Green Credit Effects of Local Debt Management System Reform:Evidence Based on Regression Discontinuity Design
Published Online: Feb 27, 2025
Received: Oct 03, 2024
Accepted: Jan 29, 2025
DOI: https://doi.org/10.2478/amns-2025-0143
Keywords
© 2025 Yuxin Zhang et al., published by Sciendo
This work is licensed under the Creative Commons Attribution 4.0 International License.
Since the ‘dual carbon’ goal was proposed in 2020, various industrial enterprises in China have also been actively seeking a more efficient and sustainable green way of consumption and productivity. Since 2021, green finance has increased rapidly and has received unprecedented attention of economic academia. China Banking and Insurance Regulatory Commission decreed the guidelines for green finance of the banking and insurance industries in June 2022, requiring to develop green finance in a more systematic and comprehensive way. Green credit of twenty one major banks in China reached a scale of 25 trillion RMB on 30th June 2023, with average annual growth rate of 33%, ranking first in the world. The chinese government has adopted many measures aimed at easing the financial pressure on local governments since 2008. Specifically, the Chinese government allows local governments to issue local bonds through ‘issuance-for-repayment’ mechanism and encourages them to set up financing platforms. Because of the implicit guarantee of local governments, a large amount of loans from commercial banks flowed into local financing platforms, which results in a rapid expansion of the issuance scale of local investment bonds. However, most of these projects are public welfare, and the cash flow situation is not ideal and the debt risk increases. So the central government began to devise a reform strategy for local debt management. Relevant departments of Chinese goverment issued the file about strengthening modern govenance of local government debt in 2014, and implemented a new budget law in 2015, marking fundamental shifts in the financial way of local governments. More funds are available for the development of green credit. The aim of this article is to investigate green credit effect and its mechanism of the local debt management system reform(abbreviated as LDMSR below) using the data of commercial banks. This article seeks to provide theoretical support and literature for the development of local debt management and green credit.
The implementation of the Chinese LDMSR in 2015 has affected public government debt and banking development. From a macro perspective, the reformation improves the quality of regional economic development (Hong Yuan and Hu Huijiao, 2023), promotes local governments to change their land transfer behavior obtaining more land fiscal revenue (Tian Shengdan et al., 2021), and alleviates the pressure of government budget constraints(Bi Sifeng and Wang Xueyuan, 2021). From the micro perspective, the reformation reduces new loans from financing platforms to alleviate the financing difficulties of enterprises (Liang Ruobing and Wang Qunqun, 2021), promotes the upgrading of corporate human capital (Hu et al., 2022) and the digital transformation of enterprises (Li Yifei et al., 2023), and inceases corporate leverage ratio (Deng Xuan et al., 2023). Huang Hao et al.(2023) concluded that enterprises have significantly reduced the proportion of investment in financial assets and increased the proportion of investment in industrial assets.
Many factors affected green credit of commercial banks. In terms of policy factors, there are macro control, environmental protection (Zhang Pingdan and Zhang Xiayi, 2017), local government environmental control (Lin Boqiang and Pan Ting, 2022), implementation of local policies (Ren Danni, 2020), perfection of the regulatory system(Mai Junhong and Xu Feng, 2015; Li Shanmin, 2019), performance evaluation standards and industry environmental performance evaluation guidelines for green credit (Zuo et al., 2017). In terms of economic factors, there are development environment of the green finance market (Yao Shuqi and Sun Hongmei, 2020), local financial development (Mai Junhong and Xu Feng, 2015; Wang Hongli, 2023), digital inclusive finance (Zhu Xiaoying, 2023), and the distortion of the capital factor market (Zhang Xin and Xia Hongwen, 2024). These factors all have impacts on commercial banks’ green credit. In terms of technological factors, Zhong Kai et al. (2023) and Kong Weiwei et al. (2023) concluded that Chinese commercial banks can promote the growth of green credit business with financial technology. Huang et al. (2023) argued that financial technology promoted the provision of green credit by alleviating information asymmetry, adjusting credit allocation, and adjusting credit risk. Jiang et al. (2023) argued that the blockchain financial technology increased green credit issuance, alleviated financial constraints and reduced financing costs of enterprises.
