Impact of Digital Finance on Green Technology Innovation: The Mediating Effect of Financial Constraints

Abstract

Green technology innovation is crucial for achieving sustainable development. This paper establishes fixed effect and mediation effect models to study how digital finance influences corporate green technology innovation and the moderating role of financial constraints using the data of Chinese A-share public businesses from 2011 to 2020. The results show that, first, green technology innovation is facilitated by digital finance, and both the coverage breadth and use depth play important roles. Second, digital finance encourages business innovation in green technology by alleviating financial constraints. Third, in state-owned businesses and businesses located in the eastern regions, digital finance has a more visible driving impact on green technology innovation. The aforementioned findings offer insightful research to encourage the balanced growth of digital finance and better enable corporate green technology innovation.

1. Introduction

With the degradation in the global environment, the new round of technological competition and the global industrial revolution pays more attention to green technology innovation. As a symbol of productivity, sustainability, and a reduction in adverse effects on the production process [1], green technology innovation includes the innovation of new products, management styles, and services [2], and it is an important driving factor in attaining sustainable development [3]. In 2020, 40% of the cities in China did not satisfy the ambient air quality standard, and about one-third of the cities did not meet the national level II criteria for PM2.5 concentration. The high-speed economic development mode at the cost of environmental pollution is difficult to sustain. Therefore, China urgently has to rely on green technology to prevent environmental pollution and achieve sustainable growth. Green technology innovation considers corporate interests and environmental factors and is the fundamental way to balance economic development and the ecological environment.

In terms of supporting green technology innovation of enterprises, traditional financial institutions have the problem of mismatching financial resources and low financing efficiency [4], which makes it difficult to meet corporate capital needs. In comparison, digital finance connects information technology with traditional finance, which complements the flaws of traditional finance and gives greater vitality to technological innovation. On the one hand, digital finance can lower the cost of credit assessment, speed up loan procedures, and open up new financing avenues for corporate innovation [4]. It enlarges the coverage of financial services [5], lowers the financial services threshold, and is helpful in enabling corporate technology innovation [6]. Green technology innovation, on the other hand, can give businesses new market opportunities, boost their competitiveness, and help realize their sustainability development [7]. Studying how digital finance promotes green technology innovation is of great significance.

Based on the listed companies’ data from 2011 to 2020, this paper researches the connection between digital finance and green technology innovation. First, this paper studies how digital finance and its three dimensions influence green technology innovation. Secondly, this paper studies the mechanism of green technology innovation driven by digital finance. Finally, heterogeneity analysis is used to deeply research the influential mechanism of digital finance on green technology innovation.

The remainder of this paper is structured as follows: Section 2 is a review of the literature. Section 3 is the theoretical analysis and hypothesis. Section 4 introduces the study design. Section 5 analyzes the empirical findings. Section 6 provides conclusions and policy implications.

2. Literature Review

2.1. Green Technology Innovation

Green technology innovation, as introduced by Brawn and Wield, is described as production and manufacturing methods that can lessen the use of resources such as energy and raw materials and environmental pollution [8]. The implementation of green technology innovation fosters the conversion of corporate green knowledge into environmental performance [9]. It helps reduce environmental deterioration and raise productivity [10]. The existing literature on green technology innovation is mostly combined with environmental regulation. Porter concluded that appropriate environmental regulations would help promote green technology innovation [11]. Based upon Porter’s argument, other academics have confirmed that green technology innovation can be encouraged by environmental regulation [12,13,14,15]. However, Lin et al. [16] concluded that environmental regulation hindered green technology innovation. They stated that regulations place higher compliance costs on businesses that pollute extensively. According to Behera et al., there’s a U-shaped changing trend between green technology innovation and environmental regulation [12].

There are also many scholars who study green technology innovation from different perspectives. Na et al. [17] found that corporate green technology innovation could be enhanced by green finance. Additionally, economic growth is crucial for fostering the development of green innovation [18]. In addition, government subsidies considerably increase manufacturing companies’ incentives to pursue green innovation [19]. Pedro et al. demonstrated that enterprises could promote green innovation activities by strengthening R&D investment intensity [20]. Zhao et al. investigated the link between board size and green innovation. The findings demonstrate that more board members can support the implementation of a green innovation strategy and expand the openness and breadth of innovation [21]. Additionally, the extended producer responsibility system has significantly promoted green technology innovation [22]. Furthermore, green technology innovation aids in lowering carbon dioxide emissions [23,24,25].

