Has Green Finance Reform and Innovation Pilot Zone Policy Improved New Quality Productive Forces? Quasi-Natural Experiment Based on Green Finance Reform and the Innovation Pilot Zone

Abstract

Using green finance reform and innovation pilot zone policy as a quasi-natural experiment, this study uses different methods to explore whether the policy implementation has improved new quality productivity in Chinese cities at the prefecture level and above. The results show that green finance reform and innovation pilot zone policy can significantly improve the level of regional new quality productivity, and the results are robust. Further heterogeneity analysis reveals that green finance reform and innovation pilot zone policy have heterogeneous impacts on the latest quality productivity of cities in different regions, scales, environmental regulatory intensities, financial development levels, and administrative levels. The parallel trend sensitivity test and function form robustness test further support the conclusions of this paper. Finally, suggestions are put forward to enhance green finance policies further.

1. Introduction

The Chinese Government attaches great importance to the issue of pollution and the environment. Although the wording varies, the reports at successive National Congresses of the Chinese Communist Party have emphasized green development. China is vigorously promoting the construction of an ecological civilization, adhering to the harmonious coexistence of human beings and nature, and steadily promoting carbon peaking and carbon neutrality. These are essential manifestations of the practice of lucid waters and lush mountains, which are invaluable assets and critical measures to accelerate the green transformation of the mode of development and promote Chinese-style modernization.

Green finance has emerged to address climate change, protect the environment, and encourage low-carbon economic transformation. The origin of green finance in China was in 2015, when the establishment of a green financial system was first proposed in the Overall Program for the Reform of the Ecological Civilization System. To further build China’s green financial system, in 2017, a State Council Standing Committee meeting designated five provinces, Zhejiang, Jiangxi, Guangdong, Guizhou, and Xinjiang, as the first pilot regions for green financial reform and innovation. Implementation of this policy aims to tilt financial resources to cleaner and more environmentally friendly projects by adjusting the allocation of financial resources, restricting the scale of financing in the “two high and one leftover” industries [1], and thus improving the level of green and sustainable development of financial institutions. The implementation of the policy can put into practice the concept of responsible investment and promote the construction of eco-civilization and green financial innovation and development [2] and gradually increase the role of green finance in the social and economic development of the country. The implementation of the policy can put into practice the concept of responsible investment, promote the construction of ecological civilization and the innovation and development of green finance, gradually increase the proportion of green finance in social financing, form upbeat guidance for high energy consumption and high pollution enterprises [3], and effectively serve the green development of the real economy.

The core of the policy of the pilot zone is to achieve the rational allocation of various resources and alleviate the problem of financial resource mismatch. Green finance reform and innovation pilot zones, by leveraging diverse green financial institutions, provide financial support for green industrial transformation and technological development, thereby fostering green economies and contributing to economic sustainability. Additionally, carbon financial markets have been established to advance carbon quota trading and carbon sink project development. This supports the transition to clean energy and enhances environmental sustainability. Moreover, new quality productive forces are the productivity of rational and optimal resource allocation. New quality productive forces are a leap in productivity achieved through scientific and technological innovation, industrial transformation and upgrading, and optimal resource allocation based on original productivity. New quality productive forces can be used as a comprehensive rational and optimal resource allocation indicator. In September 2023, when General Secretary Xi Jinping inspected Heilongjiang, he first proposed the concept of “new quality productive forces.” He later emphasized that new quality productive forces itself are green productivity. New quality productivity enhances production efficiency through intelligent and digital means, reducing resource consumption and waste emissions. They promote industrial upgrading, foster green industries, and drive the green transformation of the economy. This achieves a positive interaction between economic development and ecological conservation, providing strong momentum for sustainable development.

To investigate whether the green financial reform and innovation pilot zone policy effectively enhances the level of urban new quality productivity, this paper constructs a quasi-natural experiment framework, takes the green financial reform and innovation pilot zone policy as a “quasi-natural experiment”, and empirically examines the effect of the green financial reform and innovation pilot zone policy on new quality productivity. The marginal contributions of this paper are, first, that the effects can be assessed over the entire five-year period, whereas previous studies have mainly evaluated the impact of two to three years of policy implementation. Compared to past research, this study can capture the policy effects over a longer time span and provide a more comprehensive basis for policy evaluation. The second contribution is that the sensitivity to parallel trends and the robustness of the functional form are examined; and, third, is that the use of a comprehensive indicator, the new quality productive forces, is not limited to a single indicator such as the level of green finance, green technological innovations, and pollution and carbon reductions. This approach transcends the limitations of a single-dimensional perspective and evaluates the policy effects of green finance reform and the innovation pilot zone from a more comprehensive viewpoint.

2. Literature Review

The literature related to this paper focuses on the policy effects of green finance and new quality productive forces.

