The Effect of Financial Development on Industrial Green Technology Innovation Efficiency: Experience Analysis from 288 Cities in China

Abstract

Green industrial growth is at the core of the green and low-carbon transformation of the real economy. Financial services provide new channels for green innovation financing for enterprises. How to utilize financial development to enhance the industrial green technology innovation efficiency (GTIE) is the focus of this paper. Using data from 288 cities in China, this paper employs the Super-SBM model and the Network DEA model to measure the industrial GTIE and then investigates the complex relationship between financial development and industrial GTIE. The results show that, overall, financial development can promote the industrial GTIE. And alleviating enterprise financial constraints and increasing investment openness are effective pathways for this influence. In addition, we discuss the heterogeneous effects of financial development on industrial GTIE across different types of cities, manifested as a dual effect of “supporting” and “predation”. In regions with stronger financial power such as coastal and eastern cities, the “predation effect” is stronger, while the “support effect” is stronger in inland and central/western cities. Furthermore, the expansion of city size weakens the impact of financial development on industrial GTIE. This paper points out that financial development plays a crucial role in enhancing the industrial GTIE, but excessive financial power may crowd out the positive impact of financial development on the industrial GTIE. When promoting financial development, it should be matched with region development to avoid crowding out investment in green innovation by enterprises.

1. Introduction

The sustainable development of the economy is an unavoidable issue in the process of economic development for emerging countries. The scarcity of natural resources and environmental pollution have made society realize that economic development relies on the energy and ecological services provided by natural resources. Economic development should prioritize human development [1]. However, industrial growth interacts with energy consumption and environmental pollution in a mutually influential manner [2]. The successive industrial revolutions have made significant contributions to industrial growth, yet the pollution and energy efficiency issues inherent in traditional industrial production patterns are posing severe challenges to the global ecological environment. Data released by the International Energy Agency show that the global industrial sector exhibited notably higher coal consumption in 2021 compared to other industries, accounting for 82.59% of the total consumption. Simultaneously, carbon emissions from the industrial sector were surpassed only by those generated by the electricity and heat production sector, as well as transportation, comprising 18.89% of the aggregate emissions. Industrial production is a crucial driver of the continuously expanding ecological footprint [3]. Therefore, fostering the shift of industrial production toward environmentally friendly practices has become a prevalent choice for governments and enterprises worldwide. Moreover, enhancing the industrial GTIE is a pivotal agenda in attaining sustainable development [4].

In the wake of rapid development in emerging economies, environmental issues have gradually gained attention from various sectors, leading to a concept shift in development from “treatment after pollution” to “green development”. Ambec and Lanoie [5] introduce a green total factor productivity measurement method based on environmental efficiency, which modifies and supplements traditional measures. Henceforth, abundant literature has investigated the influence of economic and institutional environmental enhancements on the promotion of green growth [6,7,8]. The level of financial development in the market is identified as a key factor for low-income countries to narrow the gap with their high-income counterparts. This stems from the fact that a more developed financial sector increases the likelihood of low-income countries catching up with middle- to high-income nations [9]. Nawaz et al. [10] argue that green finance development is considered a primary solution for achieving sustainable development and plays a significant role in economic growth. Given the high-risk and low-return characteristics of green projects, utilizing financial instruments is paramount to propel green initiatives [11]. To advance the development of a green economy, as a leading representative of emerging economies, China has instituted a series of green financial policies, including green credit and subsidy aimed at fostering innovation in green production. It utilizes credit and subsidy policies to guide financial capital toward the environmental protection industry and expand investments that are environmentally friendly, thereby stimulating green innovation in the industry [12,13].

The existing literature generally recognizes the positive influence of green finance on fostering green innovation. Nevertheless, ongoing discussions persist regarding the broader impact of financial development on green innovation across various industries, underscoring the insufficient attention paid to the industrial GTIE. Given the significant economic role of industrial sectors and their energy consumption and emission issues, it is essential to conduct an in-depth analysis of the impact and mechanism of financial development on the industrial GTIE. Additionally, it is crucial to analyze under what circumstances financial development can promote industrial GTIE and avoid excessive financial development causing waste of resources and crowding out green innovation.

This paper offers three main contributions. Firstly, we utilize the Super-SBM model and Network DEA model to measure the industrial GTIE of 288 cities across China from both static efficiency and dynamic productivity perspectives. Secondly, it emphasizes the importance of financial development for the GTIE of industrial sectors and investigates the mediating roles of financial constraints and investment openness. Thirdly, we investigate the “support” and “predation” effects of financial development on industrial GTIE across different cities and find that excessive financial power may crowd out firms’ green technology innovation efforts.

2. Literature Review

Dangelico et al. [14] argue that green technology innovation is a form of innovation capable of reducing environmental harm while increasing human welfare. Green technology innovation can decrease energy consumption, alleviate ecological pollution, and enhance firms’ GTIE [15,16]. The greening of industrial production is a necessary path toward achieving economic sustainability. From 1990 to 2015, China experienced significant growth in green technology innovation [17], effectively mitigating the conflict between industrial growth and environmental degradation [18].

Research on the relationship between financial development and industrial GTIE generally stems from theories of financial development and economic growth. Financial development is positively correlated with per capita income growth in a country, which is an important factor affecting national economic growth [19,20]. Lv et al. [21] argue that optimizing financial structure benefits the green technology innovation of enterprises. Through functions such as risk diversification and credit support, the financial system catalyzes enterprises’ green innovation activities and thereby enhances their GTIE. Many scholars believe that financial development has significant long-term effects on environmental improvement and sustainable development [22,23]. However, the existing research still lacks sufficient attention to the industrial sector, where the green transformation in production is particularly crucial for environmental improvement. Therefore, this paper focuses on the impact of financial development on industrial GTIE and analyzes the relationship between financial development and industrial GTIE from three perspectives.

2.1. Direct Effect of Financial Developments on the Industrial GTIE

Financial development provides more funding support for enterprises, facilitating the green transformation of industries. Green finance is a crucial means to regulate the flow of funds in financial markets toward green sectors, which increases environmental investment, promotes green innovation, and contributes to sustainable economic growth [8,24]. Meo and Abd [25] argue that green finance has encouraged investments in renewable energy technologies, significantly reducing carbon dioxide emissions. Aristizabal-Ramirez et al. [26] consider that financial development enhances the allocation of resources in strategic sectors and promotes green technology innovation. Yet, only large firms can benefit from the process of financial development. Gao et al. [27] conducted a micro-level study and found that the development of green finance promotes enterprise innovation, alleviates financial constraints, enhances managers’ attention to the environment, and improves the green productivity of enterprises.