2.2. Digital Finance

Digital finance is emerging as the times demand against the backdrop of information technology and finance integration [26]. The existing literature about digital finance is very rich. On the macro level, by advancing and promoting the industrial structure, digital finance raises the quality of economic growth [27]. Additionally, digital finance positively moderates the influence of environmental regulation on high-quality economic development [28]. The growth of digital finance can encourage residents’ spending and consumption upgradation [29]. Additionally, by increasing online shopping and reducing income inequality, digital finance can also lower consumption inequality among farmers [30]. Digital finance can also support the development of inclusive finance [31]. On the micro level, by reducing the degree of information asymmetry, digital finance can encourage the investment behavior of micro and small enterprises (MSEs) [32]. Digital finance also has a positive impact on the corporate environment, society, and governance [33]. According to Xin et al., digital finance enhances the social responsibility performance of industries that produce a lot of pollution [34]. Abbasi et al. proposed that digital finance could enhance corporate performance and maintain competitiveness [35].

2.3. Green Technology Innovation and Digital Finance

Several academics concentrate on the connection between green technology innovation and digital finance. Some people concluded that green technology innovation could be promoted by digital finance [36,37,38]. The benefits of “cheap cost, quick speed, and extensive coverage” that are offered by digital finance can lower the entry barrier for financial services. Through well-informed decisions, it can lessen information asymmetry and increase the financial market’s efficiency [39]. Feng et al. discovered that digital finance could ease financing limitations, promote manufacturing expansion, and enhance the regional capacity for green innovation [38]. Han et al. [40] discovered that digital finance displays the “network effect” and “inclusive effect” when driving green innovation. Digitalization encourages gathering, storing, transmitting, and identifying information, improves a businesses’ capacity to analyze huge amounts of data, and raises risk-bearing by improving financing availability [41]. Digital finance has greatly decreased search and risk identification costs released a lot of new business space, and provided opportunities for corporate technology innovation [42].

Based on the sorting of the existing literature, there is some research regarding the influence of digital finance on green technology innovation. However, there are few articles to study on the different dimensions of digital financing. So, this paper examines the impact of digital finance and its three dimensions on green technology innovation, including the breadth of coverage, depth of use, and degree of digitization, which is a more thorough analysis. This paper also explores the intermediary role of financial constraints and examines the region and enterprise nature heterogeneity. This will offer theoretical support for encouraging the growth of green technology innovation.

3. Theoretical Analysis and Hypothesis

3.1. Digital Finance and Green Technology Innovation

Because green technology innovation has a high degree of complexity and novelty [43], enterprises must invest a large sum of money. Digital financing makes up for the flaws of traditional financing [44]. It uses the Internet to break the space limitation and includes the long-tailed groups into the service scope, which advances the growth of inclusive finance [45] and improves the accessibility and convenience of financial services [44]. Digital finance broadens financing avenues to meet the diversified financing requirements of businesses [26], which supports businesses in carrying out green technology innovation. Digital finance also gives financial institutions access to more credit data about lenders, which helps financial institutions reduce information asymmetry [46] and offers to finance more corporate innovation projects. All of this boosts the impetus of green technology innovation. Additionally, the financial market is creatively altered by digital finance, which helps improve the expected return on technology innovation [47] and raises entrepreneurs’ passion for innovation.

The following hypothesis is put forth based on the analysis mentioned above.

Hypothesis 1.

Digital finance can advance green technology innovation.

3.2. The Mediating Effect of Financial Constraints

The pecking order of financing theory put forward by Myers and Majluf shows that corporate financing follows the order of internal surplus, debt financing, and equity financing when there is a transaction cost [48]. Smaller enterprises need external financing since their internal earnings are insufficient. However, internal and external financing methods are not completely replaced [49] because of information asymmetry and agency costs in the real environment [50], which leads to the problem of financial constraints. Scholars also proposed that innovative enterprises confront stronger financial constraints [49]. Additionally, a company’s R&D intensity and innovative power are decreased by financial constraints [51,52]. Digital technology can assist with digital financing and alleviate the information asymmetry problem [46]. Based on the description, judgment, and prediction of massive data, big data can accurately portray users and judge customer reputation and risk-bearing capacity [53], which helps reduce transaction costs. Additionally, the internet of things also aids financial institutions in knowing the corporate circumstances better and enhancing their capacity for risk management. Digital financing can effectively alleviate financial constraints, then encourage corporate innovation in green technology.

The following hypothesis is put forth based on the analysis mentioned above.

Hypothesis 2.

Digital financing boosts corporate green technology innovation by alleviating financial constraints.