2.1. Research on the Policy Effects of Pilot Zones for Green Financial Reform and Innovation

Research on the policy effects of green finance reform and the innovation pilot zone is conducted at the macro and micro levels. At the macro level, urban panel data are primarily used for analysis. Green finance policies promote urban green innovation by improving human capital and the business environment [4]. At the same time, they enhance the efficiency of the green economy by attracting foreign investment, suppressing the financing of polluting enterprises [5], improving the efficiency of resource allocation, and promoting industrial structure upgrading [6]. Regarding energy conservation and emission reduction, the pilot zone has driven the decline in energy intensity [7] by improving green innovation, alleviating financing constraints, and optimizing the industrial structure. The policies of the pilot zone have reduced carbon emissions by increasing R&D investment, improving financing constraints [8], energy efficiency, and the intensity of environmental regulations [9], and improving the level of green technological innovation [10,11]. It has promoted the industrial structure’s transformation, upgrading, and high-quality economic development [12].

The analysis is conducted at the micro level using data from listed companies. The pilot zone policy promotes green innovation by easing the financing constraints of industrial enterprises [13] and increasing the strength of environmental constraints [14]. There is reputational risk to enterprises [15], but Jia and Zhang [16] argue that the pilot zone policy only promotes enterprises’ strategic green innovation and does not significantly impact substantive green innovation. Increased enterprise investment in environmental protection [17] and technological transformation [18] promoted enterprise capacity utilization rate and enhanced enterprise development efficiency [19]. It strengthens the financing constraints of heavy-polluting enterprises [20], optimizes the allocation of resources [21], and improves the degree of social responsibility fulfillment by enterprises. By strengthening external environmental supervision [22], improving the financing environment, and enhancing R&D and innovation, the ESG performance of enterprises has been improved. It is heterogeneous to enterprise development, enhances the value of green enterprises [23], and reduces the productivity of polluting enterprises [24].

2.2. Research Related to New Quality Productive Forces

Domestic scholars’ research on new quality productive forces mainly focuses on three aspects: the connotation and contemporary value of new productive forces. Xiao and Hao [25] think that the new quality productive forces are the intellectualization of the main body of labor, the intelligence of labor tools, and the digitalization of production factors. Zhou and Xu [26] believe that new quality productive forces are productivity generated by realizing the key disruptive technological break. Zhao and Ji [27] believe that the new scientific and technological revolution round and the development of emerging strategic industry clusters are the core connotations of new quality productive forces. Zhou and Li [28] believe that new technologies, economies, industries, and business forms characterize new quality productive forces. Wan [29] believes that the connotation of new quality productive forces is new kinetic energy, new governance, and new increments. At the same time, new quality productive forces have the value of promoting high-quality economic development [30] and Chinese modernization [31]. Second is the measurement of new quality productive forces. Li and Li [32] and Zhu et al. [33] measured new quality productive forces by using new quality laborers, new quality labor objects, and new quality labor materials; Lu [34] et al. measured the new quality productive forces by using three first-level indexes, six second-level indexes, and 18 third-level indexes, including science and technology productivity, green productivity, and digital productivity. Third is the path of new quality productive forces. Wei [35] thinks that new quality productive forces are the result of multi-factor system integration; in addition to the emphasis on scientific and technological innovation, we should also pay attention to the critical role of organization, management, systems, and other factors; Cheng and Chen [36] analyze the realization path of the development of new quality productive forces in terms of the top layout, the development of new industries, the transformation of the traditional industries, and the exertion of the advantages of intellectual property rights, and so on. In addition, digital technology [37], digital finance [38], scientific and technological innovation capacity [39], data elements [40], digital economy [41], and digital transformation [42] have also become realization paths to promoting new quality productive forces.

From a comprehensive view of existing studies, the current research on the effect of green financial policy is mainly focused on single indicators such as green innovation pollution reduction and carbon reduction, and the impact on comprehensive indicators is not yet familiar. This paper evaluates the policy effects of green finance reform and the innovation pilot zone from a more comprehensive perspective by constructing a composite indicator to measure new quality productivity; in addition, due to the timing of the publication of the literature, the effect period is mainly concentrated in the two to four years after the implementation of the policy. It is impossible to assess the overall effect of the five-year policy period entirely. This paper comprehensively measures the impact of green finance reform and the innovation pilot zone on the new quality productivity of Chinese cities at prefecture level and above from the five-year policy zones of the pilot zones, which improves the reference for policy makers to correctly evaluate the pilot zones. From the point of impact of green financial reform and innovation pilot zones, integrating green finance reform and the innovation pilot zone into the consideration of influencing factors for new quality productivity can both comprehensively evaluate the policy effects of these pilot zones and fill the research gap on the impact of green finance reform and the innovation pilot zone on new quality productivity. Therefore, studying the effect of green financial policies on new quality productive forces can play an important supplementary role to the existing research.