Financial development plays a pivotal role in driving green technology innovation. The advancement of digital technology enhances its positive impact on fostering green technology innovation within enterprises [28]. Optimizing resource allocation and reducing financial constraints are mechanisms for achieving the green transformation of environmentally harmful enterprises [29]. The decentralization of environmental responsibility can effectively alleviate resource misallocation issues in green investments.

2.2. Indirect Effect of Financial Development on the Industrial GTIE

2.2.1. Financial Constraint Effect

The growth of industrial sector productivity requires external funding support, and financial constraints restrict production and R&D [30]. In countries with more advanced financial development, enterprises have lower financing costs, and industrial sectors that rely on external financing develop faster [19]. As the financial system develops, the financial constraints on SMEs gradually decrease, and a developed financial system helps sectors relying on external financing to perform better economically during financial crises [31]. Additionally, the Chinese government has introduced green credit policies to alleviate enterprise financial constraints. With lower financial constraints, enterprises are more likely to invest in green R&D [32].

The R&D of green technology have characteristics of high risk, long cycle, and externality. Reducing capital market constraints and facilitating access to credit for enterprises are crucial for industrial green innovation [33]. Obstacles to external financing restrict the process of enterprise green technology innovation, and controlling financing risks is an effective mechanism for increasing enterprise innovation activities [34]. Cecere et al. [35] studied the impact of public funds on corporate ecological innovation when funds are scarce and found that after the relaxation of financial constraints, the more public funds or fiscal incentives enterprises receive, the stronger their ecological innovation capabilities, which improves the enterprise’s GTIE.

2.2.2. Investment Openness Effect

Due to information asymmetry in financial markets, financial intermediaries can assess and monitor projects as well as disseminate information, thereby alleviating financial market frictions. Better disclosure rules in financial markets can reduce moral hazard and adverse selection problems and can facilitate enterprises to access financial resources at lower costs [36]. FDI is a crucial manifestation of global financialization, where foreign investors are market-oriented and prefer to enter regions with developed financial markets. Regions with more developed financial systems exhibit greater attractiveness for foreign direct investment [37]. Inflows of green investments increase opportunities for developing economies to adopt new technology and to achieve cleaner and environmentally friendly production [38].

The entry of FDI alleviates the constraints of financial constraints on host country enterprises and mitigates the negative impact of financial market distortions on resource allocation [39]. Although FDI increases the energy consumption of host countries, it can help achieve clean production by reducing carbon dioxide emissions through green innovation [40]. The influx of green FDI from investment openness enhances the sustainable technological innovation capabilities of multinational enterprises and drives innovation activities related to green technology [41], thus improving the industrial GTIE of the host country.

2.3. Heterogeneous Impact of Financial Development on the Industrial GTIE

The institutional environments in transitional nations shape the impact of financial development on driving enterprise innovation [26]. Razzaq et al. [42] argue that the effects of digital financial development and renewable energy technological innovation on green growth vary asymmetrically across different regions of China. Financial development will enhance technological progress biased toward energy and the environment, but unreasonable and excessive financial development will hinder green innovation [43].

The principal components of the financial system in transitional countries are prone to evolve into special interest groups. The financial sector has its own interests and may use its advantages to seek maximum benefits. When financial market forces become too dominant, financial efficiency may stagnate or even decline. Financial institutions easily exploit their monopolistic position to raise loan interest rates, leading to high financing costs for enterprises and rent-seeking behavior. This results in a “predation effect”, crowding out funds for green innovation, which hampers the R&D efforts for green innovation. As financial development progresses, the dual roles of finance—serving the real economy and seeking profits—persist. The ultimate direction of finance’s impact on the real economy depends on the balance of power between these two functions.

3. Research Design

3.1. Model Setting

The industrial development levels vary significantly across different cities in China. To control for individual differences, this paper constructs a Fixed-Effects model to examine the impact of financial development on the industrial GTIE:

$$Y_{it}=\beta +c{Fd}_{it}+\gamma X_{it}+f_i+{\epsilon }_{it}$$

(1)

In Model (1), $Y$ denotes the industrial GTIE, where $i$ stands for the city, $t$ stands for the year, $Fd$ signifies the level of financial development, $X$ represents the set of control variables, $f_i$ captures the individual fixed effects, and ${\epsilon }_{it}$ stands for the error term. If the coefficient β is significantly positive, it indicates that financial development can improve the industrial GTIE.

To analyze the mechanisms through which financial development influences industrial GTIE, we construct a mediation model based on Model (1):

$$M_{it}=\beta +a{Fd}_{it}+\gamma X_{it}{+f}_i+{\epsilon }_{it}$$

(2)

Model (2) introduces $M$ as the mediating variable, representing financial constraints and investment openness, respectively. Model (2) primarily assesses how financial development affects the intermediary variables. If the coefficient $a$ is significantly positive, it indicates that financial development positively influences both enterprise financial constraints and investment openness.

$$Y_{it}=\beta +c^{\prime }{Fd}_{it}+b{M}_{it}{+\gamma X_{it}+f}_i+{\epsilon }_{it}$$

(3)

Model (3) examines the direct effect of financial development on industrial GTIE and the indirect effect through the mediating variables. If the coefficients $b$ and $c^{\prime }$ are significantly positive at the 10% level, with $c^{\prime }$ smaller than the value of $c$ in Model (1), it indicates that financial development can enhance industrial GTIE by alleviating enterprise financial constraints and promoting investment openness.

3.2. Data

This paper utilizes panel data spanning from 2008 to 2020, covering 288 cities at the prefecture level in China. These data cover 31 provinces and municipalities (excluding county-level cities, autonomous prefectures). The original data are sourced from the China Urban Statistical Yearbook, China City Statistical Bulletin, and China National Intellectual Property Administration. The data were compiled by the National Bureau of Statistics of China and the State Intellectual Property Office. The utilization of data provided by official institutions in this article ensures the authenticity of the information to a certain extent. Descriptive statistics for the relevant variables are presented in

Table 1. Descriptive statistics.