4. Study Design

4.1. Variable Selection

4.1.1. Explained Variable

Green technology innovation (GTI): corporate green patent applications are chosen to gauge green technology innovation based on prior studies [22,54].

4.1.2. Explanatory Variable

Digital finance (DF): The Peking University digital financial inclusion index is a popular choice for measuring digital finance in current research [37,38,40]. The index systematically depicts the development level of banking, payment, investment, insurance, and other industries based on comprehensively summarizing the connotation and characteristics of digital finance. The index includes three dimensions, including 33 specific indicators in total. It can better measure the digital finance level of different provinces. Therefore, this index is used to measure digital finance. This paper performs a logarithm of treatment on the data.

4.1.3. Control Variables

Reviewing the related literature [55,56], corporate size, leverage, growth, ROA, cash flow, the proportion of independent directors, and the proportion of the first shareholder are selected as control variables.

Table 1. Variables description.

Type	Name of Variable	Symbol	Definition
Explained variable	Green technology innovation	GTI	Number of green patent applications of listed companies
Explanatory variables	Digital finance index	DF	Peking University digital inclusive financial index
	breadth of coverage	DFc
	Depth of use	DFu
	Degree of digitization	DFd
Control variables	Corporate size	size	Natural logarithm of total assets
	Leverage	lev	Total debts/Total assets
	Growth	Growth	$\frac{\mathrm{operating} \mathrm{income} \mathrm{of} \mathrm{current} \mathrm{year}-\mathrm{operating} \mathrm{income} \mathrm{of} \mathrm{last} \mathrm{year}}{\mathrm{operating} \mathrm{income} \mathrm{of} \mathrm{last} \mathrm{year}}$
	Return on total assets	ROA	Net profit/Total assets
	cash flow	CF	Cash flow/Total assets
	Proportion of independent directors	Indd	The number of independent directors/the total number of directors
	Proportion of the first shareholder	Tops	Proportion of the first shareholder

is the variable interpretation.

4.1.4. Data Resource and Descriptive Statistics

The research sampled listed companies’ annual data from 2011 to 2020. The following processing was made: (1) exclude the special treatment companies, (2) exclude financial and real estate industries, (3) delete the sample with missing data of variables. The Winsorize method is used to avoid the impact of extreme values. After sorting out, 11,550 annual observation samples were finally obtained. The green patent application data were from the Chinese Research Data Services Platform. Corporate financial data came from CSMAR. The digital finance index was from the Peking University Digital Finance Research Center.

4.2. Model Setting

The following model is used in this essay to verify Hypothesis 1:

$$GF{I}_{i,t}={\alpha }_0+{\alpha }_{1}D{F}_{i,t}+\sum {\alpha }_{2}C{V}_{i,t}+\sum industry+\sum year+{\epsilon }_{i,t}$$

(1)

where t and i represent the year and enterprise, respectively. Green technology innovation is represented by GFI. Digital finance is marked as DF. The random error term is ${\epsilon }_{i,t}$. CV are the control variables. For empirical analysis, this study used a fixed effect model with controlling industry and year effects.

To study influential mechanisms, mediation effect models were established as follows to verify hypothesis 2 [57]:

$$GF{I}_{i,t}={\beta }_0+{\beta }_{1}D{F}_{i,t}+\sum {\beta }_{2}C{V}_{i,t}+\sum industry+\sum year+{\epsilon }_1$$

(2)

$$M_{i,t}={\gamma }_0+{\gamma }_{1}D{F}_{i,t}+\sum {\gamma }_{2}C{V}_{i,t}+\sum industry+\sum year+{\epsilon }_2$$

(3)

$$GF{I}_{i,t}={\lambda }_0+{\lambda }_1{M}_{i,t}+{\lambda }_{2}D{F}_{i,t}+\sum {\lambda }_{3}C{V}_{i,t}+\sum industry+\sum year+{\epsilon }_3$$

(4)

In Equations (3) and (4), M represents the mediator variables, that is, financial constraints.

5. Empirical Analysis

5.1. Descriptive Statistics

Table 2. Descriptive statistics.

