In the Context of Digital Finance, Can Knowledge Enable Manufacturing Companies to Be More Courageous and Move towards Sustainable Innovation?

Abstract

The advent of the VUCA era and the development of digital finance (DF) present opportunities and challenges for manufacturing companies to seek sustainable innovation by increasing their organizational resilience (OR) to withstand crises. The production, flow, and acquisition of corporate knowledge are indispensable to the establishment of organizational resilience. In this paper, we analyze how to make manufacturing enterprises more courageous and innovative in the context of digital finance. We used a perspective of knowledge channel acquisitions to achieve this aim. Using a sample of 1965 manufacturing companies in China from 2013 to 2020, we analyzed whether greater enterprise knowledge (internal knowledge and external knowledge) can yield higher levels of innovation performance and whether organizational resilience plays a role in the context of digital finance. The results show that (1) both internal enterprise knowledge (IEK) and external enterprise knowledge (EEK) have a significant positive impact on the sustainable innovation performance of manufacturing enterprises; (2) organizational resilience has a mediation role in the process of promoting sustainable innovation performance through enterprise knowledge; (3) digital finance significantly enhances the impact of enterprise knowledge on long-term growth and financial volatility of organizational resilience, and significantly positively moderates the mediation effect of organizational resilience; and (4) digital finance support policies issued by the government significantly improve the sustainable innovation performance of manufacturing firms. Based on these results, manufacturing firms can improve innovation performance by enhancing organizational resilience. This paper contributes to this field of research by providing an analysis of manufacturing firms, presenting a new view on the improvement of innovation performance in the context of digital finance.

1. Introduction

Innovation is not only essential for a strong country and a fundamental driver of strategy, but also a key factor for companies to achieve sustainable business growth [1]. On the one hand, the essence of innovation lies in the use of knowledge [2], and sustainable innovation requires knowledge [3]. On the other hand, in a complex environment, organizational resilience facilitates innovation and promotes future success [4]. The impact of enterprise knowledge (EK) and organizational resilience (OR) on the innovation performance of enterprises (IPE) has attracted the attention of many scholars.

Research shows that human-based knowledge is the source of competitive advantage and the basis for economic value [5]. Knowledge acquisition positively affects business performance through product and process innovation [6]. At the same time, knowledge creation and acquisition also contribute to a firm’s innovation strategy to achieve market globalization [7]. Some scholars also take a different view. Laursen et al. pointed out that there is a substitution effect of knowledge search on firm innovation performance in large industrial firms [8], and evidence from emerging market companies also shows that internal R&D and external knowledge acquisition can be substituted for each other [9]. It has also been recognized that the relationship between knowledge acquisition and firm performance is not simply linear but U-shaped [10].

Unlike hope, optimism, and confidence [8], organizational resilience, as a three-dimensional constituent of emotions, cognition, and relationships, can help firms safely survive crises [11]. Most existing studies have explored the positive effects of knowledge on community resilience [12] and urban resilience [13], while some scholars have also explored the impact of knowledge on entrepreneurial resilience based on the perspective of failed learning [14]. Previous studies have explored the role of organizational resilience in firm performance [15]. Based on time-lagged data from 188 SMEs, organizational resilience significantly improves performance [16]. In particular, in the new risk landscape of COVID-19, some companies are losing ground and others are adapting to the crisis with timely reforms [17], showing the importance of organizational resilience [18].

Therefore, in the era of volatility, uncertainty, complexity, and ambiguity (VUCA) [19], it is of paramount importance for manufacturing companies to harness corporate knowledge in order to enhance their organizational resilience against crises and seek continuous growth and innovation.

Digital finance has changed the economic growth model, driven industrial change, and helped enterprises innovate. The further integration of digital and traditional finance driven by technology has driven access to knowledge, helping businesses to better understand their customers [20]. In addition, digital finance increases the resilience of enterprises by reducing the friction of financing [21]. Existing studies have also focused on the ability of digital finance to promote economic growth by alleviating financing constraints [22] and improving resource allocation efficiency [23], especially in the Chinese context.

To summarize the existing research, there are two gaps worth filling. Firstly, few scholars have explored the impact of corporate knowledge on the innovation performance of enterprises considering organizational resilience. However, with the normalization of firms facing crisis, especially in the post-epidemic era, the effect of organizational resilience on innovation performance is becoming more and more evident. Secondly, most analyses of digital finance have focused on the regional level, but micro-firm behavior is an important foundation for high-quality macroeconomic development, and existing studies have neglected to discuss it more deeply at the micro-firm level. In summary, few studies have included firm knowledge, organizational resilience, digital finance, and innovation performance in the same framework.

According to the Global Innovation Index 2021 report, China ranked 12th in the world in 2021, jumping 31 places in just nine years, and “Chinese innovation” has become China’s new calling card. Thanks to the internet revolution, China’s digital economy, especially digital finance, has experienced rapid development thanks to advanced technology. Therefore, Chinese companies are an appropriate sample for exploring the intricate relationship between corporate knowledge and innovation performance of manufacturing enterprises in the context of digital finance.

To enrich the existing research, we conducted theoretical derivation and empirical analysis through the following methods. Firstly, starting from the perspective of knowledge channel acquisition, we constructed a theoretical model of the impact of knowledge acquisition on the sustainable innovation performance of enterprises and the mediation path of organizational resilience, and examined the role played by digital finance in this process. Secondly, we adopted a differences-in-differences method to further supplement the judgment of the impact of digital finance on corporate sustainable innovation performance by comparing the difference in corporate innovation performance before and after the release of digital finance policies.

The conclusions of this study represent a valuable source of new knowledge in research on enterprise innovation. Bridging the gaps in previous studies, this research makes several contributions. First, this study is one of the few to explore the mechanisms of the role of corporate knowledge in innovation performance of manufacturing firms. This paper introduces the elaboration and extension of the role of corporate knowledge in the innovation performance of firms while combining innovation models with organizational resilience and linking innovation management to organizational science, an area of study that scholars have called for further research on. Secondly, existing research remains poorly informed about the relationship between corporate knowledge and innovation performance dynamics in the context of digital finance. At the same time, the findings of this study offer a theoretical basis for those seeking to increase their knowledge base, enhance organizational resilience, and thus improve innovation performance. It also provides practical guidance for cities to build digital finance.

The remainder of this article is organized as follows. Section 2 presents the literature review. Section 3 presents the materials and methods. Section 4 shows the empirical analysis results. Finally, Section 5 summarizes the conclusions and limitations and proposes future research directions.

2. Theoretical Background and Literature Review

2.1. Enterprises Knowledge and Innovation Performance

Sustainable innovation is considered a new combination of elements or a new relationship between already-combined elements by a process of recombination [24]. Innovation is influenced by various forces [25], and the lack of innovation power of enterprises is related to the low importance of sustainable innovation, insufficient internal investment in R&D, and the neglect of external knowledge acquisition by enterprises. However, there is a close relationship between innovation activities and knowledge [26], and knowledge is the basis of enterprise innovation, affecting the efficiency and effectiveness of innovation activities and constituting the core of enterprise growth.

Many studies have highlighted the importance of knowledge for innovation, but the impact of knowledge sources remains unexplained from the perspective of manufacturing firms. Therefore, the first question that needs to be addressed is, “What impact do different sources of knowledge have on the innovation performance of manufacturing firms?” According to the acquisition channels of knowledge inflow [27], enterprise knowledge is divided into internal enterprise knowledge and external enterprise knowledge [28].

