The Impact of the Industrial Innovation Ecosystem on Innovation Performance—Using the Equipment Manufacturing Industry as an Example

Abstract

The innovation ecosystem can greatly enhance enterprises’ innovation performance. However, little is known about how the industrial innovation ecosystem (IIE) improves innovation ability within the equipment manufacturing industry (EMI). The fsQCA method is utilized in this study to explore the intricate causality behind innovation performance. The conclusions are as follows: (1) There are six factors for high innovation performance, including the technological innovation subject, the knowledge innovation subject, research and development (R&D) investment, R&D personnel, the industrial internet platform, and government subsidies. None of these is a standalone prerequisite for high innovation performance. (2) Four configuration paths achieve remarkable performance. Three configuration paths achieve inefficient performance, and these have an asymmetric relationship with the above four paths. (3) Under the premise that a technological innovation network is perfect, R&D investment and industrial internet platforms both play a crucial role in innovation performance. Meanwhile, neglect in the application of industrial internet platforms and a lack of innovative subjects are important factors for low innovation performance. This study enriches the theoretical applications for innovation management within the EMI from an IIE perspective. It provides practical and management reference to promote innovative ability and enhance the manufacturing performance for China and other developing countries.

1. Introduction

Manufacturing is an important carrier to undertaking innovative scientific practice in national economic development. The equipment manufacturing industry (EMI) refers to the sector that produces varied technical equipment. It is the core component of manufacturing. The EMI provides the necessary machine tools, basic parts, and components for production for all walks of life in the industrial chain. In recent years, the global industrial chain’s work distribution has faced considerable realignments, and developed countries have implemented the “reindustrialization” strategy in succession. Some developing countries are also accelerating their planning and layout to deal with this change. According to the latest available data, the total operating income of the Chinese EMI in 2021 was CNY 4205.657 billion, and its asset scale was CNY 461,368.2 billion. However, although the economy of China’s EMI is generally developing in a positive direction, there is still a gap compared with countries that have well-developed EMIs. This is mainly due to the lack of cooperative innovation in industry–education–research–practice, the low investment in technological innovation, the weakness of endogenous innovation, the absence of key technologies in high-end equipment and pivotal components, and the serious technical barriers in industry [1]. Compared with other manufacturing industries, EMIs are characterized by strong industrial correlation, high technology (knowledge) intensity, high capital intensity, long periodicity, and a strong government-oriented role. These characteristics lead the EMI’s production and development process to rely on technology accumulation and knowledge reserve far more than other industries. To support the enduring progress of innovation, enterprises often need large financial resources to support frequent innovation activities and long R&D cycles. In order to integrate innovation resources, an efficient collaborative mechanism should be established between equipment manufacturing, universities, scientific research institutions, and governments [2]. The innovation ecosystem theory emphasizes efficient collaboration among innovation subjects, resources, and environmental factors. By exchanging and leveraging material, energy, and information, it enhances the technological innovation ability of innovation subjects [3]. This theory can resolve the current EMI development dilemma. Therefore, a perfect IIE should be built to strengthen the entirety of the EMI’s technological innovation capacity. Equipment manufacturing enterprises can only achieve long-term sustainable development by establishing and maintaining their ecological niche in IIEs and adopting their ecosystem strategies [4].

There are various research perspectives regarding an innovation ecosystem. Its research scope has been expanded to multi-level perspectives, including the macro country and regional innovation ecosystem [5,6], meso IIEs [7], and the micro enterprise innovation ecosystem [8]. The literature on IIEs mainly focuses on the topics about connotation and characteristics, structural elements, construction and evolution [9], operation mechanism [10], and evaluation and countermeasures [11]. In different industrial situations, there are different definitions about the connotation of the industrial innovation ecosystem. Gawer defined IIEs as innovation platforms that provide complementary products, services, and technologies. Innovators can develop their own supporting products or services on them [12]. Tan et al. pointed out that the rail transit equipment IIE is a supporting system for close cooperation and symbiosis among the members of the entire industrial ecological chain (including raw materials, components, parts, research institutes, vehicle enterprises, users, etc.). The key links in the entire industrial ecological chain should be strong, and the system has absolute competitiveness and control [9]. Although scholars have different understandings of the constituent elements in different IIEs, the existing literature indicates that industry–university–research–government is the main constituent element of China’s innovation ecosystem [13]. Furthermore, scholars have incorporated customers, intermediaries, financial institutions, and the environment into the analysis framework of IIEs. When the research object involves a particular industry, its constituent elements are slightly different depending on disparate industrial characteristics. For example, the research on digital industry innovation ecosystems includes digital technology, digital elements, the digital platform, etc., in its core elements [14]. From the industrial chain perspective, some scholars divide the ice and snow enterprises that constitute ice and snow IIEs into upstream, midstream, and downstream ice and snow enterprises, which are site construction enterprises, service enterprises, cultural media, and commercial enterprises [15].

The extant literature has partially offered a foundational basis for the examination of the elements that influence a company’s innovation performance. Innovation within enterprises is a continuously evolving managerial endeavor. Improvements in innovation performance mainly occur due to many conditional variables from the internal and external environment [16]. These conditional variables include but are not limited to government subsidies [17,18], organizational learning capabilities [19], knowledge management capabilities, artificial intelligence (AI) [7], and digital transformation. With the gradual improvement of the digital infrastructure in countries around the world, the digital upgrading of manufacturing has reached a strategic level. Many scholars have begun to probe how digital infrastructure can influence innovative performance. Li et al. studied the high-end EMI, analyzing how digital innovation ecosystems can propel the intelligent evolution of businesses [14]. Researchers largely argue that digital transformation facilitates innovative practices within businesses [16,20]. Conversely, other scholars hold different views. Gebauer et al. stated that investments in digital technology do not always correlate with an improvement in innovation returns [21]. More recently, some authors have implied that industrial companies may reject certain types of AI, similar to the “Not Invented Here” syndrome [22].

Numerous scholars have actively explored, from different perspectives, the manner in which IIEs affect performance. In an empirical analysis, enterprises, universities, research institutions, research and development (R&D) investment, innovative environment, open degrees, and new generations of information technology have been proven to promote innovation performance [7,23,24,25]. The authors of [17] used fsQCA to analyze samples from 253 Chinese-listed manufacturing enterprises and discussed the dynamics through which enterprise knowledge management enhances innovative outcomes within an open innovation ecosystem. Artificial intelligence also affects the creation, delivery, and capture of enterprise value, thus helping manufacturing pursue an evolution in business strategies within the industrial ecosystem [7].