Variables	obs	Mean	Median	Std.Dev.	Min	Max
GTI	11,550	1.380	1.099	1.320	0.000	7.319
DF	11,550	5.309	5.485	0.634	2.786	6.068
DF-coverage	11,550	5.195	5.418	0.730	0.673	5.984
DF-use	11,550	5.340	5.456	0.588	1.911	6.192
DF-digitization	11,550	5.477	5.777	0.771	2.026	6.136
Size	11,550	22.595	22.447	1.409	16.161	28.636
Lev	11,550	0.465	0.461	0.388	−0.195	28.548
Growth	11,550	0.176	0.104	0.468	−0.608	3.317
ROA	11,550	0.035	0.033	0.055	−0.225	0.185
CF	11,550	0.044	0.044	0.122	−10.216	0.600
Indd	11,550	0.374	0.333	0.058	0.200	0.800
Tops	11,550	34.543	32.030	15.494	2.870	89.990

presents the descriptive statistics. The standard deviation and mean of green technology innovation are 1.32 and 1.38, respectively, which implies that there are huge individual differences in green technology innovation. The average number of digital finances is 5.309, which means that the digital financial development level is high. The maximum and minimum values are 6.068 and 2.786, respectively, which means the unbalanced regional development of digital finance.

Table 3. Correlation coefficient.

	GTI	DF	DFc	DFu	DFd	Size	Lev	Growth	ROA	CF	Indd	Tops
GTI	1.000
DF	0.279 ***	1.000
DFc	0.269 ***	0.983 ***	1.000
DFu	0.282 ***	0.971 ***	0.942 ***	1.000
DFd	0.234 ***	0.875 ***	0.799 ***	0.816 ***	1.000
size	0.451 ***	0.246 ***	0.236 ***	0.227 ***	0.228 ***	1.000
lev	0.096 ***	0.014	0.010	0.008	0.030 ***	0.190 ***	1.000
growth	0.001	−0.008	−0.010	−0.010	−0.002	−0.036 ***	0.016 **	1.000
ROA	0.008	−0.027 ***	−0.023 **	−0.025 ***	−0.030 ***	0.038 ***	−0.672 ***	0.001	1.000
CF	0.020 **	0.058 ***	0.055 ***	0.054 ***	0.059 ***	0.130 ***	−0.062 ***	−0.010	0.083 ***	1.000
Indd	0.069 ***	0.040 ***	0.037 ***	0.036 ***	0.040 ***	0.083 ***	0.022 **	−0.010	−0.010	−0.022 **	1.000
Tops	0.019 *	−0.096 **	−0.086 ***	−0.102 ***	−0.101 ***	0.271 ***	0.040 ***	0.007	0.025 ***	0.079 ***	0.032 ***	1.000

Notes: *, **, and *** represent significance at levels of 10%, 5%, and 1%, respectively, and t-statistics are in parentheses.

exhibits the correlation coefficients between variables. In addition, all of the variables’ variance inflation factors are under 10, demonstrating that the multicollinearity problem does not exist.

5.2. Regression Analysis

From the results in columns 1 and 2 of

Table 4. Regression results of the impact of digital finance on green technology innovation.

Variables	GFI
	(1)	(2)	(3)	(4)	(5)
DF	0.152 ***	0.095 ***
	(3.69)	(2.64)
DFc			0.054 **
			(2.20)
DFu				0.159 ***
				(4.46)
DFd					0.014
					(0.36)
Size		0.488 ***	0.488 ***	0.488 ***	0.489 ***
		(57.16)	(57.17)	(57.20)	(57.26)
Lev		0.099 ***	0.098 ***	0.102 ***	0.095 ***
		(2.79)	(2.77)	(2.88)	(2.68)
Growth		−0.001	−0.001	−0.001	−0.001 *
		(−1.62)	(−1.61)	(−1.60)	(−1.65)
ROA		0.068	0.067	0.071	0.065
		(1.56)	(1.55)	(1.62)	(1.49)
CF		−0.086	−0.086	−0.091	−0.085
		(−1.05)	(−1.04)	(−1.10)	(−1.03)
Indd		−0.064	−0.066	−0.055	−0.072
		(−0.38)	(−0.39)	(−0.33)	(−0.42)
Tops		−0.002 **	−0.002 **	−0.002 **	−0.001 **
		(−2.24)	(−2.23)	(−2.27)	(−2.15)
Constant	0.932 **	−10.102 ***	−9.933 ***	−10.389 ***	−9.774 ***
	(2.29)	(−25.06)	(−25.56)	(−25.65)	(−24.43)
Industry	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes
N	11,550	11,550	11,550	11,550	11,550
adj.R2	0.217	0.408	0.408	0.409	0.408

Notes: *, **, and *** represent significance at the levels of 10%, 5%, and 1%, respectively, and t-statistics are in parentheses.

, the coefficients of DF were significantly positive before and after adding the control variables. This implies that green technology innovation could be promoted by digital finance. Hypothesis 1 is verified.