However, the relationship between different channels of knowledge acquisition and the innovation performance of enterprise remains unclear. In fact, knowledge makes innovation more complex [29]. The positive impact of internal enterprise knowledge on organizational innovation performance has been acknowledged, but relying entirely on internal knowledge requires firms to accumulate it for a long time [30], and creating new knowledge or new knowledge combinations becomes complicated, so acquiring critical external knowledge may become an important means of firm development.

External enterprise knowledge has been proven to enhance internal innovation and improve aggressive marketing performance [31]. Likewise, external knowledge acquisition is difficult to achieve, especially core technical knowledge from outside the organization, facing many limitations [32]. Based on data from a survey of construction company managers, Duodu and Rowlinson point out that most external knowledge has no positive impact on exploitation as innovation [6]. Internal enterprise knowledge refers to the effect of 1 + 1 > 2, allowing enterprises to spread and diffuse valuable knowledge within the organization by means of sharing [33]. The internal knowledge accumulated by an enterprise is the basis of knowledge application and is the most unique and indispensable resource in the innovation process [34,35]. In order to meet external expectations, employees within an organization exchange effective knowledge with each other and not only enhance corporate learning efficiency and reduce organizational information redundancy with lower knowledge acquisition costs, but also set the prerequisites for innovation. It is worth noting that R&D personnel, as the most important carriers of knowledge, often prefer to acquire knowledge from within, implying that if they can acquire a large amount of knowledge from within the enterprise, they can give full play to their knowledge reserve and ability advantages [36], break solidified thinking, and continuously try new combinations among knowledge elements efficiently and precisely, which in turn helps to improve enterprise innovation performance. At the same time, the internal knowledge of enterprises can also determine the absorptive capacity of enterprises [37], that is, the stronger the ability of enterprises to identify and digest the value of external information and carry out commercial transformation when they have a higher level of internal knowledge [38]. Based on this, the following hypothesis is proposed.

Hypothesis 1a (H1a).

There is a positive correlation between internal enterprise knowledge and innovation performance.

External enterprise knowledge is the process of internalizing external information and technology into knowledge that can be mastered by the enterprise and synergizing it with existing internal knowledge to achieve the goal of improving enterprise performance. For external free knowledge, enterprises can absorb and internalize it through organizational learning, which can enrich their own knowledge base and provide new ideas for output solutions.

In the existing studies, scholars mostly divide external knowledge acquisition into market knowledge and technical knowledge acquisition [39]. On the one hand, market knowledge is the basic requirement of modern innovation, and enterprises can acquire external market knowledge through effective interaction and in-depth communication with consumers, suppliers, or other external partners, so as to correctly and efficiently respond to the complex demands of different customers and markets [40]. At the same time, the act of acquiring external market knowledge also releases positive signals to the outside world, promoting the investment behavior of external partners and providing more resources for innovation behavior, thus enhancing innovation performance. On the other hand, external science and technology can provide enterprises with inspirational technology sources and solid R&D reserves to promote internal innovation capabilities [41]. In particular, the proactive behavior of enterprises’ industry–academia–research interaction and cooperation can correct, improve, and acquire more scientific knowledge or effective information from the outside world, which provides knowledge that is difficult to generate internally as well as new creativity and ideas that can be more easily produced, increasing the possibility of knowledge innovation. In summary, acquiring knowledge from both internal and external sources can help enterprises break through the limitation of innovation resources to a certain extent and thus gain innovation advantages. Based on the above analysis, the following hypothesis is proposed:

Hypothesis 1b (H1b).

There is a positive correlation between external enterprise knowledge and innovation performance.

2.2. Enterprise Knowledge and Organizational Resilience

Resilience was originally a concept discussed in several disciplines, including ecology, physics, and psychology, and refers to the ability of a system to recover or even surpass its original state after suffering a trauma [42]. Subsequently, the combination of resilience with the organizational domain was considered as the ability of companies with organizational resilience to seize opportunities and exploit challenges, thus having a long-term relative advantage [43]. Based on the perspective of capability, “organizational resilience” is not only the ability of an organization to not collapse in the face of disruptive “setbacks” [44], but also the ability to learn from turbulence, thus enabling the organization to continuously change and become more courageous and resilient [45]. Resilience is the soft power and new capital of an organization, and different levels of resilience will have different corporate culture, organizational behavior, and corporate performance. When there are more and more external challenges, such as political unrest, market contraction, technological iterations, changing consumer preferences, and even natural disasters [46], there is a greater risk of business failure and less room for growth. Therefore, there is an urgent need for firms to construct organizational resilience to cope with environmental changes in order to survive and thrive in the long run.

As an unobservable holistic construct, organizational resilience reduces a firm’s sense of vulnerability and results in an increase in financial capability to achieve the ideal state of long-term viability and consistent high performance [47]. Consistent with the prevailing view, in this study, we consider organizational resilience as the higher level of long-term growth and lower level of financial volatility that can be achieved. Long-term growth is described as the ability of a firm to gain a unique advantage over competition, enhance sustainability, increase long-term earnings, and maximize corporate value from its inception to its growth and development. Financial volatility, on the other hand, is defined as the degree of fluctuations in the indicators of the financial assets of the enterprise over a certain period of time due to external or own factors, often reflecting the level of risk of the enterprise.

Organizational resilience is not an indicator of immediate utility, and its benefits are difficult to demonstrate in the short term, meaning that companies may sacrifice some temporary gains in exchange for achieving their vision. Enterprise growth requires a rational allocation of existing resources, including the continuous and coordinated management of internal knowledge. The development of internal knowledge enables an enterprise to enhance the cohesiveness of knowledge, keenly perceive changes in the external environment, and continuously reinvent itself, which has a positive effect on the growth of the enterprise [48]. Thus, we propose the following hypothesis.

Hypothesis 2a (H2a).

There is a positive correlation between internal enterprise knowledge and long-term growth.

By capturing and sharing corporate information, companies can acquire vital external knowledge, such as product improvements or perceptions of potential consumers, match the conditions required for organizational growth, and help companies identify opportunities for growth and achieve gradual and continuous development. We thus establish the following hypothesis.

Hypothesis 2b (H2b).

There is a positive correlation between external enterprise knowledge and long-term growth.

The internal knowledge already possessed helps firms to improve their ability to anticipate risks, perceive the environment, and quickly alert and correct deviant tendencies to avoid crisis effects on the firm. This study suggests that:

Hypothesis 2c (H2c).

There is a negative correlation between internal enterprise knowledge and financial volatility.

Enterprises with rich external knowledge are larger and more complex than those with weak knowledge, and can show stronger stability and have more redundancy in cushioning against shocks and avoiding the risk of damaging their financial performance.

The necessary foundation condition for value creation is that firms have knowledge, but firms need to have potential knowledge in order to cope with shocks that are free from constant pressure and discontinuity, which means that the more knowledge a firm has, the better it is [49]. Therefore, we suggest that:

Hypothesis 2d (H2d).

There is a negative correlation between external enterprise knowledge and financial volatility.

2.3. The Mediating Role of Organizational Resilience

Innovation is an essential part of business, and putting it into action is a complex “spiral” process of trial and error and incremental improvement, with no precedent for success [50]. Innovation also implies risks, and enterprises are faced with the contradiction of investing a lot of resources and gaining nothing [51].

Organizational resilience, as an important corporate capability, increases a firm’s ability to refine and use information, improves the efficiency of practice and the multiple possibilities of development by creating new ideas, stimulates the mutual collision of ideas [52], and enhances a firm’s ability to innovate. At the same time, OR enables companies to recognize the possibility of failure in innovation, to be ever-vigilant to failure, to reduce possible mistakes in the innovation process or to correct deviations in time [53], and also to take the initiative to learn from mistakes and reduce the consequences of mistakes in the innovation process. In addition, OR can build a harmonious and mutually supportive internal relationship network of employees, reducing the risk of employees breaking the “psychological contract” [54], and a stable social relationship will encourage employees to take initiative and actively engage in creative behavior, thereby achieving organizational goals.