The theory of government debt sustainability demonstrates the government should focus on controlling the scale and cost of debt in the process of performing its functions to ensure the sustainability of debt. Due to implement of the former budget law in 1995, local government’s financing funds came from commercial bank credit, in 2012 CBIRC issued the ‘Green Credit Guidelines’ requiring banks to effectivelycarry out green credit, which makes commercial bank have less funds to develop green credit after meeting the financing requirements of local governments. So the green credit development of commercial banks has relatively decreased. In 2015, LDMSR began to be implemented which changed the financing ways of local governments, reduced the recessive guarantee of local government financing loans and the scale of commercial banks’ loans to financing platform companies. Under the condition that the supply quantity of credit funds from commercial banks remains unchanged, crowding-out effect of government debt on commercial banks’ green credit is reduced, which is conducive to the moderate development of Chinese commercial banks’ green credit. So this paper propose hypothesis H1 as follow:
Based on the theory of green finance innovation, the adjustment of credit allocation is an important factor affecting the development of green credit for commercial banks (Huang et al., 2023). Liang Haisheng et al. (2020) believe that the overall return and risk of commercial banks’ assets will be changed during LDMSR, which will affect their capital allocation strategies, and ultimately it is likely to improve bank liquidity. The reformation can release more credit funds from commercial banks by reducing the debt risk of local governments, that is, local bonds of local goverment can reduce the demand of credit funda for Chinese commercial banks, thereby reduce the credit risk of Chinese commercial banks themselves, and provide full play to the role of government bond funds in driving social funds, so as to promote the flow of funds to public welfare industries and stimulate the development of regional economy. Chinese commercial banks can adjust credit allocation in reasonable and effective ways and improve the efficiency of credit allocation, that is, commercial banks will be more willing to take risks and issue more green credit than before, in order to obtain green reputation, complete green credit policy indicators and obtain more hidden benefits. So this paper propose hypothesis H2 as follow:
Debt ratio and credit risk of the local government have increased, when the local government has implicit rigid payment and China central government has been restricted transfer of payment to the local government before the reformation. The reformation influences green credit of commercial banks from the following aspects. First, for commercial bank, the risk of purchasing local bonds has been reduced, the profit has been guaranteed, and commercial banks have more funds to develop green credit. Second, with the transparency improvement of local government debt after the reform, commercial banks can more accurately assess the risk of government debt, so as to make more rational credit decisions, which will help to reduce the credit risk caused by information asymmetry. Third, the reformation has reduced the difficulty in risk management of commercial banks by regulating the borrowing behavior of local governments and strengthening debt supervision. This helps commercial banks identify, assess and control credit risks more effectively; Fourth, the reformatio has adopted debt swaps and debt restructuring methods, which has reduced debt costs of Chinese local governments and eased the capital pressure of Chinese commercial banks. This will help commercial banks maintain sufficient capital and improve their ability to resist risks of distressed Assets. So this article proposes hypothesis H3 as follow:
Considering LDMSR in 2015, this article selects 36 commercial banks listed in China’s A-share market as the research object, with a time span from 2008 to 2021. We manually collect data of green credit from the sustainable development reports and social responsibility reports of Chinese commercial banks, and We also use linear interpolation to fill in the small amount of missing data. Other control variable data come from Guotaian database, Wind database and Resset database.The article not only filters and eliminates the sample banks with missing key variables, but also performs 1% and 99% percentiles tail reduction on all continuous variables, to ensure the reliability, accuracy and stability of the empirical results. In the end, 316 samples of 36 banks were obtained. Stata 16.0 software are used for building models and empirical analysis.