This paper further selects three dimensions of the digital financial index for empirical analysis, including the digital financial breadth of coverage (DFc), depth of use (DFu), and degree of digitization (DFd). The breadth of coverage mainly means account coverage, depth of use includes indicators such as the actual total usage and activity situation, and digital degree refers to the degree of mobility, facilitation, and credit. As shown in columns (3), (4), and (5), three variables are positively connected with green technology innovation. The coefficients of DFc and DFu are at a significant level. Based on Internet technology, digital finance can serve more enterprises across time and space constraints. In addition, digital financing can provide services such as payment, insurance, and credit. A deeper exploration based on diversified business formats will help to provide a sustainable impetus for enterprise innovation [42]. Therefore, digital finance should not only focus on coverage but also pay attention to the in-depth upgrading of different service formats. This helps digital finance to provide a strong and sustainable impetus for enterprises to conduct green technology innovation.

5.3. Mediation Effect Analysis

To verify hypothesis 2, this paper tested the mediation effect and took financial constraints as intermediary variables. Hadlock and Pierce [58] established a model to analyze the KZ index and WW index, concluded that enterprise size and age are good indicators of financial constraints, and proposed the SA index as a new metric to gauge financial constraints.

Table 5. Mediation effect analysis-financial constraints.

Variables	GTI	FC	GFI
	(1)	(2)	(3)
DF	0.095 ***	−0.045 ***	0.073 **
	(2.64)	(−5.19)	(2.05)
FC			−0.477 ***
			(−12.50)
Size	0.488 ***	−0.005 **	0.485 ***
	(57.16)	(−2.40)	(57.25)
Lev	0.099 ***	0.104 ***	0.148 ***
	(2.79)	(12.05)	(4.18)
Growth	−0.001	−0.00003	−0.001 *
	(−1.62)	(−0.22)	(−1.66)
ROA	0.068	0.103 ***	0.117 ***
	(1.56)	(9.74)	(2.70)
CF	−0.086	0.076 ***	−0.050
	(−1.05)	(3.82)	(−0.61)
Indd	−0.064	−0.306 ***	−0.210
	(−0.38)	(−7.42)	(−1.24)
Tops	−0.002 **	−0.001 ***	−0.002 ***
	(−2.24)	(−7.95)	(−3.18)
Constant	−10.102 ***	3.532 ***	−8.416 ***
	(−25.06)	(36.09)	(−19.92)
Industry	Yes	Yes	Yes
Year	Yes	Yes	Yes
N	11,550	11,550	11,550
adj.R2	0.408	0.289	0.416

Notes: *, **, and *** represent significance at the levels of 10%, 5%, and 1%, respectively, and t-statistics are in parentheses.

shows that the impact coefficient of digital financing on financial constraints is significantly negative. The effect coefficient of digital financing on green technology innovation decreases after adding intermediary variables, indicating that there is an intermediary effect. By lowering information asymmetry and agency costs [34], digital financing can foster innovation in green technology by eliminating financial constraints. Hypothesis 2 is confirmed.

5.4. Robustness Test

5.4.1. Tobit Model

Since the quantity of green patent applications of some businesses is 0, the Tobit model can be used to test the data analysis of this zero-value accumulation.

Table 6. Tobit model results.

Variables	GTI
	(1)	(2)	(3)	(4)
DF	0.108 **
	(2.13)
DFc		0.065 *
		(1.86)
Dfu			0.199 ***
			(3.93)
DFd				0.007
				(0.14)
Size	0.640 ***	0.640 ***	0.640 ***	0.641 ***
	(51.29)	(51.29)	(51.38)	(51.35)
Lev	0.052	0.051	0.060	0.045
	(0.84)	(0.83)	(0.98)	(0.72)
Growth	−0.001	−0.001	−0.001	−0.001
	(−1.33)	(−1.32)	(−1.32)	(−1.35)
ROA	0.168	0.168 **	0.165 **	0.171
	(1.44)	(1.44)	(1.41)	(1.47)
CF	−0.0003	−0.001	−0.008	−0.002
	(0.00)	(0.00)	(−0.05)	(−0.01)
Indd	−0.409 *	−0.410 *	−0.399 *	−0.415 *
	(−1.73)	(−1.73)	(−1.69)	(−1.75)
Tops	−0.002 *	−0.002 *	−0.002 *	−0.002 *
	(−1.83)	(−1.82)	(−1.86)	(−1.76)
Constant	−13.185 ***	−12.933 ***	−12.933 ***	−12.610 ***
	(−30.72)	(−34.71)	(−34.71)	(−27.84)
Industry	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
N	11,550	11,550	11,550	11,550
Pseudo R2	0.1516	0.1516	0.1516	0.1515

Notes: *, **, and *** represent significance at the levels of 10%, 5%, and 1%, respectively, and t-statistics are in parentheses.

shows the empirical results. The coefficients of DF, DFc, and DFu are positive and statistically significant. This shows that the results are still valid after replacing the model.

