table 3, Gcredit was negatively associated with GTFP, and it was significant at least at the 5% level. In the twoway fixed effects model with all control variables included, for every 1% increase in Gcredit, the level of GTFP will drop by 0.038%. Because Gcredit is an inverse indicator of green credit, this finding preliminarily verifies Hypothesis 1, namely, that green credit significantly promotes GTFP.
For the control variables, the coefficients of HUM, URB, and FIN are positive and pass the significance test at least at the 5% level. This shows that the improvement in human capital, urbanization rate, and financial development level can improve GTFP. The coefficients of GDP and EXP are significantly negative at 1%. This indicates China's economic development has not yet achieved complete transformation, and rapid economic growth still depends on energy consumption and brings a certain degree of environmental pollution. Furthermore, fiscal expenditure is used more for administration, resulting in resource allocation distortions and efficiency losses. Finally, the coefficients of environmental regulation and FDI are not significant, suggesting that they are not core factors affecting GTFP. These results are generally consistent with previous research.

Replacing the dependent variable
In the robustness test, this paper recalculates the GTFP (GTFP2) using the Epsilon-based model with the Global-Malmquist-Luenberger technique (EBM-GML) proposed by Tone and Tsutsui (2010). The EBM-GML model is able to consider not only the radial proportion between the target and the actual values, but also deal with the relaxation between input factors and output factors, which enhances the comparability of decisionmaking units. Columns (1) and (2) of Table 4 display  (2) is formed by adding the first-order lag term of GTFP (GTFP it-1 ) to formula (1). From the econometric model in this paper, there may exist endogeneity problems caused by reverse causality among the explained variables, each explanatory variable, and some control variables. For example, green credit affects GTFP, and the level of GTFP may also have a significant impact on green credit. Therefore, for the purpose of overcoming the possible endogeneity issue and further verifying the robustness of the findings, two dynamic panel estimation methods are used: the differential generalized method of moments estimation (Diff-GMM) and the systematic generalized method of moments estimation (Sys-GMM). The lags of green credit were used as instrumental variables for Diff-GMM and Sys-GMM estimation.
Results of dynamic panel model regression are shown in columns (3) to columns (4) of Table 4. It can be seen that the regression coefficient of Gcredit is significantly negative whether using Diff-GMM estimation or Sys-GMM estimation. This verifies that green credit can promote GTFP. At the same time, the results of autocorrelation test and Hansen test also show that the model's instrumental variables are effective; there is no over identification problem, and the endogeneity problem is well controlled to a certain extent. In conclusion, the above-mentioned double robustness tests confirm that this paper's conclusions are solid and credible.