Variables	Obs.	Mean	Std. Dev.	Min.	Max.
$\mathrm{G}\mathrm{i}\mathrm{e}$	3744	0.3519	0.1901	0.0234	3.8664
$\mathrm{G}\mathrm{t}\mathrm{f}\mathrm{p}$	3744	1.8542	1.5067	0.1087	38.9197
$\mathrm{F}\mathrm{d}$	3744	0.9589	0.5913	0.0175	5.7483
$\mathrm{F}\mathrm{c}$	3744	1.4507	1.5218	0.0053	19.2601
$\mathrm{F}\mathrm{d}\mathrm{i}$	2967	1.3036	5.2708	0.0100	148.34
$\mathrm{P}\mathrm{g}\mathrm{d}\mathrm{p}$	3744	2.3490	1.4961	0.3284	9.9082
$\mathrm{T}\mathrm{i}$	3744	0.0156	0.0153	0.0004	0.1656
$\mathrm{T}\mathrm{d}\mathrm{o}$	3744	0.1957	0.3292	0.0000	3.2573
$\mathrm{C}\mathrm{a}\mathrm{p}$	3744	0.7828	0.3156	0.0872	2.6702
$\mathrm{H}\mathrm{c}$	3744	9.2253	1.1815	7.1211	14.8305
$\mathrm{I}\mathrm{e}\mathrm{s}$	3744	0.6438	0.2902	0.0307	3.5973
$\mathrm{G}\mathrm{o}\mathrm{v}$	3744	0.1984	0.1064	0.0426	1.0268

.

3.2.1. Dependent Variable

This paper measures the industrial GTIE across various cities in China from both static and dynamic perspectives. We utilize the Super-SBM model to assess the static industrial GTIE. Compared to traditional DEA models, the Super-SBM model addresses the slack variable issue, enabling comparisons among efficient DMUs [44,45,46]. However, the Super-SBM model only measures static efficiency and cannot capture the dynamic evolution of efficiency trends or efficiency levels at different DMUs. Therefore, we further employ the Network DEA model to measure the dynamic industrial GTIE.

The Network DEA model can measure the efficiency of innovation production activities across different stages. Innovation activities consist of continuous stages. Following the existing Innovation Production Process (IPP) analysis framework [47], this paper divides green technology innovation activities into two stages: technological research and development and achievement transformation. Therefore, the calculation of the industrial GTIE is divided into two subprocesses [48,49,50,51], as illustrated in

Table 2. The green technology innovation efficiency index system.

Stage	First-Grade Indicators	Second-Grade Indicators	Indicator Description
Technical research and development	Input index	Labor input	Number of year-end employees in the industrial sector
		Capital input	Total assets of industrial enterprises above designated size
	Output index	R&D output	Number of invention patent applications
Achievement transformation	Input index	R&D input	Number of invention patent applications
		Energy input	Regional total electricity consumption
	Desirable output	Economic output	Regional Real GDP
	Undesirable outputs	Pollutant emissions	Regional industrial wastewater discharge
			Regional industrial sulfur dioxide emissions
			Regional industrial smoke (dust) emissions

Notes: The data are sourced from the National Bureau of Statistics of China and China National Intellectual Property Administration.

.

The first process is the technological research and development stage, where we initially measure the efficiency of the green innovation process related to R&D output. As we are currently unable to separate green R&D personnel and funds from traditional R&D personnel and funds, we use the number of employees and asset scale in regional industrial sectors as proxies for labor and capital inputs in this stage. We use the number of invention patents with high technological content as the output indicator.

The second process is the achievement transformation stage, where we measure the efficiency of technological innovation in enhancing industrial green growth. Here, we use the output of the first stage of technological research and development as the input indicator for the second stage and incorporate energy input variables. The output indicators for the second stage include expected and unexpected outputs, with expected output measured using the real GDP of the city and unexpected output measured using the emissions of industrial wastewater, sulfur dioxide, and particulate matter.

Based on efficiency calculations from two subprocesses, we utilize the input-oriented Network DEA model to compute overall efficiency [52] and then use the overall efficiency adjusted for biases to measure the industrial GTIE.

To mitigate potential bias from endogenous variables affecting estimation results, we also examine the industrial GTIE from the perspective of innovation outcomes. Specifically, we use the number of green invention patent applications and the number of patents granted with a lag of one period as alternative indicators for the industrial GTIE [53]; this approach was employed to conduct a robustness test.

3.2.2. Independent Variable

Financial development serves as the core independent variable in this paper. The McKinnon Index (M2/GNP) and Goldsmith Index (FIR) are the two most prevalent methods for assessing financial development. However, Levine and Zervos [54] suggest that the McKinnon Index fails to effectively measure the sources of liabilities and resource allocation within the financial system. Therefore, the Goldsmith Index is employed as a metric to gauge financial development. Given that the influence of finance on industrial development mainly manifests in credit support, we utilize the ratio of regional financial institution loans to gross domestic product as a measure [55]. A higher ratio indicates greater loan activity and a higher level of financial development. The Goldsmith Index primarily reflects the level of financial development in a region based on changes in financial interconnectedness rates [56]. Due to the imperfections in China’s financial markets, we also examine the level of financial development in various cities from the perspective of financial institution activity. We use per capita deposit and loan amounts in cities to reflect the activity of financial institutions.

The measurement methods for the other relevant variables are detailed in

Table 3. Variable Description.

Variables	Symbols	Description	Source
Financial development	$\mathrm{F}\mathrm{d}$	the ratio of financial institution loans to GDP	China Urban Statistical Yearbook
Financial constraint	$\mathrm{F}\mathrm{c}$	the proportion of liquidity assets to fixed assets
Investment openness	$\mathrm{F}\mathrm{d}\mathrm{i}$	the quantity of foreign direct investment projects
Economic development	$\mathrm{P}\mathrm{g}\mathrm{d}\mathrm{p}$	per capita GDP adjusted for price factors
Science Technological Investment	$\mathrm{T}\mathrm{i}$	the proportion of scientific and technological expenditures to total government expenditure
Trade openness	$\mathrm{T}\mathrm{d}\mathrm{o}$	the ratio of total imports and exports to GDP
Fixed asset investment	$\mathrm{C}\mathrm{a}\mathrm{p}$	the ratio of total fixed asset investment to GDP
Human capital	$\mathrm{H}\mathrm{c}$	the average years of education attained by the population
Ownership structure	$\mathrm{I}\mathrm{e}\mathrm{s}$	the proportion of non-private unit industrial sector employees to total industrial employees at year-end
Government intervention	$\mathrm{G}\mathrm{o}\mathrm{v}$	the ratio of total government fiscal expenditure to GDP

Notes: The data are sourced from the National Bureau of Statistics of China.

, including the mediating variables and control variables. Previous studies have shown that industrial GTIE is also influenced by factors such as economic development [57], Science and Technological Investment [58], trade openness [59], fixed asset investment [60], human capital [61], ownership structure [62], and government intervention [63]. Therefore, we introduced these factors as control variables.