In line with the above derivation, organizational resilience is an important bridge in the impact of a firm’s knowledge on innovation performance.

The reliance on internal knowledge based on cross-sectorial communication facilitates in-depth and comprehensive analysis of market trends, continuous upgrading of technology and correction of shortcomings [55], effective prevention of serious deviations in innovation, and reduced innovation risks due to technological weakness or lack of knowledge, and enables the advancement of the company on a more efficient path to enhance innovation capabilities. Therefore, we propose the following hypothesis.

Hypothesis 3a (H3a).

Long-term growth plays a mediating role in the effect of internal enterprise knowledge and innovation performance.

For acquiring external knowledge, extensive and specialized external knowledge can make the enterprise knowledge diverse [56], and allow an enterprise to accumulate unique know-how, expand the enterprise’s information scope while the ability to process information is enhanced, and identify and seize the opportunities of technology turnover and market environment changes to promote innovation. Therefore, we propose the following hypothesis.

Hypothesis 3b (H3b).

Long-term growth plays a mediating role in the effect of external enterprise knowledge and innovation performance.

Enterprises accumulate and mutate new knowledge, fuse it at the component or system level, and escape from vulnerability to form an organizational memory that is agile, adaptable, and creative in its thinking. This reduces the time it takes for firms to respond to unexpected crises and delivers more expansive solutions, which in turn leads to innovation. The following hypothesis is thus proposed.

Hypothesis 3c (H3c).

Financial volatility plays a mediating role in the effect of internal enterprise knowledge and innovation performance.

By searching and absorbing external knowledge, enterprises can also promote inter-enterprise adhesion, reduce the financial sway caused by “going it alone”, and promote their initiative to embed themselves in diversified cooperative networks, thus breaking away from conventions and experience to achieve new breakthroughs. Thus, we propose the following hypothesis.

Hypothesis 3d (H3d).

Financial volatility plays a mediating role in the effect of external enterprise knowledge and innovation performance.

2.4. Moderating Effect of Digital Finance

The combination of financial services and digital technology has become a new trend [57]. Digital finance is an organic fusion of digital information technology and financial services, referring to the act of promoting finance through financial services with the support of technologies such as big data, cloud computing, artificial intelligence, etc. It also has the core characteristics of finance to achieve the optimal allocation of resources with low risk, given new attributes by digital technology. Compared with traditional finance, digital finance not only has a positive impact on individuals, but also offers new opportunities for enterprises and facilitates the global availability of products and services [58]. In addition, the unique inclusiveness of digital finance allows groups in underdeveloped areas, traditional industries, micro and small enterprises, and other organizations that are disadvantaged in traditional finance to have convenient and efficient financial services, thereby reducing the incidence of inequity [59]. The higher degree of digital finance breaks the spatial and temporal barriers to the dissemination of internal and external information and knowledge of enterprises, and effectively improves the efficiency of knowledge flow, thus laying a solid foundation for enterprises to build organizational resilience. With the support of new advanced technology, enterprises can quickly acquire and process massive external knowledge at a lower cost and increase the stock of enterprise knowledge. At the same time, digital finance also reduces the financing constraints of enterprises, which means that even when enterprises face difficulties, they can mobilize more resources at high efficiency and low cost. Resources are the basis on which companies pursue, achieve, and sustain competitive advantage [60]. More easily accessible resources mean smoother growth for the business. On the contrary, with a low degree of digital finance, the old sectors that do not have competitive advantages and development prospects still occupy the market and consume a large amount of internal knowledge, while the accumulation of production factors and the process of knowledge evolution in the production sectors of enterprises are slow. This is not conducive to long-term business development. Based on the above derivation, we propose the following hypotheses.

Hypothesis 4a (H4a).

Digital finance strengthens the positive relationship between internal enterprise knowledge and long-term growth.

Hypothesis 4b (H4b).

Digital finance strengthens the positive relationship between external enterprise knowledge and long-term growth.

Technology is moving human resources into the “Human Capital 4.0” era, where people are expected to improve technologically, socially, and psychologically, and employees are more resilient to future interactions [61]. With a high degree of digital finance, the embedding of digital technology also promotes the change in enterprise management decision paths and organizational structure, which have stronger adaptability to the environment. The four-dimensional business environment of humanities, laws, government affairs, and markets faced by enterprises in their development and even transformation will be more stable, and the probability of outside interference will be lower and the “trauma area” brought to enterprises will be smaller. At the relatively low level of digital finance, enterprises are not sensitive enough to forward-looking technology, and it is difficult to judge the value of exogenous information and data, which makes the inflow of external knowledge difficult. The solidification of knowledge elements makes enterprises become more like “islands” and it is difficult for them to resist external shocks, and because enterprises have limited knowledge at their disposal, managers focus more on the current development of enterprises with short-sightedness. At the same time, the lack of maturity of financial markets not only makes them vulnerable to capital control, but also the impact of external shocks on financial markets is amplified, and firms have unstable financial performance, leading to frequent and huge shocks. Based on the above derivation, we propose the following hypotheses.

Hypothesis 4c (H4c).

Digital finance strengthens the negative relationship between internal enterprise knowledge and financial volatility.

Hypothesis 4d (H4d).

Digital finance strengthens the negative relationship between external enterprise knowledge and financial volatility.

2.5. Mediating Role with Moderation

Based on the previous deduction and related hypotheses that organizational toughness mediates the relationship between corporate knowledge and innovation performance, and that digital finance regulates the relationship between corporate knowledge and organizational toughness, we further deduced that digital finance can regulate the mediating effect on organizational toughness between corporate knowledge and innovation performance, i.e., the first half path of the mediating effect is regulated by digital finance. Based on the above discussion, the following hypotheses are proposed.

Hypothesis 5a (H5a).

Digital finance can positively moderate the mediating effect of long-term growth in internal enterprise knowledge and innovation performance. That is, the mediating effect is more significant at higher levels of digital finance.

Hypothesis 5b (H5b).

Digital finance can positively moderate the mediating effect of financial volatility in internal enterprise knowledge and innovation performance. That is, the mediating effect is more significant at higher levels of digital finance.

Hypothesis 5c (H5c).

Digital finance is able to positively moderate the mediating effect of long-term growth in external enterprise knowledge and innovation performance. That is, the mediating effect is more significant at higher levels of digital finance.

Hypothesis 5d (H5d).

Digital finance is able to positively moderate the mediating effect of financial volatility in the external enterprise knowledge and innovation performance of the firm. That is, the mediating effect is more significant at higher levels of digital finance.

Based on the above analysis, the following theoretical model is proposed, shown in Figure 1.

3. Materials and Methods

3.1. Sample Selection and Processing

According to the industry classification standard of the China Securities Regulatory Commission, the manufacturing companies listed in Shanghai and Shenzhen A-shares from 2013 to 2020 were selected as the research sample. The listed companies are large in scale, and their financial data are strictly audited and publicly available. Moreover, compared with 12 other industries, the sample size of manufacturing enterprises is large, and the availability and authenticity of data are guaranteed. The CSMAR database, Wind Financial Database, and CNRDS China Research Data Service Platform provide support for the original data of this study, and the data of Digital Inclusive Finance Index were obtained from the Peking University Digital Inclusive Finance Index (2011–2020). To ensure the rigor of the conclusions, we cross-referenced the data and processed the raw data according to the following principles: (1) eliminating companies that have been ST and *ST during the sample period; (2) eliminating companies with less than 2 years of observations in the sample data; (3) eliminating companies with important data missing or obviously wrong data. The final panel data of 1965 firms were obtained. Meanwhile, all continuous variables were shrunken at the first 1% quintile.