3. Theoretical Analysis and Research Hypothesis

Green finance reform and the innovation pilot zone facilitate the development of new quality productivity through multifaceted mechanisms including capital supply optimization, innovation promotion, and industrial structure optimization. Green finance reform and the innovation pilot zone have established a sound green financial system by optimizing capital supply and guiding capital flows toward green industries and environmental protection projects. This provides financial support for green technology research and development as well as green industry development required by new quality productivity, thereby promoting the advancement of new quality productivity. Green finance reform and the innovation pilot zone have promoted green innovation through policy support and improving the business environment. The promotion and application of these innovations have enhanced production efficiency and resource utilization efficiency, thereby driving the development of new quality productivity. Green finance reform and the innovation pilot zone have promoted industrial structure optimization through the synergistic interaction of policy guidance and market mechanisms. They have fostered the development of strategic emerging industries such as energy conservation, environmental protection, and clean energy, driven the green transformation of traditional industries, and facilitated industrial upgrading and transformation. In this process, new production models and technologies have been widely applied, fostering the development of new quality productive forces. Based on this, the study proposes the following hypothesis:

Hypothesis 1.

The implementation of green finance reform and innovation pilot zone policy has a positive effect on enhancing new quality productivity in Chinese cities.

Green finance reform and the innovation pilot zone can influence the development of new quality productivity through multiple mechanisms. However, due to significant differences in resource endowments, industrial structures, and economic development stages across regions, the transmission mechanisms of pilot zone policy and their impact effects on new quality productivity vary significantly among regions. Green finance reform and innovation pilot zones may also be characterized by heterogeneity in new quality productivity in cities of different sizes. Cities of mega-size and above exhibit high factor agglomeration levels but lower marginal effects, while cities of other sizes demonstrate relatively lower factor agglomeration levels yet relatively higher marginal effects. Cities with varying intensities of environmental regulations exhibit heterogeneous characteristics in new quality productivity within pilot zones due to the differential effects of environmental regulation. When cities have varying levels of financial development, there may also be heterogeneity in their impact on new quality productivity. Cities with higher financial development levels are often better able to guide capital flows into green industries and provide stronger policy support for green technological innovation. Cities at different administrative hierarchy levels have varying impacts on new quality productivity. Lower-level cities have a more urgent demand for green finance, and pilot zone policies are more likely to promote the development of new quality productivity in such cities. Based on the above theoretical analysis, the following hypothesis is proposed:

Hypothesis 2.

The policy has heterogeneous effects on the new quality productivity of cities in different regions, and of different scales, environmental regulation intensities, financial development levels, and administrative hierarchies.

4. Modeling and Data Description

4.1. Modeling

To study whether green financial reform and innovation pilot zone policy effectively enhance the level of new quality productive forces in cities, this policy is treated as a “quasi-natural experiment”, with a double-difference model constructed to conduct a benchmark regression analysis. Many cities are in the green financial reform and innovation pilot zone, further refined into 10 cities in this paper. Due to the severe lack of data in Changji Prefecture and Hami City in Xinjiang Uygur Autonomous Region, eight cities, namely, Quzhou City, Huzhou City, Guangzhou City in the east, Nanchang City, Jiujiang City in the center, and Guiyang City, Anshun City, and Kelamayi City in the west, are finally taken as the experimental group in the study. The remaining 271 prefecture-level cities and above are taken as the control group. Although Lanzhou New Area and Chongqing Municipality, etc., started their green finance reforms in November 2019 and August 2022, respectively, they are not treated as experimental groups because they have not had an entire assessment period of less than five years. The model is set explicitly as follows:

$$y_{it}={\beta }_0+{\beta }_{1}DI{D}_{it}+x_{it}^{\prime }\theta +{\alpha }_i+{\lambda }_t+{\epsilon }_{it}$$

(1)

where $i$ denotes the city, $t$ denotes the year; $y_{it}$ denotes the new quality productive forces of the $i$ city in the $t$ year; $DI{D}_{it}$ denotes whether city $i$ in year $t$ is included in the pilot green finance reform and innovation test zone, and ${\beta }_1$ is the policy effect that this paper focuses on; $x_{it}$ denotes a series of control variables, $\theta$ is the coefficient of influence of the control variables, ${\alpha }_i$ denotes the city fixed effect, ${\lambda }_t$ denotes the time fixed effect, and ${\epsilon }_{it}$ denotes the randomized perturbation term.

4.2. Variable Selection

(1)

New Quality Productive Forces

From the viewpoint of Marxist economics, the new quality productive forces are divided into new quality laborers, new quality labor materials, and new quality labor objects. Combined with the two sessions’ main points and data availability at the city level, this paper constructs an indicator system including three first-level indicators, eight s-level indicators, and 17 third-level indicators. Among them, the labor dimension includes three indicators, namely, the level of science and technology expenditure, the level of education expenditure, and the level of human capital; the dimension of labor means includes nine indicators, namely, the total amount of natural gas supply, the household usage of liquefied petroleum gas, the number of Internet users per 100 people, the number of cell phone users per 100 people, the per capita income from telecommunication services, the digital inclusive finance index, the percentage of employees in the computer software industry, the total amount of green patents granted, and the number of inventions applied for in the current year; the dimension of labor object includes five indicators, including the proportion of industrial added value, green coverage rate of built-up areas, harmless treatment rate of domestic garbage, centralized treatment rate of sewage treatment plants, and the comprehensive utilization rate of general industrial solid waste. The weight of each indicator of new quality productive forces is measured by the entropy value method.