3.3. Efficiency Measurement Model

3.3.1. Super-SBM Model

This paper utilizes a Super-Efficiency Slack-Based Measure (Super-SBM) model, which includes unexpected output, to calculate the industry GTIE across different cities [64]. The calculation method is as follows:

$$\begin{gathered} Min\rho =\frac{1+\frac{1}{m}\sum _{i=1}^m\raisebox{1ex}{${s_i}^{-}$}\!\left/ \!\raisebox{-1ex}{$x_{ik}$}\right.}{1-\frac{1}{q_1+q_2}(\sum _{r=1}^{q_1}\raisebox{1ex}{${s_r}^{+}$}\!\left/ \!\raisebox{-1ex}{$y_{rk}$}\right.+\sum _{t=1}^{q_2}\raisebox{1ex}{${s_t}^{b-}$}\!\left/ \!\raisebox{-1ex}{$b_{tk}$}\right.)} \\ s.t.\sum _{j=1,j\ne k}^n{x}_{ij}{\lambda }_j-{s_i}^{-}\le x_{ik} \\ \sum _{j=1,j\ne k}^n{y}_{rj}{\lambda }_j+{s_r}^{+}\ge y_{rk} \\ \sum _{j=1,j\ne k}^n{b}_{tj}{\lambda }_j-{s_t}^{b-}\le b_{tk} \\ 1-\frac{1}{q_1+q_2}(\sum _{r=1}^{q_1}\raisebox{1ex}{${s_r}^{+}$}\!\left/ \!\raisebox{-1ex}{$y_{rk}$}\right.+\sum _{t=1}^{q_2}\raisebox{1ex}{${s_t}^{b-}$}\!\left/ \!\raisebox{-1ex}{$b_{tk}$}\right.)>0 \\ \lambda ,s^{-},s^{+}\ge 0;i=1,2\cdots m;r=1,2\cdots q;j=1,2\cdots n(r\ne k) \end{gathered}$$

(4)

In Formula (4), $\rho$ denotes the GTIE value of the evaluated DMU. $x_{ik}$, $y_{rk}$, and $b_{tk}$ respectively denote the corresponding input, expected output, and non-expected output values for the DMU. $q_1$ and $q_2$ stand for the number of elements for expected and non-expected outputs. ${s_i}^{-}$, ${s_r}^{+}$, and ${s_t}^{b-}$ represent the slack variables for input, expected output, and non-expected output, respectively. Based on the calculation result derived from Formula (4), it is denoted as $Gie$.

3.3.2. The Network DEA Model

The network DEA model provides a detailed look inside each DMU, unlike the traditional DEA model, allowing for step-by-step measurement of production process efficiency. Assuming there are n DMUs, ${DMU}_j$ ($j=1,2,\cdots ,n$), each DMU has K nodes, ($K=1,2,\cdots ,k$). $m_k$ denotes the number of input indicators at node $k$, and $r_k$ denotes the number of output indicators at node $k$. $X_j^1={(X_{1j}^1,X_{2j}^1,\cdots ,X_{m1j}^1)}^T$ represents the input vector of the ${DMU}_j$ in the first stage, and $Z_j^{(1,2)}=(Z_{1j}^{\left(1,2\right)},Z_{2j}^{\left(1,2\right)},\cdots ,Z_{tj}^{\left(1,2\right)})$ denotes the intermediate output of the ${DMU}_j$ in the first stage. $X_j^2={(X_{1j}^2,X_{2j}^2,\cdots ,X_{m1j}^2)}^T$ represents the input vector of the ${DMU}_j$ in the second stage. $Y_j^2={(Y_{1j}^2,Y_{2j}^2,\cdots ,Y_{rj}^2)}^T$ denotes the output vector of the ${DMU}_j$ in the second stage. $w^k$ represents the weight of node $k$.

Under constant returns to scale, the production possibility set during the technological research and development stage is represented as $\sum _{j=1}^n{x}_j^1{\lambda }_j^1\le X^1$, $\sum _{j=1}^n{z}^{(1,2)}{\lambda }_j^1\ge Z^{(1,2)}$. During the achievement transformation stage, the production possibility set is represented as $\sum _{j=1}^n{x}_j^2{\lambda }_j^2\le X^2$, $\sum _{j=1}^n{z}^{(1,2)}{\lambda }_j^2\le Z^{(1,2)}$, $\sum _{j=1}^n{y}_j^2{\lambda }_j^2\ge Y^2$. By introducing the constraint $\sum _{j=1}^n{\lambda }_j^2=1$(${\lambda }_j^k\ge 0,k=1,2\cdots ,K)$, we obtain the production possibility set under variable returns to scale.

At this point, the calculation method for the network DEA model is as follows:

$${\rho }^{\ast }=min\frac{\sum _{k=1}^2{w}^k(1-\frac{1}{m_k}\sum _{i=1}^{m_k}\raisebox{1ex}{${s_{i0}}^{k-}$}\!\left/ \!\raisebox{-1ex}{$x_{i0}^k$}\right.)}{\sum _{k=1}^2{w}^k(1+\frac{1}{r_k}\sum _{r=1}^{r_k}\raisebox{1ex}{${s_{r0}}^{k+}$}\!\left/ \!\raisebox{-1ex}{$y_{r0}^k$}\right.)}$$

(5)

In Formula (5), ${\rho }^{\ast }$ represents the overall efficiency of DMU. Each DMU has $m$ types of inputs denoted as $x_i$ and $r$ types of outputs denoted as $y_r$. ${s_i}^{k-}$ and ${s_r}^{k+}$ are slack variables for input and output, respectively, and they are greater than or equal to zero.

The efficiency calculated using the network DEA model above does not account for the impact of frontier sampling variation on the results [65]. To address this, we further utilize bootstrapping estimation to correct for biases in the results [50,51,66]. We employ multiple sampling methods to estimate efficiency scores, alleviating issues caused by sampling fluctuations and ultimately obtaining efficiency results corrected for biases; it is denoted as $Gtfp$.

4. Empirical Results and Analysis

4.1. Benchmark Regression Analysis

Table 4. Benchmark regression results.