3.2. Variable Design and Metrics

By constructing a mediated model with moderation, we investigated the impact of corporate knowledge on innovation performance, and how the level of digital economy moderates the degree of impact and how organizational resilience influences the mechanism of action. The specific variables are defined and measured as follows.

3.2.1. Dependent Variables

Innovation performance of enterprises (IPE). The main measures of innovation performance of enterprises in the existing literature are questionnaires, corporate R&D investment, and patents. Considering that innovation is the direct cause of changes in the profitability of firms, the purpose of innovation is to create real benefits [62]. This means that whether a firm invests in R&D activities or acquires new products or processes, the ultimate goal is still a higher financial performance. Considering the availability of data, we referred to the existing practice and chose to measure the innovation performance of manufacturing firms using the indicator of total corporate profit.

3.2.2. Independent Variables

Enterprise knowledge (EK)—internal enterprise knowledge (IEK) and external enterprise knowledge (EEK). The completely static internal knowledge cannot be effectively interacted internally and is difficult to effectively transform, while the sunken external knowledge will fall into the cycle of “backward-introduction” and imitation. For the measurement of knowledge, we focused more on the precipitation of knowledge, i.e., technological innovation. The internal knowledge measured in this study is the internal knowledge acquired by the enterprise through internal sharing and internal R&D; the external knowledge is the knowledge acquired by the enterprise through collaborative R&D between the enterprise and the organization. Drawing on and combining the metrics of Boeing et al. [63] and other scholars, the number of patents independently filed by enterprises is used to measure internal enterprise knowledge, and the number of patents filed jointly by enterprises and external organizations is used to measure external knowledge. Considering the time lag effect of patent output, the explanatory variables are taken with a lag.

3.2.3. Mediating Variables

Organizational resilience (OR)—long-term growth (Growth) and financial volatility (Volatility). Organizational resilience is divided into long-term growth and financial volatility, and higher long-term growth and lower financial volatility represent stronger organizational resilience [64]. Cumulative growth is a better measure of growth than year-over-year growth, and in this study, cumulative three-year net sales growth is considered long-term growth [65]. Financial volatility refers to the volatility of stock returns [66], and for each year, the standard deviation of weekly stock returns was calculated as a measure of financial volatility.

3.2.4. Moderating Variables

The measurement of the development status of digital finance (DF) uses the Digital Finance Index compiled by the Digital Finance R&D Center of Peking University and the Ant Financial Services Group to measure the development status of digital finance [67]. The index uses the big data of transaction accounts from Ant Financial Services and measures three dimensions, namely, coverage breadth index, usage depth index, and digitization degree index, and reliability is ensured. The larger the index, the better the development of digital finance in that province or city, and vice versa.

3.2.5. Control Variables

In this paper, with reference to relevant studies on corporate innovation performance, variables related to corporate innovation are selected as controls. Age at market, number of R&D employees [68], and R&D investment [69] have been shown to be key factors influencing firms’ innovation performance. Combining existing research, the number of years on the market (Cage), the number of corporate R&D employees (Employee), and corporate R&D investment (Research) are selected as control variables. In order to avoid the influence of individual characteristics and time trends on the research results, we also introduced the firm dummy variable (Code) and the year dummy variable (Year) for control.

The variables are specified as shown in

Table 1. Variable description and measurement.

Variable Type		Name	Abbreviation	Measurement
Dependent		Innovation performance of enterprises	IPE	Total enterprise net profit (billion yuan)
Independent	EK	Internal enterprise knowledge	IEK	Number of patents filed independently in year t − 1
		External enterprise knowledge	EEK	Number of patents filed jointly in year t − 1
Mediating	OR	Long-term growth	Growth	Cumulative sales growth over three years
		Financial volatility	Volatility	Weekly average stock price return volatility per year
Moderating		Digital finance	DF	Provincial Digital Inclusive Finance Index
Control		Corporate maturity	Cage	Year t minus the listed year plus 1 is taken as the natural logarithm
		Number of R&D employees	Employee	The number of R&D employees is added by 1 and taken as the natural logarithm
		R&D investment	Research	R&D investment as a percentage of operating revenue
		Enterprise	Code	Enterprise dummy variables
		Year	Year	Year dummy variable

.

3.3. Research Model

In order to verify whether corporate knowledge contributes to corporate innovation performance, the following benchmark model is constructed in this study:

$$IP{E}_{it}={\alpha }_1+{\beta }_{1}E{K}_{it-1}+{\beta }_{2}Control+{\epsilon }_{it}$$

(1)

where IPE denotes innovation performance; EK denotes enterprise knowledge, which is treated with a one-period lag; “Controls” denotes a set of control variables; i denotes a firm; t denotes a year; and ε_it denotes a random disturbance term.

To explore the mediating effect of organizational resilience between enterprise knowledge and innovation performance, we used a three-step test to develop estimation models (2) and (3).

$$O{R}_{it}={\alpha }_2+{\beta }_{2}E{K}_{it-1}+{\beta }_{4}Control+{\epsilon }_{it}$$

(2)

$$IP{E}_{it}={\alpha }_3+{\beta }_{5}E{K}_{it-1}+{\beta }_{6}O{R}_{it}+{\beta }_{7}Control+{\epsilon }_{it}$$

(3)

For the intermediation model with moderation to test the existing studies, models (4), (5), and (6) are constructed. Models (5) and (6) together determine whether digital finance moderates the first half path of intermediation.

$$IP{E}_{it}={\alpha }_4+{\beta }_{8}E{K}_{it-1}+{\beta }_{9}D{F}_{it}+{\beta }_{10}E{K}_{it-1}\ast D{F}_{it}+{\beta }_{11}Control+{\epsilon }_{it}$$

(4)

$$O{R}_{it}={\alpha }_5+{\beta }_{12}E{K}_{it-1}+{\beta }_{13}D{F}_{it}+{\beta }_{14}E{K}_{it-1}\ast D{F}_{it}+{\beta }_{15}Control+{\epsilon }_{it}$$

(5)

$$IP{E}_{it}={\alpha }_6+{\beta }_{16}E{K}_{it-1}+{\beta }_{17}D{F}_{it}+{\beta }_{18}O{R}_{\mathit{it}}+{\beta }_{19}O{R}_{it}\ast D{F}_{it}+{\beta }_{20}Control+{\epsilon }_{it}$$

(6)

where EK_it−1 ∗ DF_it denotes the interaction term between corporate knowledge and digital finance and OR_it ∗ DF_it denotes the interaction term between digital finance and organizational resilience.

4. Results

4.1. Descriptive and Correlation Analysis

The results of descriptive statistics and correlation analysis of each variable are shown in

Table 2. Descriptive statistics and correlation analysis results.

	Average	SD	1	2	3	4	5	6	7	8	9
1. IPE	2.95	7.48	1
2. IEKt−1	10.15	29.97	0.38 ***	1
3. EEKt−1	1.79	6.93	0.28 ***	0.27 ***	1
4. Growth	6.22	22.12	0.52 ***	0.25 ***	0.20 ***	1
5. Volatility	41.04	32.13	−0.03 ***	−0.01	−0.01	−0.01	1
6. DF	273.24	78.18	0.09 ***	0.07 ***	0.08 ***	0.17 ***	0.17 ***	1
7. Cage	1.54	1.08	0.22 ***	0.21 ***	0.20 ***	0.18 ***	0.18 ***	0.22 ***	1
8. Employee	3.71	2.78	0.22 ***	0.25 ***	0.18 ***	0.26 ***	0.39 ***	0.60 ***	0.47 ***	1
9. Research	3.85	3.74	−0.03 ***	0.14 ***	0.067 ***	−0.03 ***	0.36 ***	0.25 ***	0.21 ***	0.39 ***	1

Note: *** indicates significance at the 1% levels.

