BUSINESS( NO PLAGARISM A+ WORK, ON TIME)

Environmental Management (2020) 66:1059–1071 https://doi.org/10.1007/s00267-020-01346-w

Green Credit, Financial Constraint, and Capital Investment: Evidence from China’s Energy-intensive Enterprises

Yanli Wang1 ● Xiaodong Lei1 ● Ruyin Long1

● Jingjing Zhao1

Received: 22 February 2020 / Accepted: 30 July 2020 / Published online: 13 August 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020

Abstract The green credit policy is an important green financial tool that can achieve the win–win scenario with economic development and environmental protection through the reasonable allocation of credit resources. Using the green credit guidelines (GCGs) in China as a quasi-natural experiment, this study explored the impacts of the green credit policy on the capital investment of energy-intensive enterprises in a difference-in-differences framework and established the mediation effect model to analyze the mechanisms. The empirical results showed that the capital investment of energy-intensive enterprises was significantly reduced after the promulgation of the GCGs. Considering the intermediary paths along with the green credit policy on energy-intensive investment through financial constraints, the total bank loans and long-term bank loans played partial intermediary roles, whereas the short-term bank loans as mediator variable showed no significant intermediary effect. The findings of this study illustrated that the green credit policy has been well implemented and promoted in China. It inhibited energy-intensive investment, which is of great significance to improving the efficiency of resource utilization and promoting green and low-carbon development.

Keywords Green credit guidelines ● Financial constraint ● Capital investment ● Difference-in-differences ● Mediation effect

Introduction

In recent years, environmental problems such as excessive pollution emissions, energy depletion crisis, and severe climate change not only pose threat to human beings’ health and survival, but also hinder the sustainable development of economy (Chung et al. 2005; Hoang et al. 2019; Hu 2016; Liu et al. 2016; Swanson et al. 2001; Weaver and Miller 2019; Ye 2016). Hence, how to find effective solutions to balance economic performance and environmental protec- tion has become an urgent issue for all countries in the world. As the world’s second largest economy, China has contributed significantly to the prosperity of the global economy while accounted for a large share of energy con- sumption and carbon dioxide emissions (International Energy Agency 2019). Therefore, whether China can suc- cessfully deal with environmental problems will play an

important role in the process of green and low-carbon development of the world. China has introduced a series of environmental policies to boost green development. Speci- fically, the environmental regulation tools can be roughly divided into three categories: the first is command-and- control method (Du and Li 2019; Feng and Li 2020; Pang et al. 2019). The second are market-oriented means char- acterized by incentives (Chang et al. 2020; Ren et al. 2020), and the third is information disclosure such as environment- related corporate social responsibility (CSR) reporting1. Among them, the green credit policy implemented in China encourages banks to allocate credit resources on the basis of enterprises’ information disclosure (Sun et al. 2019), and it also combines government supervision with the market incentives, which has attracted widespread attention in the practice of environmental governance. The green credit policy in China can be traced back to the Notice on Issues Related to Implementing Credit Policy and Strengthening Environmental Protection Work in 1995, which required financial institutions to fully take environmental factors into account when implementing credit policies (The People’s* Ruyin Long

[email protected]

1 School of Economics and Management, China University of Mining and Technology, Xuzhou 221116 Jiangsu, PR China

1 We thank the anonymous reviewers for constructive suggestions about the category of environmental regulation policies.

12 34

56 78

90 () ;,:

12 34 56 78 90 (); ,:

Bank of China 1995). In February 2012, the former China Banking Regulatory Commission (CBRC) issued the green credit guidelines (GCGs), which made specific requirements and arrangements for green credit policies and clearly pro- posed to adjust the credit structure in order to prevent environmental risks (China Banking Regulatory Commis- sion 2012). The promulgation of the GCGs is a milestone, which marked the transition of China’s green credit policy to a new phase (Liu et al. 2019). By the end of 2018, the green credit balance of Chinese financial institutions exceeded 8.23 trillion yuan, with a year-on-year increase of 16%. Currently, the green credit policy has become a powerful driving force in the process of China’s ecological civilization (Ng 2018; Xie and Liu 2019).

Theoretically, the green credit can guide credit resources away from high-pollution and energy-intensive industries through stringent credit management, which will promote energy conservation and emission reduction. Furthermore, it also plays a significant role in adjusting the industrial structure and changing the mode of China’s economy. Hence, the green credit policy shows promise to balance the contradiction between economic targets and environmental protection goals. However, the results of existing research on the effectiveness of green credit policy are mixed. One line of thought argues that financial institutions as the main body of the implementation of green credit, will not fully implement green lending due to the lack of economic incentives (Biswas 2011), which is not conducive to rea- lizing the dual objectives of economy and environment. A second idea contends that for the purpose of improving capital security and obtaining differential competition (Zhang et al. 2011), financial bodies will proactively follow green credit principles, reduce the credit allocation of heavy-polluting industries (Su and Lian 2018), and curb the expansion of energy-intensive industries (Liu et al. 2017), which can improve environmental quality to some extent (Cai et al. 2019; Sun et al. 2019). Moreover, to the best knowledge of us, most of the research based on the micro- enterprise level prefer to focus on the financial constraint resulting from the green credit policy. However, curbing high-pollution and energy-intensive companies’ financing is only a means, but not the ultimate goal. The main purpose of green credit policy is to constrain resource allocation through financial instruments, eliminate backward produc- tion capacity, and achieve green economic development (Zhu et al. 2020). In addition, green credit policy does not require the direct closure of energy-intensive enterprises. Instead, it hopes to use credit constraints to force such enterprises to exit projects that may cause major environ- mental problems. Therefore, the interesting topic worthy of further study is that, when it comes to the financial con- straints imposed by the green credit policy, will the energy- intensive companies respond as expected to exit investment

projects that do not conform to the green principle? This remains to be verified by more supportive evidence.

In order to deal with the above important but rarely discussed issue, the present study takes the introduction of the GCGs in China as a quasi-natural experiment to investigate the impacts of green credit policy on capital investment of energy-intensive companies through difference-in-differences (DID) method, using China’s A- share listed companies as the research samples. On this basis, the mediating effect models are employed to examine the potential mechanisms, with the total bank loan, long- term bank loan and short-term bank loan as the mediators, respectively. The empirical results showed that the GCGs had inhibitory effects on the capital investment of energy- intensive enterprises through debt financial constraints, indicating that enterprises have taken active respond to green lending policy, which will positively boost the sus- tainable development in future.