(2)

Explanatory Variables

This paper uses 2017 as the time when the policy impacted each pilot city. If a city is a pilot city in the green finance reform and innovation pilot zone, its dummy variable is assigned a value of 1; otherwise, it is assigned a value of 0. The year of policy implementation and subsequent years are assigned a value of 1; otherwise, they are assigned a value of 0. Finally, the interaction term of the two is used as the policy variable for green finance reform and the innovation pilot zone to construct the core explanatory variable DID.

(3)

Control Variables

Concerning existing research results, when studying the policy effect of green finance reform and the innovation pilot zone on new quality productive forces, this paper selects a series of control variables at the city level, including per capita regional GDP (X1), measured by the ratio of regional GDP to the total population at the end of the year; the degree of openness to the outside world (X2), measured by the ratio of RMB-denominated import and export volume to regional gross domestic product; the proportion of expenditure on science and education of fiscal spending (X3), measured by the proportion of spending on science and technology and education in local general fiscal expenditure; the level of financial development (X4), measured by the ratio of the year-end balance of deposits and loans of urban financial institutions to GDP; and the level of industrial structure (X5), measured by the sophistication of the industrial structure.

4.3. Data Sources

The data range is 2011–2022, and the total number of cities in the sample is 279 due to missing data for some cities in most years. In the indicators for measuring new quality productivity, the Digital Financial Inclusion Index is sourced from the Institute of Digital Finance Peking University; the total amount of natural gas supply, the household usage of liquefied petroleum gas, and the green coverage rate of built-up areas are obtained from the EPS Data Platform; other variables are all sourced from the China City Statistical Yearbook; control variables are sourced from the China City Statistical Yearbook and the statistical bulletins of respective cities. Individual missing values are filled in using linear interpolation.

The descriptive statistics for each variable are shown in

Table 1. Descriptive statistics of variables.

Variable	N	Average	Standard Deviation	Min.	Median	Max.
y	3348	0.070	0.058	0.015	0.051	0.684
DID	3348	0.014	0.119	0	0	1
X1	3348	57,458.259	35,741.503	6457	47,970	467,749
X2	3348	0.193	0.318	0	0.079	3.078
X3	3348	0.193	0.042	0.048	0.194	0.372
X4	3348	2.582	1.230	0.588	2.280	21.301
X5	3348	1.078	0.608	0.114	0.941	5.650

.

5. Empirical Analysis

5.1. Parallel Trend Test

An essential prerequisite for using the double difference method is the assumption of parallel trends; that is, if the pilot city is not approved as a green financial reform and innovation pilot zone, the trend of changes in its new quality productive forces level will be the same as that of non-pilot cities. The double difference method does not require the pilot and non-pilot cities to be the same. There can be some differences between the two, but these differences cannot change over time before the policy is implemented.

There are usually two methods for testing parallel trends. The first is the time trend chart method, which involves plotting the time trend of the mean value of new quality productive forces in pilot and non-pilot cities. Although this method is intuitive, it cannot statistically test whether the trend changes in pilot and non-pilot cities before implementing policies are significant.

The second is the event study method, which is implemented by generating an interaction term between a dummy variable for the year of the event and a dummy variable for the treatment group and then adding it to the model for regression. The coefficient of these interaction terms measures the change in the difference between the pilot and non-pilot cities in other years before the policy compared with their difference in the base year. If the coefficient is not significantly different from zero, it indicates that the parallel trend assumption is valid.

The results of the parallel trend test are shown in Figure 1 and Figure 2. Figure 1 shows the results of the parallel trend test using the trend method. It can be seen from the figure that before the year of policy implementation, 2017, from 2011 to 2016, the trend of changes in the level of new quality productive forces in pilot cities (treatment group) and non-pilot cities (control group) was parallel. Figure 2 shows the results of the parallel trend test using the event study method. Before, Current, and After indicate the periods before, during, and after the policy implementation, respectively; the benchmark comparison period in the event study method can be any year before the policy implementation. In this paper, the year before the implementation of the policy, i.e., 2016, is selected as the benchmark comparison period. Therefore, the coefficients of Before6-Before2 represent the changes in the differences between the pilot cities and non-pilot cities from 2011 to 2015, and in 2016, respectively. As can be seen from Figure 2, the confidence intervals of the Before6-Before2 coefficients both include zero, indicating that the difference between the pilot cities and non-pilot cities from 2011 to 2016 before the implementation of the policy was not statistically significant, i.e., the parallel trend assumption holds, so it is reasonable and credible to use the double difference method to evaluate the policy effect of green finance reform and the innovation pilot zone.

5.2. Baseline Regression Analysis

Based on control of the urban effect, this section adopts the double difference method to estimate the policy effect of green financial reform and the innovation pilot zone on new quality productive forces. Columns (1) and (2) in

Table 2. Analysis of baseline regression results.