Because this article studies the green credit, dependent variable is green credit of commercial banks whose abbreviation is grl. Referring to Zhong Kai et al. (2023), most data of green credit were obtained from the Guotaian database, and a small part of the data are obtained by manually sorting out from the social responsibility repors of commercial banks.
Explanatory variable is LDMSR,which is represented by the product named TD of the dummy variable ‘T’ which is effective date for the local government debt management system reform and the dummy variable ‘D’ which is whether the city has implemented the reformation. Drawing on the research of Liang Ruobing and Wang Qunqun (2021), we manually queried the annual general accounts and the explanation of government debt borrowing of the government, and identified reform time of the cities where banks headquarters is located.
Bank credit allocation and bank credit risk are mediating variables. Referring to the research approach of Liu Fang et al. (2022), loan-to-deposit ratio (DR) of commercial banks is adopted to measure their credit allocation status. The higher the loan-to-deposit ratio, the higher the efficiency of credit allocation for commercial banks. Drawing on previous studies, ratio of risk-weighted assets to total loans (CR) is used to measure the credit risk level for commercial banks. The lower the ratio, the lower the credit risk level for commercial banks.
Referring to previous studies, the control variables not only include microscopic control variables of commercial banks, including: revenue growth rate (GR), liquidity ratio (FR), cost-to-income ratio (CI), the logarithm of total assets(SZ), the debt-to-equity ratio (Der), the non-performing loan ratio (Npl), but also include macroscopi control variables: inflation level (FL), the level of economic development (GD) and broad money volume (MN). Table 1 shows definition and symbol of the above variables.
| types of variables | variable | Symbol | definition |
|---|---|---|---|
| variable explained | Green credit | grl | Green credit of commercial bank |
| Explaining variable | The local debt management System reform began | DT | Local governments really implement the reformation |
| mediating variables | Bank credit allocation | DR | Average amount of total loans / Average amount of total deposits |
| Bank credit risk | CR | Risk-weighted assets / Total loans amount | |
| control variables | Revenue growth rate | GR | Revenue increase / Total revenue of last year |
| Liquidity ratio | FR | Current assets amount / Current liabilities amount | |
| Cost-income ratio | CI | Operating expenses amount / Operating income amount | |
| Total assets | SZ | The logarithm of total assets | |
| Non-performing loan ratio | Npl | Non-performing loans/gross loans | |
| Debt-to-equity ratio | Der | Total liabilities / Owners’ equity | |
| Inflation levels | FL | consumer price index | |
| Economic development level | GD | Annual growth rate of GDP | |
| Amount of broad money | MN | The growth rate of broad money |
According to the Opinions on Strengthening the Management of Local Government Debt issued by the State Council and the new Budget Law officially implemented in 2015, LDMSR has been carried out nationwide, which has further promoted the development of green credit of Chinese commercial banks. Therefore, 2015 is the demarcation point for whether the object of this paper is affected by the treatment effect of LDMSR. However, years are often discontinuous at the cut-off point. Therefore, based on the characteristics of quasi-random experiments on LDMSR, this paper adopts the accurate regression discontinuity design (RDD) to evaluate the impact of the reformation. In order to test whether the reformation has an impact on banks’ green credit, the following regression discontinuity model (1) is established to verify research hypothesis H1:
In model (1), grl is the explained variable, T is dummy variable, whose value is 1, if LDMSR was implemented,otherwise, the value is 0 respectively. D is a dummy variable, whose value is 1 when bank is located in the reform city, otherwise, D = 0. The coefficient of regression β3 reflects the degree of influence. α and β are regression coefficients. i represents the bank and ε represents random error.
In order to test the impact mechanism of LDMSR on commercial banks’ green credit, the two-step method of Jiang Dingsheng’s (2022) intermediary effect was used for analysis, regression discontinuity model(2) and (3) are built as follow to test research hypotheses H2 and H3:
In model (2), Mediator is the intermediary variable, including bank credit allocation (DR), bank credit risk (CR), α and γ are regression coefficients, εrepresents the random error.