The possible contributions of this study are twofold as follows. First, the present study enriches the existing research on the effectiveness of green credit policy. As some scholars point out that, with respect to the scanty research on quantitative, qualitative analysis, and especially the empirical studies about green credit, it is not enough to facilitate the understanding of the green lending behavior between financial institutions and enterprises (Wang et al. 2019), and thus is not conducive to timely adjustment and optimization of policies. On the basis of the existing lit- erature, with dataset of China’s enterprises as research sample, this paper can add evidence for the effect of green credit to some extent. Second, only few scholars have researched the mechanism between the green credit policy and corporate investment behavior from the perspective of financial constraints. Su and Lian (2018) and Ding (2019) study the impact of green credit policy on the investment and financing of polluting firms, without considering the impact mechanisms. Liu et al. (2019) and Chen (2019) also investigate the effect of green credit policy from the per- spective of enterprises. However, they focus on the financial constraint and ignore its further impact on the firms’ investment behavior. Although Sun et al. (2019) have examined the effect of China’s green credit policy on pol- luting firms’ environmental performance through the mechanisms of financial constraint, they focus on testing the initial effect of green credit policy during the pilot period in Jiangyin City, China. However, it is worth mentioning that since the introduction of GCGs in 2012, China’s green credit system has developed rapidly, so it is necessary to test the effect of green credit policy in the new stage. Therefore, with reference to the analysis of Sun et al. (2019), based on the perspective of capital supply and demand, this study adopts the sample dataset of the new stage of China’s green credit policy to combine the corporate financing ability with

1060 Environmental Management (2020) 66:1059–1071

the capital investment scale and further clarify the influen- tial mechanism through which the green credit policy work on micro-enterprises, and it provides new supportive evi- dence from the firm level for the effectiveness of green credit policy in China.

The remainder of the paper is organized as follows: in the second section, conceptual framework is presented followed by the proposition of research hypotheses. The third section describes the methodology. The empirical results follow in the fourth section. In the fifth section, we conclude the present study and put forward policy implications based on research findings.

Conceptual Framework and Hypothesis Development

Adopting a green and low-carbon development mode has become the consensus of countries around the world to cope with climate change and environmental pollution (Jiang et al. 2016). How to consider environmental issues while stabilizing economic development is one of the most con- cerning matters under focus in academia (Li et al. 2019). The green credit policy is an economic instrument based on the information disclosure of enterprises to solve environ- mental problems through adjusting the credit structure. In the process of green lending, financial institutions that issue loans, such as commercial banks, greatly affect whether green credit policy can be promoted as expected. At present, there are two opposite views on whether the green credit policy is being implemented. The first is about the non- executive theory, which asserts that banks will not willingly follow the principles of the green credit policy because of a lack of economic incentives (Biswas 2011). It is empha- sized that maximizing profits is the ultimate goal of banks, and most energy-intensive companies are strong enough as the main source of income from the banks’ credit. In con- trast, the green projects are lower in return (Taghizadeh- Hesary and Yoshino 2019) and may lead to high risks due to uncertainty (Yoshino et al. 2019). If the principles of the GCGs are followed, restricting credit to energy-intensive enterprises can make large portions of customers stay away, which will inevitably cause damage to the banks’ economic performance in the short term (Li et al. 2016). Therefore, financial institutions are not very willing to follow the green financial policy. However, another view supports the executive theory. This view holds that the green credit policy will provide promising opportunities for banks in the long run. Despite the short-term negative impacts (Hu and Cao 2011), the green credit policy can improve the banks’ environmental risk management, increase their brand value, tap into new markets, and enhance business efficiency (Liao et al. 2019). Furthermore, the policy will be instrumental in

achieving environmental protection and sustainable eco- nomic development simultaneously. The GCGs, as the first special policy on green credit in China, play a pivotal role (Chen and Li 2019). They put forward specific requirements to launch the green credit policy, strengthen the supervision of enforcement, and raise the cost of the compliance penalties faced by the banks. In recent years, increasingly severe environmental issues have increased the security risks of the banks’ credit, and the environmental risks faced by the banks are also rising dramatically. If the banks only pay for their immediate profit while ignoring the potential environmental risks, they are likely to suffer huge losses. In addition, energy conservation and emission reduction are conducive to the construction of the ecological civilization. For the realization of high-quality development (He 2016; Li et al. 2019), financial institutions should actively assume environmental responsibility (Liu and Wen 2019) in order to support the green economy. Under the background of fierce competition in the banking industry, green credit has undoubtedly been discovered as a blue ocean market strat- egy to occupy the commanding heights of potential new business areas, which is beneficial for banks to obtain dif- ferentiated competitive advantages through innovations with green credit products.

Energy-intensive industries are characterized by high energy consumption, large emissions, and serious pollution, which is contrary to the concept of energy conservation and emission reduction advocated by the GCGs (Bai et al. 2019; Du et al. 2018; Lin and Tan 2017; Zhang et al. 2018). Therefore, if the banks strictly follow the green credit policy and incorporate environmental factors into their credit management framework, the amount of loans for energy- intensive enterprises should be reduced, and more credit resources will be allocated to green and low-carbon enter- prises (Jin and Mengqi 2011). Under the status quo of China’s financial market, companies mainly rely on indirect financing channels dominated by the banks. Zhu et al. (2015) believe that China’s institutional environment will have impacts on enterprises’ financial ability, especially the financing constraints of industries that are affected by policies that are unfavorable to them. Therefore, the green credit policy may restrict the debt finance of energy- intensive enterprises. From the perspective of the close connection between financing and investment, credit financing is the vital source of capital supply for investment. Once enterprises are constrained by GCGs’ restrictions on their financing, they will naturally adjust their investment strategies and appropriately reduce their investment (Hao et al. 2020). Su and Lian (2018) found that the green credit policy has a significant investment inhibitory effect, and Liu et al. (2017), using the computable general equilibrium model, verified that the green credit policy can effectively curb the investment growth in energy-intensive industries.

Environmental Management (2020) 66:1059–1071 1061

Based on the following considerations, it is therefore pro- posed that energy-consuming enterprises will trim their capital investment. First, in order to ease the financing constraints brought by the green credit policy, energy- intensive enterprises will proactively respond by reducing capital investment in polluting areas. Second, because the green credit policy has strengthened the signal effect of national environmental governance, the enterprises will withdraw from the non-compliant investment projects in order to avoid the penalty of discharge. Finally, with the rise of Environment, Social Responsibility and Corporate Governance (ESG) investment, the value of green reputa- tion cannot be ignored. By reducing the investment projects with high environmental pollution, enterprises can set an example of bearing environmental responsibility, which is helpful to obtaining government’s encouragement and social financing support (Wu et al. 2020; Ye et al. 2010).

In summary, the mechanism of the GCGs on the capital investment of energy-intensive enterprises is shown in Fig. 1. Based on the above analysis, the following research hypotheses are proposed:

Hypothesis 1: The financial institutions in China strin- gently participate in the green credit policy, which imposes financing constraints on energy-intensive enterprises.