Variables	(1)	(2)	(3)	(4)
	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$
$\mathrm{F}\mathrm{d}$	0.0382 ***	0.5335 ***	0.0299 ***	0.3312 ***
	(0.0071)	(0.0720)	(0.0085)	(0.0873)
$\mathrm{P}\mathrm{g}\mathrm{d}\mathrm{p}$			0.1030 ***	0.3691 ***
			(0.0060)	(0.0296)
$\mathrm{T}\mathrm{i}$			2.3194 ***	23.3868 ***
			(0.2746)	(2.8130)
$\mathrm{T}\mathrm{d}\mathrm{o}$			−0.0208	−0.0954
			(0.0152)	(0.1554)
$\mathrm{C}\mathrm{a}\mathrm{p}$			0.0824 ***	1.1541 ***
			(0.0099)	(0.1012)
$\mathrm{H}\mathrm{c}$			−0.0103 **	−0.0005
			(0.0050)	(0.0509)
$\mathrm{I}\mathrm{e}\mathrm{s}$			0.0409 ***	−0.0515
			(0.0126)	(0.1292)
$\mathrm{G}\mathrm{o}\mathrm{v}$			0.0047	0.6047
			(0.0553)	(0.5665)
$\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{t}$	0.3153 ***	1.3425 ***	0.0666	−1.5093 ***
	(0.0071)	(0.0721)	(0.0476)	(0.4876)
FE	Yes	Yes	Yes	Yes
Obs.	3744	3744	3744	3744
${\mathrm{R}}^2$	0.0084	0.0156	0.1362	0.1302
Cities	288	288	288	288

Notes: standard errors are shown in parentheses; ** p < 0.05, *** p < 0.01.

presents the regression results using the Fixed-Effects model to identify the causal relationship between financial development and industrial GTIE. Regardless of whether other variables are controlled, the coefficients of financial development all significantly lie above the 1% level, which shows that financial development significantly drives the enhancement of industrial GTIE. Regarding controlling variables, Economic Development, Fixed Asset Investment, and Industrial Ownership Structure have positive effects on industrial GTIE. Economically developed regions have relative advantages in innovation resources and institutional environments. Moreover, the increase in regional fixed assets strengthens infrastructure construction. These factors collectively lead to reductions in the costs of industrial green R&D, production, and transportation costs. This helps enterprises reduce costs and increase innovation efficiency. In addition, green technology innovation is mainly concentrated in strategic emerging industries with a high degree of capital intensity. Compared to private enterprises, non-private entities possess stronger market power [67], with financial development exerting a greater impact on their GTIE. Trade openness is negatively correlated to the industrial GTIE. This may be due to China’s industrial export structure biased toward labor- and resource-intensive products. Export growth has led to a biased factor structure change, which is not conducive to green technology innovation [68].

4.2. Robustness Test

4.2.1. Replacing Dependent Variable

To mitigate the impact of measurement bias in the dependent variable, this paper employs the quantity of applied green invention patents ($Gupa$) and the quantity of granted green invention patents ($Gipa$) lagged by one period as proxy indicators for the industrial GTIE. After substituting the measure for evaluating industrial GTIE, the regression results remain largely unchanged (

Table 5. Regression results with replacement of dependent variable.

Variables	(1)	(2)	(3)	(4)
	$\mathbf{G}\mathbf{u}\mathbf{p}\mathbf{a}$	$\mathbf{G}\mathbf{i}\mathbf{p}\mathbf{a}$	$\mathbf{G}\mathbf{u}\mathbf{p}\mathbf{a}$	$\mathbf{G}\mathbf{i}\mathbf{p}\mathbf{a}$
$\mathrm{F}\mathrm{d}$	5.7677 ***	1.1478 ***	4.7390 ***	0.7712 ***
	(0.3335)	(0.1324)	(0.3977)	(0.1590)
Control	No	No	Yes	Yes
FE	Yes	Yes	Yes	Yes
Obs.	3744	3456	3744	3456
${\mathrm{R}}^2$	0.0797	0.0232	0.2132	0.1326
Cities	288	288	288	288

Notes: standard errors are shown in parentheses; *** p < 0.01.

). The coefficient for financial development remains robustly positive, indicating that financial development indeed enhances industrial green production capacity and improves industrial GTIE. The signs and significance of the coefficients for the control variables remain unchanged, ensuring the robustness of the conclusions.

4.2.2. Replacing Independent Variable

To mitigate the impact of measurement bias in the independent variable, we use per capita deposit ($Pcs$) and per capita loan amounts ($Pcl$) as proxy indicators of financial development. Model (1) is regressed again, with the results shown in

Table 6. Regression results with replacement of core independent variable.

Variables	(1)	(2)	(3)	(4)
	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$
$\mathrm{P}\mathrm{c}\mathrm{s}$	0.0071 ***		0.0677 ***
	(0.0006)		(0.0064)
$\mathrm{P}\mathrm{c}\mathrm{l}$		0.0074 ***		0.0727 ***
		(0.0007)		(0.0075)
Control	No	No	Yes	Yes
FE	Yes	Yes	Yes	Yes
Obs.	3744	3744	3744	3744
${\mathrm{R}}^2$	0.1639	0.1583	0.1538	0.1496
Cities	288	288	288	288

Notes: standard errors are shown in parentheses; *** p < 0.01.

. After replacing the evaluation method of financial development, the coefficients of $Pcs$ and $Pcl$ remain significantly positive, which also verifies that the benchmark regression results are robust and reliable.

4.2.3. Controlling Environmental Regulation and Firm Size

Environmental regulation and firm size are also crucial factors influencing industrial GTIE [69]. On one hand, environmental regulations increase the cost of pollution control for firms, squeezing out research and development investment and impeding the improvement of green innovation efficiency. On the other hand, appropriate environmental regulations can stimulate enterprise innovation. The benefits of innovation can offset the negative costs brought about by regulations, thereby enhancing GTIE. We construct a comprehensive indicator to measure the degree of environmental regulations from the perspective of pollutant emissions. The calculation method is as follows:

$${Er}_i=\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$\sum _j^m\frac{\raisebox{1ex}{$P_{ij}$}\!\left/ \!\raisebox{-1ex}{${Gdp}_i$}\right.}{\sum _{i=1}^n\frac{\raisebox{1ex}{$P_{ij}$}\!\left/ \!\raisebox{-1ex}{${Gdp}_i$}\right.}{n}}$}\right.$$

(6)

In Formula (6), ${Er}_i$ represents the environmental regulation intensity, $P_{ij}$ stands for the emissions of pollutant $j$ in city $i$, $P_{ij}$ divided by ${Gdp}_i$ represents the emissions of pollutant $j$ per unit output value in city $i$, and $\sum _{i=1}^n\frac{\raisebox{1ex}{$P_{ij}$}\!\left/ \!\raisebox{-1ex}{${Gdp}_i$}\right.}{n}$ represents the average level of emissions of pollutant $j$ per unit output value nationwide, where $m$ is the number of types of pollutants. We calculate the ratio of emissions of pollutant $j$ per unit output value in the city to the average emissions of pollutant $j$ per unit output nationwide. A smaller ratio indicates higher environmental regulation intensity. For convenience, reciprocal processing is adopted.