An important criterion for the establishment of breakpoint regression design is that the amount of green credit in commercial banks must have a significant jump at the breakpoint. The basic principle is as follows: If the green credit volume of commercial banks shows an obvious break point in 2015, the change is likely to be caused by LDMSR. Figure 1 below is a breakpoint effect analysis based on graph and polynomial fitting, including scatter plot and linear fitting plot. The horizontal axis is time and the vertical axis is green credit (grl). Figure 1 shows, before LDMSR, the development of green credit in commercial banks shows a slow upward trend, mainly because it was proposed in 2007 that commercial banks should develop green credit. After LDMSR in 2015, it can be seen from the linear fitting chart and the fitting chart that the development of green credit has slowly declined, which may be due to the initial implementation of the reform. As a result, it is difficult for commercial banks to recover the funds that flowed into local bonds before the reform, thus affecting commercial banks’ investment in green credit. On the whole, after LDMSR, green credit showed an obvious jump and the slope changed, that is, there was a significant breakpoint effect. This shows that this reform has a certain effect on green credit of commercial banks.
Scatter plot and the fit of Fig
From the analysis of the above breakpoint effect shown in figure 1, it can be seen that reformation can promote green credit of commercial banks. Therefore, the precise breakpoint regression model is applicable to study the effect of LDMSR on green credit of commercial banks. This article uses the method of gradually adding control variables and incorporating the commonly used 1.5 and 2 times optimal bandwidth to perform breakpoint regression, in order to enhance the reliability of the results. Table 2 shows the regression results for the imfluence of the reformation on green credit. The data in column (1) of table 2 show that the regression coefficient of the reformation is 1.051 and that the reform has a significant impact on green credit of commercial banks, when the control variables are not added to the first regression. The data in column (2) of table 2 show that the regression coefficient of the reformation after adding control variables is 0.987 and is significant, indicating that the reformation is still effective in promoting the green credit.
RDD regression results
| (1) | (2) | |
|---|---|---|
| grl | grl | |
| 1.051*** | 0.987** | |
| (2.99) | (2.25) | |
| 1.051*** | 0.970** | |
| (2.99) | (2.20) | |
| 1.051*** | 0.961** | |
| (2.99) | (2.18) | |
| 316 | 316 | |
| No | Yes |
In order to reduce the impact of other potential variables on green credit of commercial banks from LDMSR, this article adopts the method of counterfactual assumption to test the robustness, that is, the time of the reformation is advanced and lagged by 1 year and 2 years, and it is taken as the new reform time, and the new reform time is used as the breakpoint for regression. According to the placebo test results in table 3, we find the regression results are not significant whether the time of reform is 1 year and 2 years ahead or 1 year and 2 years behind, which can prove that the conclusions reached above are robust.
Placebo test results
| Variable | The reform 2 years ahead of schedule | The reform 1 year ahead | The reform lags behind 1 year | Reform lags behind 2 years |
|---|---|---|---|---|
| grl | grl | grl | grl | |
| -0.110 | 0.266 | -0.442 | -0.794 | |
| (-0.19) | (0.40) | (-0.50) | (-0.47) | |
| -0.065 | 0.233 | -0.375 | -0.557 | |
| (-0.15) | (0.44) | (-0.48) | (-0.39) | |
| -0.034 | 0.248 | -0.207 | -0.516 | |
| (-0.07) | (0.47) | (-0.29) | (-0.46) | |
| 316 | 316 | 316 | 316 | |
| Yes | Yes | Yes | Yes |
In order to avoid the problem of error in the selection of estimation methods, a simple OLS method is used for research, although the OLS method is not causal inference, but it reflects the accuracy of the results as a robustness. The article analyzes the OLS method, and its regression results are shown in table 4. Table 4 shows a obviously positive correlation between green credit and reforms, which proves that the reforms can effectively promote green credit of Chinese commercial banks.