Hypothesis 2: The capital investment level of energy- intensive enterprises has declined along with the progress of the GCGs.

Hypothesis 3: The introduction of the GCGs reduces the capital investments of the energy- intensive enterprises through financing constraints.

Research Design

Sample Selection and Data Source

This study takes Chinese A-share listed companies as its research sample, with the sample period range being from

2009 to 2014. The rationality of using GCGs as exogenous policy shock lies in that since the introduction of GCGs in 2012, the intensity of green credit policy has been con- tinuously strengthened, so with GCGs as the exogenous policy shock can more accurately capture the response behavior of enterprises. The green credit policy is in a relatively early stage in China, although the policy pilot started in 2005, as well as other initial policies may not be enough to cause the obvious response of enterprises. However, to be more cautious and robust, the placebo test with 2005 and 2007 as policy implementation times will be conducted. Since the GCGs were issued in 2012, setting 2009–2014 as the research period can get the pure policy effect because there are no interference events during these years (Ding 2019). At the same time, it can alleviate the sequence autocorrelation problem caused by too long a research period (Bertrand et al. 2004). According to the DID model, the enterprises were divided into two groups: the experimental group, which was greatly affected by the GCGs, and the control group, which was less affected. As for the grouping standard, the real energy consumption intensity (EI) is used to measure the energy situation of the industrial sectors (Tan and Lin 2018). First, the real energy consumption intensity of 36 industries in China from 1999 to 2011 is calculated according to formula (1). Considering that some industries have experienced name or classification changes, this study excludes other mining industry, arts and crafts, and other manufacturing industry, waste resources, and the waste materials recycling processing industry.

EIit ¼ Eit

Yit ; ð1Þ

where i is the industry, t is the year, EI is real energy consumption intensity (unit: 10,000 tons of standard coal/ yuan), E is energy consumption (unit: 10,000 tons of standard coal), and Y is the real industrial output value (deflated at constant prices in 1999). According to formula (1), the

Fig. 1 The mechanism of the GCGs on energy-intensive enterprises

1062 Environmental Management (2020) 66:1059–1071

average value of the actual energy consumption intensity of each industry is calculated, and the energy-intensive group (higher than the median) is selected (more detailed informa- tion about treated group can be seen in Table 1). Then, the sample is divided into the energy-intensive group (treated group) and other group (control group) in accordance with the industry classification of the listed company. At the same time, the following samples are excluded from this study: (1) financial enterprises; (2) companies that are specially processed by the exchange; (3) listed companies that listed later than 2008; (4) listed companies with missing data. In the end, the balance panel data with 10,456 observations were obtained. The data come from the CMSAR database, China Energy Statistical Yearbook, China Statistical Yearbook, China Environmental Statistics Yearbook, and annual reports disclosed by listed companies. In order to avoid the influence of outliers, the winsorization was performed on all the continuous variables.

Variable Definition

(1) Dependent variable: The dependent variable is the capital investment level of the enterprises (Invest). Referring to Chen et al. (2012), this study takes the ratio of cash paid for the purchase of fixed assets, intangible assets, and other long-term assets to total assets as an indicator of the level of the enterprises’ capital investment.

(2) Mediator variables: The mediator variables are the financial constraints. This study measures the financial constraints from three dimensions: total credit (Debt), long-term credit (Ldebt), and short-term credit (Sdebt). The ratio of total bank loans to total assets is used to express the total credit scale, the ratio of the long-term bank loans to the total assets represents the long-term credit scale, and the ratio of the short-term bank loans to the total assets expresses the short-term credit scale.

(3) Main independent variables: The main independent variables are relevant indicators of the GCGs. GC indicates the time dummy variable for the implementa- tion of this policy. With the official introduction of the GCGs in February 2012 as the time node, the year before 2012 is set to 0, while after 2012 it is set to 1. Treated is a grouping dummy variable according to the real energy consumption intensity. Energy-consuming enterprises are set to 1 and other enterprises are set to 0.

(4) Control variables: Referring to the previous literature on corporate investment and financing behaviors (Chen et al. 2012; He et al. 2019), this study selects the following control variables: the size of the company (Size), profitability (Roe), financial leverage (Lev), Growth (Growth), free cash flow (Cash), investment opportunity (Tq), property right (Soe), the bank- enterprise relationship (BC), listing age (Age), board size (Board), and the degree of equity concentration (Thr).

The detailed definition of each variable is shown in Table 2.

Model Specification

Referring to Li and Lin (2017), the DID method was applied to study the impacts of the GCGs on the energy- intensive enterprises’ capital investment. Specifically, the model was as follows:

Investit ¼ α0 þ α1GCit þ α2Treatedit þ α3GCit � Treatedit þ αControlsit þ

P Year þP

Ind þ εit :

ð2Þ

In Eq. (2), i represents the enterprise, t is the year, and Invest indicates the capital investment level of the enter- prises. GC denotes the time dummy variable for the implementation of the GCGs, which is set to 0 before 2012 and set to 1 after 2012. Similarly, Treated indicates the different types of enterprises. The energy-intensive enter- prises are the experimental group, which is set to 1, and the other enterprises are set to 0. Controls represents the

Table 1 Industrial sectors distribution of treated group and average energy consumption intensity

Industrial sector EI Industrial sector EI

Smelting and pressing of ferrous metals 2.4159 Mining and processing of ferrous metals 1.2041

Manufacture of nonmetallic mineral goods 2.2124 Smelting and pressing of non-ferrous metals 1.1734

Industry of coal mining and washing 1.7082 Extracting of petroleum and gas 0.9321

Manufacture of raw chemical materials and chemical products 1.6592 Manufacture of chemical fibers 0.9291

Production and supply of gas 1.6292 Manufacture of paper and paper products 0.9143

Production and supply of water 1.3591 Mining and dressing of non-ferrous metals 0.6954

Processing of oil, coking and nuclear fuel 1.3017 Manufacture of rubber and plastic products 0.5206

Production and supply of electric power and thermal energy 1.2943 Manufacture of textiles 0.4475

Mining and processing of nonmetal mineral 1.2211 Manufacture of foods 0.3244

Environmental Management (2020) 66:1059–1071 1063

relevant control variables. ε comprises the random error terms with the property of independence and identical dis- tribution. Year and Ind represent the year effect and industry effect, respectively. This study focuses on α3, the estimated coefficient of GC× Treated, which implies the impacts of the green credit policy on the capital investment of energy- intensive enterprises.

In addition, existing studies have shown that credit financing can affect enterprises’ investment strategies. Therefore, based on the intermediary test procedure pro- posed by Wen and Ye (2014), this study constructs models (3)–(4) to assess the mediation mechanism of financing constraints between the green credit policy and capital investment of the energy-intensive enterprises.