	(1)	(2)
	Y	Y
DID	0.4126 ***	0.1060 ***
	(0.0231)	(0.0350)
X1		6.6 × 10−6 ***
		(1 × 10−6)
X2		−0.1855 **
		(0.0797)
X3		0.2722
		(0.2143)
X4		0.0714 **
		(0.0313)
X5		0.2265 ***
		(0.0369)
Fixed effect of the city	Yes	Yes
Fixed effect of time	Yes	Yes
Constant	−0.7726 ***	−2.9048 ***
	(0.0000)	(0.2013)
N	3348	3348
R2	0.8425	0.9462

The standard error of the robust regression coefficient is in parentheses. *** and ** indicate significance at the 1% and 5% levels, respectively.

show the impact of green financial reform and innovation pilot zone policy on new quality productive forces before and after adding control variables, respectively. The coefficients of the core explanatory variable DID were 0.4126 and 0.1060 before and after adding control variables, respectively. These coefficients indicate that a per unit increase in the DID variable raises productivity levels by 0.4126 and 0.1060 units, both of which are statistically significant and positive. This confirms that green finance reform and innovation pilot zone policy have indeed significantly improved new quality productivity levels in China’s prefecture-level and above cities, supporting research hypothesis 1. After the control variables were added, the model’s goodness of fit increased from 0.8425 to 0.9462, and the DID coefficient decreased from 0.4126 to 0.1060, indicating that the DID coefficient in the model without control variables includes the influence of some control variables. Compared to the previous literature, this result focuses on the impact of pilot zone policies on comprehensive indicators of new quality productivity, instead of the single indicators such as green finance levels and green technological innovation studied in previous research.

5.3. Robustness Tests

If the differences between pilot and non-pilot cities are relatively significant, the results of the benchmark regression may be affected at this point. To avoid the influence of the difference between pilot and non-pilot cities on the double difference estimation results, we can use the propensity score to match and then carry out the double difference estimation for the matched samples, the so-called PSM-DID. The first essential step is to conduct Logit regression on all the samples with “whether it is a green financial reform and innovation pilot zone” as the dependent variable. The fitted value of the dependent variable is the probability of whether it is a green financial reform and innovation pilot zone, known as the propensity score in the causal inference. Then, according to the propensity score, we use nearest-neighbor matching to match the cities in the treatment and control groups. Then we carry out the regression analysis of the matched samples and compare the difference with the baseline regression results. Columns (1) and (2) in

Table 3. PSM-DID test results.

	(1)	(2)
	Y	Y
Treat	0.4085 ***	0.1312 *
	(0.0304)	(0.0333)
Control variable	No	Yes
Fixed effect of the city	Yes	Yes
Fixed effect of time	Yes	Yes
Constant	−2.1406 ***	−3.3727
	(0.1401)	(0.1951)
N	166	166
R2	0.9770	0.9913

The standard error of the robust regression coefficient is in parentheses. *** and * indicate significance at the 1% and 10% levels, respectively.

show the results of the impact of green financial reform and innovation pilot zone policy on new quality productive forces after propensity score matching with and without control variables, respectively. The coefficient estimates of the core explanatory variable DID are 0.4085 and 0.1312, respectively, indicating that each 1-unit increase in the DID coefficient raises productivity levels by 0.4085 units and 0.1312 units. Both results are significantly optimistic and consistent with the results of the benchmark regression, indicating that the results of the previous benchmark regression are still robust after excluding differences in the characteristics of pilot and non-pilot cities themselves. In addition, the goodness-of-fit of the DID regression model for the matched samples improves considerably compared with the benchmark regression. It confirms that the implementation of green finance reform and innovation pilot zone policy has a significant positive impact on new quality productivity in China’s prefecture-level and above cities, thereby validating the hypothesis.

6. Heterogeneity Analysis

The above research found that establishing green finance reform and innovation pilot zones can effectively improve urban new quality productive forces. However, implementing policies may have different effects due to differences in regional resource endowments, administrative hierarchy, level of financial development, institutional environment, etc. Therefore, this paper examines the heterogeneous impact of green finance reform and innovation pilot zones on urban new quality productive forces from the perspective of regional differences, city size, the intensity of urban environmental regulation, the level of urban financial development, and the administrative hierarchy of cities.

6.1. Regional Heterogeneity

The country is divided into the eastern, central, and western regions for analysis, and the results are shown in

Table 4. Analysis of regional heterogeneity in policy effects.

	Eastern Region	Central Region	Western Region
Treat	0.0318	0.1186 ***	0.0491
	(0.0412)	(0.0398)	(0.0745)
Control variable	Yes	Yes	Yes
Fixed effect of the city	Yes	Yes	Yes
Fixed effect of time	Yes	Yes	Yes
Constant	−2.9500 ***	−2.8834 ***	−3.9929 ***
	(0.0911)	(0.2391)	(0.1394)
N	948	1428	972
R2	0.9723	0.9267	0.9402

The standard error of the robust regression coefficient is in parentheses. *** indicates significance at the 1% levels.