. IEK was significantly positively correlated with IPE (β = 0.38, p < 0.01), EEK was significantly positively correlated with IPE (β = 0.28, p < 0.01), and the stronger correlation between the independent and dependent variables provided initial support for the subsequent regression. The correlation coefficients between the control variables were all less than 0.70, which was the basis for the absence of multicollinearity among the control variables. Further aided by the variance inflation factor (VIF) test, the VIF values were all less than the critical value of 10, indicating that there was no multiple cointegration among the variables, and the selection of the model and sample was more reasonable.

4.2. Hypothesis Testing

4.2.1. The Role of Enterprise Knowledge on Innovation Performance

Table 3. The influence of enterprise knowledge on innovation performance.

Variables	Dependent Variable: IPE
	Model 1-1	Model 1-2	Model 1-3	Model 1-4
IEKt−1		0.02 *** (8.67)		0.02 *** (8.52)
EEKt−1			0.03 *** (3.23)	0.03 *** (2.80)
Cage	−1.45 *** (−11.10)	−1.49 *** (−10.17)	−1.37 *** (−9.40)	−1.49 *** (−10.19)
Employee	0.33 *** (12.43)	0.30 *** (10.12)	0.32 *** (10.64)	0.30 *** (10.09)
Research	−0.17 *** (−11.31)	−0.20 *** (−11.55)	−0.19 *** (−11.03)	−0.20 *** (−11.55)
Constant	3.83 *** (23.36)	4.09 *** (21.47)	4.05 *** (21.16)	4.01 *** (21.32)
Code	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
R-squared	0.075	0.073	0.068	0.073
F-test	111.35	92.55	85.61	84.89

Note: *** indicates significance at the 1% levels. Values in parentheses indicate t-values for significance tests, as in the table above.

demonstrates the regression results of EK on IPE. In order to reduce the errors caused by individual differences among different firms, balanced panel data for the period 2013–2020 are established. To improve the validity of the data regression results, a fixed-effects model is selected based on the Hausman test results prompt. The main effects are tested by stepwise regression, and Model 1-1 is the benchmark model containing enterprise maturity, the number of enterprise R&D employees, and enterprise R&D investment. Based on the benchmark model, Model 1-2 incorporates the explanatory variable IK, and the results show that IEK has a significant positive effect on IPE (β = 0.02, p < 0.01), and the hypothesis H1a is confirmed. Model 1-3 incorporates the explanatory variable EEK, and the results show that EEK has a significant positive effect on IPE (β = 0.03, p < 0.01), and the hypothesis H1b is confirmed. Model 1-4 incorporates both explanatory variables IEK and EEK on the basis of Model 1-1, and when IEK is included, EEK still has a significant positive effect on IPE (β = 0.03, p < 0.01); meanwhile, IEK still has a significant positive effect on IPE (β = 0.02, p < 0.01).

4.2.2. The Mediating Role of Organizational Resilience

In this study, the mediating role of organizational resilience is tested according to a three-step test, and the results are shown in

Table 4. Regression analysis of the mediation effect of organizational resilience on the relationship between “enterprise knowledge and innovation performance”.

Variable	Dependent Variable: Growth		Dependent Variable: Volatility		Dependent Variable: IPE
	Model 2-1	Model 2-2	Model 2-3	Model 2-4	Model 3-1	Model 3-2	Model 3-3	Model 3-4
IEKt−1	0.11 *** (8.38)		−0.12 *** (−6.62)		0.02 *** (5.96)	0.02 *** (8.50)
EEKt−1		0.18 *** (4.23)		−0.27 *** (−4.28)			0.01 * (1.72)	0.03 *** (3.11)
Growth					0.09 *** (48.26)		0.09 *** (48.66)
Volatility						−0.00 *** (−2.65)		−0.00 *** (−3.03)
Cage	−8.53 *** (−12.20)	−8.01 *** (−11.48)	−34.47 *** (−34.71)	−35.04 *** (−35.41)	−0.76 *** (−5.68)	−1.61 *** (−10.50)	−0.68 *** (−5.11)	−1.51 *** (−9.88)
Employee	1.36 *** (9.52)	1.44 (10.03)	7.91 *** (38.88)	7.83 *** (38.54)	0.19 *** (6.82)	0.33 *** (10.42)	0.20 *** (7.16)	0.35 *** (11.05)
Research	−1.01 *** (−12.90)	−1.03 (−12.41)	4.54 *** (38.68)	4.50 *** (38.33)	−0.11 *** (−6.86)	−0.18 *** (10.00)	−0.10 (−6.48)	−0.17 (−9.40)
Constant	13.52 *** (14.81)	13.28 (14.50)	55.58 *** (42.85)	55.91 *** (43.01)	2.94 *** (16.73)	4.29 *** (20.94)	2.91 *** (16.52)	4.28 *** (20.79)
Code	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R-squared	0.11	0.11	0.36	0.36	0.23	0.07	0.22	0.07
F-test	149.36	143.49	657.22	653.27	312.52	84.82	308.71	78.71

Note: *** and * indicate significance at the 1% and 10% levels, respectively.

. Based on the significant positive effects of IEK and EEK on IPE, Model 2-1 and Model 2-2 indicate that the regression coefficients of IEK and EEK on Growth are 0.11 (p < 0.01) and 0.18 (p < 0.01), respectively, and hypotheses H2a and H2b are confirmed, i.e., IEK and EEK contribute significantly and positively to Growth. Model 2-3 and Model 2-4 indicate that the regression coefficients of IEK and EEK on Volatility are −0.12 (p < 0.01) and −0.27 (p < 0.01), respectively, and hypotheses H2c and H2d are confirmed, i.e., IEK and EEK have a significant suppression on Volatility.

On the basis of the significant effect of EK on OR, EK and OR were further put into the model simultaneously to examine the effect of EK on IPE through OR. As in Model 3-1 and Model 3-3, the regression coefficients of IEK and EEK are 0.02 (p < 0.01) and 0.01 (p < 0.1), respectively, which are still significant. The effects of IEK and EEK on IPE work through the path of increasing Growth, and Growth plays a partially mediating role, verifying hypotheses H3a and H3b. In Model 3-2 and Model 3-4, the regression coefficients of IEK and EEK are 0.02 (p < 0.01) and 0.03 (p < 0.01), respectively, which are still significant. The effects of IEK and EEK on IPE work through the path of decreasing Volatility, and Volatility plays a partially mediating role, testing hypotheses H3c and H3d. In summary, OR plays a partial mediating role in the effects of IEK and EEK on IPE.

4.2.3. The Mediating Role of Organizational Resilience

The interaction terms of the independent and moderating variables may cause covariance, so the interaction terms are constructed after centering the independent and moderating variables separately. The results of the test for the moderating effect of digital finance are shown in

Table 5. Regression analysis of the moderation effect of digital finance on the relationship between “enterprise knowledge and organizational resilience”.