DebtitðorLdebtitorSdebtitÞ ¼ β0 þ β1GCit � Treatedit þ βControlsit þ

P Year þP

Ind þ υit ;

ð3Þ

Investit ¼ η0 þ η1GCit � Treatedit þ η2Debtit orLdebtitorSdebtitð Þ þ ηControlsit þ

P Year þP

Ind þ μit ;

ð4Þ where Debt, Ldebt, and Sdebt represent the credit ability of enterprises, respectively, which can reflect the enterprises’ financial constraints. υ and μ are random error terms of

models (3)–(4); the definitions of the other variables are consistent with model (2).

The steps of the mediation effect testing are as follows: First, the estimated coefficient α3 of model (2) is tested. If the coefficient is significantly negative, it indicates that the green credit policy has a direct inhibitory effect on capital investment of the energy-intensive enterprises, and the next series of tests are needed. Otherwise, the mediation test should be stopped; The second step is to test the estimated coefficient β1 of model (3) and η2 of model (4), in turn. If both the two coefficients and η1 in model (4) are statistically significant, this indicates that there is a partial intermediary effect. However, if β1 and η2 are significant, but η1 is not significant, it can be concluded that there is a complete mediating effect. Furthermore, if at least one coefficient between β1 and η2 is insignificant, the bootstrap method should be applied. Only when the test results are significant, can the intermediary effect be considered as obvious.

Empirical Results

Descriptive Statistics

Table 3 provides the descriptive statistical analysis of the variables. During the research periods, the energy-intensive

Table 2 Definition of variables Symbol Variables Definition

Invest Capital investment Cash paid for purchase of fixed assets, intangible assets and other long-term assets/total assets

Debt Total loans Total loans/total assets

Ldebt Long-term loans Long-term loans/total assets

Sdebt Short-term loans Short-term loans/total assets

GC Green Credit Guidelines Dummy variables representing the implementation of green credit policy

Treated Enterprise group Dummy variables representing enterprise groups

Size Enterprise scale Natural logarithm of total assets at the end of the year

Roe Return on equity Net profit/net assets

Lev Financial leverage Total liabilities/total assets

Growth Growth rate of total assets (Total assets at the end of the year—total assets at the beginning of the year)/total assets at the beginning of the year

Cash Free cash flow Monetary fund/total assets

Tq Tobin Q value Market value/total assets

Soe Property right If ultimate controller is state-owned, it is set as 1, otherwise set as 0

BC banks-enterprise relationship

The experience of directors or senior executives in banking or banking supervision department is set as 1, otherwise it is set as 0

Age Listed years Natural logarithm of the enterprises’ age

Board Board size Natural logarithm of total number of board members

Thr Equity concentration Sum of shareholding ratio of top ten shareholders

Year Year Annual dummy variable

Ind Industry Industrial dummy variable

1064 Environmental Management (2020) 66:1059–1071

enterprises accounted for 28.05% of the whole sample. The average value of corporate capital investment is 0.0559, and the median value is 0.0413, which indicates that there are large differences among the enterprises’ capital investment. As for bank credit, the mean value of the total bank loans is 0.1742, and the median value is 0.1531, which shows that the total credit scale among the enterprises is unbalanced. Combined with the data of the long-term and short-term loans, it illustrates that there exist obvious differences in the credit maturity structure among the enterprises.

Analysis of DID Regression Results

Table 4 reports the basic DID regression results. Columns (1)–(3) examine the impacts of the GCGs on the capital investment of energy-intensive enterprises. The results in column (1) show that the estimated coefficient of GC× Treated is significantly negative (α3(1) =−0.0079; t= −3.3278). After including control variables, the results are basically consistent, as shown in column (2), which indicates that the GCGs negatively impacted (α3(2)= −0.0075; t=−3.1917) the capital investment of energy- intensive enterprises. Considering that the values of the enterprises’ capital investment are always positive and belong to limited data, the Tobit regression model is further adopted to reduce the deviation caused by the data characteristics. Column (3) shows that the estimated coefficient of GC× Treated is still significantly negative, implying that the green credit policy had an inhibitory effect (α3(3)=−0.0063; t=−3.7404) on the capital investment of energy-intensive enterprises. Based on the

above results, the GCGs had obviously restrictive effect on capital investment of high energy-consuming enter- prises, which supports the Hypothesis 2. Furthermore, the green credit policy could force enterprises withdraw from projects with energy-intensive and high pollution. In turn, it urged energy-intensive enterprises to follow the trend of low-carbon and green development, change investment concept, adjust investment structure, and incorporate environmental protection into the production process,

Table 4 Results of DID regression

Variables (1) (2) (3)

Invest Invest Invest

GC −0.0023* 0.0063** −0.0049***

(−1.9264) (2.1772) (−3.2051)

Treated 0.0233*** 0.0161 0.0214*

(12.3441) (0.6423) (1.8433)

GC × Treated −0.0079*** −0.0075*** −0.0063***

(−3.3278) (−3.1917) (−3.7404)

Constant 0.0520*** 0.0770 0.0311*

(54.2058) (1.3914) (1.7067)

Controls No Yes Yes

Year No Yes Yes

Ind No Yes Yes

Obs. 10,456 10,456 10,456

R2 0.0298 0.1196 –

T value in parentheses. *, **, *** denote statistical significance levels at 10%, 5%, and 1%, respectively

Table 3 The statistical description of variables

Variables Mean Median St.Dev Min Max Observations

Invest 0.0559 0.0413 0.0516 0.0002 0.2580 10,456

GC 0.6002 1.0000 0.4899 0.0000 1.0000 10,456

Treated 0.2805 0.0000 0.4493 0.0000 1.0000 10,456

Debt 0.1742 0.1531 0.1492 0.0000 0.5977 10,456

Ldebt 0.0583 0.0099 0.0919 0.0000 0.4228 10,456

Sdebt 0.1150 0.0866 0.1146 0.0000 0.4955 10,456

Size 22.0072 21.8355 1.2502 19.4507 25.5124 10,456

Roe 0.0644 0.0695 0.1259 −0.8185 0.3189 10,456

Lev 0.4704 0.4780 0.2095 0.0606 0.9100 10,456

Growth 0.1487 0.1035 0.2207 −0.3044 1.1678 10,456

Cash 0.1789 0.1456 0.1260 0.0084 0.5908 10,456

Tq 2.0178 1.6436 1.1519 0.1528 6.8571 10,456

Soe 0.5158 1.0000 0.4998 0.0000 1.0000 10,456

BC 0.4626 0.0000 0.4986 0.0000 1.0000 10,456

Age 2.0988 2.3979 0.7311 0.6931 3.0445 10,456

Board 2.1720 2.1972 0.1987 1.6094 2.7081 10,456

Thr 55.7211 56.4132 15.9435 21.3666 89.6855 10,456

Environmental Management (2020) 66:1059–1071 1065

which is of great significance to resource efficiency and sustainable development.