Influence mechanism
According to the research hypothesis, the stepwise regression method is used in this paper to identify the mediating effects of industrial structure, energy structure, and green technology innovation on GTFP. The testing process is as follows: First, the influence of green credit on GTFP is examined: Second, we examine the influence of green credit on the mediating variables (Mediator): Third, we simultaneously include the green credit and mediating variables in the regression model to test their impact on GTFP: According to models (3)-(5), the premise of the mediating effect is that the coefficient of β 1 is significant. On this basis, we continue to investigate the significance of coefficients β 2 and θ. If both β 2 and θ are significant, this indicates the presence of a mediating effect. In addition, the mediating variable plays a complete mediating effect if β 3 is not significant; otherwise, it plays a partial mediating effect. According to model (3), β 1 is significantly negative, in line with the benchmark regression in this paper. Therefore, the following will mainly test model (4) and model (5).
(2) According to , the ratio of tertiary industry's added value over secondary industry's added value is the measure of industrial structure upgrading (Advance). Columns (1) and (2) in Table 5 analyze whether green credit can improve GTFP by promoting industrial structure advancement. The coefficient of Gcredit in column (1) is − 7.988, significant at 1%, indicating that green credit can help with industrial structure upgrading. In column (2), after adding industrial structure advancement, the coefficient of Gcredit is still significant. Whereas the coefficient of advance is 0.001, significant at 5%, this indicates that industrial structure advancement has a positive and partial mediating influence between green credit and GTFP. Green credit forms industrial capital by gathering funds and guides the funds to flow to high-quality green projects, thus promoting the upgrading of industrial structure and the growth of green economy. Therefore, hypothesis 2 is verified.
In accordance with Li et al. (2022), green innovation (Innovation) is quantified by calculating the logarithm of green patent applications by province. The mediating effect test results of green technology innovation are shown in columns (3) and (4). The coefficient of Gcredit is − 0.178 in column (3), significant at 5% level, illustrating that green credit has the potential to stimulate green innovation. In column (4), the coefficient of Gcredit is no longer significant, while the coefficient of innovation is positive at 5%. Green credit guides capital flow to green enterprises through incentive and restraint mechanism, relieves financing constraints of green enterprises, and helps to improve green technology innovation investment of green enterprises. At the same time, the innovative compensation effect of green credit also promotes the two high enterprises to increase R&D investment, thus promoting regional GTFP. The above results indicate that green innovation has a positive and complete  (1) and (2), t-values in parenthesis are provided. For columns (3) to (6), Z-values are shown in brackets below explanatory variables; P values are shown in parentheses below the AR (2) and Hansen test * Significant at the 10% level; **significant at the 5% level; ***significant at the 1% level mediating influence between green credit and GTFP, which verifies hypothesis 3. The energy consumption structure (Energy) of each province is calculated based on the coal consumption percentage of total energy consumption in this study. Columns (5) and (6) of Table 5 analyze whether the energy consumption structure can be optimized so that green credit promotes green total factor productivity. For column (5), the coefficient of Gcredit is positive at 1%, illustrating that green credit can effectively decrease the proportion of coal consumption. In column (6), the coefficient of Gcredit is − 0.032, and the coefficient of Energy is − 0.078, both significant at least at the 5% level. Clearly, the increase in the energy consumption structure defined by the share of coal use has a considerable negative impact on China's GTFP. Green credit policy promotes GTFP by supporting industrial enterprises to transform traditional energy, encouraging renewable energy utilization, reducing coal consumption, and promoting green and low-carbon energy consumption structure. The above results indicate that energy structure optimization has a positive and partial mediating influence between green credit and GTFP, which verifies hypothesis 4.

Location heterogeneity
The driving effect of green credit on GTFP may vary due to different geographical conditions. The natural environment, resource endowments, and industrial bases of eastern China and mid-western China are clearly different. Therefore, the 30 provinces in China can be divided into eastern and mid-western regions. Columns (1) and (2) in Table 6 display the regression findings for the location heterogeneity of green credit on GTFP. As can be seen, the coefficient of Gcredit in the eastern regions is − 0.062 and significant at 1%. In contrast, the coefficient of Gcredit does not pass the significance test in the mid-western regions.
The following are some possible explanations: First, the eastern area has a greater level of industrialization (Xu and Li 2020). There are more polluting firms that are restricted as a result of the green credit policy. Therefore, enterprises in eastern regions are more prone to using green credit. Second, in the east, environmental regulation is more strict and extensive than in the midwest (Guan et al. 2017). As a compensatory environmental regulation, green credit policy is more likely to be favored by enterprises in the eastern regions. Green credit can offset part of the cost increase brought by administratively mandated environmental regulation. Third, although the enterprises in the eastern regions have a greater capacity for innovation and fewer constraints on innovation financing (Zhang et al. 2021a, b, c), the market is more competitive. Obtaining green finance support may send favorable signals to the whole market and increase enterprises' motivation to innovate green and maintain their competitive advantage (Taghizadeh-Hesary and Yoshino 2019). Overall, in the eastern regions, the transmission efficiency of green credit is higher than that in the mid-western regions thanks to industrial restructuring, energy structure optimization, and green innovation.