The existing literature contains relatively few studies on the correlation between “IIEs” and “performance” within the EMI, and there are some research gaps. Firstly, the essence of IIEs varies significantly against different industrial backgrounds. The existing research on IIEs is too scattered and an independent research system for IIEs has not been formed in every industry. The research system for EMI innovation ecosystems is even less formed. Throughout the research conducted worldwide, there are few systematic studies on EMI innovation ecosystems. Secondly, in terms of the research context, most publications in the literature have focused on high-tech industries [26], strategic emerging industries [27], and manufacturing [17]. The geographical scope of their empirical research is mostly divided according to the high-tech zones [19] or provinces of China [27]. In the study of the manufacturing industry specifically, most scholars choose their specific sub-sectors or enterprises of a specific scale—such as small- and medium-sized manufacturing enterprises [28], high-end equipment manufacturing industries [14], and the rail transit equipment industry [9]—as the research object. Few studies research the EMI deeply. Thirdly, when exploring the role of universities in driving innovation within an enterprise, most indicators are measured by the number of universities in the region. This measurement does not exclude universities that have no involvement in the research object’s R&D, production, and application. As a result, the data sources are not adequately rigorous. Fourthly, with its powerful data integration and analysis capabilities, the industrial internet platform provides a highly connected ecosystem for innovation in the EMI. It not only accelerates the process from concept to productization but also improves R&D efficiency and product quality by monitoring and optimizing the production process in real time. In addition, the platform promotes cross-domain collaboration, reduces the threshold of innovation, and accelerates the formation of innovation networks. However, most scholars have ignored how the industrial internet platform influences EMI innovation. Only a few studies have examined the industrial internet’s particular influence. Finally, theoretically, scholars have been more interested in exploring the particular influence of an element on performance. They have ignored the synergistic effect among elements, and they have restricted the explanatory power of innovation ecosystem theory to enterprise innovation. All of these gaps make it difficult to effectively resolve the aforementioned innovation and development challenges faced by EMIs. Therefore, identifying the key innovation elements of IIEs and analyzing these elements’ dynamic impact on innovation performance has become increasingly essential.

Our research is devoted to answering the following questions: (1) what elements in the EMI innovation ecosystem can promote innovation performance? (2) Does a single element constitute the requirement of high performance? (3) What are the effective paths for realizing high innovation performance?

To explore potential answers, the fuzzy-set qualitative comparative analysis (fsQCA) is utilized. Our specific research contributions are as follows: (1) We define the connotation of the EMI innovation ecosystem. (2) We analyze the key elements of the EMI innovation ecosystem affecting innovation performance and we optimize the conceptual model. (3) We regard the industrial internet platform as a key element in the EMI innovation ecosystem and explore its necessity and synergies with other elements. (4) This study identifies alternative configurations and explains the complex causal mechanism of high-level innovation performance by fsQCA. It provides an effective approach for equipment manufacturing to improve innovation performance.

The rest of this study is structured as follows. Section 2 introduces the materials and methods in detail. Section 3 discusses the results of our empirical analysis. Section 4 summarizes the theoretical and practical implications, limitations, and future research. Section 5 presents our conclusions.

2. Materials and Methods

2.1. Theoretical Framework

The innovation ecosystem theory emphasizes the nonlinear role and synergistic interaction between innovation elements. Its ultimate goal is to achieve value co-creation and co-evolution [29,30]. Industrial innovation ecosystems (IIEs) combine ecosystem and innovative systems under an industrial scenario [31]. IIEs offer an environment in which innovation subjects can cooperate [32]. They promote technological progress and industrial development through cooperative innovation, and they create more value [33]. The innovation ecosystem theory posits that systems possess self-organizing characteristics, and at different stages of dynamic evolution, there may be different architects or leaders within the system [9]. This study posits that equipment manufacturing IIEs can be viewed as an open innovative system formed in the EMI context, where clusters of equipment manufacturing enterprises, universities, research institutions, governments, and various service organizations—stakeholders involved in innovation—engage in close cooperation and interdependence through the exchange and transmission of innovative resources, such as materials, energy, and information, within a conducive industrial technological innovation environment. The ultimate goal of the entire IIE is to effectively integrate innovation resources and achieve a sustainable development and dynamic balance. Scholars have pointed out that the main characteristics of IIEs are dynamism, diversity, co-evolution, self-organization, and stability [3]. In conjunction with the characteristics of the EMI mentioned earlier, this study posits that an EMI innovation ecosystem also possesses systematicity, regionality, complexity, openness, connectivity, and a long-cycle nature.

When deconstructing the elements of IIEs, different analytical frameworks are presented. Based on TOE theory, various academics have examined the causality relationship between multi-level antecedent conditions and different IIEs [34,35]. Using China’s 3D printing industry as an example, Xu et al. evaluated the creative capability of IIEs through the analytical framework of “science–technology–business” [36]. As an important theory studying multi-element collaborative innovation, the spiral innovation theory offers researchers a basis for studying the impact of different innovative subjects on innovation performance. The triple helix theory holds that the pivotal innovation subjects are enterprises, universities, and governments [13]. The IIE has the characteristics of dynamism and openness. It is not only necessary to realize the benign cooperation between the subjects but also to consider the dynamic interaction between the system and its surroundings. As research into the innovation ecosystem theory continued to advance, researchers started using this dichotomy (innovation subject, innovation environment) to deconstruct the elements of the innovation ecosystem. Later, researchers separated “resources” from “environment”, forming the tripartite method (innovation subject, innovation resource, innovation environment) [24,37,38]. Therefore, this study adheres to the innovation ecosystems theory and builds an analytical framework for IIEs from the dimensions of innovation subjects, resources, and environment.