Variable	Dependent Variable: IPE		Dependent Variable: Growth		Dependent Variable: Volatility
	Model 4-1	Model 4-2	Model 5-1	Model 5-2	Model 5-3	Model 5-4
IEKt−1	0.01 *** (4.85)		−0.01 (−0.41)		−0.05 ** (−2.57)
EEKt−1		0.00 (0.21)		−0.15 *** (−3.10)		0.01 (0.16)
DF	−0.01 (−1.28)	−0.01 (−1.39)	−0.05 *** (−2.68)	−0.06 *** (−3.09)	0.11 *** (3.99)	0.11 (4.27)
IEKt−1 ∗ DF	0.00 *** (7.18)		0.00 *** (19.13)		−0.00 *** (−8.20)
EEKt−1 ∗ DF		0.00 *** (5.35)		0.01 *** (13.77)		−0.01 *** (−7.89)
Cage	−1.32 *** (−8.91)	−1.26 *** (−8.53)	−6.40 *** (−9.16)	−6.62 *** (−9.46)	−35.99 *** (−35.85)	−36.36 *** (−36.43)
Employee	0.29 *** (9.83)	0.31 *** (10.37)	1.25 *** (8.82)	1.34 *** (9.38)	7.97 *** (39.26)	7.90 *** (38.94)
Research	−0.19 *** (−11.06)	−0.19 *** (−10.70)	−0.96 *** (−11.76)	−0.96 *** (−11.67)	4.47 *** (38.09)	4.44 *** (37.84)
Constant	4.88 *** (6.24)	4.99 *** (6.36)	20.79 *** (5.63)	23.01 *** (6.17)	36.48 *** (6.87)	35.04 *** (6.59)
Code	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
R-squared	0.077	0.070	0.140	0.123	0.363	0.361
F-test	81.86	73.98	159.26	137.68	558.18	554.14

Note: *** and ** indicate significance at the 1% and 5% levels, respectively.

.

According to the regression results in Model 4-1 and Model 4-2, the direct effects of IEK and EEK on IPE are moderated by DF (β = 0.0001, p < 0.01; β = 0.0005, p < 0.01) when the mediating effects are not considered. This is the basis for the demonstration of the modulation mechanism of DF in EK–OR–IPE.

Regarding the inclusion of IEK_t−1 ∗ DF in Model 5-1, the interaction term between IEK and DF has a significant positive effect on Growth (β = 0.002, p < 0.01), and hypothesis H4a is confirmed. Model 5-2 incorporates EEK_t−1 ∗ DF, and the interaction term between EEK and DF has a significant positive effect on Growth (β = 0.01, p < 0.01), and hypothesis H4b is confirmed. Model 5-3 incorporates IEK_t−1 ∗ DF, and the interaction term between IEK and DF has a significant negative effect on financial volatility (β = −0.001, p < 0.01), and hypothesis H4c is confirmed. Model 5-4 incorporates EEK_t−1 ∗ DF, and the interaction term between EEK and DF has a significant negative effect on Volatility (β = −0.005, p< 0.01), and hypothesis H4d is confirmed.

To further show the moderating effect of DF, the moderating effect is plotted (Figure 2 and Figure 3). Figure 2 shows a simple slope diagram of the moderating effect of DF between EK and Growth, and Figure 3 shows a simple slope diagram of the moderating effect of DF between EK and Volatility. According to Figure 2, combining Model 2-1 and Model 4-1, DF has a significant strengthening effect on the relationship between IEK and Growth. Combining Model 2-2 and Model 4-2, DF has a significant strengthening effect on relationship between EEK and Growth, further verifying hypotheses H4a and H4b. According to Figure 3, combining Model 2-3 and Model 4-3, DF has a significant strengthening effect on the relationship between IEK and Volatility, and the inter-impact relationship has a significantly weaker effect. Combining Model 2-4 and Model 4-4, DF has a significant strengthening effect on the inter-impact relationship of EEK and Volatility.

4.2.4. Test for Mediating Effects with Moderation

The mediating effect is the basis of the moderated mediating effect, and the moderated mediating effect is tested on the premise that the mediating effect of OR is significant.

Table 6. Regression result analysis of moderated mediating effect.

Variable	Dependent Variable: IPE
	Model 6-1	Model 6-2	Model 6-3	Model 6-4
IEKt−1	0.01 *** (5.67)	0.02 *** (8.50)
EEKt−1			0.01 (1.60)	0.03 *** (3.11)
Growth	0.08 *** (35.62)		0.08 *** (35.70)
Volatility		−0.004 *** (−2.62)		−0.004 *** (−3.01)
DF	−0.00 (−0.95)	−0.00 (−1.03)	−0.00 (−0.89)	−0.00 (−0.89)
Growth ∗ DF	0.00 *** (5.36)		0.00 *** (5.64)
Volatility ∗ DF		0.00 (−0.52)		0.00 (−0.74)
Cage	−0.73 *** (−5.39)	−1.60 *** (−10.38)	−0.65 *** (−4.85)	−1.50 *** (−9.77)
Employee	0.18 *** (6.65)	0.33 *** (10.43)	0.19 *** (6.96)	0.35 *** (11.05)
Research	−0.11 *** (−6.78)	−0.18 *** (−10.00)	−0.10 *** (−6.41)	−0.17 *** (−9.41)
Constant	3.54 *** (4.92)	5.08 *** (6.48)	3.47 *** (4.82)	4.97 *** (6.32)
Code	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
R-squared	0.23	0.07	0.23	0.07
F-test	267.26	71.87	264.33	66.71

Note: *** indicates significance at the 1% levels.

shows the results of DF’s moderating effect test for the EK–OR–IPE model.

Combining Model 5-1, the DF and IEK interaction terms are significant and the effect of Growth on IPE in Model 6-1 is significant (β = 0.08, p < 0.01). The first half of the mediating effect IEK–Growth–IPE is moderated by DF, and hypothesis H5a is confirmed. The IEK interaction term is significant in Model 5-3, and the effect of Volatility on IPE in Model 6-2 is significant (β = −0.004, p < 0.01). The first half of the mediating effect IEK–Volatility–IPE is moderated by DF, and hypothesis H5b is confirmed.

Combining the significant DF and EEK interaction term in Model 5-2 and the significant effect of Growth on IPE in Model 6-3 (β = 0.08, p < 0.01), the first half of the mediating effect IEK–Growth–IPE is moderated by DF, and hypothesis H5c is confirmed. The EEK interaction term is significant in Model 5-4, and the effect of Volatility on IPE in Model 6-4 is significant (β = −0.004, p < 0.01). The first half of the mediating effect IEK–Growth–IPE is moderated by DF, and hypothesis H5d is confirmed.

4.3. Robustness Tests

To further ensure the reliability of the regression results, we conducted robustness tests on the main effects, mediating effects, and moderating effects.

As shown in

Table 7. Robustness test of main effects.

Variable	Dependent Variable: IPE
	Model 1-a	Model 1-b
IEKt−1	0.01 *** (8.00)
EEKt−1		0.03 *** (8.00)
Cage	−1.48 *** (−11.34)	−1.46 *** (−11.15)
Employee	0.31 *** (11.51)	0.33 *** (12.23)
Research	−0.18 *** (−12.16)	−0.17 *** (−11.40)
Code	Yes	Yes
Year	Yes	Yes
Constant	3.75 *** (22.93)	3.78 *** (3.751)
R-squared	0.08	0.08
F-test	107.50	104.52

Note: *** indicates significance at the 1% levels.

, the robustness test for the main effect is conducted using the substitution variable approach. The sum of independently filed and non-independently filed invention and utility model patents in the current period is used to denote IEK and EEK, respectively. IEK and EEK are significantly positively correlated with IPE at the 1% level in Model 1-a and Model 1-b (β = 0.01, p < 0.01; β = 0.03, p < 0.01), and hypotheses H1a and hypothesis H1b are once again confirmed.

For the robustness test of the mediating effect, we first randomly selected 70% of the sample observations, and then used the “sgmediation” command in Stata software. The results of the Sobel test are reported in

Table 8. Robustness test of mediation effects.