Least Squares regression results
| Variables | OLS | OLS |
|---|---|---|
| grl | grl | |
| TD | 1.051*** |
0.934*** |
| CP | -0.171** |
|
| GD | 0.016 |
|
| MN | 0.001 |
|
| GR | 0.008 |
|
| FL | 0.004 |
|
| CI | -0.028*** |
|
| SZ | 1.627*** |
|
| Der | -0.171** |
|
| Npl | -0.425*** |
|
| Constant | 1.187*** |
-41.277*** |
| Observations | 316 | 316 |
| R-squared | 0.021 | 0.486 |
| Observations | 316 | 316 |
Since the default trigonometric kernel function is used for estimation in the RDD regression mentioned above, the method of replacing kernel function is adopted here, that is, replacing it with moment function to verify the robustness of previous results. The regression results of replacing kernel function can be seen in Table 5 before the control variable is added for the first time, The regression coefficient of the reform is 1.051 and shows that green credit is obviously positively correlated with the reformation. Furthermore, after addition of control variables to the models, the regression coefficient of the reformation is still obviously positively correlated with green credit. According to the above results, it can still be concluded that the above conclusions are robust.
Replace the kernel function regression results
| Variables | moment function | moment function |
|---|---|---|
| grl | grl | |
| lwald | 1.051*** |
0.934** |
| lwald150 | 1.051*** |
0.934** |
| lwald200 | 1.051*** |
0.934** |
| Control variables | No | Yes |
We learn from above research conclusion that reforms of local debt management system can facilitate green credit in commercial banks. Exploring the influence mechanism of green credit of commercial banks from the reform, the article also used breakpoint regression model to test its intermediary effect by the two paths of adjusting credit allocation and adjusting credit risk.In order to avoid the endogeneity problem of explanatory variables in the traditional three-step method, this paper adopts the two-step mediation effect method of Jiang Tian (2022). The first step is the regression analysis of explanatory variable (DT) and explained variable (grl), that is, the RDD regression mentioned above, whose conclusion is very significant. In the second step, the explanatory variable green credit is replaced with the intermediary variable. From column (1) in table 6, we can see the regression coefficient of reform(DT) is 7.084 and a significant positive correlation between reformation(DT) and intermediary variable (DR) at the significance level of 1%, and we come to conclusion that the intermediary effect of bank credit allocation(DR) is significant. The reformation positively promotes green credit by adjusting the assignment of bank credit. The digits in column (2) of table 6 show that the regression coefficient of the reform(DT) is -0.116 and that there is a obvious negative correlation between the reformation(DT) and intermediary variable (CR) at the significance level of 1%. From the above analysis, we find the intermediary effect of Bank credit risk (CR) is significant. LDMSR can promote the development of green credit by reducing risks in bank credit.
Results of mediator regression
| Variable | (1) | (2) |
|---|---|---|
| DR | CR | |
| lwald | 7.084*** |
-0.116*** |
| lwald150 | 6.829*** |
-0.110** |
| lwald200 | 6.706*** |
-0.108** |
| Observations | 314 | 314 |
| Control variables | Yes | Yes |
Chinese green credit of commercial banks has increase rapidly with reforms of local debt management system towards modern governance. Based on the background, this article constructs accurate breakpoint regression models and mediation effect models, and explores the effects of LDMSR on green credit of commercial banks and its specific impact mechanism, with panel data of Chinese listed commercial banks from 2008 to 2021. Empirical analysis of effects shows that the reformation can effectively promote green credit of commercial banks. Empirical analysis of mechanism shows that bank credit allocation and bank credit risk are the impact pathes or channels that the reformation promotes green credit.
The authors acknowledge the 2023 National University Student Innovation Training Project of Hebei Finance University (Project No. 202311420005) - ‘Mechanism and Optimization Strategies of Tax and Fee Policies Supporting Enterprise Green Development under Carbon Risk Constraints’ and sponsored by Project of Hebei Finance Department in 2024 (Project No. HBCZ24KY258) - ‘Effects of Local Debt Management System Reform on Green Credit of commercial banks’