Analysis of Mediation Test Results

Table 5 shows the mediation effect of bank credit between the green credit policy and the capital investment of energy- intensive enterprises. Combined with results of column (1), Columns (2) and (3) are the test results with the total bank loans as intermediary variable. In column (1), the estimated coefficient of GC× Treated is significantly negative (α3= −0.0075, t=−3.1917), which shows the restraint effect of the green credit policy on the capital investment of energy- intensive enterprises. The estimated coefficient of Debt in column (2) is significantly negative (β1(Debt)=−0.0265, t= −6.0329), which indicates that the green credit policy had a restrictive effect on the total credit size of energy-intensive enterprises. In column (3), the estimated coefficients of GC × Treated and Debt are both significant (η1(Debt)=−0.0061, t=−2.6290; η2(Debt)= 0.0529, t= 5.7799). With reference to the mediation effect test mechanism, it shows that the total bank loans had a partial intermediary effect between the green credit policy and capital investment. Similarly, the results of columns (4) and (5) show that when the long-term bank loans are used as intermediary variable, the partial intermediary effect is also significant (β1(Ldebt)=−0.0231, t=−6.3357; η1(Ldebt)=−0.0048, t=−2.1207; η2(Ldebt)= 0.1167, t= 8.3817). Through the regression coefficients in columns (6) and (7), the mediating effect of short-term loans doesn’t pass the significance test (β1(Sdebt)=−0.0032, t=−0.8115; η1(Sdebt)=−0.0076, t=−3.2188; η2(Sdebt)=

−0.0190, t=−1.8409). Therefore, it is necessary to use the bootstrap method to check whether the intermediary effect of short-term credit between the green credit policy and capital investment exists. The results of the bootstrap test with 1000 samples show that the indirect effect coefficient was −0.0000669 (p= 0.295), which is insignificant, indi- cating that the intermediary effect of short-term credit is untenable. One possible explanation lies in the fact that the investment of energy-intensive enterprises mainly depends on long-term capital. Therefore, the green credit policy could not affect the capital investment of energy-intensive enterprises through short-term financing constraints. In summary, the green credit policy had restrained energy- intensive investment through financing constraints, and this was mainly realized along with a long-term path. These findings illustrate that the green credit policy is effective and plays a vital role in guiding resource allocation rationally, which can force enterprises to reduce energy- intensive investment; this strongly supports the Hypotheses 1 and 3.

Robustness Test

Parallel Trend Test

An essential premise of a DID model is to meet the parallel trend hypothesis; that is, to ensure that all enterprises in this study have the same change trend before the policy shock. In order to make the results more convincing, this paper employed event-study method to verify the parallel trend (Beck et al. 2010; Sun et al. 2019). Specifically, with 2012 as

Table 5 Results of mediation effect test

Variables Benchmark Total bank loan Long-term loan Short-term loan

(1) (2) (3) (4) (5) (6) (7)

Invest Debt Invest Ldebt Invest Sdebt Invest

GC × Treated −0.0075*** (−3.1917)

−0.0265*** (−6.0329)

−0.0061*** (−2.6290)

−0.0231*** (−6.3357)

−0.0048** (−2.1207)

−0.0032 (−0.8115)

−0.0076*** (−3.2188)

Debt 0.0529*** (5.7799)

Ldebt 0.1167*** (8.3817)

Sdebt −0.0190* (−1.8409)

Constant 0.0770 (1.3914) −0.1998* (−1.6550)

0.0876 (1.6226)

−0.4040*** (−4.7159)

0.1241** (2.2966)

0.2297** (2.1706)

0.0814 (1.4601)

Controls Yes Yes Yes Yes Yes Yes Yes

Year Yes Yes Yes Yes Yes Yes Yes

Ind Yes Yes Yes Yes Yes Yes Yes

Obs. 10,456 10,456 10,456 10,456 10,456 10,456 10,456

R2 0.1196 0.3451 0.1261 0.1477 0.1378 0.2297 0.1203

T value in parentheses. *, **, *** denote statistical significance levels at 10%, 5%, and 1%, respectively

1066 Environmental Management (2020) 66:1059–1071

the baseline, we set the dummy variables of the year effect before and after 2012, and then multiply the variable repre- senting the group classification (Treated) to construct inter- action terms (Before_3, Before_2, Before_1, After_1, and After_2) in regression model. If the coefficients of Before_3, Before_2, and Before_1 are not significant, then the parallel trend hypothesis can be supported. Table 6 shows that the estimated coefficients of Before_3, Before_2, and Before_1 are all insignificant, while both of the coefficients of After_1 and After_2 are negatively significant, indicating that the trend of the two groups is basically the same before the issuance of the GCGs and the green credit policy sig- nificantly affects the energy-intensive enterprises. Therefore, in the present research, the requirement of parallel trend has been verified, which indicates that the regression results of this study are reasonable to some extent.

PSM-DID estimation

Considering that the different sensitivity of capital invest- ment between the treated group and the control group to financial policy such as GCGs, the potential systematic difference may lead to the incompatibility between the two groups. Therefore, in order to enhance the comparability between the control group and experimental group in terms of investment change trend, with reference to Lu et al. (2020) and Shi et al. (2018), this paper applied propensity score matching method with DID models (PSM-DID) for robustness analysis. Before the conduction of PSM, it is necessary to ensure that there is no significant difference between the two groups after matching. Referring to the previous research literature (Chen et al. 2012; He et al. 2019), the investment characteristics of enterprises are affected by the variables at the firm level, such as the size of the company (Size), profitability (Roe), financial leverage (Lev), Growth (Growth), free cash flow (Cash), investment

opportunity (Tq), property right (Soe), the bank-enterprise relationship (BC), listing age (Age), board size (Board), and the degree of equity concentration (Thr). So, the above variables were included as covariates and the capital investment of enterprises as outcome variable to match the propensity score. Table 7 is the test results after matching, showing that the p value of all variables are <0.1, and the bias between the two groups is <10%. From this, it can be concluded that the application of PSM is effective.

Table 8 shows the results of PSM-DID. The estimated coefficient of GC×Treated is still significantly negative (α3=−0.0076, t=−3.22), which is consistent with the previous finding of this study.