• Innovation subjects

The innovation subject is composed of individuals and populations that promote and support innovation. It is a direct manifestation of industrial technology activities and innovation levels. Equipment manufacturing usually involves highly technical processes, including R&D design, engineering processes, product manufacturing processes, etc. Its production and R&D processes rely on technology accumulation and knowledge reserves far more than other industries. At present, global manufacturing is developing toward the trends of intelligence, information, and green operations. The EMI has gradually improved its degree of specialization and strengthened its high-tech attributes by merging with the latest information technology industry. We should establish an IIE in which equipment manufacturing enterprises are the technological innovation subjects. This cannot only enhance the overall innovation ability of the EMI [39] but also alleviate the impact of changes in the external environment with other participants within the ecosystem [40], thus driving a deep evolution of the entire ecosystem [41]. Some studies suggest that the innovation ecosystem consists of the knowledge ecosystem and the business ecosystem [36,42]. As the most important strategic resource of enterprises, knowledge can create a sustainable competitive advantage and improve innovation performance [43]. As a knowledge-intensive industry, the EMI places high requirements on R&D level, technical strength, and intellectual property investment. As innovation subjects that generate knowledge, universities and research institutions are essential. Colleges and universities undertake basic research and original innovation. The evolution of the EMI is subject to the quality of universities in the region [44]. As an open, sophisticated knowledge creation system, colleges and universities enhance knowledge production, exchange, and integration within the innovation ecosystem [45,46]. Collaborative innovation theory insists that the profound cooperation of higher education and industry may promote resource sharing and enhance innovative development and R&D outcomes.

In summary, this study supports the conclusion that innovation subjects within the EMI innovation ecosystem mainly include the technological innovation subject represented by equipment manufacturing enterprises and the knowledge innovation subject represented by colleges and universities. Through continuous cooperation and interaction, these innovation subjects form an innovation network and provide continuous resources for industrial innovation [23,47].

2.

Innovation resources

The necessary resources for technological innovation include experienced R&D personnel, sufficient R&D funds, and a good management structure [48]. The rational allocation of funds, personnel, and related resources invested in R&D activities is the key to effective technological innovation. Industrial entities must make rational use of resources to achieve long-term and stable processes of industrial performance. The production and R&D technology of equipment manufacturing products is complex and highly specialized. In order to produce some specific machines and equipment, enterprises usually need to invest massive funds to build plants for previous investment and purchase materials, equipment, and supporting facilities for production and R&D.

Under conditions of raging competition, enterprises must make continuous investments and formulate proper strategies in scientific and technological innovation [49]. Empirical studies have revealed that R&D investment is instrumental in advancing industrial outcomes [50]. As technology continues to improve, the industry’s own ability to accept and absorb external technological innovation has also increased. Therefore, the input of R&D funds is not only the element that affects enterprises’ technological innovation of processes and results but is also a vital index for measuring the innovation resources within the whole EMI innovation ecosystem. These funds are the basis and impetus for innovation. Under limited technical conditions, R&D personnel, as the main carrier for knowledge and technology, form a robust support system for innovative achievements. An increased number of R&D personnel are more conducive to facilitating knowledge absorption and conversion, as well as spreading technology within IIEs [23].

3.

Innovation environment

Since innovation has a social and geographic embeddedness, innovation activities must also be conducted in a particular innovation environment [51]. Enhanced innovation performance stems from the comprehensive influence of various elements in the internal and external environment [16]. Many studies have shown that a suitable innovation environment can encourage enterprises to innovate and explore actively, thus promoting the innovation ability and output of the entire innovation ecosystem [52,53].

Equipment manufacturing embodies the transformation of cutting-edge technology, and the upgrading of its products depends on advanced scientific achievements and industrial technology. In order to maintain sustainable operation, enterprises need to invest massive expenditures to support frequent innovation activities and long R&D cycles. The capital-intensive characteristics of the EMI make the risk and failure rate of its innovation activities continue to rise [54]. In order to mitigate innovation risks and uncertainties, enterprises usually need to establish links with government agencies for government resources such as government subsidies [55]. The government can effectively promote the investment of enterprises in innovation resources, whether it is through the direct provision of innovation funds or the construction of laboratories or advanced facilities. Different forms of government assistance may significantly lessen the innovative risks and costs and promote a new round of innovation investment, thus greatly improving innovation output [56]. The government can also guide EMI evolution by formulating specific policies and regulations. It requires the EMI to have an accurate pre-judgment ability and adaptability to cope with changes in the policies and regulations and to cope with the macro environment. Some studies posit that the government is an important participant in innovation ecosystems. Interestingly, there are also two arguments about how government support affects innovative performance. Numerous investigations reveal that the support of policy environments, such as government R&D subsidies and government tax reduction, promotes enterprise creativity [17,19,57]. Nevertheless, some demonstrate the opposite view. Neukam and Bollinger highlighted that R&D investments alone are insufficient, as funds can sometimes be used in ways that are detrimental to the environment without proper guidance [58]. Wang et al. pointed out that if enterprises rely on the government’s financial and tax policy support excessively, they may produce a crowding-out effect on R&D spending, leading to a lack of innovative willingness in their long-term development [59].

The digital infrastructure of innovative subjects and ecosystems accelerates the innovative process [60]. The industrial internet is not only an important part of Industry 4.0 but also an important component of new infrastructure. It can enable the digital transformation of manufacturing by reducing costs, improving efficiency, and extending the industrial chain. Meanwhile, the industrial internet has also promoted the advancement of the data services field and other industries, accelerated the data-based business ecosystem’s development [61], and provided a conceptual basis for the collaborative value creation model [62]. As an operating system, the industrial internet platform becomes the key environmental element for the manufacturing innovation ecosystem in a digital context [63]. The resource-based view insists that the industrial internet platform provides equipment manufacturing enterprises with the most important and accurate data resources. These data resources contribute to improving these enterprises’ data processing capabilities, develop their new products and services efficiently, and improve their innovation performance [64]. In this study, we argue that the industrial internet platform provides an environment for the EMI’s innovative subjects to share resources, and it accelerates the formation of IIEs. Therefore, the industrial internet platform should be incorporated into the analysis framework of EMI innovation ecosystems, and we should explore how it helps to promote performance.

This study constructs a structural model in light of the analysis on how elements of the IIE influence innovation performance (Figure 1). This structural model reflects the interactive relationships between innovation elements, intuitively demonstrating how these elements are organized and how they effectively collaborate to promote innovation.