Variable	Dependent Variable: IPE
	Model 3-a	Model 3-b	Model 3-c	Model 3-d
IEKt−1	0.16 *** (5.51)	0.03 *** (8.06)
EEKt−1			0.03 *** (2.84)	0.05 *** (4.65)
Growth	0.78 *** (39.83)		0.08 *** (40.16)
Volatility		−0.00 ** (−1.84)		−0.00 (−2.10)
Controls	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
Code	Yes	Yes	Yes	Yes
Constant	5.50 *** (3.65)	8.09 *** (4.94)	5.09 *** (3.37)	7.54 *** (4.60)
Observations	9632	9632	9632	9632
Adj-R	0.80	0.76	0.80	0.76
Sobel Z	4.49 ***	-4.12 ***	2.50 **	-2.68 ***
Sobel Z-p	(0.000)	(0.000)	0.013)	(0.007)
Goodman-1 Z	4.46 ***	-4.09 ***	2.46 **	-2.64 ***
Goodman-1 Z-p	(0.000)	(0.000)	(0.014)	(0.008)
Goodman-2 Z	4.51 ***	-4.14 ***	2.53 ***	-2.724
Goodman-2 Z-p	(0.000)	(0.000)	(0.011)	(0.006)
Intermediary effect as a percentage	0.01	0.19	0.00	0.07

Note: *** and ** indicate significance at the 1% and 5% levels, respectively.

. The significant positive Sobel Z values of Model 3-a and Model 3-c demonstrate that abundant EK can enhance innovation performance by promoting long-term growth. The negative significant Sobel Z values of Model 3-b and Model 3-d demonstrate that abundant EK can enhance IPE by reducing financial volatility, further proving that the partial mediation effect of OR holds.

Based on the robustness of the mediation test, the extracted sample is further tested for the robustness of the moderating effect and the mediated effect being moderated. The moderating variables are further refined, and the city digital inclusive finance index is used to measure the level of digital finance development of enterprises. From the regression results in

Table 9. Robustness test of moderation effects and moderated mediating effects.

Variable	Dependent Variable: Growth		Dependent Variable: Volatility		Dependent Variable: IPE
	Model 5-a	Model 5-b	Model 5-c	Model 5-d	Model 6-a	Model 6-b	Model 6-c	Model 6-d
IEKt−1	0.00 (0.07)		−0.03 (−1.38)		0.02 *** (5.34)	0.03 *** (8.04)
EEKt−1		0.07 (1.02)		0.01 (0.06)			0.03 *** (2.88)	0.05 *** (4.63)
DF	−0.03 ** (−2.21)	−0.02 (−1.04)	−0.01 (−0.60)	−0.02 (−0.84)	0.00 (0.70)	0.00 (0.88)	0.00 (0.73)	0.00 (1.03)
IEKt−1 ∗ DF	0.00 *** (15.02)		−0.00 (−6.19)
EEKt−1 ∗ DF		0.00 *** (6.03)		−0.00 *** (−4.98)
Growth					0.08 *** (35.20)		0.08 *** (35.26)
Volatility						−0.00 (−1.83)		−0.00 ** (−2.08)
Growth ∗ DF					0.00 ** (2.36)		0.00 *** (2.70)
Volatility ∗ DF						0.00 (0.02)		0.00 (−0.08)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Code	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Constant	18.01 *** (7.54)	16.58 *** (6.85)	56.00 *** (17.98)	56.48 *** (18.10)	2.69 *** (6.11)	3.97 *** (8.24)	2.63 *** (5.98)	3.87 *** (8.01)
R-squared	0.14	0.20	0.37	0.36	0.23	0.08	0.23	0.07
F-test	113.42	92.48	396.17	392.85	186.06	53.45	184.05	49.88

Note: *** and ** indicates significance at the 1% levels.

, it can be seen that the regression results remain highly consistent with the above regression results under different robustness tests.

In summary, the regression model and empirical results of this study have adequate robustness, and the conclusions obtained have strong credibility.

To clarify the confirmation or rejection of hypotheses, we built a summary table indicating which hypotheses are accepted and which are rejected, as shown in

Table 10. Assumed acceptance/rejection summary table.

Hypothesis	Specific Content	Confirmed/Rejected
H1a	There is a positive correlation between internal enterprise knowledge and innovation performance.	Confirmed
H1b	There is a positive correlation between external enterprise knowledge and innovation performance.	Confirmed
H2a	There is a positive correlation between internal enterprise knowledge and long-term growth.	Confirmed
H2b	There is a positive correlation between external enterprise knowledge and long-term growth.	Confirmed
H2c	There is a negative correlation between internal enterprise knowledge and financial volatility.	Confirmed
H2d	There is a negative correlation between external enterprise knowledge and financial volatility.	Confirmed
H3a	Long-term growth plays a mediating role in the effect of internal enterprise knowledge and innovation performance.	Confirmed
H3b	Long-term growth plays a mediating role in the effect of external enterprise knowledge and innovation performance.	Confirmed
H3c	Financial volatility plays a mediating role in the effect of internal enterprise knowledge and innovation performance.	Confirmed
H3d	Financial volatility plays a mediating role in the effect of external enterprise knowledge and innovation performance	Confirmed
H4a	Digital finance strengthens the positive relationship between internal enterprise knowledge and long-term growth.	Confirmed
H4b	Digital finance strengthens the positive relationship between external enterprise knowledge and long-term growth.	Confirmed
H4c	Digital finance strengthens the negative relationship between internal enterprise knowledge and financial volatility.	Confirmed
H4d	Digital finance strengthens the negative relationship between external enterprise knowledge and financial volatility.	Confirmed
H5a	Digital finance can positively moderate the mediating effect of long-term growth in internal enterprise knowledge and innovation performance. That is, the mediating effect is more significant at higher levels of digital finance.	Confirmed
H5b	Digital finance can positively moderate the mediating effect of financial volatility in internal enterprise knowledge and innovation performance. That is, the mediating effect is more significant at higher levels of digital finance.	Confirmed
H5c	Digital finance is able to positively moderate the mediating effect of long-term growth in external enterprise knowledge and innovation performance. That is, the mediating effect is more significant at higher levels of digital finance.	Confirmed
H5d	Digital finance is able to positively moderate the mediating effect of financial volatility in the external enterprise knowledge and innovation performance of the firm. That is, the mediating effect is more significant at higher levels of digital finance.	Confirmed

.

4.4. Further Research

The Chinese government attaches great importance to the development of digital finance, and in December 2015, the State Council issued the Plan for Promoting the Development of Inclusive Finance (2016–2020) to make specific arrangements for the development of inclusive digital finance in China. To test the impact of DF on IPE, the impact of DF on IPE is judged by comparing the difference in IPE before and after the policy release. In order to avoid the interference of other factors, the more scientific Differences-in-Differences (DID) method is chosen to assess the impact of DF in this study. Drawing on existing studies, the following equation is constructed:

$$IP{E}_{it}={\alpha }_7+{\beta }_{21}DF\_Afte{r}_{it}+{\beta }_{22}Control+{\epsilon }_{it}$$

(7)

where $DF\_Afte{r}_{it}$ is the core explanatory variable in model (7), indicating whether firms have dummy variables supported by government policies; β₁₄ > 0 and significant, which indicates that digital finance can promote firms’ innovation performance, and the opposite inhibits innovation performance; other variables are explained as above.

Table 11. PSM-DID model regression results.