Firm size is a critical factor affecting enterprise innovation capability. Schumpeter’s hypotheses suggest a close connection between firm size and innovation capability. The larger the firm, the greater the investment in R&D, consequently leading to stronger innovation capability. In this paper, regional industrial added value divided by the number of large-scale industrial enterprises is used as a proxy indicator of firm size ($Size$). Environmental regulations and firm size are separately incorporated into Model (1), and the results are shown in

Table 7. Controlling the environmental regulation and firm size.

Variables	(1)	(2)	(3)	(4)
	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$
$\mathrm{F}\mathrm{d}$	0.0309 ***	0.3708 ***	0.0203 **	0.2375 ***
	(0.0085)	(0.0861)	(0.0085)	(0.0876)
$\mathrm{E}\mathrm{r}$	0.0024 **	0.0983 ***
	(0.0010)	(0.0098)
$\mathrm{S}\mathrm{i}\mathrm{z}\mathrm{e}$			0.0454 ***	0.4456 ***
			(0.0059)	(0.0609)
Control	Yes	Yes	Yes	Yes
FE	Yes	Yes	Yes	Yes
Obs.	3744	3744	3744	3456
${\mathrm{R}}^2$	0.1377	0.1548	0.1506	0.1435
Cities	288	288	288	288

Notes: standard errors are shown in parentheses; ** p < 0.05, *** p < 0.01.

.

Regression results show that the environmental regulations and firm size have a significant positive effect on industrial GTIE. After controlling for the influence of environmental regulations and firm size, the coefficient of financial development remains significantly positive. This confirms the reliability of our conclusions.

4.3. Endogeneity Test

To address endogeneity issues, the existing literature suggests that the Difference Generalized Method of Moments (GMM) model can resolve the endogeneity problem without the need to search for external instruments [21,70]. The two-stage least squares method (2SLS) is a specific case of the GMM model. When instruments are highly exogenous and strongly correlated with the regressors, the 2SLS model outperforms the GMM model [71]. To ensure the robustness of the results, this paper employs both GMM and 2SLS estimation methods to tackle the endogeneity issues. Additionally, we use the external shock of the sci-tech finance pilot policy as a proxy variable for financial development. By regressing the benchmark model on external instruments, we try to further address the endogeneity issue.

4.3.1. Difference GMM Model

To avoid the influence of estimation method selection bias on the results, we use the Difference GMM estimation to retest the baseline results [70]. The findings are presented in

Table 8. Regression results of Difference GMM.

Variables	(1)	(2)	(3)	(4)
	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$
$\mathrm{L}.\mathrm{G}\mathrm{i}\mathrm{e}$	0.3155 ***		0.2131 ***
	(0.0133)		(0.0115)
$\mathrm{L}.\mathrm{G}\mathrm{t}\mathrm{f}\mathrm{p}$		0.6210 ***		0.5981 ***
		(0.0015)		(0.0015)
$\mathrm{F}\mathrm{d}$	0.0552 ***	0.4310 ***	0.0733 ***	0.1657 ***
	(0.0091)	(0.0395)	(0.0115)	(0.0513)
Control	No	No	Yes	Yes
FE	Yes	Yes	Yes	Yes
Obs.	3168	3168	3168	3168
Cities	288	288	288	288

Notes: standard errors are shown in parentheses; *** p < 0.01.

. The coefficients of $L.Gie$ and $L.Gtfp$ are significantly positive, indicating the persistence influence of the GTIE. Regions with higher industrial GTIE in the previous period continue to exhibit greater potential for industrial green development in the next period. Even after considering the lagged effects of GTIE, the coefficient for financial development remains significantly positive, same as the benchmark results.

4.3.2. Two-Stage Least Squares Method

Considering that the current industrial GTIE is not influenced by the previous period’s financial development, there is also a significant correlation between the current level of financial development and that of the preceding period, which meets the requirements of relevance and exogeneity for instrumental variable selection [21,71]. Therefore, we utilize the lagged one-period financial development as an instrumental variable and employ the 2SLS method for examination. The results are displayed in

Table 9. Regression results of 2SLS method.

Variables	(1)	(2)	(3)	(4)
	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$
$\mathrm{F}\mathrm{d}$	0.0418 ***	0.4399 ***	0.0573 ***	0.4438 ***
	(0.0083)	(0.0853)	(0.0106)	(0.1092)
first-stage	0.9228 ***	0.9228 ***	0.8273 ***	0.8273 ***
(IV)	(0.0087)	(0.0087)	(0.0098)	(0.0098)
Cragg-Donald Wald F	11,290.14	11,290.14	7010.42	7010.42
KP rk Wald F	1427.67	1427.67	798.09	798.09
Control	No	No	Yes	Yes
FE	Yes	Yes	Yes	Yes
Obs.	3456	3456	3456	3456
${\mathrm{R}}^2$	0.6169	0.3742	0.6592	0.4352
Cities	288	288	288	288

Notes: standard errors are shown in parentheses; *** p < 0.01.

. In the first-stage regression results, all coefficients of the instrumental variables are significantly positive, and the Cragg–Donald Wald F and KP rk Wald F statistics are significant, effectively addressing concerns regarding weak instrumental variables. The 2SLS estimation results similarly demonstrate a positive connection between financial development and industrial GTIE.

4.3.3. Instrumental Variables Method

The sci-tech finance pilot policy is implemented by China to break through the financing bottleneck of enterprises and facilitate the translation of technological achievements. After policy implementation, pilot areas innovate in their fiscal and technological investment approaches, continuously optimizing the local science and technology financial ecosystem. Under the guidance of local governments and various departments, financial and private capital are encouraged to support innovation among technology-oriented small and medium-sized enterprises as much as possible. The level of financial development in the pilot areas is much higher than that in non-pilot areas. Therefore, the sci-tech finance pilot policy can serve as a substitute metric for financial development.

During the period from 2010 to 2020, the Chinese Ministry of Science and Technology successively released two batches of lists of the sci-tech finance pilot areas, covering 25 regions and including 50 cities. This provides an effective quasi-natural experiment for examining the influence of the pilot policy on industrial GTIE. To alleviate potential endogeneity issues, we construct a multi-period DID model:

$$Y_{it}=\alpha +\beta {Treat}_i\ast {Time}_t+\gamma X_{it}+f_i+{\epsilon }_{it}$$

(7)

In Model (7), ${Treat}_i$ is a regional dummy variable indicating whether the city is the pilot city. ${Treat}_i=1$ indicates being within the scope of the pilot policy area, which is the treatment group; otherwise, they are the control group. ${Time}_t$ is a time dummy variable indicating whether the pilot policy is implemented. ${Treat}_i\ast {Time}_t$ represents the policy variable. The regression results are displayed in

Table 10. Regression results of multi-period DID (instrumental variable method).