2.2. Method

There are various research methods for exploring how disparate elements influence enterprise innovation performance. The common methods include the hypothesis testing method [65], difference-in-differences model analysis method [66], bivariate regression model analysis method [67], etc. However, scholars only discuss the individual impact of a given element on innovative performance, ignoring the interactive effects between elements. IIEs exhibit traits of both industrial and innovation ecosystems. It is a typical multi-level, complex nested system, and its influence on innovation performance generally depends on the interaction between multiple elements. The configuration theory focuses on “multiple causes and one effect”, which helps one to understand the complex problems of multiple concurrent and asymmetric causality [68,69]. Thus, fsQCA is considered to be an ideal method for exploring the influence of various elements on innovation performance within the EMI. To elaborate, the following points can be made: (1) QCA technology focuses on the “concurrent causality” across cases [70]. It has three operation methods—fuzzy set, crisp set, and multi-valued set—based on three forms of variables [71]. Crisp-set qualitative comparative analysis (csQCA) and multi-value qualitative comparative analysis (mvQCA) are suitable for dealing with binary categorical variables [72] and multi-category variables [73], respectively, while fsQCA uses a membership degree between 0 and 1 to indicate the possibility of causal conditions. FsQCA can analyze case data through Boolean algebra [74] and deal with situations involving partial membership and degree change. (2) The variables involved in this study are mostly continuous variables, and there are some problems of partial membership and degree change. Therefore, we argue that fsQCA can observe the subtle effects of variables adequately under different combinations of conditions [75]. (3) The FsQCA method effectively combines the benefits of qualitative and quantitative analysis. It is suitable for both small-sized sample studies of less than 10 or 15 and medium-sized sample studies of 10 or 15 to 50 [71]. This study takes the Chinese listed companies of the EMI in thirty provincial-level administrative regions (provinces) as samples, which is a medium sample, and the sample size matches the research method. (4) FsQCA not only precisely identifies each explanation case in a different configuration, revealing latent regional disparities by the selection of innovative strategies, but also contrasts the asymmetrical dependent variables of high and low innovation performance across regions [24,71]. This approach contributes to a deeper and broader knowledge about the complexities inherent in industrial innovation.

From the perspective of configuration, the influence of innovation subjects, resources, and environment on IIEs’ innovation performance is not independent. The linkage and matching between the influencing elements and innovative performance may also have a nonlinear effect. The key elements of the above IIEs, namely, the technological innovation subject, the knowledge innovation subject, R&D investment, R&D personnel, industrial internet platform, and government subsidies, are taken to be antecedent variables, while innovation performance is acknowledged as an outcome variable. The fsQCA method is used to verify whether a single antecedent variate becomes the essential condition of the response variable and to explore the synergy mechanism between IIEs and equipment manufacturing enterprises and the interaction among innovation elements.

2.3. Samples and Data

As an essential part of manufacturing, the listed companies of the EMI directly reflect manufacturing’s value creation power. The research samples are selected from the Chinese A-share main board equipment manufacturing quoted companies in Shanghai and Shenzhen. According to the classification of the newly revised “2012 Industry Classification Guidelines for Listed Companies” and referring to the “National Economic Industry Classification” (GB/T4754-2017) [76], the EMI belongs to the C category in the industrial category, which is composed of eight sub-sectors: C33–C40. In this study, the samples are processed in the following manner: (1) when selecting sample data, this study only selects companies in the C33–C40 category, and the quoted companies of the EMI designated by ST, * ST, and PT of the China Securities Regulatory Commission are excluded; (2) excluding companies that were not listed or had a mass of missing data in a certain year during the sample period; (3) when individual companies in the sample data lack an indicator, meaning that the company’s indicator value is small or 0, the variable is assigned a value of 0. Finally, a total of 789 eligible enterprises, belonging to 30 provinces across China, were selected. Since fsQCA is adept at analyzing medium-sized samples, during empirical analysis, we measured relevant indicators of the EMI within the boundaries of provinces. This approach facilitates the adaptation of innovation strategies by each province in China according to their local conditions and specific development circumstances. Most of the data comes from the China Stock Market Accounting Research (CSMAR) database and the China Statistical Yearbook on Science and Technology. The industrial internet platform’s application degree is assigned using “crawler keyword + manual confirmation” technology, and the relevant enterprise’s annual report data are from the official websites of the Shenzhen and Shanghai Stock Exchange.

Since the impact of the construction of IIEs on technological innovation within the EMI exhibits a noticeable temporal delay, this study determined the lag period to be two years [77]. To eliminate interference from random disturbances, the average value of three years’ data of each case was used for characterization [24]. In light of data accessibility, our study selected the average value of each antecedent condition from 2018 to 2020 and the average value of innovation performance from 2020 to 2022 for analysis.

Table 1. Declaration of variables.

Abbreviation	Variable	Description	Source	Selection Basis
Outcome
INNO	Innovation performance	Number of enterprises’ patent applications	CSMAR database	[78]
Antecedents
TIS	Technological innovation subject	Number of equipment manufacturing enterprises	CSMAR database	[24,49]
KIS	Knowledge innovation subject	Number of universities associated with enterprises innovation	List of Chinese universities and official website information	[24,49]
R&DI	R&D investment	The R&D investment within equipment manufacturing enterprises	CSMAR database	[50]
R&DP	R&D personnel	The R&D personnel within equipment manufacturing enterprises	CSMAR database	[23,24]
IIP	Industrial internet platform	The application degree of industrial internet platform	Annual reports of companies on the official websites of the Shenzhen Stock Exchange and the Shanghai Stock Exchange	[79,80]
GS	Government subsidy	The government subsidies for equipment manufacturing enterprises	CSMAR database	[55,65]

displays the selection, description, data source, and selection basis of the outcome and antecedent variable.

2.4. Measurement

Regional innovation ecosystems focus on innovation activities within a geographical area, encompassing the interactions and collaborations among different industries, organizations, and institutions within the region [27]. However, IIEs emphasize innovation activities and interactions within a specific industry, highlighting the connections between technology, market, and organization within the industry. China’s provinces vary significantly in their spatial distribution, economic status, and industrial structure, which makes the selected provinces better able to represent the characteristics of the EMI innovation ecosystem in different provinces of China.