Placebo test

The placebo test method was then implemented in this study to avoid the interference of other policies on energy con- servation and environmental protection. The practice of constructing a counterfactual experiment was as follows:

Table 6 Results of parallel trend test

Variables Invest Coef. St. Err. T value p value

Before_3 0.0038 0.0028 1.36 0.173

Before_2 −0.0044 0.0028 −1.59 0.112

Before_1 −0.0037 0.0026 −1.45 0.147

After_1 −0.0097*** 0.0023 −4.14 0.000

After_2 −0.0100*** 0.0026 −3.85 0.000

Constant 0.0399 0.0483 0.83 0.409

Controls Yes

Year Yes

Ind Yes

Obs. 10,456

R2 0.1285

T value in parentheses. *** denote statistical 1%

Table 7 Effectiveness test results of PSM-DID

Variables T value p value Bias (%)

Size 0.38 0.705 1.0

Roe 1.26 0.208 3.9

Lev 1.29 0.196 3.3

Growth 0.23 0.817 0.6

Cash −1.19 0.234 −2.7

Tq −0.87 0.385 −2.1

Soe −0.63 0.531 −1.6

BC −0.92 0.356 −2.4

Age 0.67 0.506 1.7

Board −0.87 0.383 −2.3

Thr −0.83 0.406 −2.2

Table 8 Regression results of PSM-DID

Variables Invest Coef. St. Err. T value p value

GC 0.0062** 0.0029 2.15 0.031

Treated 0.0181 0.0249 0.72 0.469

GC×Treated −0.0076*** 0.0024 −3.22 0.001

Constant 0.0744 0.0557 1.34 0.182

Controls Yes

Year Yes

Ind Yes

Obs. 10,444

R2 0.1203

T value in parentheses. **, *** denote statistical significance levels at 5% and 1%, respectively

Environmental Management (2020) 66:1059–1071 1067

First, assuming that 2005 was the year when the policy came into effect, this study repeated the previous test, with a sample range from 2003 to 2006. Second, set the 2007 as the node time when green credit policy came into effect, then repeat above analysis using sample period from 2006 to 2011. If the coefficients of GC× Treated are insignif- icant, it indicates that the effect of the virtual policy did not exist; that is, the decline of the capital investment of energy-

intensive enterprises was indeed affected by the GCGs. According to the estimated coefficients of GC × Treated in Table 9, they are not significantly negative (α3(1)= −0.0011, t=−0.2986; α3(2)=−0.0052, t=−1.4286), which supports the rationality of taking the GCGs as exo- genous shocking policy.

Alternative variables

Finally, the robustness test was carried out by changing the core variables. New capital investment was used as the dependent variable to conduct the DID model, with invest= (cash paid for purchasing fixed assets, intangible assets, and other long-term assets—cash received from disposal of fixed assets, intangible assets and other long-term assets)/total assets at the end of the period. The regression results of column (1) in Table 10 show that the GCGs have significant inhibitory effects (α3=−0.0075, t=−3.1357) on the capital investment of energy-intensive enterprises, which denotes that the results of this study are reliable to some extent.

In addition, this study analyzed the intermediary effect by using the operating revenues to standardize the mediator variables (tdebt, ldebt, sdebt), with new capital investment (invest) as the dependent variable. The results are shown in Table 10; the intermediary paths along which the GCGs affected energy-intensive investment through total bank loans (β1(debt)=−0.0917, t=−3.9169; η1(debt)=−0.0067, t=−2.8572; η2(debt)= 0.0088, t= 4.0562) and long-term bank loans (β1(ldebt)=−0.0701, t=−4.2453; η1(ldebt)=

Table 9 Results of placebo test

Variables (1) (2)

Invest Invest

GC 0.0072 0.0016

(1.3433) (0.3561)

Treated −0.0108*** −0.0196

(−2.8076) (−0.5594)

GC × Treated −0.0011 −0.0052

(−0.2986) (−1.4286)

Constant 0.0511 −0.1610**

(0.4377) (−2.5423)

Controls Yes Yes

Year Yes Yes

Ind Yes Yes

Obs. 4,275 7,594

R2 0.1023 0.0915

T value in parentheses. **, *** denote statistical significance levels at 5% and 1%, respectively

Table 10 Results of robustness test with alternative variables

Variables Benchmark Total bank loan Long-term loan Short-term loan

(1) (2) (3) (5) (6) (8) (9)

invest debt invest ldebt invest sdebt invest

GC × Treated −0.0075*** −0.0917*** −0.0067*** −0.0701*** −0.0062*** −0.0260** −0.0075***

(−3.1357) (−3.9169) (−2.8572) (−4.2453) (−2.6843) (−2.3248) (−3.1442)

debt 0.0088***

(4.0562)

ldebt 0.0182***

(5.6643)

sdebt −0.0010

(−0.2906)

Constant 0.0394 0.0630 −2.0929*** 0.0774 −0.2829 0.0391

(0.6977) (1.1349) (−4.4699) (1.3973) (−0.8797) (0.6931)

Controls Yes Yes Yes Yes Yes Yes Yes

Year Yes Yes Yes Yes Yes Yes Yes

Ind Yes Yes Yes Yes Yes Yes Yes

Obs. 10,456 10,456 10,456 10,456 10,456 10,456 10,456

R2 0.1219 0.1757 0.1270 0.1212 0.1327 0.1311 0.1220

T value in parentheses. *** denote statistical 1%

1068 Environmental Management (2020) 66:1059–1071

−0.0062, t= -2.6843; η2(ldebt)= 0.0182, t= 5.6643) are sig- nificant. However, combined with the bootstrap test, the coefficient of the indirect effect was−0.0000366 (p= 0.460), which indicates that the intermediary effect of the short-term loans between the GCGs and the energy-intensive investment is untenable (β1(sdebt)=−0.0260, t=−2.3248; η1(sdebt)= −0.0075, t=−3.1442; η2(sdebt)=−0.0010, t=−0.2906). The results are consistent with the above analysis of this study.

Conclusions and Implications

In recent years, the concept of green development has been deeply rooted in people’s minds. Determining how to explore the most effective ways for environmental protection is of great significance. Can the introduction of the green credit policy as a new environmental tool have a positive effect? This study took the promulgation of the GCGs in 2012 as a quasi-natural experiment, used the data of A-share listed companies in China to study the impacts of the green credit policy on the capital investment of energy-intensive enter- prises, and further analyzed the intermediary mechanism of financial constraints. The conclusions reached based on above empirical test results are as follows:

(1) The capital investment of energy-intensive enterprises was significantly reduced after the promulgation of the GCGs. The green credit policy could effectively curb the blind expansion of energy-intensive indus- tries, which was conducive to balancing the contra- diction between economic development and environmental protection.

(2) The GCGs had financial constraints on energy- intensive enterprises and therefore inhibited capital investment. It implied that the green credit policy has been well implemented in China to guide credit resources reasonably.

(3) Considering the intermediary paths along with the green credit policy on energy-intensive investment through financial constraints, the intermediary effect of total bank loans and long-term loans was significant, whereas the intermediary effect of short- term loans was insignificant. This result illustrated that GCGs mainly affect the investment behavior of energy-intensive enterprises by adjusting the long- term credit supply.