Variables	(1)	(2)	(3)	(4)
	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$
$\mathrm{T}\mathrm{i}\mathrm{m}\mathrm{e}\ast \mathrm{T}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{t}$	0.1324 ***	1.3045 ***	0.0891 ***	0.9232 ***
	(0.0104)	(0.1067)	(0.0103)	(0.1059)
Control	No	No	Yes	Yes
FE	Yes	Yes	Yes	Yes
Obs.	3744	3744	3744	3744
${\mathrm{R}}^2$	0.0445	0.0415	0.1514	0.1454
Cities	288	288	288	288

Notes: standard errors are shown in parentheses; *** p < 0.01.

.

In Table 10, the coefficients of the policy variable are significantly positive, and the sci-tech finance pilot policy can promote industrial GTIE. After considering endogeneity issues, the empirical analysis results remain unchanged, indicating that the benchmark results are robust.

4.4. Mechanism Analysis

Due to space constraints, we only report the regression results of the mediation model with $Gie$ as the dependent variable. In

Table 11. Regression results of mechanism analysis.

Variables	(1)	(2)	(3)	(4)	(5)
	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{F}\mathbf{c}$	$\mathbf{F}\mathbf{d}\mathbf{i}$	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{i}\mathbf{e}$
$\mathrm{F}\mathrm{d}$	0.0409 ***	0.5695 ***	0.5734 *	0.0292 ***	0.0186 *
	(0.0085)	(0.0439)	(0.3030)	(0.0087)	(0.0096)
$\mathrm{F}\mathrm{c}$				0.0205 ***
				(0.0033)
$\mathrm{F}\mathrm{d}\mathrm{i}$					0.0015 **
					(0.0006)
Control	Yes	Yes	Yes	Yes	Yes
FE	Yes	Yes	Yes	Yes	Yes
Obs.	3744	3744	2967	3744	2967
${\mathrm{R}}^2$	0.1183	0.0742	0.0441	0.1282	0.1945
Cities	288	288	255	288	255

Notes: Column (1) reflects the regression results of model (1), while columns (2)–(3) are the results of model (2); columns (4)–(5) are the results of model (3). Standard errors are shown in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

, column (1) reports the total effect of financial development on industrial GTIE. Columns (2)–(3) report the influence of financial development on firm financial constraints and investment openness. Columns (4)–(5) report the direct impact of financial development on industrial GTIE and the indirect impact mediated by mediating variables.

In columns (2)–(3), the coefficient of financial development is significantly positive, indicating that financial development effectively alleviates the financial constraints of industrial enterprises, providing them with financial support and easing the dilemma of insufficient funds in the process of green technology innovation. Additionally, financial development improves the investment environment, enhances investment openness, and attracts capital inflows for R&D. In columns (4)–(5), the coefficients of $Fd$, $Fc$, and $Fdi$ are significantly positive, and the coefficients of $Fd$ in columns (4)–(5) are smaller than that in column (1). This suggests that the mediating effects of financial constraints and investment openness are established. Financial development not only has a direct effect on industrial GTIE but also indirectly enhances industrial GTIE through the facilitation of financing and investment openness channels.

5. Heterogeneity Analysis

Owing to diverse regional development strategies in China, coastal cities exhibit a considerably higher level of financial development compared to inland cities. The eastern region generally boasts superior financial development compared to the central and western regions, while financial development tends to be more robust in large cities than in small to medium-sized ones. Therefore, we divide the 288 cities in China into different sub-samples and examine the “support” and “predatory” effects of financial development on industrial GTIE across different types of cities.

5.1. Coastal and Inland Cities

China’s regional economic policies have gone through three stages: the coastal areas leading in development, the common development of coastal and inland areas, and current initiatives aimed at achieving coordinated development across all regions. With the dual advantages of national regional economic policy tilt and unique geographical location, coastal cities have achieved rapid economic growth and far higher levels of financial development than inland areas. In this paper, the data samples are segregated into coastal and inland cities based on whether the target cities have a coastline. There are a total of 52 coastal cities and 236 inland cities. The regression results are presented in

Table 12. Regression results for coastal and inland cities.

Variables	(1)	(2)	(3)	(4)
	Coastal		Inland
	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$
$\mathrm{F}\mathrm{d}$	−0.0004	−1.3408 ***	0.0327 ***	0.5162 ***
	(0.0227)	(0.3959)	(0.0093)	(0.0736)
Control	Yes	Yes	Yes	Yes
FE	Yes	Yes	Yes	Yes
Obs.	676	676	3068	3068
${\mathrm{R}}^2$	0.1417	0.0673	0.1453	0.2401
Cities	52	52	236	236

Notes: standard errors are shown in parentheses; *** p < 0.01.

.

The results show that in coastal cities, the coefficient of financial development lacks significance in column (1). However, in column (2), it passes the significance test at the 1% level, which suggests that financial development exerts a significant inhibitory influence on the industrial GTIE in coastal cities and that overpowering financial institutions crowd out the positive impact of financial development on industrial green innovation; the “predatory effect” outweighs the “support effect” at this point. In columns (3) and (4), the coefficient of financial development exhibits positive significance at the 1% level. This signifies that the financial development significantly promotes the industrial GTIE in inland cities, aligning with anticipated outcomes.

5.2. Different Geographical Location

The extent of financial development varies widely across different regions of China, displaying distinct spatial imbalances and polarization characteristics, notably with the eastern region exhibiting higher levels compared to the central and western regions. To explore the heterogeneous impact of financial development on the industrial GTIE, we divide the sample data according to geographical location, resulting in 100 cities in the east, 109 in the center, and 79 in the west. The results are presented in

Table 13. Regression results for eastern and central/western regions.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Eastern		Central		Western
	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$
$\mathrm{F}\mathrm{d}$	0.0246	−0.4765 **	0.0561 ***	0.8981 ***	0.0371 **	0.3577 ***
	(0.0164)	(0.2282)	(0.0162)	(0.1300)	(0.0151)	(0.1208)
Control	Yes	Yes	Yes	Yes	Yes	Yes
FE	Yes	Yes	Yes	Yes	Yes	Yes
Obs.	1300	1300	1417	1417	1027	1027
${\mathrm{R}}^2$	0.1987	0.0714	0.2354	0.3428	0.0703	0.1933
Cities	100	100	109	109	79	79

Notes: standard errors are shown in parentheses; ** p < 0.05, *** p < 0.01.