. Green finance reform and innovation pilot zone policies have all increased the level of new quality productive forces in the three regions. The results indicate that the green finance reform and innovation pilot zone policies enhance the new quality productivity levels of prefecture-level and higher cities in China across the eastern, central, and western regions, which aligns with the research hypothesis. Among them, the policy’s role in the central region is more prominent, followed by the western and eastern areas, forming a situation of difference. Green finance reform and the innovation pilot zone have a heterogeneous effect on new quality productivity of cities in the eastern, central, and western regions, which satisfies hypothesis 2. In terms of significance, implementing policies significantly positively impacts cities in the central region. In contrast, the impact on the western and eastern regions is positive but not statistically significant.

6.2. Heterogeneity of Urban Size

Cities are divided into mega and above-scale cities and other scale cities according to whether the size of the city’s resident population is greater than 5 million, and the city-scale heterogeneity of the policy effect of green financial reform and the innovation pilot zone is analyzed. As seen from

Table 5. Analysis of city size heterogeneity in policy effects.

	Other Size Cities	Cities of Mega-Size and Above
Treat	0.0915 ***	0.0364
	(0.0330)	(0.0348)
Control variable	Yes	Yes
Fixed effect of the city	Yes	Yes
Fixed effect of time	Yes	Yes
Constant	−3.4534 ***	−3.5835 ***
	(0.0959)	(0.2294)
N	2289	1059
R2	0.9099	0.9705

The standard error of the robust regression coefficient is in parentheses. *** indicates significance at the 1% levels.

, green finance reform and the innovation pilot zone positively impact the new quality productive forces of cities of different sizes. The policy of green finance reform and the innovation pilot zone in cities of varying sizes has been demonstrated to enhance the level of new quality productivity in China’s prefecture-level and above cities, which conforms to research hypothesis 2. However, the effect is significant only for other cities, and the impact on megacities and above is not statistically significant. One reason may be that mega-cities and above have a higher concentration of new types of workers, new kinds of means of production, and new types of production objects. Their new quality productive forces levels are also higher. Therefore, the marginal effect of green finance reform and the innovation pilot zone policy is correspondingly lower. In contrast to mega-cities and above, other cities have a limited agglomeration effect on new types of workers, new kinds of means of production, and new types of objects of production, and their new quality productive forces levels are relatively low. Therefore, the marginal effect of the policy is higher. Green finance reform and the innovation pilot zone demonstrate significantly greater enhancement of new quality productivity in other size cities compared to cities of mega-size and above, indicating that the policy exerts heterogeneous effects on new quality productivity across cities of different scales.

6.3. Heterogeneity in the Intensity of Environmental Regulation

The strength of environmental regulation is measured using the proportion of ecological protection words in local government reports. The sample is divided into two groups with high and low environmental regulation intensity according to the 50th percentile value, and regression is performed based on the model. The results, as shown in

Table 6. Heterogeneity analysis of the environmental regulatory intensity of policy effects.

	Low Environmental Regulatory Intensity	High Environmental Regulatory Intensity
Treat	0.1422 ***	0.0106
	(0.0375)	(0.0373)
Control variable	Yes	Yes
Fixed effect of the city	Yes	Yes
Fixed effect of time	Yes	Yes
Constant	−3.2126 ***	−3.2898 ***
	(0.1149)	(0.0998)
N	1814	1534
R2	0.9479	0.9506

The standard error of the robust regression coefficient is in parentheses. *** indicates significance at the 1% levels.

, show that green finance reform and the innovation pilot zone policy significantly improve the level of new quality productive forces in cities with low environmental regulatory intensity, and the policy effect on high environmental regulatory intensity is positive but not statistically significant. It shows that green finance reform and innovation pilot zone policy for cities with different environmental regulation intensities can improve the level of new quality productivity in Chinese cities at prefecture level and above, which is in line with the research hypothesis. The policy demonstrates weaker promotion effects on new quality productivity in cities with high environmental regulation intensity compared to those with low intensity, indicating significant heterogeneity in policy outcomes, which also aligns with research hypothesis 2. Cities with high environmental regulatory intensity may put more pressure on polluting enterprises’ survival, forcing them to accelerate technological innovation, gradually eliminating backward production capacity and reducing environmental pollution. The ecological regulatory effect has weakened the policy effect of green financial reform and innovation pilot zones to a certain extent. For cities with low environmental regulatory intensity, the pilot green financial reform and innovation pilot zones have improved green financial development and provided financial support for industrial restructuring and technological innovation, thus significantly enhancing the level of new quality productive forces.

6.4. Heterogeneity in Levels of Financial Development

Different levels of financial development often correspond to varying levels of financial system construction, which further affects the implementation effect of policies in green finance reform and innovation pilot zones. The samples are divided into two groups based on the 50th percentile of the level of financial development: high and low. The results are shown in

Table 7. Analysis of Heterogeneity in Levels of Financial Development.

	Low Level of Financial Development	High Level of Financial Development
Treat	0.1168 **	0.1278 **
	(0.0497)	(0.0495)
Control variable	Yes	Yes
Fixed effect of the city	Yes	Yes
Fixed effect of time	Yes	Yes
Constant	−4.0133 ***	−2.5822 ***
	(0.0899)	(0.3117)
N	2016	1332
R2	0.9207	0.9589

The standard error of the robust regression coefficient is in parentheses. *** and ** indicate significance at the 1% and 5% levels, respectively.