Outcome variable. The explained variable is the innovation performance of the EMI. The existing research mostly uses the number of patent authorizations [81], applications [78], or new product sales [82] to measure innovation performance. However, new product salesrooms need to be commercialized to generate economic benefits before they can more reasonably measure enterprises’ innovation performance. Moreover, enterprise data cannot be collected easily, which is not considered in this study. There are many uncertainties in the process of patent authorization, and the patent application count can accurately reflect the enterprise’s innovative outputs. In consequence, this study selects patent count, which is the most common and easy to obtain, to measure innovation performance. After performing a sample selection on the CSMAR database, this study obtained the essential data from 789 listed enterprises, including enterprise names, provinces, etc. Then, their yearly number of patent applications from 2020 to 2022 was retrieved, and the average value for the three years was calculated. Finally, through the province screening function, the annual number of patent applications in the provinces where the 789 listed enterprises were located was obtained. Given that some provinces may have no patents and that the patent data are characterized by right-skewed distribution, it was processed by adding 1 and taking the logarithm [65].

Antecedent variables. The specific measurement of the antecedent variables in this study was carried out as follows.

2.4.1. Innovation Subjects

Equipment manufacturing enterprises are used to measure the technological innovation subject, and universities located in the province of equipment manufacturing enterprises are used to measure the knowledge innovation subject [49]. In the same way, this study used the total of the three-year averages of listed equipment manufacturing enterprises in each province. Universities are the main sources for creating knowledge. The proximity between universities and enterprises significantly influences the absorption and transfer of knowledge for industry and the enterprises’ innovation activities [83]. For actual measurement, this study conducted the following screening: (1) Considering that engineering has a direct association with the EMI and that fundamental research in science can provide a theoretical basis for the EMI to develop new materials, processes, and technologies, colleges and universities that do not have engineering and science majors, such as medicine, linguistics, law, and arts, were excluded. (2) According to the sample enterprises’ annual reports, the universities’ and enterprises’ official websites, media news, and other public information, we screened out the colleges and universities that are related to EMI innovation.

2.4.2. Innovation Resources

According to the theoretical framework, this study directly selects “R&D investment” and “R&D personnel” data from the financial statements of sample companies in the CSMAR database to measure the “R&D investment” and “R&D personnel” indicators. Similarly, the average values of three years of data are used and summed up according to the provinces where they are located.

2.4.3. Innovation Environment

Government subsidies can most intuitively reflect the quality of the innovation environment (as good or bad) provided by the government for enterprises. Consistent with the previous text, this study selected “government subsidies” from the financial statements published by sample enterprises in the CSMAR database and calculated the average value over three years [65]. At present, there are few quantitative studies on enterprises’ application of an industrial internet platform. Wu et al. used “crawler keyword + manual confirmation” technology to determine whether an enterprise is embedded in the industrial internet and used a “0–1” dummy variable to assign values [79]. This technical processing approach is incapable of adequately demonstrating enterprises’ “application degree” of an industrial internet platform. This study should analyze the index data of the case for 3 years, which means it cannot use the average value of the virtual value. As an important manner, whereby digital transformation occurs in an enterprise and as a major strategy for development, such characteristic information about the application of an industrial internet platform will be largely reflected in the enterprises’ annual reports. The vocabulary in the annual reports could reflect the enterprises’ future strategical directions and development paths. Thus, the text analysis method is used to summarize and sort out the characteristic words concerning the industrial internet platform (see

Table 2. Feature words of “industrial internet platform”.

Feature Words

Industrial internet platform

Internet, industry internet, intelligent manufacturing, cloud platform, mobile internet, mobile interconnection, industrial internet, Internet+, industrial 4.0, industrial cloud, informatization, information technology, information communication technology, cloud service, internet of things, blockchain, digital twin, edge computing, big data analysis, intelligent sensing, cloud collaboration, IIoT, CPS, 5G, IPv6, TSN, SDN, CAX, ERP, MES, and SCM

for details). In addition, this study refers to the works [79,80] and draws lessons from industrial internet-related policies and their connotations. Finally, using statistics regarding the frequency of the aforementioned feature words, the application degree of the industrial internet platform is objectively quantified. The detailed procedure is outlined below: (1) the crawler function of Python was used to crawl the annual reports of 789 companies selected above and extract the text. (2) this study used Python to segment all samples and perform word frequency statistics on the feature words in Table 2. (3) Some feature words, which did not meet the statistical conditions according to a judgment of the position and content of the feature words in the annual report, were removed. This study eliminated those words involved in the companies’ prospect and planning such as “commit to building industrial internet”, as well as those found in the introduction of the industry background, development trend, and government documents. (4) We standardized the word frequency data and used the entropy method to determine the weight of each index and obtain the final application degree index of the platform.

2.5. Data Calibration

Calibration is the process of assigning a set membership score to each case [84]. Referring to previous research, the three anchor points of full membership, crossing points, and full non-membership of all variables in this study were set as 75%, 50%, and 25%, respectively [14,17]. Then, all the variable data were calibrated to membership values within the fuzzy set ranging from 0 to 1 [70]. We calibrated the result variables and the antecedent variables (see

Table 3. Calibration anchors.

Variables	Full Membership	Crossover Point	Full Non-Membership
INNO	22.209	10.411	2.348
TIS	24.500	11.500	4.500
KIS	106.750	74.000	54.750
R&DI	900,374.956	427,827.538	106,880.292
R&DP	23,263.584	8144.667	3023.833
IIP	25.146	11.884	4.607
GS	17,954.479	2715.391	567.338

for details) and generated a truth table.

3. Results

3.1. Necessity Analysis

Before the conditional configuration analysis was carried out, the necessity for each antecedent condition was first tested. The necessary conditions of high- and non-high-level innovation performance were analyzed by fsQCA 4.1 (

Table 4. Necessity analysis.

Condition	High-Level Innovation Performance		Non-High-Level Innovation Performance
	Consistency	Coverage	Consistency	Coverage
TIS	0.877406	0.817397	0.26675	0.282869
~TIS	0.230221	0.216198	0.827802	0.884873
KIS	0.799002	0.755391	0.316218	0.340297
~KIS	0.302209	0.279683	0.772699	0.813984
R&DI	0.874554	0.838115	0.266124	0.290301
~R&DI	0.259444	0.236979	0.851597	0.885417
R&DP	0.878831	0.849173	0.243582	0.267906
~R&DP	0.242338	0.219638	0.862868	0.890181
IIP	0.876693	0.817276	0.266124	0.282392
~IIP	0.230221	0.216054	0.827802	0.884281
GS	0.797577	0.788028	0.296806	0.333803
~GS	0.325731	0.289241	0.811522	0.820253

). Most research measures necessary conditions according to consistency [74]. The condition is a necessary condition for response variables when a conditional variable’s consistency reaches 0.9 [85]. Therefore, the six antecedent conditions proposed here are not the necessary conditions for high innovative performance. Meanwhile, the absence of any conditions is not the necessary condition for non-high performance. This indicates that it is complicated to explain how IIEs impact innovation performance within the EMI. The innovative subjects, environment, and resources need to be coordinated and matched with each other to influence innovation performance within the EMI.