5.4.2. Replacing Explained Variable

This paper replaces the metrics of green technology innovation and uses the number of green patent authorizations.

Table 7. Empirical results after changing explained variables.

Variables	GTI	GTI	GTI	GTI	FC	GTI
	(1)	(2)	(3)	(4)	(5)	(6)
DF	0.071 **				−0.047 ***	0.046
	(2.07)				(−5.20)	(1.37)
DFc		0.048 **
		(2.09)
DFu			0.133 ***
			(3.93)
DFd				−0.073 **
				(−2.08)
FC						−0.515 ***
						(−14.08)
Size	0.413 ***	0.413 ***	0.413 ***	0.413 ***	−0.013 ***	0.406 ***
	(49.68)	(49.68)	(49.71)	(49.76)	(−5.87)	(49.26)
Lev	0.144 ***	0.144 ***	0.147 ***	0.143 ***	0.128 ***	0.210 ***
	(3.68)	(3.68)	(3.78)	(3.67)	(12.25)	(5.38)
Growth	−0.001	−0.001	−0.001	−0.001	−0.00002	−0.001
	(−1.60)	(−1.59)	(−1.61)	(−1.58)	(−0.17)	(−1.63)
ROA	0.088 *	0.085 *	0.088 **	0.084 *	0.129 ***	0.150 ***
	(1.92)	(1.92)	(2.00)	(1.92)	(10.81)	(3.42)
CF	−0.068	−0.068	−0.072	−0.069	0.073 ***	−0.031
	(−0.89)	(−0.89)	(−0.95)	(−0.90)	(3.56)	(−0.40)
Indd	−0.073	0.073	0.08	0.07	−0.312 ***	−0.088
	(0.45)	(0.45)	(0.49)	(0.43)	(−7.20)	(−0.55)
Tops	−0.001 **	−0.001	−0.001	−0.001	−0.001 ***	−0.001 *
	(−0.97)	(−0.97)	(−0.99)	(−0.95)	(−6.18)	(−1.83)
Constant	−9.010 ***	−8.870 ***	−9.354 ***	−8.192 ***	4.764 ***	−6.557 ***
	(−22.05)	(−23.17)	(−22.89)	(−19.43)	(43.70)	(−14.88)
Industry	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
N	10,395	10,395	10,395	10,395	10,395	10,395
R2	0.3752	0.3752	0.3759	0.3752	0.2659	0.3869

Notes: *, ** and *** represent significant at the significance levels of 10%, 5% and 1%, respectively, and t-statistics are in parentheses.

shows the empirical results. Columns 1–4 show that the result of green technology innovation could be encouraged by digital finance and is robust. Columns 5 and 6 prove that there is an intermediary effect.

5.4.3. Endogenetic Test

This paper used the methods of model replacement and variable replacement for robustness testing. Considering endogenous problems such as missing variables, this paper selected internet penetration as the instrumental variable when consulting the method of Xie et al. [59]. The data come from the China Internet Information Center. Regional internet penetration expansions provide the foundations for digital finance. The two are closely linked. In comparison, there is no direct link between green technology innovation and regional internet penetration. In theory, this instrumental variable is reasonable. The 2SLS method is used to carry out the endogenous test.

Table 8. Endogenetic test.

Variables	GTI	GTI
	Patent Application	Patent Authorizations
DF	0.503 ***	0.439 ***
	(5.80)	(5.65)
Size	0.479 ***	0.405 ***
	(46.16)	(42.85)
Lev	0.125 ***	0.109 ***
	(2.89)	(3.01)
Growth	−0.001 ***	−0.001 **
	(−3.10)	(−2.12)
ROA	−0.093 **	0.032
	(2.02)	(0.98)
CF	−0.344 ***	−0.331 ***
	(−3.53)	(−5.95)
Indd	−0.138	0.056
	(−0.71)	(0.31)
Tops	−0.002 **	−0.001
	(−2.06)	(−1.04)
Constant	−11.878 ***	−10.129 ***
	(−20.99)	(−19.88)
Industry	Yes	Yes
Year	Yes	Yes
N	10,395	10,395
First-stage F-statistics	707.49 ***	707.49 ***
p-value	0.000	0.000
Wald test	6761.91 ***	6002.56 ***
p-value	0.000	0.000

Notes: **, and *** represent significance at the levels of 5%, and 1%, respectively, and t-statistics are in parentheses.

shows that the digital financial coefficient is also significantly positive, which confirms the robustness of the results.