Based on above analysis, this study put forward several implications. First, in order to promote the green credit pol- icy, the government, financial institutions, and enterprises should cooperate with each other. The regulatory department should actively guide banks and enterprises to attach

importance to the green credit policy and design reasonable incentive systems to encourage banks and enterprises. Financial institutions should consider the environmental risks in the process of granting credit and devote themselves to supporting green and low-carbon development. Enterprises should raise their awareness of environmental risks and pay attention to environmental protection in order to meet their social responsibilities. Second, the green financial products are relatively single-focused, which cannot meet the diver- sified market demand. Although the green credit policy reduced energy-intensive investment in a short time, the impacts of ongoing financial constraints may hinder the normal development of enterprises, and this is not conducive to mobilizing the vitality of the market-oriented economy. Therefore, it is necessary to take advantage of big data to promote the sharing of environmental information between governments and enterprises and speed up the development of innovative green financial products relying on financial technology to establish a diversified green financial market. This would be particularly beneficial in deepening the green financial reform.

Acknowledgements This study was financially supported by the Ministry of Education of Humanities and Social Science Project of China (19YJA790086), the Key Project of National Social Science Foundation of China (No.18AZD014) and the Major project of National Social Science Foundation of China (No.19ZDA107). We would like to thank the editor (Bram F. Noble) and the anonymous reviewers for constructive and insightful comments.

Compliance with Ethical Standards

Conflict of Interest The authors declare that they have no conflict of interest.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

Bai Y, Song S, Jiao J, Yang R (2019) The impacts of government R&D subsidies on green innovation: Evidence from Chinese energy-intensive firms. J Clean Prod 233:819–829. https://doi. org/10.1016/j.jclepro.2019.06.107

Beck T, Levine R, Levkov A (2010) Big Bad Banks? The winners and losers from bank deregulation in the United States. J Financ 65:1637–1667. https://doi.org/10.1111/j.1540-6261.2010.01589.x

Bertrand M, Duflo E, Mullainathan S (2004) How much should we trust differences-in-differences estimates? Q J Econ 119:249–275. https://doi.org/10.1162/003355304772839588

Biswas N (2011) Sustainable Green banking approach: The need of the hour. Bus Spectr 1:32–38

Cai H, Wang X, Tan C (2019) Green credit policy, incremental bank loans and environmental protection effect. Account Res 3:88–95

Chang S, Yang X, Zheng H et al. (2020) Air quality and health co- benefits of China’s national emission trading system. Appl Energy 261:114226. https://doi.org/10.1016/j.apenergy.2019.114226

Environmental Management (2020) 66:1059–1071 1069

Chen D, Li S, Zhong Y (2012) Government quality, investment and capital allocation efficiency. J World Econ 35:89–110

Chen Q (2019) Has China’s green credit policy been implemented? An analysis of loan scale and costs based on “Two Highs and One Surplus” enterprises. Contemp Financ Econ 3:118–129

Chen S, Li Z (2019) An introduction to green finance theory and practice. Fudanpress, Shanghai.

China Banking Regulatory Commission (2012) Notice of the China Banking Regulatory Commission on issuing the green credit guidelines. http://www.cbrc.gov.cn. Accessed 24 Feb 2012

Chung SS, Fryxell GE, Lo CW (2005) Corporate environmental policy statements in mainland China: to what extent do they conform to ISO 14000 documentation? Environ Manag 35:468–482. https:// doi.org/10.1007/s00267-003-0085-3

Ding J (2019) Green credit policy, credit resources allocation and strategic response of enterprises. Econ Rev 4:62–75. https://doi. org/10.19361/j.er.2019.04.05

Du G, Sun C, Ouyang X, Zhang C (2018) A decomposition analysis of energy-related CO2 emissions in Chinese six high-energy inten- sive industries. J Clean Prod 184:1102–1112. https://doi.org/10. 1016/j.jclepro.2018.02.304

Du W, Li M (2019) Influence of environmental regulation on pro- moting the low-carbon transformation of China’s foreign trade: Based on the dual margin of export enterprise. J Clean Prod 244:118687. https://doi.org/10.1016/j.jclepro.2019.118687

Feng M, Li X (2020) Evaluating the efficiency of industrial environ- mental regulation in China: A three-stage data envelopment analysis approach. J Clean Prod 242:118535. https://doi.org/10. 1016/j.jclepro.2019.118535

Hao Y, Ye B, Gao M, Wang Z, Chen W, Xiao Z, Wu H (2020) How does ecology of finance affect financial constraints? Empirical evidence from Chinese listed energy- and pollution-intensive companies. J Clean Prod 246:119061. https://doi.org/10.1016/j. jclepro.2019.119061

He D-X (2016) Coping with climate change and China’s wind energy sustainable development. Adv Clim Change Res 7:3–9. https:// doi.org/10.1016/j.accre.2016.06.003

He LY, Liu RY, Zhong ZQ, Wang DQ, Xia YF (2019) Can green financial development promote renewable energy investment efficiency? A consideration of bank credit. Renew Energy 143:974–984. https://doi.org/10.1016/j.renene.2019.05.059

Hoang TC, Black MC, Knuteson SL, Roberts AP (2019) Environ- mental pollution, management, and sustainable development: strategies for Vietnam and other developing countries. Environ Manage 63:433–436. https://doi.org/10.1007/s00267-019-01144-z

Hu A-G (2016) The Five-Year Plan: A new tool for energy saving and emissions reduction in China. Adv Clim Change Res 7:222–228. https://doi.org/10.1016/j.accre.2016.12.005

Hu N, Cao D (2011) Green credit policy and environmental risk management of commercial banks. Econ Probl 3:103–107. https://doi.org/10.16011/j.cnki.jjwt.2011.03.027

International Energy Agency (2019) Global Energy & CO2 Status Report 2019. https://www.iea.org/reports/global-energy-co2-sta tus-report-2019. Accessed 26 Mar 2019

Jiang K-J, Zhuang X, He C-M, Liu J, Xu X-Y, Chen S (2016) China’s low-carbon investment pathway under the 2 °C scenario. Adv Clim Change Res 7:229–234. https://doi.org/10.1016/j.accre. 2016.12.004

Jin D, Mengqi N (2011) The paradox of green credit in China. Energy Procedia 5:1979–1986. https://doi.org/10.1016/j.egypro.2011.03. 340

Li C, Bai W, Wang Y, Li Y (2016) How can green credit policies be effectively implemented by commercial banks?—A study based on evolutionary game theory and the DID model. South China Financ 1:47–54

Li J, Lin B (2017) Environmental impact of electricity relocation: a quasi-natural experiment from interregional electricity transmis- sion. Environ Impact Assess Rev 66:151–161. https://doi.org/10. 1016/j.eiar.2017.07.002