Differences in financial development level
According to the median financial development level within each province, the sample is classified into financial developed and underdeveloped areas. Table 6 shows the results of grouped regression in columns (3) and (4). In the financial developed regions, the Gcredit coefficient is − 0.075 and strongly significant at 1%. The coefficient of Gcredit in the financial underdeveloped regions is − 0.029 and significant at 10%. After using Fisher's permutation test, it is found that Gcredit coefficients differ significantly between the two groups at the 1% level. This shows that green credit is more successful at boosting GTFP in financial developed regions.
A developed financial market is one of the basic conditions for green credit to exert its policy effect. Compared with financial developed regions, green credit policy is more likely to be a public welfare policy in financial undeveloped regions since institutions there lack adequate motive to engage in green credit activity, so the implementation effect of the policy in financial underdeveloped regions is relatively poor. The signal of energy-conservation and emission-reduction transmitted by green credit is not effective enough, and the financing constraints on polluting enterprises are limited. Therefore, although they all have certain effects, in financial developed area, green credit is implemented more efficiently, and the effect on the improvement in GTFP is more obvious. As the eastern financial markets are more developed than those in the mid-western area, this conclusion is also consistent with the results of regional heterogeneity.

Differences in legal development level
The legal development level data adopts the index of market intermediary organizations and legal system environment of each province in China Marketization Index Report. According to the median legal development level within each province, the sample is divided into legal developed and underdeveloped regions. Table 6 displays the results of grouped regression in columns (5) and (6). In the legal developed regions, the Gcredit coefficient is − 0.066, and strongly significant at 1%. The coefficient of Gcredit in the legal underdeveloped regions is − 0.013 and significant at 10%. After using Fisher's permutation test, it is found that Gcredit coefficients differ significantly between the two groups at the 1% level. This shows that green credit is more successful at boosting GTFP in legal developed regions.
The premise of implementing green credit policy is to have a sound legal environment. In legal developed regions, a sound legal environment can provide a good external environment for financial institutions to carry out green credit business. A sound legal system covers the whole process of prevention, supervision, and punishment of green credit, ensuring the smooth implementation of green credit policies and the flow of funds to low-carbon and environmental protection fields. Therefore, although all of them have certain effects, the implementation efficiency of green credit is higher, and the improvement effect on GTFP is more obvious. As the provinces with high level of legal development have high level of financial development, and mostly located in the eastern region, this conclusion is consistent with the results of regional heterogeneity and financial development level heterogeneity.