5.5. Heterogeneity Analysis

5.5.1. Regional Heterogeneity

There are 34 provincial administrative regions in China, and regional economic development levels vary greatly. To further analyze if there are regional disparities, China is classified into three parts, namely, West China, Central China, and East China. As shown in

Table 9. Empirical results of regional heterogeneity.

Variables	East China	Central China	West China
	(1)	(2)	(3)
DF	0.341 ***	0.191 ***	0.205 ***
	(16.76)	(6.55)	(6.32)
Size	0.514 ***	0.598 ***	0.506 ***
	(28.72)	(16.62)	(12.15)
Lev	−0.036	0.329 *	−0.029
	(−1.04)	(2.23)	(−0.16)
Growth	0.000	0.0003	0.003
	(−0.66)	(0.31)	(1.61)
ROA	−0.0764 *	−0.350	−0.674 *
	(−2.00)	(−1.25)	(−2.44)
CF	−0.126	0.334	0.157
	(−1.94)	(1.10)	(0.45)
Indd	−0.111	−0.193	0.131
	(−0.46)	(−0.42)	(0.29)
Tops	−0.006 ***	−0.002	−0.006 *
	(−4.39)	(−0.89)	(−2.30)
Constant	0.000	0.0002	0.000
	(−0.00)	(0.02)	(0.00)
N	7737	2263	1550
adj.R2	0.2930	0.2601	0.2336
	East versus West	East versus Central	West versus Central
Group difference test	11.37 ***	16.67 ***	0.1
p-value	0.001	0.000	0.753

Notes: *, and *** represent significance at the levels of 10%, and 1%, respectively, and t-statistics are in parentheses.

, three digital financial coefficients were positive and at a significant level. By contrast, in the eastern region, digital finance had a more obvious promotion effect on corporate green technology innovation. East China has the advantages of a dense population, a developed economy, and a high digital finance development level. This makes enterprises in the eastern region have a stronger capital gathering and technological innovation capabilities. The smooth flow of information makes them face lower financial constraints. Therefore, digital financing can more effectively encourage green technology innovation among eastern region businesses.

5.5.2. Enterprise Nature Heterogeneity

The economic activities of enterprises are impacted by their various natures. In order to determine if digital financial effects on green technology innovation differ between different corporate natures, this article separated all the samples of listed companies into two categories: non-state-owned and state-owned.

Table 10. Empirical results of enterprise-type heterogeneity.

Variables	GTI
	State Enterprise	Non-State Enterprises
DF	0.295 ***	0.219 ***
	(14.85)	(9.76)
Size	0.575 ***	0.552 ***
	(24.80)	(27.78)
Lev	−0.210 *	0.00815
	(−2.05)	(0.23)
Growth	0.00163	−0.000463
	(1.79)	(−1.18)
ROA	−0.308	−0.0576
	(−1.60)	(−1.49)
CF	0.223	−0.111
	−1.27	(−1.60)
Indd	−0.0212	−0.216
	(−0.08)	(−0.72)
Tops	−0.00914 ***	−0.00282
	(−6.14)	(−1.73)
_cons	−0.000512	−0.0000322
	(−0.06)	(−0.00)
N	5808	5742
adj.R2	0.2816	0.2792
Group difference test	5.68 **
p-value	0.0171

Notes: *, **, and *** represent significance at the levels of 10%, 5%, and 1%, respectively, and t-statistics are in parentheses.

shows that in state-owned companies, the positive impact of digital financing on green technology innovation was more visible. The group difference test also shows that this result is meaningful. Technology spillovers brought by green technology innovation make it challenging for businesses to realize maximum revenues. State-owned businesses have more benefits in capital strength and risk-taking than non-state-owned businesses and experience fewer financial constraints. They also have better financial resources [60]. Furthermore, strong external supervision may compel state-owned companies to increase their innovation in green technologies. Therefore, digital financing can better support state-owned firms’ innovation in green technology.