3.2. Configuration Analysis

To explore which mixtures of these six antecedent conditions constitute a sufficient explanation for high innovation performance, the calibrated data were input into fsQCA 4.1 for processing. Firstly, according to the existing literature, the consistency threshold of the configuration path is established at 0.8 [74]. The PRI is an alternative method for consistency in subset relationships. It is an important criterion used to evaluate and filter truth table rows in fsQCA to assure the quality and accuracy of the analysis results [71,75]; thus, we established 0.7 as the PRI [86]. Considering that the data of equipment manufacturing enterprises after screening are divided into 30 provinces, in order to retain more than 75% of the cases, the case frequency threshold was established as one [71]. Secondly, there are two qualitative implications of “TIS×R&DI” and “R&DI×IIP” in the standardized analysis of high innovation performance. The elimination of qualitative implication requires researchers to judge the importance of each choice by experience [87]. According to the innovation ecosystem theory and the actual situation for China’s EMI, “Select All” was chosen. Then, given that a unified understanding of the connection between all conditions and results is still pending and that there is a dearth of explicit theoretical expectations when conducting counterfactual analysis, “present or absent” was selected [24]. Finally, the solutions with different simplification degrees are obtained through a Boolean minimization operation. Conditions present in both the intermediate and reduced solutions are referred to as core conditions, and configurations sharing these core conditions are categorized as a class of solutions. Conditions appearing only in intermediate solutions are named auxiliary conditions [33,35].

Table 5. Antecedent configuration of high-level innovation performance.

Antecedents	High-Level Innovation Performance				Non-High-Level Innovation Performance
	H1a	H1b	H2a	H2b	N1	N2	N3
TIS	⬤	●	⬤	●	●	⊗	⊗
KIS	●	●			●	⊗	⊗
R&DI	⬤	⬤	⬤	⬤	⨂	⨂	●
R&DP	●	●	●	●	⊗	⨂
IIP	●	⬤	●	⬤	●	⊗	⊗
GS			●	●	⊗		⨂
Consistency	0.94	0.94	0.941	0.941	0.995	0.933	0.995
Raw coverage	0.696	0.696	0.728	0.728	0.126	0.649	0.135
Unique coverage	0.097	0.097	0.129	0.129	0.061	0.532	0.037
Overall consistency	0.938				0.942
Overall coverage	0.825				0.747

Note(s): (1) the larger ⬤ and ⊗ indicate the existence and absence of core conditions, respectively; the smaller ● and ⨂ indicate the existence and absence of auxiliary conditions, respectively; (2) spaces without any indication show that the conditions can be either present or omitted.

shows the final configuration results.

The results show that four configuration paths can attain high performance. The consistency of the single and overall solution in these four configuration paths are 0.940, 0.940, 0.941, 0.941, and 0.938, respectively. All the consistencies are greater than the threshold of 0.8. This shows the four paths are all sufficient conditions for high innovative performance and generally constitute sufficient conditions for high innovative performance within the EMI, and 93.8% of the EMIs demonstrate high innovation performance. The original coverage rates of the single solution are 0.696, 0.696, 0.728, and 0.728, respectively. The overall solution’s coverage rate stands at 0.825, which means that every configuration accounts for a considerable percentage of positive innovation results, and the combination of four configurations accounts for about 82.5% of the cases. Meanwhile, three configuration paths may generate low innovative performance, and the consistency of the single and overall solution reaches 0.995, 0.933, 0.995, and 0.942, respectively. This indicates that these three paths constitute sufficient conditions for non-high innovative performance.

Next, the four paths for high innovative performance are analyzed vertically. In the configuration H1a (TIS×KIS×R&DI×R&DP×IIP), the technological innovation subject and R&D investment become the core conditions. The knowledge innovation subject, R&D personnel, and industrial internet platform are the marginal conditions. It is shown that with the agglomeration of enterprises in the EMI innovation chain and the sufficient investment of R&D funds, enterprises can achieve better technological achievements if they utilize industrial internet platform applications and carry out joint R&D with surrounding universities. In the configuration H1b (TIS×KIS×R&DI×R&DP×IIP), the R&D investment and industrial internet platform become core conditions. The innovation subjects and R&D personnel play an auxiliary role. This shows that even if the innovation network of the EMI and university is not perfect, enterprises can make up for the limited R&D ability of personnel through increasing R&D investment and leveraging the industrial internet platform for resource sharing. In the configuration H2a (TIS×R&DI×R&DP×IIP×GS), the technological innovation subject and R&D investment become the core conditions. The R&D personnel and environment constitute the marginal conditions. This shows that under the empowerment of a strong innovation environment, the equipment manufacturing enterprise cluster in IIEs can realize high innovation performance by providing adequate provision for its innovation resources. In the configuration H2b (TIS×R&DI×R&DP×IIP×GS), the R&D investment and industrial internet platform are the core conditions. The technological innovation subject, R&D personnel, and the government subsidy are the marginal conditions. This shows that under the premise of ensuring sufficient supply for innovative resources, equipment manufacturing enterprises can realize better innovation performance by actively applying the industrial internet platform to strengthen resource sharing with other enterprises in the innovation chain.

According to the number of intermediate solutions, these four configurations can also be summarized into two paths: H1 (TIS×KIS×R&DI×R&DP×IIP) and H2 (TIS×R&DI×R&DP×IIP×GS). Configurations such as H1a and H1b demonstrate the characteristics of innovation performance driven by innovation subjects, resources, and industrial internet platforms, and this is referred to as the subject–resource–platform-driven type. Configuration H2a and H2b are characterized by the synergy of the technological innovation subject, resources, and environment, and this is referred to as the enterprise–resource–environment-driven type.