. Implementing the policies significantly improves the level of urban new quality productive forces. Judging from the magnitude of the coefficients, implementing the policies has a more significant promoting effect on the latest quality productive forces of cities with high levels of financial development than on those with low levels. The policy of green finance reform and the innovation pilot zone exerts heterogeneous effects on the new quality productivity of cities with different levels of financial development, consistent with hypothesis 2. Since regions with a high level of financial development have a more complete financial system and a financial environment that is more conducive to implementing green finance policies, they can significantly enhance cities’ new quality productive forces. However, cities with a low level of financial development sometimes adopt a wait-and-see attitude when implementing green finance policies, which to some extent weakens or delays the effect of policy implementation.

6.5. Heterogeneity in the Hierarchy of Administrative Areas

The level of financial development in different regions means different levels of administrative management to a certain extent. In relative terms, provincial capitals, sub-provincial cities, and municipalities directly under the central government often have better resource endowments, economic conditions, and policy environments than general prefecture-level cities, which leads to differences in the policies of green financial reform and innovation pilot zones in cities of different administrative levels. Based on this, provincial capitals, sub-provincial cities, and “larger cities” approved by the State Council are classified as high-level administrative cities. The remaining prefecture-level cities are classified as low-level cities, and an analysis of the heterogeneity of urban administrative levels is carried out accordingly.

Table 8. Heterogeneity analysis of regional administrative hierarchies.

	Low-Ranking Cities	High-Ranking Cities
Treat	0.0854 **	0.0844 **
	(0.0398)	(0.0340)
Control variable	Yes	Yes
Fixed effect of the city	Yes	Yes
Fixed effect of time	Yes	Yes
Constant	−3.4917 ***	−2.9475 ***
	(0.0980)	(0.2386)
N	2748	600
R2	0.8955	0.9632

The standard error of the robust regression coefficient is in parentheses. *** and ** indicate significance at the 1% and 5% levels, respectively.

shows that green finance reform pilot zones significantly affect the new quality productive forces of both types of cities. The effect on lower-tier cities’ latest quality productive forces is slightly higher than that of higher-tier cities, but the difference is insignificant. The green finance reform and innovation pilot zone policy is heterogeneous for new quality productivity in cities of different administrative levels, consistent with hypothesis 2. Lower-tier cities urgently need green finance to pursue development, so the effect of green finance reform pilot zones may be more significant.

7. Further Analysis

7.1. Robustness Test of the Model Functional Form Setting

The model set in the previous section is linear; if the model form is not linear, it will produce model setting bias; at this time, the least squares estimator of the coefficients will be biased and non-consistent. Double Machine Learning can utilize the advantages of machine learning algorithms in dealing with nonlinear data to avoid the model mis-setting problem effectively. Double/Debiased Machine Learning (DDML) mainly includes partial linear regression (PLR) models and the more general interaction regression (IR) models. The general form of a partial linear regression model is

$$y_{it}={\beta }_{1}DI{D}_{it}+g\left(x_{it}\right)+u_{it} E[u_{it}\left|DI{D}_{it}\right.,x_{it}]=0$$

(2)

$$DI{D}_{it}=m\left(x\right)+v_{it} E\left[v_{it}\left|x_{it}\right.\right]=0$$

(3)

The general form of an interactive regression model is

$$y_{it}=g_0(DI{D}_{it},x_{it})+u_{it }E[u_{it}\left|DI{D}_{it}\right.,x_{it}]=0$$

(4)

$$DI{D}_{it}=m_0\left(x\right)+v_{it} E\left[v_{it}\left|x_{it}\right.\right]=0$$

(5)

Double Machine Learning is used to estimate the $y_{it}$ and $DI{D}_{it}$ conditional expectations by employing a machine learning approach, which is further estimated to get the policy effects. The results of the Double Machine Learning estimation of policy effects are shown in

Table 9. Double Machine Learning Estimation.

	Partial Linear Regression		Interactivity Regression
	Lasso (5-Fold)	Lasso (10-Fold)	Lasso (5-Fold)	Lasso (10-Fold)
DID	0.1199 ***	0.1131 ***	0.1499 ***	0.1493 ***
	(0.0234)	(0.0225)	(0.0445)	(0.0503)
Control variable	Yes	Yes	Yes	Yes
Fixed effect of the city	Yes	Yes	Yes	Yes
Fixed effect of time	Yes	Yes	Yes	Yes

The standard error of the robust regression coefficient is in parentheses. *** indicates significance at the 1% levels.

; when partial linear regression is used, whether 5-fold or 10-fold-cross-validation, the policy effect of green financial reform and innovation pilot zones is significantly positive and slightly higher than the benchmark regression results. When using interaction regression, whether 5-fold or 10-fold-cross-validation, the policy effect is significantly positive and higher than the benchmark and partial linear regression results. Therefore, even considering the model function setting problem, the policy effect of green financial reform and innovation pilot zones is still robust.