.

According to Table 13, in the regression results of the eastern region, the financial development coefficient exhibits positivity but lacks significance in column (1), is negative in column (2), and passes the 1% significance threshold. This implies that financial development significantly hinders the industrial GTIE in the eastern region. In contrast, in columns (3)–(6), the financial development coefficients are notably positive, signifying that financial development substantially fosters the industrial GTIE in these regions. At this point, the service function of the financial sector to the industrial economy outweighs its profit-seeking function.

5.3. Different City Sizes

Generally, city size has a positive contribution to financial development, with larger cities having higher levels of financial development. In order to study the relationship between the effects of financial development on the industrial GTIE and city size. The paper groups the sample according to the criteria outlined in the “Notice on Adjusting the Standards for Classifying City Sizes” issued by the State Council of China in 2014. Specifically, cities with a permanent urban population exceeding 1 million are categorized as large cities, those with populations between 500,000 and 1 million are classified as medium-sized cities, and those with populations below 500,000 are deemed small cities. Following this classification, there are 75 large cities, 94 medium-sized cities, and 119 small cities. After segmenting the sample based on city size, empirical tests are conducted separately using Model (1). The regression results are detailed in

Table 14. Regression results for cities of different sizes.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Large		Medium		Small
	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$	$\mathbf{G}\mathbf{i}\mathbf{e}$	$\mathbf{G}\mathbf{t}\mathbf{f}\mathbf{p}$
$\mathrm{F}\mathrm{d}$	0.0133	0.2948 ***	0.0325 **	0.7164 ***	0.0401 **	0.0890
	(0.0126)	(0.0923)	(0.0144)	(0.1154)	(0.0159)	(0.1856)
Control	Yes	Yes	Yes	Yes	Yes	Yes
FE	Yes	Yes	Yes	Yes	Yes	Yes
Obs.	975	975	1222	1222	1547	1547
${\mathrm{R}}^2$	0.2557	0.3176	0.2583	0.3467	0.0857	0.0811
Cities	75	75	94	94	119	119

Notes: standard errors are shown in parentheses; ** p < 0.05, *** p < 0.01.

.

The findings reveal that the financial development coefficients are all positive across all city sizes. The $Gie$ coefficients for small cities exceed those of both medium-sized and large cities. The promoting effect of financial development on industrial GTIE decreases as the city size expands. This is related to the negative externality of urban size expansion, which is essentially a resource constraint issue in the process of urban development. The expansion of urban scale increases resource consumption. While social supply is limited, the crowding effect caused by resource scarcity leads to continuously rising innovation costs for enterprises, thereby diminishing resource utilization efficiency, which is unfavorable for enhancing the industrial GTIE.

6. Discussion

The green growth of industrial sectors plays a pivotal role in facilitating the real economy’s transition to a green and low-carbon model. In alignment with the objectives of achieving carbon peaking and carbon neutrality goals, the Chinese government has implemented a series of financial policies aimed at enhancing the extent of financial development to support green technological innovation in industries. Through the integration of panel data encompassing 288 cities in China from 2008 to 2020, we conducted empirical analysis using a variety of estimation methods to explore the positive impact of financial development on industrial GTIE and its mechanisms. As a result, we summarized two key implications.

6.1. Limitations and Research Perspectives

Due to limitations such as the availability of sample data and the research perspectives, the paper has certain limitations. Based on the limitations, we also propose corresponding directions for future research.

Firstly, the selection of indicators for measuring industrial GTIE is not comprehensive enough. Based on data availability, industrial total output value is chosen as a substitute indicator for expected output. However, indicators reflecting the output of green technology innovation should also include sales revenue of new products, the number of new product development projects, technology market turnover, etc. In future research, efforts can be made to construct a more comprehensive indicator system.

Secondly, there is a lack of validation at the enterprise level. This paper is based on the city level, but industrial enterprises are not only the main body of green technology innovation but also the main venue for financial activities. In the next step, validation can be conducted from a micro-perspective to examine the impact of financial development on green technology innovation efficiency in industrial enterprises.

6.2. Policy Implications

Firstly, promote the deepening development of finance. Financial development not only directly promotes industrial green development but also indirectly promotes the efficiency of industrial green technology innovation by alleviating corporate financial constraints and increasing investment openness. Therefore, the government should adopt effective policy tools to enhance the level of financial development, such as accelerating the construction of regional financial agglomeration areas, fostering financial functional areas with local characteristics, improving and implementing preferential policies for financing industrial enterprises’ green innovation projects, leveraging the guiding role of green finance, directing social funds into green and low-carbon industries, and creating clusters of green ecological industries. Increase support for environmentally friendly enterprises and energy-saving industrial enterprises, and promote industrial structural adjustment and upgrading through financial incentives.

Secondly, deepen financial-market-oriented reforms. When the financial market’s influence is too strong, financial development has an inhibitory effect on the efficiency of industrial green technology innovation. To reduce the negative impact of financial development, it is necessary to promote the transition from a “bank-dominated” to a “market-dominated” financial system, further improve the mechanism for capital market price formation, lower the actual interest rates on corporate loans, moderately reduce the entry barriers for financial markets, expand financial development, and promote healthy competition in financial markets, thereby reducing the regional monopolistic power of financial institutions, restraining their ability to extract, alleviating the dilemma of difficult and expensive financing for industrial enterprises, and better fulfilling the service function of finance for industrial green development.

7. Conclusions

This paper combines panel data from 288 cities in China from 2008 to 2020 and employs various methods such as the Fixed-Effects model, Difference GMM model, Two-Stage Least Squares, Mediation model, and Multi-period Double Difference model to examine the relationship between financial development and the industrial GTIE.

The research findings are as follows: (1) Financial development significantly promotes the industrial GTIE. (2) Alleviating financial constraints and increasing investment openness are effective mechanisms through which financial development influences the industrial GTIE. (3) The impact of financial development on industrial GTIE exhibits heterogeneity, manifested as a dual effect of “support” and “predation” across different markets. Financial development inhibits the GTIE in regions with stronger financial influence, such as coastal and eastern cities, while promoting it in inland and central-western cities. Moreover, this promotion effect decreases with the expansion of city size.

Green innovation in industrial sectors plays a crucial role in achieving the goals of carbon peaking and carbon neutrality. In future research, it is worthwhile to consider micro-level issues such as enterprise behaviors and performance. Further exploration can be undertaken to investigate the actual measures taken by enterprises in the process of low-carbon transformation and their implementation effects.