5.6. Results and Discussion

The empirical portion researched how digital financing affects green technology innovation. First, green technology innovation is positively impacted by digital financing; green technology innovation rises by 9.5% for every 1% growth in digital finance. This result is similar to Astadi Pangarso et al., Anu et al., Hung Bui Quang et al., and Rao et al., who further proved that digital finance could enable green technology innovation [61,62,63,64]. Additionally, digital technologies such as the internet of things and artificial intelligence can support the development of green technology [65,66,67]. The paper also examines the effect of three dimensions of digital financing on green technology innovation. The regression coefficients of coverage breadth and use depth are 0.054 and 0.159, while the coefficient of the digital degree is not significant; this provides some inspiration. It is necessary to advance the development of digital financing to assist green technology innovation. Digital finance contributes to the quick advancement of digital technology, which can provide more opportunities for corporate green innovation [68]. Nevertheless, we should be aware that the foundation of the digitization degree is not strong enough. Digitization has a positive effect on green economy development [69]. Hence, promoting the comprehensive development of digital finance is important.

Second, financial constraints act as an intermediary role; that is, digital finance advances green technology innovation by removing financial constraints. The coefficient of digital finance to green technology innovation reduces from 0.095 to 0.073 after adding the financial constraints variable. This result is also supported by Han et al. and Yang et al., who found that digital finance significantly relaxed financial constraints [39,70]. Insufficient funds make it difficult to better integrate financial development with technology innovation [71]. Digital finance offers solutions to this problem. It can decrease corporate financing costs and create diversified financing channels [60]. Enterprises will increase their investment with the decline in financial constraints [72]. Additionally, the mitigation of financial constraints is also conducive to corporate technology innovation [73].

Third, the paper analyzes heterogeneity from the perspective of the region and equity nature. Digital finance demonstrates a significantly positive driving influence on green technology innovation in the eastern, central, and western areas, with coefficients of 0.341, 0.191, and 0.205, respectively. This result is similar to that of Wang et al., who found that the positive effect of the digital economy on green technology innovation was more significant in the eastern region [74]. In the group of equity nature, digital finance has a stronger driving effect on green technology innovation of state-owned firms, and the coefficient is 0.295. This is different from Feng et al. [38], who concluded that digital financing could better promote the green technology innovation of non-state businesses. Therefore, to accomplish coordinated development, we need to consider the variations in the regional economic environment and individual distinctions across businesses.

6. Conclusions and Policy Implication

6.1. Conclusions

This paper investigates how digital finance affects green technology innovation and conducts robustness tests and heterogeneity analysis using data from public companies from 2011 to 2020. The results obtained showed that, first, digital finance and its breadth of coverage and use depth have positive impacts on green technology innovation. This result also passed robustness tests through variable and model replacements. Second, financial constraints are the functional channel for digital finance to encourage green technology innovation. Third, the driven effect of digital finance on green technology innovation is particularly visible in eastern regions and state-owned businesses.

From the conclusions above, the following suggestions are offered; first, in a society with increasingly developed information technology, it is essential to accelerate digital finance development and foster the balanced development of different dimensions. Meanwhile, we should tap into the diversified service capabilities of digital finance deeply, deepen various financial service functions, and perfect digital financial infrastructure construction. Second, to alleviate financial constraints, digital finance and traditional financial institutions need to work together to open up the industrial chain of digital financial services and assist in the financing of firm innovation activities. Third, we must promote a balanced development of digital financing in different regions. We should increase support for central and western enterprises, provide more innovation incentive policies for non-state-owned enterprises, ensure the smooth flow of information and capital, and enhance their capacity and drive for green technology innovation.

6.2. Policy Implication

The empirical findings of this paper have some policy implications and can be used as a reference when creating new policies.

First, the development of digital finance needs to be emphasized. The government can implement pertinent measures to enhance the macroeconomic environment and develop acceptable and adaptable digital finance regulatory policies that are helpful for digital financing to support innovation in green technology further.

Secondly, businesses should be encouraged to innovate in green technology. To provide new financing avenues for businesses, the government can create fiscal and tax incentives as well as other policies, such as tax and fee reductions, financial subsidies, and other actions. The government should boost its investment in the development of green technology.

Thirdly, there should be a focus on balanced development. To assist the green technology innovation projects, the government can implement appropriate policy incentives based on the different equity nature and provide talent introduction measures for the central and western regions. In promoting the construction of digital finance, policymakers should concentrate on boosting the development of digital degrees and adjust measures for local conditions to achieve the balanced development of different regions.

There are some limitations in this paper. First, a spatial spillover effect could result from the driving effect of digital finance on green technology innovation, which is not considered in the paper. Second, in addition to financing constraints, there may be other intermediary mechanisms. Future research can start from the perspective of spatial spillovers and mine other intermediary mechanisms.