7.2. Sensitivity Tests for Parallel Trends

A parallel trend is the most critical assumption of the double difference method. The estimated policy effect will be inaccurate if the parallel trend assumption does not hold. Roth et al. [43] pointed out that the traditional parallel trend test is less effective, with a low probability of rejection. They suggested that statistical inference and sensitivity analysis should be conducted on the confidence interval of the estimated policy effect. First, the maximum difference in the trend between the pilot and non-pilot cities before the policy implementation is calculated. Then, based on this maximum deviation, a confidence interval for the policy effect is constructed. If the confidence interval of the policy effect does not include zero at the maximum deviation, the policy effect is considered robust to deviations from parallel trends. In their study, Sarsons [44] set the maximum deviation to twice the standard error of the treatment effect. This setting can assess whether the policy effect is still significant in the most unfavorable situation, thereby enhancing the reliability and robustness of the research results. This paper adopts the same setting and obtains the confidence intervals of the policy effect at the maximum deviation level, i.e., the year of policy implementation, the second year after policy implementation, the third year after policy implementation, and the fourth year after policy implementation, as shown in Figure 3. Figure 3a shows the 90% confidence intervals for the policy effect in the year of policy implementation under various deviations from the parallel trend assumption, and Figure 3b–d shows the 90% confidence intervals for the policy effect in the second, third, and fourth periods after policy implementation, respectively. All the confidence intervals in the four sub-charts in Figure 3 do not include zero, indicating that even if there is a certain degree of deviation from the parallel trend assumption, the effect of green finance reform and the innovation pilot zone on enhancing new quality productive forces is still robust.

8. Conclusions and Recommendations

This paper studies the policy effect of the pilot zone for green financial reform and innovation on new quality productive forces by constructing a double difference model. The study shows that the impacts of green financial reform and innovation pilot zone policies on new quality productivity are all positive, indicating that green financial reform and innovation pilot zone policies can significantly improve the level of new quality productivity in Chinese cities at prefecture level and above, the research hypothesis has been confirmed, and the following conclusions and suggestions are obtained.

First, implementing the policies of green finance reform and innovation pilot zones can significantly promote improving the level of urban new quality productive forces. This conclusion effectively evaluates the policy effects of the pilot zone, thereby assisting policymakers in properly understanding and actively promoting their development. Based on a comprehensive summary of the five-year trial period’s experience, the scope of the pilot zones should be expanded further. Financial institutions and enterprises should be guided to deeply understand the policy content and advantages, while financial institutions should be encouraged to develop more financial products and services tailored to the development needs of green industries to continuously promote the implementation and integration of green financial policy. At present, Shanghai and Weihai have actively applied to establish national-level green finance reform and innovation pilot zones. Other regions should also actively act in the context of the national dual carbon target and initiate a higher positive interaction between green finance and new quality productive forces in their regions.

Second, regional heterogeneity exists in the policy effects of green finance reform and innovation pilot zones. The research conclusion demonstrates the regional heterogeneity of the experimental area, providing theoretical support for promoting regional green development tailored to local conditions. In further promoting the pilot zones for green financial reform and innovation, from an efficiency perspective, priority should be given to regions with the best foundation for green finance and strong potential policy effects. These regions should be encouraged to take the lead in piloting new green financial products and promoting the internationalization of green finance markets; from an equity perspective, priority should be given to areas with a poor foundation for green finance and an urgent need for green finance. Fiscal transfer payments to these areas should be increased to channel financial resources toward them, supporting local efforts to establish green financial service systems and foster the development of green finance; from an efficiency-first, equity-balanced perspective, priority can be given to regions with relatively underdeveloped economies, and priority can be given to cities in relatively underdeveloped areas with potential solid policy effects. A point-to-area approach should be adopted to drive green finance development in surrounding regions through cities with strong potential policy effects. When measuring regional heterogeneity, this study mainly considers the macro-level impacts of different regions on new quality productivity, but there is insufficient exploration of how intra-regional industrial structures affect policy outcomes. Future research could refine the analysis of intra-regional industrial structures to make policy formulation more aligned with regional industrial characteristics.

Third, the policy effects of green finance reform and innovation pilot zones exhibit heterogeneity in urban nature. This conclusion provides theoretical support for cities with diverse characteristics to identify their own positions and develop new quality productivity based on their unique features. When further promoting green financial reform and the innovation pilot zone, policies should be formulated in a differentiated manner based on the nature of the city. Pilot cities are encouraged to draw on the successful experiences of cities that have already piloted the approach and to choose from various options, such as corporate carbon accounts, personal carbon accounts, carbon emission rights collateralized financing, carbon-neutral funds, and forest carbon pools based on local conditions. At the same time, the green financial product system should be gradually enriched, innovation in green financial products and services promoted, and the strengthening and development of new quality productive forces actively supported. This study has paid less attention to the differences in policy transmission mechanisms among cities at different administrative levels when analyzing the heterogeneity of urban characteristics. Subsequent research could focus on this aspect to optimize policy transmission pathways across cities with varying administrative hierarchies, thereby enhancing policy implementation effectiveness.