Furthermore, this study analyzes the three configurations of non-high innovation performance. The configuration N1 (TIS×KIS×~RDI×~RDP×IIP×~GS) shows that the EMI ecosystem can neither lack government subsidies nor neglect the introduction of R&D talents. Otherwise, even if the enterprises and universities cooperate closely, and the industrial internet platform is appropriately used, it will not be able to effectively transform R&D resources into R&D results. The only typical example of this path is Liaoning Province. Configuration N2 (~TIS×~KIS×~RDI×~RDP×~IIP) indicates that if there are fewer enterprises participating in innovation in the EMI innovation ecosystem, if they fail to establish R&D cooperation with universities effectively, and if the effective synergy of an industrial internet platform is ignored, then the system’s innovation resources will not be fully utilized. This results in low innovation performance. This configuration covers the western region of China, including Guizhou, Yunnan, Shanxi, Gansu, Qinghai, Ningxia, etc., and the northeast region of China, including Jilin and Heilongjiang. The configuration N3 (~TIS×~KIS×RDI×~IIP×~GS) indicates that the low density of innovation subjects and neglect of the industrial internet platform will result in a limited innovation input that cannot be translated into innovation results. The cases covered by this configuration include Chongqing and Xinjiang.

Comparing the seven configuration paths that affect the innovation performance of the EMI, it was found that the configuration for low innovation performance is not entirely the opposite of the high-performance configuration. It indicates that the antecedent conditions that drive enterprises’ high and low innovative performance are asymmetric.

3.3. Analysis of Explanation Cases

FsQCA is effective in accurately situating cases covered by each equivalent configuration. By comparing the cases covered by high performance paths in the intermediate solution, it was found that the evolution of China’s EMI innovation ecosystem has clear regional differences in configuration paths (Figure 2).

Figure 2 shows that the explanation cases of configuration H1 (the subject–resource–platform-driven type) are primarily focused within the Yangtze River Economic Belt, covering Jiangsu, Zhejiang, Anhui, Sichuan, Hubei, as well as cities in developed areas of China, such as Beijing, Fujian, Shandong, and Guangdong. Taking Guangdong Province as a typical case, its Pearl River Delta city cluster is one of the most dynamic economic zones in the Asia–Pacific region. It has an advanced manufacturing base. It also has a complete range of industrial categories and has the largest number of listed companies in the EMI and industry-related universities in China (143 and 142, respectively, according to sample data from this study). The complete cluster of the technological innovation subjects provides an effective basic condition for innovation in the EMI. Over the past few years, Guangdong Province has primarily insisted on pursuing intelligent manufacturing and has promoted high-end, intelligent, green, and international development of the EMI. The “GuangShenFoGuan” (Guangzhou, Shenzhen, Foshan, and Dongguan) intelligent equipment cluster has become one of the nation’s advanced manufacturing clusters. To promote technological innovation and achieve transformation within the EMI, the Chinese government and Guangdong have adopted numerous political measures, including tax incentives, innovation subsidies, talent recruitment, etc. At the same time, increasing attention has been paid to the construction of an innovation platform. In a list of the first batch of demonstration projects for industrial internet pilot demonstrations (issued in 2018—the sampling period of this study), nine enterprises in Guangdong Province were selected, accounting for 12.5%. Subsequently, three and four pilot enterprises were added in 2019 and 2020, respectively. The industrial internet platform has been a significant starting point for enterprises’ digital transformation. It provides a good platform for resource sharing among innovation subjects within the industry chain and accelerates the process to build equipment manufacturing IIEs.

The explanatory cases of configuration H2 (enterprise–resource–environment-driven type) include most provinces in developed regions of China, such as Beijing, Shanghai, Jiangsu, Zhejiang, Shandong, Guangdong, as well as some underdeveloped regions, covering Henan, Hebei, Anhui, Shanxi, etc. Using Beijing as an example, as the capital of China, it has strong comprehensive strength. It is also one of the Chinese cities that is a manufacturing center and possesses large-scale and rich resources. For the past few years, the technological innovation capability and comprehensive industrial strength of Beijing’s intelligent equipment industry have been significantly enhanced. Beijing has a number of outstanding enterprises that have mastered international cutting-edge core technologies and advanced processes. Many equipment backbone enterprises have formed, especially in the field of manufacturing intelligent equipment, such as industrial robots, high-end CNC machine tools, and transportation equipment. The high concentration of technological innovation subjects and the unique innovation environment lead these equipment manufacturing enterprises to bravely face the risks of innovation. They strengthen the investment of innovative resource elements such as R&D personnel and inputs, thus forming a perfect innovation ecology. The mean annual growth rate for the output value of Beijing’s high-end EMI reached 17% from 2016 to 2020. In 2020 alone, the output value of Beijing’s high-end EMI was CNY 78.84 billion. These data confirm that Beijing, as a typical case of configuration H2, has stepped out of the enterprise–resource–environment-driven road and achieved high innovation performance.

By comparing the above high and non-high innovation performance configuration paths and explanation cases, it can be found that there are obvious gaps in resource endowment among various regions in China. The innovation environment positively influences the innovation capability for developed regions and key economic zones. However, it has not yet contributed to enhancing the innovative capability of the central and western regions.

3.4. Robustness Test

There are three common methods to test the robustness of fsQCA, namely, changing the case frequency, increasing the consistency threshold, and adjusting the anchor point of calibration data [16]. The truth table was re-edited, and the consistency threshold was raised from 0.80 to 0.85 [19,88]. The overall consistency, coverage, and configuration path results remained unchanged, which shows that our conclusions were robust.

4. Discussion

4.1. Theoretical Implications

Combining the industrial basis of equipment manufacturing and the theory of innovation ecosystems, this study provides numerous theoretical contributions. First, against different industrial backgrounds, the nature of IIEs is significantly different. This study defined the connotation and characteristics of equipment manufacturing IIEs. It offset the insufficiency of further research on IIE in the specific EMI. Second, in previous studies, scholars were keen to discuss a single element’s net effect on innovation performance. This study probed the effect of synergy among innovation elements, identified six key elements that affect performance, and optimized the conceptual model of the EMI innovation ecosystem. In particular, due to the special industrial background of the EMI, this study innovatively incorporated the industrial internet platform into the conceptual model and added a discussion of the synergistic effect of the industrial internet platform and other elements on innovation performance. Finally, the fsQCA method was applied to research IIEs. This enhanced the explanation of the innovation ecosystem theory in industrial innovation research. Using the configuration of “the same destination by different ways”, this study empirically demonstrated the matching linkage effect and synergy mechanism of innovation elements in improving innovation performance under the EMI context. It revealed that the “black box” of “subject–resource–environment” affects EMI innovation performance. Meanwhile, this study also responds to the appeal from Chen [89], Hong [3], Huang et al. [90], and other scholars regarding “the application research of innovation ecosystem in different industrial situations”, “the establishment of an independent innovation ecosystem research system for each industry”, and “the introduction of more industrial situations into the research of IIEs”.