Li J-F, Gu AL, Ma Z-Y, Zhang C-L, Sun Z-Q (2019) Economic development, energy demand, and carbon emission prospects of China’s provinces during the 14th Five-Year Plan period: appli- cation of CMRCGE model. Adv Clim Change Res 10:165–173. https://doi.org/10.1016/j.accre.2019.09.003

Liao J, Hu W, Yang D (2019) Dynamic analysis of the effect of green credit on bank operating efficiency—based on panel VAR model. Collect Essays Financ Econ 2:57–64. https://doi.org/10.13762/j. cnki.cjlc.2019.02.005

Lin B, Tan R (2017) Estimating energy conservation potential in China’s energy intensive industries with rebound effect. J Clean Prod 156:899–910. https://doi.org/10.1016/j.jclepro. 2017.04.100

Liu J-Y, Xia Y, Fan Y, Lin S-M, Wu J (2017) Assessment of a green credit policy aimed at energy-intensive industries in China based on a financial CGE model. J Clean Prod 163:293–302. https://doi. org/10.1016/j.jclepro.2015.10.111

Liu Q, Chen Y, Tian C, Zheng X-Q, Li J-F (2016) Strategic deliberation on development of low-carbon energy system in China. Adv Clim Change Res 7:26–34. https://doi.org/10.1016/j.accre.2016.04

Liu X, Wen S (2019) Should financial institutions be environmentally responsible in China? Facts, theory and evidence. Econ Res J 54:38–54

Liu X, Wang E, Cai D (2019) Green credit policy, property rights and debt financing: Quasi-natural experimental evidence from China. Financ Res Lett 29:129–135. https://doi.org/10.1016/j.frl.2019. 03.014

Lu J, Zhao Y, Su Y (2020) “Civilized City” selection and environ- mental pollution control: a quasi-natural experiment. J Financ Econ 46:109–124. https://doi.org/10.16538/j.cnki.jfe.2020.04.008

Ng AW (2018) From sustainability accounting to a green financing system: Institutional legitimacy and market heterogeneity in a global financial centre. J Clean Prod 195:585–592. https://doi. org/10.1016/j.jclepro.2018.05.250

Pang R, Zheng D, Shi M, Zhang X (2019) Pollute first, control later? Exploring the economic threshold of effective environmental regulation in China’s context, J Environ Manage 248:109275. https://doi.org/10.1016/j.jenvman.2019.109275

Ren S, Liu D, Li B, Wang Y, Chen X (2020) Does emissions trading affect labor demand? Evidence from the mining and manu- facturing industries in China. J Environ Manage 254:109789. https://doi.org/10.1016/j.jenvman.2019.109789

Shi D, Ding H, Wei P, Liu J (2018) Can smart city construction reduce environmental pollution. China Indus Econ 6:117–135. https:// doi.org/10.19581/j.cnki.ciejournal.2018.06.008

Su D, Lian L (2018) Does green credit policy affect corporate finan- cing and investment? Evidence from publicly listed firms in pollution-intensive industries. J Financ Res 12:123–137

Sun JX, Wang F, Yin HT, Zhang B (2019) Money talks: the envir- onmental impact of China’s Green Credit Policy. J Policy Anal Manag 38:653–680. https://doi.org/10.1002/pam.22137

Swanson KE, Kuhn RG, Xu W (2001) Environmental Policy Implementation in Rural China: A Case Study of Yuhang, Zhejiang. Environ Manag 27:481–491. https://doi.org/10.1007/ s002670010164

Taghizadeh-Hesary F, Yoshino N (2019) The way to induce private participation in green finance and investment. Financ Res Lett 31:98–103. https://doi.org/10.1016/j.frl.2019.04.016

Tan R, Lin B (2018) What factors lead to the decline of energy intensity in China’s energy intensive industries? Energy Econ 71:213–221. https://doi.org/10.1016/j.eneco.2018.02.019

1070 Environmental Management (2020) 66:1059–1071

The People’s Bank of China (1995) The notice on implementing credit policies and strengthening environmental protection. http://www. pbc.gov.cn. Accessed 6 Feb 1995

Wang Y, Pan D, Yu P, Liang X (2019) China’s incentive policies for green loans: a DSGE approach. J Financ Res 11:1–18

Weaver CP, Miller CA (2019) A Framework for climate change-related research to inform environmental protection. Environ Manage 64:245–257. https://doi.org/10.1007/s00267-019-01189-0

Wen Z, Ye B (2014) Analyses of mediating effects: the development of methods and models. Adv Psychol Sci 22:731–745. https://doi. org/10.3724/sp.j.1042.2014.00731

Wu W, An S, Wu C-H, Tsai S-B, Yang K (2020) An empirical study on green environmental system certification affects financing cost of high energy consumption enterprises-taking metallurgical enterprises as an example. J Clean Prod 244:118848. https://doi. org/10.1016/j.jclepro.2019.118848

Xie T, Liu J (2019) How does green credit affect China’s green economy growth? China Population. Resour Environ 29:83–90

Ye K, Zhang R, Xu H (2010) Reputation, institutional environment and debt financing: evidence from Chinese private listed com- panies. J Financ Res 8:171–183. https://doi.org/10.1360/972010- 1084

Ye Q-Z (2016) Safety and effective developing nuclear power to realize green and low-carbon development. Adv Clim Change Res 7:10–16. https://doi.org/10.1016/j.accre.2016.06.005

Yoshino N, Taghizadeh–Hesary F, Nakahigashi M (2019) Modelling the social funding and spill-over tax for addressing the green energy financing gap. Econ Model 77:34–41. https://doi.org/10. 1016/j.econmod.2018.11.018

Zhang B, Yang Y, Bi J (2011) Tracking the implementation of green credit policy in China: top-down perspective and bottom-up reform. J Environ Manag 92:1321–1327. https://doi.org/10.1016/ j.jenvman.2010.12.019

Zhang Y, Ma S, Yang H, Lv J, Liu Y (2018) A big data driven analytical framework for energy-intensive manufacturing indus- tries. J Clean Prod 197:57–72. https://doi.org/10.1016/j.jclepro. 2018.06.170

Zhu J, Han F, Lu Z (2015) Industrial policy, bank connections, and debt financing: an empirical research based on A-share listed companies. J Financ Res 3:176–191

Zhu N, Bu Y, Jin M, Mbroh N (2020) Green financial behavior and green development strategy of Chinese power companies in the context of carbon tax. J Clean Prod 245:118908. https://doi.org/ 10.1016/j.jclepro.2019.118908

Environmental Management (2020) 66:1059–1071 1071

Reproduced with permission of copyright owner. Further reproduction prohibited without permission.