Variable	Dependent Variable: IPE
	PSM	PSM-DID	PSM	PSM-DID
	Model 7-a	Model 7-b	Model 7-c	Model 7-d
DF_Afterit	0.48 *** (3.73)	0.27 *** (2.06)	0.39 *** (3.10)	0.23 * (1.89)
Controls	No	No	Yes	Yes
Code	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
Constant	1.89 *** (21.51)	1.80 *** (21.28)	4.73 *** (20.39)	4.27 *** (19.33)
N	15,712	13,031	15,712	13,031
R-squared	0.01	0.01	0.00	0.00
F-test	22.10	20.76	16.23	15.53

Note: *** and * indicate significance at the 1% and 10% levels, respectively.

shows the regression results of model (7). Models 7-a and 7-c were regressed using the DID method, and Models 7-b and 7-d were regressed using the Propensity Score Matching Differences-in-Differences (PSM-DID) method for regression. The results of Model 7-a and Model 7-b indicate that digital finance policy significantly contributes to firm innovation when other factors are not controlled for. The regression results of Model 7-c and Model 7-d confirm that the DID regression coefficient of DF is 0.39 (p < 0.01) and the PSM-DID regression coefficient is 0.23 (p < 0.1), both of which are significantly positive, controlling for the confounding factors. In summary, regardless of the exclusion of confounding factors, for both DID and PSM-DID, the results confirm that digital finance policies have a significant positive impact on IPE, implying that the government’s issuance of digital finance support policies can effectively improve IPE.

5. Conclusions and Discussion

5.1. Conclusions

In this study, we explored the underlying mechanisms by which corporate knowledge acquired through different channels affects the sustainable innovation performance of manufacturing firms at the micro level, in the context of digital finance. To achieve this research objective, we used a data sample of Chinese manufacturing companies from 2013 to 2020 through multiple data collections and integrated traditional knowledge theory and organizational theory as the theoretical basis to support our underlying hypotheses.

In particular, we first invoked this theoretical logic to hypothesize and find that knowledge from different sources of access positively affects innovation performance; at the same time, knowledge acquired from different sources has a positive impact on innovation performance through the mediation of organizational resilience. The power of knowledge is embodied. This is because knowledge resources are a major source of competitive advantage for firms [70] and are inextricably linked to corporate innovation [71]. These results also suggest that when the environment changes, organizational resilience allows firms to generate solutions to the challenges they face [16]; the innovative activity of the company is not affected and is therefore sustainable.

Next, we linked digital finance to the relevant literature to gain insight into how digital finance affects firm innovation performance through the mediating mechanism of organizational resilience, which further confirms that in the context of digital finance, knowledge can enable manufacturing firms to be more courageous and progress towards sustainable innovation. These findings suggest that digital finance has a crucial moderating role in increasing organizational resilience. Moreover, as the level of digital finance increases, the mediating effect of organizational resilience on firm knowledge and innovation performance becomes more significant. This is because increased levels of digital finance facilitate knowledge exchange [72] and creation and generate new ideas, which are key elements of organizational resilience and enterprise innovation [73].

Finally, we examined the relationship between government digital financial support policies and firm innovation performance. This study provides empirical evidence that digital financial support policies can significantly contribute to firm innovation performance [74], supporting the important role of government regulation in firm innovation [75].

Therefore, this study has several theoretical and practical implications.

5.2. Theoretical Implications

The first theoretical implication may be that the positive relationship between enterprise knowledge and innovation performance confirms findings that all knowledge from different acquisition channels can improve innovation performance [76]. While previous investigations on how knowledge affects innovation performance have mostly focused on knowledge attributes [40] and state of knowledge [77], this study focused on knowledge acquisition channels and examined the positive effects, enriching the existing research on the impact of enterprise knowledge on innovation performance.

Second, the confirmation of the mediating role of organizational resilience has enriched the research on the mechanism of enterprise knowledge’s influence on innovation performance and verifies the importance of organizational resilience. The obtained findings prove that when studying the impact of knowledge on innovation performance, it is necessary to consider not only the absorptive capacity [78] and knowledge integration capacity [79] of firms, but also the role of organizational resilience in the relationship between the two from an organizational perspective.

Third, the existing literature points to the impact of the development of digital finance on urban innovation [80] or green development [81], which is also explored by scholars at the micro level [82], but they ignore the impact of digital finance on the innovation performance of firms. By combining the mediating role of organizational resilience and the moderating role of digital finance on the mediating role, this paper provides a plausible explanation of how digital finance plays a role in innovation from a micro perspective, based on knowledge theory and organizational theory, which is a further exploration and expansion of related theoretical studies.

Previous literature on enterprise and organizational resilience mostly focuses on using questionnaires or cross-sectional data as samples [6,83,84]. This study uses secondary data from sample firms to conduct research to avoid the subjectivity of questionnaires, while using panel data to obtain more robust conclusions.

5.3. Practical Implications

In addition to theoretical contributions, this study has certain practical insights. Firstly, enterprises need to attach importance to enterprise knowledge acquisition, not only because knowledge acquisition is conducive to integrating the existing knowledge of enterprises and improving sustainable innovation performance, but also because enterprise knowledge helps to build organizational resilience of enterprises and promote enterprise innovation. Based on this, on the one hand, enterprises should not only attach importance to internal knowledge sharing activities [85], increase resource investment, and reduce process-oriented blockage to form efficient knowledge flow, but also improve internal R&D investment [86], make full use of existing internal innovation knowledge, and convert innovation resources into output as soon as possible. On the other hand, enterprises should communicate with external stakeholders or industry–academia–research institutions to acquire external knowledge, consider high-frequency and in-depth interaction with external knowledge acquisition channels on the basis of measuring their own capabilities and costs, and establish deep relationships to obtain more valuable knowledge resources.

Second, an organization’s ability to proactively seek stability in the midst of turmoil and find success in the midst of setbacks helps companies improve their sustainable innovation performance in the face of adversity. The importance of organizational resilience in turbulent environments cannot be overstated [46]. By communicating the vision to employees to make organizational commitment, improving organizational cohesion, and forming a resilient organizational culture, top executives inspire shared achievement of the organizational vision, laying the foundation for recovery from crisis, sustainable growth, and high financial performance [87]. This also encourages the acquisition and accumulation of internal and external knowledge, especially when disruptive changes occur, which helps to strengthen the resilience of the company, save the company from a crisis, and avoid the defeat of the company.

Third, to fully understand the impact of digital finance on innovation, the government should strongly support the development of the digital economy and improve the quality of digital finance-driven innovation development. Local governments should enhance their support for digital industries, cultivate and grow digital industries along the way [88], increase the construction of information networks and new infrastructure, and realize the deep integration of digital finance with other industries. At the same time, data technology should be used to help improve the capacity of financial services, increase the degree of digital support services, optimize the allocation of digital financial resources, and reduce the costs and improve the convenience for users of financial services [89]. The government should also continue to optimize the digital financial atmosphere, create a fair and just competition environment for market participants, and give more policy support to enterprises and improve the digital resource-sharing mechanism.

5.4. Research Limitations and Future Prospects

Although this study has certain theoretical value and practical guidance for enterprises to construct knowledge, there are some shortcomings. First, considering the variables and data availability, this study was conducted based on Chinese listed manufacturing enterprises, and the universality of the research findings cannot be guaranteed. Second, the selection of mediating and moderating variables and the influence mechanism in the model construction of enterprise knowledge to sustainable innovation performance must be fully discussed, and future studies can be conducted on this basis to deepen the research on enterprise knowledge and sustainable innovation performance. Thirdly, there are too many factors affecting the innovation performance of enterprises and it is difficult to include them all in the model, resulting in a relatively low R-square value in this paper, but future studies can try to improve the expected effect of the model.