Conclusions and policy implications
Based on Chinese provincial data from 2005 to 2019, this study assesses GTFP with the SBM-GML technique and investigates the influence of green credit on GTFP as well as its intrinsic mechanism using fixed effect model. The following conclusions can be drawn: (1) Green credit can significantly promote GTFP; the double robustness test also supports this conclusion. (2) In terms of the influence mechanism, green credit improves GTFP by elevating the advanced level of industrial structure, improving the ability of green innovation and reducing the proportion of coal consumption in energy consumption. (3) There exists significant regional heterogeneity in the impact of green credit on GTFP. By distinguishing the geographical locations, we find that green credit has a beneficial influence on GTFP in the eastern regions, while the impact is not evident in the mid-western regions. By distinguishing the level of financial development and legal development, this paper finds that the promoting effect of green credit on GTFP is more significant and effective in financial developed regions and legal developed regions. The theoretical contribution of this paper is to systematically explain the impact mechanism of green credit on GTFP. This paper puts forward and verifies that green credit can promote GTFP through three channels: upgrading industrial structure, stimulating green technological innovation, and optimizing energy structure. This study theoretically complements the theoretical research on green finance and enriches the related research on the economic consequences of green debt financing and the influencing factors of GTFP. The practical contribution of this paper is to provide scientific basis for decision-makers to formulate measures for promoting GTFP and help to realize the sustainable development of economy and environment.
Closely combined with the research findings in this paper, the following policy implications are put forward. First, green credit policy has already had a phased effect on GTFP promotion. The government should timely expand the scale of green credit, keep the credit threshold under close control, and maintain the policy's sustainability and stability. Green credit policy is an important strategic tool for China to develop a green economy in the new era and is also an important supplement to the traditional administrative mandatory environmental regulations. With the deepening of China's market-oriented reform, the green credit policy will play a greater role in the promotion of GTFP. We believe that the implementation experience of green credit in China is also very suitable for other countries and regions.
Second, financial institutions should continue to promote flexible and differentiated green financial policies to help green credit play its role in promoting industrial transformation and upgrading, stimulating enterprises to carry out green innovation, and reducing the proportion of coal consumption in the energy consumption structure. From the perspective of promoting industrial structure upgrading, financial institutions need to strictly implement the differential interest rate policy, which is the most basic and critical step to promote green credit to really play a role in industrial structure upgrading. On this basis, the docking platform of green credit data can be built between the government, banks, and enterprises, and various industrial financial policies and services can be efficiently integrated, so as to realize the benign flow of green credit funds among industries and make green credit funds play their due role in supporting industrial transformation and upgrading. In addition, through the publicity and guidance of policies related to green credit, it is necessary to ensure that commercial banks assess the green projects of enterprises in a timely manner, and also ensure that the green credit funds obtained by enterprises are used for the transformation and upgrading of industrial structure.
From the perspective of stimulating green technology innovation, the first is to encourage financial institutions to carry out linkage business of investment and loan aiming at green technology innovation, so as to enhance banks' willingness of credit support and preferential credit for green technology innovation. Second, the government should provide guidance and incentive mechanisms, including incubation, guarantee, and discount interest, to reduce the financing cost and risk premium of green technology enterprises and open up the channel of green finance to reduce the financing cost of green innovation. Third, the government should establish evaluation standards and information disclosure standards suitable for green technology projects, reduce the evaluation cost of financial institutions for such projects, and encourage financial institutions to provide lower-cost financial support for green innovation.
From the perspective of optimizing the energy consumption structure, the development of green finance is conducive to promoting China's investment and financing in the field of energy conservation and emission reduction, which can promote the development and utilization of green low-carbon energy by enterprises, reduce the consumption of highly polluting energy such as coal and oil, and make China's energy structure increasingly diversified and low-carbonized. Therefore, in terms of credit policy, financial institutions should continuously increase the support of credit policy for clean and renewable energy industries such as hydropower, wind power, and photovoltaic power generation. In terms of authorization policy, financial institutions should give differentiated authorization arrangements for green credit to improve the time limit for credit approval.
Third, to maximize the boosting effect of GTFP, distinct green credit policies can be adopted in various regions. For the eastern regions, financial developed regions, and legal developed regions, green credit is crucial to the promotion of GTFP. As a result, green credit policies in these areas need to be strengthened appropriately. In contrast, for the mid-western regions, financial underdeveloped regions, and legal underdeveloped regions, traditional green credit has a limited promotion effect on GTFP. All green financial entities should cooperate to explore new green credit pilot projects. For example, Bank of Jiangsu has cooperated with China Clean Development Mechanism Fund Management Center, Jiangsu Provincial Department of Finance, Jiangsu Provincial Department of Ecology and Environment, and Jiangsu Provincial Department of Water Resources to develop characteristic innovative products such as green innovation portfolio loan, environmental protection loan, and water saving loan, which have achieved good results. Financial institutions in the mid-western regions can follow the local characteristics, develop pilot green loans, and innovate green credit products, thus solving the problem of green transformation of local enterprises.
In addition, the mid-western governments and financial institutions should cooperate to build a diversified green financial product system. Apart from innovating green credit products, financial institutions could also speed up the innovation of green bonds, green insurance, green funds, and other instruments and improve service efficiency, so as to achieve the coordinated development of the breadth and depth of the green financial system and better serve the green economic growth through the diversified development and multi-level construction of the green financial system.

Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations
Ethics approval This study is ethically approved. The contents of this manuscript will not be copyrighted, submitted, or published elsewhere, while acceptance by Environmental Science and Pollution Research is under consideration. There are no directly related manuscripts or abstracts, published or unpublished, by any authors of this paper.

Consent to participate
All the authors of this study have directly participated in the planning, execution, or analysis of this study.

Consent for publication
All the authors of this paper have read and approved the final version here submitted.

Competing interest
The authors declare no competing interests.