4.2. Practical Implications

The research results provide an alternative path for the EMI innovation ecosystem to achieve high innovation performance. Meanwhile, it also puts forward guiding countermeasures and management enlightenment for China and other developing countries, which can help them build a perfect IIE and enhance the innovation ability and international competitiveness of the EMI.

First of all, we should pay attention to the configuration effect and the leadership positions of key elements. The linkage matching model for innovation subject, resources, and environment demonstrates that the manner whereby the mechanism of high innovation performance in the EMI is driven by IIEs is very complex. The seven configuration paths above show that the resource endowments of each equipment manufacturing enterprise are not exactly the same, and the development level of its IIEs is also uneven. This is mainly because China’s provinces have great differences in their geographical locations, economic development levels, and industrial structures [91]. Therefore, equipment manufacturing enterprises should first identify their own ecological niche in IIEs and select innovation strategies suitable for their own development. By using the configuration results of “high innovation performance”, enterprises can achieve differentiated improvement paths for their innovation performance by making reasonable use of innovation resources and innovation environment. Meanwhile, enterprises should focus on the configuration path of “non-high innovation performance” to avoid the result of high input and low output due to incorrect innovation decisions.

Secondly, we must focus on the positive effect of the application of the industrial internet platform on the EMI’s innovative performance. Industrial platforms can guide the trajectory of technological innovation and stimulate complementary innovation [92]. The application of industrial internet platforms can significantly improve enterprises’ innovation performance [64]. According to the configuration results, the perfect innovation subject network is the basic carrier to drive innovation performance. The industrial internet platform promotes data interconnection within the platform and accelerates knowledge sharing and knowledge innovation. The government needs to place a high priority on developing the industrial internet platform. To urge equipment manufacturing enterprises to speed up their process of applying this platform, the government can issue relevant policy documents. In this way, it can motivate the multilateral co-creation of the platform and innovation subjects and promote the efficient operation of IIEs. Therefore, it is necessary to leverage digital means and build an innovation ecosystem by applying the industrial internet platform, enhancing industrial innovation capabilities, and achieving value co-creation.

Finally, equipment manufacturing enterprises must take note of the sufficient supply of R&D investment and scientific training of R&D personnel. Enterprises should introduce their own R&D personnel based on their actual needs and establish a multi-level R&D personnel training system. Enterprises can achieve talent information sharing by actively establishing contact with universities and R&D institutions. At this stage, the government is still the main provider of innovation activities and innovation resources. In addition to broadening financing channels, enterprises can actively establish contact with the government to obtain government resources. The government should supply government subsidies reasonably to provide a good innovation environment.

4.3. Limitations and Future Research

This study has not escaped some limitations. Firstly, most of the data in this study are derived from the CSMAR database, official websites, and annual reports of listed companies. Although these data meet the requirements of multi-source research data, there may be limitations in the timeliness and accuracy of second-hand data due to the different research methods of various institutions. In addition, when obtaining the data regarding the degree of application of the industrial internet, this study applied the crawler function in Python 3.8 to crawl the annual report data of the case enterprises from the official websites of the Shenzhen and Shanghai Stock Exchange. However, with the evolution of the industrial internet policy and platform technology, this study cannot fully elaborate on the characteristic words concerning the application of the industrial internet platform. Finally, since the EMI is only a part of the manufacturing industry, there are few research data and statistical data on the IIEs in the special industrial context. Moreover, our samples are all listed companies, which may limit our explanatory power for small- and medium-sized equipment manufacturing enterprises because they are characterized by significant differences in innovation subjects, resources, and environment. For improving the research results’ universality in the future, case data from small- and medium-sized equipment manufacturing enterprises should be collected. Under new economic conditions, influencing factors for innovative performance will be more complicated. In the future, IIEs can be analyzed from more perspectives and levels, such as through the use of digital applications. In order to further refine the relevant influencing factors and obtain more data, we can consider using a questionnaire to delve deeply into equipment manufacturing enterprises.

5. Conclusions

Using EMI as the empirical background, this study answers the question of how IIEs affect innovation performance. First, this study defined equipment manufacturing IIEs’ concept and connotations and constructed a system structure model. According to the innovation ecosystem theory, this study analyzed the key elements of IIEs and identified the influencing factors using the tripartite method (subject–resource–environment). Then, taking the EMI’s listed companies in China’s 30 provinces as samples, this study applied fsQCA to reveal how the synergy mechanism of IIEs drives EMI innovation performance and demonstrated the linkage effect between innovation elements. By combining typical sample cases of fsQCA, this study analyzed the theoretical logic behind different configuration paths. It responds positively to the appeal for the combination of configuration ideas and fsQCA methods to explain and verify nonlinear relationship problems in practice [69].

In this study, six key elements were found to affect innovation performance in the EMI innovation ecosystem: technological innovation subject, knowledge innovation subject, R&D investment, R&D personnel, industrial internet platform, and government subsidy. However, these six antecedent conditions cannot be used independently as a necessary condition to improve performance. High innovation performance is due to the multi-element synergy of the entire EMI innovation ecosystem. The effective combination of various elements improves the innovation performance by “reaching the same goal by different means”. Due to the different resource endowment conditions of different IIEs, four driving paths promote high performance, which can be further condensed into the subject–resource–platform and enterprises–resource–environment-driven types. Three configuration paths for inefficient performance demonstrate an asymmetric relationship with the above four paths. According to the seven configuration paths, it was found that R&D investment and the industrial internet platform influence innovation performance significantly if the technological innovation subjects’ network is perfect. Furthermore, neglect of the industrial internet platform and a lack of innovation subjects are important factors for non-high innovation performance. The research findings answer the scientific questions and fill some of the research gaps mentioned above. These findings also provide a more detailed understanding of how innovation elements influence innovation performance and expand the study of the EMI’s innovation management from the perspective of IIEs. This study explains the theoretical implications of these findings regarding innovation. Finally, it also provides practical enlightenment for management in China and other developing countries to improve innovation performance in manufacturing.